University of Southern California Dissertations and Theses

Biological impact of fluoroquinolone resistance in Pseudomonas aeruginosa

(USC Thesis Other)

Biological impact of fluoroquinolone resistance in Pseudomonas aeruginosa

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Biological Impact of Fluoroquinolone
Resistance in Pseudomonas aeruginosa

Melissa Agnello

A Dissertation
Presented to the Faculty of The University of Southern California
Graduate School
In Partial Fulfillment of Requirements for the Degree
Doctor of Philosophy
August 2015

Clinical and Experimental Therapeutics

Advisor: Annie Wong-Beringer, Pharm.D.
Co-Advisor: Steven Finkel, Ph.D

i

Acknowledgements
The incredible amount of effort and work necessary for the completion of this
dissertation project could not have been completed without the support of many people.
First and foremost, I would like to thank my advisor, Dr. Wong-Beringer, for her years of
support and encouragement and for providing me with this exciting, clinically relevant project. I
am extremely fortunate to have her as my PhD mentor and advisor. She has made me into the
scientist I am today, and for that I will be eternally grateful.
I would like to thank my co-advisor, Dr. Steven Finkel, for his support and advice. He
never hesitated to share his expertise as well as his sense of humor, and for that I am very
thankful. I want to thank Dr. Paul Beringer, for providing guidance and advice as part of my
dissertation committee but also throughout my PhD. I owe thanks to Dr. Kathleen Rodgers for
her support as part of my qualifying exam committee as well as my dissertation committee. In
addition, I want to thank other School of Pharmacy faculty and staff that have supported me,
offered guidance, and made my time here enjoyable: Dr. Roger Duncan, Dr. Sarah Hamm-
Alvarez, Dr. Stan Louie, Dr. Ron Alkana, Dr. Steven Chen, Wade Thompson-Harper, and Rosie
Soltero.
I would like to thank the SC-CTSI for providing financial support and allowing me the
opportunity to participate in the TL1 program, which provided me with invaluable learning
experiences. Specifically, I would like to thank Dr. Cecilia Patino-Sutton for her guidance and
support.
The members of my lab, specifically Tim Bensman and Jason Yamaki, have provided
immeasurable amounts of support, laughter, advice, and friendship. In addition, I could not have

ii

finished this project without the assistance of Kristy Trinh, as well as numerous volunteers and
students, especially Nicole Schrad.
My friends, especially Anna Naito and Roslynn Stone, have provided amazing social
support.
My grandma, Lucille Albanese, supports me in all my endeavors and for that I am
extremely fortunate.
Most importantly, I need to thank my parents, Michael and Valerie Agnello, for
providing all of the opportunities I have been fortunate enough to experience in my life. My
success is solely due to their influence and their selflessness, and this dissertation is dedicated to
them.
Finally, no words can express how grateful I am to my husband, Zachary Ross, for
putting up with the many times when my work had to come before spending time with him. His
encouragement, calm demeanor, and unconditional love have been essential to my happiness and
success.

iii

TABLE OF CONTENTS
Chapter Page
Abstract…………………………………………………………………………………………iv
List of Tables and Figures………………………………………………………………………vii
1: Introduction and Background…………………………………………………………………1
2: ExoU Correlates with Increasing FQ-Resistance and Virulence in
Clinical Isolates of Pa..................................................................................................................22
3: Effects of FQ-Resistance on the Fitness of exoU and exoS Clinical Strains………………...47
4: Effects of FQ-Resistance on the Fitness and Virulence in Host Infection Models………….89
5: Summary and Future Directions……………………………………………………………..100
References……………………………………………………………………………………....110

iv

Abstract
Pseudomonas aeruginosa (Pa) is an important Gram-negative pathogen in the hospital
setting. Pa causes severe disease in susceptible patients, such as those in the intensive care unit,
and is the leading cause of nosocomial pneumonia. Unfortunately, the high prevalence of
antibiotic resistance in clinical strains has made treatment difficult. Specifically, resistance to the
fluoroquinolone (FQ) antibiotics has increased rapidly over the last few decades, and it is
estimated that 30% of all clinical strains of Pa are now FQ resistant.
Pa has the ability to cause severe disease in patients, due to its large arsenal of virulence
factors. Specifically during acute infections, Pa utilizes the type III secretion system (TTSS) to
inject toxins into host immune cells. The toxins (ExoU, ExoS, ExoT, ExoY) disrupt host cell
function, leading to cell death, and allowing for immune invasion and the establishment of
infection. Interestingly, the genes encoding the toxins ExoU and ExoS are mutually exclusive in
most strains. exoS-containing strains are more prevalent in the clinical population. However,
exoU strains are more toxic, leading to worse disease and higher mortality in patients and animal
models.
Previous clinical studies have shown that infection with FQ-resistant strains leads to
increased mortality in patients, suggesting that FQ-resistance is correlated to increased disease-
causing ability. Subsequent studies revealed that ExoU strains are more likely to be FQ-resistant
than their less-virulent, ExoS-producing counterparts. We have also confirmed this observation
in a study of a large number of clinical strains. This represents a significant clinical problem; the
strains of Pa that are capable of causing the most severe disease are most likely to be resistant to
the FQs.

v

The overall goal of this dissertation project is to investigate the biological basis for the
correlation of the highly pathogenic ExoU-producing strains with FQ-resistance. We
hypothesized that exoU strains are more adaptable to FQ exposure compared to exoS strains, due
to an underlying difference in fitness, which allows exoU strains to more readily become
resistant. To test this, we created isogenic FQ-resistant mutants of exoU and exoS clinical
strains, and compared the fitness of mutants vs. parent strains in order to investigate the effects
of a resistance-conferring mutation on the fitness of exoU vs. exoS strains.
We found that exoS strains in general have a higher fitness cost compared to exoU strains
during in vitro competition experiments. In addition, the resistant-conferring mutation affects
metabolic utilization differently in exoU and exoS strains. exoU strains may be compensating for
the fitness costs associated with the mutation, as evidenced by the increase in fitness seen of
exoU mutants at the end of a 7-day competition experiment. An assay for supercoiling changes in
mutants vs. parents showed that exoU strains were better able to maintain the wild-type level of
supercoiling compared to exoS strains, suggesting that the fitness cost of the mutation is related
to supercoiling effects, and exoU strains may compensate for these costs via better regulation of
supercoiling. The results of competition experiments in vivo also suggest that exoU strains have
less of a fitness cost of the FQ-resistance mutation.
The lower fitness cost of resistance for exoU strains allows for these strains to more
readily become resistant compared to exoS strains because there is less of a barrier to the
acquisition of resistance-conferring mutations. The implication of this is that the use of FQs will
select not only for the highly resistant strains, but specifically for the highly resistant, highly
virulent exoU strains due to their increased fitness. Understanding the fitness costs of antibiotic

vi

resistance and possibilities of compensation for these costs is essential for the rational
development of strategies to combat the problem of antibiotic resistance.

vii

Tables and Figures
Chapter 1
Figure 1: Prevalence of multi-drug resistant strains of Pa from ICU patients in the
United States……………………………………………………………………………………4
Figure 2: Basic structure of the fluoroquinolone molecule with important
characteristics annotated………………………………………………………………………..6
Figure 3: Increasing rates of ciprofloxacin resistance correlate with increasing
FQ use…......................................................................................................................................7
Figure 4: The type III secretion system…………………………………………………..........13
Figure 5: LD50 in mice of varients of strains PA99 expressing individual
TTSS effectors.............................................................................................................................18

Chapter 2
Figure 6: Proportion of patients and time required to achieve clinical stability.........................22
Figure 7: Distribution of exoU and exoS genes according to levofloxacin MIC.......................23
Table 1: Primers used for PCR and sequencing..........................................................................25
Figure 8: Example of RAPD PCR gel for clonal typing............................................................26
Figure 9: A greater proportion of exoU strains are FQ-resistant compared to
exoS strains..............................................................................................................................................27

viii

Table 2: Target site mutations in exoU and exoS strains............................................................28
Figure 10: Cumulative frequency of isolates with ≥ 2 target site mutations
at each given MIC comparing exoU- vs exoS-containing isolates..............................................30
Table 3: Patient characteristics grouped by disease severity......................................................40
Figure 11: Severity of Pa infection caused by strains with different susceptibilities................42
Figure 12: Combined genotype and FQ susceptibility traits in each clinical group..................44
Table 4: Multivariable logistic regression: Variables predictive of development of
pneumonia...................................................................................................................................44

Chapter 3
Figure 13: Schematic of recombination.....................................................................................52
Table 5: Genetic elements used..................................................................................................57
Figure 14: Functional groupings of differentially expressed genes...........................................61
Table 6: Characteristics of isolates used....................................................................................62
Figure 15: Components of the Tn7 cloning and integration system..........................................65
Figure 16: Fluorescent colonies on an agar plate.......................................................................66
Figure 17: The average mutant:parent CFU ratio per day of the experiment.........................67
Figure 18: Average CFU/ml for each competition experiment..................................................68
Table 7. Average CFU/ml and PC:parent ratios at each day of competition.............................68

ix

Figure 19. Percent of carbon- and nitrogen-containing substrates in each
category of fold difference in growth vs. parent strains..............................................................70
Table 8. Nitrogen substrates with a substantial ( > 2 fold) difference
in growth of PC* compared to parent strain................................................................................71
Figure 20. Expression of TTSS related genes, fold difference of PC*
mutants compared to parent strains.............................................................................................74
Figure 21. The strains contain a genetic insertion in which the lux operon
is under the control of a supercoiling-sensitive promoter...........................................................77
Figure 22. Schematic of experimental workflow to investigate compensation..........................78
Table 9. Results of primary competition experiments including Rif
R
frequencies.....................80
Table 10. Characteristics of aged strains.....................................................................................82
Figure 23. Model of possible evolutionary paths for bacteria under
antibiotic selection........................................................................................................................85

Chapter 4
Figure 24. Cytotoxicity against A549 cells after 4 hours…........................................................92
Figure 25. Average CFU of each strain per gram of lung….......................................................95
Figure 26. Average ratio of PC*: parent CFU per gram….........................................................95
Figure 27. PC*:Parent ratio of CFU per gram of lung…............................................................96
Table 11. Fitness and virulence of aged strains vs. un-aged in host models …..........................97

1

Chapter 1: Introduction and Background

Microbiology
Pseudomonas aeruginosa (Pa) is a Gram-negative, rod-shaped bacterium that is a
ubiquitous environmental organism. Pa is able to colonize and thrive in diverse environments,
due in part to the remarkable plasticity of its genome and its ability to utilize many diverse
energy sources. Pa behaves as an opportunistic pathogen upon encountering a host with impaired
defenses, and can infect animals including humans, as well as insects (D. melanogaster and G.
mellonella), worms (C. elegans) and plants (i.e. A. thaliana) (Sifri and Ausubel, 2005). Within
the human body, Pa can colonize and infect various body sites, including the lungs, urinary tract,
and the eye, and cause devastating infections in burn victims and cystic fibrosis patients.
Pa is a non-spore forming, non-capsular bacillus, and usually has one or two polar
flagella. In the laboratory, Pa grows readily on a variety of culture media and over a wide range
of temperatures, and can be identified by its grape-like odor and the ability of many strains to
produce bright colored diffusible pigments such as pyocyanin (blue-green) and pyoverdin
(yellow-green). Some strains may also produce red or brown pigments (pyorubrin or melanin).
Colony sizes can vary depending on the strain, and can be large and mucoid, or small and
compact. Pa is non-fermenting and derives energy from carbohydrates through an oxidative
metabolism. In addition to having minimal nutritional requirements, Pa is a facultative anaerobe,
and can utilize nitrogen as the terminal electron acceptor if oxygen is not available, thus
providing the ability to live in diverse environments (Wu et al., 2015).
Strains of Pa are traditionally classified by serotyping, based on the structurally diverse
O-antigen of LPS on the bacterial cell surface. There are currently 20 known O-serotypes.

2

Molecular typing techniques are also now being used, such as multi-locus sequence typing
(MLST), pulsed-gel-field electrophoresis (PFGE), or random-amplified-polymorphic DNA PCR
(RAPD-PCR). These techniques identify genetic patterns, or ‘fingerprints’ that allow for
characterization and grouping of strains into clonal types (Groven, 2012; Wu et al., 2015).
Typing is used clinically in order to document outbreaks and track the spread of potentially
dangerous clones.
The genome of Pa consists of a single circular chromosome, and is relatively large; it can
range from 5.5 to 7 Mb depending on the strain. Pa has a high G+C content of 65-67%. The first
Pa genome to be sequenced, strain PAO1, has a genome size of 6.3 Mb with 5570 predicted
ORFs, which made it the largest bacterial genome sequenced at that time (Stover et al., 2000).
Comparison of the genome of Pa with other closely related organisms showed that the large size
of the Pa genome is due to greater genetic diversity and complexity as opposed to genome
organization differences, such as gene duplication events. This allows for the remarkable
functional and metabolic diversity that is seen in Pa; a large number of these genes encode
transport systems and metabolic enzymes (Stover et al., 2000). 8.4% of the Pa genome consists
of predicted regulatory genes, one of the highest percentages in all bacteria (Wu et al., 2015).
This in part provides the high adaptability that allows Pa to survive in a variety of environments
and become such a formidable human pathogen.
The genome of Pa consists of highly conserved regions that make up the core genome,
and variable sequences that make up the ‘accessory’ genome. The accessory genome consists of
genes and groups of genes known as genomic islands that are horizontally acquired. Genomic
islands are large segments of DNA (>10kb) that are interspersed throughout the genome but are
identified as part of the accessory genome by a G-C content different from that of the rest of the

3

genome. Genomic islands allow for genes to be transferred en bloc, and may encode adjunct
activities that allow for survival in specific environments, or specializes genomic islands known
as pathogenicity islands may encode specific virulence factors (Hacker and Carniel, 2001;
Harrison et al., 2010).
Epidemiology and Clinical Importance
Due to its prominence as a cause of invasive infections in susceptible populations,
emergence of resistance to all available antibiotics, and broad arsenal of virulence factors, Pa is
considered one of six bacteria that pose an immediate threat to public health according to the
Infectious Diseases Society of America (Boucher et al., 2009). Pa is an extremely important
pathogen in the hospital setting, where it can readily establish infection in susceptible hosts,
especially in the ICU, defining it as the ‘epitome’ of an opportunistic pathogen of humans (Wu
and Li, 2015). According to the CDC, Pa is the fourth most commonly isolated nosocomial
pathogen, accounting for 10% of all hospital-acquired infection (Diekema et al., 1999).
Specifically, it is a leading cause of hospital-acquired acute pneumonia (Quartin et al., 2013;
Restrepo and Anzueto, 2009), accounting for 20% of all pneumonia cases in ICUs (Di Pasquale
et al., 2014; Gaynes et al., 2005) and this can come with an attributable mortality of 40-70% (Alp
et al., 2004; Richards et al., 1999). Many ICUs have more than one-third of their patients on
mechanical ventilation everyday (Baughman, 2005), making ventilator-associated pneumonia
(VAP) a common problem. Pa is the leading cause of VAP, and death among VAP patients
(Rello et al., 2005).
The ability of Pa to form biofilms on many surfaces, as well as in the lungs of patients
with cystic fibrosis (CF), contributes to its pathogenicity and difficulty of treatment. Specifically,
in VAP patients, the airways become colonized almost immediately after intubation, and most

4

endotracheal tubes from VAP patients are covered with Pa biofilms, providing a reservoir for
infection (Mason and Nelson, 2005).
Antibiotic Resistance
The increasing prevalence of drug-resistant organisms is a major worldwide public health
concern and is threatening the achievements of modern medicine. In a recent report, the CDC
estimated that nearly 2 million people in the United States are sickened with drug-resistant
infections per year (2013). The World Health Organization (WHO) estimates the yearly cost of
infections with antibiotic resistant organisms to the US health system alone is $21 to $34 billion
dollars, accompanied by more than 8 million additional days in the hospital (2014). Gram-
negative pathogens are currently the most worrisome, because of their rapidly increasing rates of
resistance to nearly all treatment options. In addition to the Enterobacteriaceae, Pa is the most
common and the most serious drug-resistant, health-care associated pathogen.
The emergence and spread of multi-drug resistant (MDR) Pa is severely limiting
treatment options, and there are reports of pan-resistant strains that have developed resistance to
the last resort antibiotic, colistin, making Pa one of the greatest therapeutic challenges. The CDC
estimates that about 13% of Pa
infections are caused by MDR strains
and recently labeled MDR-Pa as a
serious public health threat that
requires “sustained and prompt action
so that the problem does not grow”
(2013). The percentage of MDR
strains seem to be growing (Fig 1), with some sites reporting as high as 23% (Lister et al., 2009).
general trends for hospitalized patients. Based on the data in
Table 1, it is difficult to draw any strong conclusions about
trends of resistance to various !-lactams. Among the amino-
glycosides, most studies have focused on gentamicin, with re-
sistance rates ranging from 12 to 22%. Gentamicin was the
least active of the aminoglycosides, with lower rates of resis-
tance being reported for tobramycin and amikacin in most
studies (Table 1).
Although the resistance trends from large national surveil-
lance studies provide important data for consideration, these
studies do not address the potential for much higher rates of
resistance within individual communities and hospitals. For
example, during the years 2001 and 2006, rates of nonsuscep-
tibility among P. aeruginosa isolates in Brooklyn, NY, ranged
from 27 to 29% for cefepime, 30 to 31% for imipenem, 23%
for meropenem, and 41 to 44% for ciprofloxacin (113). These
rates are substantially higher than national trends focusing on
all hospital isolates of P. aeruginosa (Table 1).
Not only are rates of resistance to individual drugs or drug
classes a concern, but the prevalence of multidrug-resistant
strains(resistanttothreeormoredrugclasses)isanevenmore
serioustherapeuticchallenge.Anationalsurveillanceof13,999
nonduplicate P. aeruginosa isolates from ICU patients showed
that multidrug resistance increased significantly, from 4% in
1993to14%in2002(Fig.1A)(178).Forcomparison,another
ICU surveillance study evaluated over 37,000 P. aeruginosa
isolates from 1997 to 2002 and reported an increase in preva-
lenceofmultidrug-resistantstrainsfrom13%to21%(Fig.1B)
(132). Finally, Flamm et al. reported rates of multidrug-resis-
tant P. aeruginosa ranging from 23 to 26% among 52,000 P.
aeruginosa isolates collected in the United States from 1999 to
2002 (54). The highest prevalence of multidrug-resistant
strains was observed among isolates from lower respiratory
tract infections, whereas the lowest prevalence was observed
among isolates from upper respiratory tract infections. Not
surprisingly,multidrug-resistantstrainswereisolatedmorefre-
quently from ICU and nursing home patients.
A multidrug-resistant phenotype can arise in P. aeruginosa
FIG. 1. Increasingprevalenceofmultidrugresistanceamong P. aeruginosaisolatesfromICUpatientsintheUnitedStates.(A)Datafor13,999
nonduplicate isolates collected from 1993 to 2002 (178); (B) data for 37,390 isolates collected from 1997 to 2000 (132). Data represent the
percentageof P. aeruginosaisolatesthatexpressedaphenotypeofmultidrugresistance(resistancetothreeormoredrugclasses)duringeachyear
of the studies. (Panel A is adapted from reference 178 with permission; panel B is based on data from reference 132.)
TABLE 1. Rates of antibacterial resistance among P. aeruginosa isolates from hospitals and ICUs
Antibiotic
% of strains exhibiting resistance
a
Hospital study,
2006
(n" 606)
(211)
Hospital study,
2005
(n" 589)
(212)
Hospital study,
2002
(n" 9,896)
(54)
ICU study, 2002
(n" 951)
(178)
ICU study,
2000–2002
(n! 7,500)
(95)
Hospital study,
2001
(n! 2,157)
(99)
ICU study,
2001
(n! 543)
(99)
Hospital study,
2000
(n" 882)
(100)
!-Lactams
Cefepime 6 5 9 25 12 8 10 9
Ceftazidime 13 10 13 19 17 9 9 13
Piperacillin-tazobactam 11 9 11 10 14 8 8 13
Aztreonam 12 32
Imipenem 11 7 16 23 22 12 16 16
Meropenem 6 7 18 11 16 10
Fluoroquinolones
Ciprofloxacin 21 22 35 32 33 26 25 25
Levofloxacin 22 22 34 32 27 25 27
Aminoglycosides
Amikacin 5 10 4 3
Tobramycin 8 10 12 16
Gentamicin 12 12 16 22 15 15 14
a
Based upon CLSI interpretive breakpoints.
584 LISTER ET AL. CLIN.MICROBIOL.REV.
on March 5, 2013 by USC Norris Medical Library http://cmr.asm.org/ Downloaded from
Figure 1. Prevalence of multidrug resistant strains of Pa from
ICU patients in the United States.
(Lister et al. 2009)

5

Pa is intrinsically resistant to many antibiotics, due in part to the innate impermeability of
its membrane, which does not allow large, hydrophilic molecules from passing through.
However, it also readily acquires additional resistance phenotypes. Only three major categories
of antimicrobials have reliable anti-pseudomonal activity to begin with: β-lactams,
aminoglycosides, and the fluoroquinolones. Pa is naturally resistant to some β-lactams, such as
ampicillin, because its genome endogenously carries several penicillin-binding proteins (PBPs),
as well as β-lactamases such as AmpC, that destroy the amide bond of the drug and render it
ineffective. Acquired resistance to β-lactams usually occurs through the acquisition of extended-
spectrum β-lactamases via horizontal gene transfer. Resistance to the aminoglycosides can also
occur in this manner. Furthermore, the genes conferring resistance to multiple classes of
antibiotics are frequently co-associated and acquired together (Mesquita et al., 2013; Poole,
2011). Loss of outer membrane permeability is another commonly acquired resistance
mechanism, specifically to the carbapenems. Loss of functionality of the porin OprD blocks
carbapenems from entering the cell.
Resistance to multiple antibiotic classes can occur because of the cross effects of
resistance mechanisms, such as with efflux pumps, which can effectively confer resistance to
different antibiotic classes. Efflux pumps span both membranes of the Gram-negative bacterial
cell, consisting of a periplasmic membrane fusion protein, an outer-membrane factor, and a
cytoplasmic membrane transporter. These systems pump out certain antibiotics, decreasing the
amount of drug in the cytoplasm. The resistance-nodulation division (RND) family of efflux
pumps is a significant contributor to resistance in Pa, which encodes 12 systems of these pumps
in its genome (Wu et al., 2015). The MexAB-OprM system was the first discovered, and is most
frequently associated with β-lactam resistance; however, in a perfect example of cross-resistance

6

mechanisms, it can also export many other types of antibiotics including tetracyclines,
macrolides, fluoroquinolones, chloramphenicol, trimethoprim and novobiocin. While the
MexAB-OprM system is constitutively expressed and is part of the innate resistance of Pa, some
of the other efflux systems are naturally repressed, and mutations in regulatory genes are
required for overexpression and resistance to occur (Lister et al., 2009).
Resistance to the Fluoroquinolones
Since the discovery of the quinolones in the 1960s, there has been much clinical interest
in these drugs due to their many promising attributes: high potency, broad spectrum of activity,
and excellent oral bioavailability. Quinolones were derived from quinine, and one of the earliest
modifications was the addition of a
fluorine molecule at position 6,
which provided a huge increase in
gyrase inhibition. Figure 2 shows
the basic structure of a
fluoroquinolone (FQ). In the late
1980s- 1990s systemic agents with
activity against Gram-negatives were developed and marketed, including ciprofloxacin and
levofloxacin, both with anti-pseudomonal activity (Andersson and MacGowan, 2003).
Unfortunately, the promising attributes of the fluoroquinolones, such as broad-spectrum
activity, bioavailability, and ease of use, led to the tremendous overuse of these antibiotics and
eventual widespread resistance in Pa. In 2002, FQs were the most commonly prescribed class of
antibiotics to adults in the United States (Linder et al., 2005). Figure 3 portrays results from a
2003 study that showed the increase in FQ-resistance in Pa correlating with the increase in FQ
Essential for gyrase
binding
R
R
R
Enhances gyrase
potency
Near gyrase
binding site
Figure 2. Basic structure of the fluoroquinolone
molecule with important characteristics annotated
(adapted from Andersson et al. 2003)

7

prescribing (Neuhauser et al., 2003). This
pattern has led to the current problematic
state in which it has been estimated that
close to 35% of all clinical strains of Pa
are FQ-resistant (Edelsberg et al., 2014).
Furthermore, many FQ-resistant strains are
likely to be resistant to other classes of
antibiotics as well. The aforementioned 2003 study by Neuhauser et al. showed that a significant
proportion of the ciprofloxacin-resistant strains of Pa were also resistant to other classes of
antibiotics, e.g. 66% were resistant to gentamicin, 40% to ceftazidime, and 38% to imipenem.
Another study in a different population found that 87% of FQ-resistant Pa strains were resistant
to 2 or more classes of drugs (Hsu et al., 2005). These statistics underscore the therapeutic
challenge clinicians face when treating Pa infections.
Mechanisms of FQ- resistance
As described above, efflux pumps play a significant role in FQ resistance in Pa. Four
members of the RND family of efflux systems can accommodate the FQs. The MexEF-OprN
system has been well characterized and associated with resistance in clinical isolates. A
transcriptional activator, MexT, regulates the expression of this system. Interestingly, many
wild-type strains have inactivating mutations in mexT, and the overexpression of mexEF-oprN
occurs when these mutants ‘revert’ to a functional state. Other RND systems are controlled by
repressors, which, when mutated, allow for overexpression, leading to resistance (Lister et al.,
2009; Poole, 2011).
Resistance conferred via efflux pumps is not specific to the FQs as these pumps have
Figure 3. Increasing rates of ciprofloxacin
resistance correlate with increasing FQ use
(from Neuhauser et al. 2003)

8

broad substrate specificity; many of the efflux systems can pump out various structurally
unrelated antibiotics. Therefore, overexpression of efflux pumps confers a multidrug-resistant
phenotype. In contrast, target site mutations (TSMs) confer specific resistance by altering the
target enzyme of the FQs. The primary target of FQs in Gram-negative bacteria is DNA gyrase,
encoded by gyrA/B. The secondary target is topoisomerase IV, encoded by parC/E. Both
enzymes are tetramers, composed of 2 subunits of A/B and C/E respectively. ParC is
homologous to GyrA, and ParE is homologous to GyrB (Hooper, 2000). These enzymes are
responsible for the topological maintenance of the bacterial genome. DNA gyrase catalyzes the
negative supercoiling of DNA, and topoisomerase IV decatenates daughter replicons. Both
functions are essential for successful replication of the genome. The reaction cycle of both
enzymes includes breakage and resealing of double-stranded DNA. FQs disrupt this sequence,
binding to the enzymes and preventing the completion of the ligation reaction. This results in
stalled replication forks, double stranded DNA breaks throughout the genome, and eventual
bacterial cell death (Hooper, 2001).
Mutations in the quinolone-resistance-determining regions (QRDR) of gyrA/B and
parC/E that confer resistance to the fluoroquinolones have been well described in P. aeruginosa
(Jalal et al., 2000; Lee et al., 2005a). Mutations occur first in gyrA, as the primary
fluoroquinolone target, and resistance increases as subsequent mutations in parC are acquired.
(Akasaka et al., 1999; Higgins et al., 2003). Mutations in gyrB and parE can also contribute to
increased resistance, but are less common (Akasaka et al., 2001).
Pathogenesis of P. aeruginosa
The ability of Pa to cause severe disease in susceptible hosts is due to its broad arsenal of
virulence factors. Because of its large genome, Pa is endowed with a diverse set of virulence

9

genes. Specific virulence genes are expressed under certain conditions, allowing Pa to be a
successful pathogen in a variety of host environments. Virulence in Pa is combinatorial; different
virulence factors work together to establish infection, invade the immune system, and persist in
the host.
Adhesins
In order to establish infection in the host, Pa expresses molecules that allow for adhesion
to host tissue. Adhesion is a crucial early step in infection, as well as for colonization and
replication on abiotic surfaces. Pa has two major adhesion molecules, flagella and type IV pili.
The single flagellum of Pa allows for swimming motility which helps in dispersal through the
respiratory tract as well as chemotaxis towards substrates necessary for survival (Feldman et al.,
1998). The importance of flagella in pathogenesis is exemplified by studies of non-flagellated
mutants, in which virulence and epithelial host cell invasion is attenuated (Fleiszig et al., 2001).
Pa flagella activate a relatively robust immune response through recognition by TLR5 and
subsequent induction of IL-8 (Bucior et al., 2012; Veesenmeyer et al., 2009). Therefore, After
initial colonization, flagellar genes are usually downregulated, and in chronic infections,
aflagellar mutants are selected for (Sadikot et al., 2005). The type IV pili are surface molecules
that also play an important role in motility and adhesion. In fact, about 90% of the adhesion
capability of Pa depends on these pili (Wu et al., 2015).
Secreted factors
Pa has a number of complex secretion systems, many of which are utilized for delivering
virulence factors either into the environment or directly into host cells. There are seven known
secretion systems in Gram-negative bacteria (types I-VII) and the Pa genome contains five of
them: type I, II, III, V, and VI. Each system secretes specific substrates. The proteins secreted by

10

the type II system include the virulence factors LasA, LasB, and exotoxin A. LasA and LasB are
proteases known as elastases, which target elastin, an element of connective tissue found in lung,
vascular, and ocular tissue that has high stability against most other proteases (Mesquita et al.,
2013; Wu et al., 2015). Elastases disrupt the tight junctions of the epithelial barrier and impair
ciliary function (Sadikot et al., 2005). In a perfect example of multiple virulence factors working
in concert, another protease, protease IV, facilities pili adherence to host tissue by degrading
fibronectin, which exposes the underlying pili receptors on epithelial cells (Wu and Li, 2015).
Exotoxin A is also secreted by the type II system and enters the host cell through
receptor-mediated endocytosis. It has ADP-ribosyltransferase activity that inhibits protein
synthesis, eventually leading to cell death (El Hage et al., 2010). Exotoxin A is responsible for
tissue necrosis and assists in bacterial invasion and dissemination (Sadikot et al., 2005).
The characteristic blue-green color of many strains of Pa comes from the production of
pyocyanin. This pigmented molecule interferes with the antioxidant defenses in the lung, causing
oxidative damage to host tissues (O'Malley et al., 2003). Other pigmented secretions include
pyochelin and pyoverdin, which are important siderophores that bind iron efficiently and allow
Pa to acquire iron from host tissues. Iron is necessary for bacterial survival, and pyochelin and
pyoverdin scavenge free iron for bacterial use in environments with limited iron availability
(Takase et al., 2000).
Strains of Pa found in the lungs of CF patients exhibit many differences compared to
nosocomial strains causing pneumonia and other acute infections. A complex regulatory network
turns on and off pathways necessary for infection in the different conditions. Alginate, an
extracellular polysaccharide, is an example of virulence expression induced by adaptation to the
CF lung. Production of alginate is usually due to a mutation in the muc genes, and leads to the

11

mucoidy appearance of isolates obtained from CF patients. It has been shown to allow Pa to
evade immune clearance (Boucher et al., 1997).
Quorum Sensing
Bacteria have evolved a process of cell-to-cell communication known as quorum sensing
(QS) that is based on the detection of diffusible signaling molecules. QS allows bacterial
populations to collectively control gene expression and synchronize group behavior. This ensures
that the expression of energetically costly activities, such as virulence factors, only occurs when
a large enough population of bacterial cells is present (Lazdunski et al., 2007).
The signaling molecules are small compounds called autoinducers that are released from
the bacterial cell into the environment, and are then sensed by neighboring bacteria. In Gram-
negative bacteria, the signaling molecules are acylated homoserine lactones (AHLs). When a
threshold concentration of AHL is reached, it binds to the LasR or RhlR transcriptional
activators that induce expression of certain groups of genes (Fuqua and Greenberg, 2002). The
Pseudomonas Quinolone Signal also (PQS) is a third QS system in Pa that regulates RhlR and is
regulated by LasR. Microarray expression studies have shown that QS systems regulate 6-10%
of the Pa genome, and many of these QS-regulated genes are involved in the expression of
virulence factors such as type III secretion, exotoxinA, elastase, and biofilm formation(Brint and
Ohman, 1995; Latifi et al., 1995; Winson et al., 1995). The importance of QS regulation on
virulence expression is demonstrated by studies showing that QS deficient mutants display
decreased virulence compared to wild type (Juhas et al., 2005).
Biofilm formation
Biofilms are a structured community of bacterial cells that are enclosed in a self-
produced matrix and are adherent to a surface, either living or abiotic (Ma et al., 2009). Pa

12

biofilms commonly coat medical devices such as catheters and breathing tubes and are inherently
difficult to eradicate. Biofilms present in the body can act as a continued source of planktonic
(free-living) circulating bacterial cells, causing cycles of acute infection that cannot be treated
without the surgical removal of the infected substrate or device. Biofilm formation in Pa is also
responsible for the chronic infection of the CF lung that leads to high morbidity and mortality in
these patients (Costerton et al., 1999).
Biofilm formation is a complex process and has been extensively studied in Pa. Pa
undergoes a highly regulated transition from planktonic free living cells to highly differentiated
biofilm communities. This begins with the initial attachment to a surface, followed by the
accumulation of cell clusters, and ultimately, a mature biofilm containing large aggregates of
bacteria and the production of the extracellular matrix composed of nucleic acids, proteins, and
exopolysaccharides (Ventre et al., 2007).
The physiology of biofilm cells is complex and extremely different from planktonic cells.
Since biofilms are a community of cells, the physiological state of each cell is determined by the
location within the biofilm; those closer to the surface have access to oxygen and nutrients and
are more likely to be metabolically active, while those within the exopolysaccharide matrix are
metabolically dormant. Biofilms are protected from the hostile effects of the environment, such
as antibiotic exposure or host immune attack (Mesquita et al., 2013). This makes biofilms
tolerant to high concentrations of antibiotics, adding to the difficulty in treatment and
eradication.
The Type III Secretion System
Pa and other Gram-negative pathogens utilize the Type III Secretion System (TTSS) as a
major virulence determinant during acute infections. It is a molecular syringe-like apparatus that

13

extends through both the inner and outer bacterial
membranes and directly contacts the host cell, forming
a pore and injecting toxins directly into the host cell
cytoplasm (Lee, 1997). The Pa system is similar to
Yersinia spp., Salmonella enterica, and Shigella
flexneri, (Hauser, 2009) and consists of the secretion
apparatus, the translocation apparatus, and the secreted toxins (called effector proteins) and their
cognate chaperons (Gauthier et al., 2003) (Figure 4). The secretion apparatus is termed the
needle complex, and functions as the channel to transport the effectors through both membranes
and into the host cell, and may also serve as the sensor for host cell contact. It consists of a multi-
ring base and a filament protein, PscF. PscF polymerases to form the needle structure. The
translocation apparatus is a membrane pore that delivers the effectors across the host cell
membrane and into the cytoplasm. This translocation process is very efficient; less than 0.1% of
the secreted effector proteins escape into the extracellular space (Sundin et al., 2004). The
proteins PopB, PopD, and PcrV form the translocation machinery (Sundin et al., 2002). PopB
and PopD form the translocation pore, and PcrV forms the ‘cap’ or tip of the needle complex and
has been found to be necessary for translocation of the effector toxins (Goure et al., 2004; Lee et
al., 2010). There are four effector toxins in Pa: ExoY, ExoT, ExoU, and ExoS. Interestingly, Pa
encodes fewer effector toxins than any other well-studied bacterial TTSS, which is surprising
given its large genome and ability to adapt to many environments.
Effector toxin ExoS is a bi-functional toxin that has GTPase activating protein (GAP)
activity as well as ADP ribosyl transferase (ADPRT) activity. The ADPRT domain exerts several
negative effects on the host cell including altering morphology, inhibiting DNA synthesis, and

14

disruption of vesicular trafficking (Rocha et al., 2003), and only becomes activated through
interaction with a host-derived co-factor (Aitken, 2006), demonstrating the ability of Pa to hijack
host factors. Both functional domains of ExoS disrupt the actin cytoskeleton of the host cell and
may decrease phagocytosis as well as cell-cell adherence, which would allow Pa to penetrate
epithelial barriers (Frithz-Lindsten et al., 1997). ExoT shares 76% amino acid identity with
ExoS, and also has GAP and ADPRT bi-functionality (Yahr et al., 1996). ExoU is the most
potent of the toxins. It is a phospholipase that causes rapid host cell death by destroying the
plasma membrane (Finck-Barbançon et al., 1997; Hauser et al., 1998). Like ExoS, it requires a
host cell co-factor for activation. Superoxide dismutase 1 (SOD1) was found to be the hijacked
co-factor necessary for ExoU activity (Sato et al., 2006). ExoY is an adenylyl cyclase (Yahr et
al., 1998) that results in changes in cAMP concentrations inside the host cell, leading to changes
in expression of genes effecting the actin cytoskeleton and other cell signaling functions.
The TTSS is important mainly in acute infections, and enhances the pathogenicity of Pa,
leading to a tremendous increase in disease severity. The significance of the TTSS in infection
has been shown in various infections in patients as well as animal models. A functional TTSS
has been associated with poor outcome in patients with ventilator-associated pneumonia, higher
mortality, and higher relapse rates (El Solh et al., 2008; Hauser et al., 2002; Roy-Burman et al.,
2001). In a study of Pa bloodstream infections, expression of TTSS proteins was the strongest
predictor of 30-day mortality (Hattemer et al., 2013). In a mouse burn model, Pa mutants
defective in the TTSS translocation apparatus had a nearly complete loss of virulence, allowing
for almost 100% survival, while wild type strains led to <20% survival (Holder et al., 2001). A
study of keratitis using the mouse scratched cornea model (Lee et al., 2003) showed that TTSS
mutants had reduced bacterial colonization and penetration compared to wild type strains.

15

During Pa acute lung infection, the expression of the TTSS leads to excessive numbers
of neutrophils and macrophages recruited to the lung and can cause collateral damage to the
tissue. The effect on host cell adhesion of the bacterial effector toxins results in the loss of
integrity of the epithelial barriers, which allows bacterial dissemination into the bloodstream and
the leakage of pro-inflammatory cytokines, leading to septic shock (Kurahashi et al., 1999).
The regulation of the TTSS is complex and still not fully understood. It is a highly
ordered process that involves sequential expression, secretion, and assembly of the rod, needle,
and translocation machinery. Regulation can occur at the intrinsic/genetic level and at the
extrinsic/environmental level. Expression is kept at a low basal level until host cell contact
activates secretion, and then expression is rapidly increased. Directly controlling expression of
all the TTSS operons is activator ExsA (Hovey and Frank, 1995; Yahr et al., 1995). ExsA is
controlled by a system of regulator proteins (ExsC, ExsD, and ExsE) that work in concert to
sequester ExsA away from its target genes until secretion is activated, in which time ExsA is
then freed (Dasgupta et al., 2004).
Global regulation of the TTSS in response to external stimuli is still poorly understood.
Two systems are known to work through exsA: the intracellular cyclic AMP system and the
GacS/GacA two component signal transduction system. cAMP is thought to regulate Virulence
factor regulator (Vfr). The Vfr regulon consists of ~200 virulence- associated genes, such as type
IV pili, quorum sensing, and TTSS (Wolfgang et al., 2003).
When activated, the GacS/GacA signal ultimately induces expression of two small
regulatory RNAs, RsmY and RsmZ. These RNAs sequester the RsmA regulator, which is
required for increased levels of TTSS expression. The GacS/GacA system, however, is
controlled by two other two-component signaling systems that exert opposing regulation, LadS

16

and RetS. Certain environmental cues activate LadS, which induces the GacS signaling cascade,
leading to decreased TTSS expression, while other cues activate RetS, which blocks GacS
signaling, leading to decreased rsmY/Z transcription and therefore increased TTSS expression.
Interestingly, while RsmA is an inducer of TTSS translation, it is an inhibitor of biofilm-related
genes. Therefore, the GacS/GacA system is seen as the pathway through which Pa integrates
environmental cues in order to determine when to switch from expression of genes needed for
acute infection (TTSS) to genes needed for chronic infection (biofilm) (Brencic and Lory, 2009;
Brencic et al., 2009; Ventre et al., 2006)
exoU vs exoS
Almost all strains of Pa contain the genes for the type III secretion machinery (Hauser,
2009); however, most strains do not harbor the genes for all four cytotoxins. The exoY and exoT
genes are found in 90-100% of clinical isolates; but interestingly, the genes encoding ExoU and
ExoS are mutually exclusive in most strains. exoS strains are more prevalent and account for
about 70% of clinical isolates, while the remaining 30% contain exoU. Less than 1% strains
carry both or neither genes (Feltman et al., 2001). This bias towards mutual exclusivity points to
a strong selection for strains to carry one of the genes, and a strong selection against carrying
both, but the evolutionary explanation for this is not currently known.
Although whole-genome comparisons of environmental and clinical strains show that
about 97% of genes are conserved (Wolfgang et al., 2003), the Pa genome does consists of a
highly conserved core genome interspersed with a variable accessory genome (Mathee et al.,
2008). The exoU gene is part of Pa’s accessory genome, while exoS is part of the conserved, core
genome. exoU and the gene encoding its chaperone protein SpcU are contained on a discrete
section of the genome on a type of genomic island called a pathogenicity island. The G/C content

17

is much lower in this region than in the rest of genome, it is flanked by symmetrical sequences,
and it lies next to a known integration site at a tRNA gene (Kulasekara et al., 2006; Wolfgang et
al., 2003). These characteristics are common for mobile genetic elements that have been
horizontally acquired relatively recently. It is thought that the exoU and spcU genes were
combined with the other genes contained in the pathogenicity island into a mobile genetic
element in a closely related species. Subsequently, the island was introduced into an ancestral
Pa strain as a whole, eventually losing its ability to excise itself and becoming permanently
locked in the genome (Kulasekara et al., 2006). exoU and spcU have been identified on different
pathogenicity islands in different strains. In the common exoU-containing laboratory reference
strain PA14, exoU is contained on a small, 11-kb island called PAPI-2 (He et al., 2004). Other
islands containing exoU and spcU are as large as 80Kb (Wolfgang et al., 2003).
The mutual exclusivity of the exoS and exoU genes points to an inherent incompatibility.
In all but a small percentage, the acquisition of an exoU pathogenicity island comes with the
subsequent deletion of the exoS gene from the genome, regardless of the strain, its environmental
niche, or the type of pathogenicity island that exoU is contained on. This suggests that there are
conserved selective pressures and some universal reason that does not allow the exoU and exoS
genes to coexist in the same genome, most likely due to a strong incompatibility of the gene
products and their functions. In support of this, one study found a trend towards less bacterial
persistence in a mouse model with a strain secreting both ExoU and ExoS compared to an
isogenic strain secreting only one or the other (Shaver and Hauser, 2006).
Although accounting for the minority of clinical isolates, exoU strains pose a serious
clinical problem. It has been shown in vitro, in vivo and clinically that ExoU is the most virulent
of the effector toxins. A study using an in vitro tissue culture model of microbial keratitis using

18

strain PA14 demonstrated that functional ExoU was required for full virulence, defined in this
study as traversal of corneal epithelia (Ramirez et al., 2012). A tissue culture study using various
TTSS mutants and manipulations of the laboratory PAK strain (naturally exoS/T/Y) showed that
the most dramatic difference in virulence occurred when the exoU operon was inserted into a
strain that had the rest of the TTSS effectors deleted; the recombinant strain was 100 times more
virulent than the wild type PAK (Lee et al., 2005b).
Mouse models of pneumonia comparing strains secreting just one of the effector toxins
have been able to show the relative contributions of each toxin in the host. Studies by Hauser and
colleagues were performed using a strain of Pa that naturally contained the genes for all four
effector toxins. By creating various mutants of this strain so that each mutant contained only one
functional effector toxin, the relative potency of each toxin in the host was compared in a
controlled background. Infection with the exoU-secreting strain caused the greatest pathogenicity
in the animals; 40-fold fewer bacteria were required for 50% mortality compared to the null (no
effectors) strain (Figure5). Furthermore, infection with the exoU strain increased bacterial burden
in the lungs of the mice and increased
dissemination to the liver and spleen (Shaver and
Hauser, 2004). Other studies in mouse models
have shown that exoU activity induces excessive
inflammation and tissue damage in the host as
well as increased bacterial dissemination that can
lead to septic shock (Kurahashi et al., 1999).
In patients, exoU strains are associated with poor outcomes of pneumonia (Roy-Burman
et al., 2001) and with persistence and severity of disease (El Solh et al., 2008; Schulert et al.,

19

2003). A clinical study showed that bacteremic patients infected with strains expressing the
TTSS compared to those infected with non-TTSS expressing strains had an increased 30-day
mortality; specifically, none of the patients infected with ExoU-secreting strains survived past 30
days (El-Solh et al., 2012).
Correlation of exoU and FQ-resistance
Previously, my advisor, Dr. Wong-Beringer, and her clinical research group investigated
outcomes of patients infected with Pa, and showed that patients infected with FQ-resistant
strains had a 3-fold higher mortality and prolonged illness compared to those infected with
susceptible strains (Hsu et al., 2005). In a separate study, they reported a significant association
between FQ-resistance and the exoU genotype in 45 clinical isolates of Pa obtained from various
body sites exhibiting a range of susceptibilities to fluoroquinolones (Wong-Beringer et al.,
2008). This suggested that there may be a co-selection of exoU with FQ-resistance in the clinical
setting, leading to worse outcomes for patients. Others have reported similar findings; a study by
Garey et al. analyzed 120 bloodstream isolates of Pa and found that exoU+ strains were more
frequently multi-drug resistant compared to exoS+ strains (Garey et al., 2008). Further analysis
on the data indicated that the highest frequency of resistance among the exoU+ strains was
towards the FQs compared to other antibiotics (personal communication). Similar results were
observed by others examining corneal isolates (Borkar et al., 2014; Zhu et al., 2006).
Interestingly, Fleiszig et al. has shown that resistance of Pa to contact lens disinfectant was
linked to acute cytotoxic activity from ExoU towards corneal epithelial cells and exsA, a
transcriptional regulator of genes encoding effectors of the TTSS (Lakkis and Fleiszig, 2001). In
a study of population structure of Pa isolates in Europe, the exoU+ genotype was found to be
significantly associated with multidrug resistance and ciprofloxacin resistance compared to the

20

exoS+ genotype (Maatallah et al., 2011). More recently, a higher proportion of exoU strains were
found to be multi-drug resistant than exoS strains in a study of patients with diabetic foot
infections (Zhang et al., 2014).
The evidence described above suggests a co-selection of virulence (presence of the exoU
gene) and antibiotic resistance. While the processes of virulence and antibiotic resistance may
seem to be distinct biological pathways, investigators are starting to realize that these two traits
may be interdependent (Hauser, 2014). Studies of antibiotic resistance in other drug-organism
pairs present a complex view of bacterial regulation mechanisms. Resistance has been shown to
either increase or decrease virulence, depending on the species and drug studied (Beceiro et al.,
2013).
Summary
Pa is an important opportunistic pathogen in the clinical setting. It is highly adaptable to
a variety of environments and host infection sites. Furthermore, it can form biofilms on medical
devices and surfaces, making it extremely difficult to eradicate. The high prevalence of antibiotic
resistance in Pa limits treatment options; specifically, resistance to the once-effective
fluoroquinolones is widespread.
The many virulence factors encoded in the Pa genome make it a formidable pathogen
with the ability to cause severe disease. Specifically, the type III secretion system is the major
virulence determinant during acute infections, and allows Pa to inject effector toxins
(exoU,S,T,Y) directly into the host cell cytoplasm, leading to immune evasion and the
establishment of infection. Most clinical isolates of Pa contain the TTSS secretion machinery;
however, it is very rare that a single strain contains the genes for all four toxins. Most strains
contain exoT and exoY, but the genes encoding ExoU and ExoS are generally mutually exclusive.

21

exoS strains are more prevalent, but exoU strains are more virulent and cause more severe
disease in animal models and worse outcomes in patients.
Published studies with small sample sizes have shown that exoU-containing strains are
more likely to be FQ-resistant, suggesting a possible co-selection of virulence and resistance in
Pa. Therefore, the goals of this dissertation were to 1) confirm the observed correlation of FQ-
resistance and TTSS genotype in a larger collection of clinical isolates and 2) investigate the
biological impact of FQ-resistance on the fitness and virulence of Pa, specific to strain
background of exoU vs. exoS strains.

22

Chapter 2:
ExoU Correlates with Increased FQ-resistance and Virulence in Clinical Isolates of Pa

Introduction
Although exoU strains account for about 30% of clinical isolates (Feltman et al., 2001),
they represent a significant clinical problem. Infections with exoU strains are often associated
with poor outcomes. In a study of 103 pneumonia patients, those infected with exoU strains had a
relative risk of death of 2.3 (compared to infection with a non-TTSS secreting strain), while the
relative risk of death for infections with exoS strains was 1.2 (Roy-Burman et al., 2001). In a
smaller study of patients with ventilator-associated pneumonia, those infected with exoU strains
were more likely to have persistence of Pa than those infected with exoS strains (El Solh et al.,
2008). In a study of Pa bacteremia, all patients who did not survive to 30 days were infected with
exoU strains, while none of the patients who survived the first 30 days were infected with an
exoU strain. Clearly, infection with an exoU strain decreases the chance of survival and prolongs
the duration of illness.
Clinical studies investigating infections with fluoroquinolone-resistant (FQ-R) strains of
Pa have shown worse outcomes for these patients
compared to FQ-susceptible (FQ-S) strains.
In a case-control study of 177 patients comparing
FQ-R (cases) and FQ-S (controls), those infected
with FQ-R strains had three-fold greater mortality.
Among those who achieved clinical stability, the
time required for vital signs to return to baseline

23

was prolonged by 5 days in the FQ-R group (Figure 6). The underlying severity of illness was
controlled for in this study (Hsu et al., 2005) suggesting that FQ-R is independently associated
with increased virulence.
Further investigation into the correlation of FQ-R and poor patient outcomes led to the
observation that the highly virulent exoU strains are associated with increased FQ-resistance. In a
study of Pa isolates collected from 45 patients at
Huntington Memorial Hospital in Pasadena, CA,
exoU isolates were associated significantly with
FQ-R when compared with exoS isolates (92% of
exoU strains were FQ-R vs. 61% of exoS strains),
and the proportion of exoU-containing isolates
increased as the level of FQ- resistance (based on
the minimum-inhibitory concentration [MIC] to levofloxacin, an FQ) increased (Figure 7). The
mechanisms of resistance were also investigated, and 91% of the exoU strains had a gyrA
mutation with overexpression of efflux pumps, while only 59% of exoS strains portrayed the
same mechanisms (Wong-Beringer et al., 2008).
Study 1: Differentiation in Quinolone Resistance by Virulence Genotype in Pseudomonas
aeruginosa. (Agnello M., PLoS ONE, 2012)
Based on the above results from a relatively small sample of 45 patients, we further
investigated TTSS genotype and FQ-resistance phenotype and mechanism in a larger sample of
270 clinical isolates also collected from patients admitted to Huntington Memorial Hospital. The
isolates used for this study were obtained from respiratory sources, and most of the patients had
pneumonia. We hypothesized that Pa strains, depending on their TTSS effector genotypes,

24

differentially develop resistance-conferring mutations, which results in the observed selection
bias for an FQ-resistant population predominated by exoU strains. The goal of the study was to
compare the FQ-resistance phenotypes and mechanisms of exoU and exoS strains, specifically:
1) prevalence and degree of FQ-resistance, 2) efflux pump overexpressed phenotype, and 3)
mutations in target genes conferring FQ resistance.
Materials and Methods.
Ethics Statement
This study has been approved by the institutional review boards at both Huntington
Hospital and University of Southern California. Informed consent was waived from all
participants since bacteria cultures were saved as part of a longitudinal epidemiological
surveillance study for resistance trend. Respiratory cultures from hospitalized patients growing
Pa were later identified retrospectively from microbiology records. All data was analyzed
anonymously.
Susceptibility Testing and Sequencing for Target Site Mutations
Bacterial isolates were stored in cryovials at -80C until ready for testing. Clonality of the
clinical isolates was assessed by random amplified polymorphic DNA polymerase chain reaction
(RAPD-PCR) using a primer sequence published previously (Mahenthiralingam et al., 1996)
(Table 1). Susceptibility testing to levofloxacin was performed by broth microdilution in 2-fold
dilution at concentrations ranging from 0.25 to 128 µg/ml according to guidelines recommended
by CLSI (Institute, 2007). FQ resistance is defined as levofloxacin MIC ≥ 4 µg/ml. Both the
TTSS effector genotype (exoU and exoS) and the QRDR region of respective target genes, gyrA,
gyrB, parC and parE was assayed by polymerase chain reaction (PCR) using previously
published primers and conditions (Jalal and Wretlind, 1998; Lee et al., 2005a). The latter was

25

sequenced to identify active and silent mutations compared to wild-type strains PAO1
(exoS+/exoT+/exoY+) and PA103 (exoU+/exoT+/exoY+). Primer sequences are listed in Table
1. Efflux pump overexpressed phenotype was determined by use of a commercially purchased
efflux pump inhibitor (EPI, MC-0228) (Sigma) previously shown to broadly inhibit the
multidrug Mex efflux pumps of Pa in which the FQs are substrates (Hsu et al., 2005;
Lomovskaya et al., 2001). An overexpressed phenotype is defined as ≥ 8-fold decrease in
levofloxacin MIC when tested in the presence of EPI at 20 µg/ml previously established based
on the change in MICs between wild type and efflux pump knockout mutants (Lomovskaya et
al., 2001).
Statistical Analysis
All respiratory isolates were grouped by TTSS effector genotype and compared for their
degree of resistance to levofloxacin as measured by MIC, efflux pump involvement (magnitude
of decrease in levofloxacin MIC in the presence of an efflux pump inhibitor), and number and
type of target site mutations. Chi-square or Fisher’s exact tests were used where appropriate. A p
value ≤ 0.05 denotes statistical significance.
TABLE 1. Primers used for PCR and sequencing
Name Primer Sequence (5’-3’)
GyrA Forward: ttatgccatgagcgagctgggcaacgact
Reverse: aaccgttgaccagcaggttgggaatctt
GyrB

Forward: gcgcgagatgacccgccgca
Reverse: ctggcggaagaagaaggtcaaca
ParC

Forward: cgagcaggcctatctgaactat
Reverse: gaaggacttgggatcgtccgga

26

ParE

Forward: cggcgttcgtctcgggcgtggtgaagga
Reverse: tcgagggcgtagtagatgtccttgccga
ExoS

Forward: atggcgtgttccgagtca
Reverse: aggtgtcggttcgtgacgtct
ExoU

Forward:ggcacatatctccggttccttc
Reverse: tcaactcagctgccaaccatgc
RAPD 272 agcgggccaa

Results
Bacterial Strains
A total of 270 Pa isolates obtained from respiratory specimens from different patients
were evaluated in this study. Majority of the strains (93%) caused pneumonia while the
remainder caused bronchitis or colonized the airways. About 20% of the patients with
pneumonia had ventilator-associated pneumonia. Clonality was determined for most strains
(n=223); those strains carrying both or neither genes encoding ExoU and/or ExoS were excluded.
Strains with identical band patterns
were considered related and
assigned a unique random
amplified polymorphic type.
Clonal distribution was similar
between FQ-resistant and FQ-sensitive strains with 51% (65/126) and 45% (43/95) of strains
belonging to unique RAPD types, respectively. The 63 FQ-resistant exoU strains sequenced for
FQ target site mutations fell into 34 clonal groups, and the 27 FQ-sensitive exoU strains

27

sequenced fell into 17 clonal groups. Among the exoS isolates, the 65 strains sequenced fell into
37 groups, and the 68 FQ-sensitive strains are in 26 groups. Overall, there are no more than 9
isolates in any given clonal group. Figure 8 is an example of one RAPD PCR analyzed on an
agarose gel.
Prevalence and Degree of FQ resistance
Overall, less than half of the 270 isolates (38%, 103/270) contained the exoU gene while
56% (152/270) were exoS+; five isolates were positive for both effector genes (2%) while 10
(6%) had neither. Similar proportion of exoU and exoS strains caused pneumonia (85%, 87%
respectively). Slightly over half (54%, 145) of the isolates were FQ-resistant.
The distribution of minimum-inhibitory concentration (MIC) values to levofloxacin was
similar between exoU and exoS populations. MIC50 and MIC90, defined as the minimum
concentration necessary to inhibit 50% and 90%, respectively, of all strains, were found to be 8
and 64 µg/ml. Notably, a significantly higher proportion of exoU strains were FQ-resistant
compared to exoS strains (63%, 65/103 vs 49%, 75/152; p =0.03) (Figure 9).
Figure 9. A greater proportion of exoU strains are FQ-resistant compared to exoS strains

0
20
40
60
80
100
exoS exoU
Percent of Isolates
Sensitive
Resistant

28

Efflux Pump Overexpressed Phenotype
Efflux pump overexpression due to mutations in regulator genes is one mechanism of
resistance in Pa. The use of an efflux pump inhibitor while testing susceptibility defines the
contribution of efflux pump overexpression to the overall resistance of the bacterial strain. The
efflux pump overexpressed (EPO) phenotype was defined previously (Lomovskaya et al., 2001)
as a decrease in MIC > 8 fold when testing in the presence of an efflux pump inhibitor.
Most FQ-resistant strains (92%, 133/145) exhibited the EPO phenotype; only 8 strains
had < 8-fold decrease in MIC. A similar proportion of exoU strains exhibited the EPO phenotype
compared to exoS strains (97%, 63/65 vs. 91%, 68/75 respectively; p =0.13). The addition of an
efflux pump inhibitor reduced the MIC in 59% (n=145) of FQ-resistant strains to “susceptible”
range at MIC < 4 µg/ml overall; including 55% (n=65) of the exoU strains and 71% (n=75) of
exoS FQ-resistant strains.
Target Site Mutations
The QRDR of the 4 target site genes (gyrA, gyrB, parC, and parE) was amplified by PCR
and sequenced for resistance-conferring target site mutations (TSMs) in a subset of 108 isolates;
59 (55%) of which were exoU and 49 (45%) were exoS isolates. Table 2 summarizes all 108
isolates that have been sequenced for target site mutations.
Table 2. Target Site Mutations in exoU and exoS strains
Target Protein Amino acid substitution exoU n=59 (%) exoS n=49 (%)
No TSM 9/59 (15) 8/49 (16)
Single TSM 19/59 (32) 22/49 (45)
GyrA 17/19 (89) 19/22 (86)
Thre83Ile 14 16
Asp87Asn 3 2

29

Lys65Arg 0 1
GyrB 1/19 (5) 1/22 (4)
Glu468Asp 1 0
Ser266Tyr 0 1
ParC 1/19 (5) 1/22 (4)
Ser87Leu 1 1
ParE 1/19 (5) 1/22 (4)
Asp419Asn 0 1
Combined TSMs 31/59 (53) 19/49 (39)
GyrA 1/31 (3) 1/19 (5)
Thr83Ile+Asp87Glu 1 0
Thr83Ile+Asp87Tyr 0 1
GyrA+GyrB 1/31 (3) 2/19 (10)
Thr83Ile+Glu468Asp 1 2
GyrA+ParC 27/31 (87)
a
11/19 (58)
a

Thr83Ile + Ser87Leu 19/27 (70) 3/11 (27)
Thr83Ile + Glu91Lys 5/27 (19) 8/11(72)
Thr83Ile,Asp87Asn + Ser87Leu 3/27(11) 0
GyrA + ParE 1/31 (3)
b
4/19 (21)
b

Thr83Ile+ Ile463Phe 1 4
GyrA+ParC+ParE 1/31 (3) 1/19 (5)
Thr83Ile+ Ser87Leu+ Ile463Phe
1 1

In general, MIC increased as the number of active mutations increased. Out of all the
isolates with at least one target site mutation, only 4 had levofloxacin MIC < 4 µg/ml. A total of
5 exoU and 8 exoS strains had no target site mutations; all showed the EPO phenotype. Among
the 91 isolates with active mutations, a trend towards more single TSM was observed in exoS

30

strains (54%, 22/41 vs 38%, 19/50; p=0.135) while more exoU strains had combined TSMs
(62%, 31/50 vs 46%, 19/41; p=0.135). Interestingly, exoU strains were more likely to acquire 2
or more TSMs than exoS strains at lower MICs (≤ 8 µg/ml), 13% (4/31) vs. none (Figure 10).
Figure 10. Cumulative
Frequency
of
Isolates
with
≥
2
target
site
mutations
at
each
given
MIC

comparing
exoU- vs exoS-containing isolates.
The proportion of isolates that acquired at least two target
site mutations is greater in exoU than exoS isolates and increases in a linear fashion as MIC increases,
with exoU isolates starting at lower MICs. At a levofloxacin MIC ≤ 8 ug/ml, 25% of exoU isolates
compared to 0% of exoS isolates have 2 or more TSMs whereas for isolates with MICs ranging from 16-
32 ug/ml, 36% exoU isolates compared to 22% of exoS isolates, and for isolates with MICs ranging from
64-128 ug/ml, 52% exoU isolates compared to 39% of exoS isolates have 2 or more TSMs.

Substitution of isoleucine for threonine at position 83 in gyrA was the most commonly
noted mutation among those with active mutations (88%, 80/91) and in similar proportion
between exoU and exoS isolates (90%, 45/50 vs. 85%, 35/41, respectively). Of those with only
this mutation at position 83 of gyrA, all except one had levofloxacin MIC ≥ 8 µg/ml. Even after
addition of the efflux pump inhibitor (EPI), about 37% (11/30) of those isolates remain resistant,
with MIC ranging from 4 to 16 µg/ml. A total of 5 isolates have a different gyrA amino acid
substitution, asparagine for aspartic acid at position 87. This mutation contributed only
moderately to resistance in strains with combined TSM and EPO phenotype; levofloxacin MICs
ranged from 1-16 µg/ml in those isolates, which lowered to ≤ 0.5 µg/ml after addition of EPI. In
total, only 2 isolates each had any active mutations in gyrB, parC, or parE without a concurrent
0
10
20
30
40
50
60
70
80
90
100
≤ 8 16-32 64-128
Cumulative % Isolates
MIC (µg/ml)
ExoU
ExoS

31

gyrA mutation; the corresponding levofloxacin MICs after addition of EPI for those isolates
were: gyrB (2, 0.25 µg/ml), parC (8 µg/ml in both), and parE (2 µg/ml in both).
Notable differences in specific TSMs (single or combined) were observed between exoU
and exoS isolates. Mutations unique to the exoU strains included a double amino acid change in
GyrA at positions 83 and 87, a single substitution at position 468 in GyrB, and a triple mutant
with two GyrA changes (positions 83 and 87) and a ParC substitution at position 87. Mutations
unique to the exoS isolates result in a change from lysine to arginine at position 65 of GyrA, and
a double gyrA mutant similar to the one found in the exoU strains at positions 83 and 87. One
exoS isolate also carried a unique mutation in gyrB which resulted in a substitution of tyrosine
for serine at position 466. Unlike any exoU isolate in this study, one exoS+ isolate had a single
mutation in parE only, which resulted in a levofloxacin MIC of 16 µg/ml and 2 µg/ml in the
presence of EPI.
Interestingly, among isolates with combined TSMs, mutation in gyrA is present in all
isolates, but the proportion of exoU and exoS strains differ significantly in the subunit of
topoisomerase IV affected. Strains containing the exoU gene were twice more likely to have
combined mutations in both gyrA and parC than exoS strains (48%, 28/59 vs. 24%, 12/49,
p=0.0439, OR 2.369 (95%CI: 1.014-5.536). On the other hand, exoS strains were more likely to
have combined mutations in gyrA and parE.
Many silent mutations were also found in the target genes; gyrB and parE were the most
affected, with 84% of all the isolates carrying one or more silent mutations in gyrB and 86%
carrying one or more in parE. Silent mutations occur more frequently in the gyrA gene among
exoS than exoU isolates (74%, 36/49 vs. 20%, 12/59; p<.0001). On the contrary, significantly
more exoU isolates have silent mutations in the parE gene than exoS isolates (93%, 55/59 vs.

32

78%, 38/49. p=0.019). The presence or number of silent mutations did not correlate with
levofloxacin MIC.
Discussion
Clinical isolates of Pa carrying the gene encoding for either the ExoU or the ExoS TTSS
effector proteins have been documented to differentially impact virulence in pneumonia as well
as in the development of antimicrobial resistance. Compared to exoS strains, exoU strains have
been shown to have higher cytotoxicity correlated with increased resistance, in both in vitro
(Lakkis and Fleiszig, 2001; Wong-Beringer et al., 2008) as well as in vivo murine models of
infection (Allewelt et al., 2000; Schulert et al., 2003). In a study of population structure P.
aeruginosa isolates in Europe, the exoU genotype was found to be significantly associated with
multidrug resistance and ciprofloxacin resistance compared to the exoS genotype (Maatallah et
al., 2011). This study sought to explore the observed link between TTSS virulence and
fluoroquinolone resistance in Pa. We showed that exoU and exoS strains differentially acquire
resistance-conferring mutations and that exoU strains may be genetically favored in a
fluoroquinolone-rich environment that stemmed from heavy prescribing.
Pa develops resistance to the fluoroquinolones primarily from acquiring target site
mutations or overexpression of multidrug efflux pumps from the Resistance-Nodulation-Division
(RND) family (Poole, 2011). Confirming our previously published findings, we have shown in
this study that the fluoroquinolone-resistant subpopulation is predominated by exoU strains in a
large sample of respiratory isolates, despite the higher prevalence of exoS strains overall. This
may be related to our observation that exoU strains more readily acquire multiple target site
mutations and at a lower MIC than exoS strains. Overexpression of multidrug efflux pumps
belonging to the RND family of proteins results from mutations in genes that encode for the

33

pumps or in genes that regulate their expression (Poole, 2011). Here, we utilized an efflux pump
inhibitor to assess the extent of efflux pump involvement in resistance by comparing the MICs of
levofloxacin before and after the addition of efflux pump inhibitor. Efflux pump-overexpression
significantly contributed to FQ-resistance in both exoU and exoS-containing isolates.
With respect to target site mutations, those observed in our isolates are consistent with
previously published mutations (Akasaka et al., 2001; Higgins et al., 2003; Jalal and Wretlind,
1998; Lee et al., 2005a; Nakano et al., 1997). Substitution of isoleucine for threonine at position
83 in GyrA is the most frequent TSM in our cohort of FQ-resistant isolates. Higher MICs are
associated with an accumulation of mutations, the most common being a double mutation in
gyrA and parC. Others have found strains with a double gyrA-parC mutation to be 3-4 times
more resistant to ciprofloxacin than strains with a single mutation in gyrA (Mouneimné et al.,
1999). Similar results were found in Salmonella sp. where mutations in gyrA and parC as well
as double mutations in gyrA conferred high level resistance (Ling et al., 2003). Our results
provide further support for this stepwise accumulation of mutations mechanism (Jalal and
Wretlind, 1998); only 2 resistant isolates have single mutations in parC without the “first step”
threonine 83 mutation in gyrA. This shows that although GyrA is the main target of
fluoroquinolones, a mutation in gyrA is not a strict prerequisite for the acquisition of mutations at
other target genes leading to resistance. Despite previous results describing single mutations in
the gyrB and parE genes not conferring much resistance (Akasaka et al., 2001), we found
relatively high levofloxacin MICs ranging from 8-32 µg/ml (0.25-2 µg/ml in the presence of an
EPI) in 2 isolates with a single gyrB mutation and one with a single parE mutation.
One striking difference observed between TTSS genotype in FQ resistance relates to the
presence of specific active target site mutations conferring resistance. Isolates containing the

34

exoU gene were twice as likely to have mutations in both gyrA and parC than exoS isolates.
Furthermore, the type of mutations in the parC gene differed with respect to TTSS effector
genotype. The serine to leucine change at amino acid 87 in ParC was more likely to occur in
exoU strains, while the glutamate to lysine change at position 91 was more likely to occur in
exoS strains. This observed difference may point to a differential adaptive response of exoU and
exoS strains to fluoroquinolone exposure.
Synonymous mutations have also been shown to have an effect on gene expression.
When a library of green fluorescent protein genes that differed only in codon usage but not in the
amino acid sequence was expressed, the levels of expression varied among the polymorphic
genes, indicating that cellular processes can be affected without changing the amino acid
sequence (Kudla et al., 2009). This is likely a result of a change in the mRNA secondary
structure. Interestingly, we observed in this study differences in the prevalence and type of silent
mutations that occurred in the QRDRs of the four target genes by TTSS effector genotype. The
gyrB and parE genes had the highest number of silent mutations possibly because they are not
the main contributors to enzymatic action and therefore can tolerate the polymorphisms. exoS-
containing strains have a significantly higher rate of silent mutations in gyrA than exoU strains
while the opposite is observed for the exoU strains with the parE gene; both TTSS effector
genotypes have similar rates of silent mutations in gyrB and parC.
The results presented in this study show that resistance-conferring mutations developed
differently in exoU and exoS strains. We speculate that the mechanism for this difference may
involve the SOS response; a global response that allows bacteria to tolerate sudden increases in
DNA damage. Fluoroquinolones induce the SOS response by poisoning the topoisomerase
enzymes, leading to double stranded DNA breaks. Sub-MIC levels of ciprofloxacin have been

35

shown to rapidly induce the SOS response in S. typhimurium (Wang et al., 2010). A subset of
genes induced by the SOS response include the error prone polymerases (Pol II, Pol IV, and
PolV) (Cirz et al., 2006). These low-fidelity polymerases are able to perform translesion
synthesis, enabling stalled replication forks to continue past DNA damage, but at the expense of
introducing errors into the genome. By preventing induction of the SOS response in vivo and in
vitro, pathogenic E. coli were unable to acquire resistance mutations (Cirz et al., 2005). This
same study also showed that if any of the SOS-induced polymerases were made non-functional,
the mutation rate was much less than in wild type. Relating those results to our findings leads us
to question whether a difference in the SOS response in exoU and exoS strains could account for
the differences in target site mutations observed.
This study demonstrated evidence of differentiation in quinolone resistance by virulence
genotype in a large sample of respiratory isolates of Pa. Importantly, negative outcomes reported
in the literature in association with fluoroquinolone resistance in Pseudomonas infections may be
attributable in part to the shift in population dynamics in which resistance and virulence traits co-
evolved, favoring the more virulent exoU strains. The current practice of widespread
fluoroquinolone prescribing has far reaching negative consequences beyond antimicrobial
resistance.

Study #2: Risk of Developing Pneumonia Is Enhanced by the Combined Traits of
Fluoroquinolone Resistance and Type III Secretion Virulence in Respiratory isolates of
Pseudomonas aeruginosa. Sullivan, E., Bensman, J., Agnello M., et al. Critical Care
Medicine, 2014.

The previously discussed study published in PloS ONE investigated the relationship
between genotype (exoU vs exoS) and FQ-resistance and found that exoU strains were more

36

likely to be FQ-resistant, and also more likely to acquire multiple target site mutations leading to
resistance. The next study, published in Critical Care Medicine, focused on the clinical
consequences of exoU vs exoS genotype and FQ-resistance in patients.
This study investigated the invasiveness of different strains of Pa; levels of invasiveness
can range from merely colonizing the airways, to causing bronchitis, and most severely, causing
pneumonia. The goal of the study was to identify factors contributing to the invasiveness in
patients, focusing on TTSS genotype (exoU vs exoS) and FQ-resistance. An adult cohort with a
range of respiratory syndromes caused by Pa was examined to identify the microbial
characteristics that may be associated with invasiveness of disease.
Materials and Methods
Study Design
The microbiology laboratory database was used to identify adult patients hospitalized
between January 2005 and January 2010 who had growth of Pa from a monomicrobial or
polymicrobial respiratory specimen; for those with multiple admissions during the study period
with positive respiratory culture for Pa, only the first admission was included. Patients whose
medical charts were incomplete or unavailable or had a diagnosis of cystic fibrosis or
bronchiectasis were excluded from the study.
Antimicrobial susceptibility testing was performed by broth microdilution (Vitek,
bioMerieux). The agents tested included piperacillin-tazobactam, ceftazidime, cefepime,
ciprofloxacin, imipenem, gentamicin, tobramycin, and amikacin. Isolates with minimum
inhibitory concentration (MIC) corresponding to intermediately susceptible or resistant
breakpoints were considered “resistant.” Specifically, FQ-R included strains with levofloxacin
MIC greater than or equal to 4 mg/L. Multidrug resistance (MDR) was defined as resistance to at

37

least one agent in each of the three or more classes of antipseudomonal agents, with the
following defined as distinct classes: piperacillin-tazobactam; ceftazidime or cefepime;
imipenem or meropenem; aztreonam; ciprofloxacin or levofloxacin; gentamicin, tobramycin, or
amikacin (Magiorakos et al., 2012).
Bacterial isolates were cultured overnight in 200 µL of Luria broth in 96-well plates at
37ºC without shaking, followed by dilution to either 1:10 or 1:100 in H2O. The presence of
TTSS genes exoU, exoS, and pcrV was detected by polymerase chain reaction (PCR) using
previously published primers (Ajayi et al., 2003). Products were visualized following
electrophoresis on agarose 3% w⁄v gels stained with ethidium bromide and viewed under
ultraviolet light. Clonality of the bacterial isolates was assessed by random amplified
polymorphic DNA (RAPD)-PCR as described in the previous Materials and Methods section.
Data Collection
Medical and laboratory records were reviewed for pertinent demographic, laboratory,
radiographic, and clinical information. A standardized data collection form was used to record
the following information extracted from patients’ medical charts and entered into a database
management program (Microsoft Access; Microsoft, Redmond, WA): age, gender, comorbid
conditions, residence prior to admission, place of Pa acquisition, Acute Physiology and Chronic
Health Evaluation (APACHE) II score calculated based on the worst values within 24 hours of
admission, need for enteral feeding or mechanical ventilation, recent non-Pa pneumonia (within
30 days of positive Pa culture), admission to ICU, and death. Comorbid conditions included
diabetes mellitus, cardiovascular disease, cerebrovascular disease, pulmonary disease (chronic
obstructive pulmonary disease, asthma), chronic tracheostomy, renal insufficiency (serum
creatinine > 2 mg/dL or 34.2 mmol/L), hepatic dysfunction or known liver disease, corticosteroid

38

use (prednisolone ≥ 20 mg/d or equivalent), and malignancy. Details of antimicrobial therapy
(agent, dose, and duration) prior to and after isolation of Pa were recorded.
Study Definitions
Colonization was defined as the absence of respiratory signs and symptoms of infection
and without receipt of antimicrobial therapy active against Pa following isolation. Bronchitis
was defined as presence of respiratory signs and symptoms (e.g., cough, sputum production,
dyspnea) with negative findings on chest radiograph (Horan et al., 2008). Ventilator-associated
tracheobronchitis (VAT) was defined as the presence of respiratory signs and symptoms with
radiographic findings consistent with atelectasis, acute respiratory distress syndrome and
congestive heart failure without new infiltrate in those who have been mechanically ventilated
for at least 48 hours (Craven et al., 2011). Pneumonia was defined according to Centers for
Disease Control and Prevention criteria (Horan et al., 2008) as radiographic evidence consistent
with infection and with clinical signs and symptoms of respiratory infection. Hospital-acquired
pneumonia (HAP) was defined according to American Thoracic Society (ATS)/Infectious
Diseases Society of America (IDSA) criteria as pneumonia that occurs 48 hours after admission
(22). VAP was defined according to ATS/IDSA criteria as pneumonia that developed more than
48 hours after endotracheal intubation (Society and America, 2005).
Data Analysis
Patients were grouped according to their respiratory syndrome: colonization, bronchitis,
or pneumonia for comparison of host and microbial (FQ-R, MDR, and TTSS effector genotype)
characteristics. Continuous variables were compared using Kruskal-Wallis test, whereas
categorical variables were compared using chi-square test or Fisher exact test where applicable.
Host and microbial factors which differed between groups (p < 0.2) on univariate analysis were

39

modeled in a multivariate logistic regression analysis (modeling pneumonia vs no pneumonia)
controlling for age and APACHE II score to identify independent predictors of pneumonia. After
forcing in age and APACHE II score, additional predictors were selected by forward and
backward selection procedure, with p-values of 0.15 entry and 0.1 removal, respectively. All
statistical tests were two tailed; p-value of less than 0.05 was considered significant. Statistical
analyses were performed using Graph-Pad Prism version 5.0 (San Diego, CA) and SAS version
9.2 (Cary, NC).
Results
Over a 5-year study period, a total of 403 adult hospitalized patients with Pa positive
culture from a respiratory specimen were screened. A total of 218 patients met inclusion criteria;
of those, 117 patients (54%) had pneumonia, 70 patients (32%) had bronchitis, and 31 patients
(14%) were colonized. Reasons for exclusion were diagnosis of cystic fibrosis or bronchiectasis
and unavailable medical records or chest radiograph.
Host Characteristics
Patient characteristics are shown in Table 3. Patients with pneumonia (n = 117) were
older (p < 0.027), more likely to have been admitted from a long-term care facility (p < 0.0001),
had nosocomial or healthcare-associated acquisition of Pa (p < 0.0001), and had higher
APACHE II scores (p < 0.0001) than patients in bronchitis or colonization groups. Of note, 31%
of patients (22 of 70) in the bronchitis group had VAT. Among those with pneumonia, VAP
developed in 38% of patients, whereas the remainder had HAP (29%), healthcare-associated
pneumonia (26%), and community-acquired pneumonia (7%). As expected, overall in-hospital
mortality rate was highest for patients with pneumonia than bronchitis or colonization (35% vs
11% vs 0%, respectively; p = 0.0003).

40

Comorbid conditions significantly associated with the development of pneumonia rather
than bronchitis or colonization were cerebrovascular disease with a documented risk of
aspiration (p = 0.005), malignancy (p = 0.016), and cardiovascular disease (p = 0.025). Study
groups were also compared with respect to the presence of risk factors for pneumonia and
antibiotic exposure prior to isolation of Pa. Overall length of stay and duration of ICU stay did
not differ between groups. However, a significantly greater proportion of patients with
pneumonia compared with the other study groups had the following risk factors: presence of
enteral feeding tube (p=0.0028), mechanical ventilation within 5 days (p=0.029) and for a longer
duration (p=0.044), and recent history of non-Pa pneumonia within 30 days (p=0.001). The
proportion of patients exposed to antibiotics within 5 days prior to isolation of Pa was similar
between groups (68–74%), and the median duration of antibiotic therapy in days was 1.5, 1, and
2 for the colonization, bronchitis, and pneumonia groups, respectively (p=0.84). Notably, a
significantly greater number of patients with pneumonia received an anti-pseudomonal
containing regimen than those in the bronchitis or colonizing groups (40% vs 34% vs 3%, p <
0.0001) with a significantly longer duration of anti-pseudomonal therapy between the groups
(p=0.0014).
Table 3. Patient Characteristics Group by Disease Severity
Variables Colonized
n=31
Bronchitis
n=70
Pneumonia
n=117
P-value
Age, yr (median, IQR) 74 (52, 82) 70 (55, 78) 78 (65, 83) 0.027
Male, n (%) 14 (45) 35 (50) 73 (62) 0.11
Residence, n (%)
Home 23 (74) 36 (51) 47 (40) 0.0029
Skilled nursing
facility/long-term care
5 (16) 29 (41) 69 (59) <0.0001
APACHE II* (median,
IQR)
9 (4, 13) 13 (9, 17) 18 (13, 24) <0.0001
Place of Pa acquisition, n (%)
Community 13 (42) 10 (14) 10 (9) <0.0001

41

Healthcare 4 (13) 18 (26) 41 (35) 0.042
Nosocomial 14 (45) 42 (60) 66 (56) 0.38
Healthcare + Nosocomial 18 (58) 60 (86) 107 (92) <0.0001
Ventilator-associated
pneumonia, n (%)
NA NA 45 (38) NA
Co-morbidities, n (%)
Diabetes mellitus 6 (19) 26 (37) 48 (41) 0.084
Cardiovascular disease 22 (71) 47 (67) 98 (84) 0.025
Cerebrovascular disease 3 (10) 7 (10) 32 (27) 0.005
Pulmonary disease 14 (45) 30 (43) 37 (32) 0.19
Chronic tracheostomy 2 (6) 8 (11) 11 (9) 0.73
Renal insufficiency 3 (10) 10 (14) 25 (21) 0.22
Hepatic dysfunction 0 2 (3) 2 (2) 0.63
Corticosteroid use 4 (13) 1 (1) 7 (6) 0.062
Malignancy 0 8 (11) 21 (18) 0.016
In-hopsital mortality, n (%) 0 8 (11) 41 (35) 0.0003
*Acute Physiology and Chronic Health Evaluation II
Microbial Characteristics
Clonality was determined for most strains (n = 186, 85%). Strains with identical band
patterns were considered related and assigned a unique random amplified polymorphic type.
Overall, 102 unique RAPD types were identified. The majority of RAPD types contained only
one strain (n = 62, 61%). The 38 FQ-R exoU strains tested belonged to 28 clonal groups.
Similarly, the 35 FQ-sensitive exoU strains assessed belonged to another set of 28 clonal groups.
Among the exoS isolates, the 38 FQ-R strains and 45 FQ-sensitive strains analyzed belonged to
30 and 33 groups, respectively. No significant differences in clonal distributions were found
between FQ-R and FQ-sensitive strains with 72% (42 of 58) versus 66% (33 of 50) types
containing only one strain (p = 0.53) or among strains that caused pneumonia, bronchitis, or
colonization (73% [50 of 69] vs 85% [35 of 41] vs 88% [21 of 24], p = 0.14).
When antimicrobial susceptibility was compared between isolates obtained from patients
with different respiratory syndromes, a significantly greater proportion of FQ-R (57% vs 34% vs
16%, p < 0.0001) and MDR strains (36% vs 26% vs 7%, p = 0.0045) caused pneumonia than

42

bronchitis or colonization, respectively (Figure 11). When the susceptibility pattern was analyzed
for other antipseudomonal agents, patients with pneumonia were more likely to also harbor
cephalosporin-resistant strains (53% vs 39% vs 13%, p = 0.0001) and aminoglycoside-resistant
strains (26% vs 20% vs 3%, p = 0.014) compared with bronchitis or colonized groups. FQ-R
overlapped in 50–79% of the cephalosporin-resistant strains and 79–100% of the
aminoglycoside-resistant strains when separated by respiratory syndromes.
Figure 11. Severity of Pa infection caused by strains with different susceptibilities. The proportion of
patients with FQ-S strains was highest in the colonizing group, while the proportion of FQ-R and MDR
strains increased with increasing severity.

Presence of co-pathogens was significantly different between the three groups: 26% (8 of
31), 51% (36 of 70), and 65% (76 of 117) of colonized, bronchitis, and pneumonia patients,
respectively, (p = 0.0004) had growth of at least one other pathogen from their respiratory
specimens. Methicillin-resistant Staphylococcus aureus was the co-pathogen most frequently
isolated in 6% (2 of 31), 14% (10 of 70), and 25% (29 of 117) (p = 0.034) followed by Candida
species in 0% (0 of 31), 11% (8 of 70), and 13% (15 of 117) (p = 0.11) of colonized, bronchitis,
and pneumonia patients, respectively.

Of all the Pa isolates tested for TTSS effector genotype by PCR analysis, 208 isolates
0
10
20
30
40
50
60
70
80
90
100
colonization bronchitis pneumonia
Percent
of
Pa=ents

FQ-S
FQ-R
MDR

43

(95%) harbored the pcrV gene encoding a component of the secretion apparatus essential for
translocation of the TTSS effectors (e.g., ExoU, ExoS) into host cells. Genes encoding either the
ExoU or ExoS effector were present in 94% (n = 196) of the pcrV+ isolates. The presence of the
exoU gene in Pa isolates was similarly distributed among our study groups (pneumonia,
bronchitis, colonization) (44% [52 of 117] vs 30% [21 of 70] vs 48% [15 of 31], respectively; p=
0.093). Interestingly, when both FQ-R and TTSS virulence characteristics were considered
together, FQ-R strains that harbored the exoU gene were significantly more likely to cause
pneumonia than bronchitis (32% [38 of 117] vs 9% [6 of 70], p < 0.0001), whereas FQ-R strains
containing the exoS gene were equally likely to cause pneumonia or bronchitis (21% [25 of 117]
vs 23% [16 of 70], p = 0.81) (Figure 12).
Host and microbial factors that differed between groups on univariate analysis were
modeled in a multivariate logistic regression analysis to identify significant predictors for the
development of pneumonia. After controlling for age and APACHE II score, we found that the
presence of a Pa strain characterized by FQ-R and TTSS effector genotype exoU to have the
highest odds for development of pneumonia (odds ratio [OR], 3.82 [95% CI, 1.58–9.25]; p =
0.003) followed by pneumonia within 30 days prior to isolation of Pa (OR, 3.25 [95% CI, 1.10–
9.61]; p = 0.033). In addition, we retained the variable FQ-R, exoS genotype in the model for
comparison where it was found to be non-significant (OR in predicting the development of
pneumonia) (Table 4).

44

Figure 12. Combined genotype and FQ susceptibility traits in each clinical group. exoU, FQ-R
strains predominate in the pneumonia group.
*No exoS, FQ-R strains in the colonized group

Table 4. Multivariable Logistic Regression: Variables predictive of development of pneumonia
Variables OR (95% CI) p-value
Age (yr) 1.02 (0.99–1.04) 0.096
APACHE II 1.13 (1.07–1.19) < 0.0001
Non-Pa pneumonia within
30 days prior to initial
isolation
3.25 (1.10–9.61) 0.033
FQ-resistant, exoU genotype 3.82 (1.58–9.25) 0.003
FQ-resistant, exoS genotype

1.96 (0.87–4.42) 0.11

Discussion
Infections with Pa can range in severity from merely colonizing the airways, to
bronchitis, and most severely, pneumonia. This study linked the previous findings that adverse
outcomes were associated with FQ-R and that exoU strains were more likely to be FQ-R with the
novel finding that both traits together are predictive of disease severity.
0
5
10
15
20
25
30
35
40
45
50
colonization bronchitis pneumonia
Percent of Patients
exoS, FQ-S
exoS, FQ-R
exoU, FQ-S
exoU, FQ-R

45

The results of the univariate analysis were consistent with other findings, in that the need
for mechanical ventilation (Bonten et al., 1996) and feeding tube use (von Baum et al., 2010) are
risk factors for the development of pneumonia. In addition, patients with recent non-Pa
pneumonia and those who had a history of cerebrovascular accident with documented risk for
aspiration were also more likely to develop pneumonia rather than bronchitis. Therefore, it
appears that recent lung injury may allow Pa to establish in the airways and eventually lead to
pneumonia. We also confirmed prior results of a high frequency of cross-resistance in FQ-R
isolates. In this study, 64% of FQ-R strains were also multi-drug resistant (resistant to > 3
categories of drugs).
This was the first study to demonstrate both resistance and virulence impacting disease
severity in Pa respiratory isolates. exoU alone was not a significant risk factor for the
development of pneumonia; only the combined FQ-R, exoU phenotype predicted pneumonia.
The combined exoS, FQ-R phenotype did not predict pneumonia. Although FQ-R strains were
significantly more likely to cause infection (bronchitis or pneumonia) than colonization, presence
of the exoU gene in combination with the FQ-R phenotype had the highest odds for the
development of pneumonia in a multivariate analysis, after controlling for age and severity of
underlying disease.
The finding that Pa strains with FQ-R and exoU are critical for the development of
pneumonia, particularly in those with impaired lung epithelium, has important implications on
diagnostics, therapeutics, and antibiotic stewardship. Rapid diagnostics to identify the genotype
of Pa isolated from patient airways would allow for more aggressive intervention for exoU, FQ-
R strains. Also, Therapeutics to protect or repair injured lung epithelium could prove beneficial

46

in preventing Pa invasion in susceptible hosts, such as those with recent history of pneumonia or
those requiring mechanical ventilation.
Summary
Clinical studies such as these are essential for documenting the characteristics Pa strains
in the host and assessing the impact on patients. Based on the findings in both of these studies,
exoU strains are genotypically favored in acquiring an FQ-R phenotype in the clinical setting.
Overall, these studies show that not only are exoU strains more likely to develop resistance as
well as the mutations leading to FQ-resistance, but that these co-selected traits have significant
negative outcome on patients. This effect on severity of disease in patients necessitates further
investigation that can provide an explanation as to why exoU strains more frequently acquire the
FQ-R phenotype.
Citations for the studies described in this chapter:
1. Agnello, M., Wong-Beringer, A. Differentiation in Quinolone Resistance by Virulence Genotype in
Pseudomonas aeruginosa. PLoS One, 2012; 7:e42973.
2. Sullivan, E., Bensman, J., Lou, M., Agnello, M., Shriner, K., Wong-Beringer, A. Risk of Developing
Pneumonia is Enhanced by the Combined Traits of Fluoroquinolone Resistance and Type III Secretion
Virulence in Respiratory Isolates of Pseudomonas aeruginosa. Critical Care Medicine 2014; 42(1):48-56.

47

Chapter 3: Effects of FQ-Resistance on the Fitness of exoU and exoS Clinical Strains

Introduction
The clinical evidence presented in the preceding chapter suggests that exoU strains of Pa
more frequently become FQ-resistant and cause severe disease in patients compared to exoS
strains. We hypothesize that exoU strains are more adaptable to FQ- exposure than exoS strains,
allowing for the development of FQ-resistance and the ability to maintain substantial
pathogenicity without a fitness cost of resistance. In addition, we hypothesize that exoU strains
have an increased overall biological adaptability to FQ-exposure that may include altered
metabolic, genetic, and regulatory functions.
Biological adaptability can also be defined as biological fitness. In the Darwinian sense,
fitness is the ability of an organism to survive and reproduce, therefore passing on its genes. In
the case of pathogens, fitness is the ability of the organism not only to survive and reproduce, but
also to adapt to the host environment, cause disease, and propagate to the next host. In
pathogenic bacteria, the capacity for change in the face of selective pressures exerted by the host
environment determines overall fitness (Hacker and Carniel, 2001). We hypothesize that Pa
strains with the exoU-specific strain background have increased fitness compared to exoS
strains, which allows them to better adapt to the selective pressures exerted during FQ-exposure
and predominate among FQ-resistant subpopulations in the clinical environment.
In general, antibiotic resistance is mediated in one of two ways: through the acquisition
of a specific gene whose product degrades or otherwise renders the drug incapable of killing the
bacterium, or through the generation of mutations in genes that alter the target site of the
antibiotic. Antibiotic resistance is generally thought to be accompanied by a fitness cost to the

48

bacterium in an antibiotic-free environment due to a genetic burden from the resistance
mechanism (Andersson and Levin, 1999). The costs associated with the acquisition of genetic
elements include the cost of maintaining and replicating the elements themselves; therefore, this
mode of resistance probably confers a greater fitness cost than the generation of mutations in
target genes. For the purposes of this project, we will be focusing solely on the fitness costs
associated with target site mutations.
A recent meta-analysis of studies conducted on fitness costs of resistance to a variety of
antibiotic classes in different organisms showed that most point mutations are generally costly
(Melnyk et al., 2014). However, there is great variability depending on organism, drug, and
mechanism of resistance. For example, rifampicin resistance can either decrease or increase
fitness, depending on the environment. In aging bacterial colonies of E. coli and Salmonella
enterica, rifampicin-resistant mutants have increased fitness (Wrande et al., 2008), but in
Mycobacterium tuberculosis, all rifampicin mutants had a fitness defect (Billington et al., 1999).
A study of rifampicin resistance in Pa showed that subsequent rifampicin resistance mutations
act to buffer the high fitness costs of single mutations (Hall and MacLean, 2011), suggesting that
fitness effects are complex and multi-factorial.
Resistance to the fluoroquinolones can occur via mutations in the regulatory genes
controlling efflux pump expression, leading to overexpression and efflux of the FQs, or through
target site mutations. Efflux pump overexpression has been shown to cause a decrease in fitness
in Pa (Abdelraouf et al., 2011); however, a separate study found that the more noteworthy effect
of efflux overexpression (specifically MexEF-OprN) was that it altered the Pa transcriptome
(Olivares et al., 2012) and led to physiologic changes in the cell, such as reduced expression of
several quorum-sensing genes.

49

Target site mutations that lead to FQ resistance occur in the topoisomerase genes gyrA/B
and parC/E, in a region labeled the ‘quinolone resistance determining region’ (QRDR).
Relatively high-level FQ-resistance can occur through one or two point mutations that lead to
single amino acid substitutions. The topoisomerase enzymes are essential for the topological
maintenance of the genome by adjusting supercoiling during replication and decatenating the
daughter replicons. Mutations that alter the structure of these enzymes are assumed to have a
fitness cost to the cell. However, studies done in a variety of organisms have shown variable
fitness costs from common FQ-resistance conferring mutations. A study of the fitness cost
associated with a common FQ-resistance gyrA mutation in Camplyobacter jejuni, showed a high
fitness cost for one strain, but led to enhanced fitness in a different C. jejuni strain, suggesting
fitness effects may be dependent on strain background (Luo et al., 2005). Clostridium difficile
also acquires resistance to ciprofloxacin via mutations in gyrA, and very minimal fitness effects
were seen in a variety of gyrA mutants (Wasels et al., 2014). In a study of norfloxacin resistance-
conferring gyrA mutations selected for in a strain of E. coli, there was little to no fitness cost
associated with commonly occurring single point mutations in gyrA, but a more dramatic
decrease in fitness occurred with the acquisition of subsequent point mutations in gyrB and marR
(Komp Lindgren et al., 2005). A recent study of ciprofloxacin-resistant E. coli strains that were
engineered from a parental wild-type strain to reflect many combinations of common gyrA and
parC mutations showed very little fitness effects for gyrA mutants, and actually an increase in
fitness in some combinations of gyrA + parC mutations (Machuca et al., 2014). A similar study
looking at a variety of combinations of gyrA and parC mutations found that for some
combinations, there is a dramatic increase in fitness with multiple mutations (Marcusson et al.,
2009). Similarly in Streptococcus pneumoniae, strains with mutations both in gyrA and parC did

50

not show a decrease in fitness compared to strains with just a gyrA mutation (Balsalobre and de
la Campa, 2008; Rozen et al., 2007). A contradictory effect was found in Neisseria gonorrhoeae,
in which ciprofloxacin resistance is also commonly caused by mutations in gyrA and parC. One
study found that a single mutation in gyrA conferred a fitness benefit, and not much change
occurred with a subsequent mutation in parC, but some combinations of multiple mutations
decreased fitness (Kunz et al., 2012).
In Pa, there have only been a few studies investigating the fitness costs of topoisomerase
mutations. One study investigated in vitro selected mutants of laboratory reference strain PAO1
as well as isolates from patients with cystic fibrosis with mutations in gyrA, gyrB, and parC and
found that single mutations led to a slight reduction in fitness in some strains but not all, and
additional mutations in parC led to a more severe decrease in fitness (Kugelberg et al., 2005).
Overall, the literature suggests that fitness costs of FQ-resistance may depend on the specific
combinations of multiple mutations, and different combinations may affect fitness in different
ways. However, there are no in-depth investigations of FQ-resistant clinical isolates of Pa, and
no published investigations into the potential differential fitness effects specific to exoU and
exoS strain background. Therefore, the studies presented in this dissertation reflect a novel
investigation into the biological differences between strains with the exoU vs. exoS genetic
background that lead to the differential development of resistance observed in clinical isolates.
Evidence from the clinical studies described in the previous chapter suggest that exoU
and exoS strains differ in the clinical setting in terms of development of FQ-resistance. As a
preliminary investigation, we chose 4 clinical isolates based on strain background (exoU vs.
exoS) and FQ-resistance level for an exploratory microarray study to examine global differences

51

in gene expression. The insights gleaned from this study supported further investigations to study
resistance mutations in a controlled genetic background using isogenic mutants.
Results from the studies examining clinical isolates also showed that exoU strains are
more likely than exoS strains to acquire two or more mutations, specifically in gyrA + parC. We
hypothesized that exoU strains are more fit after acquisition of these resistance mutations than
exoS strains, and therefore our goal was to compare the effects of these mutations on fitness in
exoU and exoS strains. To do this, we created isogenic mutants from exoU and exoS clinical
isolates by inserting a point mutation into the parC gene that leads to a SeràLeu substitution at
position 87. This mutation was the most common mutation found in the parC gene in clinical
isolates from our large collection (Agnello and Wong-Beringer, 2012) and has been found to be
common in other studies of clinical FQ-resistant isolates (Akasaka et al., 2001; Higgins et al.,
2003; Jalal and Wretlind, 1998; Lee et al., 2005a; Nakano et al., 1997). Because this mutation in
parC rarely occurs in clinical isolates without a primary, pre-existing mutation in gyrA (Jalal and
Wretlind, 1998), we chose clinical strains that had already naturally acquired a mutation in gyrA
(leading to threonineàisoleucine change at position 83) for mutagenesis in order to mimic FQ-
resistant strains encountered in the clinical setting.
To create the isogenic mutants, we utilized a novel technique called oligonucleotide
recombination. This technique takes advantage of the innate bacterial mechanism of homologous
recombination. Homologous recombination is a highly conserved process that facilitates genetic
exchange between identical or nearly identical DNA molecules (Lovett et al., 2002). Swingle et
al. (2010) demonstrated that bacterial recombination can be achieved experimentally by
introducing high concentrations of single stranded synthetic oligonucleotides via electroporation.
The authors successfully utilized this procedure to introduce site-specific mutations into the

52

genomes of Pseudomonas syringae, Shigella flexneri, Escherichia coli, and Salmonella
typhimurium.
To utilize this technique, we rationally designed a short oligo identical to the parC gene
sequence, save for the specific point mutation of interest, and after introduction via
electroporation, the mutation successfully incorporated into the genome (Figure 13). This
method differs from other laboratory techniques for generating resistant mutants in that we
rationally designed an oligonucleotide containing a point mutation of interest, chosen based on
data from clinical isolates to mimic resistance mutations that were selected from the host
environment, and incorporated it into the genome of our strains via electroporation, instead of the
common method of serially passaging and selecting randomly generated mutations. In vitro
selection does not reflect what occurs in the host and the clinical setting, resulting in the
selection for mutations that are different than those that clinical isolates commonly acquire
(Komp Lindgren et al., 2005; Kugelberg et al., 2005).

Figure 13. Schematic of recombination. The oligonucleotide used was 60 base pairs in length and
identical in sequence to a portion of the wild type parC gene, save for the point mutation indicated. The
strains integrated the oligo into their genomes at the site of similarity within the parC gene.

With oligonucleotide recombination, we have successfully created 6 isogenic, FQ-
resistant strains from 3 exoU and 3 exoS strains (mutants are denoted PC*). Differences in fitness

53

revealed between the parent and mutant strains can be attributed directly to the effects of the
parC mutation.
To investigate the effects of this common target site mutation in FQ-resistant gyrA
mutants, we tested several components of fitness. First, we performed head-to-head competition
experiments of each mutant vs. its respective parent strain in vitro. Competition assays are
classic experiments used to investigate fitness of isogenic mutants; mutants vs. parent strains
must compete for limited resources in the same culture. Next, we investigated how the mutation
affects the ability of the strains to grow on a variety of metabolic substrates using a Metabolic
Phenotype Microarray. This allowed us to observe if there were specific conditions in which the
parC mutation affected the growth ability of exoU and exoS strains. Finally, we investigated the
effect of the parC mutation on virulence via measurement of expression of TTSS-specific genes
using qRT-PCR. The virulence ability of a strain, specifically the expression of virulence genes,
is an important component of fitness due to the high energy demands of a complex system such
as the TTSS. Importantly, the TTSS is the primary cause of injury to the lung during pneumonia,
and therefore it is of great interest to examine whether FQ-resistance affects the expression of
these genes differently in exoU vs. exoS strain backgrounds.
Further, to investigate the potential mechanisms by which the parC mutation may cause
fitness effects, we examined the supercoiling levels of the PC* strains compared to isogenic
parent strains using a luminescence-based reporter assay. The parC gene is one of the
topoisomerase enzymes that are responsible for regulating supercoiling of the bacterial genome,
and a mutation in this gene may affect overall supercoiling levels. Perturbations of the
supercoiling of the genome can effect multiple aspects of the cell due to global changes in gene
expression caused by DNA topology. This has the potential to affect many cellular processes

54

including important metabolic and virulence pathways (Fournier and Klier, 2004; Redgrave et
al., 2014).
We hypothesize that the exoU strains are better able to adapt to the fitness costs
associated with FQ-resistance due to their ability to compensate for the potential costs
associated with resistance. Therefore, we investigated potential compensation by examining
whether fitness of the strains is enhanced after being ‘aged’ through a 7-day competition
experiment. We also examined mutation frequency of the strains as a potential mechanism to
explain compensation; increased mutation frequency would allow for potential beneficial
mutations to arise.
Finally, we investigated the effect of sub-inhibitory levels of fluoroquinolone exposure
on competitive fitness and compensation by collecting aged strains from primary competition
experiments in which 1/8 MIC of levofloxacin was added to the media.
Materials and Methods
Bacterial Strains and Culture Conditions
Strains of Pa used in this study (Table 6) were selected from a collection of isolates
obtained from the respiratory tract of hospitalized patients at Huntington Hospital, Pasadena, CA
and were stored in cryovials at -80ºC in 30% glycerol until ready for use. All strains had been
previously characterized by: the presence of the exoU or exoS gene, resistance to levofloxacin,
presence of mutations in the quinolone- resistance determining regions of gyrA/B and parC/E,
and clonal relatedness by RAPD PCR, as described in the previous chapter, and in (Agnello and
Wong-Beringer, 2012). Strains were routinely grown in LB broth at 37ºC and 250 rpm shaking
for liquid culture, or grown on Pseudomonas Isolation Agar (PIA) plates at 37ºC. Strains were
routinely sub-cultured twice from the frozen stock before use in any experiment.

55

Susceptibility Testing and Sequencing
Susceptibility testing to levofloxacin and rifampicin was performed by broth
microdilution in 2-fold dilutions at concentrations ranging from 0.25 to 128 mg/L according to
guidelines recommended by CLSI (CLSI, 2007). In order to characterize the involvement of the
multidrug Mex efflux pumps to resistance, MIC of levofloxacin was also measured with the
addition of a commercially purchased efflux pump inhibitor (EPI, MC-0228) (Sigma) at 20
µg/ml (Hsu et al., 2005).
For sequencing, genomic DNA was extracted and purified from isolates using the
DNeasy Mini Kit (Qiagen). The quinolone resistance determining regions of target genes gyrA,
gyrB, parC and parE were amplified by PCR using previously published primers and conditions
and sequenced to identify mutations compared to wild-type strain PAO1. (Jalal and Wretlind,
1998; Lee et al., 2005a; Winsor et al., 2011) (Described in Chapter 2, Table 1).
Microarray
Isolates used for microarray were grown in Inducing Media (LB broth plus 5mM EGTA
and 20mM MgCl
2
) to induce expression of the TTSS to late exponential phase, measured by an
OD
600
nm reading of 0.7. Total RNA was extracted using PureZOL reagent as per
manufacturer’s instructions. mRNA was isolated using the RNeasy Mini-Kit (Qiagen). mRNA
was converted to cDNA using iScript cDNA synthesis kit (Biorad, Hercules, CA).
Microarray analysis was performed at the Genomics Core at Children’s Hospital Los
Angeles with the Affymetrix Gene Chip Pseudomonas aeruginosa Genome Array. Analysis was
performed using Partek Genomics Suite.

56

Creation of Mutants
Isolates (3 exoU and 3 exoS) with a pre-existing gyrA FQ-resistance mutation (ThràIle
substitution) were selected for mutagenesis. Single-stranded oligos 22 nucleotides and 60
nucleotides in length were used and were designed based on the parC gene sequence of strain
PAO1 (Winsor et al., 2011) and shown in Table 5. The oligos were identical to the PAO1
sequence from nucleotides 249 to 270 and 230 to 289 respectively, save for the point mutation
TCGàTTG at locations 12 and 31 of the oligos, corresponding to nucleotide 260 in the parC
gene (Figure 13). This point mutation, which is the most common parC mutation observed
among fluoroquinolone-resistant clinical strains, gives rise to the Ser87àLeu amino acid change
in the ParC protein.
For electroporation, a protocol for P. aeruginosa was adapted (Choi et al., 2006). An
overnight culture was diluted 100 fold, incubated with shaking at 37°C and harvested at OD
600
=
0.3-0.5 by centrifugation at 3200 x g. Cells were resuspended twice in 300 mM sucrose at room
temperature. Either the 22-nt or the 60 nt oligo (5-6 µg in 2-3 µl) was added to 40 µl of
electrocompetent cells and transformed by electroporation at 2.5 kV in a .2 cm cuvette using a
Micropulser (Bio-Rad). SOC medium (1 ml) was added immediately and the cells were
outgrown overnight at 37°C on Pseudomonas Isolation Agar plus levofloxacin at concentrations
2, 4, 8, and 16 fold above the original minimum inhibitory concentration (MIC) of the isolates.
For each isolate, a mock experiment was performed as a control in which cells underwent
electroporation and selection without the addition of any oligo.
Growth Rate Measurements
Independent overnight cultures were diluted to OD
600
~ 0.1 and grown at 37 degrees with
shaking in 50 ml LB broth in a 100 ml flask. An aliquot of the culture (150 µl) was sampled

57

every 30 min for 8 hours and turbidity was measured at OD
600
using a Tecan Sunrise microplate
reader (Tecan Group Ltd., Switzerland). Results were reported as an average of at least 3
independent experiments in terms of doubling time in hours.
Table 5. Genetic Elements Used.
Primer Name
PCR/qPCR
Sequence Ref
rpoD Forward: GGGCGAAGAAGGAAATGGTC
Reverse: CAGGTGGCGTAGGTGGAGAA
_
pcrV Forward: CGAGTTCCTGGTGTCGGCCT
Reverse: CGCTTGCCGTCCTGGGTCTG
_
exsA Forward:
GAGAAGCACTACCTCAACGAGTGG
Reverse: AAGGTGGTCAGCCCCATGCC
_
rsmZ Forward: CACGCAACCCCGAAGGAT
Reverse: GTATTACCCCGCCCACTCT
_
rsmY Forward: CGCAGGAAGCGCCAAAGA
Reverse: GCGGGGTTTTGCAGACCT
_
glmS- up CTGTGCGACTGCTGGAGCTGA

Choi 2006
glmS-down GCACATCGGCGACGTGCTCTC

Choi 2006
Tn7-R CACAGCATAACTGGACTGATTTC

Choi 2006
Tn7-L ATTAGCTTACGACGCTACACC

Choi 2006
Oligo for
Recombination
Sequence Ref
PC*- 22 nt GCACGGCGACTTGGCCTGCTAC Agnello 2014
PC*-60 nt
(successful)
TGCTCGGCAAGTTCCACCCGCACGGCGA
CTTGGCCTGCTACGAGGCCATGGTGCTG
ATGG
Agnello 2014
Plasmid Function Ref
pTNS3 Helper plasmid, encodes Tn-7 transposition
pathway
Choi 2008
pUC18T-mini-
Tn7T-Gm-eyfp
Delivery plasmid for Gm
R
marker and YFP Choi 2006
pUC18T-mini-
Tn7T-Gm-ecfp
Delivery plasmid for Gm
R
marker and CFP Choi 2006
TOP10/mini-Tn7-
PA0614 promoter-
Gm-luxCDABE
Delivery plasmid for Gm
R
marker and lux
operon
Choi 2006

58

Insertion of fluorescent tag
Mini-Tn7 vectors developed and generously shared with us by Dr. Schweizer (Choi et al.,
2005; Choi et al., 2008) were utilized to insert cassettes containing the genes encoding YFP or
CFP as well as a cassette containing the lux operon into strains used for competition experiments
and supercoiling experiments. The cassette inserts into the Pa genome at a single location
(attTn7 site downstream of the glmS gene) and contains the gene for the fluorescent protein
under the control of a constitutive promoter or the lux operon under the control of a supercoiling-
sensitive promoter (Moir et al., 2007), as well as a gentamicin resistance marker for selection.
The delivery plasmid was co-electroporated with a helper plasmid (pTNS3) encoding the
necessary transposase function for insertion. The electroporation protocol used for
transformation is described in detail in (Choi and Schweizer, 2006); briefly, strains were made
electro-competent through a series of washes with 300 mM sucrose, subjected to electroporation
as described above, and plated on LB plates containing 30 µg/ml gentamicin for selection. To
confirm insertion had occurred, PCR was performed to amplify the insertion region using
primers in Table 5, and protocol as described previously ((Choi and Schweizer, 2006).
Competition Assays
To investigate the effects on fitness of the parC mutation, mutants were directly
competed against isogenic parent strains in co-culture. Strains were tagged with either CFP or
YFP for differentiation as described above. Strains were grown overnight in LB, and 10
5
CFU of
each strain were co-inoculated into 10 ml LB in a 50 ml flask, and grown with shaking at 37ºC.
Every 24 hours, 10 µl of the culture was transferred to a new flask containing 10 ml fresh LB,
and a sample of the culture was serially diluted then plated on PIA for CFU enumeration.
Colonies of each color were counted using a fluorescent wide-field microscope (Zeiss Axio

59

Zoom.V16). Experiment duration ranged from 4-7 days. The ratio of the number of CFUs of the
mutant: parent strain at each day of the experiment was calculated and used to determine the
relative fitness of the mutant. Experiments were repeated 7 times. Assays were also performed
with the addition of levofloxacin equal to 1/8 the MIC (2 or 0.5 µg/ml). To later investigate
compensation, colonies were selected from plates at 4 and 7 days of each experiment, inoculated
into LB, grown overnight, and saved at -80ºC in 30% glycerol until further experimentation.
Student’s t-test was performed to compare the difference in average CFU counts of mutant vs.
parent strains and aged vs. un-aged strains.
Metabolic Utilization Assays
Phenotype Microarray Plates PM1, PM2A, and PM3B (Biolog Co., Hayward, CA) were
used to analyze utilization of carbon and nitrogen substrates. Each plate consists of 96 wells,
each well containing a different substrate. Experiments were conducted according to the
manufacturer’s instructions. Briefly, colonies from an overnight agar plate were inoculated into
minimal media with ammonia as the sole nitrogen source for the carbon experiments, and
succinate as the sole carbon source for the nitrogen experiments. The solution was adjusted to an
OD
600
of .05-.07, and 100 µl of this inoculate was pipetted into each well of the 96-well plates.
Growth on each substrate was monitored through the presence of a redox active dye in the
inoculating fluid. Development of purple coloration reflected the amount of metabolic
utilization of the specific substrate and was measured via absorption at 600nm at two time
points: immediately following inoculation, and after 24 hours. Initial readings were subtracted
from final readings to obtain a value for growth on each substrate. These values were compared
between the strains to obtain the fold differences reported.

60

qRT-PCR
RNA was collected and cDNA created as described for the microarray experiments.
Primers were created for exsA, pcrV, rsmY, rsmZ, and rpoD based on the P.aeruginosa PAO1
genome (Winsor et al., 2011) (Table 5). The housekeeping gene rpoD was used as a standard for
comparing relative expression levels between the isolates. qPCR was performed with iQ SYBR
SuperMix in the iQ5 Real-Time PCR Detector (Bio-Rad, Hercules, CA). Samples were run in
triplicate, and results confirmed in two independent experiments.
Supercoiling Assay
We adapted a reporter assay to estimate the ability of strains to maintain supercoiling
levels in mutants compared to parent strains by inserting a Tn-7 cassette containing the lux
operon under the control of a supercoiling-sensitive P. aeruginosa promoter from the gene
PA0614 (Moir et al., 2007). The cassette was inserted as described above. Luminescent strains
were grown to mid exponential phase, then diluted 1:4 in LB+ levofloxacin at ¼, ½, and 1x the
measured MIC, and grown in triplicate in a deep well 96-well plate for 7 hours, to mid-to-late
exponential phase. Luminescence was measured using the Envision Multi-Label plate reader
(Perkin-Elmer) as well as OD
600
as described above. Relative Luminescence Units (RLU) were
normalized to OD
600
for all comparisons.
Mutation Frequency Assay
Mutation frequency was estimated using the spontaneous rifampicin resistance method
(Oliver et al., 2000). Strains were grown independently in LB overnight, spun down,
resuspended in MH (Mueller-Hinton) broth, serially diluted, and plated on MH agar with and
without rifampicin at 5x the MIC (ranging from 40-500 µg/ml depending on the strain). The

61

mutation frequency reported is the number of colonies that grew on MH agar + rifampicin
divided by the number of colonies on MH agar alone.
Mutation frequency was also assessed during competition experiments by plating on
PIA+ rifampicin in addition to plating on PIA at each 24-hour interval, and fluorescent colonies
were counted as described above.
Results and Discussion
Microarray
We chose 4 isolates from our collection of clinical strains for an exploratory microarray.
Isolates were chosen based on number and type of target site mutations in the gyrA and parC
genes, as well as the presence of exoU or exoS (Table 6). The goal of the microarray analysis was
to get a broad overview of the differences in gene expression that occur among exoU and exoS
clinical isolates with varying levels of resistance to the fluoroquinolones and with 1 or 2 target
site mutations (TSMs) that are most representative of those commonly encountered among all
clinical isolates.
Since we are interested in the effects of the 2 TSMs (in gyrA and parC) on exoU and exoS
strains, we primarily focused on differential expression in the two exoU strains (2 TSMs vs 1
TSM) and the two exoS strains (2 TSMs vs 1 TSM). A summary of functional groups of genes
differentially expressed is presented in Figure 14. Upregulation is defined as ≥ 2 fold increase in
expression and downregulation as ≥ 2 fold decrease.

62

Figure 14. Functional groupings of differentially expressed genes. Comparisons made between the
strains based on number of target site mutations (1 or 2) separated for exoU and exoS strain background.
Arrows denote pathogenesis category in which there is an opposite pattern of upregulated and
downregulated genes in exoU vs. exoS.

Notably, in exoU strains with 2 TSMs vs. 1 TSM, there were 29 pathogenesis-related
genes upregulated. In contrast, 25 pathogenesis-related genes were downregulated in exoS strains
as the number of TSMs increased. These results suggest an association between upregulation of
certain virulence genes and the exoU strain background, especially in exoU strains with 2 target
site mutations (TSMs). Only limited conclusions can be made, however, due to the non-
isogenicity of the strains.
Table 6. Characteristics of Isolates Used
Microarray
Name
exoU/exoS
LVX MIC
(µg/ml)
LVX + EPI
a
MIC (µg/ml)
Mutations Doubling
Time (h)
b

205
exoU
64 4 gyrA: T83I
parC: S87L
-
91 exoU 2 2 gyrA: T83I

-
36
exoS
64 1 gyrA: T83I
parC: S87L
-
139
exoS
32 4 gyrA: T83I

-
Isogenic strains (All fitness assays)
Name
exoU/exoS
LVX MIC
(µg/ml)
LVX + EPI
a
MIC (µg/ml)
Mutations Doubling
Time (h)
b

37
exoU
16 0.5 gyrA: T83I 2.0
37-PC*
16 0.5 gyrA: T83I
parC: S87L
1.7
91
exoU
2 2 gyrA: T83I 3.9
91-PC*
16 2 gyrA: T83I
parC: S87L
3.9
92
exoU
16 0.5 gyrA: T83I 2.5
92-PC*
16 0.5 gyrA: T83I
parC: S87L
2.8
139
exoS
4 0.5 gyrA: T83I 3.4
139-PC* 32 0.25 gyrA: T83I
parC: S87L
3.1

63

247
exoS
16 1 gyrA: T83I 2.5
247-PC*
16 8 gyrA: T83I
parC: S87L
2.5
215
exoS
16 1 gyrA: T83I 2.4
215-PC*
16 1 gyrA: T83I
parC: S87L
2.3
Mutants with Ser87Leu substitution in ParC denoted –PC*
a
EPI= efflux pump inhibitor
b
Doubling time during exponential phase; average of at least 3 independent experiments

Creation of isogenic pairs

We adapted the technique of oligonucleotide recombination to create isogenic strains to
study the effects of the resistance-conferring mutation in a controlled background (Agnello and
Wong-Beringer, 2014; Swingle et al., 2010). To optimize the recombination procedure, we
attempted to transform the strains using both a 22 nucleotide oligo and a 60 nucleotide oligo
(Table 5). Strains transformed with the oligo 22 nucleotides in length did not grow on any
selection plates. “Mock” control strains did not grow on any selection plate in 4 out of the 7
experiments. In the other 3 experiments, the mock cultures were able to grow on plates
containing levofloxacin at a concentration 2 times higher than the strain’s MIC; however,
subsequent sequencing of these colonies showed that no change had occurred in the parC gene,
demonstrating that growth of the control strains on selection was not due to the presence of a
spontaneous target site mutation. Strains electroporated with the 60-nucleotide oligo grew on
selection plates with a concentration of up to 16 times higher than their original MIC. Single
colonies were selected from the highest concentration of levofloxacin-containing plates and
sequenced. Sequencing results confirmed that all strains had incorporated the TCGàTTG
mutation in the parC gene at nucleotide 260, and no extraneous recombination had occurred in
the gyrA gene, despite the high degree of sequence similarity to parC. In total, we created 6
isogenic pairs (3 exoU and 3 exoS).

64

Minimum inhibitory concentration (MIC) to levofloxacin was measured by broth
microdilution and was also measured with the addition of an efflux pump inhibitor (Phe-Arg β-
naphthylamide dihydrochloride, Sigma) at 20 µg/ml in order to more accurately reflect the
resistance phenotype conferred by the specific point mutation introduced (Lomovskaya et al.,
2001). The parC mutation increased the MIC in 2 out of the 6 isolates compared to the parent
strains, by 8 fold (Table 6).
The technique of oligonucleotide recombination that we have adapted for use in clinical
isolates of Pseudomonas aeruginosa is an efficient and practical approach for inserting point
mutations into specific sections of the bacterial genome. As laboratory-derived mutants do not
reflect real world pathogenesis (Fux et al., 2005), this technique proves useful in creating
isogenic mutant strains from clinical isolates for in depth investigations.
Growth Rates
To measure growth rates, overnight cultures were diluted 10 fold and grown at 37°C with
shaking. Cultures (150 µl) were sampled every 30 min for 8 hours and turbidity was measured in
triplicate at OD
600
using a microplate reader (Tecan Group Ltd., Switzerland). Doubling time
during exponential phase was calculated by dividing the time interval over the number of
generations. Interestingly, results of an average of at least 3 independent experiments showed
that growth rates of the recombinants were not significantly affected; the average difference
between the generation times of the recombinants compared to the parents was less than 1 hour
(Table 6).
Fitness in vitro
Head-to-head competition assays are a standard method for investigating the relative
fitness of a mutant strain compared to its isogenic parent strain, and it is possible to detect

65

differences in fitness as small as 1%. Strains are grown together in co-culture, and must compete
for the limited resources available. In this way, pairwise competition experiments simultaneously
compare several components of the competing strains: lag periods, rates of exponential growth,
resource utilization efficiencies, and death rates, which cannot be done in comparisons of growth
in solo culture (Andersson and Levin, 1999).
In most studies, strains are differentiated based on selective growth on antibiotic plates.
However, since the parC mutation we inserted did not increase the levofloxacin MIC in all
strains, we had to develop a different approach. We took advantage of a Tn-7 based system
developed by Schweizer’s group (Choi et al.,
2005; Choi and Schweizer, 2006) in order to
insert a cassette containing either yellow
fluorescent protein (YFP) or cyan fluorescent
protein (CFP) under the control of a strong
promoter. The cassette is contained within a
suicide delivery plasmid. Through co-
electroporation with a helper plasmid that encodes
the transposition machinery, the cassette site-
specifically and orientation-specifically inserts into a neutral att site downstream from the glmS
gene (Figure 15). It has been shown that there is only a single site where insertion will occur in
Pa (Choi et al., 2005), making this a reliable and predictable approach to tagging our strains. The
cassette contains a gentamicin resistance marker, which allowed for initial selection; however,
since the cassette inserts in the genome, it is stable without the need for continued selection for
over 100 generations.

66

After electroporation of both plasmids and selection on gentamicin-containing plates,
strains were assessed via PCR to ensure the insertion had occurred, and in the correct location. In
addition, transformants could be easily identified with fluorescent microscopy. Colonies of CFP-
and YFP-containing strains are easily differentiated on plates using a wide-field microscope with
the appropriate filters (Figure 16), which allows for the enumeration of the number of colony
forming units (CFUs) of each strain during the competition experiment.
To investigate the fitness costs of the parC
mutation, we performed head-to-head competition assays in
which each mutant was directly competed against its
isogenic parent strain for 4 days. Colonies were counted
every 24 hours, and the relative fitness of the mutant is
determined from the ratio of mutant:parent colony forming
units (CFU) per ml. Each experiment was independently
repeated a minimum of 7 times. We confirmed the
neutrality of the YFP and CFP tags by repeating each experiment with each strain carrying the
opposite tag. No difference was found, and therefore results from all experiments were
combined. In each individual competition experiment, one strain is tagged with CFP, and the
other with YFP. This allows for the direct enumeration of the number of colonies of each strain
after serial dilution onto PIA.
Figure 17 depicts the average mutant/parent ratios over time for each strain, and Figure
18 shows the average CFU counts over time for each pair. A reduction in fitness is considered
when the mutant/parent ratio is < 1. The results show that each strain has a unique pattern of
fitness costs associated with this mutation; however, in general there seems to be a higher fitness

67

cost for the exoS strains with the acquisition of the parC mutation. Strain 92 (exoU) has the least
fitness effects, with the mutant/parent ratios ranging from 1.2-2.02 from 2-4 days of competition.
Strain 37 (exoU) shows a similar pattern except the mutant starts with low fitness (ratio 0.5 at
day 2), but catches up to the parent strain by day 4 with a ratio of 0.99 (Table 7). Strain 91
(exoU) is unique and interestingly shows a tremendous fitness increase due to the mutation; the
mutant rapidly outcompetes the parent strain and completely takes over the culture by day 3. The
exoS strains 139-PC*, 247-PC* and 215-PC* show a consistent fitness defect over all time
points, with strain 139-PC* the most severe.

Figure 17. The average mutant:parent CFU ratio per day of the experiment. Results are plotted on a log10
scale. Points above zero indicate the mutant is more fit, while below zero indicate a fitness cost. The ratio
for exoS strains generally decreases over the course of the experiment, while the ratios for exoU strains
either increases or remains stable. Results are an average of 7 independent experiments. Error bars=SEM

68

Figure 18. Average CFU/ml for each competition experiment. Error bars = SEM. Results are an average
of 7 independent experiments.

Table 7. Average CFU/ml and PC:parent ratios at each day of competition
Day 1 Day 2 Day 3 Day 4
exoU
92 2.7E+09 3.8E+09 2.9E+09 2.1E+09
92-PC* 3.0E+09 4.2E+09 4.3E+09 3.8E+09
Ratio
(PC*/Parent)
1.12 1.21 1.73 2.02

91 7.7E+08 2.6E+08 0 -
91-PC* 4.7E+09 6.9E+09 4.42E+09 -
Ratio
(PC*/Parent)
8.55

111.0

- -
37 5.0E+09 3.8E+09 5.7E+09 2.6E+09
37-PC* 2.1E+09 1.9E+09 4.3E+09 1.9E+09
Ratio
(PC*/Parent)
0.48

0.47

0.88

0.99

exoS
139 5.0E+09 3.5E+09 4.3E+09 3.4E+09
139-PC* 2.9E+09 9.2E+08 7.8E+08 4.0E+08
Ratio 0.80 0.28 0.2 0.12

69

(PC*/Parent)
247 5.37E+09 4.04E+09 4.05E+09 3.93E+09
247-PC* 4.10E+09 1.73E+09 1.82E+09 1.65E+09
Ratio
(PC*/Parent)
0.76

0.48

0.52

0.62

215 3.70E+09 4.47E+09 5.53E+09 4.06E+09
215-PC* 3.62E+09 2.82E+09 3.58E+09 2.09E+09
Ratio
(PC*/Parent)
1.03

0.68

0.81

0.66

Overall, exoS strains suffered a fitness cost, while exoU strains were able to better
tolerate the mutation. Fitness of exoU-PC* strains ranged from maintaining the wild-type level of
fitness to outcompeting the parent strain by more than 100 fold, whereas exoS-PC* strains were
consistently less fit than parent strains. Extrapolating these results to other conditions is difficult,
but if exoU strains are better able to survive after acquiring an FQ-resistance mutation in parC,
and exoS strains are less able to survive after gaining this mutation, it could explain the findings
that exoU strains are more likely to be FQ-resistant in the clinical setting.
Metabolic Utilization Assays
Previous studies have shown that the development of resistance is associated with overall
metabolic changes (Abdelraouf et al., 2011; Linares et al., 2010). To gain an overview of the
effect of the parC mutation on metabolic function, Biolog Phenotype Microarray Plates were
used. This allowed us to compare growth of the exoU vs. exoS PC* mutants and their respective
parent strains on a variety of carbon and nitrogen substrates.
Growth was measured after 24 hour incubation in the presence of 190 different carbon
substrates and 96 different nitrogen substrates. The amount of growth measured via color change
was calculated by subtracting the initial absorbance reading from the reading at 24 hours. Next,

70

the growth of each PC* strain was divided by the growth of its respective parent strain, to get the
PC*/parent fold difference in growth.
On the carbon substrates, there was a slight skew towards average greater fold difference
in growth for the exoS-PC* strains compared to parent strains, while the average distribution of
substrates for exoU-PC* strains clustered around .75-1 fold on average across substrates,
indicating slight decrease to no difference on most carbon substrates compared to the parent
strains (Figure 19). However, the exoS distribution is heavily weighted by strain 139-PC*, which
had >1.5 fold increase in growth on 89 carbon substrates, while exoS-PC* strain 247 only had 15
substrates in the >1.5 fold increase category. For comparison, exoU-PC* strains 37 and 91 had
39 and 45 substrates with >1.5 fold increase, respectively. Therefore, there seems to be large
strain-specific variations of the effect of the –PC* mutation on growth ability on different carbon
substrates.

Figure 19. Percent of carbon- and nitrogen-containing substrates in each category of fold difference in
growth vs. parent strains. The percent of substrates that fell into each category was averaged for all exoU
strains (n=3) and all exoS strains (n=3).

Interesting, a notable difference was observed in the pattern of growth on nitrogen
substrates. exoU-PC* strains had increased growth (>1.0 fold difference) on an average of 44

71

nitrogen substrates and decreased growth (<1.0) on an average of 53 substrates, while exoS-PC*
strains had increased growth on an average of about 24 substrates, and decreased growth on 69
substrates. Specifically, exoU strain 92-PC* did not differ in growth much from its parent strain,
while exoU-PC* strains 37 and 91 had 1.5 fold increased growth on an average of 9 nitrogen-
containing substrates. In contrast, exoS-PC* strains (139, 247, and 215) had 1.5 fold increased
growth only on an averge of 3 substrates. Specific nitrogen substrates that showed the most
dramatic increases or decreases in growth in the PC* strains are presented in Table 8.

Table 8. Nitrogen substrates with a substantial ( > 2 fold) difference in growth of PC* compared to parent
strain.
> 2 fold decrease vs. parent > 2 fold increase vs. parent
exoU
92-PC* Ethylamine
D-Glucosamine
Thymidine
Galactosamine
-
37-PC* Alanine–Histidine
N-Butylamine
D,L-α-Amino-N-Butyric acid
ε-Amino-N-Caproic acid
L-Citrulline
Glycine–Glutamate
Methionine-Alanine
L-Tyrosine
β-Phenylethylamine
Hydroxylamine
Formamide
91-PC* N-Phthaloyl-L-Glutamic acid
Alloxan
Formamide
D-Glutamic acid
L-Methionine
L-Pyroglutamic acid
Ethylenediamine
N-Acetyl-D-Glucosamine
exoS
139-PC* - -
247-PC* Cytosine
L-Lysine
D-Glutamic acid
Alanine-Leucine
D,L-α-Amino-N-Butyric acid
Ethanolamine
L-Citrulline
Uracil
L-Threonine
L-Homoserine
N-Acetyl-D-Glucosamine
Uridine
Hydroxylamine
N-Butylamine

72

L-Tryptophan
215-PC* Alanine–Glutamate
Alanine–Histidine
Alanine-Threonine
L-Tyrosine

Not included in Table 8, growth on β-phenylethylamine was increased >1.5 fold in all
three exoU-PC* strains compared to parent strains. This small molecule is abbreviated PEA
(Irsfeld et al., 2013) and is produced by bacteria that contain decarboxylases, such as
Pseudomonas spp. It has been found to inhibit growth as well as biofilm production and other
virulence factors in E. coli and other enteric bacteria (Irsfeld et al., 2013; Stevenson et al., 2013).
A study of a pycR mutant of P. aeruginosa showed increased growth on this substrate as well;
pycR is an important transcriptional regulator that has been shown to indirectly affect virulence
factors, biofilm growth, and other processes (Kukavica-Ibrulj et al., 2008). It is unclear what the
significance is of this increased growth in our findings, but it is interesting that it was increased
in all exoU-PC* strains compared to their parent strains, and in none of the exoS-PC* strains.
Overall, these results suggest better utilization of nitrogen-containing substrates for
exoU-PC* strains and worse utilization for the exoS-PC* strains compared to their respective
parents. Pa is a facultative anaerobe that utilizes nitrogen as the final electron acceptor in low-
oxygen and anaerobic conditions (Arai, 2011). Therefore, enhanced utilization of nitrogen
substrates on the BIOLOG plates may suggest better survival in low oxygen conditions. This
would have implications on the ability of the strains to survive in the low-oxygen environment of
the infected lung. Furthermore, differential regulation of metabolism can have downstream
effects on virulence, specifically on TTSS gene expression in Pa (Rietsch et al., 2004).
Specifically, growth in low-oxygen conditions has been reported to stimulate the TTSS
(O'Callaghan et al., 2011), and upregulation of TTSS genes has been correlated with
upregulation of denitrification genes (Mikkelsen et al., 2009). A recent study revealed that one of

73

the fitness effects of MexEF-OprN efflux overexpression in Pa is the untimely activation of the
nitrate respiratory chain in aerobic conditions (Olivares et al., 2014). The strains in our study do
show the efflux overexpressed phenotype, as do most FQ-resistant clinical isolates; and therefore
increase in nitrogen metabolism may be due to efflux pump overexpression, with the effect more
apparent in exoU-PC* than in exoS strains. This could potentially reflect a combinatorial effect
of the parC mutation and efflux pump overexpression specific to strain background, in exoU but
not exoS strains.
Expression of TTSS genes
Virulence is the ability of a pathogen to cause disease, and therefore may be linked to
fitness. The fitness of the pathogen will dictate its ability to thrive in the hostile host
environment, and the ability to transmit to other hosts. The expression of virulence factors, such
as the type III secretion system, is energetically costly, as such, strains that are more fit will be
better able to express these genes (Beceiro et al., 2013).
Pa contains many virulence factors; however, a major contributor to acute infections is
the type III secretion system (TTSS). Therefore, we were interested in investigating if FQ-R
strains with the PC* mutation have increased expression of the TTSS. To specifically investigate
expression of the TTSS, we used qRT-PCR to measure expression of pcrV, an essential
component of the needle complex (Goure et al., 2004; Lee et al., 2010), as well as activator exsA
which directly controls expression of all TTSS components (Dasgupta et al., 2004). Interestingly,
all 3 exoS-PC* strains had lower expression of both exsA and pcrV compared to the wild type
parent strains. In contrast, all exoU-PC* strains had higher expression of pcrV than the parent
strains (Figure 20). This suggests that acquisition of this mutation can exert an effect on

74

expression of TTSS, and this effect differs for strains depending on exoU vs exoS background,
enhancing expression specifically in exoU strains.

Figure 20. Expression of TTSS related genes, fold difference of PC* mutants compared to parent strains.
Housekeeping gene rpoD was used as a reference for expression normalization.

Overall, the observation that exoU-PC* strains and exoS-PC* strains show opposite
expression patterns compared to their parent strains is interesting and warrants further
investigation. The parC mutation may affect TTSS expression via overall regulatory changes
that may reflect either changes in supercoiling from the mutation, and/or global metabolic
changes.
Supercoiling
In bacteria, chromosomal DNA exists in a very condensed state due to twisting of the
DNA into supercoils. The topoisomerase enzymes control and regulate the level of supercoiling
in the cell, which is not fixed and constantly changing in response to cellular processes (such as
DNA replication) and environmental stressors (Redgrave et al., 2014). FQ-R target site
mutations occur in the genes encoding for the topoisomerases; therefore, these mutations may
lead to perturbations in the regulation of supercoiling.

75

Regulation of supercoiling reflects the level of fitness, and may explain the fitness
differences seen in our strains. Other studies have shown the supercoiling effects of FQ-
resistance mutations and their effects on fitness. In C. jejuni (which does not have the gene for
ParC), the Thr86Ile mutation in gyrA, which correlates to the Thr83Ile mutation in Pa that our
strains carry, led to reduced DNA supercoiling compared to wild type (Han et al., 2012). Studies
in E. coli have shown that gyrA mutations lead to differences in supercoiling compared to wild
type. Slightly contradicting results from these studies revealed that mutations in both gyrA and
parC either further decrease supercoiling or return the supercoiling level closer to that of wild
type (Bagel et al., 1999; Marcusson et al., 2009). In a study of the fitness of FQ-R Pa strains, the
strains with decreased fitness also had less supercoiling than wild type; however, this study only
focused on mutations in gyrA (Kugelberg et al., 2005).
We assessed supercoiling of parent vs. mutants in exoU strains 92 and 37 and exoS strains
215 and 139. Due to challenges inserting the reporter, only 2 strains of each background were
chosen as representative strains. Common methods for investigating supercoiling levels rely on
reporter plasmids that need to be selected for and maintained. Because we are using clinical
isolates, there is a high level of multi-drug resistance that precludes the use of standard selection
antibiotics. To circumvent this, we inserted a Tn7 genetic element in which a supercoiling
sensitive P. aeruginosa promoter controls the lux operon. This genetic element was created and
validated for use as a reporter screening assay for gyrase inhibitors (Moir et al., 2007). The
reporter is the lux operon, and it is under the control of a supercoiling sensitive promoter from
the gene PA0614, which was selected due to its high responsiveness to changes in supercoiling
induced by gyrase inhibition. We took advantage of this tool in order to investigate if the parC
mutation affects the strains’ ability to regulate supercoiling. Strains with the insertion were

76

grown under levels of levofloxacin at and below the MIC to maximally induce expression of the
reporter, and luminescence values were normalized to OD
600
. In this way, the level of
luminescence in the parent strain is an indirect reporter of the baseline level of supercoiling, and
any difference in luminescence expression for the PC* strain reflects a change in supercoiling
due to the mutation. Results show that the PC* mutants in general have decreased luminescence
compared to parent strains; however, this decrease was significantly more pronounced in exoS-
PC* compared to exoU-PC* (Figure 21). This suggests that exoU-PC* mutants were better able
to maintain the supercoiling levels of the parent strains, while exoS-PC* mutants showed a more
drastic change in supercoiling, as observed by decreased lux expression.
The ability of exoU-PC* strains to maintain “wild-type” supercoiling levels under the
stress of FQ exposure may reflect increased overall fitness of these strains. The supercoiling
level of the cell affects global gene expression, and changes in supercoiling may alter response to
environmental stressors, or modulate pathogenesis in the host (Redgrave et al., 2014). Promoters
of certain genes have been shown to be especially sensitive to changes in the level of
supercoiling, affecting expression levels. Altered levels of supercoiling via inhibition of gyrase
were shown to alter the expression levels of virulence genes in Salmonella typhimurium (Galán
and Curtiss, 1990) as well as Staphylococcus aureus (Fournier and Klier, 2004), suggesting that
transcription of virulence genes may be regulated through changes in supercoiling. In that case,
the ability of FQ-R exoU strains to maintain wild-type levels of supercoiling may allow for
increased pathogenicity in the host.

77

Figure 21. The strains contain a genetic insertion in which the lux operon is under the control of a
supercoiling-sensitive promoter. Strains were grown in concentrations of levofloxacin equal to 1/4, 1/2,
and 1 x the MIC. Luminescence was normalized to cell count using OD
600.
At 1/4 and 1/2 MIC, fold
changes in exoU strains 92 and 37 are significantly different from those of exoS strain 215 (p=0.004 for
both). At 1x MIC, both exoU strains have significantly different fold changes than both exoS strains (92
vs. 139, p=0.02; 92 vs. 215, p=0.03; 37 vs. 139, p=0.03; 37 vs. 215, p=0.04). Results are an average of 3
independent experiments, and error bars represent SEM.

Compensation
The results of the fitness experiments suggest that exoU strains have a lower fitness cost
associated with FQ-resistance than exoS strains. To investigate whether stable changes were
occurring during competition that allowed the strains to compensate for fitness costs, we
collected colonies of parent and PC* strains from a primary competition experiment (parent vs.
mutant) after 7 days (Figure 22) and labeled these ‘aged’ strains. We investigated potential
compensation by comparing fitness of each aged strain to its un-aged counterpart. We
investigated fitness of the aged strains in terms of mutation frequency, competitive growth,
growth on nitrogen substrates, and TTSS expression in exoU strains 92 and 37, and exoS strains
215 and 139. We were not able to collect aged colonies from the primary competition
experiments of strain 91 due to the vast difference in fitness between the mutant and parent
strains; the mutant strain completely overtook the culture by day 3. Because this left us with only

78

2 exoU strains in which to test compensation, we chose 2 exoS strains in order to have an equal
number in each group. exoS strains 215 and 139 represent both ends of the spectrum in terms of
fitness of the PC* mutants; strain 215-PC* shows the least fitness cost of all the exoS strains,
while strain 139-PC* has the greatest fitness cost. In addition, we investigated the effect of sub-
inhibitory levels of fluoroquinolone exposure on competitive fitness and compensation by
collecting aged strains from primary competition experiments in which 1/8 MIC of levofloxacin
was added to the media.

Figure 22. Schematic of experimental workflow to investigate compensation. From the primary
competition experiment, colonies of both PC* and parent strains were collected separately into liquid
media and saved, then competed against the un-aged strains in a secondary competition.

Primary competition experiments of mutants vs. parents were performed for 7 days, and
we collected both aged parent and aged PC* strains at 7 days. The primary competition
experiments in which aged strains were collected are longer than the 4 days reported above for
initial competitive fitness experiments because we needed to enrich the population of potentially
compensated cells in the culture to be able to collect them. Although compensation may have
occurred at 4 days the probability of collecting those cells was too low and therefore we needed
Primary Competition
PC* vs. Parent

c

c

Day 7

Collection of PC* and
parent colonies
Secondary Competition
Aged-PC* vs. Un-aged PC*
Secondary Competition
Aged Parent vs. Un-aged Parent

79

to extend the experiment until 7 days. These primary competition experiments were performed
with and without the addition of 1/8 MIC of levofloxacin. One parent and one mutant aged strain
collected from the 92, 37, 215, and 139 primary competition experiments (+ and – levofloxacin)
were chosen for further analysis. Sequencing the QRDRs of the aged strains revealed that no
changes occurred in the genes associated with FQ-R except for exoU strain 92, in which the
parent acquired the Ser87Leu substitution in ParC during competition under levofloxacin
exposure (+LVX) (Table 10). The MIC to levofloxacin did not change in any of the aged strains.
Mutation Frequency
Since compensation occurs through the accumulation of beneficial mutations, we
investigated whether the rate of spontaneous mutation frequency could explain the increase in
fitness during competition that was seen in exoU-PC* strains. Mutation frequency was estimated
by calculating the frequency of spontaneous resistance to rifampicin (Rif
R
) (Oliver et al., 2000).
We estimated mutation frequency during the primary competition experiments for strains 92, 37,
and 215 by plating the mixed culture on agar containing 5x the MIC of rifampicin at each time
point (Days 1-4), and counting the number of colonies based on fluorescence of each strain that
were able to grow on the Rif plates compared to the overall CFU counts. We were not able to use
this method for exoS strains 139 or 247 due to high innate Rif resistance; 5x the MIC of Rif
turned the agar plates a dark shade of red, which prevented us from discriminating the blue and
yellow fluorophore-containing colonies on the plates. Conversely, exoU strain 91 was not able to
grow on any concentration of Rif, rendering this experiment useless for that strain. Average
mutation frequencies for strains 92, 37, and 215 along with the corresponding competition results
at each day are presented in Table 9. Overall, there does not seem to be a pattern associated with
mutation frequency and competitive growth. However, exoU-PC* strains had higher mutation

80

frequencies compared to parent strains by Day 4 of the competition, while the exoS-PC* strain,
although steadily increasing in mutation frequency throughout the experiment, did not reach the
level of the wild- type parent strain at Day 4.
Table 9. Results of primary competition experiments including Rif
R
frequencies. (Competition results are
repeated here from Table 7 for clarity and ease of comparison to Rif
R
frequencies).
Day 1 Day 2 Day 3 Day 4
Avg
CFU/ml
Avg
Rif
R

Freq.
Avg
CFU/ml
Avg
Rif
R

Freq.
Avg
CFU/ml
Avg
Rif
R

Freq.
Avg
CFU/ml
Avg
Rif
R

Freq.
exoU 92 2.7E+09 4.7E-08 3.8E+09 4.9E-08 2.9E+09 6.8E-08 2.1E+09 3.1E-08
92-PC* 3.0E+09 2.4E-08 4.2E+09 3.4E-08 4.3E+09 2.6E-08 3.8E+09 3.2E-08
Ratio
(PC*/
Parent)
1.12 0.8

1.21 0.99

1.73 0.48

2.02

1.30

91 7.7E+08 - 2.6E+08 - 0 - - -
91-PC* 4.7E+09 - 6.9E+09 - 4.42E+0
9
- - -
Ratio
(PC*/
Parent)
8.55

- 111.0

- - - - -
37 5.0E+09 7.1E-08 3.8E+09 3E-07 5.7E+09 1.7E-07 2.6E+09 7.8E-08
37-PC* 2.1E+09 1E-07 1.9E+09 4.6E-07 4.3E+09 1.1E-07 1.9E+09 8.4E-08
Ratio
(PC*/
Parent)
0.48

1.61

0.47

1.47

0.88

0.71

0.99

1.74

exoS 139 5.0E+09 - 3.5E+09 - 4.3E+09 - 3.4E+09 -
139-
PC*
2.9E+09 - 9.2E+08 - 7.8E+08 - 4.0E+08 -
Ratio
(PC*/
Parent)
0.80

- 0.28

- 0.2

- 0.12

-
247 5.4E+09 - 4.0E+09 - 4.05E+0
9
- 3.9E+09 -
247-
PC*
4.1E+09 - 1.7E+09 - 1.82E+0
9
- 1.7E+09 -
Ratio
(PC*/
Parent)
0.76

- 0.48

- 0.52

- 0.62

-
215 3.7E+09 2.1E-05 4.5E+09 3.5E-05 5.5E+09 1.0E-05 4.1E+09 2.1E-05
215-
PC*
3.6E+09 1.5E-05 2.8E+09 3.4E-05 3.6E+09 1.0E-05 2.1E+09 2.0E-05
Ratio
(PC*/
Parent)
1.03

0.73

0.68

0.86

0.81

0.99

0.66

0.91

81

In addition, we estimated mutation frequency of aged strains compared to un-aged strains
in separate experiments. Results varied based on strain and are summarized in Table 10.
Notably, exoU strain 92-PC* had an increased mutation frequency of almost 6 fold after aging.
exoU strain 37-PC* collected from a primary competition experiment under sub-inhibitory
levofloxacin increased in mutation frequency over 3.5 fold, while aged exoS strain 139-PC* had
decreased mutation frequency of 2 fold.
Competitive fitness of aged strains
Aged strains were subjected to a secondary competition experiment in which they were
competed against the original, un-aged strain. Results reflect the aged:un-aged ratio at Day 2 and
are summarized in Table 10. Because each aged strain was competed directly against an un-aged
version of itself, we were able to directly observe if fitness had changed in the aged strains. All
parent strains had a negligible change in fitness. Both exoU-PC* aged strains increased in fitness
greater than 4 fold, while the exoS-PC* aged strains showed decreased fitness greater than 50%;
139-PC* specifically was greater than 100 times less fit than before aging.
We also investigated whether growth under sub-inhibitory levels of levofloxacin affected
fitness and compensation. Sub-inhibitory concentrations of antibiotics are routinely present
during treatment, due to insufficient dosing or inadequate penetration to certain areas of the body
(Andersson and Hughes, 2014; Baquero and Negri, 1997). We were interested in investigating
whether the resistance conferring mutation in parC added survival benefits during exposure at
low concentrations; therefore a low level of levofloxacin was added to the primary competition
experiments. The addition of the drug did not affect the results of primary competition
experiments, but notable results were seen in the secondary competition experiments of the aged
strains vs. un-aged, which also included 1/8MIC of levofloxacin. Interestingly, in this subsequent

82

exposure to the drug, exoU-PC* strains were much more fit, outcompeting the un-aged strain
rapidly. For exoU strains 37 and 92, the PC* mutants outcompeted the un-aged strains
significantly and rapidly, much more so than in the competition experiments without the drug.
This is surprising, because the measured MICs of the aged and un-aged strains are the same.
exoS-PC* strains showed conflicting results; 215-PC* had gained fitness (4 fold), while 139-PC*
had slightly decreased fitness (2.4 fold) (Table 10).
Table 10. Characteristics of Aged Strains

Aged Strain
(day collected)
Fold difference
in Rif
R
freq. vs.
un-aged

Target site
mutations
gained
Fold difference
in fitness in
vitro vs. un-
aged

exoU

92
(7 days)
-3.3 none +3.5
92-PC*
(7 days)
+5.8 none +5.8
92
(7 days +LVX)
-1.25
parC:
TCGàTTG
+9.4
92-PC*
(7 days +LVX)
+1.2 none +91.6
37
(7 days)
-1.2 none -2.9
37-PC*
(7 days)
-5.0 none +4.0
37
(7 days +LVX)
-2.1 none +17.5
37-PC*
(7 days +LVX)
+3.6 none
+18.4
exoS
139
(7 days)
+1.1 none
+1.3
139-PC*
(7 days)
+1.4 none
-121.7
139
(7 days +LVX)
1.0 none
+4.8
139-PC*
(7 days +LVX
-2.1 none
-2.4
215
(7 days)
+1.4 none
+1.1
215-PC*
(7 days)
+1.1 none
-9.0

83

215
(7 days +LVX)
+1.4
none
-16.4
215-PC*
(7 days +LVX)
+2.9
none
+4.0

Interestingly, aged exoU-PC* strains that had been collected after 7 days of competition
in the presence of the drug also showed an increase in mutation frequency, which may allow for
beneficial compensatory mutations to arise. However, none of the aged PC* strains had an
increase in MIC, nor any additional FQ-resistance mutations in target site genes. These results
suggest that although the level of levofloxacin was much below the MIC, highly fit strains were
selected for rapidly, more so than in conditions without drug. Also, the highly fit strains were not
more resistant, suggesting that perhaps the presence of levofloxacin accelerated the process of
compensation for the already present resistance mutations. The implications of this are alarming
and suggest that low levels of antibiotic can rapidly select for highly fit strains, preferentially the
highly virulent exoU-containing strains. Furthermore, when these highly fit, highly virulent
strains are re-introduced to the antibiotic, they will rapidly outcompete all other strains. The sub-
inhibitory concentration was also surprisingly able to select for a parC mutation in an exoU
parent strain, emphasizing the known phenomenon that selection for resistance mutations can
still occur at sub-inhibitory levels of antibiotic (Andersson and Hughes, 2014; Andersson and
Levin, 1999; Baquero et al., 1998; Gullberg et al., 2011).
Metabolic Utilization Assay in Aged Strains
Because we observed differences in growth on nitrogen-containing substrates between
exoU- and exoS-PC* mutants compared to parent strains, we compared the growth of the aged
PC* strains 92 and 139 to the un-aged strains to see if compensation affected this aspect of
metabolism. Interestingly, aged exoS strain 139-PC* showed a tremendous increase in growth on

84

nitrogen substrates compared to the un-aged strain, with >1.5 fold increase in growth on 59
substrates. Aged exoU strain 92 also had increased growth than the un-aged strain, although not
as dramatic, with >1.5-fold higher growth on 13 substrates. The changes that occurred in these
strains over the course of the primary competition experiments allow for increased growth in
these conditions. It is interesting that aged exoS strain 139-PC*, which is 100-fold less fit than
the un-aged strain in secondary competition experiments, has better growth on nitrogen
substrates than the un-aged strain. The changes that occurred in this strain during aging are
probably complex and conditional; the fitness defects only appear during head-to-head
competition in rich media.
Summary and Conclusions
In order to investigate the clinical observation that exoU strains are more likely to be FQ-
resistant and acquire multiple FQ-R mutations than exoS strains, we compared the fitness of
exoU and exoS FQ-R mutants. A preliminary exploration of global gene expression in non-
isogenic clinical isolates with the exoU or exoS genetic background and single or multiple FQ-R
mutations revealed increased expression of pathogenicity-related genes for exoU strains with
mutations both in gyrA and parC.
Head-to-head competition assays of isogenic mutants vs. parent strains revealed a fitness
defect for exoS-PC* strains and a low- to no-cost fitness effect for exoU-PC* strains. Metabolic
activity was affected in PC* mutants as well, mainly in the ability to utilize nitrogen-containing
substrates for growth. exoU-PC* mutants showed less of a growth defect compared to parent
strains than the exoS-PC* mutants on the nitrogen-containing substrates. In addition, exoU-PC*
strains showed higher expression of type III secretion related genes compared to parent strains,
while exoS-PC* strains had lower expression.

85

In head-to-head competition experiments, exoU-PC* mutants tended to increase in fitness
over the course of the 4 day experiment. This suggested that exoU strains may be compensating
for the fitness costs associated with the PC* mutation. Bacteria have the ability to rapidly evolve
compensatory mechanisms to mitigate the fitness costs associated with antibiotic resistance.
Compensatory mechanisms can reverse fitness costs without any loss of resistance by either
reducing the need for the affected function of the resistance protein, substituting the affected
function with an alternative function, or directly or indirectly restoring the efficiency of the
affect function . Compensatory mechanisms include
the accumulation of additional beneficial mutations
that allow bacterial populations to evolve towards a
more fit state that simultaneously reduces antibiotic
susceptibility (Figure 26) (Marcusson et al., 2009).
For example, studies of fluoroquinolone resistance in
other organisms have shown that compensation
readily occurs during in vitro or in vivo selection in
the absence of antibiotics; low fitness resistant mutants revert to wild-type levels of fitness
without changes in FQ-resistance. In one study, the acquisition of an additional mutation in gyrA
led to the reduction of the fitness cost of an FQ-resistant gyrA mutation of E. coli. The authors
suggest the secondary mutation somehow restores function to the enzyme (Komp Lindgren et al.,
2005). The compensatory mechanisms are not always easily identified; however, and probably
reflect multiple mutations and regulatory changes that affect fitness and resistance in a complex
and indirect manner. This has been shown in studies of FQ resistance in Salmonella sp in which
Figure 23. Model of possible evolutionary paths
for bacteria under antibiotic selection. Mutual
compensation includes mutations that
simultaneously improve fitness and increased
resistance. (from Marcusson 2009)

86

serial passage of resistant mutants led to the reversion of fitness costs via unidentified
compensatory mutations (Giraud et al., 2003; O'Regan et al., 2010).
The results of the fitness experiments suggest an increased ability of exoU strains to
compensate for the fitness costs associated with FQ-resistance, but that the specific phenotypes
associated with compensation are dependent on strain background. In head-to-head competition
experiments, which are the standard method for investigating overall fitness in vitro , exoU-
PC*strains that had been ‘aged’ through a primary competition experiment for 7 days have
increased fitness, while aged exoS-PC* strains have a decreased fitness. Because we saw less of
a change in fitness in aged parent strains, the compensation mechanism most likely acts as
repressor of the negative effects of the parC mutation, instead of just conferring a general gain in
fitness. Compensatory effects on other measures of fitness, such as metabolic growth and
virulence, differed based on specific strain background. Interestingly, exoS strain 139-PC*,
although much less fit than the un-aged strain as determined by the competition experiments, had
much higher growth on nitrogen-containing substrates than the un-aged strain. This could
represent a conditional fitness effect of aging.
We showed that exoU-PC* mutants are better able to maintain wild-type levels of
supercoiling compared to exoS-PC* mutants. This suggests that the compensatory mechanisms in
exoU strains may be acting to maintain wild-type supercoiling levels in the PC* mutants, as has
been reported in other studies of compensation in costs of FQ-resistance (Bagel et al., 1999;
Kugelberg et al., 2005; Marcusson et al., 2009).
Although we did not identify the mechanisms responsible for the compensatory
phenotypes, we know that they are stable, since strains were frozen and grown before testing for
compensation in secondary competition experiments, and the results were repeated many times.

87

Sequencing of the quinolone-resistance determining regions of gyrA/B and parC/E revealed no
additional mutations had occurred during competition; however, beneficial mutations may have
arisen elsewhere
With such large gains in fitness for exoU-PC* aged strains in the competition
experiments, it is surprising that we were unable to detect any pattern of differences or increases
in mutation frequency during the 4 day primary competition experiment. It is interesting to note
that a similar study in E. coli showed compensation had occurred that most likely helped to
regulate supercoiling, but no mutations were found in any genes known to be involved in
supercoiling (Kugelberg et al., 2005). Therefore, it is possible that other stable mechanisms
besides genetic mutation may be involved in compensation. Results of a recent study suggest that
Pa may re-wire its metabolic networks in order to compensate for fitness effects, not necessarily
requiring additional mutations (Olivares et al., 2014).
Taken together, these results suggest a lower fitness burden associated with FQ-
resistance for exoU strains than for exoS strains, which in part provides a biological explanation
for exoU strains’ greater tendency to acquire FQ-resistance in the clinical setting. The ability of
exoU strains to compensate for the costs of FQ- resistance has many clinically negative
consequences. Compensation in clinical populations leads to the stabilization of resistant
populations without the presence of drug . Therefore, a reduction in antibiotic use as a strategy
for reducing resistance will not necessarily be effective, once resistant mutants have been
selected. Clinical studies have shown that in some cases, stopping antibiotic use did not decrease
the abundance of resistant bacteria in the population (Melnyk et al., 2014), presumably due to the
high level of fitness of resistant strains. In the case of Pa, it is not just highly fit FQ-resistant
strains that are being selected for; it is the combination of highly fit and highly virulent exoU,

88

FQ-resistant strains. The increased fitness of these strains after acquisition of FQ-resistance adds
to the already complicated problem of treating infections with highly virulent exoU strains.

The results presented in this chapter are described in part in the following publication:
Agnello, M., and Wong-Beringer, A. (2014). The use of oligonucleotide recombination to
generate isogenic mutants of clinical isolates of Pseudomonas aeruginosa. J Microbiol Methods
98, 23-25.

89

Chapter 4: Effects of FQ-resistance in ExoU vs. ExoS strain on the Fitness and Virulence in
Host Infection Models

Introduction
Although the results described in Chapter 3 provide evidence suggestive of an increased
ability of Pa strains with the exoU background to compensate for fitness costs associated with
FQ-resistance, the experiments were done in vitro, without taking into account the effect of the
host environment on the bacterial cell. Pa is an opportunistic pathogen that has evolved together
with the host immune system in order to exploit host weaknesses. The type III secretion system,
for example, is not induced until the bacterial cell is in direct contact with a host cell. In vitro
systems attempt to mimic induction through manipulation of culture conditions, as described in
Chapter 3, but this is an approximation at best of the actual conditions that induce expression in
the host. Therefore, to fully investigate the effects of the FQ-resistance conferring mutation in
parC on the fitness and virulence of exoU and exoS strains, it is essential to use host models to
more closely mimic the host environment.
To compare the virulence ability of PC* vs. parent strains, we used a cell culture model
and measured the ability of the strains to induce cell death. We utilized a human lung epithelial
cell line, A549, which is widely used to study Pa virulence (Chemani et al., 2009; Ichikawa et
al., 2000; O'Loughlin et al., 2013), especially the type III secretion system (Coburn and Frank,
1999; Rietsch et al., 2005; Sawa, 2014; Vance et al., 2005). It has also been shown that virulence
against A549 cells is correlated with virulence in vivo (Sawa et al., 1998; Schulert et al., 2003).
Therefore, this was an appropriate model to compare the virulence of FQ-resistant mutants to
parent strains. We hypothesized that exoU-PC* mutants would have greater cytotoxicity than

90

parent strains and vice versa for exoS-PC* mutants, in accordance with our in vitro TTSS
expression results (see Chapter 3). In addition, we used this model to test the virulence of aged
strains compared to un-aged strains, to investigate if compensatory mechanisms allowing for
greater virulence had evolved.
To investigate the effects of the parC mutation on the competitive fitness of exoU and
exoS strains in vivo, we utilized a murine model of acute pneumonia. This allowed us to
investigate fitness in a manner that leads to more clinically relevant results. Animal models are
essential to pre-clinical investigations and have contributed significantly to the knowledge of the
pathogenesis of lung infections (Bakker-Woudenberg, 2003; Mestas and Hughes, 2004).
Specifically, we used an established murine model that leads to the rapid establishment of Pa in
the lungs (Allewelt et al., 2000; Comolli et al., 1999). This model has been used extensively to
mimic acute pneumonia caused by Pa in patients (Diaz et al., 2008; Howell et al., 2013). The in
vitro fitness results described in Chapter 3 warranted further investigation into the ability of the
strains to compete for survival under the stresses of the host environment and immune system.
We hypothesized that exoS strains would have a significant fitness cost associated with the parC
mutation, while exoU-PC* strains would either have enhanced fitness or not be affected.
Methods
Cytotoxicity against A549 cells
The human epithelial cell line A549 (lung carcinoma cells) was used to measure the
cytotoxicity via Pa-induced lactate dehydrogenase (LDH) release. Cells were routinely grown in
Ham’s F12K media with 10% fetal bovine serum. For the assay, the manufacturer’s directions
were followed (Cytotox 96, Promega, Madison, WI). 96-well plates were seeded with 50,000
A549 cells per well, and bacteria were added at an MOI of 50:1 for exoU strains and 65:1 for

91

exoS strains. LDH release was measured at 2, 4, and 6 hours after infection. Samples were run in
triplicate, with controls per the manufacturer’s instructions.
Competition assays in vivo
Individual Pa strains from frozen stocks were grown overnight in LB, diluted 1:100 in 50
ml LB, and grown to mid-exponential phase. 25 ml of each culture was then centrifuged at 12000
x g for 7 min, and resuspended in 1-2 ml PBS, to an OD
600
of 1.8 for exoS strains, and 1.5 for
exoU strains. 500 µl of each culture was then mixed and used for infection. The aspiration
pneumonia model was used (Allewelt et al., 2000; Comolli et al., 1999) as described previously.
Briefly, 6-8 week old male C57BL/6 mice (5 per group) were anesthetized with inhaled
isoflurane, and 20 µl of the mixed culture, corresponding to approximately 3x10
9
CFU for exoS
strains and 1x10
9
CFU for exoU strains, was placed on the nares with a pipette tip while the mice
were held upright. At 18 hours after infection, the mice were sacrificed, and the lungs, liver, and
spleen aseptically removed and weighed. The organs were homogenized in 2 ml sterile PBS,
serially diluted, and plated on PIA for CFU enumeration. Colonies were counted using the
fluorescent microscope as described above. Results are reported as CFU/gram of each organ.
The Institutional Animal Care and Use Committee of the University of Southern
California approved the animal studies.

92

Results and Discussion
Cytotoxicity
We investigated virulence of mutants compared to parent strains by co-inoculating the
strains with A549 cells, and the amount of lactate dehydrogenase (LDH) released was measured
via a color change reaction and used as an indicator of cell death.
Figure 24. Cytotoxicity against A549 cells after 4 hours. Only strain 91-PC* shows more than a slight
difference compared to the parent strain. Average of 3 replicates, error bars = standard deviation.

In general, there was minimal difference in cytotoxicity of mutants compared to parent
strains (Figure 27). Two exoU-PC* strains showed a slight increase over the parent strains. exoU
strain 91-PC* showed a significant decrease in cytotoxicity compared to the parent strain. For
exoS strains, 139-PC* was slightly decreased, no difference was seen in 215-PC*, and 247-PC*
was slightly increased. Overall, it appears that the addition of the parC mutation in either exoU
or exoS strains did not affect cytotoxicity to epithelial cells, with the exception of strain 91.
Cytotoxicity results did not correlate well with expression of TTSS genes (see Chapter 3),
especially in exoU strain 91, in which 91-PC* was substantially less cytotoxic than the parent
strain, but had much higher expression of pcrV. Pa virulence is multifactorial, and cytotoxicity is

93

determined by the combination of all virulence factors. Therefore, while TTSS may be highly
expressed in strain 91-PC*, in that particular strain it may play a minor role in overall virulence
under cell culture conditions. In addition, when RNA is collected for qPCR to measure
expression of TTSS genes, it is under an artificial environment that induces TTSS (the presence
of a calcium chelator) and may not accurately reflect the complex host-pathogen interactions that
lead to TTSS expression during direct host cell contact, as in cell culture assays.
As described in Chapter 3, we collected aged PC* and parent strains at the end of a 7-day
competition experiment. We investigated the fitness of aged vs. un-aged strains to determine if
compensatory mechanisms had evolved that led to increased fitness and virulence. Interestingly,
aging affected the cytotoxicity of the exoU-PC* and exoS-PC* strains differently. Both aged
exoU-PC* strains (37-PC* and 92-PC*) had <1.5 fold difference in cytotoxicity compared to the
un-aged 37-PC* and 92-PC*, while exoS aged strains had decreased cytotoxicity compared to
un-aged strains; 139-PC* decreased 2.7 fold and 215-PC* decreased 1.7 fold (Table 11). Both
these exoS strains are also less fit than their un-aged counterparts in the head-to-head competition
assays, suggesting that in this genetic background, the changes that occurred in aging affected
both fitness and virulence ability
Competition assays in a murine model
Mice were co-infected with both the parent and the PC* mutant, and after 18 hours of
infection, mice were sacrificed, the lungs, liver, and spleens removed and homogenized, and
colonies of each strain were counted. The livers and spleens were collected in addition to the
lungs in order to investigate dissemination to other organs as a marker of fitness and virulence.
As expected, infection with exoU strains overall led to higher bacterial burden compared
to exoS strains (Figure 28) and exoU strains were more likely to disseminate to the liver and

94

spleen. exoS strain 247 was the exception; levels of bacteria similar to those from the exoU-
infected mice were recovered from the lungs and an average of 2x10
4
total bacteria per gram in 3
out of the 5 mice were recovered from the liver. Infection with exoS strains 139 and 215 did not
lead to dissemination to either organ in any of the mice. exoU strain 37 had the highest rate of
dissemination, with an average of 3x10
4
total bacteria per gram recovered from the liver and
9x10
4
recovered from the spleen from all 5 mice. There was no difference detected in the ability
of the PC* mutants and parent strains to disseminate beyond the lungs.
The results of the competition experiments did not show a significant fitness effect of
the PC* mutation in any strain. The average CFUs per gram of lung of each PC* mutant and its
parent strain is shown in Figure 28. Surprisingly, exoU pair 91 and 91-PC* and exoS pair 139
and 139-PC*, which had the greatest fitness differences in vitro, only had slight differences in
vivo. Interestingly, however, was that in vivo, the fitness effect for strain 91-PC* compared to its
parent was opposite of that measured in vitro. In vitro, 91-PC* rapidly outcompeted its parent
strain, but in the lungs, 91-PC* was slightly less fit (on average about 2 fold less CFU/gram in
the lungs collected of 91-PC* compared to 91). These contradictory results mirror those seen in
the virulence assay: expression of TTSS genes was much more highly expressed in 91-PC*
compared to 91, but the cytotoxicity against human cells of strain 91-PC* was severely
decreased compared to 91. This may reflect some sort of conditional fitness effects of the PC*
mutation in strain 91, where fitness defects are only seen in response to the host. In addition, this
strain may have other characteristics (mutations, regulatory defects) that react poorly with the
addition of a PC* mutation when in conditions where host cells are present. A more in depth
investigation of this strain is needed to determine the cause of the contradictory observations.

95

Figure 25. Average CFU of each strain per gram of lung. Mice (n=5 mice for 92, 91, 37, 215, and 139.
n=4 for strain 247) were co-infected with each strain plus its corresponding PC* mutant. Error bars=SEM.

Although the difference in CFU per gram of PC* mutants vs. parent strains in the lungs
did not reach statistical significance in any experiment, when comparing the PC*:parent strain
ratios in the lungs, an interesting trend was noted. exoU strains 92 and 37 had ratios above 1 (1.2
and 1.8 on average, respectively), while all three exoS strains had PC*:parent ratios less than 1
(ranging from 0.7-0.8) (Figure 29). When the PC*:parent ratios are plotted for each individual
mouse, only 3 of the 11 mice with ratios above 1 are from an exoS experiment (Figure 30). This
suggests that exoU-PC* strains are more frequently able to out-compete parent strains than exoS-
PC* strains.

96

Figure 26. Average ratio of PC*: parent CFU per gram. Ratio over 1 indicates more of the PC* strain was
recovered from the lung, while ratio less than 1 indicates more of the parent strain was recovered. n=5
mice for 92, 91, 37, 215, and 139; n=4 for strain 247.

Figure 27. PC*:Parent ratio of CFU per gram of lung. Each symbol represents an individual mouse.
We also performed in vivo competition preliminarily with two aged strains to determine
if the aged effects could be confirmed in the host, in exoU strain 92-PC* and exoS strains 139-
PC*. Results for aged strain 92-PC* did not reflect the large gain in fitness seen in vitro; the
aged strain was actually on average almost 2 fold less fit (n=5) than the un-aged strain in vivo.
Results of the competition experiments of exoS aged strain 139-PC* did confirm the in vitro
results, although not as dramatic; aged 139-PC* was on average 9 fold less fit than the un-aged
strain in vivo (Table 11). In vivo results may not have mimicked the results seen in vitro for the
exoU strain 92 for a number of reasons. First, there are many different stressors in the host
environment than in a culture flask that may obscure the fitness effects seen in vitro. Secondly,
because we are sacrificing the mice at only a single time point, the ratio of aged:un-aged may
reflect the endpoint of adaptation to the environment. Specifically for the aged vs. un-aged 92-
92
91
37
215
139
247
0
1
2
3
4
5
PC*:Parent ratio
exoU exoS

97

PC* strains, the un-aged strain may be evolving during the 18hr infection to the fitness level of
the aged strain, which would explain the negative results seen in vivo.

Table 11. Fitness and Virulence of Aged Strains vs. Un-Aged in Host Models
Aged Strain
(Day collected)
Fold difference in
cytotoxicity vs. un-aged
Fold difference in fitness
in vivo vs. un-aged
exoU
92-PC*
(7 days)
+1.1 -1.6
37-PC*
(7 days)
-1.4 not tested
exoS
215-PC*
(7 days)
-1.7 not tested
139-PC*
(7 days)
-2.7 -9.0

Summary and Conclusions
The utilization of host models for the investigation into the underlying mechanisms
leading to the differential development of FQ-resistance in strains with the exoU and exoS
genetic backgrounds is an essential component of this research project. The clinical studies in
Chapter 2 described the observations that clinical strains obtained from patients with the exoU
background are more likely to be FQ-resistant, and exoU, FQ-R strains were more likely to cause
severe disease. Investigations into the fitness and virulence of FQ-resistant vs. wild-type strains
would not be complete without the use of an experimental host model. Lung epithelial cell
culture and the murine model of pneumonia provided the experimental models that most closely
mimic the host environment for our investigations.
Results generally followed the trend of the results found in vitro, but the differences were
not significant. The lack of statistically significant results in the in vivo competition may be due
to a number of factors. Infection in a host is not a controlled environment, as is the flask of an in

98

vitro experiment. There are many host factors that may affect growth of the strains, and these
may be variable from mouse to mouse. In vitro, the strains are presumably not expressing
virulence genes or genes necessary for establishment in the host, which would be necessary in
vivo. The expression of these factors could affect fitness in an unknown manner.
The environment of a culture tube in vitro does not mimic the complex host-pathogen
interactions, which may affect the strains in numerous ways. However, other studies
investigating fitness in Pa and other organisms have mostly seen good reliability of in vitro and
in vivo fitness results (Abdelraouf et al., 2011; Luo et al., 2005; Marcusson et al., 2009). With the
exception of strain 91, the trend of our in vivo results follows that of the in vitro experiments.
More mice per group may yield a significant difference; but another explanation is that this
method may not be sensitive enough to detect a small fitness difference. In the in vitro
experiments, we have the ability to monitor the competition of the strains at different time points,
while in vivo, the results only reflect the end point. During infection, the ‘battle’ of the strains
may have already been fought before we collect the lungs at the time of sacrifice, with both
strains evolving to the point of mutual survival. A shorter infection time, or a time series with
mice sacrificed at different time points, may reveal more information.
Importantly, Pa can live in the hospital setting not just in the lungs of patients, but in
other organs and on various surfaces inside and outside the body (Kramer et al., 2006).
Therefore, fitness effects seen in vitro may have relevance to the clinical setting; it is likely that
there are some environmental conditions that exist that mimic the results of our in vitro
experiments in which exoU, FQ-R strains are more fit and exoS, FQ-R strains are less fit
(Andersson and Levin, 1999). Importantly, patient data shows that exoU strains predominate in
the FQ-resistant population as well as cause worse disease. Therefore, experimental methods

99

currently employed in the laboratory are likely not adequate and will need further adjustments in
order to mimic the clinical environment to allow for full exploration of fitness and virulence
effects of the parC mutation in FQ-resistant Pa strains in a controlled setting.

100

Chapter 5: Summary and Future Directions

Summary of Findings
Pseudomonas aeruginosa (Pa) is the leading cause of nosocomial pneumonia,
specifically affecting ICU patients. Besides its high mortality rate, Pa is extremely difficult to
treat because of the ever-increasing rates of antibiotic resistance. The CDC has labeled multi-
drug resistant Pa a serious threat that requires monitoring and prompt action to prevent further
increases in resistance; currently it estimates that 13% of all Pa infections are caused by multi-
drug resistant strains (2013). These strains are resistant to 3 or more antibiotics, which severely
limits treatment options for patients.
The fluoroquinolone (FQ) antibiotics were once highly effective against Pa. However,
because of their ease of use and effectiveness, they became the most commonly prescribed class
of antibiotics in the U.S. (Linder et al., 2005). Not surprisingly, resistance to the FQs rapidly
increased, and currently it is estimated that 35% of all strains of Pa are FQ-resistant (Linder et
al., 2005; Rosenthal et al., 2012). Resistance to the FQs occurs through mutations controlling the
regulation of efflux pumps, as well as mutations in target site genes. The target sites of the FQs
are the topoisomerase enzymes, GyrA/B and ParC/E. Resistance to the FQs occurs when
mutations arise in the quinolone-determining-resistance regions of gyrA/B and parC/E.
The ability of Pa to cause both chronic and acute infections, and adapt to a wide array of
environments and infection sites, is due in part to its broad metabolic capabilities and its
diversity of virulence factors. Specifically, Pa utilizes its type III secretion system (TTSS)
during acute infections to evade phagocytosis, invade host cells, and cause cell death
(Veesenmeyer et al., 2009). The TTSS consists of a molecular syringe-like apparatus that

101

extends through the inner and outer membranes and directly contacts the host cell. This allows
effector toxins (ExoU, ExoS, ExoY, and ExoT) to be directly injected into the cytoplasm of host
cells. The genes encoding the cytotoxins ExoU and ExoS are mutually exclusive in most strains,
with the exoS genotype predominating the wild-type population (Feltman et al., 2001). However,
exoU strains are more virulent, as has been shown in animal models of acute pneumonia (Shaver
and Hauser, 2004), and importantly, lead to poor outcomes in patients with ventilator-associated
pneumonia (El Solh et al., 2008; Roy-Burman et al., 2001) as well as persistence and severity of
disease (Schulert et al., 2003).
Clinical studies have shown that patients infected with FQ-resistant strains of Pa have 3-
fold higher mortality and prolonged illness by an additional 5 days compared to those infected
with antibiotic susceptible strains (Hsu et al., 2005). More in-depth investigations revealed that
exoU strains were more likely to be FQ-resistant than exoS strains in a sample of 45 clinical
isolates (Wong-Beringer et al., 2008).
The results of 2 large clinical studies, described in Chapter 2, provide further evidence for
a correlation of FQ-resistance with virulence and the exoU genotype. In a large study of clinical
isolates, we found that the combined traits of FQ-resistance and exoU genotype among
respiratory isolates of Pa are significantly associated with the development of pneumonia rather
than bronchitis or colonization (Sullivan et al., 2014). This indicates that FQ-resistance and the
exoU genotype together predict more severe disease than the exoS genotype and FQ-resistance,
or either trait alone. The second study described in Chapter 2 was an investigation into the
genotype, level of FQ-resistance, and target site mutations in 270 clinical strains. This study
showed that 63% of exoU strains were FQ-resistant, compared to 49% of exoS strains, a
statistically significant difference. Furthermore, exoU strains were more likely to acquire two or

102

more target site mutations, specifically in the parC gene. These studies together suggest a co-
selection of FQ-resistance and virulence in clinical strains of Pa. The next step was to investigate
the underlying mechanisms responsible for this correlation. We hypothesized that the differences
in the development of FQ-resistance seen in exoU vs. exoS strains is due to an underlying
difference in fitness specific to the strain background: exoU strains are more adaptable to FQ-
exposure than exoS strains, allowing them to more readily acquire FQ-resistance.
The investigation of clinical isolates showed that exoU and exoS strains were equally
likely to acquire a mutation in the gyrA gene, but exoU strains were more likely to have a
mutation in both gyrA and parC compared to exoS strains. GyrA is the primary target of the FQs,
and mutations usually arise in gyrA as the first step in resistance (Higgins et al., 2003).
Therefore, we focused on a specific FQ-resistance mutation in parC, and tested the effect that
mutation had on the fitness of exoU and exoS strains. We were interested in comparing the
magnitude of the fitness effect in exoU vs. exoS strains.
In order to investigate the fitness effects of the mutation in parC in a controlled genetic
background, we inserted the point mutation into the genomes of clinical exoU and exoS strains (3
each) using the technique of oligonucleotide recombination, in order to create isogenic pairs of
strains. We then tested the fitness of the mutant vs. parent strains in variety of experiments. The
standard method of investigating the fitness of isogenic strains is with head-to-head competition
experiments. We performed these both in vitro and in vivo. As described in Chapter 3, results
suggest that exoU-PC* strains have less of a fitness cost, both in vitro and in vivo, compared to
exoS-PC* strains, and in some cases, the mutation may actually confer a benefit to exoU strains.
In addition to head-to-head competition, we compared growth on an array of carbon and nitrogen
substrates. No difference was seen in growth on carbon substrates overall, but exoU-PC* mutant

103

strains had higher growth on more nitrogen-containing substrates than exoS strains, suggesting
that the PC* mutation may provide a benefit to exoU strains in some growth conditions. Next,
because the target site mutations occur in the topoisomerase genes, we investigated the effect on
supercoiling. Results showed that exoU-PC* strains are better able to maintain supercoiling
levels compared to exoS-PC* strains. Finally, because virulence may be linked to the fitness of a
pathogen, we compared in vitro expression of the type three secretion system (TTSS) as well as
cytotoxicity against human epithelial cells of mutant vs. parent strains. Interestingly, exoU-PC*
mutants had higher expression of TTSS genes compared to parent strains, while exoS-PC* strains
had lower expression. However, cytotoxicity against A549 epithelial cells was not different in
any of the strains.
The results of the head-to-head competition experiments suggested that exoU-PC*
mutants may be compensating for the fitness costs associated with the mutation, since in some
strains the mutants started out less fit than the parents, but increased in fitness over the course of
the 4-day experiment. Therefore, to investigate compensation, we collected mutant and parent
strains at the end of 7-day competition experiments. We then re-competed these ‘aged’ strains
against un-aged versions of themselves in order to test if they had evolved stable compensatory
mechanisms. These secondary competition experiments revealed that aged exoU-PC* mutants
had in fact increased in fitness substantially, while aged exoS-PC* strains had either no change,
or a substantial decrease in fitness after aging. These results were more pronounced when we
tested strains from a primary competition experiment in which sub-inhibitory levels of
levofloxacin had been added, suggesting that low levels of antibiotic can select for highly fit,
exoU strains. Overall, it appears that exoU strains can compensate for the fitness costs of the
parC mutation, while exoS strains cannot.

104

In addition to in vitro experiments, we investigated virulence and fitness using host
infection models, as described in Chapter 4. First, we compared cytotoxicity against human lung
epithelial cells of mutants vs. parent strains. Overall, there was only a slight difference between
the mutants and parents, and no difference between exoU and exoS strains in terms of the effect
of the parC mutation on cytotoxicity. We also utilized an established murine model of acute
pneumonia in order to compare the in vivo fitness of mutants vs. parent strains. The results did
not mimic the in vitro results; in vivo, we generally found no difference in the ability of the
mutant vs. parent strains to compete for survival in the lungs; similar numbers of each strain
were recovered from the lungs of the mice. Although in individual mice there was a trend
towards higher fitness of exoU-PC* strains, the results did not reach statistical significance. This
is most likely due to the experimental model not being sensitive enough to detect small fitness
differences that we are able to detect in vitro. Adjustment of the model will most likely allow for
better resolution.
In general, we know from clinical data that exoU strains are more prevalent in the FQ-
resistant sub-population and more likely to acquire a mutation in parC. Results of the in vitro
fitness experiments suggest that the parC mutation exerts a different fitness effect depending on
exoU vs. exoS strain background. Further fine-tuning of the host models will allow for
confirmation of these results in vivo.
Conclusions and Significance
The observed correlation of virulence and FQ-resistance in clinical strains of Pa poses a
serious challenge in the control and treatment of nosocomial Pa infections. Investigating this
correlation led us to the finding that the more virulent exoU strains may be more likely to acquire
FQ-resistance due to a greater ability to compensate for the fitness costs associated with an FQ-

105

resistance- conferring mutation. This study is a novel investigation into the differential responses
of clinical isolates of Pa to FQ-resistance depending on strain background of exoU vs. exoS.
An additional important finding is that we were able to select for high fitness and high
resistance with sub-inhibitory concentrations of antibiotic (1/8xMIC). This is not a novel finding,
but further underscores the dangers of the misuse and overuse of antibiotics. Not only will the
presence of FQs in the clinical setting select for highly resistant strains, but specifically will
select for the highly virulent exoU strains due to their increased fitness.
An important practical consequence of these findings is that since FQ-resistant exoU
strains may have higher fitness than FQ-sensitive strains in the absence of antibiotic, the removal
of the antibiotic from the environment may not be enough to stop the spread of FQ-resistant
strains. Removing antibiotics from use, or ‘cycling’ antibiotics is a common practice in the
clinical setting in order to decrease the prevalence of antibiotic resistant strains and has been met
with mixed success (Arason et al., 2002; Enne et al., 2001; Sundqvist et al., 2010). However, this
relies on the assumption that without the selective pressure of antibiotics, resistant strains will
disappear. If, however, as the results of this dissertation show, the resistant strains have higher
fitness than susceptible strains, the removal of the antibiotic alone will not effectively decrease
prevalence of the resistant strains. Therefore, the compensation of fitness costs of resistance, as
is occurring in exoU strains, acts to stabilize resistant strains in the environment (Andersson and
Hughes, 2011).
The potential fitness effects of resistance should also be taken into account during the
development of new antimicrobial agents. This can be accomplished by choosing antibiotic
targets in which when resistance is acquired, the fitness cost is high and the chance of
compensation low. Finally, the experimental procedure presented here can be used as a model to

106

study the fitness effects of resistance mutations in a variety of drug-pathogen pairs using clinical
isolates. Understanding the fitness costs of antibiotic resistance and possibilities of compensation
for these costs is essential for the rational development of strategies to combat the problem of
antibiotic resistance.
Future Directions
Whole genome sequencing
The next steps of these investigations into the fitness effects of FQ-resistance in exoU and
exoS strains will be to focus on the specific genetic differences that may be contributing to the
fitness differences observed. First, whole genome sequencing should be performed on each exoU
and exoS strain (6 total). Because of the plasticity of the Pa genome, different clinical isolates
can have variable genetics. The Pa genome consists of a highly conserved core genome,
however; variability is introduced in the form of genomic islands, which make up the accessory
genome. Genomic islands are large pieces of DNA (>10kb) that are interspersed within the
genome. They are identified by a substantially different G-C content, identifying the sequences
as having been horizontally acquired. Genes within the islands usual encode accessory activities
such as specific pathogenicity or symbiosis factors (Harrison et al., 2010). In Pa, genomic
islands can make up approximately 10% of the genomes of some strains (He et al., 2004).
Therefore, whole genome sequencing will be able to identify the conserved and accessory
genomes of our strains and reveal the genetic differences that may be contributing to the
differences in fitness observed.
The exoU gene, along with its chaperone spcU, is located on a pathogenicity island,
which is a specialized genomic island. exoU has been identified as part of a few different
pathogenicity islands. The most highly studied is from the reference strain PA14, in which the

107

exoU gene is on an island named PAPI-2 (Kulasekara et al., 2006). Studies with PA14 have
shown that virulence is dependent on the presence of the entire island and not just the exoU gene
alone; therefore, other as yet unknown genes contained on the pathogenicity islands contribute
combinatorially to the increased virulence of exoU strains (Harrison et al., 2010). Furthermore,
pathogenicity in strain PA14 requires the coordinated action of multiple virulence factors,
associated with both the core and accessory genomes (Lee et al., 2006). Therefore, it is possible
that other genes in the accessory genome in combination with exoU may provide fitness benefits
to exoU strains that allow for increased ability to adapt to the fitness costs of FQ-resistance.
Genomic islands have been shown to confer fitness benefits, and the accessory genome of Pa is
an important driver of the ability of strains to persist in a particular environment (Hacker and
Carniel, 2001; Ozer et al., 2014).
RNA-sequencing
The results of an exploratory microarray expression study described in Chapter 3
highlighted differences in gene expression based on the number of target site mutations in exoU
and exoS strains, specifically in pathways relating to pathogenicity and certain aspects of
metabolism; however, this was a small study that did not include isogenic strains.
The creation of isogenic mutants from exoU and exoS parent strains has allowed us to
investigate the specific biological effects of this mutation in controlled genetic backgrounds.
RNA-seq on these strains will allow us to investigate gene expression in a controlled fashion;
differences in expression of mutants compared to parents can be directly attributed to the parC
mutation. Following whole genome sequencing with RNA-seq will also simplify analysis;
transcripts can be directly aligned to each genome sequence.

108

Results described in Chapter 3 suggest that PC* mutants may have changes in the overall
supercoiling structure of the genome, which may affect fitness via the regulation of genes with
supercoiling-sensitive promoters, as has been shown in other organisms (Fournier and Klier,
2004; Galán and Curtiss, 1990; Steck et al., 1993). We hypothesize that exoU strains are able to
compensate for the potential supercoiling disruption imposed by this mutation, allowing for
increased fitness compared to exoS resistant mutants. RNA-seq will allow us to investigate
potential effects on supercoiling through levels of expression of known supercoiling-sensitive
genes, as well as investigate possible compensatory mechanisms such as increased expression of
other DNA topology-regulating genes. In addition, it will be important to collect RNA for these
experiments from the lungs of mice during infection in order in order to obtain data that is as
relevant as possible to clinical infection (Mandlik et al., 2011; Westermann et al., 2012).
Common patterns of expression between different exoU and exoS pairs will confirm an important
pathway warranting further investigation.
Virulence in a Murine Model
An integral component of fitness is the ability of a pathogen to reproduce and cause
disease in the host. The next step of our investigations into fitness will be to further utilize the
murine model described in Chapters 3 and 4 to include investigations into lung injury caused by
parent vs. PC* strains. The toxin ExoU specifically increases neutrophil-mediated inflammation
in the lung, as well as primarily targets and lyses neutrophils (Diaz and Hauser, 2010; Diaz et al.,
2008). It is not known how FQ-resistance affects the ability of exoU and exoS strains to induce
an innate immune response. We can investigate inflammatory response by measuring and
identifying the cytokines released into the lungs by ELISA (Kalle et al., 2012). By also
measuring bacterial burden in the lungs, liver, and spleen, and histologically examining sections

109

of the lung for injury, we can compare the virulence potential of PC* vs. parent strains, and
assess if the PC* mutation affects the virulence of exoU strains differently than exoS strains. This
is important for demonstrating that the PC* mutation affects the virulence of exoU strains in a
clinically relevant manner.
As described in Chapter 3, the 6 pairs of isogenic strains are also tagged with fluorescent
proteins. An interesting extension of the murine infection model will be to visualize the strains
inside the animal during infection using whole body, in vivo imaging. Using this technique, we
would be able to monitor the progress and dissemination of the infection in real time, without
sacrificing the mice. Interestingly, if monitored during a co-infection experiment, as in a
competition experiment, we would potentially be able to identify any spatial differences of the
PC* strains compared to the parent strains, which could highlight fitness attributes that are
affected by the PC* mutation relating to the location of the strains in the lungs and the rest of the
body.
Overall, the experiments and results presented in this dissertation represent a novel
investigation into the effects of FQ-resistance in P. aeruginosa. The most benefit in future
studies will come in the form of replicating these results in many additional exoU and exoS
clinical isolates, to increase generalizability of the results.

110

References
(2013). Antibiotic Resistance Threats in the United States (Centers for Disease Control).
(2014). Antimicrobial Resistance Global Report in Surveillance (France: World Health
Organization).
C.L.S.I. (2007). Performance standards for antimicrobial susceptibility testing: 17
th

Informational supplement (Wayne, PA), pp. M100-S117.
Abdelraouf, K., Kabbara, S., Ledesma, K.R., Poole, K., and Tam, V.H. (2011). Effect of
multidrug resistance-conferring mutations on the fitness and virulence of Pseudomonas
aeruginosa. J Antimicrob Chemother 66, 1311-1317.
Agnello, M., and Wong-Beringer, A. (2012). Differentiation in Quinolone Resistance by
Virulence Genotype in Pseudomonas aeruginosa. PLoS One 7, e42973.
Agnello, M., and Wong-Beringer, A. (2014). The use of oligonucleotide recombination to
generate isogenic mutants of clinical isolates of Pseudomonas aeruginosa. J Microbiol Methods
98, 23-25.
Aitken, A. (2006). 14-3-3 proteins: a historic overview. Semin Cancer Biol 16, 162-172.
Ajayi, T., Allmond, L.R., Sawa, T., and Wiener-Kronish, J.P. (2003). Single-nucleotide-
polymorphism mapping of the Pseudomonas aeruginosa type III secretion toxins for
development of a diagnostic multiplex PCR system. J Clin Microbiol 41, 3526-3531.
Akasaka, T., Onodera, Y., Tanaka, M., and Sato, K. (1999). Cloning, expression, and enzymatic
characterization of Pseudomonas aeruginosa topoisomerase IV. Antimicrob Agents Chemother
43, 530-536.
Akasaka, T., Tanaka, M., Yamaguchi, A., and Sato, K. (2001). Type II topoisomerase mutations
in fluoroquinolone-resistant clinical strains of Pseudomonas aeruginosa isolated in 1998 and

111

1999: role of target enzyme in mechanism of fluoroquinolone resistance. Antimicrob Agents
Chemother 45, 2263-2268.
Allewelt, M., Coleman, F.T., Grout, M., Priebe, G.P., and Pier, G.B. (2000). Acquisition of
expression of the Pseudomonas aeruginosa ExoU cytotoxin leads to increased bacterial
virulence in a murine model of acute pneumonia and systemic spread. Infect Immun 68, 3998-
4004.
Alp, E., Güven, M., Yildiz, O., Aygen, B., Voss, A., and Doganay, M. (2004). Incidence, risk
factors and mortality of nosocomial pneumonia in intensive care units: a prospective study. Ann
Clin Microbiol Antimicrob 3, 17.
Andersson, D.I., and Hughes, D. (2011). Persistence of antibiotic resistance in bacterial
populations. FEMS Microbiol Rev 35, 901-911.
Andersson, D.I., and Hughes, D. (2014). Microbiological effects of sublethal levels of
antibiotics. Nat Rev Microbiol 12, 465-478.
Andersson, D.I., and Levin, B.R. (1999). The biological cost of antibiotic resistance. Curr Opin
Microbiol 2, 489-493.
Andersson, M.I., and MacGowan, A.P. (2003). Development of the quinolones. J Antimicrob
Chemother 51 Suppl 1, 1-11.
Arai, H. (2011). Regulation and Function of Versatile Aerobic and Anaerobic Respiratory
Metabolism in Pseudomonas aeruginosa. Front Microbiol 2, 103.
Arason, V.A., Gunnlaugsson, A., Sigurdsson, J.A., Erlendsdottir, H., Gudmundsson, S., and
Kristinsson, K.G. (2002). Clonal spread of resistant pneumococci despite diminished
antimicrobial use. Microb Drug Resist 8, 187-192.

112

Bagel, S., Hullen, V., Wiedemann, B., and Heisig, P. (1999). Impact of gyrA and parC mutations
on quinolone resistance, doubling time, and supercoiling degree of Escherichia coli. Antimicrob
Agents Chemother 43, 868-875.
Bakker-Woudenberg, I.A. (2003). Experimental models of pulmonary infection. J Microbiol
Methods 54, 295-313.
Balsalobre, L., and de la Campa, A.G. (2008). Fitness of Streptococcus pneumoniae
fluoroquinolone-resistant strains with topoisomerase IV recombinant genes. Antimicrob Agents
Chemother 52, 822-830.
Baquero, F., and Negri, M.C. (1997). Selective compartments for resistant microorganisms in
antibiotic gradients. Bioessays 19, 731-736.
Baquero, F., Negri, M.C., Morosini, M.I., and Blázquez, J. (1998). Antibiotic-selective
environments. Clin Infect Dis 27 Suppl 1, S5-11.
Baughman, R.P. (2005). Microbiologic diagnosis of ventilator-associated pneumonia. Clin Chest
Med 26, 81-86.
Beceiro, A., Tomás, M., and Bou, G. (2013). Antimicrobial resistance and virulence: a successful
or deleterious association in the bacterial world? Clin Microbiol Rev 26, 185-230.
Billington, O.J., McHugh, T.D., and Gillespie, S.H. (1999). Physiological cost of rifampin
resistance induced in vitro in Mycobacterium tuberculosis. Antimicrob Agents Chemother 43,
1866-1869.
Bonten, M.J., Bergmans, D.C., Ambergen, A.W., de Leeuw, P.W., van der Geest, S.,
Stobberingh, E.E., and Gaillard, C.A. (1996). Risk factors for pneumonia, and colonization of
respiratory tract and stomach in mechanically ventilated ICU patients. Am J Respir Crit Care
Med 154, 1339-1346.

113

Borkar, D.S., Acharya, N.R., Leong, C., Lalitha, P., Srinivasan, M., Oldenburg, C.E., Cevallos,
V., Lietman, T.M., Evans, D.J., and Fleiszig, S.M. (2014). Cytotoxic clinical isolates of
Pseudomonas aeruginosa identified during the Steroids for Corneal Ulcers Trial show elevated
resistance to fluoroquinolones. BMC Ophthalmol 14, 54.
Boucher, H.W., Talbot, G.H., Bradley, J.S., Edwards, J.E., Gilbert, D., Rice, L.B., Scheld, M.,
Spellberg, B., and Bartlett, J. (2009). Bad bugs, no drugs: no ESKAPE! An update from the
Infectious Diseases Society of America. Clin Infect Dis 48, 1-12.
Boucher, J.C., Yu, H., Mudd, M.H., and Deretic, V. (1997). Mucoid Pseudomonas aeruginosa in
cystic fibrosis: characterization of muc mutations in clinical isolates and analysis of clearance in
a mouse model of respiratory infection. Infect Immun 65, 3838-3846.
Brencic, A., and Lory, S. (2009). Determination of the regulon and identification of novel
mRNA targets of Pseudomonas aeruginosa RsmA. Mol Microbiol 72, 612-632.
Brencic, A., McFarland, K.A., McManus, H.R., Castang, S., Mogno, I., Dove, S.L., and Lory, S.
(2009). The GacS/GacA signal transduction system of Pseudomonas aeruginosa acts exclusively
through its control over the transcription of the RsmY and RsmZ regulatory small RNAs. Mol
Microbiol 73, 434-445.
Brint, J.M., and Ohman, D.E. (1995). Synthesis of multiple exoproducts in Pseudomonas
aeruginosa is under the control of RhlR-RhlI, another set of regulators in strain PAO1 with
homology to the autoinducer-responsive LuxR-LuxI family. J Bacteriol 177, 7155-7163.
Bucior, I., Pielage, J.F., and Engel, J.N. (2012). Pseudomonas aeruginosa pili and flagella
mediate distinct binding and signaling events at the apical and basolateral surface of airway
epithelium. PLoS Pathog 8, e1002616.

114

Chemani, C., Imberty, A., de Bentzmann, S., Pierre, M., Wimmerová, M., Guery, B.P., and
Faure, K. (2009). Role of LecA and LecB lectins in Pseudomonas aeruginosa-induced lung
injury and effect of carbohydrate ligands. Infect Immun 77, 2065-2075.
Choi, K.H., Gaynor, J.B., White, K.G., Lopez, C., Bosio, C.M., Karkhoff-Schweizer, R.R., and
Schweizer, H.P. (2005). A Tn7-based broad-range bacterial cloning and expression system. Nat
Methods 2, 443-448.
Choi, K.H., Kumar, A., and Schweizer, H.P. (2006). A 10-min method for preparation of highly
electrocompetent Pseudomonas aeruginosa cells: application for DNA fragment transfer
between chromosomes and plasmid transformation. J Microbiol Methods 64, 391-397.
Choi, K.H., Mima, T., Casart, Y., Rholl, D., Kumar, A., Beacham, I.R., and Schweizer, H.P.
(2008). Genetic tools for select-agent-compliant manipulation of Burkholderia pseudomallei.
Appl Environ Microbiol 74, 1064-1075.
Choi, K.H., and Schweizer, H.P. (2006). mini-Tn7 insertion in bacteria with single attTn7 sites:
example Pseudomonas aeruginosa. Nat Protoc 1, 153-161.
Cirz, R.T., Chin, J.K., Andes, D.R., de Crécy-Lagard, V., Craig, W.A., and Romesberg, F.E.
(2005). Inhibition of mutation and combating the evolution of antibiotic resistance. PLoS Biol 3,
e176.
Cirz, R.T., O'Neill, B.M., Hammond, J.A., Head, S.R., and Romesberg, F.E. (2006). Defining the
Pseudomonas aeruginosa SOS response and its role in the global response to the antibiotic
ciprofloxacin. J Bacteriol 188, 7101-7110.
Coburn, J., and Frank, D.W. (1999). Macrophages and epithelial cells respond differently to the
Pseudomonas aeruginosa type III secretion system. Infect Immun 67, 3151-3154.

115

Comolli, J.C., Hauser, A.R., Waite, L., Whitchurch, C.B., Mattick, J.S., and Engel, J.N. (1999).
Pseudomonas aeruginosa gene products PilT and PilU are required for cytotoxicity in vitro and
virulence in a mouse model of acute pneumonia. Infect Immun 67, 3625-3630.
Costerton, J.W., Stewart, P.S., and Greenberg, E.P. (1999). Bacterial biofilms: a common cause
of persistent infections. Science 284, 1318-1322.
Craven, D.E., Hudcova, J., and Lei, Y. (2011). Diagnosis of ventilator-associated respiratory
infections (VARI): microbiologic clues for tracheobronchitis (VAT) and pneumonia (VAP). Clin
Chest Med 32, 547-557.
Dasgupta, N., Lykken, G.L., Wolfgang, M.C., and Yahr, T.L. (2004). A novel anti-anti-activator
mechanism regulates expression of the Pseudomonas aeruginosa type III secretion system. Mol
Microbiol 53, 297-308.
Di Pasquale, M., Ferrer, M., Esperatti, M., Crisafulli, E., Giunta, V., Li Bassi, G., Rinaudo, M.,
Blasi, F., Niederman, M., and Torres, A. (2014). Assessment of severity of ICU-acquired
pneumonia and association with etiology. Crit Care Med 42, 303-312.
Diaz, M.H., and Hauser, A.R. (2010). Pseudomonas aeruginosa cytotoxin ExoU is injected into
phagocytic cells during acute pneumonia. Infect Immun 78, 1447-1456.
Diaz, M.H., Shaver, C.M., King, J.D., Musunuri, S., Kazzaz, J.A., and Hauser, A.R. (2008).
Pseudomonas aeruginosa induces localized immunosuppression during pneumonia. Infect
Immun 76, 4414-4421.
Diekema, D.J., Pfaller, M.A., Jones, R.N., Doern, G.V., Winokur, P.L., Gales, A.C., Sader, H.S.,
Kugler, K., and Beach, M. (1999). Survey of bloodstream infections due to Gram-negative
bacilli: frequency of occurrence and antimicrobial susceptibility of isolates collected in the

116

United States, Canada, and Latin America for the SENTRY Antimicrobial Surveillance Program,
1997. Clin Infect Dis 29, 595-607.
Edelsberg, J., Weycker, D., Barron, R., Li, X., Wu, H., Oster, G., Badre, S., Langeberg, W.J.,
and Weber, D.J. (2014). Prevalence of antibiotic resistance in US hospitals. Diagn Microbiol
Infect Dis 78, 255-262.
El Hage, T., Lorin, S., Decottignies, P., Djavaheri-Mergny, M., and Authier, F. (2010).
Proteolysis of Pseudomonas exotoxin A within hepatic endosomes by cathepsins B and D
produces fragments displaying in vitro ADP-ribosylating and apoptotic effects. FEBS J 277,
3735-3749.
El Solh, A.A., Akinnusi, M.E., Wiener-Kronish, J.P., Lynch, S.V., Pineda, L.A., and Szarpa, K.
(2008). Persistent infection with Pseudomonas aeruginosa in ventilator-associated pneumonia.
Am J Respir Crit Care Med 178, 513-519.
El-Solh, A.A., Hattemer, A., Hauser, A.R., Alhajhusain, A., and Vora, H. (2012). Clinical
outcomes of type III Pseudomonas aeruginosa bacteremia. Crit Care Med 40, 1157-1163.
Enne, V.I., Livermore, D.M., Stephens, P., and Hall, L.M. (2001). Persistence of sulphonamide
resistance in Escherichia coli in the UK despite national prescribing restriction. Lancet 357,
1325-1328.
Feldman, M., Bryan, R., Rajan, S., Scheffler, L., Brunnert, S., Tang, H., and Prince, A. (1998).
Role of flagella in pathogenesis of Pseudomonas aeruginosa pulmonary infection. Infect Immun
66, 43-51.
Feltman, H., Schulert, G., Khan, S., Jain, M., Peterson, L., and Hauser, A.R. (2001). Prevalence
of type III secretion genes in clinical and environmental isolates of Pseudomonas aeruginosa.
Microbiology 147, 2659-2669.

117

Finck-Barbançon, V., Goranson, J., Zhu, L., Sawa, T., Wiener-Kronish, J.P., Fleiszig, S.M., Wu,
C., Mende-Mueller, L., and Frank, D.W. (1997). ExoU expression by Pseudomonas aeruginosa
correlates with acute cytotoxicity and epithelial injury. Mol Microbiol 25, 547-557.
Fleiszig, S.M., Arora, S.K., Van, R., and Ramphal, R. (2001). FlhA, a component of the
flagellum assembly apparatus of Pseudomonas aeruginosa, plays a role in internalization by
corneal epithelial cells. Infect Immun 69, 4931-4937.
Fournier, B., and Klier, A. (2004). Protein A gene expression is regulated by DNA supercoiling
which is modified by the ArlS-ArlR two-component system of Staphylococcus aureus.
Microbiology 150, 3807-3819.
Frithz-Lindsten, E., Du, Y., Rosqvist, R., and Forsberg, A. (1997). Intracellular targeting of
exoenzyme S of Pseudomonas aeruginosa via type III-dependent translocation induces
phagocytosis resistance, cytotoxicity and disruption of actin microfilaments. Mol Microbiol 25,
1125-1139.
Fuqua, C., and Greenberg, E.P. (2002). Listening in on bacteria: acyl-homoserine lactone
signalling. Nat Rev Mol Cell Biol 3, 685-695.
Fux, C.A., Shirtliff, M., Stoodley, P., and Costerton, J.W. (2005). Can laboratory reference
strains mirror "real-world" pathogenesis? Trends Microbiol 13, 58-63.
Galán, J.E., and Curtiss, R. (1990). Expression of Salmonella typhimurium genes required for
invasion is regulated by changes in DNA supercoiling. Infect Immun 58, 1879-1885.
Garey, K.W., Vo, Q.P., Larocco, M.T., Gentry, L.O., and Tam, V.H. (2008). Prevalence of type
III secretion protein exoenzymes and antimicrobial susceptibility patterns from bloodstream
isolates of patients with Pseudomonas aeruginosa bacteremia. J Chemother 20, 714-720.

118

Gauthier, A., Thomas, N.A., and Finlay, B.B. (2003). Bacterial injection machines. J Biol Chem
278, 25273-25276.
Gaynes, R., Edwards, J.R., and System, N.N.I.S. (2005). Overview of nosocomial infections
caused by Gram-negative bacilli. Clin Infect Dis 41, 848-854.
Giraud, E., Cloeckaert, A., Baucheron, S., Mouline, C., and Chaslus-Dancla, E. (2003). Fitness
cost of fluoroquinolone resistance in Salmonella enterica serovar Typhimurium. J Med
Microbiol 52, 697-703.
Goure, J., Pastor, A., Faudry, E., Chabert, J., Dessen, A., and Attree, I. (2004). The V antigen of
Pseudomonas aeruginosa is required for assembly of the functional PopB/PopD translocation
pore in host cell membranes. Infect Immun 72, 4741-4750.
Groven, J. (2012). Pseudomonads and Non-Fermenters. In Medical Microbiology, D.
Greenwood, M. Barer, R. Slack, and W. Irving, eds. (Elsevier Ltd), pp. 298-304.
Gullberg, E., Cao, S., Berg, O.G., Ilbäck, C., Sandegren, L., Hughes, D., and Andersson, D.I.
(2011). Selection of resistant bacteria at very low antibiotic concentrations. PLoS Pathog 7,
e1002158.
Hacker, J., and Carniel, E. (2001). Ecological fitness, genomic islands and bacterial
pathogenicity. A Darwinian view of the evolution of microbes. EMBO Rep 2, 376-381.
Hall, A.R., and MacLean, R.C. (2011). Epistasis buffers the fitness effects of rifampicin-
resistance mutations in Pseudomonas aeruginosa. Evolution 65, 2370-2379.
Han, J., Wang, Y., Sahin, O., Shen, Z., Guo, B., Shen, J., and Zhang, Q. (2012). A
fluoroquinolone resistance associated mutation in gyrA Affects DNA supercoiling in
Campylobacter jejuni. Front Cell Infect Microbiol 2, 21.

119

Harrison, E.M., Carter, M.E., Luck, S., Ou, H.Y., He, X., Deng, Z., O'Callaghan, C., Kadioglu,
A., and Rajakumar, K. (2010). Pathogenicity islands PAPI-1 and PAPI-2 contribute individually
and synergistically to the virulence of Pseudomonas aeruginosa strain PA14. Infect Immun 78,
1437-1446.
Hattemer, A., Hauser, A., Diaz, M., Scheetz, M., Shah, N., Allen, J.P., Porhomayon, J., and El-
Solh, A.A. (2013). Bacterial and clinical characteristics of health care- and community-acquired
bloodstream infections due to Pseudomonas aeruginosa. Antimicrob Agents Chemother 57,
3969-3975.
Hauser, A.R. (2009). The type III secretion system of Pseudomonas aeruginosa: infection by
injection. Nat Rev Microbiol 7, 654-665.
Hauser, A.R. (2014). Pseudomonas aeruginosa virulence and antimicrobial resistance: two sides
of the same coin? Crit Care Med 42, 201-202.
Hauser, A.R., Cobb, E., Bodi, M., Mariscal, D., Vallés, J., Engel, J.N., and Rello, J. (2002). Type
III protein secretion is associated with poor clinical outcomes in patients with ventilator-
associated pneumonia caused by Pseudomonas aeruginosa. Crit Care Med 30, 521-528.
Hauser, A.R., Kang, P.J., and Engel, J.N. (1998). PepA, a secreted protein of Pseudomonas
aeruginosa, is necessary for cytotoxicity and virulence. Mol Microbiol 27, 807-818.
He, J., Baldini, R.L., Déziel, E., Saucier, M., Zhang, Q., Liberati, N.T., Lee, D., Urbach, J.,
Goodman, H.M., and Rahme, L.G. (2004). The broad host range pathogen Pseudomonas
aeruginosa strain PA14 carries two pathogenicity islands harboring plant and animal virulence
genes. Proc Natl Acad Sci U S A 101, 2530-2535.

120

Higgins, P.G., Fluit, A.C., Milatovic, D., Verhoef, J., and Schmitz, F.J. (2003). Mutations in
GyrA, ParC, MexR and NfxB in clinical isolates of Pseudomonas aeruginosa. In Int J
Antimicrob Agents (Netherlands), pp. 409-413.
Holder, I.A., Neely, A.N., and Frank, D.W. (2001). Type III secretion/intoxication system
important in virulence of Pseudomonas aeruginosa infections in burns. Burns 27, 129-130.
Hooper, D.C. (2000). Mechanisms of action and resistance of older and newer fluoroquinolones.
Clin Infect Dis 31 Suppl 2, S24-28.
Hooper, D.C. (2001). Emerging mechanisms of fluoroquinolone resistance. Emerg Infect Dis 7,
337-341.
Horan, T.C., Andrus, M., and Dudeck, M.A. (2008). CDC/NHSN surveillance definition of
health care-associated infection and criteria for specific types of infections in the acute care
setting. Am J Infect Control 36, 309-332.
Hovey, A.K., and Frank, D.W. (1995). Analyses of the DNA-binding and transcriptional
activation properties of ExsA, the transcriptional activator of the Pseudomonas aeruginosa
exoenzyme S regulon. J Bacteriol 177, 4427-4436.
Howell, H.A., Logan, L.K., and Hauser, A.R. (2013). Type III secretion of ExoU is critical
during early Pseudomonas aeruginosa pneumonia. MBio 4, e00032-00013.
Hsu, D.I., Okamoto, M.P., Murthy, R., and Wong-Beringer, A. (2005). Fluoroquinolone-resistant
Pseudomonas aeruginosa: risk factors for acquisition and impact on outcomes. J Antimicrob
Chemother 55, 535-541.
Ichikawa, J.K., Norris, A., Bangera, M.G., Geiss, G.K., van 't Wout, A.B., Bumgarner, R.E., and
Lory, S. (2000). Interaction of Pseudomonas aeruginosa with epithelial cells: identification of

121

differentially regulated genes by expression microarray analysis of human cDNAs. Proc Natl
Acad Sci U S A 97, 9659-9664.
Institute, C.L.S. (2007). Performance standards for antimicrobial susceptibility testing: 17
th

Informational supplement (Wayne, PA), pp. M100-S117.
Irsfeld, M., Spadafore, M., and Prüß, B.M. (2013). β-phenylethylamine, a small molecule with a
large impact. Webmedcentral 4.
Jalal, S., Ciofu, O., Hoiby, N., Gotoh, N., and Wretlind, B. (2000). Molecular mechanisms of
fluoroquinolone resistance in Pseudomonas aeruginosa isolates from cystic fibrosis patients.
Antimicrob Agents Chemother 44, 710-712.
Jalal, S., and Wretlind, B. (1998). Mechanisms of quinolone resistance in clinical strains of
Pseudomonas aeruginosa. Microb Drug Resist 4, 257-261.
Juhas, M., Eberl, L., and Tümmler, B. (2005). Quorum sensing: the power of cooperation in the
world of Pseudomonas. Environ Microbiol 7, 459-471.
Kalle, M., Papareddy, P., Kasetty, G., Mörgelin, M., van der Plas, M.J., Rydengård, V.,
Malmsten, M., Albiger, B., and Schmidtchen, A. (2012). Host defense peptides of thrombin
modulate inflammation and coagulation in endotoxin-mediated shock and Pseudomonas
aeruginosa sepsis. PLoS One 7, e51313.
Komp Lindgren, P., Marcusson, L.L., Sandvang, D., Frimodt-Møller, N., and Hughes, D. (2005).
Biological cost of single and multiple norfloxacin resistance mutations in Escherichia coli
implicated in urinary tract infections. Antimicrob Agents Chemother 49, 2343-2351.
Kramer, A., Schwebke, I., and Kampf, G. (2006). How long do nosocomial pathogens persist on
inanimate surfaces? A systematic review. BMC Infect Dis 6, 130.

122

Kudla, G., Murray, A.W., Tollervey, D., and Plotkin, J.B. (2009). Coding-sequence determinants
of gene expression in Escherichia coli. Science 324, 255-258.
Kugelberg, E., Löfmark, S., Wretlind, B., and Andersson, D.I. (2005). Reduction of the fitness
burden of quinolone resistance in Pseudomonas aeruginosa. J Antimicrob Chemother 55, 22-30.
Kukavica-Ibrulj, I., Sanschagrin, F., Peterson, A., Whiteley, M., Boyle, B., Mackay, J., and
Levesque, R.C. (2008). Functional genomics of PycR, a LysR family transcriptional regulator
essential for maintenance of Pseudomonas aeruginosa in the rat lung. Microbiology 154, 2106-
2118.
Kulasekara, B.R., Kulasekara, H.D., Wolfgang, M.C., Stevens, L., Frank, D.W., and Lory, S.
(2006). Acquisition and evolution of the exoU locus in Pseudomonas aeruginosa. J Bacteriol
188, 4037-4050.
Kunz, A.N., Begum, A.A., Wu, H., D'Ambrozio, J.A., Robinson, J.M., Shafer, W.M., Bash,
M.C., and Jerse, A.E. (2012). Impact of fluoroquinolone resistance mutations on gonococcal
fitness and in vivo selection for compensatory mutations. J Infect Dis 205, 1821-1829.
Kurahashi, K., Kajikawa, O., Sawa, T., Ohara, M., Gropper, M.A., Frank, D.W., Martin, T.R.,
and Wiener-Kronish, J.P. (1999). Pathogenesis of septic shock in Pseudomonas aeruginosa
pneumonia. J Clin Invest 104, 743-750.
Lakkis, C., and Fleiszig, S.M. (2001). Resistance of Pseudomonas aeruginosa isolates to
hydrogel contact lens disinfection correlates with cytotoxic activity. J Clin Microbiol 39, 1477-
1486.
Latifi, A., Winson, M.K., Foglino, M., Bycroft, B.W., Stewart, G.S., Lazdunski, A., and
Williams, P. (1995). Multiple homologues of LuxR and LuxI control expression of virulence

123

determinants and secondary metabolites through quorum sensing in Pseudomonas aeruginosa
PAO1. Mol Microbiol 17, 333-343.
Lazdunski, A.M., Ventre, I., and Bleves, S. (2007). CELL–CELL COMMUNICATION:
QUORUM SENSING AND REGULATORY CIRCUITS IN PSEUDOMONAS
AERUGINOSA In Pseudomonas L. Ramos, and A. Filloux, eds. (Springer), pp. 279–310. .
Lee, C.A. (1997). Type III secretion systems: machines to deliver bacterial proteins into
eukaryotic cells? Trends Microbiol 5, 148-156.
Lee, D.G., Urbach, J.M., Wu, G., Liberati, N.T., Feinbaum, R.L., Miyata, S., Diggins, L.T., He,
J., Saucier, M., Déziel, E., et al. (2006). Genomic analysis reveals that Pseudomonas aeruginosa
virulence is combinatorial. Genome Biol 7, R90.
Lee, E.J., Evans, D.J., and Fleiszig, S.M. (2003). Role of Pseudomonas aeruginosa ExsA in
penetration through corneal epithelium in a novel in vivo model. Invest Ophthalmol Vis Sci 44,
5220-5227.
Lee, J.K., Lee, Y.S., Park, Y.K., and Kim, B.S. (2005a). Alterations in the GyrA and GyrB
subunits of topoisomerase II and the ParC and ParE subunits of topoisomerase IV in
ciprofloxacin-resistant clinical isolates of Pseudomonas aeruginosa. Int J Antimicrob Agents 25,
290-295.
Lee, P.C., Stopford, C.M., Svenson, A.G., and Rietsch, A. (2010). Control of effector export by
the Pseudomonas aeruginosa type III secretion proteins PcrG and PcrV. Mol Microbiol 75, 924-
941.
Lee, V.T., Smith, R.S., Tümmler, B., and Lory, S. (2005b). Activities of Pseudomonas
aeruginosa effectors secreted by the Type III secretion system in vitro and during infection.
Infect Immun 73, 1695-1705.

124

Linares, J.F., Moreno, R., Fajardo, A., Martínez-Solano, L., Escalante, R., Rojo, F., and
Martínez, J.L. (2010). The global regulator Crc modulates metabolism, susceptibility to
antibiotics and virulence in Pseudomonas aeruginosa. Environ Microbiol 12, 3196-3212.
Linder, J.A., Huang, E.S., Steinman, M.A., Gonzales, R., and Stafford, R.S. (2005).
Fluoroquinolone prescribing in the United States: 1995 to 2002. Am J Med 118, 259-268.
Ling, J.M., Chan, E.W., Lam, A.W., and Cheng, A.F. (2003). Mutations in topoisomerase genes
of fluoroquinolone-resistant salmonellae in Hong Kong. Antimicrob Agents Chemother 47,
3567-3573.
Lister, P.D., Wolter, D.J., and Hanson, N.D. (2009). Antibacterial-resistant Pseudomonas
aeruginosa: clinical impact and complex regulation of chromosomally encoded resistance
mechanisms. Clin Microbiol Rev 22, 582-610.
Lomovskaya, O., Warren, M.S., Lee, A., Galazzo, J., Fronko, R., Lee, M., Blais, J., Cho, D.,
Chamberland, S., Renau, T., et al. (2001). Identification and characterization of inhibitors of
multidrug resistance efflux pumps in Pseudomonas aeruginosa: novel agents for combination
therapy. Antimicrob Agents Chemother 45, 105-116.
Lovett, S.T., Hurley, R.L., Sutera, V.A., Aubuchon, R.H., and Lebedeva, M.A. (2002). Crossing
over between regions of limited homology in Escherichia coli. RecA-dependent and RecA-
independent pathways. Genetics 160, 851-859.
Luo, N., Pereira, S., Sahin, O., Lin, J., Huang, S., Michel, L., and Zhang, Q. (2005). Enhanced in
vivo fitness of fluoroquinolone-resistant Campylobacter jejuni in the absence of antibiotic
selection pressure. Proc Natl Acad Sci U S A 102, 541-546.
Ma, L., Conover, M., Lu, H., Parsek, M.R., Bayles, K., and Wozniak, D.J. (2009). Assembly and
development of the Pseudomonas aeruginosa biofilm matrix. PLoS Pathog 5, e1000354.

125

Maatallah, M., Cheriaa, J., Backhrouf, A., Iversen, A., Grundmann, H., Do, T., Lanotte, P.,
Mastouri, M., Elghmati, M.S., Rojo, F., et al. (2011). Population structure of Pseudomonas
aeruginosa from five mediterranean countries: evidence for frequent recombination and
epidemic occurrence of CC235. PLoS One 6, e25617.
Machuca, J., Briales, A., Labrador, G., Díaz-de-Alba, P., López-Rojas, R., Docobo-Pérez, F.,
Martínez-Martínez, L., Rodríguez-Baño, J., Pachón, M.E., Pascual, A., et al. (2014). Interplay
between plasmid-mediated and chromosomal-mediated fluoroquinolone resistance and bacterial
fitness in Escherichia coli. J Antimicrob Chemother 69, 3203-3215.
Magiorakos, A.P., Srinivasan, A., Carey, R.B., Carmeli, Y., Falagas, M.E., Giske, C.G.,
Harbarth, S., Hindler, J.F., Kahlmeter, G., Olsson-Liljequist, B., et al. (2012). Multidrug-
resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert
proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect 18, 268-
281.
Mahenthiralingam, E., Campbell, M.E., Foster, J., Lam, J.S., and Speert, D.P. (1996). Random
amplified polymorphic DNA typing of Pseudomonas aeruginosa isolates recovered from
patients with cystic fibrosis. J Clin Microbiol 34, 1129-1135.
Mandlik, A., Livny, J., Robins, W.P., Ritchie, J.M., Mekalanos, J.J., and Waldor, M.K. (2011).
RNA-Seq-based monitoring of infection-linked changes in Vibrio cholerae gene expression. Cell
Host Microbe 10, 165-174.
Marcusson, L.L., Frimodt-Møller, N., and Hughes, D. (2009). Interplay in the selection of
fluoroquinolone resistance and bacterial fitness. PLoS Pathog 5, e1000541.
Mason, C.M., and Nelson, S. (2005). Pulmonary host defenses and factors predisposing to lung
infection. Clin Chest Med 26, 11-17.

126

Mathee, K., Narasimhan, G., Valdes, C., Qiu, X., Matewish, J.M., Koehrsen, M., Rokas, A.,
Yandava, C.N., Engels, R., Zeng, E., et al. (2008). Dynamics of Pseudomonas aeruginosa
genome evolution. Proc Natl Acad Sci U S A 105, 3100-3105.
Melnyk, A.H., Wong, A., and Kassen, R. (2014). The fitness costs of antibiotic resistance
mutations. Evolutionary Applications.
Mesquita, C.S., Soares-Castro, P., and Santos, P.M. (2013). Pseudomonas aeruginosa:
phenotypic flexibility and antimicrobial resistance. In Microbial Pathogens and Strategies for
Combating Them: Science, Technology, and Education, A. Mendez-Vilas, ed., pp. 650-665.
Mestas, J., and Hughes, C.C. (2004). Of mice and not men: differences between mouse and
human immunology. J Immunol 172, 2731-2738.
Mikkelsen, H., Bond, N.J., Skindersoe, M.E., Givskov, M., Lilley, K.S., and Welch, M. (2009).
Biofilms and type III secretion are not mutually exclusive in Pseudomonas aeruginosa.
Microbiology 155, 687-698.
Moir, D.T., Ming Di, Opperman, T., Schweizer, H.P., and Bowlin, T.L. (2007). A high-
throughput, homogeneous, bioluminescent assay for Pseudomonas aeruginosa gyrase inhibitors
and other DNA-damaging agents. J Biomol Screen 12, 855-864.
Mouneimné, H., Robert, J., Jarlier, V., and Cambau, E. (1999). Type II topoisomerase mutations
in ciprofloxacin-resistant strains of Pseudomonas aeruginosa. Antimicrob Agents Chemother 43,
62-66.
Nakano, M., Deguchi, T., Kawamura, T., Yasuda, M., Kimura, M., Okano, Y., and Kawada, Y.
(1997). Mutations in the gyrA and parC genes in fluoroquinolone-resistant clinical isolates of
Pseudomonas aeruginosa. Antimicrob Agents Chemother 41, 2289-2291.

127

Neuhauser, M.M., Weinstein, R.A., Rydman, R., Danziger, L.H., Karam, G., and Quinn, J.P.
(2003). Antibiotic resistance among Gram-negative bacilli in US intensive care units:
implications for fluoroquinolone use. JAMA 289, 885-888.
O'Callaghan, J., Reen, F.J., Adams, C., and O'Gara, F. (2011). Low oxygen induces the type III
secretion system in Pseudomonas aeruginosa via modulation of the small RNAs rsmZ and rsmY.
Microbiology 157, 3417-3428.
O'Loughlin, C.T., Miller, L.C., Siryaporn, A., Drescher, K., Semmelhack, M.F., and Bassler,
B.L. (2013). A quorum-sensing inhibitor blocks Pseudomonas aeruginosa virulence and biofilm
formation. Proc Natl Acad Sci U S A 110, 17981-17986.
O'Malley, Y.Q., Reszka, K.J., Rasmussen, G.T., Abdalla, M.Y., Denning, G.M., and Britigan,
B.E. (2003). The Pseudomonas secretory product pyocyanin inhibits catalase activity in human
lung epithelial cells. Am J Physiol Lung Cell Mol Physiol 285, L1077-1086.
O'Regan, E., Quinn, T., Frye, J.G., Pagès, J.M., Porwollik, S., Fedorka-Cray, P.J., McClelland,
M., and Fanning, S. (2010). Fitness costs and stability of a high-level ciprofloxacin resistance
phenotype in Salmonella enterica serotype enteritidis: reduced infectivity associated with
decreased expression of Salmonella pathogenicity island 1 genes. Antimicrob Agents Chemother
54, 367-374.
Olivares, J., Alvarez-Ortega, C., Linares, J.F., Rojo, F., Köhler, T., and Martínez, J.L. (2012).
Overproduction of the multidrug efflux pump MexEF-OprN does not impair Pseudomonas
aeruginosa fitness in competition tests, but produces specific changes in bacterial regulatory
networks. Environ Microbiol 14, 1968-1981.

128

Olivares, J., Álvarez-Ortega, C., and Martinez, J.L. (2014). Metabolic compensation of fitness
costs associated with overexpression of the multidrug efflux pump MexEF-OprN in
Pseudomonas aeruginosa. Antimicrob Agents Chemother 58, 3904-3913.
Oliver, A., Cantón, R., Campo, P., Baquero, F., and Blázquez, J. (2000). High frequency of
hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288, 1251-
1254.
Ozer, E.A., Allen, J.P., and Hauser, A.R. (2014). Characterization of the core and accessory
genomes of Pseudomonas aeruginosa using bioinformatic tools Spine and AGEnt. BMC
Genomics 15, 737.
Poole, K. (2011). Pseudomonas aeruginosa: resistance to the max. Front Microbiol 2, 65.
Quartin, A.A., Scerpella, E.G., Puttagunta, S., and Kett, D.H. (2013). A comparison of
microbiology and demographics among patients with healthcare-associated, hospital-acquired,
and ventilator-associated pneumonia: a retrospective analysis of 1184 patients from a large,
international study. BMC Infect Dis 13, 561.
Ramirez, J.C., Fleiszig, S.M., Sullivan, A.B., Tam, C., Borazjani, R., and Evans, D.J. (2012).
Traversal of multilayered corneal epithelia by cytotoxic Pseudomonas aeruginosa requires the
phospholipase domain of exoU. Invest Ophthalmol Vis Sci 53, 448-453.
Redgrave, L.S., Sutton, S.B., Webber, M.A., and Piddock, L.J. (2014). Fluoroquinolone
resistance: mechanisms, impact on bacteria, and role in evolutionary success. Trends Microbiol
22, 438-445.
Rello, J., Díaz, E., and Rodríguez, A. (2005). Etiology of ventilator-associated pneumonia. Clin
Chest Med 26, 87-95.

129

Restrepo, M.I., and Anzueto, A. (2009). The role of Gram-negative bacteria in healthcare-
associated pneumonia. Semin Respir Crit Care Med 30, 61-66.
Richards, M.J., Edwards, J.R., Culver, D.H., and Gaynes, R.P. (1999). Nosocomial infections in
medical intensive care units in the United States. National Nosocomial Infections Surveillance
System. Crit Care Med 27, 887-892.
Rietsch, A., Vallet-Gely, I., Dove, S.L., and Mekalanos, J.J. (2005). ExsE, a secreted regulator of
type III secretion genes in Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 102, 8006-8011.
Rietsch, A., Wolfgang, M.C., and Mekalanos, J.J. (2004). Effect of metabolic imbalance on
expression of type III secretion genes in Pseudomonas aeruginosa. Infect Immun 72, 1383-1390.
Rocha, C.L., Coburn, J., Rucks, E.A., and Olson, J.C. (2003). Characterization of Pseudomonas
aeruginosa exoenzyme S as a bifunctional enzyme in J774A.1 macrophages. Infect Immun 71,
5296-5305.
Rosenthal, V.D., Bijie, H., Maki, D.G., Mehta, Y., Apisarnthanarak, A., Medeiros, E.A.,
Leblebicioglu, H., Fisher, D., Álvarez-Moreno, C., Khader, I.A., et al. (2012). International
Nosocomial Infection Control Consortium (INICC) report, data summary of 36 countries, for
2004-2009. Am J Infect Control 40, 396-407.
Roy-Burman, A., Savel, R.H., Racine, S., Swanson, B.L., Revadigar, N.S., Fujimoto, J., Sawa,
T., Frank, D.W., and Wiener-Kronish, J.P. (2001). Type III protein secretion is associated with
death in lower respiratory and systemic Pseudomonas aeruginosa infections. J Infect Dis 183,
1767-1774.
Rozen, D.E., McGee, L., Levin, B.R., and Klugman, K.P. (2007). Fitness costs of
fluoroquinolone resistance in Streptococcus pneumoniae. Antimicrob Agents Chemother 51,
412-416.

130

Sadikot, R.T., Blackwell, T.S., Christman, J.W., and Prince, A.S. (2005). Pathogen-host
interactions in Pseudomonas aeruginosa pneumonia. Am J Respir Crit Care Med 171, 1209-
1223.
Sato, H., Feix, J.B., and Frank, D.W. (2006). Identification of superoxide dismutase as a cofactor
for the pseudomonas type III toxin, ExoU. Biochemistry 45, 10368-10375.
Sawa, T. (2014). The molecular mechanism of acute lung injury caused by Pseudomonas
aeruginosa: from bacterial pathogenesis to host response. J Intensive Care 2, 10.
Sawa, T., Ohara, M., Kurahashi, K., Twining, S.S., Frank, D.W., Doroques, D.B., Long, T.,
Gropper, M.A., and Wiener-Kronish, J.P. (1998). In vitro cellular toxicity predicts Pseudomonas
aeruginosa virulence in lung infections. Infect Immun 66, 3242-3249.
Schulert, G.S., Feltman, H., Rabin, S.D., Martin, C.G., Battle, S.E., Rello, J., and Hauser, A.R.
(2003). Secretion of the toxin ExoU is a marker for highly virulent Pseudomonas aeruginosa
isolates obtained from patients with hospital-acquired pneumonia. J Infect Dis 188, 1695-1706.
Shaver, C.M., and Hauser, A.R. (2004). Relative contributions of Pseudomonas aeruginosa
ExoU, ExoS, and ExoT to virulence in the lung. Infect Immun 72, 6969-6977.
Shaver, C.M., and Hauser, A.R. (2006). Interactions between effector proteins of the
Pseudomonas aeruginosa type III secretion system do not significantly affect several measures of
disease severity in mammals. Microbiology 152, 143-152.
Sifri, C., and Ausubel, F. (2005). In Cellular Microbiology, P. Cossart, P. Boquet, S. Normark,
and R. Rappuoli, eds. (ASM Press), pp. 543-560.
Society, A.T., and America, I.D.S.o. (2005). Guidelines for the management of adults with
hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit
Care Med 171, 388-416.

131

Steck, T.R., Franco, R.J., Wang, J.Y., and Drlica, K. (1993). Topoisomerase mutations affect the
relative abundance of many Escherichia coli proteins. Mol Microbiol 10, 473-481.
Stevenson, L.G., Szostek, B.A., Clemmer, K.M., and Rather, P.N. (2013). Expression of the
DisA amino acid decarboxylase from Proteus mirabilis inhibits motility and class 2 flagellar
gene expression in Escherichia coli. Res Microbiol 164, 31-37.
Stover, C.K., Pham, X.Q., Erwin, A.L., Mizoguchi, S.D., Warrener, P., Hickey, M.J., Brinkman,
F.S., Hufnagle, W.O., Kowalik, D.J., Lagrou, M., et al. (2000). Complete genome sequence of
Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 406, 959-964.
Sullivan, E., Bensman, J., Lou, M., Agnello, M., Shriner, K., and Wong-Beringer, A. (2014).
Risk of developing pneumonia is enhanced by the combined traits of fluoroquinolone resistance
and type III secretion virulence in respiratory isolates of Pseudomonas aeruginosa. Crit Care
Med 42, 48-56.
Sundin, C., Thelaus, J., Bröms, J.E., and Forsberg, A. (2004). Polarisation of type III
translocation by Pseudomonas aeruginosa requires PcrG, PcrV and PopN. Microb Pathog 37,
313-322.
Sundin, C., Wolfgang, M.C., Lory, S., Forsberg, A., and Frithz-Lindsten, E. (2002). Type IV pili
are not specifically required for contact dependent translocation of exoenzymes by Pseudomonas
aeruginosa. Microb Pathog 33, 265-277.
Sundqvist, M., Geli, P., Andersson, D.I., Sjölund-Karlsson, M., Runehagen, A., Cars, H.,
Abelson-Storby, K., Cars, O., and Kahlmeter, G. (2010). Little evidence for reversibility of
trimethoprim resistance after a drastic reduction in trimethoprim use. J Antimicrob Chemother
65, 350-360.

132

Swingle, B., Markel, E., Costantino, N., Bubunenko, M.G., Cartinhour, S., and Court, D.L.
(2010). Oligonucleotide recombination in Gram-negative bacteria. Mol Microbiol 75, 138-148.
Takase, H., Nitanai, H., Hoshino, K., and Otani, T. (2000). Impact of siderophore production on
Pseudomonas aeruginosa infections in immunosuppressed mice. Infect Immun 68, 1834-1839.
Vance, R.E., Rietsch, A., and Mekalanos, J.J. (2005). Role of the type III secreted exoenzymes
S, T, and Y in systemic spread of Pseudomonas aeruginosa PAO1 in vivo. Infect Immun 73,
1706-1713.
Veesenmeyer, J.L., Hauser, A.R., Lisboa, T., and Rello, J. (2009). Pseudomonas aeruginosa
virulence and therapy: evolving translational strategies. Crit Care Med 37, 1777-1786.
Ventre, I., Goodman, A.L., Filloux, A., and Lory, S. (2007). Modulation of bacterial lifestyles
via two-component regulatory networks. In, L. Ramos, and A. Filloux, eds. (Springer).
Ventre, I., Goodman, A.L., Vallet-Gely, I., Vasseur, P., Soscia, C., Molin, S., Bleves, S.,
Lazdunski, A., Lory, S., and Filloux, A. (2006). Multiple sensors control reciprocal expression of
Pseudomonas aeruginosa regulatory RNA and virulence genes. Proc Natl Acad Sci U S A 103,
171-176.
von Baum, H., Welte, T., Marre, R., Suttorp, N., Ewig, S., and group, C.s. (2010). Community-
acquired pneumonia through Enterobacteriaceae and Pseudomonas aeruginosa: Diagnosis,
incidence and predictors. Eur Respir J 35, 598-605.
Wang, P., Zhang, X., wang, L., zhen, Z., Tang, M., and Li, J. (2010). Subinhibitory
concentrations of ciprofloxacin induce SOS response and mutations of antibiotic resistance in
bacteria. 511-517.

133

Wasels, F., Kuehne, S.A., Cartman, S.T., Spigaglia, P., Barbanti, F., Minton, N.P., and
Mastrantonio, P. (2014). Fluoroquinolone resistance does not impose a cost on the fitness of
Clostridium difficile in vitro. Antimicrob Agents Chemother.
Westermann, A.J., Gorski, S.A., and Vogel, J. (2012). Dual RNA-seq of pathogen and host. Nat
Rev Microbiol 10, 618-630.
Winson, M.K., Camara, M., Latifi, A., Foglino, M., Chhabra, S.R., Daykin, M., Bally, M.,
Chapon, V., Salmond, G.P., and Bycroft, B.W. (1995). Multiple N-acyl-L-homoserine lactone
signal molecules regulate production of virulence determinants and secondary metabolites in
Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 92, 9427-9431.
Winsor, G.L., Lam, D.K., Fleming, L., Lo, R., Whiteside, M.D., Yu, N.Y., Hancock, R.E., and
Brinkman, F.S. (2011). Pseudomonas Genome Database: improved comparative analysis and
population genomics capability for Pseudomonas genomes. Nucleic Acids Res 39, D596-600.
Wolfgang, M.C., Kulasekara, B.R., Liang, X., Boyd, D., Wu, K., Yang, Q., Miyada, C.G., and
Lory, S. (2003). Conservation of genome content and virulence determinants among clinical and
environmental isolates of Pseudomonas aeruginosa. Proc Natl Acad Sci U S A 100, 8484-8489.
Wong-Beringer, A., Wiener-Kronish, J., Lynch, S., and Flanagan, J. (2008). Comparison of type
III secretion system virulence among fluoroquinolone-susceptible and -resistant clinical isolates
of Pseudomonas aeruginosa. Clin Microbiol Infect 14, 330-336.
Wrande, M., Roth, J.R., and Hughes, D. (2008). Accumulation of mutants in "aging" bacterial
colonies is due to growth under selection, not stress-induced mutagenesis. Proc Natl Acad Sci U
S A 105, 11863-11868.

134

Wu, M., and Li, X. (2015). Klebsiella pneumoniae and Pseudomonas aeruginosa In Molecular
Medical Microbiology, W. Tang, M. Sussman, D. Liu, I. Poxton, and J. Schwartzman, eds.
(Academic Press), pp. 1547–1564.
Wu, W., Jin, Y., Bai, F., and Jin, S. (2015). Pseudomonas aeruginosa In Molecular Medical
Microbiology, W. Tang, M. Sussman, D. Liu, I. Poxton, and J. Schwartzman, eds. (Academic
Press), pp. 753–767.
Yahr, T.L., Barbieri, J.T., and Frank, D.W. (1996). Genetic relationship between the 53- and 49-
kilodalton forms of exoenzyme S from Pseudomonas aeruginosa. J Bacteriol 178, 1412-1419.
Yahr, T.L., Hovey, A.K., Kulich, S.M., and Frank, D.W. (1995). Transcriptional analysis of the
Pseudomonas aeruginosa exoenzyme S structural gene. J Bacteriol 177, 1169-1178.
Yahr, T.L., Vallis, A.J., Hancock, M.K., Barbieri, J.T., and Frank, D.W. (1998). ExoY, an
adenylate cyclase secreted by the Pseudomonas aeruginosa type III system. Proc Natl Acad Sci
U S A 95, 13899-13904.
Zhang, J., Chu, Y., Wang, P., Ji, X., Li, X., Wang, C., and Peng, Y. (2014). Clinical outcomes of
multidrug resistant Pseudomonas aeruginosa infection and the relationship with type III
secretion system in patients with diabetic foot. Int J Low Extrem Wounds 13, 205-210.
Zhu, H., Conibear, T.C., Bandara, R., Aliwarga, Y., Stapleton, F., and Willcox, M.D. (2006).
Type III secretion system-associated toxins, proteases, serotypes, and antibiotic resistance of
Pseudomonas aeruginosa isolates associated with keratitis. Curr Eye Res 31, 297-306.

Abstract (if available)

Abstract Pseudomonas aeruginosa (Pa) is an important Gram‐negative pathogen in the hospital setting. Pa causes severe disease in susceptible patients, such as those in the intensive care unit, and is the leading cause of nosocomial pneumonia. Unfortunately, the high prevalence of antibiotic resistance in clinical strains has made treatment difficult. Specifically, resistance to the fluoroquinolone (FQ) antibiotics has increased rapidly over the last few decades, and it is estimated that 30% of all clinical strains of Pa are now FQ resistant. ❧ Pa has the ability to cause severe disease in patients, due to its large arsenal of virulence factors. Specifically during acute infections, Pa utilizes the type III secretion system (TTSS) to inject toxins into host immune cells. The toxins (ExoU, ExoS, ExoT, ExoY) disrupt host cell function, leading to cell death, and allowing for immune invasion and the establishment of infection. Interestingly, the genes encoding the toxins ExoU and ExoS are mutually exclusive in most strains. exoS‐containing strains are more prevalent in the clinical population. However, exoU strains are more toxic, leading to worse disease and higher mortality in patients and animal models. ❧ Previous clinical studies have shown that infection with FQ‐resistant strains leads to increased mortality in patients, suggesting that FQ‐resistance is correlated to increased disease‐causing ability. Subsequent studies revealed that ExoU strains are more likely to be FQ‐resistant than their less‐virulent, ExoS‐producing counterparts. We have also confirmed this observation in a study of a large number of clinical strains. This represents a significant clinical problem

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Methicillin-resistant Staphylococcus aureus: molecular epidemiology, virulence in disease, and antibiotic modulation of virulence

PDF

Antibiotic immunomodulation in Staphylococcus aureus bloodstream infection

PDF

Therapeutic potential of rhesus theta (θ) defensin-1 in cystic fibrosis airway infection and inflammation

PDF

Assessing the distribution and role of the CRISPR-Cas system on carbapenem resistance in e. coli and k. pneumoniae clinical strains

PDF

Therapeutic potential of host defense peptides for the treatment of cystic fibrosis lung disease

PDF

Investigation of in vitro activity and in vivo safety of imipenem-relebactam alone and in combination with other anti-microbial agents against clinical isolates of M. abscessus from CF patients

PDF

Therapeutic potential of rhesus theta defensin-1 in the treatment of inflammatory lung diseases

PDF

Understanding the contribution of microbial resistance and virulence towards outcomes of Staphylococcus aureus bacteremia

PDF

Pharmacokinetic and pharmacodynamic optimization of CFTR modulator therapy to mitigate potential drug interactions and adverse events in people with cystic fibrosis

PDF

To evaluate the in vitro activity and potential synergistic effects of Imipenem-Relebactam alone and in combination with other antimicrobial agents against clinical isolates of M. abscesses…

PDF

Averting dementia: renin-angiotensin system and angiogenic cells in cognitive decline

PDF

Clinical, functional and therapeutic analysis of CD99 in acute myeloid leukemia

PDF

Neuroinflammation and ApoE4 genotype in at-risk female aging: implications for Alzheimer's disease

PDF

An iPSC-based biomarker strategy to identify neuroregenerative responders to allopregnanolone

PDF

Therapeutic potential of Rhesus theta defensin-1 for the treatment of COVID-19 pneumonia

PDF

Renin-angiotensin system modulation for the prevention and treatment of metabolic dysfunction

PDF

Utilization of a THP-1 macrophage model to test efficacy of single drug and combination therapy against mycobacterium abscessus

PDF

Social isolation anxiety induced synaptic failure restored by dihydromyricetin

PDF

Polymyxins retain in vitro activity and in vivo efficacy against “resistant” Acinetobacter baumannii strains when tested in physiological conditions

PDF

Inflammatory responses during infection: implications for neurodegeneration

Asset Metadata

Creator Agnello, Melissa (author)

Core Title Biological impact of fluoroquinolone resistance in Pseudomonas aeruginosa

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Clinical and Experimental Therapeutics

Publication Date 07/20/2015

Defense Date 05/13/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag antibiotic resistance,fitness,fluoroquinolones,OAI-PMH Harvest,Pseudomonas Aeruginosa

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Finkel, Steven (committee chair), Wong-Beringer, Annie (committee chair), Beringer, Paul (committee member), Rodgers, Kathleen E. (committee member)

Creator Email agnellom@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-600617

Unique identifier UC11300472

Identifier etd-AgnelloMel-3648.pdf (filename),usctheses-c3-600617 (legacy record id)

Legacy Identifier etd-AgnelloMel-3648-0.pdf

Dmrecord 600617

Document Type Dissertation

Format application/pdf (imt)

Rights Agnello, Melissa

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA