SURVIVAL OF UROPATHOGENIC ESCHERICHIA COLI UNDER PROTOZOAN
PREDATION
A Thesis
Presented to the faculty of the Department of Biological Sciences
California State University, Sacramento
Submitted in partial satisfaction of
the requirements for the degree of
MASTER OF SCIENCE
in
Biological Sciences
(Molecular and Cellular Biology)
by
Peter Chee Wai Yu
FALL
2013
© 2013
Peter Chee Wai Yu
ALL RIGHTS RESERVED
ii
SURVIVAL OF UROPATHOGENIC ESCHERICHIA COLI UNDER PROTOZOAN
PREDATION
A Thesis
by
Peter Chee Wai Yu
Approved by:
__________________________________, Committee Chair
Susanne W. Lindgren, Ph.D
__________________________________, Second Reader
Nicholas N. Ewing, Ph.D
__________________________________, Third Reader
Enid T. Gonzalez-Orta, Ph.D
________________________
Date
iii
Student: Peter Chee Wai Yu
I certify that this student has met the requirements for format contained in the University
format manual, and that this thesis is suitable for shelving in the Library and credit is to
be awarded for the thesis.
_________________________, Graduate Coordinator
Jamie M. Kneitel, Ph.D
Department of Biological Sciences
iv
_________________
Date
Abstract
of
SURVIVAL OF UROPATHOGENIC ESCHERICHIA COLI UNDER PROTOZOAN
PREDATION
by
Peter Chee Wai Yu
The ability of bacteria to persist in the environment is a critical step in the
transmission of many pathogens. The protozoan, Acanthamoeba is a known bacterial
predator in the environment that is also host to several human pathogens, including
Legionella pneumophila and enterohaemorrhagic Escherichia coli (EHEC). Recent
research has shown that virulence factors required during human pathology are also
crucial for the survival of these pathogens in environmental protozoans. Although
significant research has been done on L. pneumophila lifestyle in protozoa, and initial
research has begun on diarrheagenic EHEC, uropathogenic E. coli (UPEC) has yet to be
examined. UPEC is the leading cause of community acquired urinary tract infections in
the United States and accounts for 70-95% of these cases. This study examines the
predator-prey interactions of Acanthamoeba castellanii and UPEC as a possible source
for the environmental persistence of UPEC. The primary hypothesis of this study is that
UPEC, with its suite of virulence factors, has a fitness advantage over non-pathogenic
human enteric commensal E. coli when dealing with predation. The goal of this research
was to determine if there are differences in fitness between uropathogenic and nonv
pathogenic strains of E. coli by examining populations after long-term co-culture, as well
as, investigating whether Acanthamoeba can act as an environmental reservoir for UPEC.
This research demonstrates that in long-term co-cultures, clinical isolates of UPEC
exhibit a greater ability to survive predation with no detriment to amoebal growth. In
addition, UPEC isolates demonstrated lower ability to associate with A. castellanii than
the non-pathogenic E. coli, with no clinical isolate being able to survive phagocytosis by
A. castellanii. Clinical isolates of UPEC had greater fitness than a non-pathogenic E. coli
K-12 strain, with the enhanced fitness likely due to avoidance of predation instead of
cytotoxicity towards or invasion of protozoa in a low nutrient environment.
__________________________________, Committee Chair
Susanne W. Lindgren, Ph.D
vi
ACKNOWLEDGEMENTS
I would like to thank the technical staff of the Biology Department, Nancy
Burford, Roxanne Philips, and Sulie Ober for being so accommodating with use of
equipment and supplies throughout my research. I would also like to thank Dr. Susanne
Lindgren for being so patient and understanding with this project. It has been a long
process and I am grateful for all the help that the Biological Science technical staff and
faculty have given me over the years.
Thank you to the staff of the Clinical Microbiology section of the University of
Davis Medical Center Department of Pathology; especially Lynda Braun, Lynne Lenhart,
and Janet Kashiwada for providing the clinical isolates of Escherichia coli associated
with urinary tract infections. Additionally, I have to thank Kim Orth and Herman
Gonzalez of University of Texas Southwestern for providing the Acanthamoeba
castellanii strain used in this study.
I would like to acknowledge my family for supporting me repeatedly over the last
few years. Without their continued moral support it would have been difficult to finish
my excursion into graduate education. Especially to my sisters, Emily and Shirley and
my brother, George; I want to thank you for being there.
vii
TABLE OF CONTENTS
Page
Acknowledgements ........................................................................................................... vii
List of Tables ..................................................................................................................... ix
List of Figures ......................................................................................................................x
INTRODUCTION ...............................................................................................................1
MATERIALS AND METHODS .........................................................................................7
RESULTS ..........................................................................................................................17
DISCUSSION ....................................................................................................................47
CONCLUSIONS................................................................................................................54
Literature Cited ..................................................................................................................58
viii
LIST OF TABLES
Page
Table 1. Protozoa and Bacterial Strains used......................................................................8
Table 2. Fitness and Association comparisons of Clinical E. coli isolates
to E. coli K-12, strain C600. ...............................................................................52
ix
LIST OF FIGURES
Page
Figure 1. Percent survival of Strains PYUCD01 to PYUCD05........................................19
Figure 2. Percent survival of Strains PYUCD06 to PYUCD10........................................20
Figure 3. Percent survival of Strains EHEC EDL 932 and CTMDRUPEC. ................................................................................................................21
Figure 4. Bacterial population of Strains PYUCD01 to PYUCD05 cocultured with A. castellanii ...............................................................................23
Figure 5. Bacterial population of Strains PYUCD06 to PYUCD10 cocultured with A. castellanii ................................................................................24
Figure 6. Bacterial population of Strains EHEC EDL 932 and CTMDRUPEC co-cultured with A. castellanii ...............................................................25
Figure 7. Percent A. castellani population change during co-culture with
Strains PYUCD01 to PYUCD05 .......................................................................27
Figure 8. Percent A. castellani population change during co-culture with
Strains PYUCD06 to PYUCD10 .......................................................................28
Figure 9. Percent A. castellani population change during co-culture with
Strains EHEC EDL 932 and CTMDR-UPEC ...................................................29
Figure 10. A. castellanii population during co-culture with Strains
PYUCD01 to PYUCD05. ................................................................................31
x
Figure 11. A. castellanii population during co-culture with Strains
PYUCD06 to PYUCD10. ................................................................................32
Figure 12. A. castellanii population during co-culture with Strains EHEC
EDL 932 and CTMDR-UPEC. ........................................................................33
Figure 13. Percent E. coli associated with A. castellanii recovered during
association assays with Strains PYUCD01 to PYUCD05,
EHEC EDL 932 and CTMDR-UPEC ..............................................................35
Figure 14. Percent E. coli associated with A. castellanii recovered during
association assays with Strains PYUCD06 to PYUCD10 ...............................36
Figure 15. Ratio E. coli per A. castellanii recovered during association
assays with Strains PYUCD01 to PYUCD05, EHEC EDL 932
and CTMDR-UPEC .........................................................................................37
Figure 16. Ratio E. coli per A. castellanii recovered during association
assays with Strains PYUCD06 to PYUCD10 ..................................................38
Figure 17. Percent E .coli invaded or phagocytized by A. castellanii
recovered during invasion and intracellular survival assays
with Strains C600, PYUCD03, and PYUCD09 ..............................................43
Figure 18. Ratio E. coli per A. castellanii recovered during invasion and
intracellular survival assays with Strains C600, PYUCD03, and
PYUCD09 ........................................................................................................44
xi
1
INTRODUCTION
Environmental persistence of pathogens is a major concern, as environmental
reservoirs can act as continuous sources of infection in a community. While the vast
majority of research on pathogens focuses on the pathology within the human host,
understanding the life cycle in the environment is just as warranted. Examining
environmental persistence would aid in understanding sources of infection and how
pathogen populations are sustained. A significant amount of research has been done on
Legionella pneumophila and its relationship with its environmental host species
Acanthamoeba (Fields et al. 1984; Barker & Brown 1994). Researchers have also begun
examining the relationship of amoeba with enterohaemorrhagic Escherichia coli (EHEC)
(Barker et al. 1999), with the theory that virulence factors aid survival of these pathogens
inside of amoeba. In this new field, little research has been performed on uropathogenic
Escherichia coli (UPEC) with regard to how they persist in the environment. It can be
hypothesized that UPEC, with its unique suite of virulence factors, experiences some
fitness advantage over non-pathogenic E. coli when under predation by amoeba. The
goal of this study was to examine and compare the survival and persistence of UPEC
versus non-pathogenic E. coli under A. castellanii predation.
To understand environmental persistence of pathogenic bacteria, bacteriaprotozoa interactions such as predation and parasitism are of interest. Predation by
protozoa has significance in controlling non-pathogenic E. coli populations in the wild as
E. coli is a viable food source for many free-living protozoans (Watson et al. 1981;
2
Weekers et al. 1993; de Moraes & Alfieri 2008). Researchers have previously
demonstrated that in the presence of wild protozoa, E. coli populations are significantly
lower than populations without predation (Enzinger & Cooper 1976).
In order to mimic natural predator-prey interactions, Acanthamoeba castellanii
was chosen as the model predator for this study. Acanthamoeba is a species of freeliving amoeboid protozoan that is ubiquitous in the environment, but is also known to be
an opportunistic human pathogen. Human infections with A. castellanii typically present
as keratitis due to contaminated contact lenses, and can lead to life threatening
encephalitis in the immunocompromised (Khan 2006).
The life cycle of Acanthamoeba is well characterized and is divided between the
replicating and feeding trophozoite stage and a resistant, inactive cyst stage (Khan 2006).
Under harsh conditions, such as starvation or extreme temperatures, cysts are formed.
Acanthamoeba cysts are capable of remaining viable for years and can be transmitted
through the air before reemerging as a trophozoite once conditions become more
favorable (Khan 2006). As mentioned, A. castellanii is a known predator of nonpathogenic E. coli and has been co-cultured with these bacteria in numerous studies as a
model predator (Alsam et al. 2006; Pickup et al. 2007).
In some instances, the bacteria-protozoa interaction can result in parasitism,
where contact with protozoa does not result in ingestion by the amoeba but results in
active invasion on the part of the bacterium. The classic model for this parasitic-type
interaction is L. pneumophila and its use of protozoa to replicate and spread (Fields et al.
1984). Intracellular parasitism promotes the survival and spread of Legionella by forcing
3
a protozoan to act as an environmental host, and can effectively function as a Trojan
horse for transmission to humans (Barker & Brown 1994). By living inside of
Acanthamoeba, pathogenic Legionella are able to survive in the water supply and are
refractive to most water purification treatments (Snelling et al. 2006). Parasitism of
amoeba has also been observed in other pathogenic bacteria. For example, invasive E.
coli K1, which causes neonatal meningitis, was able to survive and replicate in
Acanthamoeba (Jung et al. 2007). These researchers found that the expression of the K1
capsule, that surrounds these bacteria, enhanced the microbes’ ability to invade and
survive within Acanthamoeba. Similar to E. coli K1, E. coli O157, an EHEC strain that
is commonly characterized as the most important diarrheagenic outbreak strain in the
United States, has been found to replicate readily within Acanthamoeba polyphaga
(Barker et al. 1999).
The life cycle of pathogenic E. coli in the environment warrants study, as
contaminated water and soil are sources of infection (Fremaux et al. 2008), and each
species would logically need to possess adaptations to persist in water and soil occupied
by bacterial predators and competitors. Shiga toxin-producing E. coli (STEC), a group of
bacteria related to EHEC, for example, are capable of surviving for extended periods of
time in water but when placed in competition with indigenous aquatic microflora, they
experience a significant reduction in ability to persist (Wang & Doyle 1998). Several
studies have found that under protozoan predation, virulent E. coli strains experience
enhanced survival compared to avirulent strains (Alsam et al. 2006 and Jung et al. 2007),
which suggests that virulence factors do play a role in their interactions. Virulence
4
factors from STEC, like those of invasive E. coli K1, have been documented to enhance
survival in the presence of protozoan predators by means of toxicity towards predators
and the ability to survive ingestion (Jung et al. 2007; Steinberg & Levin 2007; Lianhart et
al. 2009).
Another enteric group of E. coli that warrants study is the extraintestinal
pathogenic E. coli, UPEC. UPEC was chosen as the main subject of this study due to the
relative lack of research into its environmental persistence and the presence of a range of
virulence factors that could enhance environmental fitness. As UPEC is the number one
cause of community acquired urinary tract infections (Foxman 2003), understanding
UPEC ability to survive in the environment would be beneficial to its control. UPEC
typically inhabits the gastrointestinal tract of humans, but causes disease when they
colonize the urinary tract (Wiles et al.2008). UPEC is distinct from EHEC with regards
to the site of infection and virulence factors employed for pathology. Additionally,
UPEC, EHEC, and commensal E. coli strains share only 39.2% of their genomic
sequence (Brzuszkiewicz et al. 2006). Much of the observed differences between the
genomic sequences is located among pathogenicity islands (PAIs) unique to each strain,
and within these PAIs are a wide range of virulence factors UPEC uses to successfully
colonize the urinal tract (Brzuszkiewicz et al. 2006). The majority of UPEC virulence
research has focused on pathology of human cells, with their influence on protozoa being
relatively unknown. Previous research has shown that virulence factors of a variety of
bacterial pathogens elicit similar responses in protozoa as they do in human cells (Segel
& Shuman 1999, Steinberg & Levin 2007, and Lianhart et al. 2009), thus suggesting that
5
UPECs unique PAIs may have a role in bacteria-protozoa interactions.
UPEC is a heterogeneous group of E. coli, with strains expressing a wide variety
of virulence factors (Wiles et al. 2008). UPEC secretes a variety of siderophores to aid in
iron acquisition from the host in the urinary tract environment, which may confer to
UPEC nutrient advantage in the extraintestinal environment. Adhesins, such as FimH,
that are required for attachment to host uroepithelial cells may have applications for
initiating adherence to host cells preceding invasion. A small selection of toxins are also
employed by UPEC, which includes α-hemolysin, autotransporter toxin and cytotoxic
necrotizing factor 1 (CNF-1) (Wiles et al. 2008 and Bower et al. 2005). These toxins
cause cell death of epithelial cells but might also act as toxins against protozoan predators
in the environment. Additionally, colibactin, a recently discovered bacterial polyketide in
extraintestinal pathogenic E. coli that is constitutively expressed as a membrane bound
molecule, was capable of inducing dramatic contact-dependent cytopathic effects on
human epithelial cells (Nougayrede et al. 2006). This colibactin molecule was found in
53% of extraintestinal pathogenic E. coli and may have a similar cytopathic effect against
predatory protozoa.
In contrast to cytotoxicity, invasion and persistence within eukaryotic cells is an
alternate method for UPEC to interact with protozoa. Recently, it has been suggested that
UPEC is capable of invading host uroepithelial cells to form intracellular bacterial
communities (IBC) (Justice et al. 2004). UPEC’s ability to invade and form IBCs are
heavily virulence factor dependent, with adhesins and polysaccharide capsule being just a
few of the factors involved in the process (Reigstad et al. 2007; Nicholson et al. 2009;
6
Anderson et al. 2010). As seen with invasive E. coli K1, capsular expression holds
protective and invasive properties in bacterial interactions with amoeba, which suggests
that UPEC may share similar capabilities. The intracellular capability of UPEC, although
studied within human epithelial cells, suggests the possibility that UPEC could also be an
intracellular parasite of amoeba.
As a whole, various studies have shown that virulence factors, known to have a
role in human pathology, also play a part in the environmental persistence of a pathogen.
Therefore, it can be hypothesized that the suite of virulence factors employed by UPEC
in human disease may confer a fitness advantage to them in an environment with
predation. As bacteria to amoebal interactions have never been examined in UPEC, our
study will focus on characterizing fitness and associative abilities of UPEC in comparison
to non-pathogenic E. coli. To examine the possibility of UPEC virulence factor
involvement in bacteria-protozoa interactions, a broad range of newly isolated E. coli that
were attributed to urinary tract infections were tested to characterize UPEC. This
research examined the interactions of predatory protozoa with fresh clinical UPEC
isolates in co-culture, with the objective to determine whether there were any differences
in fitness over time between UPEC and non-pathogenic E. coli. The second objective
was to determine whether UPEC can utilize protozoa as an environmental reservoir.
7
MATERIALS AND METHODS
Escherichia coli, bacterial growth media, and culture techniques
All strains used in this study are listed in Table 1. Three defined bacterial strains
were used in this study: CTMDR-UPEC, a clinical isolate of a multiple drug resistance
UPEC obtained from the Sacramento County Public Health Laboratory (Lindgren
laboratory collection); EHEC O157:H7 strain CDC EDL 932 (Escherichia coli ATCC
43894), a clinical isolate from the first documented O157:H7 outbreak (Wells et al.1983);
and C600, a non-pathogenic general purpose cloning strain of E. coli K-12. For all coculture assays, C600 was used as the non-pathogenic standard, while O157:H7 strain
EDL 932 and CTMDR-UPEC were used as two pathogenic standards. Additionally, ten
new clinical isolate strains of E. coli from urinary tract infections, were obtained from
UCDMC during the summer of 2011, and were designated PYUCD01 through
PYUCD10.
Subcultures of experimental E. coli were performed monthly on LB agar plates
and stored at 4°C. All experimental E. coli strains were cultured overnight in LuriaBertani (LB) broth at 37°C prior to experimental use. For co-culture experiments,
overnight cultures of bacteria were prepared by washing cultures twice in Amoeba
Infection Media (AIM) and centrifuged at 1500g for five minutes. An estimate of E. coli
cellular density was made by spectrophotometer, and washed E. coli was diluted with
AIM to approximately 3x108 CFU/ml for co-culturing with amoeba.
8
Organism
Strain Designation
Reference
ATCC 30234
Daggett et al. 1982
E. coli K-12
C600
Appleyard 1954
EHEC O157:H7
ATCC 43894 (CDC EDL 932)
Wells et al. 1983
UPEC
CTMDR-UPEC b
This study
UPEC
PYUCD01 c
This study
UPEC
PYUCD02
This study
UPEC
PYUCD03
This study
UPEC
PYUCD04
This study
UPEC
PYUCD05
This study
UPEC
PYUCD06
This study
UPEC
PYUCD07
This study
UPEC
PYUCD08
This study
UPEC
PYUCD09
This study
UPEC
PYUCD10
This study
Acanthamoeba
Acanthamoeba castellanii
Escherichia coli a
Table 1. Protozoa and Bacterial Strains used
a
EHEC – Enterohaemorrhagic Escherichia coli; UPEC – Uropathogenic Escherichia coli
b
CTMDR-UPEC is a clinical isolate obtained from Sacramento County Public Health
Laboratory, Lindgren laboratory collection.
c
All PYUCD strains are clinical isolates obtained from UC Davis Medical Center from
urinary tract infections collected Summer 2011.
9
Acanthamoeba castellanii growth media and culture techniques
All handling of A. castellanii was restricted to the Class II biosafety cabinet in a
BL2 laboratory to prevent exposure of individuals to A. castellanii and to reduce
contamination of eukaryotic cell cultures. A. castellanii trophozoites (ATCC 30234)
were cultured axenically in 25 cm2, vented tissue culture flask in ATCC Medium 712 at
room temperature (20-25°C) without agitation. Subcultures of the amoeba were
performed every two weeks by removing spent media and adding fresh sterile media to
the flask. The flasks were gently tapped to detach cells and then diluted at a ratio of 1:5
into a new flask of fresh PYG medium.
ATCC Medium 712 is composed of a basal PYG media, glucose stock solution
and five inorganic salt solutions. Basal PYG media consists of 20 g proteose peptone and
1 g yeast extract brought to 900 ml volume in distilled deionized water that was then
autoclaved for sterility. Glucose stock solution is composed of 18 g glucose and 1 g
sodium citrate (Na3C6H5O7•2H20) brought to 50 ml volume in distilled deionized water
before being filter sterilized through a 0.45 µm cellulose acetate filter. Basal media was
made “complete” with the addition of the glucose stock solution and inorganic salts.
Sterile inorganic salt solutions in distilled deionized water were added at the following
volumes and concentrations: 10 ml of 0.4 M MgSO4•7H2O, 8 ml of 0.05 M CaCl2, 10 ml
of 0.005 M Fe(NH4)2(SO4)2•6H2O, 10 ml of 0.25 M Na2HPO4•7H2O, and 10 ml of 0.25
M KH2PO4. All inorganic salt solutions were heat sterilized by autoclave, except for
0.005 M Fe(NH4)2(SO4)2•6H2O. The 0.005 M Fe(NH4)2(SO4)2•6H2O solution was filter
sterilized through a 0.45 µm cellulose acetate filter prior to use.
10
For co-culture assays and washes, a non-nutrient solution was prepared to
promote predator-prey interactions over long-term co-culture. This solution was
designated Amoeba Infection Media (AIM) and was composed of 10 ml of 0.4 M
MgSO4•7H2O, 8 ml of 0.05 M CaCl2, 10 ml of 0.005 M Fe(NH4)2(SO4)2•6H2O, 10 ml of
0.25 M Na2HPO4•7H2O, and 10 ml of 0.25 M KH2PO4 added to 950 ml of an autoclaved
solution of distilled deionized water with 1 g sodium citrate (Na3C6H5O7•2H20).
For all experiments utilizing protozoa, the adherent trophozoites forms were
harvested by gently tapping flasks to resuspend cells. For fitness assays, harvested A.
castellanii were centrifuged at 600g for five minutes and washed with sterile AIM; this
wash was repeated a second time before 0.5 ml of washed A. castellanii was aliquoted
into each well of a 24-well culture plates. Plates were allowed to settle for 24 hours at
room temperature (20-25°C) before experimentation to allow A. castellanii to adhere to
the well surface and become acclimated to AIM. In association and invasion assays, A.
castellanii were grown to confluence in PYG on 24-well plates at room temperature (2025°C) for two days prior to co-culture. Plates were washed once with AIM 24 hours
before experimentation and suspended in 0.5 ml AIM to acclimate A. castellanii to AIM.
Long-Term Predator-Prey Population Assay
For the purposes of this study, an assay was established in order to determine the
impact of long-term predator-prey interactions between UPEC and Acanthamoeba. To
this end, approximately 1.5x105 cells/ml A. castellanii was co-cultured with
approximately 1.5x108 CFU/ml E. coli for three days. The ratio of 1:1000 protozoa to
11
bacteria was set as the baseline for this assay, similar to levels of co-culture used by
Huws et al. (2008) for observing A. polyphaga predation on common pathogenic bacteria.
Population experiments were performed in AIM to simulate low nutrient conditions and
to encourage predation. Briefly washed A. castellanii, were resuspended in AIM and
adjusted to 3x105 cells/ml before being aliquoted at 0.5 ml/well into 24-well plates.
Amoebae were allowed to settle for 24 hours at room temperature (20-25°C). Every
clinical E. coli isolate was tested in triplicate, with each well treated with 0.5 ml of
overnight E. coli culture, washed twice as described above and resuspended in AIM to a
concentration of approximately 3x108 CFU/ml. Each independent experiment included
the non-pathogenic E. coli K-12 C600 strain as a control alongside clinical UPEC
isolates.
Wells were sampled at day 0, immediately after inoculation, and at day 3 post coculture. Plates were incubated between samplings at room temperature (20-25°C)
without agitation. At every time point, each co-culture sample was enumerated for
population of A. castellanii by hemocytometer, and for population of E. coli by plate
count using serial dilution and spread plating of 100 µl of suspension onto LB nutrient
agar plates. Bacterial fitness was assessed by percent survival of bacteria after three days
co-culture. Percent survival of bacteria was calculated by E. coli (CFU/ml) recovered at
day 3 divided by average initial E. coli (CFU/ml) sampled at day 0 times 100. Response
of A. castellanii to bacterial co-culture was assessed by percent population change of
amoeba by calculating A. castellanii (cells/ml) recovered at day 3 divided by average
initial A. castellanii (cells/ml) at day 0 times 100.
12
Short-term Predator-Prey Association Assays
Three different short-term predator-prey assays were performed to evaluate
association, invasion, and intracellular interactions of E. coli with A. castellanii. The
short-term predator-prey assays were performed according to methods detailed by Alsam
et al. (2006) for co-culture, with the exception that throughout all assays, instead of PBS,
sterile AIM was used as the medium for culturing and washing. Final bacterial
concentrations of 1.5x108 CFU/ml were used for all short-term assays. Each independent
experiment included the non-pathogenic E. coli K-12 C600 strain as a control beside
clinical UPEC isolates.
Association assay - Association assays were done in order to evaluate E. coli
adherence to protozoa after co-culturing for two hours. To prepare A. castellanii for coculture, amoeba were grown at room temperature (20-25°C) until confluent on 24-well
plates, and washed once with AIM 24 hours prior to co-culture. Each well was filled
with 0.5 ml AIM post-wash. Every clinical E. coli isolate was tested in triplicate, with
each well treated with 0.5 ml of overnight E. coli culture, washed twice with AIM and
resuspended in AIM to a concentration of approximately 3x108 CFU/ml. Each
independent experiment included the non-pathogenic E. coli K-12 C600 strain as a
control alongside clinical UPEC isolates. Co-cultures were allowed to incubate for two
hours at room temperature. After incubation, each well was gently washed with AIM
three times to remove non-adherent bacteria and resuspend in one ml of AIM. Amoeba
in each sample were enumerated with a hemocytometer by removing 0.5 ml from each
well, with the remaining sample lysed with the addition of 0.5 ml of 1% SDS for 20
13
minutes to release bound and intracellular bacteria. Recovered bacteria were enumerated
by serial dilution in AIM and plated onto LB nutrient agar plates. To account for loss of
A. castellanii during the washing step, which might affect recovery of bacteria, the
bacterial association was assessed using two calculations. First, percent association for
determining the overall amount of the E. coli inoculum associated with or being
internalized by A. castellanii was determined by calculation of recovered E. coli
(CFU/ml) divided by final concentration E. coli (CFU/ml) used in inoculation. Second,
the E. coli to A. castellanii ratio for assessing number of bacterium attached to each
amoeba, was calculated by recovered E. coli (CFU/ml) divided by recovered A.
castellanii.
Invasion assay - Invasion assays were used to determine the ability of bacteria to
invade protozoa. To prepare A. castellanii for co-culture, amoebas were grown with
PYG at room temperature (20-25°C) until confluent on 24-well plates. Wells were
washed once with AIM, 24 hours prior to co-culture. Each well was filled with 0.5 ml
AIM, post-wash. Every clinical E. coli isolate was tested in triplicate, with each well
treated with 0.5 ml of overnight E. coli culture, washed twice with AIM and resuspended
in AIM to a concentration of approximately 3x108 CFU/ml. Each independent
experiment included the non-pathogenic E. coli K-12 C600 strain as a control alongside
clinical UPEC isolates. To compare invasive capabilities between clinical isolates, cocultures were first centrifuged, to synchronized contact of bacteria to amoeba, at 600g for
five minutes and allowed to co-culture for one hour at room temperature. Wells were
then washed twice with AIM and treated with 100 μg/ml gentamicin for 45 minutes to
14
kill any extracellular bacteria. After gentamicin treatment, wells were washed once with
AIM to remove the gentamicin. Amoeba in each sample were then enumerated by
hemocytometer by removing 0.5 ml from each well, and then the remaining sample was
lysed with the addition of 0.5 ml of 1% SDS for 20 minutes to release bound and
intracellular bacteria. Recovered bacteria were enumerated by serial dilution in AIM and
plated onto LB nutrient agar plates. Similar to the association assay where loss of
amoeba due to repeated washes can skew bacterial counts, bacterial invasion was
assessed by two calculations. First, percent invasion was determined by calculation of
recovered E. coli (CFU/ml) divided by total concentration E. coli (CFU/ml) used in
inoculation. Second, the E. coli to A. castellanii ratio was calculated by recovered E. coli
(CFU/ml) divided by recovered A. castellanii.
Intracellular survival assay - Intracellular survival assays were used to determine
the viability of E. coli within protozoa over time. To prepare A. castellanii for co-culture,
amoeba were grown at room temperature (20-25°C) until confluent on 24-well plates and
washed once with AIM, 24 hours prior to co-culture. Each well was filled with 0.5 ml
AIM post-wash. Every clinical E. coli isolate was tested in triplicate, with each well
treated with 0.5 ml of overnight E. coli culture, washed twice with AIM and resuspended
in AIM to a concentration of approximately 3x108 CFU/ml. Each independent
experiment included the non-pathogenic E. coli K-12 C600 strain as a control alongside
clinical UPEC isolates. Invasion was synchronized by centrifugation at 600g for five
minutes and allowed to co-culture for one hour at room temperature. Wells were then
washed twice with AIM and treated with 100 μg/ml gentamicin for 45 minutes to kill
15
extracellular bacteria. As with the invasion assay stated above, after gentamicin
treatment, wells were washed once with AIM to remove the gentamicin; however, wells
were then allowed to incubate further. Viable bacteria remaining in the wells were
enumerated at three hours and 24 hours post gentamicin treatment. As with the invasion
assay, A. castellanii in each sample were enumerated by hemocytometer by removing 0.5
ml from each well, and then the remaining sample was lysed with the addition of 0.5 ml
of 1% SDS for 20 minutes to release intracellular bacteria. Recovered bacteria were
enumerated by serial dilution in AIM and plated onto LB nutrient agar plates. Percent
intracellular survival was determined by calculation of recovered E. coli (CFU/ml)
divided by total concentration E. coli (CFU/ml) used in inoculation. The E. coli to A.
castellanii ratio was calculated by recovered E. coli (CFU/ml) divided by recovered A.
castellanii.
Gentamicin Minimum Inhibitory Concentration (MIC) assay
Because gentamicin was used to evaluate intracellular invasion and survival of
clinical UPEC isolates, antibiotic susceptibility of these strains to gentamicin was
assessed. Overnight bacterial cultures were prepared in LB broth at 37°C. Five tubes of
varying antibiotic concentrations were made by serial two-fold dilutions of gentamicin.
A stock gentamicin solution of 50 mg/ml was diluted to generate the first tube of 10 ml of
400 µg/ml gentamicin. Five ml of the 400 µg/ml gentamicin solution were transferred to
five ml of LB broth to create the next dilution. Five 2-fold serial dilutions were made to
final concentrations of 400 µg/ml, 200 µg/ml, 100 µg/ml, 50 µg/ml and 25 µg/ml of
16
gentamicin in LB broth. All tubes were inoculated with 20 µl of overnight bacterial
culture and incubated overnight at 37°C. Observation of the lowest concentration that
prevented growth, as observed by turbidity, was noted as the MIC of gentamicin.
Statistical Analysis of Data
All data analysis employed the statistical program SPSS Statistics 17.0 with result
analyzed by one-way ANOVA or paired t-test. Post hoc analysis was set to Bonferroni
assuming equal variances with a significance level of 0.05. Post hoc analysis focused on
comparing clinical isolates to the non-pathogenic C600 and to wells with no treatment.
Graphs were generated in SPSS Statistic 17.0 and presented as mean value with error
bars ± two Standard Error of Mean (SEM).
17
RESULTS
Co-culture Fitness of Clinical Isolates
In order to assess the fitness of UPEC under protozoan predation, long-term cocultures were established to compare the survival of pathogenic and non-pathogenic
strains with A. castellanii. Experiments were designed to culture several E. coli strains
concurrently, and examined the survival capabilities of each strain relative to another by
comparing percent bacterial survival and change in amoebal population of A. castellanii.
Percent survival of bacteria after three days of co-culture was used to assess the viability
of UPEC under predation, with values normalized for variations in inoculum
concentration. Percent population change in amoeba was used to assess the ability of A.
castellanii to use co-cultured bacteria as a food source and the ability of bacteria to prey
on the amoeba. If virulence plays a role in predator-prey interactions of UPEC, we
would expect the pathogenic UPEC to show greater survival in co-culture, as well as,
support less amoebal growth than non-pathogenic C600. If there are no benefits to
possessing virulence factor associated with human pathology, we would expect to see
UPEC survive at rates equal to or less than that of C600, in addition to supporting
amoeba growth.
For each bacterial strain, co-culture experiments were performed in triplicate over
the course of three independent experiments with C600 as a non-pathogenic control. We
observed that independent experiments varied dramatically, as seen by percent survival of
C600 across three independent experiments. A possible cause for this discrepancy could
18
be due to observations that not all Acanthamoeba inoculated into cell culture wells were
adhering to the well surface, thereby resulting in morphological and behavioral
differences within each amoeba population that could not be controlled between
experiments. It was previously observed by Pickup et al. 2007, that settled and floating
trophozoites exhibit differences in feeding behavior. Even though values varied between
independent experiments, the overall relationships between strains were consistently
observed across three independent experiments. The relative differences between strains
within the same experiment were therefore used to compare overall fitness of strains
towards one another. Comparisons of values across experiments are not recommended
due to variances between independent experiments. All tables are representative of the
relationships observed over three independent experiments.
In order to assess the viability of UPEC over long-term co-culture, percent
bacterial survival of strains after three days of co-culture, were compared with the
survival of a non-pathogenic standard, C600. Significant differences (p < 0.05) in
percent survival of E. coli were observed for most clinical isolates, with the exceptions
being strains PYUCD03 (Figure 1), PYUCD07 and PYUCD09 (Figure 2) and EDL 932
(Figure 3). This suggests that as a group, UPEC tends to have greater fitness in amoebal
co-culture than the non-pathogenic C600 and EHEC EDL 932. In these long-term fitness
experiments, no isolate exhibited significantly lower percent survival than C600 (Figures
1, 2 and 3). This indicates that UPEC isolates either had greater or equivalent fitness
with the non-pathogenic C600. Every E. coli strains examined during long-term coculture were observed to have negligible change in population when cultured over seven
19
Figure 1. Percent survival of Strains PYUCD01 to PYUCD05. Figure 1 is
representative of three independent co-culture experiments, with each strain tested
in triplicate. Error bars represent mean ± 2 SEM. PYUCD01, PYUCD02,
PYUCD04 and PYUCD05 showed statistically significant difference compared to
C600 (p < 0.05).
20
Figure 2. Percent survival of Strains PYUCD06 to PYUCD10. Figure 2 is
representative of three independent co-culture experiments with each strain tested
in triplicate triplicate Error bars represent mean ± 2 SEM. PYUCD06, PYUCD08
and PYUCD10 showed statistically significant different compared to C600 (p <
0.05).
21
Figure 3. Percent survival of Strains EHEC EDL 932 and CTMDR-UPEC.
Figure 3 is representative of three independent co-culture experiments with each
strain tested in triplicate. Error bars represent mean ± 2 SEM. CTMDR-UPEC
showed statistically significant different compared to C600 (p < 0.05).
22
days in AIM alone, without amoebal predation (data not shown). The decreased
population exhibited by each E. coli strain during amoebal co-cultures can therefore be
attributed to amoeboid predation and not to population decline. To further clarify the
impact predation by A. castellanii had on E. coli strains, Figures 4, 5 and 6 were
included. Figures 4, 5 and 6 presents the population counts of each strain at day 0 and
day 3 of the long-term fitness assay, and displayed a two-log drop in population among
all bacterial strains. This further suggests protozoan predation was occurring in bacterial
co-culture.
In contrast to monitoring bacterial population change, bacterial fitness can also be
gauged by whether bacteria were detrimental or beneficial to amoebal growth during coculture. In order to determine the viability of A. castellanii, we monitored amoeba
population change after three days of co-culture. By examining changed in amoeba
population we can determine whether Acanthamoeba benefitted from co-culture with
UPEC or were preyed upon by the pathogenic E. coli. To elucidate the impact each
UPEC strain had on amoeba population, we analyzed the percent change in amoebal
population versus results from amoeba co-cultured with C600. If UPEC were less fit
than C600 we would expect amoebal growth to be greater, due to higher bacterial
predation and utilization of bacteria as a food source. Conversely, greater bacterial
fitness would be represented by a lower change in amoeba population due to lower
bacterial predation. Comparisons of percent change in amoeba population of clinical
isolates against no bacterial treatment were also used to ensure any change in amoeba
population could be attributed to the bacterial co-culture.
23
Figure 4. Bacterial population of Strains PYUCD01 to PYUCD05 co-cultured
with A. castellanii. Figure 4 is representative of results from three independent
co-culture experiments with each strain tested in triplicate. Values depict the
Log10 of the mean population taken at day 0 and day 3 of co-culture. Open bars
represent values from day 0 and filled bars represent values from day 3. Error
bars represent mean ± 2 SEM.
24
Figure 5. Bacterial population of Strains PYUCD06 to PYUCD10 co-cultured
with A. castellanii. Figure 5 is representative of results from three independent
co-culture experiments with each strain tested in triplicate. Values depict the
Log10 of the mean population taken at day 0 and day 3 of co-culture. Open bars
represent values from day 0 and filled bars represent values from day 3. Error
bars represent mean ± 2 SEM.
25
Figure 6. Bacterial population of Strains EHEC EDL 932 and CTMDR-UPEC
co-cultured with A. castellanii. Figure 6 is representative of results from three
independent co-culture experiments with each strain tested in triplicate. Values
depict the Log10 of the mean population taken at day 0 and day 3 of co-culture.
Open bars represent values from day 0 and filled bars represent values from day
3. Error bars represent mean ± 2 SEM.
26
Percent population change of amoeba in co-culture is shown in Figures 7, 8 and 9.
Each figure is representative of percent change in A. castellanii, from initial inoculation
to three days of co-culture with E. coli. Results were attained from three independent
experiments done in triplicate for each E. coli strain. As noted in Figures 7, 8 and 9, no
population growth was detected in the no bacterial treatment control, which indicated that
co-culture media, AIM, was not a source of nutrients for A. castellanii and did not
contribute to amoebal growth. For all strains, co-culture of E. coli with A. castellanii
resulted in a significance increase in A. castellanii population compared to no bacterial
treatment (Figures 7, 8 and 9) (p < 0.05). The exceptions to this were strains PYUCD06
and PYUDC07 (Figure 8), which presents amoebal growth levels that were not
considered significantly different from no bacterial treatment. However, further analysis
of the two other experiments represented by Figure 8, indicate that amoebal growth on
PYUCD06 and PYUCD07 were significantly different even though Figure 8 does not
present it (data not shown).
To compare fitness between clinical isolates of E. coli, analysis on percent change
in amoebal population with UPEC and C600 were performed (Figures 7, 8 and 9). In
these comparisons, A. castellanii population growth with different E. coli strains showed
no significant difference between pathogenic isolates and C600 (p < 0.05). The amoeba
population results indicate that although virulent with known cytotoxic pathways, none of
the UPEC strains significantly impacted the growth of A. castellanii in comparison to
C600. No clinical UPEC strains exhibited the ability to decrease amoebae presence and
all appear to support amoebal growth. To further clarify the impact co-culture of E. coli
27
Figure 7. Percent A. castellani population change during co-culture with Strains
PYUCD01 to PYUCD05. Figure 7 is representative of three independent coculture experiments with each strain tested in triplicate. Error bars represent
mean ± 2 SEM. No statistically significant difference was detected among UPEC
strains compared to C600 (p < 0.05). Significant differences were observed
between A.castellanii treated with E. coli compared to no bacterial treatment (p <
0.05).
28
Figure 8. Percent A. castellani population change during co-culture with Strains
PYUCD06 to PYUCD10. Figure 8 is representative of three co-culture
experiments with each strain tested in triplicate. Error bars represent mean ± 2
SEM. No statistically significant difference was detected among UPEC strains
compared to C600 (p < 0.05). Significant differences were observed between A.
castellanii treated with E. coli compared to no bacterial treatment for strains
PYUCD08, PYUCD09 and PYUCD10 (p < 0.05). No statistically significant
difference was observed with PYUCD06 and PYUCD07 compared to no bacterial
treatment.
29
Figure 9. Percent A. castellani population change during co-culture with Strains
EHEC EDL 932 and CTMDR-UPEC. Figure 9 is representative of three coculture experiments with each strain tested in triplicate. Error bars represent
mean ± 2 SEM. No statistically significant difference was detected among these
strain compared to C600 (p < 0.05). Significant differences were observed
between A. castellanii treated with E. coli compared no bacterial treatment (p <
0.05).
30
strains had on A. castellanii populations, Figures 10, 11 and 12 were included. Figures
10, 11 and 12 presents the population counts of A. castellanii at day 0 and day 3 of the
long-term fitness assay, and revealed a half-log increase in population during co-culture
with bacteria. This further suggests that all E. coli were subject to predation and were
consumed. Growth of A. castellanii during co-culture with UPEC isolates is an
indication that cytotoxins, typically employed by UPEC in human pathology (Wiles et al.
2008), does not have a significant impact on amoebal populations. Therefore cytotoxicity
induced by UPEC on amoeba was not further investigated in this study.
Bacterial Association with Acanthamoeba
Although it was clear that UPEC were not cytotoxic towards amoeba over longterm co-cultures, we wanted to assess the ability of bacteria to adhere, invade, and grow
within amoeba. As noted previously in work done by Justice et al. (2004), UPEC have
exhibited invasive capabilities that aid in their persistence in the urinary tract. As UPEC
was observed to have greater survival than C600 during the fitness assays, contact
dependent interactions like adherence and invasion were further investigated. The
association assays were employed to clarify whether the differences in fitness were due
bacteria adhering to, invading, and subsequently persisting within amoeba.
To gauge the associative abilities of UPEC clinical isolates, association assays
were designed to use a known concentration of bacteria to interact with a confluent lawn
of amoeba for two hours, with washes to remove unbound bacteria. Multiple bacterial
strains were evaulated together in triplicate within a experiment over three independent
31
Figure 10. A. castellanii population during co-culture with Strains PYUCD01 to
PYUCD05. Figure 10 is representative of results from three independent coculture experiments with each strain tested in triplicate. Values depict the Log10
of the mean amoeba population taken at day 0 and day 3 of co-culture. Open bars
represent values from day 0 and filled bars represent values from day 3. Error
bars represent mean ± 2 SEM.
32
Figure 11. A. castellanii population during co-culture with Strains PYUCD06 to
PYUCD10. Figure 11 is representative of results from three independent coculture experiments with each strain tested in triplicate. Values depict the Log10
of the mean amoeba population taken at day 0 and day 3 of co-culture. Open bars
represent values from day 0 and filled bars represent values from day 3. Error
bars represent mean ± 2 SEM.
33
Figure 12. A. castellanii population during co-culture with Strains EHEC EDL
932 and CTMDR-UPEC. Figure 12 is representative of results from three
independent co-culture experiments with each strain tested in triplicate. Values
depict the Log10 of the mean amoeba population taken at day 0 and day 3 of coculture. Open bars represent values from day 0 and filled bars represent values
from day 3. Error bars represent mean ± 2 SEM.
34
experiments. For each experiment, C600 was utilized as the non-pathogenic control with
results from representative experiments presented in Figures 13, 14, 15, and 16.
Similar to results seen in long-term co-cultures, the percent E. coli associated with
A. castellanii were variable between experiments, with typically less than 0.5% of the
inoculation dose of approximately 1.5x108 CFU/ml bacteria associating with A.
castellanii. The same problem with variable A. castellanii morphology seen in the fitness
assays may also explain the variance seen between independent association experiments,
and may be exacerbated due to loss of both amoeba and bacteria during wash steps.
Association was measured in two ways. One was by percentage of E. coli
inoculum associated with A. castellanii after two hours co-culture and washing, and the
second by the ratio of E. coli to A. castellanii recovered. Percent E. coli associated with
amoeba was representative of the degree by which the initial inoculation of E. coli came
in contact with and has adhered to amoeba, and accounts for variations in initial E. coli
dose. The ratio of E. coli to A. castellanii is a metric to measure the approximate ratio of
bacteria interacting with amoeba during experimentation and accounts for variations in
recovered amoeba population. If UPEC is capable of utilizing invasion and adherence to
amoeba to improve survival, we would expect to see the strains that exhibited greater
fitness to associate readily to amoeba and have a high association ratio.
Comparisons of the percent bacteria associated with amoeba between pathogenic
and non-pathogenic E. coli, is meant to provides insight into whether UPEC differs in its
ability to come into contact with and remain adherent to amoeba compared to C600.
Significant differences in percent E. coli associated with A. castellanii were only
35
Figure 13. Percent E. coli associated with A. castellanii recovered during
association assays with Strains PYUCD01 to PYUCD05, EHEC EDL 932 and
CTMDR-UPEC. Figure 13 is representative of three co-culture experiments with
each strain tested in triplicate. Error bars represent mean ± 2 SEM. Strains
PYUCD04, EDL 932 and CTMDR-UPEC showed statistically significant
difference compared to C600 (p < 0.05).
36
Figure 14. Percent E.coli associated with A. castellanii recovered during
association assays with Strains PYUCD06 to PYUCD10. Figure 14 is
representative of three co-culture experiments with each strain tested in triplicate.
Error bars represent mean ± 2 SEM. Strains PYUCD08 and PYUCD10 showed
statistically significant difference compared to C600 (p < 0.05).
37
Figure 15. Ratio E. coli per A. castellanii recovered during association assays
with Strains PYUCD01 to PYUCD05, EHEC EDL 932 and CTMDR-UPEC.
Figure 15 is representative of three co-culture experiments with each strain tested
in triplicate. Error bars represent mean ± 2 SEM. All but PYUCD05 strains show
statistically significant difference compared to C600 (p < 0.05).
38
Figure 16. Ratio E.coli per A. castellanii recovered during association assays
with Strains PYUCD06 to PYUCD10. Figure 16 is representative of three coculture experiments with each strain tested in triplicate. Error bars represent
mean ± 2 SEM. Strains PYUCD08 and PYUCD10 showed statistically
significant difference compared to C600 (p < 0.05).
39
observed in three clinicial isolates: PYUCD04, PYUCD8 and PYUCD10, as well as in
EDL 932 and CTMDR-UPEC when compared to C600 (Figures 13 and 14) ( p < 0.05).
These strains showed that a lower percentage of the inoculum was associating with
amoeba compared to the non-pathogenic C600 (Figures 13 and 14). All other strains
appear to have percent bacterial association close to that of C600 (Figures 13 and 14).
With half of the high fitness UPEC showing low association and the remainder showing
equal association to C600, it may be possible that different mechanisms are employed by
UPEC to improve fitness in co-culture. The profile of high fitness with high association
would match how we would expect an invasive UPEC to behave, while a low association
with high fitness would indicate evasion. No strains tested exhibited greater percent
bacterial association compared to C600, which suggest that UPEC were not proactively
associating with amoeba (Figures 13 and 14). As a whole, virulent UPEC clinicial isolate
strains either had equal or less bacterium associated with amoeba then the non-pathogenic
C600.
Ratio of bacteria associated with amoeba may be an more accurate assessment of
association as it accounts for the population of A. castellanii involved. The result from
Figures 15 and 16, indicates that most clinical UPEC strains tested have a significantly
lower bacteria to amoeba ratio than C600. Strains PYUCD01 to PYUCD04, PYUCD8,
PYUCD10, EDL 932 and CTMDR-UPEC had lower ratios of E. coli per amoeba when
compared with C600 (p < 0.05). These results suggest that for most clinical isolates,
fewer E. coli are associated with each A. castellanii cell in comparison to the nonpathogenic C600 strain.
40
In order to fully assess the association results, comparisons to fitness results were
also taken into account. Strains with lower association ratio compared to C600 appear to
be strains that also had greater survival compared to C600. These strains, which may
represent a evasive phenotype, were PYUCD01, PYUCD02, PYUCD04, PYUCD08,
PYUCD10 and CTMDR-UPEC (Figures 1, 2 and 3). Additionally, there were strains that
exhibited association values equal to C600 and showed either greater or equal survival.
PYUCD05 and PYUCD06 were two strains that had greater survival than C600 but equal
association ratio (Figures 1 and 2). PYUCD05 and PYUCD06 results showed high
variance within and between experiments and conclusions on these strains may not be
representative. PYUCD03, PYUCD07, and PYUCD09 were strains that had equal
survival to C600 and equal association ratios (Figures 1 and 2) and are representative of
UPEC that does not differ significantly from C600. Taken together, the results from
these studies demostrate that the UPEC clinical isolates tested fall into two categories,
those showing lower association than E.coli K-12 strain C600 and those exhibiting equal
association. Furthermore, these results present a scenario that suggests that UPEC
clinical isolates are not invasive but rather are adapted to evade predation.
Invasion and Intracellular survival
Results from the association assays indicated that UPEC strains displayed two
distinct association behaviors; either lower or equivalent to C600 (Figures 13, 14, 15, and
16). Two UPEC strains that exhibited higher association and two UPEC strains that
exhibited lower association with A. castellanii were assessed for invasion and
41
intracellular survival to ascertain whether association lead to invasion. If UPEC employs
protozoan invasion as a survival mechanism, we would expect that strains that exhibited
higher associative ability to be internalized and be capable of surviving within A.
castellanii. Conversely, the low associative UPEC would not be expected to invade nor
survive within amoeba.
The invasion and intracellular survival assays employed are a modified
association assay designed to eliminate bacteria that are associated extracellularly to
Acanthamoeba, in order to determine the amount of bacteria phagocytized by amoeba
during association. Invasion and intracellular survival assay enumeration methodologies
were identical to that of the association assay. To assess intracellular bacteria, cultures
were treated with gentamicin after one hour of synchronized association and washing.
Gentamicin, a bacterial aminoglycoside antibiotic that does not penetrate eukaryotic cell
membrane, was used to effectively kill extracellular bacteria. The antibiotic treatment
was designed to eliminate bacteria associated extracellularly to amoeba, with results to be
representative of the intracellular population of bacteria in co-culture assays. Because the
UPEC strains were recent clinical isolates, we confirmed that they were in fact sensitive
to gentamicin at levels used in the invasion assay. All strains were determined to be
gentamicin susceptible at the final concentration of 100 µg/ml (data not shown).
Intracellular survival assays differed from invasion in that sampling occurred at three and
24 hours post wash and are designed to assess bacterial survival after invasion or
phagocytosis by amoeba. Bacterial strains were evaulated concurrently in triplicate, over
three independent experiments. Each experiment used C600 as the non-pathogenic
42
control.
Invasion and intracellular survival were measured in two ways. One was by
percentage of E. coli inoculum invaded or phagocytized by A. castellanii after one hour
of co-culture and accounts for variations in initial E. coli concentration. The second by
the ratio of E. coli to A. castellanii recovered during the invasion assays, which accounts
for variations in recovered amoeba population. Percent E. coli phagocytized by amoeba
was representative of the degree by which the initial inoculation of E. coli came in
contact with and phagocytized by amoeba. The ratio of E. coli to A. castellanii was used
to measure the approximate ratio of bacteria interacting with amoeba during invasion or
internalization of bacteria.
Invasion and intracellular survival assays were performed on PYUCD03,
PYUCD04, PYUCD9, PYUCD10, EDL 932, CTMDR-UPEC and C600, based on each
strains behavior observed during association assays (Figures 13, 14, 15, and 16).
PYUCD04 and PYUCD10 were UPEC deemed as lower associative strains, while
PYUCD03 and PYUCD09 were UPEC deemed as high associative strains (Figures 13,
14, 15 and 16). Unfortunately, in many cases the results were below the level of
detection of our method to enumerate E. coli by plate count methodology (<30 CFU/ml)
and therefore a limited set of results were analyzed for significance.
Invasion assay results for PYUCD03, PYUCD09 and C600 are shown in Figures
17 and 18, and are representative of three independent experiments, within which each
strain was evaluated in triplicate. PYUCD04, PYUCD10, EDL 932 and CTMDR-UPEC
were also tested but bacterial counts were below the limit of detection for our assay (< 30
43
Figure 17. Percent E. coli invaded or phagocytized by A. castellanii recovered
during invasion and intracellular survival assays with Strains C600, PYUCD03,
and PYUCD09. Figure 17 is representative of three co-culture experiments,with
each strain tested in triplicate, on EDL 932, CTMDR-UPEC, C600, PYUCD03,
PYUCD04, PYUCD09 and PYUCD10. Data from EDL 932, CTMDR-UPEC,
PYUCD04 and PYUCD10 are not shown as values obtained were below the level
of detection (< 30 CFU/plate). Error bars represent mean ± 2 SEM. Line at
0.0005 was set as limit of detection for percent bacteria recovered by the invasion
assay. Statistical analysis was performed in a limited manner due to bacteria
enumeration at three hours and 24 hours falling below limits of detection.
Bacterial invasion (time zero) by PYUCD03 showed significant difference
compared to C600 and PYUCD09 (One-way ANOVA, p < 0.05). PYUCD03
three hours post inoculation was significantly different compared to time zero
(Paired t-test, p < 0.05).
44
Figure 18. Ratio E. coli per A. castellanii recovered during invasion and
intracellular survival assays with Strains C600, PYUCD03, and PYUCD09.
Figure 18 is representative of three co-culture experiments, with each strain tested
in triplicate, on EDL 932, CTMDR-UPEC, C600, PYUCD03, PYUCD04,
PYUCD09 and PYUCD10. Data from EDL 932, CTMDR-UPEC, PYUCD04
and PYUCD10 are not shown as values obtained were below the level of
detection (< 30 CFU/plate). Error bars represent mean ± 2 SEM. Line at 0.00048
was set as limit of detection for the ratio of bacteria per amoeba for the invasion
assay. Statistical analysis was performed in a limited manner due to bacteria
enumeration at three hours and 24 hours were below limits of detection. Ratio of
bacteria to amoeba (time zero) of PYUCD03 showed significant difference
compared to C600 and PYUCD09 (One-way ANOVA, p < 0.05). PYUCD03
three hours post inoculation was significantly different compared to time zero
(Paired t-test, p < 0.05).
45
CFU/plate). They were therefore excluded from Figures 17 and 18. Statistical analysis
was run on the limited data set obtained and are presented in Figures 17 and 18. As such,
one-way ANOVA analysis was performed to compare the percent invasion and ratio of
bacteria per amoeba between strains PYUCD03, PYUCD09 and C600. A paired t-test
was performed on PYUCD03 to compare bacterial recovery at zero hours and three
hours. Values at 24 hours could not be compared due to enumerations falling below
values considered statistically significant (<30 CFU/plate).
Invasion results indicate that strains which exhibited high association during our
association assays were the only strains recoverable to any significant degree. In Figures
17 and 18, invasion or uptake of bacteria is represented by values at time zero. This was
the time at which sampling was taken immediately after the final wash and represents the
amount of bacteria found within amoeba. Invasion results for PYUCD03 indicate that
this strain was able to invade or be phagocytozed at significantly higher amounts than
either C600 or PYUCD09 (Figure 11) (p < 0.05). Additionally, the ratio of bacteria per
amoeba recovered with PYUCD03 was significantly higher than with C600 or PYUCD09
at time zero (Figure 18). This is an indication that PYUCD03 was being internalized by
amoeba at higher concentrations compared to C600 and PYUCD09. E. coli strains
PYUCD04, PYUCD10, EDL 932 and CTMDR-UPEC had recovery levels below the
assays’ limit of detection (<30 CFU/plate), suggesting negligible levels of amoeboid
phagocytosis.
Intracellular survival is depicted in Figures 17 and 18 as results obtained at three
and 24 hours post-inoculation and gentamicin treatment. Survival results indicate that all
46
strains fell below the limits of detection by 24 hours of co-culture (<30 CFU/plate).
PYUCD03 showed a significant drop in intracellular bacterium by three hours of coculture, with levels falling below limits of detection by 24 hours (<30 CFU/plate) (Figure
17). PYUCD09 and C600 fell below limits of detection by three hours post-wash and
gentamicin treatment (<30 CFU/plate) (Figure 17). These results indicate that
intracellular survival beyond three hours post-infection was not occurring for any strain
of E. coli at levels by which we could significantly detect.
Taken together the results obtained in the invasion and intracellular survival
assays suggests that much of the bacteria recovered during association testing were E.
coli that were present extracellular to A. castellanii, as recovery of bacteria was much
greater in association assays. Results from the invasion and intracellular survival assays
of the tested E. coli strains indicates internalization of bacteria was low, with little to no
bacteria surviving after 24 hours post invasion. As a whole, the results suggest that each
E. coli strain was being phagocytized instead of invading amoeba, as internalized bacteria
were subsequently digested by amoeba within 24 hours of internalization.
47
DISCUSSION
In this study we wanted to evaluate the overall fitness of UPEC in comparison to
a non-pathogenic standard. To this end we tested ten clinical isolates of UPEC alongside
the non-pathogenic strain of E. coli K-12 known as C600 through several assays. We
first examined the overall bacterial fitness of UPEC by performing co-culture
experiments of E. coli with A. castellanii incubated together for three days. These
experiments examined the capability of bacteria to survive under predation by A.
castellanii. To elucidate differences observed in survival within long-term co-culture
experiments, we also examined the association, invasion, and intracellular survival
capabilities of UPEC to characterize and explain the differences observed in UPEC
fitness.
Bacterial fitness was gauged by observing the percentage of bacteria that was
recovered after three days of co-culture with A. castellanii. Results from these long-term
co-cultures demonstrated that under predation, clinical UPEC isolates appear to have a
slightly enhanced ability to survive predation over the non-pathogenic C600 strain
(Figures 1 and 2). Although heavily preyed upon, as indicated by the low percent survival
of each bacterial population (Figures 1, 2 and 3), significant differences were observed in
seven of the ten UCDMC UPEC isolates and the CTMDR-UPEC in comparison to C600.
These strains exhibited higher percent survival of bacteria than C600 and were therefore
considered as having greater fitness under predation than the C600 E. coli K-12 strain.
The overall decrease in population, a roughly two-log drop in population for all strains
48
(Figures 4, 5 and 6), was consistent with the decrease in population Huws et al. (2008)
observed in bacteria under A. polyphaga predation. As the primary difference between
the tested UPEC strains and C600 are the virulence factors employed in urinary tract
infection (Brzuszkiewicz et al. 2006), further investigation into the genetic profile of the
UPEC clinical isolates with high fitness would aid in uncovering the source of the
difference in long-term fitness between UPEC and C600.
Bacterial co-culture with amoeba also showed that significant levels of amoebal
growth were observed with bacteria versus when amoeba were cultured in AIM alone
(Figures 7, 8 and 9). AIM contains only inorganic chemicals with no real carbon source
for amoeba to utilize for energy and so the results indicate that growth was due to
predation of bacteria. The levels of amoebal growth observed were similar to that seen
with de Moraes & Alfieri (2008) study with A. castellanii and E. coli K-12, with half a
log increase in amoeba at the concentration used. The increased population of A.
castellanii during co-culture is suggestive that none of the tested bacterium had a
cytotoxic effect on the amoeba and that bacteria were utilized as the primary food source
for the amoeba during experimentation. Although E. coli was heavily predated on by the
amoeba in our study, E. coli did persist in co-culture with amoeba over a three day period
(Figures 4, 5 and 6). While the observe bacterial persistence is consistent with prior
studies (Huws et al. 2008 and de Moraes et al. 2008), further examinations of survival to
later time points would be worthwhile to fully characterize UPEC.
The fitness assays in this study revealed that EHEC EDL 932 exhibited a low
survival rate under predation, consistent with work by Wang & Doyle (1998), who
49
observed that predation had a negative impact on EHEC population. The tested EHEC
EDL 932, did not exhibit cytotoxic or intracellular behavior like that seen with Barker et
al (1999), Lainhart et al. (2009) or Chekbab et al. (2013), as no negative impact on A.
castellanii population was observed, nor was EHEC recovered from within A. castellanii.
The discrepancy in EHEC response from other studies may be due to use of different
EHEC serotypes and protozoan species for observing bacterial-protozoa interactions.
Additionally, it was recently observed that Shiga toxin expression may reduce EHEC
survival within A. castellanii (Chekabab et al. 2013). EHEC strain EDL 932 produces
both Stx1 and Stx2, which may have factored into the observed results in this study.
Bacterial association and invasion was also examined in this study as a possible
method UPEC clinical isolates employ to survive predation and persist in the co-culture
environment. However, association with amoeba appears to be rare, as percent E. coli
associated with amoeba after two hours was lower than one percent of the initial
inoculum (Figures 13 and 14). In our study, many of the strains had low percent
association and ratio of E. coli to A. castellanii when compared to C600. This was in
contrast to Alsam et al (2006) and Jung et al (2007) work with invasive E. coli K1, where
they observed a high association rate compared to their K-12 isolate. As a whole, the
UPEC isolates were less likely to associate with A. castellanii than C600 (Figures 13, 14,
15 and 16). The lower ratio of bacteria per amoeba seen in UPEC would suggest these
strains were in contact with A. castellanii less frequently, and with survival taken into
account, were less likely to be ingested by the predator. This is supported by results from
strains that showed lower association than C600 having greater percent survival than
50
C600. Conversely, strains that had bacteria per amoeba ratios close to that of C600 did
have similar percent survival to C600 in the long-term fitness assays.
Although UPEC is known for its broad range of virulence factors that aid in
adhesion to host cells in the urinary tract (Bower et al. 2005), the association results
suggest that the tested clinical isolates of UPEC do not employ adhesion mechanisms in
bacteria-protozoa interactions (Figures 13, 14, 15 and 16). No UPEC strains showed
greater association with amoeba than C600, with either equal or lower association being
observed (Figures 13, 14, 15 and 16). The association results suggest evasion of predation
rather than invasion like that observed with E. coli K1 (Alsam et al. 2006 and Jung et al.
2007). The varying degrees of association the clinical isolates of UPEC exhibited with A.
castellanii may be explained by UPEC genetic diversity (Brzuszkiewicz et al. 2006 &
Vejborg et al. 2011) and can be clarified by characterizing the virulence profile of each
UPEC isolate.
In addition to association, invasion/internalization of bacteria was assessed to
further elucidate bacteria-protozoa interactions. Various E. coli species have been shown
to be ingested by amoeba but remain viable within their predator (Alsam et al. 2006; Jung
et al. 2007, Nelson et al. 2007). The results from this study indicate that UPEC does not
invade nor survive within A. castellannii to any significant degree (Figures 17 and 18).
Of the four clinical isolates tested, the highly associative PYUCD03 was the only strain
capable of generating values of significance. PYUCD03 results showed that while initial
uptake by A. castellanii was present, by 24 hours no significant number of bacteria could
be recovered (Figures 17 and 18). The other clinical isolates of UPEC, CTMDR-UPEC,
51
EDL 932 and C600 examined in the invasion assays, showed low to no invasion at time
zero, with no significant levels of E. coli being recovered at three or 24 hours post
bacterial invasion/amoeba uptake (Figures 17 and 18). Altogether, these results indicate
that UPEC were not invading amoeba but instead were being phagocytized and digested.
Intracellular survival results were consistent with the known virulence factors
profile of UPEC (Wiles et al. 2008), as UPEC are not typically known to possess a
capsule, like the invasive E. coli K1 strain, for use in surviving phagocytosis. Without a
capsule or similar feature, UPEC strains would not be expected to survive within the
digestive vacuoles of A. castellanii. In addition, most studies of UPEC invasion into host
cells involved epithelial cells of the bladder with UPEC as the initiating party (Justice et
al. 2004, Bower et al. 2005). The morphological differences between epithelial cells to
amoeba are likely the reason why invasion was not observed in this study. It should also
be noted that Acanthamoeba are morphologically and functionally similar to
macrophages due to similarities in phagocytic activity and interactions with bacteria
(Siddiqui and Khan 2012, & Yan et al. 2004). Justice et al. (2004) made the observation
that UPEC were readily targeted and cleared by macrophages during in vivo infections,
with evasion in the form of IBCs as their conclusion for UPEC persistence in urinary
tract infections. As such, UPEC invasion behavior may rely on selective invasion by
bacteria into a susceptible host cell and not through survival and modification of the
phagolysosome.
Comparisons of both fitness and association assays results, allows us to place
each isolate into one of four groupings (Table 2). The first grouping was for strains that
52
Strain
% Bacterial
Survival
Bacteria to Amoeba
Association Ratio
% Bacteria
Associated
PYUCD01
>a
<
=b
PYUCD02
>
<
=
PYUCD03
=
<
=
PYUCD04
>
<
<
PYUCD05
>
=
=
PYUCD06
>
=
=
PYUCD07
=
=
=
PYUCD08
>
<
<
PYUCD09
=
=
=
PYUCD10
>
<
<
EHEC EDL 932
=
<
<
CTMDR-UPEC
>
<
<
Table 2. Fitness and Association comparisons of Clinical E. coli isolates to E. coli K-12,
strain C600.
a
> Or < Strains showed statistically significant difference with E. coli K-12, strain C600
and were either greater than or less than values observed with C600.
b
= Strains were consider equal if no statistically significant difference was observed
53
displayed lower association with amoeba and equal fitness compared to C600. The sole
strain tested that met these two parameters was EHEC EDL 932. The next grouping was
of strains that had association and fitness comparable to C600, where association levels
and long-term survival were equivalent. This grouping is represented by UPEC clinical
isolates: PYUCD03, PYUCD07 and PYUCD09 and closely resembled the nonpathogenic C600 in assay performance. The third grouping encompassed strains that had
lower association and higher fitness compared to C600. This pattern was observed in
CTMDR-UPEC and the clinical isolates: PYUCD01, PYUCD02, PYUCD04, PYUCD08
and PYUCD10. The fourth grouping was of strains that had association equal to C600
and exhibited greater fitness. This grouping was made up of PYUCD05 and PYUCD06.
PYUCD05 and PYUCD06 were not used for invasion assays as results, had high variance
in values within and between independent experiments. A majority of the clinically
isolated UPEC fell into the grouping that had greater fitness and were associating with A.
castellanii at lower rates than the non-pathogenic E. coli K-12. As most clinical UPEC
strains exhibit low association but high percent survival under predation, our results
suggest that UPEC typically persists in the environment by evasion of predators. This is
supported by the lack of intracellular survival and presents UPEC as being unable to use
A. castellanii as an environmental reservoir. The remainder of UPEC strains exhibited
results that indicated that they were equivalent to the non-pathogenic E. coli, C600.
54
CONCLUSIONS
It has been theorized that pathogenic strains of bacteria employ virulence factors
in use for environmental survival, the case in point being invasive E. coli K1 with the use
of its capsule to survive ingestion and invade Acanthamoeba (Alsam et al. 2006; Jung et
al. 2007). Environmental reservoirs may also be a key to continued infections of the
human populace by various bacteria, much like with Legionella (Fields et al.1984). This
study on UPEC interaction with A. castellanii has demonstrated that UPEC does not
invade nor associate with A. castellanii in any significant number, thereby indicating that
it does not utilize protozoa as an environmental reservoir. While UPEC did associate
with A. castellanii, levels were lower than the non-pathogenic E. coli K-12, with invasion
results suggesting that few of the associated bacteria were internalized by amoeba. In
terms of fitness, it was observed that some strains of UPEC can survive under long-term
protozoan predation better than non-pathogenic E. coli, with association results
suggesting that these very UPEC were also less likely to be associated with amoeba.
Taken together, the results in this study suggest that UPEC cannot use A. castellanii as a
reservoir for environmental persistence, and thus environmental control of this group of
E. coli would be easier than with a bacterium that employs an amoebal reservoir such as
L. pneumophila (Winiecka-Krusnell and Linder 2001).
The results obtained in this study indicate that virulent strains of bacteria hold a
slight advantage in interactions with predatory protozoa. Although not fully elucidated,
this study suggests that virulence factors in UPEC strains may play a role in the
55
interactions of E. coli with protozoa and thus how E. coli persist in the environment.
What was discovered was that UPEC does not appear to be capable of using amoeboid
protozoa as an environment reservoir but rather is effective in evasion of predators as an
alternative mode of bacteria-protozoa interactions. UPEC interaction with protozoa was
that of a food source and that it persisted extracellularly to amoeba. This conclusion
indicates that water treatment should still be an effective method to deal with UPEC
contamination, as they do not use protozoa as a reservoir, which can protect intracellular
bacteria from disinfectants (Winiecka-Krusnell and Linder 2001).
Future work with UPEC is suggested that would explore long-term interactions
beyond that examined in this study. Investigation into long-term bacteria-protozoa
interactions could look at later time points past three days of co-culture, such as those
presented in experiments by Weekers et al. (1993) and Barker et al. (1999). Extended coculture experiments could determine whether UPEC are continuously preyed on, whether
the two populations eventually reach a steady state, or if UPEC population eventually
rebounds. UPEC could be sampled at 24-hour intervals to enhance the comparison of
their response to predation. As this study was geared towards screening multiple strains
at once to gauge overall response of UPEC to predation, following the population
changes of a few select strains over time, could aid in characterizing the differences in
survival observed in this study. Additionally, testing different initial inoculation doses
could also be done as the dosage may have an impact on interactions. In Wang and
Ahearn (1997), they found that ratio of bacteria to amoeba exceeding 10000:1 in coculture, had an inhibitory effect on amoebal growth. Lower inoculum was used in Alsam
56
et al. (2006) and Jung et al. (2007) studies than that employed in this study, which may
explain the dramatic difference between their results and ours in regards to percent E. coli
K-12, strain C600, associated with amoeba.
Future investigation into the tested strains through genetic profiling may shed
light on the differences observed between clinical isolates. Characterization of what
virulence factors were present in each isolate may determine why certain strains had high
fitness and low association in comparison to C600. The isolates used in this study were
of unknown profile and thus no judgment could be made on any individual strain. With
defined profiles, more accurate conclusions can be made to explain differences in fitness,
association and intracellular survival observed with UPEC and C600. Thus, the presence
of known virulence factors of UPEC such as cytotoxic virulence factors like, colibactin
(pks), α-hemolysin (HlyA) and cytotoxic necrotizing factor 1 (CNF-1), as well as
adhesions like type 1 pili (FimH) and P pili (pap) (Wiles et al 2008), may be of interest.
Further testing with knock-out mutants of particular virulence factors or UPEC strains
with known genetic profiles would be the next logical avenue for future work.
Additionally, it may be of interest to further examine the lack of cytotoxicity and
invasion observed during this study. Other methodologies more sensitive to cytotoxicity
should be employed as enumeration may not fully account for lysis of cells.
Additionally, Nelson et al. (2003) use of bioluminescent E. coli to monitor the
intracellular activities within Tetrahymena could be adapted to examine how UPEC
interact with A. castellanii. Furthermore, it may be possible to monitor the location of
UPEC and determine whether they are within digestive vacuoles or on the cell surface of
57
the amoeba.
This study attempted to determine how UPEC interacts with amoeba in conditions
that favored predation. With long-term population assays, we examined UPECs ability to
persist in co-culture with a predatory amoeba. In association assays, we observed for
UPECs ability to adhere and contact A. castellanii. Through invasion and intracellular
survival assays, we studied UPECs ability to be internalized and survive within
Acanthamoeba. All these experiments together culminate to show that UPEC appears to
have a broad but distinct response to predation. Generally, UPEC exhibited greater
survival under predation than the non-pathogenic E. coli K-12, with less UPEC bacteria
associating with amoeba during co-culture in comparison to E. coli lacking virulence
factors. Cytotoxicity and invasion of A. castellanii were not observed, indicating these
two methods are not likely employed to aid in UPEC fitness under predation. Results of
this study suggest that UPEC evades predation to a greater degree than bacteria lacking
virulence factors. While further work is necessary to fully decipher the differences
observed between strains in this study and to characterize the mechanism by which
UPEC persists under predation, this study does show that UPEC as a human pathogen
survives protozoa predation better than non-virulent E. coli.
58
LITERATURE CITED
Alsam, S., S.R. Jeong, J. Sissons, R. Dudley, K.S. Kim, and N.A. Khan. (2006).
Escherichia coli interactions with Acanthamoeba: a symbiosis with environmental
and clinical implications. Journal of Medical Microbiology. 55:689-694.
Anderson, G.G., C.C. Goller, S. Justice, S.J. Hultgren, and P.C. Seed. (2010).
Polysaccharide Capsule and Sialic Acid-Mediated Regulation Promote BiofilmLike Intracellular Bacterial Communities during Cystitis. Infection and Immunity.
78:963-975.
Appleyard, R.K. (1954). Segregation of new lysogenic types during growth of a doubly
lysogenic strain derived from Escherichia coli K-12. Genetics 39:440-452.
Barker, J. and M.R.W. Brown. (1994). Trojan Horses of the microbial world: protozoa
and the survival of bacterial pathogens in the environment. Microbiology.
140:1253-1259.
Barker, J., T.J. Humphrey, and M.W.R. Brown. (1999). Survival of Escherichia coli
O157 in a soil protozoan: implications for disease. FEMS Microbiology Letters.
173:291-295.
Bower, J.M., D.S. Eto, and M.A. Mulvey (2005). Covert Operations of Uropathogenic
Escherichia coli within the Urinary Tract. Traffic. 6:18-31.
59
Brzuszkiewicz, E., H. Bruggemann, H. Liesegang, M. Emmerth, T. Olschlager, G. Nagy,
K. Albermann, C. Wagner, C. Buchrieser, L. Emody, G. Gottschalk, J. Hacker,
and U. Dobrindt. (2006). How to become a uropathogen: Comparative genomic
analysis of extraintestinal pathogenic Escherichia coli strains. PNAS. 103:1287912884.
Chekabab, S.M., F. Daigle, S.J. Charette, C.M. Dozois, and J. Harel. (2013). Shiga
toxins decrease enterohaemorrhagic Escherichia coli survival within
Acanthamoeba castellanii. FEMS Microbiol Lett. 344:86-93.
Daggett, P.M., T.K. Sawyer, and T.A. Nerad. (1982). Distribution and possible
interrelationships of pathogenic and nonpathogenic Acanthamoeba from aquatic
environments. Microb. Ecol. 8: 371-386.
de Moraes, J. and S.C. Alfieri. (2008). Growth, encystment and survival of
Acanthamoeba castellanii grazing on different bacteria. FEMS Microbiol Ecol.
66:221-229.
Enzinger, R.M. and R.C. Cooper. (1976). Role of Bacteria and Protozoa in the Removal
of Escherichia coli from Estuarine Waters. Applied and Environmental
Microbiology. 31:758-763.
Fields, B.S., E.B. Shotts Jr, J.C. Feeley, G.W. Gorman, and W.T. Martin. (1984).
Proliferation of Legionella pneumophila as an Intracellular Parasite of the Ciliated
Protozoan Tetrahymena pyriformis. Applied and Environmental Microbiology.
47:467-47.
60
Foxman B. (2003). Epidemiology of urinary tract infections: incidence, morbidity, and
economic costs. Dis Mon. 49:53–70.
Fremaux, B., C. Prigent-Combaret, and C. Vernozy-Rozand. (2008). Long-term survival
of Shiga toxin-producing Escherchia coli in cattle effluents and environment: An
updated review. Veterinary Microbiology. 132:1-18.
Huws, S.A., R.J. Morley, M.V. Jones, M.R.W Brown, and A.W. Smith. (2008).
Interactions of some common pathogenic bacteria with Acanthamoeba polyphaga.
FEMS Microbiol Lett. 282:258-265.
Jung, S.Y., A. Matin, K.S. Kim, and N.A. Khan. (2007). The capsule plays an important
role in Escherichia coli K1 interactions with Acanthamoeba. International
Journal for Parasiology. 37:417-423.
Justice, S.S., C. Hung, J.A. Theriot, D.A. Fletcher, G.G. Anderson, M.J. Footer, and S.J.
Hultgren. (2004). Differentiation and developmental pathways of uropathogenic
Escherchia coli in urinary tract pathogenesis. Proceedings of the National
Academy of Sciences of the United States of America. 101:1333-1338.
Khan, N.A. (2006). Acanthamoeba: biology and increasing importance in human health.
FEMS Microbiol Rev. 30:564-595.
Lainhart, W., G. Stolfa, and G.B. Koudelka. (2009). Shiga Toxin as a Bacterial Defense
against a Eukaryotic Predator, Tetrahymena thermophila. Journal of
Bacteriology. 191:5116-5122.
61
Nelson S.M., A.A.A Cooper, E.L Taylor, and V.C. Salisbury. (2003). Use of
bioluminescent Escherchia coli O157:H7 to study intra-protozoan survival of
bacteria within Tetrahymena pyriformis. FEMS Microbiology Letters. 223:95-99.
Nicholson, T.F., K.M. Watts, and D.A. Hunstad. (2009). OmpA of Uropathogenic
Escherichia coli Promotes Postinvasion Pathogenesis of Cystitis. Infection and
Immunity. 77: 5245-5251.
Nougayrede, J.P., S. Homburg, F. Taieb, M. Boury, E. Brzuskiewicz, G. Gottschalk, C.
Buchrieser, J. Hacker, U. Dobrindt, and E. Oswald. (2006). Escherichia coli
Induces DNA Double-Strand Breaks in Eukaryotic Cells. Science. 313:848-851.
Pickup, Z.L., R. Pickup and J.D. Parry. (2007). A comparison of the growth and
starvation responses of Acanthamoeba castellanii and Hartmannella vermiformis
in the presence of suspended and attached Escherichia coli K-12. FEMS
Microbiol Ecol. 59: 556-563.
Reigstad, C.S., S.J. Hultgren, and J.I. Gordon. (2007). Functional Genomic Studies of
Uropathogenic Escherichia coli and Host Urothelial Cells when Intracellular
Bacterial Communities Are Assembled. The Journal of Biological Chemistry.
282: 21259-21267.
Segal G. and H.A. Shuman. (1999). Legionella pneumophila Utilizes the Same Genes To
Multiply within Acanthamoeba castellanii and Human Macrophages. Infect.
Immun. 67:2117-2124.
Siddiqui, R. and N.A. Khan. (2012). Acanthamoeba is an evolutionary ancestor of
macrophages: a myth or reality? Exp Parasitol. 130:95-97.
62
Snelling, W.J., J.E. Moore, J.P. McKenna, D.M. Lecky, and J.S.G. Dooley. (2006).
Bacterial-protozoa interactions; an update on the role these phenomena play
towards human illness. Microbes and Infection. 8:578-587.
Steinberg, K.M. and B.R. Levin. (2007). Grazing protozoa and the evolution of the
Escherichia coli O157:H7 Shiga toxin-encoding prophage. Proceedings of the
Royal Society B. 274:1921-1929.
Vejborg, R.M., V. Hancock, M.A. Schembri, and P. Klemm. (2011). Comparative
Genomics of Escherichia coli Strains Casuing Urinary Tract Infections. Appl.
Environ. Microbiol. 77:3268-3278.
Wang, G. and M.P. Doyle. (1998). Survival of enterohaemorrhagic Escherichia coli
O157:H7 in water. J. Food Protect. 61:662–667.
Wang, X. and D.G. Ahearn (1997). Effect of Bacteria on Survival and Growth of
Acanthamoeba castellanii. Current Microbiology. 34:212-215.
Watson, P.J., K. Ohtaguchi, and A.G. Fredrickson. (1981). Kinetics of Growth of Ciliate
Tetrahymena pyriformis on Escherichia coli. Journal of General Microbiology.
122:323-333.
Weekers, P.H.H., P.L.E. Bodelier, J.P.H. Wijen, and G.D. Vogels. (1993). Effects of
Grazing by the Free-Living Soil Amoebae Acanthamoeba castellanii,
Acanthamoeba polyphaga, and Hartmannella vermiformis on Various Bacteria.
Applied and Environmental Microbiology. 59:2317-2319.
63
Wells, J.G., B.R. Davis, I.K.Wachsmuth, L.W. Riley, R.S. Remis, R. Sokolow, and G.K.
Morris. (1983). Laboratory investigation of hemorrhagic colitis outbreaks
associated with a rare Escherichia coli serotype. J. Clin. Microbiol. 18: 512-520.
Wiles, T.J., R.R. Kulesus, and M.A. Mulvey. (2008). Origins and Virulence Mechanisms
of Uropathogenic Escherichia coli. Exp Mol Pathol. 85:11-19.
Winiecka-Krusnell, J. and E. Linder. (2001). Bacterial infections of free-living amoebae.
Res. Microbiol. 152:613-619.
Yan, L., R.L. Cerny, and J.D. Cirillo. (2004). Evidence that hsp90 is involved in the
altered interactions of Acanthamoeba castellanii variants with bacteria. Eukaryot
Cell. 3:567-578.