UNIVERSITY OF CINCINNATI
Date:______
I, ______, hereby submit this work as part of the requirements for the degree of: in:
It is entitled:
This work and its defense approved by:
Chair: ______
Antibiotic Therapy in the Treatment of E. coli O157:H7.
A dissertation submitted to the
Division of Graduate Studies and Research
of the University of Cincinnati
In partial fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY (Ph.D.)
In the Department of Molecular Genetics, Biochemistry, & Microbiology
of the College of Medicine
2008
Colleen M. McGannon
B.S., Ashland University, 2002
Committee Chair: Alison A. Weiss, Ph.D.
Abstract
Escherichia coli O157:H7 causes an estimated 73,000 of food borne illness annually. Varying levels of disease severity exist and include diarrhea, bloody diarrhea, and hemolytic uremic syndrome which can result in kidney damage or death. E. coli
O157:H7 produces Shiga toxins 1 and/or 2, although it is Shiga toxin 2 that is linked to severe disease. Currently, a Shiga toxin-producing isolate will yield a positive diagnostic result, regardless of which Shiga toxin variant is produced. We have developed an
ELISA that can differentiate between Shiga toxin 1 and Shiga toxin 2 in the presence of fecal material, and if used in the clinical setting, can increase the accuracy of prognosis for a patient infected with E. coli O157:H7. Epidemiological studies have suggested that antibiotics may be linked to HUS development. As such, we have completed an extensive study to determine which, if any, antibiotics are safe for the treatment of E. coli
O157:H7. The data suggest that antibiotic-mediated Shiga toxin induction or reduction is based on the mechanism. Antibiotics that target DNA increase Shiga toxin production while protein synthesis inhibitors decrease Shiga toxin production. Human intestinal E. coli have been shown to amplify Shiga toxin. Our data suggest that protein synthesis inhibitors may be effective, even if the patient harbors a Shiga toxin-amplifying isolate, provided that the commensal isolate is susceptible to the antibiotic. Conversely, antibiotics would likely be of no benefit if the Shiga toxin-amplifier is resistant to the antibiotic. Future studies should focus on characterizing the flora of HUS patients, and examining antibiotic treatments within the context of these strains.
iii
iv
Follow your passion.
~ Alison A. Weiss
v Acknowledgements
Alison said, in a heart-warming fashion,
Colleen, I encourage you to follow your passion.
Kind in allowing me this latitude, there’s
Not enough thanks and/or gratitude.
Of course my committee supported me too,
Willingly sharing their thoughts based on their point of view.
Lyndsay, Mark, Shantini, Mike, Cindy, and Scott
Eternally treasured, never forgot.
Dad and Mom were supportive, and helped keep me sane.
Goodness, I know it, bacteria must have a brain!
Exuding such patience as the time it did crawl.
Mr. O’Malley often took me to the Serpentine Wall.
Emily Bradford, my friend, the dearest, the best. We’re
Not really that different, who would have guessed?
To RIVER and Skyline and St Monica St George.
Sad to leave Cincinnati, but now to Albuquerque I’ll forge.
vi Table of Contents
Abstract……………………………………………………………………..………...iii
Acknowledgements……………………………………………………………………v
Table of Contents………………………………………………………………….....vi
List of Figures and Tables……………………………………………………..…….ix
List of Abbreviations…………………………………………………………..……..xi
Chapter I. E. coli O157:H7 literature review………………………………..……..1
Introduction 2 E. coli O157:H7 is Associated with Disease 3 Food borne Transmission 4 Other Routes of Transmission 5 National Monitoring Systems 9 Inspection Protocols 11 Disease 13
Shiga toxins 15 Discovery 15 Toxin Nomenclature 16 Genetics and Expression 16 Structure 18
Toxin Receptor (Gb3) 20 Toxin-Gb3 Interaction 23 Toxin-Gb3 Internalization 27 Shiga toxin 2 Glycosidase Activity 28
Antibiotic Treatment of E. coli O157:H7 29 Ciprofloxacin 30 Trimethoprim/Sulfamethoxazole 34 Ampicillin 36 Ceftriaxone 37 Fosfomicin 38 Gentamicin 39 Doxycyline 39 Azithromycin 40 Rifampicin 41
vii Role of Normal Intestinal Flora in Disease 42 Lipopolysaccharide Fecal-Isolate 29 45
Laboratory Assays 48 Sorbitol MacConkey Agar 48 Latex Agglutination Assay 48 Polymerase Chain Assay 49 Western blot Assay 50
Rationale for Studies Performed 51
Chapter II. Glycoconjugates that specifically bind Stx1 or Stx2………………....53
Abstract 54
Introduction 55
Materials and Methods 59
Results 62
Discussion 67
Chapter III. Antibiotic treatment to prevent the systemic complications of
Escherichia coli O157:H7 Infection………………………………………………...71
Abstract 72
Author Summary 74
Introduction 75
Materials and Methods 79
Results 85
Discussion 102
viii
Chapter IV. Discussion………………………………………………………. 106
Summary 107
Future Directions 108
Concluding Remarks 113
References……………………………………………………………………… 114
ix Figures and Tables
Chapter I.
Figure 1. The structure of Stx. (Page 19)
Figure 2. The structure of Gb3. (Page 21)
Figure 3. Sequence alignment of the Stx1 and Stx2 B subunits. (Page 26)
Figure 4. The structure of lipopolysaccharide (LPS). (Page 47)
Chapter II.
Figure 1. Structure of Gb3 and the O117 LPS. (Page 57)
Figure 2. Representation of the tailored biotinylated glyconjugate and the structures of the five molecules. (Page 58)
Figure 3. Antibodies specific to O107 and O117 block Stx2 binding to LPS. (Page 64)
Figure 4. Differential binding of Stx1 or Stx2 to synthetic glycoconjugates. (Page 66)
Chapter III.
Table 1. Plasmids and strains used in this study. (Page 84)
Figure 1. The Luc2p-MigR1 retroviral vector. (Page 87)
Figure 2. Quantifying Shiga toxin production. (Page 88)
Table 2. The Minimum Inhibitory Concentrations (MIC) of antibiotics. (Page 90)
Figure 3. The effect of ciprofloxacin and trimethoprim/sulfamethoxazole on the production of Stx. (Page 92)
Figure 4. Antibiotics that induce production of Stx. (Page 93)
Figure 5. Determining if antibiotics induce Stx1, Stx2, or both in E. coli O157:H7 strain PT-32 (stx1, stx2). (Page 95)
Figure 6. Determining the effect of antibiotics on the production of Stx. (Page 96)
Figure 7. The efficacy of antibiotics in the presence of a susceptible amplifer strain. (Page 99)
x Figure 8. The efficacy of antibiotics in the presence of a resistant amplifier strain. (Page 101)
xi Abbreviations
CAT scan computer axial tomography scan CIP ciprofloxacin CDC Centers for Disease Control E. coli Escherichia coli ED50 effective dose 50% ELISA enzyme-linked immunosorbent assay EHEC enterohemorhagic Escherichia coli EPEC enteropathogenic Escherichia coli FDA Food and Drug Administration FI fecal isolate Gb3 globotriaosylceramide GFP green fluorescent protein HUS hemolytic uremic syndrome HuSAP human serum amyloid P IL interleukin kDa kilodalton KDEL lysine-glutamic acid-aspartic acid-lysine KDO 3-deoxy-D-manno octulosonic acid LB Luria Bertani LEE Locus of Enterocyte Effacement LPS lipopolysaccharide MH Mueller-Hinton MIC minimum inhibitory concentration PEST proline-glutamic acid-serine-threonine PBS phosphate buffered saline RPLA reversed passive latex agglutination RT-PCR reverse transcriptase-polymerase chain reaction SDS-PAGE sodium dodecyl sulfate poly acrylamide gel electrophoresis SAP serum amyloid P SOS save-our-ship Stx Shiga toxin Tir translocated intimin receptor TMP-SMX trimethoprim-sulfamethoxazole USDA United States Department of Agriculture
xii Chapter I.
Literature Review
1 INTRODUCTION
Escherichia coli is a Gram-negative bacterium present within the human
microbiota. E. coli is a component of animal gut flora as well. For example, cattle are
natural carriers of E. coli serotype O157:H7 [1-3]. While adult cattle harbor this strain
asymptomatically, [4], E. coli O157:H7 is responsible for a likely underestimated 73,000
cases of human disease per year, and is the primary cause for the development of
hemorrhagic colitis and hemolytic uremic syndrome (HUS) in children [5].
First associated with disease in 1982 [6] [7], E. coli O157:H7 is classified as a
newly emerging food borne pathogen, with ground beef being the primary source for
human infection. After ingestion of an infectious dose as low as 50 organisms, the
bacteria colonize the large intestine, causing a disruption of the microvilli. Subsequently,
the bacteria produce its major virulence factors called Shiga toxins, which inhibit protein
synthesis within target cells. Unfortunately, antibiotic treatment has been linked to the
development of HUS in children infected with E. coli O157:H7 [8]. Therefore, the current treatments are limited to supportive therapies such as rehydration.
This review will first provide background information regarding E. coli O157:H7
and its Shiga toxins, then introduce topics related to the two major courses of
experimental study reported in this document. Those topics include the development of
an assay that can distinguish between the two types of Shiga toxin, and in addition
determining which, if any, antibiotics are safe for use in the treatment of E. coli
O157:H7.
2 E. COLI O157:H7 IS ASSOCIATED WITH DISEASE
E. coli O157:H7 was initially isolated in 1975 from a sporadic case of
hemorrhagic colitis. This strain was not seen again until 1983 when it was first linked to
disease as a food borne pathogen [7]. Two groups of patients, one in Oregon and one in
Michigan, displayed what are now known to be the classic symptoms of E. coli O157:H7-
related disease; abdominal cramping and grossly bloody diarrhea. This strain was
isolated from patient stool samples in both locations. In addition, E. coli O157:H7 was
isolated from a hamburger patty taken from a fast-food chain that all the infected persons
had eaten at prior to becoming ill, suggesting that consumption of contaminated ground
beef was the source of the infection.
A second paper published in that same year described the discovery of the Shiga
toxins produced by E. coli O157:H7. Supernatants from E. coli O157:H7 isolated from the stool samples of patients with sporadic cases of HUS [6] were shown to contain a
compound that was toxic to the Vero kidney cell line. In addition, patient serum was able
to neutralize the toxicity on Vero cells, suggesting that the toxin had produced an immune response within these patients. These data suggested that the toxin produced by
E. coli O157:H7 was correlated to disease development.
To determine if E. coli O157:H7 was emerging for the first time or if it had just
gone unnoticed until that point, the CDC analyzed the 3,000 strains of E. coli that had
been collected and serotyped in the previous ten years. Other than the 1975 isolate, no
other E. coli O157:H7 strains had been previously identified [7]. Similar studies were
carried out in the United Kingdom and Canada and out of 17,000 strains analyzed, only
3 seven were identified as E. coli O157:H7 isolates [9-11]. Thus, E. coli O157:H7 is truly
an emerging pathogen.
FOOD BORNE TRANSMISSION
E. coli O157:H7 is one of the top four pathogens associated with food borne
disease within the United States [12]. Between 1982 and 2002, 49 states have reported a total of 350 outbreaks of E. coli O157:H7. Of these, 52% were due to the consumption of contaminated food [13]. Consumption of undercooked, contaminated ground beef was
responsible for 41% of the food borne-related cases [13].
However, according to a recent report by the United States Department of
Agriculture (USDA), cases originating from the consumption of contaminated ground
beef have been reduced to 34% [12]. This is perhaps in part due to an increasing
knowledge that thorough cooking (160ºF) of ground beef will kill any E. coli O157:H7
present. While initially associated with fast food chains, contaminated ground beef can
also be found in grocery stores. For example, the Kroger Company has recently recalled
ground beef thought to contain E. coli O157:H7 that was distributed in the Michigan and
central/northwestern regions of Ohio [14].
During the slaughter and processing of cattle, portions of the carcass intended for
consumption are in close proximity to those that harbor pathogenic bacteria. Therefore, it
is easy to understand how E. coli O157:H7 from the cattle gut might easily incorporated
into the ground beef food supply during processing [15]. Various measures have been
taken to decrease the risk of contamination during this process and will be discussed in an
upcoming section.
4 Produce was responsible for a surprising 21% of food borne outbreaks in 1982-
2002 [13]. Lettuce was responsible for 34% of these produce-linked outbreaks, followed
by apple cider/juice (18%), salad (16%), coleslaw (11%), melons (11%), sprouts (8%),
and grapes (3%) [13]. In the majority of these instances, the mishandling of food within
the kitchen resulted in cross-contamination with meat products [13]. Spinach, not
mentioned in these outbreaks, is now the leafy green most commonly associated with E.
coli O157:H7 due to a recent outbreak [16]. The details of this outbreak will be
discussed within the context of the National Monitoring Systems in an upcoming section.
OTHER ROUTES OF TRANSMISSION
Other routes of infection include, but are not limited to, contact with infected
animals and person-to-person spread. It has been well established that cattle harbor E.
coli O157:H7 [1-3] and can shed it in their feces [17-19]. Therefore, there is an increased
risk of becoming infected with E. coli O157:H7 in environments such as agricultural zoos
and petting zoos where close interactions with cattle and their feces are prevalent.
For instance, three separate outbreaks linked to petting zoos occurred in North
Carolina, Florida, and Arizona between 2004 and 2005. Of 173 confirmed cases, 15 developed HUS. Pulse field gel electrophoresis (PFGE) identified the strains present in
both the animal and patient samples as being identical in each respective outbreak. As
the presence of actual manure on the hands of the ill children was determined to be a risk
factor, The Centers for Disease Control suggested that restricting small children from
5 entering open-interaction areas in which feces are abundant may help to reduce instances
of transmission of E. coli O157:H7 to people within this environment [20].
Two outbreaks that occurred in 2000 and 2001 at a farm-based day camp run by a university to promote the interest in agricultural animals through hands-on activities further supports the danger of facilitating the interaction of young children with cattle.
At this camp, children ages 5-10, the exact age group classically considered at-risk for the development of HUS after E. coli O157:H7 infection, were responsible for the 1 on 1 care of a calf, even if the animal was ill. Campers bottle fed the calves, groomed them, and cleaned the pens. Throughout the first summer, 84 children became infected with bacteria commonly associated with cattle, including E. coli O157:H7. Risk factors associated with disease included the care of an ill calf, and (as seen with the petting zoos), the presence of visible manure on the hands of the child.
In response to the illnesses, the camp removed ill animals from the site. To
supplement the alcohol gels currently being used, hand-washing stations were installed
within the barns. Counselors were also provided with information regarding farm-related
diseases. The instances of disease then decreased.
The following summer, these preventative mechanisms remained in place, but
were sadly not enough to prevent the new campers from becoming ill. The camp then
closed to reassess the protective protocols being utilized. Fewer numbers of calves were
used at the camp. Only camp supervisors were able to enter the pen and feed the animals.
Campers were required to wear short sleeves to minimize the contact of clothes with
contaminated sources. Individual hand-washing stations were upgraded to accommodate
one counselor and the eight children under his supervision. A specialist trained the staff
6 on proper hand-washing techniques, and campers watched a video on these practices as
well. After reopening, no further illnesses were seen [21].
These events illustrate the importance of not only limiting the contact between
children and agricultural animals, but that the simple act of hand-washing is extremely
effective in the prevention of disease transmission. In addition, these illnesses occurred
in 2000-2001, a time in which the basic knowledge of E. coli O157:H7 was well
established. This indicates that perhaps the general public is not as aware of agricultural-
based pathogens as one would hope.
The danger posed by the agricultural setting is not limited to children petting
animals. At the Medina, Ohio County Fair in 1999, persons who had consumed
beverages made with water supplied from an on-site distribution system from vendors
near the animal display arena had a significantly higher rate of E. coli O157:H7 infection
that those who had not. Although the water system tested negative for the presence of E.
coli O157:H7, a second group became infected during a Halloween event held at the fairgrounds. Extensive testing of the water system indicated faulty pressure, the presence of backflow, and the exposure to contaminated environments allowed E. coli O157:H7 to enter the system.
An agricultural fair-related outbreak that deviates from the norm is also of interest. At the Lorain County fair In 2001, E. coli O157:NM (nonmotile) was found to be present in the sawdust within a building used to temporarily house animals during shows or practices. Infected patients were found to have consumed food or drink within the building, attended a dance there, or handled the sawdust directly [22].
7 Person-to-person contact was responsible for 14% of the outbreaks analyzed by
Rangel [13]. These contacts primarily occurred in the day care setting. In Minnesota, interviews of parents with infected children were conducted between 1988 and 1989.
Throughout nine different institutions, 68 cases of E. coli O157:H7 infection were reported. Person-to-person transmission was implicated as the removal of the child from the day care centers prevented the spread to other children [23]. The sources of the outbreaks were not identified.
Day care center outbreaks peak between June and August, a time when the
interaction with farm animals is high [13]. This can put children that live in rural settings
at a higher risk of infection, as seen in a day care in, again, Minnesota. In this instance,
E. coli O157:H7 was isolated from a boy who had developed HUS. Within ten days,
disease was confirmed in 17 additional children of the 64 attendees. 5 days later, the day
care closed. It was hypothesized that the initial infection likely occurred during one of
two farm visits taken by the boy prior to the outbreak, and that the bacteria was
transmitted from child to child during unsupervised bathroom visits [24].
Only one laboratory outbreak has been reported. This occurred in 2002 when two
technicians were ironically infected while testing an E. coli O157:H7 sterilization
technique [13].
In summary, E. coli O157:H7 can spread through a variety of different means,
most of which can be traced back to cattle. In order to prevent E. coli O157:H7-related
disease, the public must be aware of the contacts made near the presence of cattle, ground
beef or otherwise, and be willing to partake in simple preventative measures such as
hand-washing.
8
NATIONAL MONITORING SYSTEMS
FoodNet, an organization designed to monitor food borne illness within the
United States, was set up in 1992 after an outbreak of E. coli O157:H7. This outbreak
was the impetus needed to make gathering data to fully determine the prevalence of food
borne pathogens within the country a priority within the national budget [25].
FoodNet has determined that a series of steps need to be completed in order for an infection to be reported. First, the general public must recognize the symptoms of food
borne illness and seek medical attention. Secondly, the proper samples must be collected and subsequently tested/verified for the presence of a given pathogen. Lastly, the information must be reported to the CDC. To both monitor the success of this process and eliminate the state-specific variations on reporting requirements, the organization routinely conducts surveys of the general public, physicians, and clinical laboratories regarding knowledge, medical practices, and diagnostic procedures. Specific to E. coli
O157:H7, FoodNet has included pediatric nephrologists within the network to specifically monitor the prevalence of HUS, a strong indicator of E. coli O157:H7
infection [25].
Through this process, it was discovered that between 1995 and 2000, the number of clinical laboratories routinely testing for E. coli O157:H7 had increased from 59% to
68%, indicating that FoodNet is able to successfully monitor the E. coli O157:H7-related
laboratory activities [26].
9 The structure of FoodNet has also enabled case-controlled studies to be
conducted that would have otherwise been impossible. It was this organization that
determined that contact with cattle and the consumption of pink hamburgers, knowledge
that is now seemingly commonplace, were significant risk factors in contracting E. coli
O157:H7 infections during the 1996 outbreak [27].
PulseNet was established in 1996, again, as a result of an E. coli O157:H7
outbreak. At this time, the CDC began using pulse field gel electrophoresis (PFGE) to
determine if the E. coli O157:H7 strains isolated from patients were identical. Because the request for subtyping from different states was so overwhelming, the CDC decided to
coordinate a network of state laboratories that utilized standardized PFGE techniques to
conduct their own subtyping [28]. FoodNet laboratories recognized the value of this new
organization, and encouraged states to participate in PulseNet. This resulted in the
collaboration of both epidemiologists and microbiologists, a valuable pairing regarding
the monitoring of food borne pathogens.
The profiles collected from each subtyping reaction are entered into an electronic
database. This system is useful as related outbreaks that would have otherwise been
classified as separate based on geographic location can now be quickly linked. This is
especially critical now that the distribution of food has evolved to include not only local
areas, but national areas as well [29].
The effectiveness of PulseNet to quickly identify and link food borne illnesses
nationwide was illustrated in the 2006 outbreak in which prewashed, bagged spinach was
implicated as the contaminated food source. Only four days after the identification of a cluster of E. coli O157:H7 cases, PulseNet confirmed that a multistate outbreak of a
10 single strain of E. coli O157:H7 was occurring. Patients had identified eating fresh spinach, which was then quickly removed from grocery store shelves. Even though cases were reported in states from the east to the west coast, the outbreak was stopped within one month [16]. Although the process in which the spinach had first become contaminated was not decisively determined, PulseNet confirmed that both cattle and feral swine in a pasture near the spinach field harbored the same E. coli O157:H7 strain as those infected persons [30].
INSPECTION PROTOCOLS
As mentioned earlier, 34% of food borne cases of E. coli O157:H7 infection originate from the consumption of contaminated ground beef [12]. While the aforementioned monitoring systems are successful in dealing with ill patients, new protocols to be used by slaughterhouse inspectors were established by the Food Safety and Inspection Service in spring of 2008 in order to reduce the rate of initial contamination of ground beef [12].
The protocols currently implemented by slaughterhouse inspectors were created
prior to the knowledge that pathogens are related to food borne illness, and rely on the
visible illness of the animal. The new set of protocols is designed to standardize the inspection process to focus on vulnerable points within the slaughter process in which inspectors electronically answer yes or no to standardized questions. If a violation is found, the inspector will be provided with a standardized set of instructions to follow to
11 specifically address the correction of the violation This automated process will both standardize and allow for the electronic documentation of each inspection [12].
It was generally believed that E. coli O157:H7 can easily be incorporated into ground beef if the intestines are accidentally nicked during the butchering process. A thorough study of the slaughter/processing procedure indicates that it is not that simple.
This procedure is divided into categories including transport/receiving/holding, stunning/bleeding, head skinning and removal, esophagus and hoof removal, skinning, evisceration, carcass splitting, chilling, head and cheek meat processing, labeling, and storage/shipping. The rate of contamination within a plant can be directly correlated to the physical layout, speed of the kill, and skill level of the employees [12].
While contamination can occur throughout the process, select steps were identified as being most vulnerable. The main source of contamination was actually identified to be the hide removal. E. coli O157:H7 adheres to the gloves, knives, and clothing of the workers as they remove the dust and fecal material from the hide prior to skinning. Therefore, workers must be made aware of the consequences of mishandling the carcass at this point to prevent cross-contamination.
During evisceration, the both the esophagus and the anus must be properly tied shut to prevent the contents from leaking out. Before the removal of the gastrointestinal tract, the ventral side of the carcass must be precisely cut as to prevent slicing open the intestinal tract. Finally, the carcass is hung upside down during neck and cheek processing, resulting in a concentration of microbes within this area [12]. Therefore, employees must handle the neck and cheeks areas with care in order to prevent cross- contamination.
12 The final ground beef products are also tested. An electronic database called the
System for Tracking E. coli O157:H7 Positive Suppliers has been instituted to track those plants that have produced contaminated ground beef, and expedite the identification of
those repeat offenders [12].
The Office of the Inspector General has analyzed the meat packing industry in
great detail. By identifying the steps in the butchering process that are vulnerable to
bacterial contamination, effective prevention measures can be put into place. The
electronic inspection procedures will standardize both the protocols that look for and
respond to violations. This plan was released in Spring 2008 and a report on the
implementation has yet to be released.
DISEASE
If, in spite of the preventative mechanisms just described, a person were to ingest
E. coli O157:H7, disease would likely develop. Disease encompasses varying levels of
severity. Initially, patients exhibit abdominal cramps and watery diarrhea. Of these
patients, 70% progress to bloody diarrhea, a disease state that is termed hemorrhagic
colitis [31,32]. In severe cases, patients pass actual blood clots [33].
During hemorrhagic colitis, damage to the colon is evident. The colon wall
thickens due to edema (increased fluid in the tissue) [33]. Colonoscopic exploration
(video footage) revealed that mucosal hyperemia (bloody mucous), ulcerations, and
lengthy ulcer-like lesions are also present [33]. Most patients resolve hemorrhagic colitis
within 14 days.
13 Ten percent of hemorrhagic colitis patients go on to develop the renal
complication hemolytic uremic syndrome (HUS) which develops approximately one
week after the initial onset of diarrhea. E. coli O157:H7-related HUS occurs primarily in
children under the age of 10 and the elderly [34]. Classic symptoms of HUS include
thrombocytopenia (a reduction in platelet cells in the blood) [35] and hemolysis [36].
Proper functioning of the kidney relies on the adequate blood flow through the
organ, a process that is blocked in HUS patients due to the accumulation of fibrin within the microvasculature [37]. This is evidenced by the presence of high levels of the
fibrinogen degradation product D-dimer in children with E. coli O157:H7-related HUS as
compared to those children who did not develop HUS [38] and the presence of fibrin
within the microvasculature of HUS patients [37]. It is not uncommon for a HUS
survivor to exhibit renal damage or failure anywhere from 16-24 years later [39], further
exhibiting the lasting devastation that can result from this disease.
Recent advances have been made in establishing a small animal model that could
be used for the study of E. coli O157:H7, a tool that has remained elusive as murine
models do not adequately mimic human disease. Oral inoculation of germ-free Swiss
Webster mice with pathogenic E. coli O157 strains induced cecal fluid accumulation,
kidney lesions, the accumulation of fibrin, and death [40], indicating that this model may
have promise within the field.
14
SHIGA TOXINS
After E. coli O157:H7 colonizes the human host gut, the bacteria can produce and release different toxins which are described next.
DISCOVERY
Shigella dysenteriae is a bacterium that is responsible for 30-50% of cases of dysentery and 70% of diarrheal fatalities [41]. Dr. Shiga, whose mentor was trained under Koch, was the first to isolate this bacterium from a patient and fulfill Koch’s postulates [42]. After showing that this organism agglutinated in the presence of patient serum, Dr. Shiga went as far as to test on himself a heat-killed, whole cell vaccine that was later administered to thousands even though an effective vaccine remains to be created [42]. It was shown that one of the major virulence factors produced by S. dysenteriae is a toxin, now termed Shiga toxin [43].
Like S. dysenteriae, E. coli also produces Shiga toxin. In 1977, culture filtrates from E. coli strains isolated from infants with diarrhea were shown to produce a toxin that exerted a permanent cytotoxic effect on Vero cells (African green monkey kidney cells) as reflected in the unique morphology of the cells [44]. Shortly thereafter, it was shown that the cytotoxic effect of this Verotoxin could be neutralized by antiserum to
Shiga toxin isolated from S. dysenteriae, demonstrating the close similarity between the toxin produced by the two bacterial pathogens [45].
15
TOXIN NOMENCLATURE
E. coli O157:H7 strains are now known to produce two different Verotoxins
[5,46]. The nomenclature of these toxins in the literature varies based on historical
preference. Verotoxin 1 and Verotoxin 2 resulted from Konowalchuk’s work, while
Shiga-like toxin 1 and Shiga-like toxin 2 resulted from O’Brien’s studies. Yet a third pairing of names includes Shiga toxin 1 (Stx1) and Shiga toxin 2 (Stx2), and these will be used in this document. It is important to note that while E. coli O157:H7 strains can contain Stx1, Stx2, or both [5,46,47], those strains expressing Stx2 are most commonly isolated from severely ill patients [48].
GENETICS AND EXPRESSION
Nucleotide sequence comparison of Stx from S. dysenteriae and Stx1 indicate
they are over 99% homologous at the amino acid level, differing by only one amino acid
[49], while Stx1 and Stx2 are only 60% homologous [50].
The Stx proteins are comprised of two subunits, termed A and B. The gene stxA
encodes for the A subunit and is 959 nucleotides in length [51]. The stxB gene encodes for the B subunit and is 269 nucleotides in length [51]. These genes are separated by 12 noncoding nucleotides [49,51].
The genes for Stx1 and Stx2 are present on lambdoid phages which have
integrated into the E. coli O157:H7 chromosome. Two examples of Shiga toxin-
expressing phages that are used extensively in my studies are H19B (Stx1) [52] and
933W (Stx2) [51]. Sequence comparison of H19B and 933W indicate that Stx1 and Stx2
16 are present in the same location within these individual phages, and are thus regulated by
the same transcriptional control system described next [51].
Infection of a bacterium with the Stx-encoding phage can be classified as either
being lysogenic or lytic. Phage-infected bacteria can immediately progress to the lytic
phase, and this event is more common than progression to lysogeny. In the lytic
infection, the phage will undergo replication in the host bacterium leading to the transcription and translation of genes that lead to the production of new phage and lysis of the bacterial host. As H19B and 933W phage encode for Stx1 and Stx2, these toxins will also be produced and released during this process.
Alternatively, during lysogeny, the phage becomes integrated into the host
chromosome and has little affect on the bacterial cell. The CII protein initiates the state
of lysogeny by allowing for the transcription of the CI repressor [53]. The CI repressor
binds to an operator sequence and prevents transcription of all phage genes except CII
and itself [53]. During lysogeny, neither phage nor toxin is produced until the host
bacteria undergoes stress such as exposure to antibiotics [54], or UV light [55]. These
stressors initiate the lytic phase, allowing the phage to find a new host cell.
As just stated, the exposure of lysogenic bacteria to antibiotics can induce the
lytic cycle [54,56] by activating the bacterial protein RecA [53]. RecA cleaves the CI
repressor, leaving the operator sequence accessible to RNA polymerase and transcription
of the phage genes such as cro occurs [53]. Cro allows for the expression of the
antiterminator protein, Q, subsequently followed by the read-through of the terminator
that lies between the late promoter and the phage structural genes, toxin genes, and lysis
genes, mediating their expression. Thus, in the presence of stressors such as antibiotics,
17 phage and toxin are produced, the cell lyses, and the phage and toxin are released from
the bacteria.
STRUCTURE
Shiga toxin is an AB5 toxin. AB5 toxins have two structural components, termed
the A and B subunits. The A subunit rests on a pentameric complex of B subunit (Figure
1) [57].
The A subunit of Shiga toxin contains two cysteines that are joined by a disulfide
bond [58]. Dissolution of the disulfide bond occurs after furin cleaves the A subunit at the Arg-Val-Ala-Arg trypsin recognition sequence, resulting in the generation of two individual 28-kDa A1 and 4-kDa A2 fragments [58,59]. This cleavage is necessary for
toxicity as demonstrated in a study in which cleaved Stx was 50 times more cytotoxic to
cells lacking the ability to produce furin than uncleaved Stx [59].
Differences in the primary amino acid sequence between Stx1 and Stx2 affect
antigenicity. Stx1 is 99% identical and immunologically cross-reactive with the toxin
produced by Shigella dysenteriae [49,60]. However, Stx1 antibodies cannot neutralize
Stx2 and vice versa, even though they are 60% homologous [50,61,62].
18 A
B
Figure 1. The structure of Stx. The A subunit of Stx rests on a pentamer of B subunits.
19
TOXIN RECEPTOR (Gb3)
Stx was first shown to bind to a mammalian receptor in 1986 when 125I-labelled
toxin was found to bind to rabbit ileum microvilli [63]. Toxin binding to the ileum is
significant as disease in humans begins in the intestine. A second study later that year
showed that B subunit of Stx alone could bind to the receptor but the A subunit could not
[64]. This work is consistent with subsequent findings that the B subunit of Stx1 and
Stx2 mediates binding to host cells [65-67].
Globotriaosylceramide [Gb3; Galα(1-4)-Galβ(1-4)glucosyl ceramide] (Gb3) was
shown to be the likely receptor for both Shiga toxins 1 and 2 [64,68-70]. Gb3, present in
lipid rafts, is a membrane component of several tissue types throughout the body [71,72],
and most notably, the human kidney [73].
Gb3 contains a polar carbohydrate group and a nonpolar ceramide component that
contains a fatty acid chain. The terminal Galα1-4Gal carbohydrates of Gb3 serve as the
binding site for Stx1 as replacement of the terminal galactose on Gb3 with N-
acetylgalactosamine results in the abrogation of Stx1 binding [69]. In addition, digalactosyl diglyceride, which contains the same terminal galactose moiety as Gb3, did
not bind to the toxin [69]. This indicates that Stx1 binds to those galactose moieties only
within the context of Gb3. It was then shown that Stx2 binds to Gb3, which was prepared from a human kidney.
20 Figure 2. Structure of Gb3.
21
[70]. Stx2 also requires the Galα(1-4)-Galβ linkage as when the sugar linkage was
removed or modified, Stx2 did not bind Gb3 [70].
In the studies that will be described in Chapter 2, we have also seen that Stx1
binds to a synthetic Gb3, yet not to an N-acetylated form of Gb3 or to a compound
containing only the terminal Gb3 dissacharides. The reverse is true for Stx2 in that it
bound to those compounds that Stx1 did not. This suggests that factors that influence
receptor binding are complex, and will be discussed further in Chapter 2.
It is unclear why some individuals develop HUS and others do not. It has been
hypothesized that Gb3 may be expressed at higher levels in patients with severe illness.
GalT6, or Gb3 synthase, is involved in the synthesis of Gb3, therefore, differential
expression of Gb3 on tissues may correlate to the levels of the enzyme GalT6 present within the tissues [74]. It was shown that levels of GalT6 were higher in Stx-susceptible
cultured human proximal kidney cells expressing Gb3 as opposed to Stx2-resistant
cultured brain endothelial microvascular which express low levels of Gb3 [75],
suggesting that disease development may result from higher levels of Gb3 expression in
HUS patients.
Linking in vitro data such as these to actual outcome in patients has been difficult.
For instance, one study showed that expression of Gb3 in the kidney was higher in adults
than infants[73]. This contradicts the preferential development of HUS in children,
which only rarely manifests in adults. In contrast, a second group analyzed samples from
patients ranging in age from 7 months to 85 years and found that the Gb3 was expressed equally in all patients, and as a result, Stx1 bound equally to all tissue samples [76].
22 While it is of importance to study the location of Gb3 in human tissues, it should be noted
that Stx2, the toxin associated with disease development, was not tested. In addition, these samples were not taken from HUS patients, therefore the receptor expression may not correlate to levels seen in patient tissues. This may explain why the data do not reflect the age-specific development of severe disease.
That being said, the acquisition of the appropriate samples is logistically
challenging yet one group was able to attain kidney sections from two patients, aged 21
months and 81 years, who died from E. coli O157:H7-related HUS [77], and one control
case, a 40-year old renal cancer patient. Although the study was small, the data begin to
reflect the Stx binding activities within the patient. Tissue sections were incubated with
labeled Stx1 or Stx2, and binding was detected using immunoperoxidase staining. It was
found that Stx2, not Stx1, bound to both samples, and neither Stx bound to the control.
TOXIN-Gb3 INTERACTION
As described earlier, E. coli O157:H7 can produce Stx1, Stx2, or both [5,46,47].
Strains most commonly isolated from diseased patients express Stx2 [48]. Although 125I-
labelled Stx1 binds 50 times greater than Stx2 to human endothelial cells, Stx2 is more
toxic [78]. The differential binding affinities for the toxins for this receptor may play a
role in disease development. For instance, Stx1 has been shown to localize in the lung
[79]. Stx1 may thus bind tightly to Gb3 in areas of the body not affected by the toxin,
preventing it from binding to the Gb3 expressed in the renal system. The lower affinity of
Stx2 to Gb3 may also allow the toxin to travel more freely throughout the body.
23 It is therefore critical to study the differences in the binding between Stx1 and
Stx2 in order to gain knowledge about disease development. The X-ray crystal structure
of Stx1B complexed with the Gb3 receptor analog Pk-MCO, which contains the terminal
disaccharide of Gb3, was determined [80]. The crystal structures of both Stx and Stx1
indicate that each individual toxin B subunit has three predicted Gb3 binding sites, thus,
the pentameric complex has 15 potential binding sites [80-82]. All of the binding sites are
located on the plane of the pentamer opposite the A subunit [80].
Binding site 1 is present in the cleft formed between adjacent subunits.
Within site 1, Asp17 and Phe30 (Figure 3) mediate the Stx interaction with Gal2 of the
receptor [80]. X-ray crystal structure of the toxin in which Phe30 had been mutated to
an alanine indicated that the overall structure of the mutant was identical to that of wild
type, yet the mutation resulted in a 4 fold decease in the Gb3 binding affinity over wild
type, in addition to a marked reduction in the cytotoxicity on Vero cells [65]. This evidence reveals the critical role of Phe30 in Stx1 binding to Gb3. It is imperative to note
that Stx2 has a tryptophan residue at position 30, not a phenylalanine (Figure 3).
Binding site 2 is the major binding site for Gb3 [80,82,83] as evidenced by the
fact that all three sugars of Pk-MCO are able to interact at this binding site [80]. Stx1B with mutations in binding sites 1 and 3 but an intact site 2 was able to bind to Pk-MCO based on surface plasmon resonance [82]. Interaction between Stx1 and Pk-MCO involves hydrogen bonding of seven residues between the toxin and Gal1 [80]. Sequence
comparison (Figure 3) reveals that only four amino acids in the binding site differ
between Stx1 and Stx2. Those that vary are as follows, and the number designates the
position of the residues based on the location in Stx1. Stx1 contains Glu16, Asn32,
24 Asn55, and Ala56 while Asp16, Ser32, Ser55, and Thr56 are present in Stx2. The amino acids common to both toxins include arginine, glycine, and phenylalanine.
Binding site 3 is near the center of the B subunit where the B-pentamer binds the
A subunit. This site, located at the central pore of the pentamer, has few hydrogen bonds and is highly dependent on hydrophobic interactions [80]. The Pk-MCO sugars have the least interaction with the toxin at site 3 as the orientation of Pk-MCO is nearly perpendicular to the toxin [80]. Mutation of site 3 did not seem to affect the binding of
Stx1B to Gb3 based on surface plasmon resonance [82]. Sequence comparison showed that the tryptophan and asparagine in site 3 are conserved between Stx1 and Stx2.
25 1 20 Stx1 T P D C V T G K V E Y T K Y N D D D T F T V K V G Stx2 - A D C A K G K I E F S K Y N E D D T F T V K V D
β1 β2 β3
40 Stx1 D K E L F T N R W N L Q S L L L S A Q I T G M T Stx2 G K E Y W T S R W N L Q P L L Q S A Q L T G M T
β4 α
60 Stx1 V T IKTN A CHNG G G F S E V I F R- - Stx2 V T IKSS T CESG S G F A E V Q F NDD
β5 β6
Figure 3. Sequence alignment of Stx1 and Stx2 B subunits. Amino acids conserved between subunits are shaded in pink. Secondary structure designations are beneath the sequence. Residues important in binding, yet differ between Stx1 and Stx2, are outlined in black.
26
TOXIN-Gb3 INTERNALIZATION
Gb3 colocalizes with lipid rafts markers on the cellular membrane [84]. Lipid
rafts are present in cellular membranes and are distinguished by a lipid composition that
varies from the rest of the membrane [85]. For instance, in polarized epithelial cells, lipid
rafts composed of glycosphingolipids and cholesterol are localized to the apical
membrane [85].
After binding to the receptor, Stx1 and Stx2 may enter the cell through
endocytosis. Proteins taken into a cell through this process are normally routed to the
lysosome and degraded. However, Stx has established a mechanism to avoid the
degradation process in some cells. In toxin-sensitive cells, Shiga toxin avoided the
endocytic pathway, proceeding through what is termed retrograde transport, traveling
through the trans- Golgi network to the Golgi apparatus [86]. In contrast, cells resistant
to toxin have been shown to route the toxin to the lysosomes [86].
The exact mechanism of entry is still debated. Initial data indicated that after the
toxin binds to Gb3, the complex is quickly internalized into clatharin-coated pits [87,88].
Yet, Stx1 can localize to Golgi apparatus in clathrin-mediated endocytosis-deficient cells
[89,90].
A second route of entry has been proposed to occur via caveolae, which are
plasma membrane invaginations with a composition high in cholesterol, similar to that of
lipid rafts [85]. Stx1 can bind to Gb3 in cells in which the cholesterol has been disrupted,
yet the toxin-receptor complex cannot be internalized, suggesting a calveoli-dependent mechanism of entry [84,90]. This route of internalization is also actin-dependent, and it
27 has been shown that Stx1 binding induces a recruitment and phosphorylation of proteins
that link the plasma membrane to actin to the membrane, resulting in a cytoskeletal
rearrangement [91,92]. Based on these two lines of data, a consensus as to the exact
mechanism, or combination thereof, of entry has yet to be determined.
Proteolytic cleavage of the A-subunit occurs during retrograde transport, and is required for enzymatic activity [93,94]. Furin cleaves the A-subunit into the
enzymatically active A1 and the B-subunit-binding A2 fragments at a trypsin-sensitive site
between a disulfide bridge at the carboxyl terminus of the A subunit [59]. Cleavage
occurs in the presence of brefeldin A, a chemical that disrupts the Golgi apparatus. This
suggests that cleavage occurs prior to toxin entry into the Golgi apparatus.
After translocation through the Golgi, Shiga toxin proceeds to the endoplasmic
reticulum (ER) [95]. This process was visualized using fluorescently labeled B-subunit
in HeLa cells [96]. Travel from the Golgi to the ER may occur through a microtubule-
dependent pathway [97]. Interestingly, Shiga toxin does not contain the ER-targeting sequence KDEL. Sandvig and colleagues have suggested that the Gb3, as opposed to the
toxin, may contain the retrograde transport signal within the composition of the fatty acid
chain [98].
SHIGA TOXIN GLYCOSIDASE ACTIVITY
Stx inhibits protein synthesis within the cell. As an N-glycosidase, Stx cleaves
the adenine at position 4324 within the 28S ribosomal subunit, rendering the release of
28 the ribosomes from the mRNA, inhibiting translation [99]. The N-glycosidase activity of
Stx and Stx2 was verified in vivo with the use of Xenopus oocytes [100].
ANTIBIOTIC TREATMENT OF E. COLI O157:H7
The effectiveness of antibiotics in the treatment of E. coli O157:H7 has been
called into question. Several studies have suggested that antibiotics may induce Stx
production in vitro [54,56], Wong et al. initiated a retrospective study to determine the
effect of antibiotics on disease severity in patients [8]. In collaboration with labs from
Washington, Oregon, Idaho, and Wyoming, data from patient labs and parent/hospital
questionnaires were compiled for all children who were infected between 1997-9. It was
shown that administration of antibiotics increased the development of HUS in those
children from 8% to 56% [8].
Several labs have studied the ability of antibiotics to induce Stx production in
vitro. In most cases, only a few antibiotics were analyzed per study, and the majority of
techniques used did not measure biologically active Stx [54,56]. The antibiotics tested in
my studies target the DNA (ciprofloxacin), folate metabolism
(trimethoprim/sulfamethoxazole), cell wall synthesis (ampicillin, ceftriaxone,
fosfomicin), transcription (rifampicin), and translation (gentamicin, doxycyline, and azithromycin). The following sections describe the mechanisms of action, therapeutic information, and E. coli O157:H7-related data on the antibiotics tested.
29 Ciprofloxacin
Therapeutic information
Ciprofloxacin, taken orally, is commonly prescribed for adult patients with
diarrhea, and is not recommended for children. The antibiotic is primarily absorbed
through the intestinal duodenum and jejunum, passes through the kidneys, and is then
excreted unchanged in the urine in healthy patients. In patients with severe renal defects,
the serum half life of ciprofloxacin increases from 3-6 hours to 6-9 hours, and exit from
the kidneys is delayed [101].
Mechanism of Action
DNA gyrase is a 400 kDa bacterial protein that relaxes the coiled DNA during
replication, making the strand accessible to polymerase. DNA gyrase is comprised of
two subunits, termed GyrA and GyrB. GyrA cleaves both strands of DNA, 4 bases apart, forming covalent bonds with free 5’ ends of the DNA. The DNA is then passed through the break and religated, creating what is termed a negative supercoil. The GyrB carries the ATPase activity needed for this process. [102].
Quinolones, of which ciprofloxacin is an example, inhibit this process likely by
binding to the free 3’ end of the DNA, stabilizing the gyrase-DNA intermediate.[102].
Even though DNA repair machinery is recruited to the site, the DNA remains bound to
the gyrase, ultimately inhibiting the production of replication forks down stream of the
cleavage site. This initiates the SOS response which in turn activates the expression of
Stx [103].
30 E. coli O157:H7-related studies
Several in vitro studies have linked ciprofloxacin to an increase in Stx production.
For instance, incubation of E. coli O157:H7 with subinhibitory concentrations of ciprofloxacin increased Stx1 production up to 400-fold based on an Gb3 Enzyme-linked immunosorbent assay (ELISA) [104]. Isolates producing Stx2 were not included in the study, nor was the biological activity tested in a cell-based assay.
Further studies by Kimmitt et al. approached the question by monitoring the
activity of β-galactosidase (lacZ) under the control of the Stx2A subunit promoter. Both
wild type lacZ and stx2A were removed from a clinical strain of E. coli O157:H7 and
lacZ was inserted 3’ of the Stx promoter. The strain was incubated on agar with
quinolone strips or disks in the presence of X-gal. β-galactosidase produced in the
presence of the ciprofloxacin diffused into the media, cleaving the X-gal, resulting in a
blue color around the antibiotic source. Kimmitt concluded that quinolones induced β-
galactosidase at concentrations above the MIC because cleavage of the X-gal occurred
within the zone of bacterial growth inhibition and thus, likely induce Stx as well.
Production of β-galactosidase suggests that the antibiotics have induced transcriptional
activity, yet does not necessarily reflect the amount of biologically active Stx being
produced. To test this, strains were incubated with quinolone antibiotics and a cell death
assay was used to determine the toxicity of each sample. Stx production increased 200-
fold. Specifically, ciprofloxacin, a quinolone Stx inducer that is used as a control
throughout my studies, was not tested [54] [56].
31 Other studies have confirmed these results as E. coli O157:H7 strains isolated
from patients in Cincinnati had a ~15-fold increase in Stx2 production when incubated
with ciprofloxacin [105].
While the data supporting the ability of ciprofloxacin to induce Stx is extensive, it
should also be noted that induction is strain-specific. For instance, while subinhibitory
concentrations of ciprofloxacin increased Stx production in E. coli O157:H7 strains with
stx1 and stx1, stx2, ciprofloxacin inhibited Stx production in an E. coli O157:H7 strain
with stx2 alone [106]. Data will be presented in Chapter 3 that supports strain-specific
induction. We have found that ciprofloxacin significantly induces Stx production in
clinical E. coli O157:H7 strains containing either stx1 or both stx1 and stx2. Yet, in control strains lysogenized with either stx1 or stx2-containing phage, Stx production was increased only slightly, and in some instances, was even decreased.
Grif measured toxicity in two ways. Stx was diluted and incubated with Vero
cells for two days. After the Vero cells were fixed, washed extensively, stained with
crystal violet, then washed again, the absorbance of the stained cells that remained
adhered to the plate was measured. Ultimately, toxicity is being measured by the ability
of Vero cells to adhere to the wells, under the assumption that the Vero cells affected by
the Stx will detach from the plate and be washed away. We have used a similar protocol
in the lab, and have found that there are various stages of Vero cell plate adherence, and
that this protocol is too harsh to be uniform. The second method measured by amount of
lactate dehydrogenase released by the dying cell, again, an arguably ambiguous process.
While these tests may reflect downstream effects of the Stx, in my studies I have
measured protein synthesis inhibition, the actual mechanism of Stx toxicity.
32 In vivo studies have also verified ciprofloxacin-induced Stx production. Mice infected with E. coli O157:H7 then treated with ciprofloxacin did in fact have a 3-log
reduction in the amount of bacteria present in the feces, but they had 10-40 fold more
fecally-associated Stx than controls [107].
Similarly, studies by Gamage et al. also indicated that ciprofloxacin both reduced
the presence of E. coli O157:H7 is the feces of infected mice by 2 logs while levels of
Stx2 significantly increased. It should also be noted that the increased Stx2 production varied greatly from trial to trial. In one instance, Stx2 levels increased nearly 100-fold, while in a second trial, the increase was less than 10-fold [108].
In 2000, Israeli doctors tested the safety of ciprofloxacin on children with diarrhea
because it can be administered orally unlike the current antibiotic of choice, ceftriaxone, which must be injected. Because 99% of children were cleared of the infections of which
E. coli was one, they suggested that ciprofloxacin might be safe for use in children with invasive diarrhea as defined by the presence of fecal leukocytes. Tests were run to classify the types of E. coli isolated from patients as being either diarrheagenic, enterpathogenic, or enteroinvasive. Tests to identify E. coli O157:H7 were not performed. It is therefore not surprising that the study failed to mention that fecal leukocytes are also present in Shiga toxin-producing E. coli infections [109]. So alarming was this oversight, that Acheson immediately published an article which he emphatically deemed the conclusion of the study unacceptable. The method of diagnosis suggested by these doctors could include children infected with E. coli O157:H7, who would then be given the known Stx-inducer ciprofloxacin as a treatment, putting them a high risk of developing severe disease [110].
33
Trimethoprim/Sulfamethoxazole
Therapeutic information
Trimethoprim and sulfamethoxazole are administered as a pair either orally or
intravenously for patients with urinary tract infections and more importantly, children
with diarrhea. These antibiotics are readily absorbed in the gastrointestinal tract. Serum levels peak two hours after administration and hold constant for about 6 hours. These antibiotics can be inactivated in the liver, yet it has been found that even in patients with severe renal disease, these antibiotics are present in the urine at concentrations high enough to clear urinary tract infections [101].
Mechanism of Action
Folate is a B vitamin. Acquisition of folate in the diet is critical as folate
metabolism is required for the synthesis of the DNA base, thymine. During the first step
of metabolism, folate is reduced to dihydrofolate. Dihydrofolate reductase then catalyzes
a reduction reaction to produce tetrahydrofolate. Tetrahydrofolate is a precursor for
thymine. Trimethoprim targets this metabolic activity by blocking the action of the
dihydrofolate reductase, ultimately inhibiting DNA synthesis [103].
Unlike humans, bacteria are able to synthesize folate. A reaction between p-
aminobenzoate (PABA) and dihydropteroate diphosphate is catalyzed by the enzyme
dihydropteroate synthase, producing dihydropteroate. Dihydropteroate is then
metabolized into tetrahydrofolate, the thymine precursor also produced in humans.
Sulfamethoxazole, which structurally mimics the PABA active site, acts as an alternative
34 substrate for the dihydropteroate synthase, inhibiting DNA synthesis within the cell
[103].
E. coli O157:H7-related studies
In vitro studies have examined the effect of trimethoprim/sulfamethoxazole on
Stx production. Production of Stx1 by E. coli O157:H7 increased 2-fold in the presence
of subinhibitory concentrations of this pair of antibiotics, based on a Gb3 ELISA [104].
Grif et al. also showed that this antibiotic combination increased Stx in E. coli O157:H7
isolates that contained stx1, stx2, or both [106].
The in vivo effect of trimethoprim/sulfamethoxazole was much more dramatic.
E. coli O157:H7-infected mice that were treated with trimethoprim/sulfamethoxazole
three days after infection had increased levels of Stx in the feces compared to control animals. In addition, the animals displayed a 95% mortality rate as compared to the control mortality rate of 85% [111]. Interestingly, if the antibiotics were administered earlier post-infection, the mice survived, suggesting that antibiotic-linked disease progression may be associated with the timing of antibiotic administration.
Data regarding the effect of trimethoprim/sulfamethoxazole on disease progression in patients is also available. In one instance, both workers and residents at an
institution for mentally retarded persons became infected with E. coli O157:H7 [112]. Of the 8 patients with HUS, 5 received trimethoprim/sulfamethoxazole while those without
HUS did not receive the antibiotic. While this may suggest that trimethoprim/sulfamethoxazole is associated with disease development, it should be noted that patients who developed HUS were on average 13 years old, while those who
35 did not were 27 years old. Therefore, other age-related factors may contribute to the
disease state.
Two additional studies that analyzed US patient records both nationally [113] and
in the northwest [8] indicated that sulfa-containing drugs of which
trimethoprim/sulfamethoxazole is an example were linked to the development of HUS. It was shown that HUS development in patients under 13 years of age was linked to antibiotic administration within three days of the onset of diarrhea [113].
Ampicillin
Therapeutic information
Ampicillin can be administered orally, intramuscularly, or intravenously. This
antibiotic is safe for people of all ages, including newborns and infants. Peak serum
levels are reached between 30 mins and 2 hours, based on the method of delivery.
Ampicillin can be inactivated in the liver, and 30% of the antibiotic is excreted in the
urine 6 hours post-administration [101].
Mechanism of action
The peptidoglycan layer of a bacterium’s cell wall is composed of the sugars N-
Acetyl-Glucosamine and N-Acetyl-Muramic acid, which are bound to peptides.
Transpeptidases catalyze the linkage of peptides between sugar chains, providing structural support within the layer. Ampicillin binds to the transpeptidases, inhibiting the linkage formation. Absence of the peptide cross-linking weakens the bacterial membrane, allowing the cell to become increasingly sensitive to osmotic pressures.
36 E. coli O157:H7-related studies
Two studies have examined the effect of ampicillin on the production of Stx.
The in vitro study indicated that subinhibitory concentrations of the antibiotic increased
Stx production in an stx2-containing isolate, yet had no effect on Stx production in
isolates containing stx1 or stx1 and stx2 compared to the control [106]. These
conclusions differ from the animal study in which fecally-associated Stx in mice treated
with ampicillin one day post-infection with an stx1/stx2 strain was below the limit of
detection as compared to controls which contained 40 pg/mg of Stx in the stool [111].
.
Ceftriaxone
Therapeutic information
Ceftriaxone is highly potent against enterobacteriacea. This antibiotic is most
suited for intramuscular delivery; it will not be absorbed if given orally. Due to its long
half-life, ceftriaxone is administered every 12-24 hours. Ceftriaxone is excreted from the
body unchanged. 40-50% of the antibiotic is present in the urine after 48 hours while the
remainder is present in bile [101].
Mechanism of action
The mechanism of action is similar to that of ampicillin.
E. coli O157:H7-related studies
The ability of ceftriaxone to increase or decrease Stx production is strain-specific.
One study has shown that subinhibitory concentrations of ceftriaxone increased Stx2 production in one strain, yet decreased Stx production in stx1 and stx1/stx2 strains [106].
37 Fosfomicin
Therapeutic information
Currently, fosfomycin is not used therapeutically in the United States. Because studies in Japan suggest that it is beneficial in the treatment of E. coli O157:H7 infections, it has been included in this study. Fosfomycin is administered orally and is quite effective against E. coli, for example, in urinary tract infections [101].
Mechanism of action
Peptidoglycan synthesis begins cytoplasmically and begins with the conversion of the sugar UDP-N-Acetyl-Glucosamine into UDP-N-Acetyl-Muramic acid. Fosfomicin covalently binds to the active site of MurA, one of two enzymes involved in this initial reaction [103].
E. coli O157:H7-related studies
The effect of fosfomicin on Shiga toxin production varies based on the experimental conditions. For instance, reversed passive latex agglutination assay (RPLA) analysis of the supernatant of an E. coli O157:H7 strain incubated with fosfomicin indicated that Stx1 production increased 49-fold while Stx2 levels were reduced by half
[114]. A second study showed that while Stx production was unaffected by either fosfomicin or kanamycin alone, the antibiotics acted synergistically to decrease Stx production [115]. In vivo, fosfomicin significantly reduced fecally-associated Stx in infected mice compared to untreated control animals [116].
38
Gentamicin
Therapeutic information
Gentamicin is administered either intramuscularly or intravenously. 40% of the total gentamicin is present in the kidneys. The half-life of gentamicin is 4-8 hours in a normal patient, and is increased up to 1000-fold in patients with severe renal disease.
Gentamicin is excreted from the body in its active form in the urine [101].
Mechanism of action
tRNAs bind to the ribosome and deliver an amino acid to a growing chain of mRNA, first by binding to the A-site of the ribosome, then moving to the P position.
Gentamicin binds to the 30S ribosome, inhibiting the tRNA from accessing the P position, halting protein synthesis [103].
E. coli O157:H7-related studies
The one study that examined the effect of gentamicin on Stx production found that subinhibitory concentrations of the antibiotic increased Stx production in isolates that contained stx1, stx2, or both stx1 and stx2 [106].
Doxycyline
Therapeutic information
The lipid solubility of doxycyline allows for an increased ability to cross the bacterial membrane. After oral or intravenous administration, this antibiotic is absorbed in the intestinal epithelium. 20% of orally administered doxycyline is excreted in the
39 urine or feces. The amount present in the feces increases in patients with renal impairment.
Mechanism of action
tRNAs bind to the ribosome and deliver an amino acid to a growing chain of mRNA. Doxycyline binds to the 30S subunit of the ribosome. While the tRNA can bind to the ribosome, it cannot assume the correct conformation needed due to the allosteric inhibition by doxycycline, and is released, halting protein synthesis [103].
E. coli O157:H7-related studies
Currently, there are no reports on the effect of doxycyline in the treatment of E. coli O157:H7-related disease.
Azithromycin
Therapeutic information
Azithromycin is administered orally. It is effective against gastrointestinal pathogens, yet is commonly administered to children for inner ear infections and respiratory diseases. Azithromycin is present in the tissues more so than the serum. The majority of the antibiotic is metabolized by the liver [101]. Compared to similar antibiotics, azithromycin causes less gastrointestinal irritation, has an increased stability in the pH of the stomach, better tissue penetration, and an increased half-life.
Mechanism of action
Azithromycin binds to the 23S RNA. Azithromycin binds to the ribosomal exit tunnel, causing the early release of the tRNA, again inhibiting protein synthesis [103].
40 E. coli O157:H7-related studies
One in vitro study conducted in Japan has shown that various concentrations of
azithromycin did not stimulate the production of Stx1 or Stx2 [117]. Azithromycin was also found to be protective in an in vivo study [118]. Two mouse experiments were performed. In the first, mice were injected with Stx1, then azithromycin was administered. The choice to test Stx1 is questionable as it is Stx2 that is associated with severe disease, and the authors described preparing Stx2 in the methods section. Even so, mice receiving the antibiotic had a survival rate of 40%, unlike the control group in which all mice died. A second group of mice were then infected with a Stx-producing strain of E. coli. All the mice receiving azithromycin survived as compared to the controls which again, had a 100% mortality rate. It should be noted that although the authors had access to E. coli O157:H7 strains, the strain tested in this study was serotype
O86, and there is not mention as to whether or not the strain produces Stx1 or Stx2.
Rifampicin
Therapeutic information
Rifampicin is commonly used to treat tuberculosis. It is administered orally
although the intravenous route is also an option. Rifampicin is administered in
combination with other antibiotics since resistance to rifampicin is established quickly.
This antibiotic is absorbed within the gastrointestinal tract. More so than the serum,
rifampicin is present within the lungs, liver, and stomach wall. Rifampicin is metabolized
by the liver. While a small fraction is excreted in the urine, the majority will be excreted
in the feces [101].
41 Mechanism of action
RNA polymerase is comprised of αββγ and σ subunits and its function is to transcribe DNA into RNA. Rifampicin inhibits this process by binding to the β subunit
of the enzyme. This allosteric inhibition blocks the RNA polymerase tunnel through
which the DNA/RNA are threaded, inhibiting transcription.
E. coli O157:H7-related studies
The ability of rifampicin to inhibit the transcription of Stx was measured in vitro.
Seven E. coli O157:H7 isolates were incubated with a concentration of rifampicin equal
to the MIC. While RT-PCR (reverse transcriptase polymerase chain reaction) indicated
that neither stx1 or stx2 were transcribed, an RPLA (reversed passive latex agglutination)
assay indicated that toxin production was decreased by 12-16% [119].
ROLE OF NORMAL INTESTINAL FLORA IN DISEASE
Two facts lead to the hypothesis that E. coli O157:H7 could instruct the
nonpathogenic E. coli present in the natural flora of the human gut to produce toxin.
First, other bacteria have dedicated toxin secretion systems used to release AB5 toxins
from the bacteria [120]. Yet, E. coli O157:H7 does not have a secretion system for Shiga
toxin. In E. coli O157:H7, toxin production is under the regulation of a phage late
promoter as are the lysis genes [51]. Toxin is thereby released not by secretion, but
during lysis of the host cell. Secondly, the infectious dose of E. coli O157:H7 can be as
low as 50 organisms [121]. In contemplating how so few bacteria could result in
massive, systemic disease, it was noted that the Stx is phage encoded, and phage are
42 released during lysis. It was hypothesized that toxin amplification would occur if the
phage released by E. coli O157:H7 in the intestine were able to infect the host E. coli.
Phage and toxin production would then ensue not only in the E. coli O157:H7, but within
the host E. coli as well.
To test this hypothesis, human E. coli isolates from healthy volunteers were
incubated with Stx2-encoding phage prepared from E. coli O157:H7. After overnight
incubation, the supernatants were characterized. Although 33 of the tested strains were
resistant to the phage infection, three isolates were shown to have a significant increase in the amount of Stx2 present in the supernatants as tested by the ELISA assay [122]. This
indicated that the commensal E. coli strains amplified Stx2 production.
The effect of the commensal E. coli was also tested in vivo [108]. Mice were
treated with streptomycin to reduce their intestinal flora, and then were colonized with a
human intestinal isolate that was either phage susceptible or resistant to phage infection.
Mice were subsequently infected with E. coli O157:H7. Colonization was measured by
resuspending the feces and plating on media supplemented with antibiotics. E. coli
O157:H7 in the absence of the commensal isolate colonized at a rate of ~108 – 109
CFU/g. The human intestinal isolates colonized the intestine at a level of 107 – 108, and
reduced E. coli O157:H7 colonization between 10-100 fold.
Fecally-associated Stx was measured via an ELISA. Stx was isolated from mice
infected with E. coli O157:H7 in the presence or absence of the phage susceptible
commensal isolate at a similar rate. Because colonization of E. coli O157:H7 was
decreased in the presence of the intestinal isolate, it can be argued that the commensal
isolate may have amplified Stx production although the raw Stx values are similar to E.
43 coli O157:H7 alone. In contrast, Stx was never recovered from an animal infected with
E. coli O157:H7 in the presence of a phage resistant intestinal isolate. This argues that
human intestinal isolates may play a role in disease both by influencing colonization and
Stx production by the pathogenic bacteria.
In my studies, I have tested a panel of antibiotics to determine which, if any, are
safe for use in the treatment of E. coli O157:H7. Several labs have studied this question
as well. In incubating E. coli O157:H7 isolates with subinhibitory concentrations of
antibiotic and measuring Stx production, my data can be compared to what other labs
have observed [104,106,117]. Traditionally, these studies test only a small number of
antibiotics on one strain of E. coli, unlike the nine that I have used to test four strains, two
of which are clinical E. coli O157:H7 isolates. In addition, the MIC values and Stx
induction levels are usually determined using stationary phase cultures. As bacteria grow
logarithmically during infection, I have included this growth phase in both the
determination of the MIC and the induction studies.
The second phase of the study begins to examine the effect of commensal isolates
on Stx production in the presence of antibiotics, a study which has not yet been done. To
begin to examine the effectiveness of antibiotic treatment in patients harboring a
commensal Stx-amplifying strain of E. coli, my studies also include measuring Stx levels
produced in a coculture of E. coli O157:H7, a phage-susceptible, Stx-amplifying strain,
and antibiotic.
44 LIPOPOLYSACCHARIDE OF FECAL-ISOLATE 29
The unexpected discovery of one toxin-neutralizing isolate of commensal E. coli
was most interesting [105]. The isolate was termed Fecal Isolate-29, or FI-29.
Preincubation of FI-29 supernatant with Stx2 decreased the toxicity on Vero cells, yet in
the same assay, Stx1 remained toxic [105].
Fractionation of the FI-29 supernatant using ammonium sulfate and the FPLC
Sephadex-200 column indicated that the neutralizing activity was in the void volume
[105]. Moieties present in this fraction are large, over 1,000 kDa, and a Bradford assay indicated that there were no proteins present [105]. It was also found that the neutralizing
activity was present in a crude preparation of membrane vesicles. This lead to the hypothesis that the neutralizing component could be the lipopolysaccharides (LPS) present on the bacterial membrane [105] as it may resemble the Gb3 receptor.
Lipopolysaccharide (LPS), otherwise known as endotoxin, is a component of Gram-
negative bacterial outer membrane (Figure 4). LPS has a conserved structure. Lipid A is
the hydrophobic component, and is responsible for the endotoxic activity [123]. The
inner core of LPS is comprised of 3-deoxy-D-manno octulosonic acid (KDO) and l-
glycero-D-manno- heptose (Hep). The outer core consists of sugars such as hexoses and
hexosamines. Attached to the core, extending away from the bacterial membrane, are
chains of repeating subunits called O-polysaccharides. Each subunit may contain
between one and eight glycosyl residues. The number of O-polysaccharide subunits per
molecule of LPS varies greatly, ranging between 0 to 50, and the LPS on a single
bacterium is heterogeneous in length. The O-sugars are antigenic and serotyping of the
45 LPS type is useful for distinguishing between strains of E. coli. For example, E. coli
O157:H7 was the 157th distinct O-serotype discovered. Currently, there are over 170 serotypes of E. coli.
Lipopolysaccharide serotyping of FI-29 indicated that it belongs to two cross-
reacting serotypes: O107 and O117 [124]. ELISAs, in which FI-29, O107, and O117
LPS were bound to a hydrophobic microtiter plate via the lipid A, orienting the LPS as it
would be in the bacterial membrane, indicated that FI-29 and O117 LPS bound Stx2 with
the highest affinity [124]. Preincubation of either Stx1 or Stx2 with FI-29, O107, and
O117 LPS reduced the Stx2 toxicity to Vero cells, yet Stx1 remained fully toxic [124].
Sugar chains are a major component of both LPS and Gb3. The O-polysaccharide
subunit structure from O117 has been elucidated by nuclear magnetic resonance (NMR)
[125] and was found to resemble the toxin binding sugars of Gb3. Both have a terminal
and penultimate galactose moieties. However, O117 is different in that both galactose
molecules are N-acetylated. Determination of the structure of FI-29 and O107 LPS is
underway, and preliminary results suggest that all three LPS molecules possess an N-
acetylated moiety, but are not identical. Thus, FI-29 LPS may neutralize Stx2 by acting as a Gb3 mimic. Various compounds were designed based on the LPS structures, and
have been created by Dr. Suri Iyer’s laboratory in the Department of Chemistry here at
the University of Cincinnati. The analysis of these compounds will be described in an
upcoming chapter.
46 n O-polysaccharide Outer core Outer
Hep Hep Inner core
Hep
KDO KDO
Glc Glc Lipid A
Figure 5. The structure of lipopolysaccharide (LPS). LPS is anchored to the bacterial membrane via Lipid A. The inner core is comprised of both KDO and heptose while the outer core contains hexoses. Varying numbers of O- polysaccharide subunits occupy the terminal position. Adapted from Caroff, Microbes and Infection, 2002.
47 LABORATORY ASSAYS
Sorbitol-MacConkey Agar
In 1986, March and Ratnam exploited the inability of E. coli O157:H7 to ferment
sorbitol in order to develop an assay that could be used to differentiate the organisms
from other normal E. coli flora [126]. By replacing the lactose in MacConkey media
with sorbitol, the authors created a differential media on which E. coli O157:H7 strains
are visible as white colonies, while normal floral isolates appear pink in color. This diagnostic was immediately put into routine practice worldwide [127,128] and is still
used today both scientifically and diagnostically [129-132].
Latex agglutination assay
While useful, plating on differential media is now performed diagnostically only
as a starting point, followed by a battery of additional assays that will be discussed
throughout this section. For instance, the latex agglutination test, which relies on visible
bacterial clumping as they bind to the test antibody was deemed effective in recognizing
the O157 antigen on this pathogenic bacterium [133]. The O157 polysaccharide latex
agglutination test has recently been used to test for the presence of E. coli O157 in both cattle [134] and human serum [135].
Over time, an agglutination assay specific for the detection of Stx1 or Stx2 was
developed. The effectiveness of the assay to detect Stx1 and Stx2 was compared to the
Vero assay was tested. The limit of detection for purified Stx1 was 0.7 ng/ml and for
purified Stx2 was 0.6 ng/ml and these values were 2-fold more sensitive than the Vero
48 assay. Culture filtrates of Stx-producing strains of varying E. coli serotypes were also
tested. The agglutination assay specifically identified Stx1 and Stx2, and samples were
toxic to Vero cells [136].
The Stx agglutination assay has been used to measure the amount of Stx released
by an E. coli O157:H7 isolate in the presence of antibiotics [119,137]. For instance, it
was shown that rifampicin decreased Stx1 and Stx2 production 12-16-fold compared to
controls [119]. This method is beneficial in that it specifically detects either Stx1 or
Stx2. But because of certain limitations, this assay is better suited for the research labs as
opposed to diagnostic labs. For instance, the duration of the assay is 24 hours [119], which is less than optimal for E. coli O157:H7 diagnosis as the level of disease severity can increase rapidly over a few days time. In addition, the assay requires the use of antibodies. Antibodies used in these experiments are likely isolated from mice or rabbits
that had been injected with the toxin. As with all organisms, the immune system
response will differ greatly from individual to individual and therefore, the antibodies
produced by each animal will not necessarily be identical. Hence, the recognition capabilities of the antibodies may vary from lot to lot, introducing variability into each new set of experiments. Data from these assays also indicate the level of antigenic Stx present, not necessarily the amount of biologically active toxin produced.
Polymerase chain reaction
In 1992, Brian et al. developed a PCR-based assay in which primers specific for
stx1 or stx2 were used to determine if Shiga toxin producing E. coli (including O157:H7) bacteria were present in the stools of children with HUS [138]. Not unlike the other
49 assays described in this section, PCR has the added advantage of detecting Stx-producing strains with serotypes other than O157 [139]. Stx-specific primers can be used to assay for Stx genes present E. coli strains in a variety of sources such as cattle [17,140] the environment [141,142], patients [143-146], and even test foodstuffs [147].
In summary, PCR is advantageous in that the assay can be performed in one day. In
addition, primer sets can be used to identify Stx-producing E. coli strains regardless of
serotype or source of isolation. Yet, strains that carry stx1 and/or stx2 genes produce
varying amounts of Stx [148]. Other PCR positive isolates have been shown to be
negative for Stx production using assays that detect antigenic [149] and active forms of
the toxin [150]. Data will be presented in Chapter 3 that indicates that a clinical E. coli
O157:H7 stx1+/stx2+ isolate produces Stx2 alone. Therefore, the presence of the either
stx gene does not guarantee the expression of functional toxin.
Western blot
To identify proteins of interest in a Western blot assay, eletrophoresed samples
are bound to a membrane, and proteins are ultimately detected via antibody binding activities. This process is normally carried out over the course of two days, and is therefore more appealing to the research scientist as compared to the clinical
microbiologist. For instance, monoclonal antibodies that recognize either Stx1 or Stx2
were used to study Stx binding to LPS using Western blot analysis [124]. This assay is
both beneficial and limiting for the reasons linked to the use of antibodies as described in
the agglutination assays.
50
MY STUDIES
E. coli O157:H7 can produce Stx1, Stx2, or both, yet it is Stx2 that is most
commonly associated with severe disease [48]. The ELISA-based diagnostic currently
used to identify E. coli strains as Stx producers cannot distinguish between Stx1 and
Stx2, and rely on the use of capture antibodies [151]. In collaboration with the Iyer lab, we have developed an ELISA that specifically recognizes Stx1 or Stx2 through the use of
Stx-specific glycoconjugates. Characterization of this assay will be described in Chapter
2.
Immortalized African green monkey kidney cells, called Vero cells, are used to
measure biologically active Stx. Various types of assays are designed to use this cell
line, and each has its own limitations. For instance, one procedure requires a three-day
incubation period [122]. Others measure toxicity by monitoring cellular processes such
as metabolism [152] and apoptosis [153]. Based on a previous study [154], we have
designed a Vero cell line that stably expresses luciferase. Stx mediates protein synthesis
inhibition [99], and this mechanism can observed via luciferase in four hours. The ability
of these cells to monitor the mechanistic process of Stx within hours will be a significant
contribution to this field of study. Characterization of this cell line will be described in
Chapter 3.
Various factors may play a role in disease severity. It has been suggested that antibiotic treatment has been linked to the development of HUS in patients [8,108]. As a result, patients with E. coli O157:H7 do not receive antibiotics. Both in vitro and in vivo studies have confirmed that antibiotics such as ciprofloxacin can in fact increase Stx
51 production [54], yet, it was discovered that antibiotics such as rifampicin can reduce Stx
production [119].
Other studies have indicated that ten percent of the isolates attained from
volunteers harbor phage susceptible commensal E. coli strains that amplify Stx [122]. As
ten percent of people with E. coli O157:H7-related disease develop HUS, this study
suggests that commensal E. coli present in the patient gut may play a significant role in
disease development. Therefore, it remains to be determined if HUS development is due
to antibiotic administration, the presence of a Stx-amplifying commensal E. coli, or other
factors.
Chapter 3 includes data that begins to address the appropriateness of antibiotic therapies within the context of commensal E. coli.
52 Chapter II.
Glycoconjugates that specifically bind Stx1 or Stx2 1
1 Portions of the following text have been previously published in:
Gamage, S.D., McGannon, C.M., and Weiss, A.A. 2004. Escherichia coli serogroup
O107/O117 lipopolysaccharide binds and neutralizes Shiga toxin 2. Journal of
Bacteriology. 186:5506-12. My experimental contribution: Figure 3.
and
Kale, R.R., McGannon, C.M., Fuller-Schaefer, C., Hatch. D.M., Flagler, M.J., Gamage,
S.D., Weiss, A.A., and Iyer, S.I. 2008. Differentiation between structurally homologous
Shiga 1 and Shiga 2 toxins by using synthetic glycoconjugates. Angewandte Chemie
International Edition. 47:1265-8. My experimental contribution: Figure 4.
53 Abstract
Strains of E. coli O157:H7 that produce the major virulence factor Shiga toxin 2
are most frequently isolated from patients with severe disease. E. coli isolates naturally present in the human gut may influence disease severity by influencing Shiga toxin production or availability. For instance, the lipopolysaccharide of a normal isolate of E.
coli, termed FI-29, was shown to neutralize the cytotoxicity of Stx2, not Stx1, on Vero cells, indicating that persons harboring this strain may be protected from disease. Here we demonstrate that purified lipopolysaccharide (LPS) from FI-29 and the two cross- reacting serotypes O107 and O117 directly bind to Stx2 in an ELISA assay. The recognition elements of both the Shiga toxin receptor Gb3 and O117 LPS are similar in
structure. Gb3 contains two terminal galactose moieties while O117 LPS has two
terminal N-acetylated galactose residues. Using these structures as models, synthetic
glycoconjugates were constructed by a collaborating laboratory. We have shown that
synthetic glycoconjugates base on Gb3 specifically recognize Stx1, while synthetic
glycoconjugates base on O117 LPS specifically recognize Stx2, and that spacer length
between the recognition element and the biotin affects the binding properties.
54 Introduction
E. coli O157:H7 is a food borne pathogen that is the leading cause of hemolytic
uremic syndrome in children. These bacteria produce a major virulence factor termed
Shiga toxin (Stx). Stx is an AB5 toxin in that the catalytically active A subunit rests on a
pentamer of B subunits. E. coli O157:H7 can produce Stx1, Stx2 or both, yet it is Stx2
that is most commonly associated with disease development [48].
Stx binds to the globotiaoslyceramide (Gb3) receptor [80], is internalized [88], and inhibits protein synthesis within the cell [99]. The terminal structure of Gb3 is
Galα(1→4)Galβ(1→4)Glc-ceramide, with the galactose moieties acting as the Shiga toxin recognition element [155] (Figure 1). Kinetic analysis of the binding of 125I-Stx1 or
125 I-Stx2 to the receptor was studied by incubating serial dilutions of either Stx with Gb3 bound to microtiter wells, then measuring the radioactivity present in the well after
-8 -7 washing [67]. The Kd for Stx1 was determined to be 4.6 x 10 M and for Stx2, 3.7 x 10
M [67]. The higher affinity of Stx1 for Gb3 as compared to Stx2 may be a factor in
disease development. One in vivo study has shown that the vast majority of Stx1
localized to the mouse lung and that the kidney contained more Stx2 than Stx1 [79].
Therefore, Stx1 may bind tightly to unaffected organs, whereas Stx2 may have more
freedom to circulate systemically, localizing to organs susceptible to disease such as the
kidney [79].
African green monkey kidney cells, or Vero cells, are susceptible to the cytotoxic
activity of Stx1 and Stx2 [124]. Previous studies characterized the effect of
nonpathogenic human intestinal E. coli isolates on the ability of E. coli O157:H7 to
55 produce Shiga toxin and discovered an isolate, termed Fecal isolate 29 (FI-29), that
neutralized the cytotoxicity of Stx2, not Stx1, on Vero cells [122]. It was determined that
the LPS O-antigen of FI-29 neutralized Stx2 likely by binding to the toxin [124]. The
generic structure of LPS is similar to that of Gb3. Both are anchored to the membrane via
lipid components and both contain sugar moieties that extend away from the cellular
surfaces. The LPS of FI-29 was serotyped and cross-reacted with two serotypes, O107
and O117. Fortuitously, the structure of serotype O117 was solved [125], and is depicted
in Figure 1. While Gb3 possesses two terminal Galα1-4Gal moieties, O117 LPS
possesses two terminal N-acetylated galactose moieties with the same linkage.
The Suri Iyer Laboratory from the Department of Chemistry used these structures
as a model to synthesize receptor mimics. These glycoconjugate compounds (pictured in
Figure 2) were biotinylated to facilitate the binding to a streptavadin-coated microtiter
well. GC-1 is a disaccharide with a 6 carbon spacer. GC-2a and GC-2b mimic Gb3 as trisaccharides with 6 and 12 carbon spacers, respectively. GC-3a and GC-3b mimic
O117 LPS as N-acetylated trisaccharides with 6 and 12 carbon spacers, respectively.
In this chapter, the ability of these natural and synthetic ligands to bind to Stx1 and Stx2 was examined.
56 Figure 1. Structure of Gb3 (top) and the O117 LPS (bottom). The hydroxy and N-acetyl groups have been shaded for emphasis.
57 Figure 2. Representation of the tailored biotinylated glycoconjugate and the structures of the five molecules. The blue ellipse is the carbohydrate recognition element, the biotinlyated scaffold and the spacer are colored in purple and red, respectively.
58
Materials and Methods
Bacterial Strains and Reagents - E. coli strains FI-29 and FI-4, isolated from healthy human volunteers, were described previously [122]. Serotyping was performed by the
Gastroenteric Disease Center at The Pennsylvania State University
(http://ecoli.cas.psu.edu E. coli type strains O107 and O117 were obtained from the
Gastroenteric Disease Center at The Pennsylvania State University. Typing sera were obtained from Statens Serum Institut (Copenhagen, Denmark). All strains were grown at
37°C on Luria-Bertani (LB) agar or in LB broth with shaking.
For binding studies, Stx1 and Stx2 were obtained from filter-sterilized culture
supernatants of C600::933W (Stx2) or C600::H19B (Stx1) induced with 30 μg/ml of ciprofloxacin as previously described [122].
Purification of LPS - E. coli were grown in 1 L of LB broth at 37°C overnight with
shaking. Cultures were centrifuged (7,500 x g, 20 min) and the hot phenol-water
protocol of Westphal and Jann [156] was used to extract LPS from the pellets. Briefly,
bacteria were suspended in 200 ml deionized water, homogenized (Kinematica
homogenizer, model PT 10/35 setting #4, Brinkmann Instruments, Inc., Westbury, NY),
and centrifuged (7,500 x g, 20 min). The pellet was suspended in 16 ml water,
homogenized, and heated to 65°C in a water bath. An equal volume of phenol at 65°C
was added to the suspension and homogenized, and the mixture was incubated on ice for
30 minutes, followed by centrifugation (7,500 x g, 1 hour). The aqueous phase was
59 collected, solid sodium acetate (to 1 mg/ml) and acetone (3 volumes) were added, and the
sample was incubated at 4°C overnight. The sample was centrifuged (7,500 x g, 20 min) and the pellet was suspended in 25 ml 80% acetone, followed by centrifugation (10,000 x g, 15 min). The pellet was suspended in 20 ml deionized water and centrifuged at
100,000 x g, 16 hours. The final pellet of purified LPS was suspended in 1 ml water and stored at -20°C.
The weight of the LPS was determined from dried samples, and the amount of
LPS in each preparation was also quantified using the purpald assay to detect
unsubstituted terminal vicinal glycols on 2-keto-3-deoxyoctonate and heptose in the LPS
inner core [157].
LPS ELISAs - Purified LPS from FI-29, FI-4, E. coli O107 and E. coli O117 were diluted
in coating buffer [1.59g Na2CO3 and 2.93g NaHCO3 per liter, pH9.6] and added to
microtiter plates (Pro-Bind, BD Biosciences, Bedford, MA) for a final LPS KDO
concentration of 38 nM/well and incubated at 37°C for 2 hours. Wells were washed with
phosphate buffered saline containing 0.5% Tween-20 (PBS-Tween) using an ELX405 microplate washer (Bio-Tek Instruments, Inc., Winooski, VT), blocked with 200 μl of
1% BSA and washed again with PBS-Tween. Duplicate wells for each LPS sample were
incubated with either wash buffer (negative control) or Stx2 from 933W culture
supernatant (15 μg per well) for 30 minutes at room temperature. Wells were washed
with PBS-Tween, incubated with anti-Stx2 monoclonal antibody 11E10 (1:100) at room
temperature for 1 hour, washed, and incubated with alkaline phosphatase-conjugated
secondary antibody (1:13,000) at room temperature for 1 hour. After washing the wells,
60 p-nitrophenyl phosphate (Sigma Fast™ tablets, Sigma, St. Louis, MO) was added and plates were incubated for 15 minutes in the dark at room temperature. The absorbance
(405 nm) was read using the ELX800 microplate reader (Bio-Tek Instruments, Inc.). For blocking the LPS-Stx2 interactions with specific antisera, LPS coated on the microtiter wells were incubated with antisera (1:100) for 1 hour at room temperature. FI-4 LPS was incubated with α-O25 antisera, FI-29 and E. coli O107 were incubated with α-O107 antisera, and E. coli O117 was incubated with α-O117 antisera. Wells were washed with
PBS-Tween, followed by incubation with Stx2 and detection of bound toxin as described above.
Glycoconjugate ELISAs - Binding of Stx1 and Stx2 to the synthetic glycoconjugates was assessed by ELISA analysis as described previously [124]. Toxin-containing culture supernatants, sterilized by filtration, were prepared from C600:H19B (Stx1) and
C600:933W (Stx2). The biotinylated compounds were diluted in phosphate-buffered saline (PBS) and added in excess to commercially available precoated streptavidin microwell plates with a binding capacity of around 125 pM per microwell. Binding was allowed to proceed at room temperature for 2 hours. The wells were washed three times with PBS and incubated with PBS (negative control) or Stx-containing culture supernatant at room temperature for 2 hours. The color was developed by using commercially available polyclonal antibody to Shiga toxin (Meridian Bioscience, Inc.,
Cincinnati, OH), alkaline phosphatase conjugated goat antirabbit IgG, and p-nitrophenyl phosphate (Meridian Bioscience, Inc. or Sigma, St. Louis, MO). The absorbance (
=405 nm) was read by an ELX800 microplate reader (Bio-Tek Instruments, Inc).
61 Results
Stx2 binding to LPS O-polysaccharide.
LPS ELISAs were used to evaluate Stx2 binding to LPS. ELISAs were
performed using hydrophobic microtiter plates to orient the LPS such that the Lipid A
portion would bind the plate and the O-polysaccharide would be exposed. The wells
were coated with LPS from FI-4, FI-29, E. coli O107 and E. coli O117 and incubated
with Stx2. Stx2 binding was detected using anti-Stx2 monoclonal antibodies. The LPS
from FI-29, E. coli O107, and E. coli O117 was able to bind Stx2, as evidenced by a
significant (P<0.005) increase in absorbance compared to wells to which no Stx2 was
added (Figure 3, compare gray and white bars). While E. coli O107 LPS did bind Stx2,
the binding was significantly (P<0.02) less compared to the binding of Stx2 to FI-29 or
E. coli O117 LPS (Figure 3). No increase in absorbance was observed for the toxin non-
neutralizer, FI-4.
The ability of the serotype-specific antibody to block toxin binding was assessed by incubating the LPS-coated wells with anti-LPS antibody prior to the addition of Stx2.
The increase in absorbance due to Stx2 binding to LPS was inhibited significantly
(P<0.05) for all three LPS types when the LPS was preincubated with the typing antisera
(Figure 3). For FI-29 and 107 LPS, the absorbance was reduced to background levels.
Stx2 binding to O117 LPS was also reduced, however the antibody did not completely
abrogate toxin binding (Figure 3). The O107 antiserum was not processed; however the
O117 antiserum was adsorbed with other E. coli strains to increase specificity. The
62 adsorption of the O117 antiserum with the O107 strain by the manufacturer may have removed some of the blocking activity.
63
0.20 * Control * m) +Stx2 0.15 * # ** +Ab +Stx2 ** 0.10 **
0.05
(405nAbsorbance 0.00 LPS Source: FI-4 FI-29 E. coli O107 E. coli O117
Blocking Ab: α-O25 α-O107 α-O107 α-O117
Figure 3. Antibodies specific to O107 and O117 block Stx2 binding to LPS. Microtiter wells coated with purified LPS (38 nM KDO/well) were incubated with wash buffer (control, no Stx2), Stx2 (15 μg/well), or type O-
specific blocking antibody (Ab as indicated) followed by Stx2. Stx2 binding was determined by using anti-Stx2 monoclonal antibody 11E10 and alkaline
phosphatase-conjugated anti-mouse secondary antibody. Results are the average of at least three trials. For each LPS source, an asterisk denotes a statistically higher (P < 0.005) level of Stx2 detected than that for control samples, and a double asterisk denotes a statistically lower (P < 0.05) level of Stx2 detected compared to wells with LPS and Stx2. Between LPS sources, a pound sign denotes a statistically lower (P < 0.02) level of Stx2 binding by E. coli O107 LPS compared to Stx2 binding by FI-29 and O117 LPS types.
64 Stx1 and Stx2 binding to the synthetic compounds.
Similar to observations for neutralizing LPS, Stx2 bound to the di- and mono-N- acetyl substituted galactosamine GC-1 and GC-3 a, respectively (Figure 4B), whereas
Stx1 failed to bind to either compound (Figure 4A). The trisaccharide GC-3 a appears to be a better substrate for Stx2 than the disaccharide GC-1, as the former exhibited enhanced toxin binding and was more sensitive (toxin binding was observed at lower concentrations of toxin). In contrast, GC-2 a, a trisaccharide analogue of Gb3, bound exclusively to Stx1 and not to Stx2. The limit of detection was 10 ng of toxin per microwell. We attribute the exclusive selectivity of GC-2 a towards Stx1 to the architecture of the biantennary analogue; possibly, the shorter spacer permits binding to
Stx1, but constrains binding to Stx2 [158]. Interestingly, increase in the spacer length leads to loss of sensitivity; GC-2 b, the Gb3 analogue with a 12-carbon-atom spacer bound to Stx1 with lower affinity than GC-2 a. The binding studies involving the N- acetylated analogue GC-3 b were very intriguing. Increase in spacer length from 6 (GC-
3 a) to 12 carbon atoms (GC-3 b) led to loss of selectivity and sensitivity; GC-3 b bound to both Stx2 and Stx1 with equal affinity. Clearly, the role of the spacer in the binding event remains to be determined.
65 A. Stx1 prefers Gal
1.4 GC-1 1.2 GC-3a GC-3b 1.0 GC-2a 0.8 GC-2b 0.6 0.4
Absorbance (405nm) 0.2 0.0 1000 100 10 ng Stx1 per well
B. Stx2 prefers GalNAc
1.4 GC-1 1.2 GC-3a 1.0 GC-3b GC-2a 0.8 GC-2b 0.6 0.4
Absorbance (405nm) 0.2 0.0 1000 100 10 ng Stx2 per well
Figure 4. Differential binding of A) Stx1 or B) Stx2 to synthetic glycoconjugates. Biotinylated glycoconjugates were added to streptavidin-coated microtiter wells and incubated with decreasing concentrations of Stx. Binding was determined by using anti-Stx polyclonal antibody and alkaline phosphatase conjugated goat anti-rabbit secondary antibody. Results are the average of three independent trials.
66 Discussion
FI-29, a human intestinal isolate of E. coli, was able to neutralize the cytotoxicity
of Stx2, not Stx1, on Vero cells. Characterization of this isolate revealed that LPS was
the neutralizing activity. Serotyping of the FI-29 strain indicated that it cross-reacted
with two previously identified LPS types O107 and O117. The structure of O117 had
been previously determined [125]. Both Gb3 and the O117 LPS contained terminal
galactose moieties, yet the O117 LPS has an N-acetyl modification on both of the
galactose subunits.
My studies have shown that the mechanism for the neutralization activity of FI-29
LPS is likely the ability of the LPS to bind to Stx2, preventing interaction with the intended cellular receptor. All three LPS types bound to Stx2 in an ELISA assay, yet the binding rate differed. In that FI-29 and O117 LPS had an increased binding to Stx2 as compared to O107, we can conclude that the structure of FI-29 LPS most likely resembles that of O117.
The significance of these studies is great as antibiotics are thought to be unsafe for
patients infected with E. coli O157:H7 [8]. Currently, probiotics are receiving much
attention as potential alternative therapies for gastrointestinal disease. Probiotics are live, non-pathogenic bacteria that are ingested with the intent that they colonize the gastrointestinal tract and subsequently create an environment that is unfavorable to pathogenic bacteria. Probiotics would outcompete pathogens for both attachment sites and nutrients [159]. Additional mechanisms of action include altering the pH of the local environment, initiating the immune response, or secreting antimicrobial peptides [159].
The acceptance of probiotics as a health benefit is increasing within the general
67 population due to products such as Activia© and DanActive©. It may be possible to
include FI-29 in these types of yogurt products.
The FI-29 LPS isolate has promise as a probiotic for the following reasons. First, it
was obtained from a healthy infant, the age group most at risk for developing severe
disease. Studies by S. Gamage using PCR confirmed the absence of the following
virulence factor genes of enterohemorrhagic E. coli (intimin (eae), Shiga toxin, and
virulence plasmid (pO157)), and uropathogenic E. coli (aerobactin, hemolysin, P fimbriae
(pap), S and F1C fimbriae, and cytotoxic necrotizing factor (CNF)).
Secondly, it binds to and neutralizes the cytotoxic activity of Stx2, the Stx variant
most commonly associated with the development of severe disease. Furthermore, FI-29
was found to be susceptible to commonly used antibiotics, including amikacin,
ampicillin, ampicillin/sulbactam, azetronam, cefazolin, cefepime, cefotetan, ceftazidime,
ceftriaxone, cefuroxime, cefuroxime axetil, ciprofloxacin, gentamicin, imipenem,
levofloxacin, meropenem, nitrofurantoin, piperacillin, piperacillin/tazobactam,
tobramycin, trimethoprim/sulfamethoxazole, amoxicillin, amoxicillin/clavulanic acid and
cefotaxime. This is important because the number of antibiotic resistant bacterial isolates
is rising. Establishment of FI-29 in the gut will not confer antibiotic resistance to either
the normal E. coli isolates present or any E. coli O157:H7 isolates it may encounter. As
with other bacteria, the antibiotic susceptibility will have to monitored regularly.
Lastly, it has been shown that commensal E. coli susceptible to phage infection
can amplify Stx2 production in the presence of Stx2-encoding phage from E. coli
O157:H7 both in vitro [122] and in vivo [108]. FI-29 is resistant to phage infection, and
therefore would not amplify Stx2 when encountering E. coli O157:H7.
68 The protective mechanism of action of the probiotic requires the FI-29 LPS to
bind to Stx2. In order to confer protection to the infected person, this binding interaction must occur in the gut in order to prevent Stx2 from circulating throughout the body.
Daily administration of the FI-29 probiotic is optimal as this would increase the likelihood that FI-29 would be present in the gut prior to E. coli O15:H7 infection.
Watery diarrhea is an early symptom of E. coli O157:H7 infection. As watery
diarrhea is a somewhat common gastrointestinal disruption, this symptom of E. coli
O157:H7 infection is often recognized without too much concern from the patient.
Medical treatment is usually sought upon the development of bloody, mucousy diarrhea.
Unfortunately, the blood present in the stool is indicative of intestinal epithelial
destruction [160] and Stx2 disruption of the kidney microvasculature [37]. Therefore,
patients who present symptoms of severe disease will not benefit from the administration
of the FI-29 probiotic because the Stx2 has already exited the gut and exerted systemic effects.
Realistically speaking, persons with relatively low risk for becoming infected
with E. coli O157:H7 will most likely be unwilling to take this probiotic. Conversely, those people with an immediate risk of infection, such as the asymptomatic family members of infected patients, especially young siblings, would likely be more willing to take the FI-29 probiotic.
The synthetic glycoconjugates have significant diagnostic potential. Currently, E.
coli O157:H7 infection is confirmed using an ELISA that utilizes capture antibodies that
recognize both Stx1 and Stx2. Providing hospital laboratories with a test that can
determine if a patient is infected with an E. coli O157:H7 strain that produces either Stx1
69 or Stx2 will result in a clearer prognosis for the patient. Unlike antibodies, preparation of the glycoconjugates is inexpensive and does not require the use of animals, the recognition capabilities are consistent between lots, and the glycoconjugates can be stored at 4º to -20ºC without consequence. In addition, no additional training for the technologists would be required, increasing the acceptance of the new assay in the laboratories.
In conclusion, these studies have used ELISA assays to assist in larger projects that may result in both new diagnostics and therapeutics for E. coli O157:H7 and other
Stx-releated disease.
70 Chapter III.
Antibiotic treatment to prevent the systemic complications of Escherichia coli O157:H7 infection.
71 Abstract
Antibiotics such as ciprofloxacin and trimethoprim/sulfamethoxazole have been
shown to increase Shiga toxin production by Escherichia coli O157:H7. In this study, we
have tested a panel of antibiotics to determine under which conditions, if any, antibiotics
are safe for use in the treatment of E. coli O157:H7. We have created a novel luciferase-
expressing Vero cell line that can be used to monitor Shiga toxin-mediated reductions in
protein synthesis. In the antibiotic studies, subinhibitory concentrations of ciprofloxacin
and trimethoprim/sulfamethoxazole increased toxin production in clinical E. coli
O157:H7 isolates. Using the ELISA described in Chapter 2, it was shown that although
the test strain had both stx1 and stx2, these two antibiotics induced only Shiga toxin 2.
While ampicillin, ceftriaxone, rifampicin, and gentamicin largely left Shiga toxin levels
unaffected compared to controls, subinhibitory concentrations of fosfomicin,
azithromicin, and doxycyclin reduced Shiga toxin production. Nonpathogenic E. coli
susceptible to Stx-encoding phage infection have been shown to amplify Shiga toxin
production. In a coculture of E. coli O157:H7 and a commensal Stx-amplifying E. coli
strain resistant to the tested antibiotics, gentamicin, amplicillin, an doxycycline, Stx
production remained significantly high. In the presence of an antibiotic sensitive
amplifier, ciprofloxacin increased Shiga toxin production as expected. Most interesting
was the ability of azithromicin to significantly reduce Shiga toxin in the presence of the
amplifier to levels below those produced by E. coli O157:H7 alone. These data indicate that current protocol of withholding antibiotic treatment during E. coli O157:H7 infection
should be reevaluated to include both the mechanism targeted by the antibiotic and the
72 composition of the patient flora as antibiotics that inhibit translation may be classified as potential therapeutics for those patients who do not harbor Shiga toxin-amplifying strains of E. coli resistant to the antibiotic.
73 Author Summary
Escherichia coli O157:H7 is a food-borne pathogen that is the leading cause of hemolytic uremic syndrome in children. Currently, a therapeutic remains to be developed. A major virulence factor, termed Shiga toxin, is linked to severe disease.
Bacterial stress, for instance exposure to antibiotics such as ciprofloxacin or trimethoprim/sulfamethoxazole, induces the lytic cycle at which time Shiga toxin is transcribed, translated, and released from the bacterial cell along with new phage.
Nonpathogenic strains of E. coli susceptible to phage infection have been shown to amplify Shiga toxin production. Persons harboring these amplifying strains may be at a higher risk for severe disease development. In this study, we have used a novel, luciferase-based cellular assay to test a panel of antibiotics to determine which conditions, if any, antibiotics are safe for use in the treatment of E. coli O157:H7.
Translation-inhibiting antibiotics reduced Shiga toxin production in E. coli O157 incubated either alone or with a nonpathogenic amplifier strain sensitive to the antibiotic.
In contrast, Shiga toxin production remained largely unaffected in the presence of a nonpathogenic, resistant amplifier. Therefore, the effectiveness of antibiotic treatment may vary from patient to patient based on the composition of their normal flora.
74 Introduction
The foodborne pathogen, Escherichia coli O157:H7, the causative agent of
hemorrhagic colitis [7] and hemolytic uremic syndrome (HUS) [6,34], is responsible for
about 73,000 cases of disease annually within the United States [161]. Shiga toxin is necessary for the development of HUS. Two antigenically distinct versions of the Stx exist, termed Stx1 and Stx2 [31]. These toxins share 60% amino acid identity. While strains of E. coli O157:H7 can produce Stx1, Stx2 or both, epidemiological studies [162] suggest that Stx2 is linked to the development of severe disease. Furthermore, primates injected with purified Stx2 developed symptoms of HUS, while animals receiving a similar dose of Stx1 did not develop disease [163]. These results suggest that Stx2 is sufficient to produce the symptoms of HUS.
Stx is an AB5 toxin. Five identical B, or binding, subunits bind to
globotriasylceromide (Gb3) to gain entry into host cells. After internalization and
retrograde transport to the cytoplasm, the enzymatically active A subunit inhibits protein
synthesis by cleaving a ribosomal adenine residue [99]. However, the systemic
complications from Stx2 intoxication aren’t due to protein synthesis inhibition per se, but
rather are due to the activation of distress pathways, different from but analogous to
septic shock [164]. Children with high levels of plasminogen activator inhibitor 1 (PAI-
1) have a high risk of developing HUS. PAI-1 inhibits the lysis of the blood-clotting
protein fibrin, resulting in the accumulation of fibrin deposits within the microvasculature
blocking the blood flow [35].
The genetic regulation of Stx is unusual and has implications for the management
of patients infected with Stx-producing E. coli. The genes for both Stx1 and Stx2 are
75 encoded within the late region of lysogenic phage integrated in the bacterial chromosome
[51,165]. The Stx genes are silent during lysogeny, but are expressed during the lytic cycle, resulting in the production and release of both Stx and phage [166]. The Stx phage released by E. coli O157:H7 during lytic infection can recruit other strains of E. coli to produce Shiga toxin. The presence of susceptible E. coli can amplify the amount of Stx produced by up to 1000 times in vitro [122] and in vivo [108], suggesting that the composition of an individual’s intestinal flora can influence the amount of Shiga toxin produced during infection, and ultimately the course of disease.
Antibiotic treatment can also influence Shiga toxin production and the risk of developing severe disease [107,167]. The switch from the lysogeny to the lytic phase is controlled by bacterial SOS stress response. Interestingly quinolone antibiotics, which are commonly prescribed for infectious diarrhea, have been shown to induce the SOS stress response, resulting in increased production of Stx [107], and it has been suggested that antibiotics could promote the progression to severe disease in patients infected with
E. coli O157:H7. While some epidemiological studies suggest that antibiotic treatment has been successful in resolving infection [168-170], other studies have linked antibiotic administration to the development of HUS [167].
The influence of antibiotics on Stx expression in vitro has been examined in several studies. Ciprofloxacin, trimethoprim/sulfamethoxazole, and ampicillin have been reported to increase Stx production [54,106,171-173]. Rifampicin has been shown to reduce toxin [56,171]. Depending on the study, fosfomicin, gentamicin, doxycycline, and azithromicin have been reported either increase [106,171,174]or decrease [56,171] Stx production. Other studies have reported differences between Stx1 and Stx2 expression.
76 Ciprofloxacin, ampicillin, ceftriaxone, and doxycycline induced Stx2, not Stx1 [106,174],
while fosfomicin, gentamicin, and doxycycline induced Stx1 and not Stx2. In all but two
studies [106,114], levels of Stx expression were determined antigenically (ELISA or
Western blot), or using transcriptional reporters such as β-galactosidase. These indirect
assays may not accurately reflect how much biologically active toxin is being made since
toxin protein may be present but not assembled in to the functional AB5 complex.
In this study, we characterized the ability of a panel of antibiotics commonly
used to treat infectious diarrhea to influence Stx production by clinical isolates of E. coli
O157:H7. Levels of Stx secretion by bacteria incubated in the presence of sub-inhibitory
levels of antibiotic was quantified by measuring luciferase expression in Vero cells
engineered to express a destabilized version of luciferase. Ciprofloxacin and
trimethoprim/sulfamethoxazole were shown to induce Stx production. However some
antibiotics prevented bacterial growth without inducing production of Stx (ampicillin,
ceftriaxone, gentamicin, and rifampicin), while other antibiotics decreased production of
Stx at sublethal doses (fosfomicin, azithromicin, and doxycycline). These antibiotics
show promise for treatment of E. coli O157:H7.
We have also studied the effect of natural flora on the effectiveness of the antibiotics. A phage susceptible commensal isolate known to amplify toxin was incubated with E. coli O157:H7 in the presence of antibiotic. Azithromycin reduced Stx levels beneath control levels when both strains were sensitive to the antibiotic. In contrast, gentamicin, doxycycline, and ampcillin were unable to significantly reduce Stx production if the commensal isolate was resistant to the antibiotic. This suggests that the
77 normal flora may influence Stx production, and that antibiotic treatments may only be successful in those patients whose flora are susceptible to the antibiotic.
78 Materials and Methods
Bacterial strains and reagents - Shiga toxin producing strains are described in Table 1. E. coli strains were propagated at 37°C in Mueller Hinton (MH) or Luria Bertani (LB) broth or agar. In studies examining phage infection of susceptible commensal E. coli, modified
LB agar containing 10 mM CaCl2 was used. Fosfomicin was a kind gift from Dr. Mineo
Watanabe. Ciprofloxacin was purchased from Serologicals, trimethoprim, sulfamethoxazole, ampicillin, ceftriaxone, gentamicin, doxycycline, azithromicin, and rifampicin were purchased from Sigma.
Creation of a stable luciferase-expressing Vero cell line - The luciferase gene was excised from the pGL4.11[luc2p] vector (Promega) with HindIII and XbaI, blunted using
Klenow and cloned into the HpaI site of the retroviral vector MigR1 [175]. The HEK
GP2-293 packaging strain (Clontech) was incubated with 3 µg Luc2p-MigR1 DNA and 1
µg pVSV-G envelope (a kind gift from Dr. William Miller), and virus was collected as described [176]. The virus-containing supernatant was added to 2x105 Vero cells in 1.0ml of Eagle’s media (Sigma) in a 6-well tissue culture plate. The plate was centrifuged at room temperature for 20 mins at 470 X g then incubated at 37ºC with 5% CO2 overnight.
GFP-expressing Luc2p-positive cells were sorted by flow cytometry (FACS Vantage
Sorter, 90µm nozzle, Becton Dickinson) at the Cell Sorting Core Facility at Cincinnati
Children's Hospital. GFP positive cells were propagated, and sorted a second time to obtain cells expressing 10-fold more GFP over the average population. These cells have been deposited with the Biodefense and Emerging Infections Research Resources
79 Repository (BEI Resources) which was established by the National Institute of Allergy
and Infectious Disease (NIAID) http://www.bioresources.org/.
Dose response curve of antibiotics - The minimum inhibitory concentration (MIC) for
different doses of antibiotics was determined using the macrobroth dilution assay
following standard protocols [177]. Briefly, Shiga toxin producing E. coli were grown to
stationary phase overnight in MH broth, diluted 1:1000, then 1:200. For log phase studies, stationary phase bacteria were diluted 1:1000, grown to early log phase (OD600nm
approximately 0.04) and diluted as described above. Two-fold serial dilutions of
antibiotic were made in 1.0 ml of MH broth in test tubes. Trimethoprim and
sulfamethoxazole, were used at the ratio administered to patients (1:5). One ml of bacteria was added to the antibiotic dilutions and incubated at 37ºC shaking overnight.
The MIC was visually determined. To quantify Stx procuction, the overnight cultures were centrifuged at 2000 X g for 5 minutes, and the supernatant was filter sterilized using
0.22 μM SPIN-X filters (Costar, Cambridge, Mass).
Luciferase assay to assess Stx-mediated inhibition of protein synthesis - The effective
dose of toxin needed to inhibit protein synthesis by 50% (ED50) was determined using the
luciferase expressing Vero cells. Toxin-containing culture supernatants were diluted
1:100 followed by half-log serial dilutions in PBS. Samples (25μl) were placed in white,
opaque, Falcon 96-well microtiter plates (Becton Dickinson, Franklin Lakes, NJ). 100 μl
of 105 luciferase-expressing Vero cells were added to the toxin-containing wells and
incubated for 4 hours at 37ºC with 5% CO2. The media was removed and the cells were
80 washed three times with PBS using the ELX405 microplate washer (Bio-Tek Instruments,
Inc., Winooski, Vt.). 25-100μl Superlite Luciferase Substrate (Bioassay Systems,
Hayward, California) was injected into each well, and light was measured using the
Luminoskan Ascent (Thermo Labsystems, Helsinki, Finland). Filter-sterilized culture
supernatant from C600::933W containing 258 µg/ml of Stx2 or C600::H19B containing
108 μg/ml of Stx1 as determined by ELISA were used as standards [124]. The point at
which protein synthesis was reduced by 50%, or ED50, was determined. To determine
relative differences in Stx production the two points within the sample dilution series that
fell above and below the ED50 were used to determine the x intercept. The x intercept
values of samples incubated in the presence of antibiotic was compared to control
cultures grown without antibiotic.
Western blot to determine the amount of unassembled Stx subunits - Stx2-containing
supernatants (64 ng) were loaded onto a 8-16% Tris-glycine precast gel (Cambrex
Bioscience Rockland Inc., Rockland, Maine) and separated by SDS-PAGE. Gels were
run in Tris-glycine buffer (25 mM Tris, 190 mM glycine, 0.1% SDS) by using the Mini-
Protean II system (Bio-Rad Laboratories). Stx protein was transferred to polyvinylidene difluoride (PVDF) membranes using the wet transfer apparatus (Hoeger Scientific
Instruments, San Fransicso, CA). Membrane was blocked in 5% nonfat dry milk and
washed in buffer (125 mM Na2HPO4, 25 mM NaH2PO4, 100 mM NaCl, 0.5% Tween 20,
0.25% nonfat dry milk). Membranes were incubated for overnight in polyclonal anti-Stx2
antibody (1:12,000) (Meridian), washed, then incubated with goat anti-rabbit antibody
conjugate to horseradish peroxidase (Cappel).
81
ELISA assay to determine if E. coli O157:H7 PT-32 produces Stx1 and/or Stx2 in the
presence of Stx-inducing antibiotics - Biotinylated Stx receptor mimics were diluted in water and added to streptavidin-coated, blocked, microtiter plates with a binding capacity of ~125 pmol biotin per well (Reacti-Bind; PIERCE, Rockford, IL) and incubated at room temperature for 2 h as previously described [178]. Wells were washed in triplicate
with PBS by using an ELX405 microplate washer (Bio-Tek Instruments, Inc., Winooski,
VT) then blocked with 0.1% BSA and 0.01% Tween in PBS. Compounds were
incubated with either PBS (negative control), 1μg Stx1 from H19B culture supernatant
(Stx1 standard), 1μg Stx2 from 933W culture supernatant (Stx2 standard), or undilute
sample culture supernatant at room temperature for 1 h. Wells were washed with PBS,
incubated with anti-Stx1 or anti-Stx2 polyclonal antibody (Meridian Bioscience, Inc.,
Cincinnati, OH) at room temperature for 30 mins, washed, and incubated with alkaline-
phosphatase conjugated secondary antibody for 30 mins. After washing, p-nitrophenyl
phoshate (Fast tablets; Sigma, St. Louis, MO) was added and the plate was incubated in
the dark for 20 mins. The absorbance was read by using an ELX800 microplate reader
(Bio-Tek Instruments, Inc.)
Coincubation experiments – Human intestinal E. coli isolate ECOR-4, previously shown
to be susceptible to lytic infection with the Stx2 phage of PT-32 [108], was transformed
with plasmids to generate various antibiotic-resistant derivatives as indicated in Table 1.
Strain 185, an antibiotic resistant derivative of PT-32 and appropriate ECOR-4
derivatives were incubated overnight in LB broth, harvested by centrifugation, and
82 adjusted to an OD600 of 1. Prior to adjustment, the pellet of the PT-32 derivative was washed with LB to remove residual phage and Stx from the culture. 70μl of strain 185 and 7 mls of phage-susceptible strain (either CMUC171 or CMUC172) were added to test tubes containing 1 ml of antibiotic in LB broth. Cultures were overlaid onto LB- modified agar in deep Petri dishes, and incubated overnight at 37ºC as a static culture.
Bacterial CFUs for both strains were determined by plating serial dilutions on selective media. Filter-sterilized supernatants to assess Stx production were prepared as described above.
83
TABLE 1. Plasmids and strains used in this study.
Strain or plasmid Relevant characteristicsa Reference or Source E. coli strains C600::933W E. coli strain C600 lysogenized with Stx2-converting [179] phage C600::H19B E. coli strain C600 lysogenized with Stx1-converting [179] phage
PT-32 Clinical isolate of E. coli O157:H7; stx1, stx2 [122]
185 Streptomycin resistant mutant of PT-32 transformed with pBBR1MCS-2 Kanr
PT-40 Clinical isolate of E. coli O157:H7; stx2 [122]
CMUC-170 E. coli intestinal isolate ECOR-4 susceptible to Stx This study phage, transformed with pUW2138, Gentr This study CMUC-172 E. coli intestinal isolate ECOR-4 susceptible to Stx phage, transformed with pBBR1MCS-3, AmprTetr
Plasmids pBBR1MCS-2 Kanr [180] pBBR1MCS-3 AmprTetr [180] pGL4.11[luc2p] Luciferase gene tagged with PEST degradation sequence Promega vector pMig-R1 Retroviral vector, multiple cloning site, IRES, GFP [175] pUW2138 Gentr, OriT in pBluescript SK+ [181] a Strr, streptomycin resistant; Gentr, gentamicin resistant, AmprTetr; ampicillin and tetracycline resistant, and Kanr, kanamycin resistant.
84 Results
Generation of Vero cells that stably express luciferase.
We developed an improved assay to assess Stx-mediated inhibition of protein
synthesis based on the method of Zhao et al [154], using Vero cells engineered to stably
express the Luc2p variant of luciferase. Luc2p is tagged with a PEST (proline- glutamic
acid-serine-threonine) sequence, which promotes rapid turnover by targeting the protein
to the proteosome for degradation. All measurable luciferase activity is due to newly
synthesized protein, and luciferase levels immediately decrease when protein synthesis is
inhibited. The MigR1 retrovirus [175] was used to introduce the Luc2p gene into the
Vero cell chromosome. Luc2p was cloned upstream of the GFP gene (Fig. 1A). GFP and luciferase expression are transcriptionally coupled, and Vero cells tranfected with
MigR1-Luc2p were sorted for those expressing 10x higher levels of GFP compared to the
average population (Fig. 1B).
To determine the optimal number of Vero cells for protein synthesis inhibition assays, serial dilutions of Vero cells were plated, harvested at two-hour intervals, and luciferase levels were was measured (Fig. 1C). At low densities (102-103 cells per well)
luciferase production was equivalent to background. At high densities (105 cells per
well), luciferase expression decreased with time in the absence of toxin. Decreased
luciferase production was associated with a concomitant change in pH indicator,
suggesting the cells were plated too densely and were acidifying the medium. Cell plated at 104 produced high levels of luciferase which remained constant between 4 and 8 hours in culture. Cells were used at this concentration in all subsequent studies. To determine
85 the earliest time point at which Stx-mediated protein synthesis inhibition could be
measured, the luciferase activity of Vero cells incubated with 0.6ng of Stx1 or Stx2 was compared to control cells (Fig. 1D). The response to Stx1 and Stx2 was similar at 4 hours, after which time Stx1 caused a significantly greater reduction in protein synthesis than Stx2. Protein synthesis was evaluated after 4 hours in all subsequent experiments.
Limiting dilution assays were used to quantify Shiga toxin production (Fig. 2A).
In this example, untreated Vero cells displayed about 6 relative light units (RLU) of
luciferase activity (open triangles), while cells incubated with serial dilutions of a standard containing 258μg/ml of Stx2 exhibited a concentration dependent decrease in
luciferase expression (open diamonds). The ED50 (or concentration of toxin which caused
50% inhibition of protein synthesis) for the Stx2 standard was determined to be
0.6ng/well. Culture supernatant from E. coli O157:H7 grown in the absence of
antibiotics (closed squares) displayed less activity (a leftward shift) compared to the Stx2
standard, and the ED50 was determined to be 34-fold less than the Stx2 standard or 7.6
µg/ml of Stx2. The impact of trimethoprim/sulfamethoxazole on Stx production was also
assessed. E. coli O157:H7 incubated with enough trimethoprim/sulfamethoxazole to
inhibit growth (2 times the MIC), did not produce detectable level of Shiga toxin (stars).
In contrast, culture supernatant from E. coli O157:H7 incubated a subinhibitory dose of
trimethoprim/sulfamethoxazole (one fourth the MIC), produced more Shiga toxin than
the untreated control E. coli (closed circles), an amount equivalent to that seen in the Stx2
standard. The limiting dilution assay was used to quantify Stx production in all
subsequent studies.
86
15 A C 2 hours 4 hours 6 hours 8 hours LTR Luc2p IRES GFP LTR 10
5 Relative Light Units Relative
0 105 104 103 B D Cells per well 100 Stx1 80 Stx2
60
40 **
20
Percent protein synthesis 0 102 103 104 105 2 4 6 8 GFP Hours
Figure 1. The Luc2p-MigR1 retroviral vector. A) The Luc2p gene was cloned into the multiple cloning regions located upstream of the internal ribosomal entry site (IRES) and the enhanced green florescent protein (GFP) gene. B) Transfected Vero cells were analyzed using FACS and cells expressing 10x more GFP than the average population (bottom right quadrant) were collected. C) Luciferase activity was assessed as a function of time and concentration. D) Vero cells were incubated with 0.6 ng of Stx1 or Stx2 and luciferase activity was assessed as a function of time. Statistical significance between Stx1 and Stx2 is indicated by an asterisk (P<0.05) according to the Student’s t test. The results are the average of three trials.
87
A B 7 Vero cells 6 Stx2 Std TMP-SMX 5 O157, Ctrl 4 O157, 2x MIC O157, 0.25x MIC kD 3 No antibiotic ½MIC ¼MIC 2 37
Relative Light Units Light Relative 1 34x 0 26 Cells 64 6.4 0.64 0.06 ng/well Stx2 Std 11x 28x alone 102 103 104 105 Sample dilution Fold increase over control
Figure 2. Quantifying Shiga toxin production. A) Luciferase-containing Vero cells were incubated with serial dilutions of culture supernatant from Stx-producing cells or in the absence of Stx for 4 hours at 37°C. Luciferase substrate was added, and relative light units (RLU) were measured. The Stx2 standard contains 258 μg/ml of Stx2. The average for 6 wells of untreated Vero cells was determine to be 5.8 RLU, and the level of 50% inhibition or ED50 (2.9 RLU) is indicated by the horizontally dashed line. The two points within the dilutions series that fell on either side of the ED50 were used to determine the x intercept value (vertically dashed lines) which was then used to compare fold difference in toxicity of each sample (arrow). B) E. coli O157:H7 strain PT-40 was grown in the presence or absence (control) of trimethoprim/ sulfamethoxazole as indicated, and equivalent amounts of toxin (64ng/well) based on the Luc2p Vero cell toxicity assay was loaded onto a 8-16% Tris-glycine gel, and blotted with polyclonal antibody to Stx2.
88 Presence of unassembled Stx subunits.
Previous studies have used antigenic assays such as ELISAs to determine the
influence of antibiotics on Stx production [104,182]. However, the presence or
unassembled, inactive toxin subunits could give false positive results. Western blot
analysis was used to assess the presence of unassembled Stx2 (Fig 2B). Filter-sterilized
supernatants from PT-40 incubated without trimethoprim/sulfamethoxazole (control) or
with ½ or ¼ the MIC were loaded at 64ug/well Stx based on the Vero cell assay. The
amount of antigenic A subunit was highly variable. Compared to the control, the Stx2 A
subunit was increased 11-fold at ½ the MIC and 28-fold at ¼ the MIC. Values were
determined using ImageQuant.
The effect of antibiotics on Shiga toxin production.
MIC values for various antibiotics, including ones commonly prescribed for
infectious diarrhea, were determined for both stationary and log phase bacteria (Table 2).
The strains used in this study were all susceptible to the antibiotics. The standard broth
dilution method uses stationary phase bacteria, and the MIC values were comparable to
previously reported values [106,111,169,183-186]. During infection bacteria are more
likely to be actively growing and we also tested bacteria in logarithmic phase; reduced
MICs were observed for ampicillin and rifapmicin compared to stationary phase bacteria.
89 TABLE 2. The Minimum Inhibitory Concentrations (MIC) of antibiotics.
MIC μg/ml
Antibiotic (Target) *Reported C600::933W (Stx2) C600::H19B (Stx1) PT-32 (Stx1 and Stx2) PT-40 (Stx2) values Stationary Log Stationary Log Stationary Log Stationary Log
Ciprofloxacin (DNA) 0.01-8 0.01-0.02 0.01-0.03 0.01 0.01 0.02 0.02 0.02 0.02
Trimeth/Sulfa (M) 0.25-16 0.03-0.06 0.01-0.03 0.03 0.004-0.02 0.13 0.13-0.25 0.13 0.13-0.25
Ampicillin (CW) 1-64 8-16 2-8 8-16 4-8 4-8 2-8 8-16 2-4
Ceftriaxone (CW) 0.1-2 0.03-0.06 0.03-0.06 0.06-0.13 0.03 0.03-0.06 0.03-0.06 0.06 0.06
Fosfomicin (CW) 12.5-32 2 2-4 0.5-2 2 16 16-64 16-32 16-32
Rifampicin (TS) NR 8-32 8-32 16-32 4-8 2-8 4-8 16 4-16
Gentamicin (TL) 0.1-16 0.5-1 0.25-1 0.5-1 0.5 0.25-0.5 0.5 0.5-1 0.25-1
Doxycycline (TL) NR 1 0.5-1 0.5-1 0.25-0.5 0.5-1 1-2 1-2 1-2
Azithromicin (TL) 2 2-4 2-8 2-16 2-8 2.6 2.6-4 2.6-5.2 4
aTrimethoprim/Sulfamethoxazole were used at a 1:5 ratio; only the concentration of trimethoprim is indicated. Antibiotic target abbreviations:
Deoxyribose nucleic acid (DNA), metabolism (M), cell wall (CW), transcription (TS), and translation (TS). *References as follows. Cipro:
[19,22,36-38], trimeth/sulfa:[19,34,35,38], amp: [22,34,35,37], cef: [22], fos:[22,36,37], gent: [19,22,35,38], azith:{Gordillo, 1993 #1279}.
90 Ciprofloxacin and trimethoprim/sulfamethoxazole cause a dose-dependent increase in Stx
expression.
Previous studies have shown that ciprofloxacin induces Stx production in vitro
[54] and in vivo [107,108]. To verify this result, E. coli O157:H7 strain PT-32 was incubated with serial dilutions of ciprofloxacin or trimethoprim/sulfamethoxazole at 37ºC overnight and the fold increase in Stx production compared to control cells incubated in the absence of antibiotic was determined (Figure 3). As expected, no Stx was produced when bacterial growth was inhibited by concentrations of antibiotic at or above the MIC
(data not shown). Increased expression of Stx was seen when the bacteria grown in sub- inhibitory concentrations of antibiotic. A dose-response relationship was generally observed, and the most Stx was produced by bacteria grown in highest concentration of antibiotic that supported bacterial growth.
The levels of production of Stx by different Stx-encoding phage can vary [187]
and the influence of ciprofloxacin on Stx production was tested in other strains, as well as
bacteria in stationary or logarithmic phase (Figure 4A). The strains were incubated with
ciprofloxacin at one half the MIC. Laboratory strain, C600 lysogenized with the Stx2-
producing phage, 933W, or the Stx1 producing phage, H19B, was not reproducibly induced to produce Stx in the presence of ciprofloxacin. In contrast, two clinical isolates of E. coli O157:H7, one encoding only Stx2, and one encoding both Stx1 and Stx2, were induced to express significantly high levels of Stx in the presence of ciprofloxacin,
Trimethoprim/sulfamethoxazole also caused a strain-dependent increase in Stx production (Fig. 4B). Stx production was significantly increased by both PT-40 and PT-
91
1000 CIP * * TMP-SMX 100 * *
10 * *
1 compared tocontrol Fold difference in toxicity in difference Fold -10 1/2 1/4 1/8 1/16 Concentration of antibiotic compared to the MIC
Figure 3. The effect of ciprofloxacin and trimethoprim/sulfamethoxazole on the production of Stx. E. coli O157:H7 strain PT-32 was incubated with concentrations of ciprofloxacin or trimethoprim/sulfamethoxazole at 1/2, 1/4, 1/8, and 1/16 the MIC, and toxicity of the antibiotic-treated cells was compared to cells incubated in the absence of antibiotic as described in Figure 2. Symbols represent individual trials and the geometric mean is plotted. Statistical significance compared to the no antibiotic control is indicated by an asterisk (P<0.05) according to the Student’s t-test.
92
A Ciprofloxacin 1000 ** 100 ** 10 1 -10 erence compared to control
ff -100 -1000 Stationary Log Fold di Fold -10000 933W PT-40 PT-32 H19B (Stx2) (Stx2) (Stx1, Stx2) (Stx1) Strain
B Trimethoprim/sulfamethoxazole 1000 * Stationary 100 Log 10 * 1 -10 * to control -100 -1000
Fold difference compared difference Fold -10000 933W PT-40 PT-32 H19B (Stx2) (Stx2) (Stx1, Stx2) (Stx1) Strain
Figure 4. Antibiotics that induce production of Stx. E. coli strains were incubated with serial dilutions of ciprofloxacin or trimethoprim/sulfamethoxazole overnight and Stx was determined. The fold difference in toxin production compared to the control is plotted for sample incubated in antibiotic at ½ the MIC. Symbols represent individual trials and the geometric mean is plotted. Statistical significance compared to the no antibiotic control is indicated by an asterisk (P< 0.05), according to the Student’s t test. C600::933W, PT-32, and PT-40 control strains produced 8-27 ug/ml Stx2 in both stationary and log phase while C600::H19B produced 32-58 ug/ml Stx1.
93 32 in log and stationary growth phases respectively when bacteria were incubated with antibiotic at one half the MIC. Induction by trimethoprim/sulfamethoxazole was not observed for the other strains. The growth state of the bacteria did not appear to influence Stx production.
Influence of other antibiotics on Stx production.
The ability of other antibiotics to influence Stx production was also tested. In contrast to ciprofloxacin and trimethoprim/sulfamethoxazole, under some conditions azithromicin, doxycyline, fosfomicin, gentamicin and ampicillin caused a statistically significant decrease Stx production for bacteria treated at one half the MIC (Figure 5 A-
E). Ampicillin caused a modest, but statistically significant (1.5-fold) increase in Stx2 production for PT-40 in logarithmic phase (Figure 5E). Sub-inhibitory concentrations of ceftriaxone, and rifampicin did not affect Stx production (Figure 5 F-G).
Stx2 is expressed by PT-32.
Strain PT-32 has the genes for both Stx1 and Stx2. An ELISA using receptor mimics that capture either Stx1or Stx2 [178] was used to determine if Stx1 and/or Stx2 was produced by strain PT-32. The Stx1 and Stx2 standards added at 1μg/well bound to the respective compounds. Toxin levels in culture supernatant for bacteria grown in the absence of antibiotic were below the limit of detection (Figure 6, Control). In addition,
Stx1 toxin levels were below the limit of detection for bacteria incubated with ciprofloxacin or trimethoprim/sulfamethoxazole (Fig 6A). However, an increase in Stx2 production was seen for PT-32 incubated in the presence of either antibiotic (Figure 6B).
94 A Azithromicin E Ampicillin 1000 1000 Stationary Stationary 100 100 Log Log 10 10 # 1 1 * # -10 * * -10 to control * to control * -100 * -100 -1000 *# -1000 Fold difference difference compared Fold -10000 differenceFold compared -10000 933W PT-40 PT-32 H19B 933W PT-40 PT-32 H19B (Stx2) (Stx2) (Stx1, Stx2) (Stx1) (Stx2) (Stx2) (Stx1, Stx2) (Stx1) Strain Strain B Doxycyline F Ceftriaxone 1000 1000 Stationary Stationary 100 Log 100 Log 10 10 # 1 # 1 * # # -10 * -10 to control to to control -100 ** -100 -1000 -1000 Fold differenceFold compared Fold differenceFold compared -10000 -10000 933W PT-40 PT-32 H19B 933W PT-40 PT-32 H19B (Stx2) (Stx2) (Stx1, Stx2) (Stx1) (Stx2) (Stx2) (Stx1, Stx2) (Stx1) Strain Strain C Fosfomicin G Rifampicin 1000 1000 Stationary Stationary 100 Log 100 Log 10 10
1 # 1 -10 * -10 to control -100 * to control -100 -1000 -1000 # Fold differenceFold compared -10000 differenceFold compared -10000 933W PT-40 PT-32 H19B 933W PT-40 PT-32 H19B (Stx2) (Stx2) (Stx1, Stx2) (Stx1) (Stx2) (Stx2) (Stx1, Stx2) (Stx1) Strain Strain Figure 5. Determining the effect of antibiotics on the D Gentamicin production of Stx. E. coli strains were incubated with 1000 serial dilutions of A) azithromycin, B) doxycycline, C) Stationary 100 fosfomicin, D) gentamicin, E) ampicillin, F) ceftriaxone, Log or G) rifampicin overnight. Toxicity of the filter-sterilized 10 sample supernatants was measured using the 1 luciferase-containing Vero cells. The fold difference in -10 toxin production compared to the control is plotted for
to control to the sample containing antibiotic concentrations at ½ the -100 # * MIC. Statistical significance is indicated by an asterisk -1000 (P<0.05) according to the Student’s t test. Symbols
Fold differenceFold compared -10000 represent individual trials and the geometric mean is 933W PT-40 PT-32 H19B plotted. A pound sign (#) indicates that one sample in (Stx2) (Stx2) (Stx1, Stx2) (Stx1) the series was below the limit of detection (> 2.4 ng/ml). Strain
95
A. Stx1 (Gb3) 2.5
2.0
1.5
1.0
0.5 Absorbance (405nm) 0.0 Stx1 No Control CIP TMP-SMX 1μg Stx1 E. coli O157:H7
B. Stx2 (N-Acetyl Gb3)
2.5
2.0
1.5
1.0
0.5 Absorbance (405nm) 0.0 Stx2 No Control CIP TMP-SMX 1μg Stx2 E. coli O157:H7
Figure 6. Determining if antibiotics induce Stx1, Stx2, or both in E. coli O157:H7 strain PT-32(stx1, stx2). E. coli O157:H7 strain PT-32 was grown in the presence or absence (control) of ciprofloxacin (CIP) or trimethoprim/sulfamethoxazole (TMP-SMX). Undiluted filter-sterilized supernatants at ½ the MIC were incubated with Stx1 or Stx2-specific compounds and developed with anti-Stx1 or Stx2 polyclonal antibodies.
96
Stx production during co-culture with phage-sensitive Stx2-amplifying strains.
Previous studies have shown that Stx-encoding phage can infect commensal E.
coli, leading to enhanced Shiga toxin production both in vitro[122] and in vivo [108]. In this study, we wanted to assess ability of antibiotics to induce phage production in antibiotic sensitive E. coli O157:H7, and the subsequent ability of the phage to induce Stx production by phage-sensitive, antibiotic-sensitive normal intestinal flora (Figure 7), or phage-sensitive, antibiotic-resistant normal intestinal flora (Figure 8). Overnight cultures of E. coli O157 were washed by centrifugation to remove toxin and free phage particles, added to the phage susceptible strain at a ratio of 1:100 and incubated overnight in the presence of various levels of antibiotic. Bacterial replication was determined, but only the values for E. coli O157:H7 have been reported. In Table 2, MICs of the O157:H7 strains were determined visually, which corresponds to bacterial levels below 107 per ml.
Since visual determination of an antibiotic strain cannot be determined in the presence of
an antibiotic resistant strain, we used the value of 107 cfu to assign the MIC in these
studies.
Co-culture when both strains are antibiotic-sensitive.
In co-culture, MIC for the E. coli O157:H7 incubated ciprofloxacin or
azithromycin (Figure 7A and C shaded boxes) was similar to the MIC determined in
Table 2. In the absence of antibiotic, co-incubation caused about a 10-fold decrease in
the number of O157:H7 compared to pure cultures of O157:H7, likely due to competition
with the other strain. However, one hundred fewer O157:H7 bacteria were added to the
97 culture, suggesting the O157:H7 out-competed the other strain. A likely explanation is
that some of the commensal bacteria were killed by lytic infection by the Stx2 phage.
Lytic phage infection is further supported by the fact that co-incubation in the absence of antibiotic resulted in a statistically significant 10-fold increase in Stx production (Figure
7B and D). It is important to note that Stx production was increased in the co-incubation experiments, even though fewer E. coli O157:H7 were recovered.
The influence of co-incubation of Stx was also assessed. The ability of
ciprofloxacin to induce Stx production was less dramatic in the co-culture experiments.
The amount of Stx produced in the presence of sub-inhibitory levels of ciprofloxacin was
similar to that produced in the absence of antibiotic, while ciprofloxacin concentrations
above the MIC resulted in decreased Stx production (Figure 7B).
In contrast to ciprofloxacin, azithromycin had a significant impact on Stx production. Sub-inhibitory levels of azithromycin caused a statistically significant
decrease in Stx production compared to the pure culture of O157:H7, and in some cases
even compared to the amplified levels of Stx seen in the no antibiotic co-culture control
(Figure 7D). Interestingly, 1000-fold less Stx is produced in the presence of 16 μg/ml of
azithromycin compared to the no antibiotic control, but bacteria numbers are only reduce
about 100-fold.
98
A Ciprofloxacin (CFU) B Ciprofloxacin (toxin) 1012 10000 10 10 ** 1000 * ** 8 ## 10 **## 100 # 6 10 10 4 g/ml Stx 10 g/ml Stx μ 1 O157 colonies per ml 102 >MIC #* 0.1 100
E. coli E. 0.01 E. coli 0 0.02 0.05 0.10 0.22 0.44 E. coli 0 0.02 0.05 0.10 0.22 0.44 O157 O157 μg/ml antibiotic μg/ml antibiotic (E. coli O157 + sensitive amplifier) (E. coli O157 + sensitive amplifier)
C Azithromicin (CFU) D Azithromicin (toxin) 1012 10000 10 MICMIC 10 #* 1000 * #* *# 8 ** * 10 ## 100 # 6 10 10 #* g/ml Stx 4 g/ml Stx 10 μ 1 #*
O157 colonie per s ml 2 10 0.1 0 10 0.01 E. coli E. coli 0 1 2 4 8 16 E. coli 0 1 2 4 8 16 O157 O157 μg/ml antibiotic μg/ml antibiotic (E. coli O157 + sensitive amplifier) (E. coli O157 + sensitive amplifier)
Figure 7. The efficacy of antibiotics in the presence of a susceptible strains. E. coli O157:H7 strain 185 was grown overnight in the presence of a Stx-amplifying E. coli isolate susceptible to A) ciprofloxacin or B) azithromicin. Cocultures were plated, and filter-sterilized supernatants were analyzed using the luciferase assay. Symbols represent individual trials and the geometric mean is plotted. Statistical significance is designated with an asterisk (compared to E. coli O157:H7) or a pound sign (compared to the no antibiotic coculture control) according to the Student’s t-test (P<0.05). The MIC of E. coli O157:H7StrepKan in LB is noted by the grey box.
99 Co-culture with antibiotic-resistant Stx2-amplifying strains.
In these studies we tried to replicate the most confounding clinical scenario,
treatment of antibiotic-susceptible E. coli O157:H7 in the presence of antibiotic-resistant intestinal flora capable of producing Stx following exposure to the Stx encoding phage
(Figure 8). To generate antibiotic resistance, the phage susceptible strain was transformed with a plasmid encoding antibiotic resistance to amicpillin, gentamicin or doxycycline, and resistant phenotype was verified (data not shown).
Even in the presence of antibiotic resistant E. coli, the MIC for E. coli O157:H7
incubated with gentamicin and doxycycline (Figure 8A and C, shaded boxes) was only
slightly elevated compared to that seen for pure cultures. In contrast, antibiotic-sensitive
E. coli O157:H7 incubated with ampicillin resistant E. coli (Figure 8E, shaded box) was
completely resistant to ampicillin, likely due to the ability of β-lactamase to inactivate the
antibiotic.
As seen in Figure 7, in the absence of antibiotic Stx production was significantly
increased in the presence of the antibiotic-resistant amplifier strains compared to pure
cultures of E. coli O157:H7 (Fig. 8,B, D and F). In the presence of sub-inhibitory concentrations of gentamicin, ampicillin, or doxycycline, Stx produced under co-cultured conditions were similar to those observed in the absence of antibiotic (Fig. 8B, D, F).
However, in co-culture experiments with gentamicin and doxycycline at concentrations approaching the MIC, Stx production was highly variable, with greatly reduced expression observed in some cases.
100
A Gentamicin (CFU) B Gentamicin (toxin) 1012 10000 10 MIC 10 1000 * * ** 8 * * 10 * * * 100 106 10 4 g/ml Stx 10 g/ml Stx *# μ 1 O157 colonies per ml coloniesper O157 102 0.1 100
E. coli E. 0.01 E. coli 0 4 8 16 32 64 E. coli 0 4 8 16 32 64 O157 O157 μg/ml antibiotic μg/ml antibiotic (E. coliO157 + resistant amplifier) (E. coliO157 + resistant amplifier)
C Doxycycline (CFU) D Doxcyc yline (toxin) 1012 10000 10 MIC 10 #* #* * 1000 *** 8 *** # 10 100 6 10 10
4 g/ml Stx
10 μ 1
O157 colonies per ml per colonies O157 2 10 0.1 100 0.01 E. coli E. coli 0 0.5 1 2 4 8 E. coli 0 0.5 1 2 4 8 O157 O157 μg/ml antibiotic μg/ml antibiotic (E. coliO157 + resistant amplifier) (E. coliO157 + resistant amplifier)
E Ampicillin (CFU) F Ampicillin (toxin) 1012 10000 10 10 1000 8 *** * 10 100 *** 6 10 10 4 g/ml Stx g/ml Stx
10 μ 1
ml per colonies O157 2 10 0.1 100 E. coli 0.01 E. coli 0 128 256 512 1024 2048 E. coli 0 128 256 512 1024 2048 O157 O157 μg/ml antibiotic μg/ml antibiotic (E. coliO157 + resistant amplifier) (E. coliO157 + resistant amplifier)
FIG. 8. The efficacy of antibiotics in the presence of a resistant amplifier strain. E. coli O157:H7StrepKan was grown overnight in the presence of a Stx-amplifying E. coli isolate resistant to A,B) gentamicin, C,D) doxycycline, or E,F) ampicillin. Cocultures were plated, and filter-sterilized supernatants were run in the luciferase assay. Symbols represent individual trials and the geometric mean is plotted. Statistical significance is designated with an asterisk (compared to E. coli O157:H7) or a pound sign (compared to the no antibiotic coculture control) according to the Student’s t-test (P<0.05). The MIC of E. coli O157:H7StrepKan in LB is noted by the grey box.
101
Discussion
In order to develop a rapid detection method for the cytotoxicity of Stx on Vero
cells, Zhao et al. transiently transfected Vero cells with a adenovirus carrying a destabilized gene for luciferase [154], allowing one to monitor loss of luciferase production as an indicator for Stx-mediated reduction in protein synthesis. To eliminate the need to transfect Vero cells before every experiment, we created a Vero cell line in which luciferase was stably integrated into the chromosome. We used this assay to assess the influence of antibiotics on Stx production.
Currently, there are no therapeutics for the treatment of E. coli O157:H7.
Antibiotics are not administered because it has been suggested that patients who received antibiotics developed HUS at a higher rate than those that had not [8]. It has been demonstrated that ciprofloxacin and trimethoprim/sulfamethoxazole do in fact increase
Shiga toxin production [54,111]. The bacterial SOS-response is required for activation of the Stx phage lytic cycle, and the production of Shiga toxin. Ciprofloxacin inhibits DNA gyrase [102], initiating the SOS response. Similarly, trimethoprim/sulfamethoxazole inhibits folate synthesis [103]. Folate is a precursor for DNA nucleotides, and blockage of DNA synthesis due lack of nucleotide precursors can also induce the SOS response, and phage induction by trimethoprim/sulfamethoxazole is likely similar to that of ciprofloxacin.
To investigate whether any antibiotic can reduce Stx production, we incubated
four E. coli strains with nine different antibiotics. In every case, no Stx was produced
when bacterial growth was prevented. However, in vivo it is not always possible to keep
102 the concentration of antibiotics above the inhibitory level, and in this study we examined
the influence of sub-inhibitory levels of antibiotic on Stx production. It was not
surprising that both ciprofloxacin and trimethoprim/sulfamethoxazole significantly increased Stx production in the two clinical isolates of E. coli O157:H7. Sub-inhibitory
levels of ampicillin, ceftriaxone, rifampicin, and gentamicin did not significantly alter
Shiga toxin production, while fosfomicin, azithromicin, and doxycycline were shown to
decrease toxin production. These antibiotics target cell wall synthesis and translation.
Unlike the mechanism of ciprofloxacin, which inhibits bacterial DNA synthesis, yet
results in the initiation of Stx production, the data suggest that those antibiotics that
specifically target mechanisms such as protein synthesis will inhibit Stx production
directly, and these antibiotics may be potential therapeutics.
E. coli O157:H7 strains are often reported to have stx1 and/or stx2, based on PCR
detection [138]. Using an ELISA that can discriminate between Stx1 and Stx2 [178], we
have shown that a clinical isolate of E. coli O157:H7, although it has both toxin genes,
only produced Stx2. This is consistent with Stx2 being a major virulence factor in O157-
related disease. Future characterizations of E. coli O157:H7 isolates should include
assaying for the production of Shiga toxins specifically in order to establish a more
comprehensive picture of the mechanism of virulence of the individual isolate.
This model of testing the effects of antibiotics in the presence of a single Stx-
producing strain is likely oversimplified. Many types of bacteria are found in the gut,
and the commensal isolates themselves may play a role in E. coli O157:H7-related
disease development. For instance, in one study, it was shown that 10% of natural E.
coli isolates isolated from healthy volunteers were susceptible to E. coli O157:H7 phage
103 infection, at which point the natural isolates amplified Stx production [122]. It is possible
that those patients that develop HUS may harbor amplifying strains in the gut. We
therefore asked what effect antibiotics have on Shiga toxin production in the presence of
a commensal amplifier strain that was either resistant or sensitive to the antibiotic being
tested.
When testing the antibiotic susceptible Stx-amplifying isolate in the presence of
ciprofloxacin, Stx levels were high as expected. Most interesting was the ability of
azithromicin to significantly decrease the amount of Stx produced when compared to
either the coculture alone or more importantly, E. coli O157:H7 alone. Therefore, if a patient harboring a Stx-amplifying commensal isolate receives azithromicin, the antibiotic may greatly reduce disease severity in the patient.
In contrast, even though gentamicin, ampicillin, and doxycycline reduced E. coli
O157:H7 colony formation, Stx levels remained comparable to the no antibiotic control.
This suggests that a patient with antibiotic resistant Stx-amplifying commensals would
likely be in a severe disease state in which antibiotics would be of no benefit.
Many questions regarding the effectiveness of antibiotics as therapeutics in
treating E. coli O157:H7-related disease remain. For instance, it has yet to be determined
if Stx-amplifying isolates can be isolated from the gut of HUS patients. If it is shown that
amplifying strains play a primary role in HUS development, we may find that the risk of
administering protein synthesis inhibiting antibiotics to patients may not be as high as is
currently thought.
The key to antibiotic treatment will be early administration to those at a high risk
of developing disease. As described in Chapter 1, FoodNet has established that it is
104 critical that people must be aware of and recognize the symptoms linked to disease before medical diagnosis and treatments can be administered. Unfortunately, persons infected with E. coli O157:H7 often dismiss the early symptoms of disease as they consists simply of diarrhea. Medical attention is usually not sought until disease progresses to bloody diarrhea, at which point the Stx is already circulating throughout the body. Therefore, antibiotic therapies are not a likely treatment for these patients.
Public awareness of E. coli O157:H7-related symptoms will be heightened during an outbreak via information distributed by the media. The rate of early detection will be increased as people will be aware of not only the symptoms, but of the likely source of infection as well. It has yet to be determined if the timing of antibiotic therapies relative to infection will influence their effectiveness. We may find that antibiotics are ineffective in treating patients with bloody diarrhea, a symptom indicative of the systemic spread of Stx throughout the body. That being said, we cannot rule out the potential of using antibiotics proactively. For example, if a child in day care were to become infected, antibiotics could be administered to the asymptomatic children at the facillity.
Antibiotics have varying side effects, and as such, should only be given to those children who have a likelihood of developing severe disease. To date, this population cannot be identified. Therefore, diagnostics need to be developed before antibiotics can be administered to asymptomatic individuals.
In summary, antibiotics influence Stx production based on the mechanism of action. Ciprofloxacin and trimethoprim/sulfamethoxazole target DNA, and induce Stx production and should not be administered to patients. Those antibiotics that inhibit Stx production target protein synthesis, and have therapeutic potential.
105 In addition, it has been suggested that commensal E. coli can increase Stx production. We have shown that azithromycin reduces Stx levels beneath controls, indicating its potential for a patient harboring the amplifier strain. Conversely, antibiotics will be of no benefit for those persons with antibiotic resistant amplifier strains.
106 Chapter IV.
Discussion
107 Summary
My studies have focused on both the development of assays that can be used in
the research and diagnostic settings, and the examination of potential therapeutics that
can be used in the treatment of E. coli O157:H7. We have developed an ELISA that can
distinguish between Stx1 and Stx2 in the presence of fecal material. Unlike the diagnostic ELISA used now, which yields a positive result in the presence of either Stx,
this assay will lead to a more accurate prognosis for a patient infected with E. coli
O157:H7 as it can be determined if the isolate is producing Stx2, the toxin linked to severe disease, Stx1, or both.
We have also created a Vero cell line that can be used to quantify Stx within
hours. This is a vast improvement over the current methods used which can require days
to complete, reflect mechanisms other than Stx-mediated protein synthesis inhibition, and
in some instances rely on the subjective, visual determination of data points.
We have also completed an extensive study to determine which, if any, antibiotics
are safe for the treatment of E. coli O157:H7. We have shown that protein synthesis inhibitors reduce Stx production. This study has been extended beyond previously published research in that we have begun to examine antibiotics within the context of the effect of commensal, Stx-amplifying isolates. The data suggest that protein synthesis inhibitors may reduce Stx production, even if the patient harbors a Stx-amplifying isolate, provided that the commensal isolate is susceptible to the antibiotic. Conversely, antibiotics would likely be of no benefit if the Stx-amplifier was resistant to the antibiotic. Future studies should focus on characterizing the flora of HUS patients, and
examining antibiotic treatments using these strains.
108 Future Directions
Several labs have studied the effect of antibiotics in E. coli O157:H7-infected
mice [107,108,111,116,117], at which point the research projects seem to end with the same lackluster conclusion: antibiotic X may or may not be useful, and that further discussion is needed. In order for in vivo data regarding antibiotic therapeutics to be
novel and accepted with enthusiasm, the medical and research communities must first
determine more specifically which factors influence the development of HUS. Therefore,
although it may logically follow that the next step in my research is to utilize an in vivo
model, I have chosen to first focus on further research into the influence of commensal E.
coli on disease severity.
Isolation of commensal E. coli from severely ill patients to determine if disease is linked
to Stx-amplifying isolates.
The commensal isolate used in my studies was tested based on the hypothesis that
those that develop HUS may harbor Stx-amplifying commensal E. coli. This is a solid
hypothesis in that amplifiers were isolated from 10% of healthy volunteers, and 10% of
persons infected with E. coli O157:H7 develop HUS. Yet, these persons were just that,
healthy. Therefore, it remains to be determined if amplifiers are present in the gut of
severely ill patients.
To perform these studies, samples of commensal E. coli isolates would first be
acquired from healthy persons and those with watery diarrhea, hemorrhagic colitis or
HUS. A preliminary study in collaboration with Children’s Hospital could facilitate the
acquisition of samples from those infected individuals upon diagnosis. Our lab has
109 worked with persons at this facility in the past to attain samples for similar studies,
therefore, the hospital will likely be open to such a dialog.
After isolation of the commensal E. coli strains, the ability of these isolates to
amplify Stx production would be tested by incubating the isolate with a clinical strain of
Stx-producing E. coli O157:H7. Stx production in the presence and absence of the amplifer would be quantified using the luciferase-expressing Vero cell line.
Fully testing the hypothesis requires a large sample number, and the likelihood of
acquiring these from Cincinnati alone is low. Evidence that Stx amplifiers are present in
the gut of HUS patients may help to facilitate a relationship with an organization such as
the CDC. This collaboration would be needed in order to facilitate access to large
numbers of samples, for instance during an outbreak. During the American Society for
Microbiology’s General Meeting in Spring of 2008, a member of the CDC was
encouraged by my work, indicating that she felt that antibiotics were dismissed too
quickly. Therefore, this organization may be receptive to providing us with the needed
samples.
Using a murine model to test the efficacy of antibiotics in the presence of both E. coli
O157:H7 and Stx-amplifying commensal isolates.
If it is found that commensal strains isolated from severely ill patients amplify
Stx, the effect of antibiotics on the production of Stx in the presence of commensal E.
coli can be tested in vivo. A murine model has been used has been used to study the
effect of commensal E. coli on the production of Stx in animals infected with E. coli
O157:H7 [108]. Although mice do not mimic human disease, this is the logical first step
110 to in vivo studies. Mice are first treated with streptomycin to clear the gastrointestinal
tract of the animals of intestinal flora, and are then colonized with the commensal E. coli
strain of interest. The mice are subsequently infected with E. coli O157:H7. Antibiotics
can be administered both pre- and post infection to determine the effect of antibiotic at various stages of disease. Bacterial colonization and infection, in addition to Stx production, will be measured via fecal analysis.
Development of a Diagnostic Biomarker
Antibiotics are known to have detrimental side effects. Therefore, antibiotics
should be administered with discretion. Varying levels of E. coli O157:H7-related
disease exists, and most people resolve the disease without treatment. As such, antibiotic
therapies would optimally be given only to those at risk of developing HUS.
The glycoconjugate ELISA just described will be beneficial in diagnosing infected patients, but an assay that indicates that a patient may have an increased propensity for developing HUS due to the presence of a biomarker has yet to be developed. The following section will describe the potential of Gb3 receptor to act as a
biomarker.
Both Stx1 and Stx2 bind to Gb3 [188]. It has been hypothesized that the at-risk populations may harbor increased levels of the receptor within the kidney. Attempts to study this question have centered on the analysis of tissue sections taken from people of all age groups, some of which died of E. coli O157:H7-related HUS [73,76]. The results remain inconclusive.
111 Gb3 is also has a role in Fabry disease. In Fabry disease, patients have a
deficiency in α-galactosidase [189]. This results in the accumulation of Gb3 within the
body. Interestingly, researchers have developed a way to measure this receptor in the
urine [190,191]. Samples were first collected by having both healthy persons and those
with Fabry disease urinate on Whatman paper. These papers were dried and mailed back
to the lab. Labelled creatine and a known isoform of Gb3 were spotted onto the paper as
controls. Incubation of the paper with methanol removed the sample from the paper, and
samples were analyzed using liquid-chromatography-tandem mass spectrometry. Data
were expressed as a ratio of the receptor to creatine to control for the fact that different
amounts of urine were deposited onto the papers by each patient. Elevated amounts of
Gb3 were present in the urine of a Fabry patient as compared to a normal control child.
Currently, there are no reports of the presence or absence of Gb3 in the urine of
HUS patients. It would be interesting to determine if levels of Gb3 in HUS patients differ
from that of those will less severe illness. Again, development of this protocol included the analysis of samples taken from children, the population that develops severe E. coli
O157:H7-related disease. Samples will be acquired in a noninvasive manner and can be simply mailed back to the lab. The Mass Spectrometry Core Facillity at UC has the facillities needed to analyze the samples, and the results can be compared to the aforementioned published data.
If Gb3 levels were found to differ between healthy and E. coli O157:H7-infected patients, using Gb3 as a biomarker could help facilitate determining if antibiotic treatment
is appropriate for a given patient.
112 Concluding Remarks
Since E. coli O157:H7 was first identified, we have discovered much about the
biology of the organism. It is known that the bacterium can intimately attach to the gut, produce Stxs, and instruct commensal E. coli to produce Stx as well. Patients have varying levels of disease, including diarrhea, bloody diarrhea, kidney damage, and even death. Yet, in those 20 years, we have not yet determined with a high degree of certainty why some people resolve the infection without consequence, while others have lifelong complications or die. As for therpeutic options, in reality, there are none, as we are currently wrought with the fear that they will enhance disease. And therefore, patients are merely left with a heartfelt “I wish there was something else we could do.” In the next 20 years, it is my hope that the seemingly endless perseverance of research will prevail, and we will come to fully understand this bacteria we call O157. Only then will be able to join hands and sing, “Oh Shiga, do-do-do-do-do-do, O157...you’re not my toxin, yeah!, And I’m not poopin’ you!”
113
References
1. Beutin L, Geier D, Steinruck H, Zimmermann S, Scheutz F (1993) Prevalence and some properties of verotoxin (Shiga-like toxin)-producing Escherichia coli in seven different species of healthy domestic animals. J Clin Microbiol 31: 2483-2488.
2. Burnens AP, Frey A, Lior H, Nicolet J (1995) Prevalence and clinical significance of vero-cytotoxin-producing Escherichia coli (VTEC) isolated from cattle in herds with and without calf diarrhoea. Zentralbl Veterinarmed B 42: 311- 318.
3. Samadpour M, Ongerth JE, Liston J, Tran N, Nguyen D, et al. (1994) Occurrence of Shiga-like toxin-producing Escherichia coli in retail fresh seafood, beef, lamb, pork, and poultry from grocery stores in Seattle, Washington. Appl Environ Microbiol 60: 1038-1040.
4. Blanco M, Blanco JE, Blanco J, Mora A, Prado C, et al. (1997) Distribution and characterization of faecal verotoxin-producing Escherichia coli (VTEC) isolated from healthy cattle. Vet Microbiol 54: 309-319.
5. Ostroff SM, Tarr PI, Neill MA, Lewis JH, Hargrett-Bean N, et al. (1989) Toxin genotypes and plasmid profiles as determinants of systemic sequelae in Escherichia coli O157:H7 infections. J Infect Dis 160: 994-998.
6. Karmali MA, Steele BT, Petric M, Lim C (1983) Sporadic cases of haemolytic- uraemic syndrome associated with faecal cytotoxin and cytotoxin-producing Escherichia coli in stools. Lancet 1: 619-620.
7. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, et al. (1983) Hemorrhagic colitis associated with a rare Escherichia coli serotype. N Engl J Med 308: 681-685.
8. Wong CS, Jelacic S, Habeeb RL, Watkins SL, Tarr PI (2000) The risk of the hemolytic-uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 infections. N Engl J Med 342: 1930-1936.
9. Day NP, Scotland SM, Cheasty T, Rowe B (1983) Escherichia coli O157:H7 associated with human infections in the United Kingdom. Lancet 1: 825.
114 10. Griffin PM, Tauxe RV (1991) The epidemiology of infections caused by Escherichia coli O157:H7, other enterohemorrhagic E. coli, and the associated hemolytic uremic syndrome. Epidemiol Rev 13: 60-98.
11. Johnson WM, Lior H, Bezanson GS (1983) Cytotoxic Escherichia coli O157:H7 associated with haemorrhagic colitis in Canada. Lancet 1: 76.
12. Office of the Inspector General U (2008) Public Health Risk-Based Inspection System for Processing and Slaughter. 1-41 p.
13. Rangel JM, Sparling PH, Crowe C, Griffin PM, Swerdlow DL (2005) Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982- 2002. Emerg Infect Dis 11: 603-609.
14. United States Department of Agriculture FSIS (2008) Ohio Firm Recalls Ground Beef Products Due To Possible E. coli O157:H7 Contamination.
15. Murphy RY, Seward RA (2004) Process control and sampling for Escherichia coli O157:H7 in beef trimmings. J Food Prot 67: 1755-1759.
16. CDC CfDCaP (2006) Ongoing multistate outbreak of Escherichia coli serotype O157:H7 infections associated with consumption of fresh spinach--United States, September 2006. Morbidity and Mortality Weekly 55: 1045-1046.
17. al-Saigh H, Zweifel C, Blanco J, Blanco JE, Blanco M, et al. (2004) Fecal shedding of Escherichia coli O157, Salmonella, and Campylobacter in Swiss cattle at slaughter. J Food Prot 67: 679-684.
18. Stephens TP, McAllister TA, Stanford K (2008) Development of an experimental model to assess the ability of Escherichia coli O157:H7-inoculated fecal pats to mimic a super shedder within a feedlot environment. J Food Prot 71: 648- 652.
19. Depenbusch BE, Nagaraja TG, Sargeant JM, Drouillard JS, Loe ER, et al. (2008) Influence of processed grains on fecal pH, starch concentration, and shedding of Escherichia coli O157 in feedlot cattle. J Anim Sci 86: 632-639.
115 20. CDC CfDCaP (2005) Outbreaks of Escherichia coli O157:H7 associated with petting zoos--North Carolina, Florida, and Arizona, 2004 and 2005. MMWR Morb Mortal Wkly Rep 54: 1277-1280.
21. Smith KE, Stenzel SA, Bender JB, Wagstrom E, Soderlund D, et al. (2004) Outbreaks of enteric infections caused by multiple pathogens associated with calves at a farm day camp. Pediatr Infect Dis J 23: 1098-1104.
22. Varma JK, Greene KD, Reller ME, DeLong SM, Trottier J, et al. (2003) An outbreak of Escherichia coli O157 infection following exposure to a contaminated building. Jama 290: 2709-2712.
23. Belongia EA, Osterholm MT, Soler JT, Ammend DA, Braun JE, et al. (1993) Transmission of Escherichia coli 0157:H7 infection in Minnesota child day- care facilities. Jama 269: 883-888.
24. Galanis E, Longmore K, Hasselback P, Swann D, Ellis A, et al. (2003) Investigation of an E. coli O157:H7 outbreak in Brooks, Alberta, June-July 2002: the role of occult cases in the spread of infection within a daycare setting. Can Commun Dis Rep 29: 21-28.
25. Allos BM, Moore MR, Griffin PM, Tauxe RV (2004) Surveillance for sporadic foodborne disease in the 21st century: the FoodNet perspective. Clin Infect Dis 38 Suppl 3: S115-120.
26. Voetsch AC, Angulo FJ, Rabatsky-Ehr T, Shallow S, Cassidy M, et al. (2004) Laboratory practices for stool-specimen culture for bacterial pathogens, including Escherichia coli O157:H7, in the FoodNet sites, 1995-2000. Clin Infect Dis 38 Suppl 3: S190-197.
27. Kassenborg HD, Hedberg CW, Hoekstra M, Evans MC, Chin AE, et al. (2004) Farm visits and undercooked hamburgers as major risk factors for sporadic Escherichia coli O157:H7 infection: data from a case-control study in 5 FoodNet sites. Clin Infect Dis 38 Suppl 3: S271-278.
28. Barrett TJ, Lior H, Green JH, Khakhria R, Wells JG, et al. (1994) Laboratory investigation of a multistate food-borne outbreak of Escherichia coli O157:H7 by using pulsed-field gel electrophoresis and phage typing. J Clin Microbiol 32: 3013-3017.
116 29. Swaminathan B, Barrett TJ, Hunter SB, Tauxe RV (2001) PulseNet: the molecular subtyping network for foodborne bacterial disease surveillance, United States. Emerg Infect Dis 7: 382-389.
30. Jay MT, Cooley M, Carychao D, Wiscomb GW, Sweitzer RA, et al. (2007) Escherichia coli O157:H7 in feral swine near spinach fields and cattle, central California coast. Emerg Infect Dis 13: 1908-1911.
31. Slutsker L, Ries AA, Greene KD, Wells JG, Hutwagner L, et al. (1997) Escherichia coli O157:H7 diarrhea in the United States: clinical and epidemiologic features. Ann Intern Med 126: 505-513.
32. Cohen MB, Giannella RA (1992) Hemorrhagic colitis associated with Escherichia coli O157:H7. Adv Intern Med 37: 173-195.
33. Ina K, Kusugami K, Ohta M (2003) Bacterial hemorrhagic enterocolitis. J Gastroenterol 38: 111-120.
34. Karmali MA (1989) Infection by verocytotoxin-producing Escherichia coli. Clin Microbiol Rev 2: 15-38.
35. Tarr PI, Gordon CA, Chandler WL (2005) Shiga-toxin-producing Escherichia coli and haemolytic uraemic syndrome. Lancet 365: 1073-1086.
36. Karmali MA, Petric M, Lim C, Fleming PC, Arbus GS, et al. (1985) The association between idiopathic hemolytic uremic syndrome and infection by verotoxin-producing Escherichia coli. J Infect Dis 151: 775-782.
37. Tsai HM, Chandler WL, Sarode R, Hoffman R, Jelacic S, et al. (2001) von Willebrand factor and von Willebrand factor-cleaving metalloprotease activity in Escherichia coli O157:H7-associated hemolytic uremic syndrome. Pediatr Res 49: 653-659.
38. Chandler WL, Jelacic S, Boster DR, Ciol MA, Williams GD, et al. (2002) Prothrombotic coagulation abnormalities preceding the hemolytic-uremic syndrome. N Engl J Med 346: 23-32.
117 39. Gagnadoux MF, Habib R, Gubler MC, Bacri JL, Broyer M (1996) Long-term (15-25 years) outcome of childhood hemolytic-uremic syndrome. Clin Nephrol 46: 39-41.
40. Eaton KA, Friedman DI, Francis GJ, Tyler JS, Young VB, et al. (2008) Pathogenesis of renal disease due to enterohemorrhagic Escherichia coli in germ-free mice. Infect Immun 76: 3054-3063.
41. Niyogi SK (2005) Shigellosis. J Microbiol 43: 133-143.
42. Trofa AF, Ueno-Olsen H, Oiwa R, Yoshikawa M (1999) Dr. Kiyoshi Shiga: discoverer of the dysentery bacillus. Clin Infect Dis 29: 1303-1306.
43. Flexner S (1900) On the etiology of tripical dysentery. Bulletin of Johns Hopkins Hospital 11: 231-242.
44. Konowalchuk J, Speirs JI, Stavric S (1977) Vero response to a cytotoxin of Escherichia coli. Infect Immun 18: 775-779.
45. O'Brien AD, LaVeck GD, Thompson MR, Formal SB (1982) Production of Shigella dysenteriae type 1-like cytotoxin by Escherichia coli. J Infect Dis 146: 763-769.
46. Karmali MA, Petric M, Louie S, Cheung R (1986) Antigenic heterogeneity of Escherichia coli verotoxins. Lancet 1: 164-165.
47. Hashimoto H, Mizukoshi K, Nishi M, Kawakita T, Hasui S, et al. (1999) Epidemic of gastrointestinal tract infection including hemorrhagic colitis attributable to Shiga toxin 1-producing Escherichia coli O118:H2 at a junior high school in Japan. Pediatrics 103: E2.
48. Nataro JP, Kaper JB (1998) Diarrheagenic Escherichia coli. Clin Microbiol Rev 11: 142-201.
49. Strockbine NA, Jackson MP, Sung LM, Holmes RK, O'Brien AD (1988) Cloning and sequencing of the genes for Shiga toxin from Shigella dysenteriae type 1. J Bacteriol 170: 1116-1122.
118 50. Jackson MP, Newland JW, Holmes RK, O'Brien AD (1987) Nucleotide sequence analysis of the structural genes for Shiga-like toxin I encoded by bacteriophage 933J from Escherichia coli. Microb Pathog 2: 147-153.
51. Plunkett G, 3rd, Rose DJ, Durfee TJ, Blattner FR (1999) Sequence of Shiga toxin 2 phage 933W from Escherichia coli O157:H7: Shiga toxin as a phage late-gene product. J Bacteriol 181: 1767-1778.
52. Smith HW, Green P, Parsell Z (1983) Vero cell toxins in Escherichia coli and related bacteria: transfer by phage and conjugation and toxic action in laboratory animals, chickens and pigs. J Gen Microbiol 129: 3121-3137.
53. Johnson AD, Poteete AR, Lauer G, Sauer RT, Ackers GK, et al. (1981) lambda Repressor and cro--components of an efficient molecular switch. Nature 294: 217-223.
54. Kimmitt PT, Harwood CR, Barer MR (1999) Induction of type 2 Shiga toxin synthesis in Escherichia coli O157 by 4-quinolones. Lancet 353: 1588-1589.
55. Shaikh N, Tarr PI (2003) Escherichia coli O157:H7 Shiga toxin-encoding bacteriophages: integrations, excisions, truncations, and evolutionary implications. J Bacteriol 185: 3596-3605.
56. Kimmitt PT, Harwood CR, Barer MR (2000) Toxin gene expression by shiga toxin-producing Escherichia coli: the role of antibiotics and the bacterial SOS response. Emerg Infect Dis 6: 458-465.
57. Fraser ME, Fujinaga M, Cherney MM, Melton-Celsa AR, Twiddy EM, et al. (2004) Structure of shiga toxin type 2 (Stx2) from Escherichia coli O157:H7. J Biol Chem 279: 27511-27517.
58. Olsnes S, Reisbig R, Eiklid K (1981) Subunit structure of Shigella cytotoxin. J Biol Chem 256: 8732-8738.
59. Garred O, Dubinina E, Holm PK, Olsnes S, van Deurs B, et al. (1995) Role of processing and intracellular transport for optimal toxicity of Shiga toxin and toxin mutants. Exp Cell Res 218: 39-49.
119 60. Calderwood SB, Auclair F, Donohue-Rolfe A, Keusch GT, Mekalanos JJ (1987) Nucleotide sequence of the Shiga-like toxin genes of Escherichia coli. Proc Natl Acad Sci U S A 84: 4364-4368.
61. Newland JW, Strockbine NA, Neill RJ (1987) Cloning of genes for production of Escherichia coli Shiga-like toxin type II. Infect Immun 55: 2675-2680.
62. O'Brien AD, Tesh VL, Donohue-Rolfe A, Jackson MP, Olsnes S, et al. (1992) Shiga toxin: biochemistry, genetics, mode of action, and role in pathogenesis. Curr Top Microbiol Immunol 180: 65-94.
63. Fuchs G, Mobassaleh M, Donohue-Rolfe A, Montgomery RK, Grand RJ, et al. (1986) Pathogenesis of Shigella diarrhea: rabbit intestinal cell microvillus membrane binding site for Shigella toxin. Infect Immun 53: 372-377.
64. Jacewicz M, Clausen H, Nudelman E, Donohue-Rolfe A, Keusch GT (1986) Pathogenesis of shigella diarrhea. XI. Isolation of a shigella toxin-binding glycolipid from rabbit jejunum and HeLa cells and its identification as globotriaosylceramide. J Exp Med 163: 1391-1404.
65. Clark C, Bast D, Sharp AM, St Hilaire PM, Agha R, et al. (1996) Phenylalanine 30 plays an important role in receptor binding of verotoxin-1. Mol Microbiol 19: 891-899.
66. Shin IS, Nishikawa K, Maruyama H, Ishii S (2006) Histidine-tagged shiga toxin B subunit binding assay: simple and specific determination of gb3 content in mammalian cells. Chem Pharm Bull (Tokyo) 54: 522-527.
67. Head SC, Karmali MA, Lingwood CA (1991) Preparation of VT1 and VT2 hybrid toxins from their purified dissociated subunits. Evidence for B subunit modulation of a subunit function. J Biol Chem 266: 3617-3621.
68. Lindberg AA, Brown JE, Stromberg N, Westling-Ryd M, Schultz JE, et al. (1987) Identification of the carbohydrate receptor for Shiga toxin produced by Shigella dysenteriae type 1. J Biol Chem 262: 1779-1785.
69. Lingwood CA, Law H, Richardson S, Petric M, Brunton JL, et al. (1987) Glycolipid binding of purified and recombinant Escherichia coli produced verotoxin in vitro. J Biol Chem 262: 8834-8839.
120 70. Waddell T, Head S, Petric M, Cohen A, Lingwood C (1988) Globotriosyl ceramide is specifically recognized by the Escherichia coli verocytotoxin 2. Biochem Biophys Res Commun 152: 674-679.
71. Gillard BK, Jones MA, Marcus DM (1987) Glycosphingolipids of human umbilical vein endothelial cells and smooth muscle cells. Arch Biochem Biophys 256: 435-445.
72. Kang JL, Rajpert-De Meyts E, Wiels J, Skakkebaek NE (1995) Expression of the glycolipid globotriaosylceramide (Gb3) in testicular carcinoma in situ. Virchows Arch 426: 369-374.
73. Boyd B, Lingwood C (1989) Verotoxin receptor glycolipid in human renal tissue. Nephron 51: 207-210.
74. Kojima Y, Fukumoto S, Furukawa K, Okajima T, Wiels J, et al. (2000) Molecular cloning of globotriaosylceramide/CD77 synthase, a glycosyltransferase that initiates the synthesis of globo series glycosphingolipids. J Biol Chem 275: 15152-15156.
75. Hughes AK, Ergonul Z, Stricklett PK, Kohan DE (2002) Molecular basis for high renal cell sensitivity to the cytotoxic effects of shigatoxin-1: upregulation of globotriaosylceramide expression. J Am Soc Nephrol 13: 2239-2245.
76. Ergonul Z, Clayton F, Fogo AB, Kohan DE (2003) Shigatoxin-1 binding and receptor expression in human kidneys do not change with age. Pediatr Nephrol 18: 246-253.
77. Chaisri U, Nagata M, Kurazono H, Horie H, Tongtawe P, et al. (2001) Localization of Shiga toxins of enterohaemorrhagic Escherichia coli in kidneys of paediatric and geriatric patients with fatal haemolytic uraemic syndrome. Microb Pathog 31: 59-67.
78. Jacewicz MS, Acheson DW, Binion DG, West GA, Lincicome LL, et al. (1999) Responses of human intestinal microvascular endothelial cells to Shiga toxins 1 and 2 and pathogenesis of hemorrhagic colitis. Infect Immun 67: 1439- 1444.
121 79. Rutjes NW, Binnington BA, Smith CR, Maloney MD, Lingwood CA (2002) Differential tissue targeting and pathogenesis of verotoxins 1 and 2 in the mouse animal model. Kidney Int 62: 832-845.
80. Ling H, Boodhoo A, Hazes B, Cummings MD, Armstrong GD, et al. (1998) Structure of the shiga-like toxin I B-pentamer complexed with an analogue of its receptor Gb3. Biochemistry 37: 1777-1788.
81. Ling H, Pannu NS, Boodhoo A, Armstrong GD, Clark CG, et al. (2000) A mutant Shiga-like toxin IIe bound to its receptor Gb(3): structure of a group II Shiga-like toxin with altered binding specificity. Structure Fold Des 8: 253- 264.
82. Soltyk AM, MacKenzie CR, Wolski VM, Hirama T, Kitov PI, et al. (2002) A mutational analysis of the globotriaosylceramide-binding sites of verotoxin VT1. J Biol Chem 277: 5351-5359.
83. Nyholm PG, Magnusson G, Zheng Z, Norel R, Binnington-Boyd B, et al. (1996) Two distinct binding sites for globotriaosyl ceramide on verotoxins: identification by molecular modelling and confirmation using deoxy analogues and a new glycolipid receptor for all verotoxins. Chem Biol 3: 263- 275.
84. Kovbasnjuk O, Edidin M, Donowitz M (2001) Role of lipid rafts in Shiga toxin 1 interaction with the apical surface of Caco-2 cells. J Cell Sci 114: 4025-4031.
85. Hanzal-Bayer MF, Hancock JF (2007) Lipid rafts and membrane traffic. FEBS Lett 581: 2098-2104.
86. Sandvig K, van Deurs B (1992) Toxin-induced cell lysis: protection by 3- methyladenine and cycloheximide. Exp Cell Res 200: 253-262.
87. Sandvig K, Olsnes S, Brown JE, Petersen OW, van Deurs B (1989) Endocytosis from coated pits of Shiga toxin: a glycolipid-binding protein from Shigella dysenteriae 1. J Cell Biol 108: 1331-1343.
88. Sandvig K, Prydz K, Ryd M, van Deurs B (1991) Endocytosis and intracellular transport of the glycolipid-binding ligand Shiga toxin in polarized MDCK cells. J Cell Biol 113: 553-562.
122 89. Schapiro FB, Lingwood C, Furuya W, Grinstein S (1998) pH-independent retrograde targeting of glycolipids to the Golgi complex. Am J Physiol 274: C319-332.
90. Nichols BJ, Kenworthy AK, Polishchuk RS, Lodge R, Roberts TH, et al. (2001) Rapid cycling of lipid raft markers between the cell surface and Golgi complex. J Cell Biol 153: 529-541.
91. Katagiri YU, Mori T, Nakajima H, Katagiri C, Taguchi T, et al. (1999) Activation of Src family kinase yes induced by Shiga toxin binding to globotriaosyl ceramide (Gb3/CD77) in low density, detergent-insoluble microdomains. J Biol Chem 274: 35278-35282.
92. Takenouchi H, Kiyokawa N, Taguchi T, Matsui J, Katagiri YU, et al. (2004) Shiga toxin binding to globotriaosyl ceramide induces intracellular signals that mediate cytoskeleton remodeling in human renal carcinoma-derived cells. J Cell Sci 117: 3911-3922.
93. Sandvig K (2001) Shiga toxins. Toxicon 39: 1629-1635.
94. Sandvig K, Garred O, van Helvoort A, van Meer G, van Deurs B (1996) Importance of glycolipid synthesis for butyric acid-induced sensitization to shiga toxin and intracellular sorting of toxin in A431 cells. Mol Biol Cell 7: 1391-1404.
95. Johannes L, Goud B (1998) Surfing on a retrograde wave: how does Shiga toxin reach the endoplasmic reticulum? Trends Cell Biol 8: 158-162.
96. Sandvig K, Ryd M, Garred O, Schweda E, Holm PK, et al. (1994) Retrograde transport from the Golgi complex to the ER of both Shiga toxin and the nontoxic Shiga B-fragment is regulated by butyric acid and cAMP. J Cell Biol 126: 53-64.
97. White J, Johannes L, Mallard F, Girod A, Grill S, et al. (1999) Rab6 coordinates a novel Golgi to ER retrograde transport pathway in live cells. J Cell Biol 147: 743-760.
98. Sandvig K, van Deurs B (1996) Endocytosis, intracellular transport, and cytotoxic action of Shiga toxin and ricin. Physiol Rev 76: 949-966.
123 99. Endo Y, Tsurugi K, Yutsudo T, Takeda Y, Ogasawara T, et al. (1988) Site of action of a Vero toxin (VT2) from Escherichia coli O157:H7 and of Shiga toxin on eukaryotic ribosomes. RNA N-glycosidase activity of the toxins. Eur J Biochem 171: 45-50.
100. Saxena SK, O'Brien AD, Ackerman EJ (1989) Shiga toxin, Shiga-like toxin II variant, and ricin are all single-site RNA N-glycosidases of 28 S RNA when microinjected into Xenopus oocytes. J Biol Chem 264: 596-601.
101. Kucers A (1997) The use of antibiotics : a clinical review of antibacterial, antifungal, and antiviral drugs. Oxford ; Boston: Butterworth-Heinemann. x, 1950 p.
102. Maxwell A (1997) DNA gyrase as a drug target. Trends Microbiol 5: 102-109.
103. Walsh C (2003) ANTIBIOTICS: Actions, Origins, Resistance. Washington D.C.: ASM Press. 335 p.
104. Walterspiel JN, Ashkenazi S, Morrow AL, Cleary TG (1992) Effect of subinhibitory concentrations of antibiotics on extracellular Shiga-like toxin I. Infection 20: 25-29.
105. Gamage SD, Patton AK, Hanson JF, Weiss AA (2004) Diversity and host range of Shiga toxin-encoding phage. Infect Immun 72: 7131-7139.
106. Grif K, Dierich MP, Karch H, Allerberger F (1998) Strain-specific differences in the amount of Shiga toxin released from enterohemorrhagic Escherichia coli O157 following exposure to subinhibitory concentrations of antimicrobial agents. Eur J Clin Microbiol Infect Dis 17: 761-766.
107. Zhang X, McDaniel AD, Wolf LE, Keusch GT, Waldor MK, et al. (2000) Quinolone antibiotics induce Shiga toxin-encoding bacteriophages, toxin production, and death in mice. J Infect Dis 181: 664-670.
108. Gamage SD, Patton AK, Strasser JE, Chalk CL, Weiss AA (2006) Commensal bacteria influence Escherichia coli O157:H7 persistence and Shiga toxin production in the mouse intestine. Infect Immun 74: 1977-1983.
124 109. Leibovitz E, Janco J, Piglansky L, Press J, Yagupsky P, et al. (2000) Oral ciprofloxacin vs. intramuscular ceftriaxone as empiric treatment of acute invasive diarrhea in children. Pediatr Infect Dis J 19: 1060-1067.
110. Acheson DW, Sears CL (2001) Dangers of empiric oral ciprofloxacin in the treatment of acute inflammatory diarrhea in children. Pediatr Infect Dis J 20: 817-818.
111. Kurioka T, Yunou Y, Harada H, Kita E (1999) Efficacy of antibiotic therapy for infection with Shiga-like toxin-producing Escherichia coli O157:H7 in mice with protein-calorie malnutrition. Eur J Clin Microbiol Infect Dis 18: 561-571.
112. Pavia AT, Nichols CR, Green DP, Tauxe RV, Mottice S, et al. (1990) Hemolytic-uremic syndrome during an outbreak of Escherichia coli O157:H7 infections in institutions for mentally retarded persons: clinical and epidemiologic observations. J Pediatr 116: 544-551.
113. Slutsker L, Ries AA, Maloney K, Wells JG, Greene KD, et al. (1998) A nationwide case-control study of Escherichia coli O157:H7 infection in the United States. J Infect Dis 177: 962-966.
114. Yoh M, Frimpong EK, Honda T (1997) Effect of antimicrobial agents, especially fosfomycin, on the production and release of Vero toxin by enterohaemorrhagic Escherichia coli O157:H7. FEMS Immunol Med Microbiol 19: 57-64.
115. Kondo K, Doi H, Adachi H, Nishimura Y (2004) Synergistic effect of CMP/KDO synthase inhibitors with antimicrobial agents on inhibition of production and release of Vero toxin by enterohaemorrhagic Escherichia coli O157:H7. Bioorg Med Chem Lett 14: 467-470.
116. Isogai E, Isogai H, Hayashi S, Kubota T, Kimura K, et al. (2000) Effect of antibiotics, levofloxacin and fosfomycin, on a mouse model with Escherichia coli O157 infection. Microbiol Immunol 44: 89-95.
117. Yoh M, Frimpong EK, Voravuthikunchai SP, Honda T (1999) Effect of subinhibitory concentrations of antimicrobial agents (quinolones and macrolide) on the production of verotoxin by enterohemorrhagic Escherichia coli O157:H7. Can J Microbiol 45: 732-739.
125 118. Ohara T, Kojio S, Taneike I, Nakagawa S, Gondaira F, et al. (2002) Effects of azithromycin on shiga toxin production by Escherichia coli and subsequent host inflammatory response. Antimicrob Agents Chemother 46: 3478-3483.
119. Matar GM, Rahal E (2003) Inhibition of the transcription of the Escherichia coli O157:H7 genes coding for shiga-like toxins and intimin, and its potential use in the treatment of human infection with the bacterium. Ann Trop Med Parasitol 97: 281-287.
120. Weiss AA, Johnson FD, Burns DL (1993) Molecular characterization of an operon required for pertussis toxin secretion. Proc Natl Acad Sci U S A 90: 2970-2974.
121. Tilden J, Jr., Young W, McNamara AM, Custer C, Boesel B, et al. (1996) A new route of transmission for Escherichia coli: infection from dry fermented salami. Am J Public Health 86: 1142-1145.
122. Gamage SD, Strasser JE, Chalk CL, Weiss AA (2003) Nonpathogenic Escherichia coli can contribute to the production of Shiga toxin. Infect Immun 71: 3107-3115.
123. Tanamoto K, Zahringer U, McKenzie GR, Galanos C, Rietschel ET, et al. (1984) Biological activities of synthetic lipid A analogs: pyrogenicity, lethal toxicity, anticomplement activity, and induction of gelation of Limulus amoebocyte lysate. Infect Immun 44: 421-426.
124. Gamage SD, McGannon CM, Weiss AA (2004) Escherichia coli serogroup O107/O117 lipopolysaccharide binds and neutralizes Shiga toxin 2. J Bacteriol 186: 5506-5512.
125. Leslie MR, Parolis H, Parolis LA (2000) The structure of the O-specific polysaccharide of Escherichia coli O117:K98:H4. Carbohydr Res 323: 103- 110.
126. March SB, Ratnam S (1986) Sorbitol-MacConkey medium for detection of Escherichia coli O157:H7 associated with hemorrhagic colitis. J Clin Microbiol 23: 869-872.
126 127. Krishnan C, Fitzgerald VA, Dakin SJ, Behme RJ (1987) Laboratory investigation of outbreak of hemorrhagic colitis caused by Escherichia coli O157:H7. J Clin Microbiol 25: 1043-1047.
128. Cryan B (1989) Enterohaemorrhagic Escherichia coli serotype O157:H7 as an enteric pathogen in Irish children. Ir J Med Sci 158: 18-19.
129. Wu VC, Qiu X, Peggy Hsieh YH (2008) Evaluation of Escherichia coli O157:H7 in apple juice with Cornus fruit (Cornus officinalis Sieb. et Zucc.) extract by conventional media and thin agar layer method. Food Microbiol 25: 190-195.
130. Graumann GH, Holley RA (2007) Survival of Escherichia coli O157:H7 in needle-tenderized dry cured Westphalian ham. Int J Food Microbiol 118: 173-179.
131. Al-Jader L, Salmon RL, Walker AM, Williams HM, Willshaw GA, et al. (1999) Outbreak of Escherichia coli O157 in a nursery: lessons for prevention. Arch Dis Child 81: 60-63.
132. Olsen SJ, Miller G, Breuer T, Kennedy M, Higgins C, et al. (2002) A waterborne outbreak of Escherichia coli O157:H7 infections and hemolytic uremic syndrome: implications for rural water systems. Emerg Infect Dis 8: 370-375.
133. March SB, Ratnam S (1989) Latex agglutination test for detection of Escherichia coli serotype O157. J Clin Microbiol 27: 1675-1677.
134. Galland JC, Hyatt DR, Crupper SS, Acheson DW (2001) Prevalence, antibiotic susceptibility, and diversity of Escherichia coli O157:H7 isolates from a longitudinal study of beef cattle feedlots. Appl Environ Microbiol 67: 1619- 1627.
135. Chart H (1999) Evaluation of a latex agglutination kit for the detection of human antibodies to the lipopolysaccharide of Escherichia coli O157, following infection with verocytotoxin-producing E. coli O157. Lett Appl Microbiol 29: 434-436.
136. Karmali MA, Petric M, Bielaszewska M (1999) Evaluation of a microplate latex agglutination method (Verotox-F assay) for detecting and characterizing verotoxins (Shiga toxins) in Escherichia coli. J Clin Microbiol 37: 396-399.
127 137. Murakami J, Kishi K, Hirai K, Hiramatsu K, Yamasaki T, et al. (2000) Macrolides and clindamycin suppress the release of Shiga-like toxins from Escherichia coli O157:H7 in vitro. Int J Antimicrob Agents 15: 103-109.
138. Brian MJ, Frosolono M, Murray BE, Miranda A, Lopez EL, et al. (1992) Polymerase chain reaction for diagnosis of enterohemorrhagic Escherichia coli infection and hemolytic-uremic syndrome. J Clin Microbiol 30: 1801- 1806.
139. Banatvala N, Debeukelaer MM, Griffin PM, Barrett TJ, Greene KD, et al. (1996) Shiga-like toxin-producing Escherichia coli O111 and associated hemolytic-uremic syndrome: a family outbreak. Pediatr Infect Dis J 15: 1008-1011.
140. Guler L, Gunduz K, Ok U (2008) Virulence Factors and Antimicrobial Susceptibility of Escherichia coli Isolated from Calves in Turkey. Zoonoses Public Health 55: 249-257.
141. Muniesa M, Jofre J (1998) Abundance in sewage of bacteriophages that infect Escherichia coli O157:H7 and that carry the Shiga toxin 2 gene. Appl Environ Microbiol 64: 2443-2448.
142. Shelton DR, Karns JS, Park CH (2008) A multiple protocol to improve diagnosis and isolation of Shiga toxin-producing Escherichia coli from human stool specimens. Diagn Microbiol Infect Dis.
143. Schmidt H, Karch H (1996) Enterohemolytic phenotypes and genotypes of shiga toxin-producing Escherichia coli O111 strains from patients with diarrhea and hemolytic-uremic syndrome. J Clin Microbiol 34: 2364-2367.
144. Paton AW, Paton JC (1998) Detection and characterization of Shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaeA, enterohemorrhagic E. coli hlyA, rfbO111, and rfbO157. J Clin Microbiol 36: 598-602.
145. Paton AW, Paton JC (1999) Direct detection of Shiga toxigenic Escherichia coli strains belonging to serogroups O111, O157, and O113 by multiplex PCR. J Clin Microbiol 37: 3362-3365.
128 146. Hilborn ED, Mshar PA, Fiorentino TR, Dembek ZF, Barrett TJ, et al. (2000) An outbreak of Escherichia coli O157:H7 infections and haemolytic uraemic syndrome associated with consumption of unpasteurized apple cider. Epidemiol Infect 124: 31-36.
147. Elizaquivel P, Aznar R (2008) A multiplex RTi-PCR reaction for simultaneous detection of Escherichia coli O157:H7, Salmonella spp. and Staphylococcus aureus on fresh, minimally processed vegetables. Food Microbiol 25: 705- 713.
148. Wagner PL, Acheson DW, Waldor MK (1999) Isogenic lysogens of diverse shiga toxin 2-encoding bacteriophages produce markedly different amounts of shiga toxin. Infect Immun 67: 6710-6714.
149. Koitabashi T, Vuddhakul V, Radu S, Morigaki T, Asai N, et al. (2006) Genetic characterization of Escherichia coli O157: H7/- strains carrying the stx2 gene but not producing Shiga toxin 2. Microbiol Immunol 50: 135-148.
150. Zhang W, Bielaszewska M, Friedrich AW, Kuczius T, Karch H (2005) Transcriptional analysis of genes encoding Shiga toxin 2 and its variants in Escherichia coli. Appl Environ Microbiol 71: 558-561.
151. Kehl KS, Havens P, Behnke CE, Acheson DW (1997) Evaluation of the premier EHEC assay for detection of Shiga toxin-producing Escherichia coli. J Clin Microbiol 35: 2051-2054.
152. Hormansdorfer S, Gareis M, Bauer J, Mayr A (1995) [Determination of Escherichia coli Shiga-like toxins by means of the MTT bioassay]. Z Lebensm Unters Forsch 201: 293-297.
153. Emoto K, Toyama-Sorimachi N, Karasuyama H, Inoue K, Umeda M (1997) Exposure of phosphatidylethanolamine on the surface of apoptotic cells. Exp Cell Res 232: 430-434.
154. Zhao L, Haslam DB (2005) A quantitative and highly sensitive luciferase-based assay for bacterial toxins that inhibit protein synthesis. J Med Microbiol 54: 1023-1030.
155. Lingwood CA (1999) Glycolipid receptors for verotoxin and Helicobacter pylori: role in pathology. Biochim Biophys Acta 1455: 375-386.
129 156. Westphal O, Jann K (1965) Bacterial lipopolysaccharides: extraction with phenol-water and further applications of the procedure. In: Whistler RL, editor. Methods in Carbohydrate Chemistry. New York, NY: Academic Press, Inc. pp. 83-91.
157. Lee CH, Tsai CM (1999) Quantification of bacterial lipopolysaccharides by the purpald assay: measuring formaldehyde generated from 2-keto-3- deoxyoctonate and heptose at the inner core by periodate oxidation. Anal Biochem 267: 161-168.
158. Kitova EN, Kitov PI, Paszkiewicz E, Kim J, Mulvey GL, et al. (2007) Affinities of Shiga toxins 1 and 2 for univalent and oligovalent Pk-trisaccharide analogs measured by electrospray ionization mass spectrometry. Glycobiology 17: 1127-1137.
159. Hord NG (2008) Eukaryotic-microbiota crosstalk: potential mechanisms for health benefits of prebiotics and probiotics. Annu Rev Nutr 28: 215-231.
160. Sherman P, Soni R, Karmali M (1988) Attaching and effacing adherence of Vero cytotoxin-producing Escherichia coli to rabbit intestinal epithelium in vivo. Infect Immun 56: 756-761.
161. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, et al. (1999) Food-related illness and death in the United States. Emerg Infect Dis 5: 607-625.
162. Boerlin P, McEwen SA, Boerlin-Petzold F, Wilson JB, Johnson RP, et al. (1999) Associations between virulence factors of Shiga toxin-producing Escherichia coli and disease in humans. J Clin Microbiol 37: 497-503.
163. Siegler RL, Obrig TG, Pysher TJ, Tesh VL, Denkers ND, et al. (2003) Response to Shiga toxin 1 and 2 in a baboon model of hemolytic uremic syndrome. Pediatr Nephrol 18: 92-96.
164. Valles PG, Pesle S, Piovano L, Davila E, Peralta M, et al. (2005) Postdiarrheal Shiga toxin-mediated hemolytic uremic syndrome similar to septic shock. Medicina (B Aires) 65: 395-401.
165. Neely MN, Friedman DI (1998) Arrangement and functional identification of genes in the regulatory region of lambdoid phage H-19B, a carrier of a Shiga- like toxin. Gene 223: 105-113.
130 166. Wagner PL, Neely MN, Zhang X, Acheson DW, Waldor MK, et al. (2001) Role for a phage promoter in Shiga toxin 2 expression from a pathogenic Escherichia coli strain. J Bacteriol 183: 2081-2085.
167. Safdar N, Said A, Gangnon RE, Maki DG (2002) Risk of hemolytic uremic syndrome after antibiotic treatment of Escherichia coli O157:H7 enteritis: a meta-analysis. Jama 288: 996-1001.
168. Brewer DC, Willis LL (1998) Bloody diarrhea caused by Escherichia coli O157:H7. N C Med J 59: 84-85.
169. Shiomi M, Togawa M, Fujita K, Murata R (1999) Effect of early oral fluoroquinolones in hemorrhagic colitis due to Escherichia coli O157:H7. Pediatr Int 41: 228-232.
170. Proulx F, Turgeon JP, Delage G, Lafleur L, Chicoine L (1992) Randomized, controlled trial of antibiotic therapy for Escherichia coli O157:H7 enteritis. J Pediatr 121: 299-303.
171. Ito T, Akino E, Hiramatsu K (1997) [Evaluation of antibiotics used for enterohemorrhagic Escherichia coli O157 enteritis--effect of various antibiotics on extracellular release of verotoxin]. Kansenshogaku Zasshi 71: 130-135.
172. Nakata K, Tozu T, Hoshikawa Y, Sakai A, Tanaka N, et al. (1997) [Suppressive effect of clarithromycin on the production of verotoxin by E. coli O157]. Kansenshogaku Zasshi 71: 437-442.
173. Takata T, Tabata M, Tsuruoka T, Watabe H (2003) [Activity of fosfomycin against Escherichia coli O157:H7--morphological changes and production of Shiga toxins]. Jpn J Antibiot 56: 691-696.
174. Yoh M, Honda T (1997) The stimulating effect of fosfomycin, an antibiotic in common use in Japan, on the production/release of verotoxin-1 from enterohaemorrhagic Escherichia coli O157:H7 in vitro. Epidemiol Infect 119: 101-103.
175. Pear WS, Miller JP, Xu L, Pui JC, Soffer B, et al. (1998) Efficient and rapid induction of a chronic myelogenous leukemia-like myeloproliferative disease
131 in mice receiving P210 bcr/abl-transduced bone marrow. Blood 92: 3780- 3792.
176. Sherrill JD, Miller WE (2006) G protein-coupled receptor (GPCR) kinase 2 regulates agonist-independent Gq/11 signaling from the mouse cytomegalovirus GPCR M33. J Biol Chem 281: 39796-39805.
177. Allen S, and Marler (1985) Manual of Clinical Microbiology, 4th ed.; In Lennette B, Hausler and Shadomy, editor. Washington, D.C.: American Society for Microbiology.
178. Kale RR, McGannon CM, Fuller-Schaefer C, Hatch DM, Flagler MJ, et al. (2008) Differentiation between structurally homologous Shiga 1 and Shiga 2 toxins by using synthetic glycoconjugates. Angew Chem Int Ed Engl 47: 1265-1268.
179. O'Brien AD, Newland JW, Miller SF, Holmes RK, Smith HW, et al. (1984) Shiga-like toxin-converting phages from Escherichia coli strains that cause hemorrhagic colitis or infantile diarrhea. Science 226: 694-696.
180. Kovach ME, Elzer PH, Hill DS, Robertson GT, Farris MA, et al. (1995) Four new derivatives of the broad-host-range cloning vector pBBR1MCS, carrying different antibiotic-resistance cassettes. Gene 166: 175-176.
181. Fernandez RC, Weiss AA (1998) Serum resistance in bvg-regulated mutants of Bordetella pertussis. FEMS Microbiol Lett 163: 57-63.
182. Kohler B, Karch H, Schmidt H (2000) Antibacterials that are used as growth promoters in animal husbandry can affect the release of Shiga-toxin-2- converting bacteriophages and Shiga toxin 2 from Escherichia coli strains. Microbiology 146: 1085-1090.
183. Lukasova J, Abraham B, Cupakova S (2004) Occurrence of Escherichia coli O157 in raw material and food in Czech Republic. J Vet Med B Infect Dis Vet Public Health 51: 77-81.
184. Oie S, Kawakami M, Kamiya A, Tomita M (2002) In vitro susceptibility of four serotypes of enterohaemorrhagic Escherichia coli to antimicrobial agents. Biol Pharm Bull 25: 671-673.
132 185. Tsuboi I, Ida H, Yoshikawa E, Hiyoshi S, Yamaji E, et al. (1998) Antibiotic susceptibility of enterohemorrhagic Escherichia coli O157:H7 isolated from an outbreak in Japan in 1996. Antimicrob Agents Chemother 42: 431-432.
186. You JY, Moon BM, Oh IG, Baek BK, Li LG, et al. (2006) Antimicrobial resistance of Escherichia coli O157 from cattle in Korea. Int J Food Microbiol 106: 74-78.
187. Muniesa M, Blanco JE, De Simon M, Serra-Moreno R, Blanch AR, et al. (2004) Diversity of stx2 converting bacteriophages induced from Shiga-toxin- producing Escherichia coli strains isolated from cattle. Microbiology 150: 2959-2971.
188. Lingwood CA (1993) Verotoxins and their glycolipid receptors. Adv Lipid Res 25: 189-211.
189. Clarke JT (2007) Narrative review: Fabry disease. Ann Intern Med 146: 425- 433.
190. Auray-Blais C, Cyr D, Ntwari A, West ML, Cox-Brinkman J, et al. (2008) Urinary globotriaosylceramide excretion correlates with the genotype in children and adults with Fabry disease. Mol Genet Metab 93: 331-340.
191. Auray-Blais C, Cyr D, Mills K, Giguere R, Drouin R (2007) Development of a filter paper method potentially applicable to mass and high-risk urinary screenings for Fabry disease. J Inherit Metab Dis 30: 106.
133