In vitro Inhibition of the Expression of Escherichia coli 0157:H7
Genes Encoding the Shiga Like Toxins by Antimicrobial Agents:Potential
Use in the Treatment
of Human Infections
Ali Kanbar, MD
Elias Rahal, MS
Ghassan M.Matar, PhD
Department of Microbiology and Immunology, Faculty of
Medicine, American University
KEY WORDS: Rifampicin, gentamicin, shiga-like toxins
We assessed the in vitro effect of rifampicin and gentamicin
at both the minimal inhibitory concentration (MIC) and the minimal bactericidal
concentration (MBC) on the expression of shiga-like toxins (SLT I and
SLT II) in 7 Escherichia coli O157:H7 strains obtained from outbreaks
in the United States. DNA extraction and polymerase chain reaction (PCR)-based
detection of stx1 and stx2 genes, encoding, respectively, SLT I and
SLT II, were performed on all strains. RNA extraction and RT-PCR of
stx1 and stx2 genes were performed on the same strains incubated at
37˚C for 18 hours with and without rifampicin at its MIC of
4 mg/mL. RT-PCR was also
performed on the same strains incubated with rifampicin at its MIC for
18 hours followed by a 4-hour incubation with rifampicin at its MBC
value (32 mg/mL).Toxins released
were detected using the VTE C-RPLA "SEIKEN" kit in the presence and
absence of rifampicin or gentamicin at their MIC values. Subsequently,
toxin release was measured when E. coli strains were incubated separately,
initially with rifampicin or gentamicin, at their MIC of 4 mg/mL
for 18 hours followed by a 4-hour incubation at the MBC (rifampicin:32
mg/mL or gentamicin: 8 mg/mL). Control strains incubated directly with
rifampicin or gentamicin at MBC values were also performed. Our data
have shown that rifampicin at its MIC inhibited the transcription of
the genes encoding the shiga-like toxins I and II as shown by a negative
RT-PCR of stx1 and stx2. Strains incubated with the MIC of rifampicin
followed by incubation with the MBC showed similar transcription inhibition
profiles. Control strains incubated directly with the MBC of rifampicin
showed a positive RT-PCR. Toxin release was also shown to be decreased
when the strains were incubated with the MIC of rifampicin (12-fold
decrease for SLT I and 16-fold decrease for SLT II) or that of gentamicin
(8-fold decrease for SLT I and II). Similar results were obtained when
the strains were incubated first at the MIC of the drug and subsequently
at its MBC. However, when the strains were incubated directly with the
antibiotic (rifampicin or gentamicin) at its MBC, no decrease in the
level of toxins released was detected. These results may denote a possible
use of these antibiotics in the treatment of E. coli O157:H7 infections.
Additional studies are underway to assess the process in vivo.
Escherichia coli O157: H7 is one of the most notorious
pathogens associated with hemorrhagic colitis leading to bloody diarrhea,
and it has been correlated with the hemolytic uremic syndrome (HUS),
a fatal disease known to be the leading cause of acute renal failure
in children1 and an important cause of stroke and central nervous system
dysfunction. Infection with E. coli O157:H7 seems to be prevalent in
the developed world as well as in the Central African Republic and southern
Africa. Transmission is via food, water, and by direct contact from
person-to-person.2 The main source of infection is undercooked ground
beef, owing to the organism's existence as a colonizer of healthy cattle.3
Infectivity of this organism is quite high; only a low
dose is needed to cause human infection.4 The mechanism of infection
is rather simple. The bacterium attaches to the mucosa of the distal
ileum and colon through its outer membrane protein intimin, encoded
by the eaeA gene located on the locus of enterocyte effacement (LEE).5
Attachment leads to the production of typical attaching and effacing
(A/E) lesions in the host intestinal mucosa.6 The organism produces
one or both shiga-like toxin I (SLT I) and shiga-like toxin II (SLT
II), respectively, encoded by stx1 and stx2 genes on lambdoid lysogenic
bacteriophages. These toxins inhibit protein translation in the host
cell and lead to its death.7-9 Thus, it leads to microvascular damage
in the wall of the intestine.10 Once the blood circulation has been
compromised, SLTs and other bacterial products like lipopolysaccharide
(LPS) gain entrance to the circulation. This increases the levels of
inflammatory mediators, disrupts hemostatic systems, and injures several
cell types, culminating in a myriad of symptoms called the hemolytic
uremic syndrome (HUS). This progression usually takes place in patients
at the extremes of age. The disease displays a triad of hemolytic anemia,
thrombocytopenia, and renal failure. However, other organs may be affected,
including the brain.11 The mortality rate is 5%. Conversely, 5% of those
who survive the acute phase progress to severe sequelae such as end-stage
renal disease or permanent neurologic damage.1,12
Proven risk factors for progression from colitis to HUS
include infection with strains that produce SLT II, which has greater
toxicity than SLT I,13 use of an antimotility agent, bloody diarrhea,
fever, vomiting, elevated blood leukocyte counts, extremes of age, and
female gender.4 Whether the use of antimicrobial agents is a risk factor
or not remains debatable.
Treatment of the infection with E. coli O157:H7 is mainly
based on rehydration and supportive therapy. Antimicrobial treatment
is not currently recommended and may even be harmful.14 In vitro trials
have shown that the use of certain antibiotics, mainly the quinolones
trimethoprim and furazolidone, appears to augment the production of
SLTs from E. coli O157:H7.15 Thus, an agent that is effective in combating
the pathogen without exacerbating the disease or triggering detrimental
sequelae is needed. This is possible by inhibiting toxin synthesis and
release before administration of a bactericidal dose of an antimicrobial
agent to avoid unwanted sequelae that can be fatal. Because rifampicin
acts on inhibition of gene transcription, we assessed the in vitro effect
of rifampicin, at a minimal inhibitory concentration, on transcription
inhibition of genes encoding SLT I and SLT II in E. coli O157:H7. Likewise,
we assessed the in vitro effect of gentamicin, known inhibitor of protein
synthesis at the level of translation, in inhibiting the production
of shiga-like toxins. Inhibition of production of these proteins by
rifampicin or gentamicin may constitute a first step in an antimicrobial
treatment regimen of E. coli 0157:H7 infections.
MATERIALS AND METHODS
E. coli O157:H7 Strains
Seven E. coli O157:H7 strains were obtained from the Centers
for Disease Control and Prevention (CDC). They were isolated during
outbreaks of hemorrhagic colitis in the United States and all have distinct
pulsed field gel electrophoresis profiles.
DNA Extraction and PCR
Total DNA was extracted from the 7 E. coli O157:H7 strains
using the GFXTM Genomic Blood DNA Purification kit (AmershamPharmacia
Biotech, Piscataway, NJ). PCR was performed on DNA extracts to amplify
the stx1 and stx2 genes, in separate reactions, using previously published
primers.16 PCR amplification was performed in 100 mL
reaction mixtures consisting of 10 mL
DNA (2 mg/mL) and 90 mL of the amplification
mix containing 16 pmol of each primer, 200 mM concentrations of each deoxynucleoside triphosphate,
10 mL of PCR buffer (Amersham
Pharmacia Biotech, Piscataway, NJ) and 2.5 U of Taq DNA polymerase (Amersham
Pharmacia Biotech, Piscataway, NJ). A thermal cycler (PTC-100, MJ Research
Inc. Watertown, MA) was used for amplification for 35 cycles. Each cycle
consisted of denaturation at 95˚C for 15 seconds, primer annealing
at 60˚C for 45 seconds and extension at 72˚C for 45 seconds.
The cycles were followed by a final extension step at 72˚C for
Amplicons of the stx 1 (475 Bp) and stx2 (863 Bp) genes
were electrophoresed on 1% agarose (Sigma. St. Louis, MO) gel in 1x
Tris-borate-EDTA buffer at 117 V for 45 minutes. Ethidium bromide (Sigma.
St. Louis, MO) of 1 ug/ml was incorporated into the gel for staining.
Amplicons were detected under UV light and photographed with type 667
Determination of MIC and MBC of rifampicin and gentamicin
for E. coli O157: H7. Rifampicin (Sigma. St. Louis, Mo.) was dissolved
in methanol whereas gentamicin (Sigma.St.Louis,Mo) was dissolved in
sterile distilled water and both were serially diluted in sterile distilled
water in a series of tubes. A calibrated amount of 5x105 CFU of E. coli
O157: H7 were added to each tube and the minimum inhibitory concentrations
(MIC) of the drugs for the 7 strains were determined according to NCCLS
recommendations.17 On each supra MIC tube, a 1:100 dilution was performed
and subcultured on a SMAC agar plate. The antimicrobial concentration
that induced a 99.99% killing was considered to be the minimal bactericidal
RNA Extraction and RT-PCR
RNA was extracted using the Rneasy Mini Kit (QIAGEN Gmbh,
Germany) according to manufacturer's specifications. RT-PCR was performed
using The Ready-To-Go kit (Amersham Pharmacia Biotech) for c-DNA synthesis
according to manufacturer's specifications, primers and PCR conditions
as described previously for PCR. Extracted RNA of each strain was used
as a control reaction in PCR to rule out DNA contamination in RNA extracts.
Amplicons were detected on ethidium bromide-stained gel and photographed
as described previously for PCR.
Determination of The Effect of
Rifampicin on RT-PCR
Determining the lowest concentration of rifampicin that
inhibits the transcription of stx1 and stx2 genes in E. coli O157:H7
was performed as follows: a single bacterial colony from a sorbitol
MaConkey Agar plate was inoculated in 5 mL trypticase soy broth (TSB)
(either rifampicin-free or containing rifampicin at subMIC concentrations)
and incubated at 37˚C for 18 hours. The cells, being in the log
phase, were then adjusted to 3.5 McFarland (equivalent to 109 CFU/mL
using a Densimat Densitometer for inoculum standardization [BioMerieux,
Marcy, L'Etoile, France]) and then subjected to the RNA extraction procedure
as described previously. When RNA extraction was to be performed on
the strain incubated in MIC concentrations of rifampicin, 109 CFU/mL
of the strain was incubated at 37˚C for 18 hours in TSB containing
the MIC concentration of the
The procedure was first performed on a single randomly
selected strain. RT-PCR for the two genes was performed on extracted
RNA as previously described. The lowest concentration of rifampicin
that inhibited transcription of both genes as shown by a negative RT-PCR
in one E. coli 0157:H7 strain, was applied to all 7 strains as previously
described. RT-PCR was also performed on the same strains when incubated
initially with the MIC of rifampicin for 18 hours followed by addition
of the drug to the culture to attain the MBC (32 mg/mL) and incubated for 4 more hours.
Reversed Passive Latex Agglutination (RPLA) Assay of SLT
I and SLT II
kit (Denka Seiken, LTD., Tokyo, Japan) was used as a confirmatory test
for SLT I and II release inhibition, by E. coli O157:H7 strains incubated
in the presence of rifampicin and gentamicin after adjustment to a concentration
of 3.5 McFarland. The procedure was performed according to manufacturer's
specifications on strains grown for 16 to 18 hours at 37˚C both
in rifampicin-free TSB and in TSB cultures at a concentration of 4 mg/mL of rifampicin. That concentration inhibited
mRNA transcription of stx1 and stx2 genes by RT-PCR. The same procedure
was performed on strains incubated with gentamicin at an MIC of 4 mg/mL. The second step was to assess the effect of
combining the MIC and MBC in the process. The strains were incubated
for 16 to 18 hours with the antibiotic at its MIC, followed by the same
antibiotic brought to its MBC (32 mg/mL for rifampicin and 8
mg/mL for gentamicin). The cultures were incubated
for 4 more hours, followed by detection of the toxins by RPLA.
All strains were positive by PCR for the stx1 and stx2
genes. This step was performed to verify that the genes were not lost
due to subculturing, as is known to frequently occur for stx1 and stx2
genes.18 RT-PCR revealed that all 7 strains transcribed stx1 and stx2
of stx2 was inhibited in all strains at rifampicin concentrations that
were < 0.5 mg/mL,
as revealed by a negative RT-PCR. Although there was a marked decrease
in band intensity for stx1 at concentrations <
2 mg/mL, complete inhibition
of transcription occurred at a concentration of 4 mg/mL
of the antimicrobial agent.
Because the concentration of 4 mg/mL
of rifampicin was the lowest to potently inhibit transcription of both
genes in all 7 strains, this concentration was subsequently used on
all strains. Data have shown that transcription of both genes was inhibited
at a concentration of 4 mg/mL
in all 7 E. coli 0157:H7 strains (Figures 1 and 2). Bacterial strains
incubated with MIC followed by MBC also showed a negative RT-PCR for
stx 1 and stx2 (data not shown). The MIC of gentamicin was also found
to be equal to 4 mg/mL. The
MBC values for rifampicin and gentamicin were found to be 32 and 8 mg/mL, respectively.
To confirm data obtained
by RT-PCR, the RPLA assay was performed on culture supernatants of E.
coli O157:H7 strains grown in both rifampicin-free TSB and in TSB containing
4 mg/mL of rifampicin. This assay has shown an average
of 12-fold decrease in titers of SLT I and 16-fold decrease in titers
of SLT II in TSB containing rifampicin as compared with SLT I and II
controls in drug-free TSB. Similar data were obtained when strains were
incubated with the MIC of rifampicin for 18 hours followed by the MBC
for 4 hours. However, when the strains were incubated with rifampicin
at its MBC of 32 mg/mL directly, no inhibition of toxin production
was detected (Figures 3 and 4).19
Because gentamicin acts at the mRNA translation level,
toxin inhibition was assessed using the VTEC-RPLA kit only. When the
strains were incubated with gentamicin at an MIC of 4 mg/mL,
or when incubated with MIC for 18 hours followed by a 4-hour incubation
period with an MBC of 8 mg/mL,
results showed an 8-fold decrease in the titers of SLT I and SLT II.
Similar to the results obtained with rifampicin, there were no decreases
in the titers of both toxins when the strains were incubated directly
with gentamicin at its MBC (Figures 5 and 6).
Our data revealed that the MICs of rifampicin and gentamicin
for E. coli O157:H7 is not only capable of inhibiting growth of the
bacterium but also of decreasing the release of toxins necessary for
virulence and pathogenicity. A rifampicin or gentamicin concentration
of 4 mg/mL appears effectively
capable of inhibiting the production of SLT I and II. RPLA assay revealed
that incubation of the organisms with the MIC of rifampicin or gentamicin
showed a marked decrease in the titer of SLTs released when compared
with controls not incubated in the presence of the antimicrobial agent.
The inhibited genes are necessary for the infectious phenotype.
The role of SLTs extends beyond damaging the microvasculature of the
intestinal wall because their cellular receptor, globotriosyl ceramide,
is distributed over a plethora of cell types. These include enterocytes,
renal, aortic, and brain endothelial cells, glomerular endothelial cells,
mesangial cells, proximal and distal renal tubular epithelial cells,
monocytes and cells derived from the monocytic cell line, astrocytoma
cells, lung epithelial cells, polymorphonuclear cells, erythrocytes,
platelets, and B lymphocytes.20 The effect of SLTs on these cells is
lethal. Therefore, prevention of disease progression is of ultimate
importance. Otherwise, severe sequelae, such as hemolytic uremic syndrome
or HUS, may result and consequently complicate the case.
Treatment of this infection at the diarrheal phase with
bactericidal doses of an antibiotic may kill the pathogen and lead to
lysis of the bacterial cells. In fact, this was proven by our results,
because when the strains were incubated directly with the MBC of the
antibiotic, the titer of SLT I and SLT II was very high. This can be
explained by the fact that bacteria have presynthesized toxins that
can be released on cell lysis. The toxins will damage the microvasculature
of the colon wall and gain entrance, with other bacterial toxins and
antigens, to the host blood stream.11 Likewise, an alternative manner
by which antimicrobial treatment could aggravate the disease is by triggering
an SOS response within the pathogen can be considered. DNA damage to
the bacterium induces expression of the toxin genes on the bacteriophages
hosting stx1 or stx2. An SOS response is also followed by bacterial
lysis and release of toxins and antigens.15
Our study attempted to separately assay the in vitro inhibitory
effect of rifampicin and gentamicin on the production of SLTs. Our goal
was to assess the effectiveness of these antibiotics for future implementation
in treatment. Rifampicin exerts
its effect by binding the bacterial DNA-dependent RNA polymerase and
blocking transcription.21 Although gentamicin binds the 30S ribosomal
subunit irreversibly, thus inhibiting protein synthesis, our data have
shown that transcription of stx1 occurred at several rifampicin concentrations
that inhibited transcription of the stx2 gene. This may reflect differences
in affinities of the gene promoters of the toxins to the RNA polymerase,
certain structural properties of the enzyme itself, or some other factors.
This was not the case for gentamicin, which generated similar inhibition
profiles for both SLTs. This may due to the fact that a different mechanism
of inhibition is implicated with no differences in affinities to the
Monotherapy with rifampicin results in rapid resistance
via one-step mutations that alter the subunit structure of the bacterial
RNA polymerase.22 Consequently, if rifampicin is to be recommended,
then it should not be used alone but incorporated into a course of treatment
that involves other drugs like gentamicin. Additional studies are needed
to implement a therapeutic strategy for infection caused by this organism.
The authors would like to thank Dr. B. Swaminathan, Food
borne and Diarrheal Diseases Branch, Centers for Disease Control and
Prevention, for provision of primers and E. coli strains.
1. Boyce TG, Swerdlow DL, Griffin PM. Escherichia
coli O157 and the hemolytic-uremic syndrome. N Engl J Med 333:364-368,
2. Griffin PM, Tauxe RV. The epidemiology of infections
caused by Escherichia coli 0157:H7, other enterohemorrhagic E. coli,
and the hemolytic uremic syndrome. Epidemiol Rev 13:60-98, 1991.
3. Griffin P. Escherichia coli O157:H7 and other
enterohemorrhagic Escherichia coli. In: Infections of the Gastrointestinal
Tract. Blaser MJ, Smith PD, Ravdin JI, et al. (eds). New York: Raven
Press, 739-761, 1995.
4. Mead PS, Griffin PM. Escherichia coli O157:H7.
Lancet 352:1207-1212, 1998.
5. Kresse AU, Schulze K, Deibel C, et al. Pas, a
novel protein required for protein secretion and attaching and effacing
activities of enterohemorrhagic Escherichia coli. J Bacteriol 180:4370-4379,
6. DeVinney R, Stein M, Reinscheid D, et al. Enterohemorrhagic
E.coli O157:H7 produces Tir, which is translocated to the host cell
membrane but is not tyrosine phosphorylated. Infect Immunol 67:2389-2398,
7. Smith HR, Scotland SM. Verocytotoxin-producing
strains of Escherichia coli. J Med Microbiol 26:77-85, 1988.
8. Jackson MP. Structure function analyses of shiga
toxin and the shiga-like toxins. Microb Pathog 8:235-242, 1990.
9. Hofmann SL. Southwestern internal medicine conference:
Shiga-like toxins in hemolytic-uremic syndrome and thrombotic thrombocytopenic
purpura. Am J Med Sci 306:398-406, 1993.
10. Tesh VL, O'Brien AD. Adherence and colonization
of mechanisms of enteropathogenic and enterohemorrhagic E. coli. Microb
Pathog 12:245-254, 1992.
11. Siegler RL. The hemolytic uremic syndrome. Pediatr
Nephrol 42:1505-1529, 1995.
12. McLigeyo SO. Haemolytic uraemic syndrome: A review.
E African Med J 76:148-153, 1999.
13. Tesh VL, Burris JA, Owens JW, et al. Comparison
of the relative toxicities of shiga-like toxins type I and type II for
mice. Infect Immun 61:3392-3402, 1993.
14. Qadri SMH, Kayali S. Enterohemorrhagic Escherichia
coli: A dangerous food-borne pathogen. Postgrad Med 103:179-187, 1998.
15. Kimmit PT, Harwood CR, Barer MR. Toxin gene expression
by shiga toxin-producing Escherichia coli: The role of antibiotics and
the bacterial SOS response. Emerg Infect Dis 6:458-465, 2000.
16. Olsvik O, Rimstad E, Hornes E, et al. A nested
PCR followed by magnetic separation of amplified fragments for detection
of Escherichia coli shiga-like toxin genes. Mol Cell Probes 5:429-435,
17. NCCLS. Performance standards for antimicrobial
susceptibility testing. Twelfth informational supplement NCCLS document
M100-S12. Wayne PA: NCCLS, 2002.
18. Karch H, Meyer T, Russman H, Heesemann J: Frequent
loss of shiga-like toxin genes in clinical isolates of Escherichia coli
upon subcultivation. Infect Immun 60:3464-3467, 1992.
19. Matar GM, Rahal E. Transcription inhibition of
shiga-like toxins and intimin genes in Escherichia coli O157:H7 as partial
treatment of the infection. Ann Trop Med Parasitol In press, 2003.
20. Meyers KEC, Kaplan BS. Many cell types are shiga
toxin targets. Kidney Int 57:2650-2651, 2000.
21. Naryshkina T, Mustaev A, Darst SA, Severinov K.
The subunit of Escherichia coli RNA polymerase is not required for interaction
with initiating nucleotide but is necessary for interaction with rifampicin.
J Bio. Chem 276:13308-13313, 2001.
22. Reese RE, Betts RF. Rifampin. In: Handbook of
antibiotics. Reese RE, Betts RF (eds). Boston, MA: Little, Brown and
Company; 359-369, 1993.
Figure 1. Effect of rifampicin on RT-PCR of the
stx1 gene in the 7 E. coli O157: H7 strains. Lane 1: 100 bp ladder;
lane 2: negative control; lanes 3, 5, 7, 9, 11, 13, 15: amplicons of
the strains when not exposed to rifampicin; lanes 4, 6, 8, 10, 12, 14,
16: disappearance of the amplicons in the strains when incubated in
4 ug/ml of rifampicin.
Figure 2. Effect of rifampicin on RT-PCR of the
stx2 gene in the 7 E. coli O157: H7 strains. Lane 1: 100 bp ladder;
lane 2: negative control; lanes 3, 5, 7, 9, 11, 13, 15: amplicons of
the strains when not exposed to rifampicin; lanes 4, 6, 8,10, 12, 14,
16: disappearance of the amplicons in the strains when incubated in
4 ug/ml of rifampicin.
Figure 3. Effect of rifampicin on the release of
Shiga-Like Toxin I,as detected by the VTEC-RPLA kit.
Figure 4. Effect of rifampicin on the release of
Shiga-Like Toxin II,as detected by the VTEC-RPLA kit.
Figure 5. Effect of gentamicin on the release of
Shiga-like toxin I using the VTEC-RPLA kit.
Figure 6. Effect of gentamicin on
the release of Shiga-like toxin II using the VTEC-RPLA kit.