Chapter 27

Foodborne Gastroenteritis Caused
by Escherichia coli

Escherichia coli was established as a foodborne pathogen in 1971 when imported cheeses turned
up in 14 American states that were contaminated with an enteroinvasive strain that caused illness in
nearly 400 individuals. Prior to 1971, at least five foodborne outbreaks were reported in other countries,
with the earliest being from England in 1947. As a human pathogen, evidence suggests that it was
recognized as a cause of infant diarrhea as early as the 1700s.60 Since the meatborne outbreaks in the
United States of 1982 and 1993, the status of this bacterium as a foodborne pathogen is unquestioned.
Escherichia coli as an indicator of fecal contamination is discussed in Chapter 20, culture and isolation
methods are covered in Chapter 10, and molecular and bioassay methods for its detection are covered
in Chapters 11 and 12. For a more detailed history of E. coli 0157:H7, see reference 69.

SEROLOGICAL CLASSIFICATION
Pathogenic strains of Escherichia are serologically typed in the same way as other Enterobacteriaceae, and the procedure is described in Chapter 11. For E. coli, over 200 O serotypes have been
recognized. Because the flagellar proteins are less heterogeneous than the carbohydrate side chains
that make up the O groups, considerably fewer H antigenic types exist (around 30).

THE RECOGNIZED VIRULENCE GROUPS
Based on disease syndromes and characteristics, and also on their effect on certain cell cultures and
serological groupings, five virulence groups of E. coli are recognized: enteroaggregative (EAggEC),
enterohemorrhagic (EHEC), enteroinvasive (EIEC), enteropathogenic (EPEC), and enterotoxigenic
(ETEC).
Enteroaggregative E. coli (EAggEC)
This group (also designated enteroadherent) is related to EPEC but the aggregative adherence
displayed by these strains is unique. Strains exhibit a “stacked-brick-type” of adherence to HEp-2

637

638

Modern Food Microbiology

Table 27–1 Some of the O Serotypes Found Among the Five
Virulence Groups
EAggEC
3
4
6
7
17
44
51
68
73
75
77
78
85
111
127
142
162

EHEC

EIEC


EPEC

ETEC

2
5
6
4
22
26
38
45
46
82
84
88
91
103
113
104
111
116
118
145
153
156
157
163

28ac

29
112a
124
135
136
143
144
147
152
164
167

18ab
19ac
55
86
111
114
119
125
126
127
128ab
142
158

6
8
15
20

25
27
63
78
80
85
101
115
128ac
139
141
147
148
149
153
159
167

Note: Some serotypes (e.g., 111) are listed under more than one virulence group.

cells, and carry a 60-MDa plasmid that is needed for the production of fimbriae that are responsible
for the aggregative expression, and for a specific outer membrane protein (OMP). Antibodies raised
against the OMP of a prototype strain prevented adherence to HEp-2 cells.20 An EAggEC DNA probe
has been constructed by using a 1.0-kilobase (kb) fragment from the 60-MDa plasmid of the prototype
strain (03:H2), and it was found to be 99% specific for these strains.4 Some EAggEC strains produce a
heat-stable enterotoxin (ST), which has been designated EAST1.75 The plasmid-borne gene for EAST1
is astA, which encodes a 38-amino-acid molecule in contrast to estA, which encodes the 72 amino acid
enterotoxin STa (see reference 75). They produce an enterotoxin/cytotoxin that is about 108 kDa, and
it is located on the large virulence plasmid. The distinguishing clinical feature of EAggEC strains is a
persistent diarrhea that lasts >14 days, especially in children. These strains are not the primary cause

of traveler’s diarrhea.15
It is unclear whether members of this group are foodborne pathogens. Some of the serotypes in
which EAggEC strains have been found are listed in Table 27–1. Two serotypes that were designated as prototype are O3:H2 and O4:H7, and one serotype (O44) contains both EAggEC and EPEC
strains.76


Foodborne Gastroenteritis Caused by Escherichia coli

639

Enterohemorrhagic E. coli (EHEC)
These strains are both similar and dissimilar to EPEC strains. They are similar to EPEC in their
possession of the chromosomal gene eaeA (or one that is similar) and in the production of attachment–
effacement lesions (see the subsection on EPEC). In contrast to EPEC, EHEC strains affect only
the large intestine (in piglet models) and produce large quantities of Shiga-like toxins (SLT, Stx, see
below). EHECs produce a 60-MDa plasmid that encodes fimbriae that mediate attachment to culture
cells, and they do not invade HEp-2 or INT407 cell lines, although some strains have the ability to
invade some human epithelial cell lines.64 Some EHEC strains produce curli fimbriae that facilitate
attachment of cells to surfaces.
The Toxins
Shigella dysenteriae produces a potent toxin that is referred to commonly as Shiga toxin (after K.
Shiga who first isolated and studied the organism). The toxins of EHEC strains of E. coli have been
referred to as Shiga-like toxins (verotoxin, verocytotoxin) and the two prototypes as SLT-I and SLT-II.
However, new terminology has been applied, and what was once SLT-I is now Stx1 and the former
SLT-II is Stx2.9 The genes for Stx1 and Stx2 are encoded by temperate bacteriophages in some EHEC
strains. Stx1 differs from Stx (Shiga-toxin) by three nucleotides and one amino acid, and is neutralized
by antibodies to Stx. Stx1 and Stx2 are differentiated by a lack of cross-neutralization by homologous
polyclonal antisera, and by a lack of DNA–DNA cross-hybridization of their genes under conditions
of high stringency.9 Both Stx2 and Stx2e (formerly SLT-IIv, VTe) are neutralized by antisera against
Stx2 but not by anti-Stx toxin. Stx2e is a variant of Stx2 that is more toxic to Vero cells than HeLa cells,

and like Stx its gene is chromosomal.52,63 All Stxs are cytotoxic for Vero cells and lethal for mice,
and produce positive rabbit ileal loop responses. All Stxs consist of a single enzymatically active A
subunit and multiple B subunits. Stx-sensitive cells possess the toxin receptor, globotriaosylceramide
(Gb3 ), and sodium butyrate appears to play a role in sensitizing cells to Stxs.53 Once toxins bind to
Gb3 , internalization follows with transport to the trans-Golgi network. Once inside host cells, the A
subunit binds to and releases an adenine residue from the 28S ribosomal RNA (rRNA) of the 60S
ribosomal subunit and this inhibits protein synthesis. The B subunits form pentamers in association
with a single A subunit and, thus, they are responsible for the binding of the toxin to the neutral
glycolipid receptors. Although serotype O157:H7 is the prototype for this group, Stxs are produced by
a number of serotypes, some of which are listed in g 27–1. For reasons that are not clear, Stx2 appears
to be more significant in the etiology of hemorrhagic colitis (HC) and hemolytic uremic syndrome
(HUS) than Stx1.63
Growth and Stx Toxin Production
The nutritional requirements of Stx-producing strains are not unlike those for most other E. coli
strains (see Chapter 20). Reports on the effect of temperature on Stx production vary. In an early
study, temperature was found to have no effect on Stx1 synthesis, whereas iron repressed synthesis.91
In another study, Stx production occurred at all temperatures that supported growth (Figure 27–1),
although less toxin was found when cells were grown at 21◦ C than at 37◦ C even though cell numbers
were similar.1 In a ground roasted beef slurry, strain O157:H7 was found to produce Stx at either
21◦ or 37◦ C within 24 hours.1 In an earlier study employing an O157:H7 strain in milk and fresh
ground beef, Stx1 was found to be produced at maximum levels at 37◦ C in both products but at


640

Modern Food Microbiology

Figure 27–1 Growth and verotoxin production by E. coli A9124-1 at various temperatures: (a) = 37◦ C; (b) =
12◦ C; (c) = 10◦ C. Source: From Palumbo et al.,68 copyright c 1995 by International Association of Milk, Food,
and Environmental Sanitarians.



Foodborne Gastroenteritis Caused by Escherichia coli

641

25◦ or 30◦ C only traces were found.90 Highest levels were attained in fresh ground beef (452 ng/g),
whereas in milk with agitation for 48 hours, the highest level found was 306 ng/ml. No toxin was
detected in ground beef held at 8◦ C for up to 14 days. There is no evidence that preformed Stx plays
any role in diseases caused by EHEC strains. The mean optimum growth temperature for 20 E. coli
0157:H7 strains was 40.2◦ C in Mueller–Hinton broth compared to a mean of 41.7◦ C for non-0157:H7
strains.30
With regard to minimum temperature for Stx production in brain heart infusion (BHI) broth, 4 of
16 strains grew at 8◦ C but not at 5◦ C, whereas 12 of these grew at 10◦ C but not at 8◦ C.68 Three of 16
strains increased 1000-fold in numbers in 4–6 days at 10◦ C, and as noted above, Stx was produced
at all temperatures that supported growth.68 Concentrations of Stx1 were 63 and 85 ng/ml of slurry
following incubations at 21◦ and 37◦ C, respectively.1 Unlike most strains of E. coli, the O157:H7
strains do not grow in EC (Escherichia coli) medium at 44.5◦ C and their maximum in EC medium is
around 42◦ C.71
Effect of Environmental and Physical Agents
Interest in the acid sensitivity of EHEC strains increased following an outbreak traced to fresh-pressed
apple cider.6 That product had a pH range of 3.7–3.9. In one study, EC O157:H7 survived for up to
56 days at pH ≥4.0 using tryptic soy broth and various acids for pH adjustment.16 In another study,
in which the pH of Luria broth was adjusted with HCl, no loss of viability of an EC O157:H7 strain
was seen for at least 5 hours at pH 3.0–2.5 at 37◦ C.5 In a more detailed study using apple ciders with
pH values between 3.6 and 4.0 and EC O157:H7 inocula of 102 –105 , cells survived for 2–3 days at
25◦ C.98 At 8◦ C, a 105 /ml inoculum increased only about 1 log over 12 days and survived for 10–31
days at this temperature. Although potassium sorbate was only minimally effective, sodium sorbate
shortened survival time at 8◦ C to 2–10 days, and to 1–2 days at 25◦ C.98 Growth of EC O157:H7 was
demonstrated in trypticase soy broth at pH 4.5 when HCl was used, but no growth occurred at this pH

when lactic acid was used—the minimum was pH 4.6.29
In a study of EC O157:H7 survival in commercial mayonnaise, survival was noted for 35 days for
products stored at 5◦ or 7◦ C, but cells could not be detected after 72 hours when stored at 25◦ C.89 The
mayonnaise had a pH of 3.65 and the inoculum was ∼107 colony-forming units (cfu)/g. With inocula
as high as log 6.23/g in commercial mayonnaise and storage at 5◦ , 20◦ , or 30◦ C, strain 0157:H7 did
not grow and was approaching undetectable levels after 93 days at 5◦ C.37 In another study, ≥6 log cfu
of an EHEC strain was inoculated into five commercial real-mayonnaise-based and reduced-calorie
and/or fat mayonnaise dressings and stored at 25◦ C.24 The pH ranged from 3.21 to 3.94, and the
products with pH <3.6 rapidly inactivated EHEC, producing a ≥7 log cfu decrease in ≤1–3 days.
EHEC cells have been shown to have increased survival in acidic foods if they are first cultured in an
acidic environment at around a pH of 5.0.50 Two EHEC strains survived for 18 days at 4◦ C in four
varieties of ground apples and the final pH of the four ranged from 3.91 to 5.11.27 Fallen apples may be
contaminated by EHEC strains in pastures, and also by contaminated fruit flies.40 More information
on the acid tolerance of E. coli strains is presented in Chapter 22.
With regard to salt tolerance of an EC O157:H7 strain, 4.5% NaCl in broth caused a threefold
increase in doubling time, whereas at 6.5%, a 36-hour lag was noted with a generation time of 31.7
hours.29 These investigators found that no growth occurred at ≥8.5% NaCl. In the same study, the EC
O157:H7 survived sausage fermentation but did not grow when stored at 4◦ C for 2 months following
inoculation at a level of 4.8 × 104 .29
The thermal resistance of EHEC strains is not unlike that of most Gram-negative bacteria, and,
in fact, these strains appear to be more heat sensitive than most salmonellae. A recent study found