Professional Documents
Culture Documents
2018 - Evaluation of MPCR For Detection of The Top 7 STEC Serogroups
2018 - Evaluation of MPCR For Detection of The Top 7 STEC Serogroups
2018 - Evaluation of MPCR For Detection of The Top 7 STEC Serogroups
6, 2018
Background: Ready-to-eat (RTE) meats, fruits, and recent decades, especially as STEC serotype, O157:H7, has
vegetables contaminated by Shiga toxin producing become recognized as a food-borne pathogen (1). Additionally,
Escherichia coli (STEC) raise serious concerns as non-O157 strains are increasingly detected as important
because they are often consumed directly without food-borne pathogens in outbreaks, six other STEC serogroups
further processing. Objective: To evaluate a multiplex (O26, O45, O103, O111, O121, and O145) are now being
PCR for the detection of STEC across food categories. regulated by the U.S. Department of Agriculture (USDA), Food
Methods: Samples (25 g) from seven RTE meat and Safety and Inspection Service (FSIS) in raw beef products (2).
nine fruit and vegetable matrices were inoculated with A number of methods have been developed and evaluated by
each of seven STEC (O157:H7, O26, O121, O145, O45, research groups, such as for the detection of STEC O104 in
O103, O111) strains targeting 10 CFU/25 g, enriched sprouts (3); O157:H7 and non-O157 in leafy green vegetables,
in 225 mL of modified tryptone soya broth (mTSB), cilantro, sprouts and milk (4, 5); O157 in various food samples
and tested by a multiplex real-time PCR for stx and (6); O157:H7 in alfalfa sprouts (7); non-O157 in pure culture and
eae genes, following U.S. Department of Agriculture food (8), and the USDA Microbiology Laboratory Guidebook
(USDA) Food Safety and Inspection Service (FSIS) (MLG) 5B method for raw meat (9). However, despite these
Microbiology Laboratory Guidebook (MLG) 5B, efforts, information on the evaluation of methods for the
which was originally validated for meat products detection of these seven STEC serogroups in different food
and environmental sponge. Results: The mTSB was categories remains limited (10). To our knowledge, methods for
successful at enriching for STEC in RTE meat, fruit, the detection of both O157 and non-O157 STEC across broad
and vegetable matrices, except for sprouts; however, food categories have not been reported. Fruits, vegetables, and
mEHEC resulted in successful enrichment of target ready-to-eat (RTE) meats contaminated by STEC raise a more
organisms in mung bean sprouts. Suppression of eae serious concern to public health, unlike raw beef, which is
results by stx in PCR was observed in six fruit and cooked prior to consumption. The objective of this study was
vegetable matrices. Conversely, suppression of stx to develop a screening method for the detection of these STEC
gene by eae was not observed. PCR solely targeting serogroups in fruits, vegetables, and RTE meats prior to bacterial
eae is recommended if a fruit or vegetable sample culture. This work presents an adopted multiplex real-time PCR
tested positive for stx and negative for eae. Despite screening method that can be used to detect all seven regulated
the significant effect from food matrix, strain, and STEC serogroups across different food categories.
experimental batch, the cycle threshold of PCR was
<30 in inoculated samples, and mostly 30–42 and up Materials and Methods
in uninoculated samples. Conclusions: The multiplex
PCR can be adopted for detection of all seven Experimental Design
regulated STEC in RTE meat, fruit and vegetable
matrices after validation with cut off value selected
Seven RTE meats (cooked chicken, cooked deli, cooked
and justified based on real samples.
ham, cooked sausage, dry cured ham, salami, and sausage) and
nine fruit and vegetable (apple cider, broccoli, carrot, lettuce,
C
raspberry, spinach, mung bean sprouts, strawberry, and tomato)
ontamination from Shiga toxin producing Escherichia samples were inoculated with each of the seven STEC strains
coli (STEC) is a significant and highly relevant concern (Table 1) in a factorial design. The experiments were carried
in food safety. The development of methods for the out in eight batches. Each batch consisted of one blank control
detection of STEC has been the topic of significant interest in sample and two inoculated replicate samples from each of the
16 food matrices, with only one STEC strain employed for
artificial inoculation. Seven batches were conducted to cover
Received January 9, 2018. Accepted by AH March 9, 2018. each of the seven STEC strains. An additional batch of mung
This research is funded by Ontario Ministry of Agriculture, Food
bean sprout samples was inoculated with the seven STEC
and Rural Affairs (OMAFRA) via University of Guelph–OMAFRA
Partnership. strains individually and enriched in a proprietary media mEHEC
Corresponding author’s e-mail: agao@uoguelph.ca (BioControl) when it was observed that modified tryptone soya
DOI: https://doi.org/10.5740/jaoacint.18-0010 broth (mTSB; Oxoid) did not provide successful enrichment of
Gao et al.: Journal of AOAC International Vol. 101, No. 6, 2018 1829
Table 1. Shiga toxin-producing Escherichia coli strains of a Ct value, “undetermined” is given to a PCR reaction result
used for inoculation of food samples when there is not enough target DNA present. A value of 45 was
assigned to the Ct value of stx gene when it was “undetermined”
No. Serotype Strain eae stx1 stx2
in PCR. Otherwise, the samples that were “undetermined”
1 O26:H11 EC910114 + + – by PCR would be excluded by SAS in data analysis because
2 O145:NM EC19960923 + – + “undetermined” is not a recognizable quantitative value. The
3 O45:H2 EC19940040 + + – effect of the food matrix and STEC strain and their interaction
on the Ct value in PCR was analyzed by mixed-model analysis
4 O103:H2 EC930004 + + –
of variance (ANOVA) in SAS 9.3 with Enterprise Guide 4.3
5 O111:NM STXX + + +
(SAS Institute Inc., Cary, NC). Uninoculated samples and
6 O121:H19 EC19990161 + – + inoculated samples were analyzed separately. The effect of the
7 O157:H7 ATCC 43889 + + + food matrix and strains was based on inoculated samples only.
A significant difference was accepted at P < 0.05. Scatterplots
of stx gene Ct value from inoculated and uninoculated samples
STEC in sprouts. Randomly selected samples including both
for individual matrix and for individual STEC strains, as well
uninoculated and inoculated samples were tested by PCR in
as the scatterplot of stx gene Ct value between duplicate tests,
duplicate to determine the uncertainty of PCR detection.
were created in Microsoft Excel 2010. The uncertainty of PCR
detection, or the repeatability SD of Ct value, was estimated
Sample Preparation
based on the samples tested in duplicate following a procedure
described in P19-CALA Measurement Uncertainty Policy (12).
Samples were purchased from retail stores and artificially
inoculated by targeting 10 CFU/25 g sample. The actual
Results
inoculation levels turned out to be from 10 to 40 CFU/25 g, or
0.25 to 1.6 CFU/g of sample, as monitored by quantitative plate
counts of the inoculum preparation. Inoculated and uninoculated Method for DNA Extraction
samples (25 g) were enriched with 225 mL of mTSB containing
mitomycin C (100 ng/mL), homogenized for 30 s in a The Ct values of stx gene amplification in multiplex PCR
stomacher, and incubated at 35 ± 2°C for 20 to 24 h. Mung were compared between the template DNA obtained following
bean sprout samples were also enriched in mEHEC, followed the MLG 5B protocol and the DNA prepared using InstaGene
by incubation at 42 ± 1°C for 48 h in the last batch following matrix (Figure 1). The Ct values of stx gene ranged from 13 to 25
the manufacturer’s instructions after it was realized that mTSB in 80 inoculated RTE meat, fruit, and vegetable samples (except
did not work for sprouts, as mentioned above. All uninoculated sprouts) and from 25 to “undetermined” in 40 uninoculated
samples were negative by enzyme-linked immunosorbent assay samples. The template DNA samples prepared using InstaGene
and immunoblotting following the protocol described (11). resulted in lower Ct values for inoculated samples. The gap of
Ct values between uninoculated and inoculated samples seemed
DNA Extraction and PCR larger when DNA was prepared by InstaGene than the gap when
the DNA was prepared following MLG 5B. When a Chi-square
The enrichment cultures were subjected to DNA extraction as test was conducted using cutoff values between 28 and 30, both
described in USDA FSIS MLG 5B (9). Subsamples of cultures methods detected a similar number of positive samples (data
without large debris (1.4 ± 0.1 mL) were centrifuged at 10 000 × g not shown). This suggests that for screening purposes, these two
for 5 min. The pellet was then suspended in 500 ± 50 μL of DNA extraction methods are not significantly different.
0.85% saline. Samples were then again centrifuged at 10 000 × g
for 3 min. The pellet was suspended in 90 ± 9 μL TE buffer
Ct value of stx gene from DNA following MLG 5B
Effect of Food Matrix and STEC Strain Strain/batch, instead of strain alone, is defined in ANOVA
(Table 2) because our experiment was conducted in batches,
The stx gene Ct value was affected by the type of food matrix, in which samples from each of the 16 food matrices were
the STEC strain, and separate batches (Table 2). The larger an inoculated with one STEC strain as mentioned above.
F value, the smaller a P value, and the higher confidence to accept Therefore, variation between batches also contributed to
a significant difference. The F value corresponding to a P = 0.05 the significance. In other words, strain/batch covers both the
for the effect of matrix in Table 2 ranges from 2.01 to 2.07, and variation of strains and the variation of experimental batches.
the F value for P = 0.05 for strain/batch ranges from 3.67 to 3.70. There is no way to mathematically separate them in this study.
The calculated F values for the effect of matrix and strain/batch Although a level of 10 CFU/25 g sample had been targeted at
(152.25, 85.79) are much larger than their F values at P = 0.05, inoculation, the difference of inoculum level between strains/
resulting in P values (<0.0001) much less than 0.05. The interaction batches had been noticed within a range of 10 to 40 CFU/25 g
between food matrix and STEC strain is also significant; however, (Figure 3). Hopefully, the difference in inoculum level within
the mean square explained by the interaction (8.78) is much such a range will not introduce significant variation in the
smaller than that explained by either matrix (323.86) or strain/ PCR Ct value after a 20 to 24 h enrichment. If any, this type
batch (182.49; Table 2). The scatterplot of stx gene Ct values of variation should have been caught in the variation of strain/
against food matrix in both inoculated and uninoculated samples batch in Table 2.
is presented in Figure 2, and that against strain is in Figure 3.
A cutoff value of 30 was selected for all matrices and yielded 4 Amplification of eae Gene in Multiplex PCR
false-positives and 0 false-negative (Figures 2 and 3).
Multiplex PCR yielded Ct values for the amplification
of both stx and eae genes. The Ct value of stx and eae genes
correlated significantly with each other among the samples
Table 2. ANOVA for the effect of food matrix and STEC
strain/batch on stx cycle threshold (Ct) value in PCR based
yielded a Ct value < 45 from both genes (y = 0.96x + 1.38;
on inoculated samples R2 = 0.98, where y is the Ct value of eae and x is of stx; N = 501
tests). At the same time, among the tests of inoculated fruit and
Sum of Mean vegetable samples (n = 221), 19 (9%; 4 carrot, 3 lettuce, 1 sprout,
Source DFa squares square F value P value 3 raspberry, 5 strawberry, and 3 tomato) produced stx Ct values
Matrix 16 5181.83 323.86 152.25 <0.0001 from 13 to 25, while their eae gene was “undetermined.” The 19
Strain/batch 6 1094.92 182.49 85.79 <0.0001 samples consisted of eight PCR tests of samples inoculated with
O157:H7, seven tests with O111:NM, two with O145:NM, one
Matrix × strain/batch 96 843.05 8.78 4.13 <0.0001
with O45:H2, and one with O121:H19. All of the seven STEC
Residual 273 580.73 2.13
strains are known carriers of both stx and eae genes (Table 1).
a
Degree of freedom. Further PCR reactions targeting only the eae gene revealed
45
40
35
30
stx gene Ct Value
25
20
Sprouts- in mEHEC**
15
Cooked Ham
Cooked Deli
Apple Cider
Strawberry
Raspberry
Dry Cured
10
Sausage
Sausage
Sprouts*
Spinach
Chicken
Broccoli
Cooked
Tomato
Cooked
Lettuce
Salami
Carrot
Ham
5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Food Matrix Uninoculated Inoculated Proposed Cutoff
Figure 2. The stx gene cycle threshold (Ct) value in ready-to-eat meat, fruit, and vegetable matrices enriched in mTSB except where
indicated. *mTSB was ineffective for mung bean sprouts. **mEHEC worked significantly better than mTSB for mung bean sprouts.
Gao et al.: Journal of AOAC International Vol. 101, No. 6, 2018 1831
45
40
35 Uninoculated RTE
meat
Uninoculated fruit
30 & vegetable
stx gene Ct value
Inoculated RTE
meat
25
Inoculated fruit &
vegetable
20
Proposed Cutoff
15
USDA & InstaGene
10
InstaGene
InstaGene
O121:H19
O111:NM
O145:NM
O103:H2
O26:H11
O15:H7
O45:H2
USDA
USDA
USDA
5
0 1 2 3 4 5 6 7
STEC strain and batch DNA extraction method
Figure 3. The stx gene cycle threshold (Ct) value, DNA extraction protocol, and inoculum level of STEC strain in samples from
ready-to-eat meat, fruit, and vegetable matrices (except sprouts) enriched in mTSB. Strain/batch inoculum at CFU/25 g: O103:H2, 31; O111:NM,
10; O121:H19, 31; O145:NM, 31; O157:H7, 29; O26:H11, 11; and O45:H2, 40. USDA, DNA extracted following USDA FSIS MLG 5B. InstaGene,
DNA extracted using InstaGene Matrix (BioRad).
that these samples were actually positive for the eae gene as 45
well. Suppression of eae gene in multiplex PCR reaction was
stx gene Ct value of replicate 2
Discussion
and pre-enrichment to resuscitate injured cells against
Unlike clinical samples, pathogens in foods are generally competition from high levels of normal flora (13). Therefore,
present at much lower levels and are frequently injured during a successful enrichment is critical to selectively increase target
food processing, which demands improved detection sensitivity organisms to levels above the LOD of a method. The effect of
1832 Gao et al.: Journal of AOAC International Vol. 101, No. 6, 2018