lable at ScienceDirect

Food Microbiology 28 (2011) 478e483

Contents lists avai

Food Microbiology

journal homepage: www.elsevier .com/locate/ fm

O serogroup specific real time PCR assays for the detection and identificationof nine clinically relevant non-O157 STECs

Andrew Lin a,*, Omar Sultan b, Henry K. Lau a, Evelyn Wong a, Gary Hartman a, Carol R. Lauzon b

aU.S. Food and Drug Administration, San Francisco District Laboratory, 1431 Harbor Bay Parkway, Alameda, CA 94502, United StatesbCalifornia State University, East Bay, Department of Biological Sciences, 25800 Carlos Bee Blvd., Hayward, CA 94542, United States

a r t i c l e i n f o

Article history:Received 22 April 2010Received in revised form14 October 2010Accepted 19 October 2010Available online 27 October 2010

Keywords:Shiga toxinFoodborne diseaseEscherichia coli surveillance

* Corresponding author. Tel.: þ1 510 337 6893; faxE-mail address: andrew.lin@fda.hhs.gov (A. Lin).

0740-0020/$ e see front matter Published by Elseviedoi:10.1016/j.fm.2010.10.007

a b s t r a c t

TaqMan� real time PCR assays were designed for each of the non-O157 STEC O serogroups mostcommonly associated with human illness: O26, O45, O91, O103, O111, O113, O121, O128, and O145. Thenine RT-PCR assays can be run as single assays when a known pathogen is of concern, or multiplexed inthree reactions, to quickly screen for the most clinically relevant O serogroups. All assays included aninternal amplification control constructed from the green fluorescent protein gene as an indicator of PCRinhibition. Of 103 strains tested, the inclusive tests accurately identified the O serogroup for 101 strains.The exclusive tests for each assay yielded no false positives for over 120 Escherichia coli strains and 23non-E. coli bacteria tested. Furthermore, the RT-PCR assays were tested by inoculating four food matrices,milk, apple juice, lettuce, and ground beef, at �30 CFU/25 g or mL. Following a 24 h selective enrichment,the RT-PCR assays detected STECs in all foods except for one ground beef sample inoculated with O111,and all apple juice samples inoculated with O113. The assays could also detect each O serogroup inhuman stool specimens inoculated with STECs at 1000 CFU/0.5 g of stool following 24 h enrichment.

Published by Elsevier Ltd.

1. Introduction

Escherichia coli containing one or more of the Shiga toxin genesare characterized as Shiga toxin producing E. coli (STEC), and can bepathogenic to humans. STEC infections cause mild to bloody diar-rhea, hemorrhagic colitis (HC), hemorrhagic uremic syndrome(HUS) and can lead to death of individuals with compromisedimmune systems (Levine et al., 1987; Gyles, 2007). STECs have beenrecovered in the intestines of many wild and domestic animals,particularly ruminants. Transmission may occur via the fecal-oralroute, leading to the contamination of foods and water (Shooteret al., 1970; Griffin and Tauxe, 1991; Karmali, 2004; Gyles, 2007).STECs are classified by their highly variable O antigens intodifferent O serogroups. E. coli serogroup O157 remains the mostcommon STEC in the U.S., but there are more than 100 STEC Oserogroups, over 60 of which have been associated with humanillness (Bettelheim, 2003). Some of these non-O157 STECs may beas virulent as O157, concealing a significant public health concernthat should command more attention and underscores the need todevelop better detection methods for non-O157 STECs (Bettelheim,2007; Cohen et al., 2008).

: þ1 510 337 6704.

r Ltd.

Epidemiological studies suggest that 20e50% of STEC infectionsare caused by non-O157 STECs, accounting for an estimated 37,000illnesses annually in the U.S. (Brooks et al., 2005; Johnson et al.,2006). Two recent reviews in 2006 by Beutin and Johnson et al.identified O26, O103, O111, O145 and O157 as a “gang of five”(Beutin, 2006; Johnson et al., 2006). These O serogroups wereconsistently associated with increased virulence in humans, andmore frequently associated with HC and HUS. Other STECserogroups that are commonly associated with human illnessinclude O91, O113, O128 (Bettelheim, 2007), O45, and O121 (Brookset al., 2005; Gyles, 2007). Although there are assays to detect theShiga toxins (Chart et al., 2001) or Shiga toxin genes (Feng andWeagant, 2009), these assays do not identify the O serogroup,and cannot distinguish between the more virulent STECs and lessharmful or non pathogenic STECs. The difficulty in isolating STECs isthat the only distinguishing characteristic may be their ability toferment different carbohydrate-like substances (Bettelheim, 2003),such that each O serogroup may require a distinct serogroupspecific selective agar. For instance, O157 does not ferment sorbitoland can be distinguished from other STECs on TC-Sorbitol Mac-Conkey agar (March and Ratnam, 1986; Feng and Weagant, 2009).Additionally, O26 does not ferment rhamnose, and can be distin-guished on TC-Rhamnose MacConkey agar (Hiramatsu et al., 2002).Further studies on the other O serogroups may result in distinctmedia for each of the clinically relevant STECs.

A. Lin et al. / Food Microbiology 28 (2011) 478e483 479

Hereinwe describe O serogroup specific TaqMan� real time PCRassays (RT-PCR) for O26, O45, O91, O103, O111, O113, O121, O128, andO145. The assays can be run as single assays to identify O serogroupsor multiplexed to facilitate screening for the clinically relevantSTECs. Use of these O serogroup specific RT-PCR assays could indi-cate which selective agars to use to facilitate isolating some of theSTECs, and also be used to confirm the identity of possible STECisolates. The assays were effective in detecting each O serogroup infoods and human stool specimens. Improvements in detectionmethods of these pathogens may help prevent consumption ofcontaminated products, and aid in the diagnosis and treatment of

Table 1E. coli strains used in this study.

O serogroup Strain identification O serogroup

O26 TB206Ab O111TB285Ab

TB352Ab

EK29b

97-3250b

ATCC 12795c

H11b

H19b,f

DEC10Bb

DEC10Cb

DEC9Fb

TB285Cb

VP30b

O113O45 D88-28058b

DEC11Cb

5431-72b,f

4309-65b

2566-58b

B8227-C1b

E-D371b

DA-21b

B8026-C1b

B8227-C8b

B8228-C2b

08-000-17b,g

O121O91 B2F1b

23/67b

87-2927b,f

M710b

988/2b

1120/3b

852/3b

226-1b

68-Il-38b

O103 TB154Ab,f

87-2931b O128MT#82b

MT#80b

EK30b

EK31b

EK32b

109-494b

107-226b

PMK5b

RW1372b

RW1374b

DA40b

D55b,h

a O serogroups were not identified for these strains.b Strains obtained from the STEC center at Michigan State University’s National Foodc Strains obtained from the American Type Culture Collection, Manassas, VA.d Strains obtained from Robert Mandrell at the USDA-Agricultural Research Services, Ae Strains obtained from the Ohio State Department of Agriculture, Reynoldsburg, OH.f Strains used for the food matrix interference experiments and stool interference expg This previously unidentified STEC was positive for O45 by qPCR and confirmed by Oh This strain was classified as O111 by the STEC center at Michigan State Unversity’s Na

agglutination confirm that this strain was actually O103.

infected patients. Rapid identification of O serogroups would enableinvestigators to link clinical, environmental, and food isolates tomanage the spread of foodborne outbreaks.

2. Materials and methods

2.1. O antisera agglutination

Isolates were grown in brain heart infusion (BHI) broth (HardyDiagnostics, Santa Maria, CA) for 24 h at 35 �C. One milliliter of

Strain identification O serogroup Strain identification

CL-37b O145 GS G5578620b,f

DEC 8Bb IHIT 0304b

3007-85b,f IH 16b

TB226Ab 02-3422b

928/91b 4865/96b

412/55b TB269Cb

DEC8Cb MT#66b

C412b BCL73b

ED-31b

EK35b O8 06-00236e

3215-99b a 07-00006e

DA-18b O130 08-00007e

O1 08-00011e

CL-3b,f a 08-00015e

DEC16Ab O1 08-00016e

CL-15b a 08-00021e

MDCH-4b O136 08-00022e

87-307b O8 08-00024e

24-35b O73 08-00025e

DEC16Cb O15 RDEC-1b

M2101b ON BCL19b

M2415b O70 DEC10Jb

M2443b O104 G5506b

D167b O146 DEC16Eb

ECOR-30b O15 88-1509b

O125 8153B-86b

F6173b O156 M2113b

DA-1b O5 BCL17b

DA-5b OX03 90-1787b

DA-37b a ATCC 25922c

MT#2b O157 ATCC 43888c

87-2914b O157 ATCC 43894c

MT#11b a 75-83b

MT#22b a 314Sb

DA-69b,f a DEC 101b

MT42b a 403-3b

3-524b a 3377-85ba 88-4110Hb

BM2018d,f a 848/1b

BM2020d a 907/1b

BM2021d

DEC11Db

DEC13Eb

DEC14Db

DEC13Db

DEC14Eb

DEC13Cb

DEC13Ab

DEC14Bb

DEC14Ab

Safety and Toxicology Center, East Lansing, MI.

lbany, CA.

eriments.antisera agglutination.tional Food Safety and Toxicology Center. The O specific qPCR assays and O antisera

Table 3Primers and probes used in this study.

O serogroup Gene Sequence Gene position (bp)

O26 wzx ttttatctggcgtgctatcg 557e577cggggttgctatagactgaa 784e804a

6FAM-tggcactcttgcttcgcctg-BHQ1 720e740

O45 wzy tacgatttcacaagcttcca 769e789tgcaatcgcataaggaaata 1003e1023a

6FAM-tcgcgggctcccttattgtg-BHQ1 917e937TxRed-tcgcgggctcccttattgtg-BHQ2 917e937

O91 wzx catgctgctcattcttctca 266e286tggagtttgcaacaaacaaa 380e400a

A. Lin et al. / Food Microbiology 28 (2011) 478e483480

overnight culture was boiled for 1 h. One hundred twenty micro-liters of boiled culture were added to microtitre plates. Eightymicroliters of O serogroup specific antisera (MiraVista/StatensSerum Institut, Indianapolis, IN) was added to each well. Microtitreplates were then covered, and incubated at 50 �C for 24 h andobserved for agglutination.

2.2. DNA boil preparation

E. coli strains and non-E. coli bacteria listed in Tables 1 and 2were grown in BHI broth for 24 h at 35 �C. DNA was preparedfrom the overnight cultures by boiling as previously described(Feng and Weagant, 2009) and stored at �20 �C until used.

2.3. gfp internal control

Internal control DNA was prepared from the pGLO plasmid(Biorad, Hercules, CA) that contains the gene coding for greenfluorescent protein (gfp). PCR was performed using 1� Takara HotStart PCR Buffer, 200 mM each dNTP, 400 nM primers GFP549F, andGFP709R, 1.5 U Takara Taq Hot Start DNA polymerase (Takara BioInc., Madison, WI), and 1 mL of pGLO DNA as a template. PCRreactions were run on an Applied Biosystems GeneAmp 9700(Applied Biosystems Inc., Foster City, CA) with an initial denatur-ation at 94 �C for 5 min, 30 cycles of 94 �C for 30 s, 56 �C for 30 s,and 72 �C for 1 min, followed by a final extension at 72 �C for 7 min.PCR products were confirmed by staining with ethidium bromideand 300 nm transillumination of a 1% agarose gel in 1� TAE buffer.DNA was gel extracted using a Qiagen Gel Extraction kit (Qiagen,Valencia, CA). Gel purified gfp DNA was quantified using a Nano-drop 3300 Fluorospectrometer (Thermo Scientific, Wilmington,DE). gfp DNA was serially diluted 10-fold with sterile water andstored at �20 �C until used.

2.4. RT-PCR single assays

RT-PCR reactions were conducted using 1� Takara Hot StartPCR /buffer, 200 mM each dNTP, 1.5 U Takara Taq Hot Start DNA

Table 2Non-E. coli strains used in this study.

Microorganisms ATCC no.

Bacillus cereus 14579a

Bacillus subtilis 6633a

Bacillus thuringiensis 10792a

Enterobacter aerogenes 13048a

Enterococcus faecalis 19433a

Listeria innocua 33090a

Listeria monocytogenes LS82b

Listeria ivanovii 19119a

Listeria seelegeri 35967a

Salmonella enterica 9700a

Salmonella enterica 8324a

Salmonella enterica 9842a

Salmonella enterica diarizonae 12325a

Salmonella enterica diarizonae 29934a

Proteus mirabilis 25933a

Pseudomonas aeruginosa 27853a

Rhodococcus equi 6939a

Staphylococcus aureus 25923a

Staphylococcus epidermidis 14990a

Shigella dysenteriae 13313a

Shigella sonnei 9290a

Shigella boydii 9207a

Shigella flexneri 12022a

a Strains obtained from the American Type Culture Collection (ATCC).b Strain obtained from Atin Datta at the U.S. Food and Drug Admin-

istration’s Center for Food Safety and Applied Nutrition, College Park,MD.

polymerase, 400e600 nMO serogroup specific forward and reverseprimers, 100 nM O serogroup specific 6-carboxyfluorescein (FAM)probes (listed in Table 3), 45 pg gfp internal control DNA, 400 nMinternal control forward and reverse primers, and 100 nM gfpinternal control Cy5 probe. Primer concentrations for O26 and O111RT-PCR reactions were 600 nM, while all other reactions consistedof 400 nM. Two microliters of overnight culture or sample enrich-ment DNA template were added to 25 mL RT-PCR reactions. RT-PCRreactions were performed in a Cepheid smartcycler II (Cepheid,Sunnyvale, CA), with an initial denaturation at 94 �C for 5 minfollowed by 40 cycles at 94 �C for 30 s, 52 �C for 30 s with optics on,and 72 �C for 60 s. Themanual thresholdwas set for 15 fluorescenceunits using the FAM-TET-TxR-Cy5 dye set.

2.5. RT-PCR multiplex assays

RT-PCR reactions were combined in a multiplex format intothree groupings. Group 1 consisted of assays for serogroups O26,O45, and O103; Group 2 for O111, O128, and O145; and Group 3 forO91, O121, and O113. Illustra puReTaq Ready-To-Go PCR Beads (GEHealthcare, Piscataway, NJ) were used for the multiplex reactions,and all assays included the internal amplification control whichconsisted of 45 pg of gfp internal control DNA, 400 nM of forwardand reverse internal control primers, and 100 nM Cy5 internalcontrol probe. The multiplex RT-PCR assay for O26, O45, and O103

6FAM-aaatggtttgctgcgacgct-BHQ1 358e378

O103 wzx gggcttgtattgttgtaccg 896e916agtggcaaacagccaactac 1045e1065a

6FAM-tcggggattttctgcggatt-BHQ1 1025e1045Cy3-tcggggattttctgcggatt-BHQ2 1025e1045

O111 wzx caatccaatttgcatcttca 75e95accgcaaatgcgataataac 294e314a

6FAM-tggaggatgttccgcatgga-BHQ1 189e209

O113 wzx tgaccttacttcctgcgaat 752e772agcaccacgataggattgaa 977e997a

6FAM-cctgggaggaggctgcaaaa-BHQ1 953e973Cy3-cctgggaggaggctgcaaaa-BHQ2 953e973

O121 wzy tggatggcattcctcagtat 809e829agcaagccaaaacactcaac 1043e1063a

6FAM-ttaacacgggcgtggttgga-BHQ1 920e940TxRed-ttaacacgggcgtggttgga-BHQ2 920e940

O128 wzx tcgatcgtcttgttcaggtt 1123e1143gaatgcaatgggcaattaac 1298e1318a

6FAM-gggttgcacaattggcctcc-BHQ1 1184e1204TxRed-gggttgcacaattggcctcc-BHQ2 1184e1204

O145 wzy tgttcctgtctgttgcttca 224e244atcgctgaataagcaccact 495e515a

6FAM-tgggctgccactgatgggat-BHQ1 442e462Cy3-tgggctgccactgatgggat-BHQ2 442e462

Internal Control gfp gcgttcaactagcagaccat 549e569cccagcagctgttacaaact 689e709a

Cy5-tggcgatggccctgtccttt-BHQ1 589e609

a Reverse compliment.

Table 4Inclusive and exclusive studies of O serogroup specific assays.

O serogroup Inclusiveb Exclusivea

Other STEC Microorganisms

O26 13/13 0/121 0/23O45 11/12 0/122 0/23O91 8/9 0/125 0/23O103 14/14 0/120 0/23O111 12/12 0/122 0/23O113 12/12 0/122 0/23O121 11/11 0/123 0/23O128 12/12 0/122 0/23O145 8/8 0/126 0/23

a Exclusive studies are listed as number of false positive results/number ofexclusive strains tested.

b Inclusive studies are listed as number of positive results/number of inclusivestrains tested.

A. Lin et al. / Food Microbiology 28 (2011) 478e483 481

was conducted using 600 nM O26 forward and reverse primers,100 nM FAM O26 probe, 400 nM O45 forward and reverse primers,100 nM Texas Red O45 probe, 400 nM O103 forward and reverseprimers, and 100 nM Cy3 O103 probe. The multiplex RT-PCR assayfor O111, O128, and O145 was conducted using 800 nM O111forward and reverse primers, 100 nM FAM O111 probe, 400 nMO128 forward and reverse primers, 100 nM Texas Red O128 probe,400 nM O145 forward and reverse primers, and 100 nM Cy3 O145probe. The multiplex RT-PCR assay for O91, O121, and O113 wasconducted using 400 nM O91 forward and reverse primers and100 nM FAMO91 probe, 400 nM O121 forward and reverse primersand 100 nM Texas Red O121 probe, 400 nM O113 forward andreverse primers, and 100 nM Cy3 O113 probe. Two microliters ofovernight culture or sample enrichment template DNA were usedin 25 mL RT-PCR reactions. RT-PCR reactions were performed ina Cepheid smartcycler II, with an initial denaturation at 94 �C for5 min followed by 40 cycles at 94 �C for 30 s, 52 �C for 30 s withoptics on, and 72 �C for 60 s. The manual threshold was set for 15fluorescence units using the FAM-Cy3-TxR-Cy5 dye set.

2.6. Food matrix interference

Non-fat milk, apple juice, lettuce, and ground beef were inoc-ulated with one representative STEC strain from each O serogroup.Each strain was grown in BHI for 24 h at 35 �C. Cultures wereserially diluted in Butterfield’s Phosphate Buffered Saline (BPBS)(3 M, St. Paul, MN) to the desired concentration and added to eachfood sample. Lettuce and ground beef samples were allowed to airdry for 10 min. Milk and apple juice samples were mixed bypipetting. Uninoculated samples were also tested. Twenty-fivemilliliters or twenty-five grams of each inoculated sample wereenriched with 225 mL of Tryptic Soy Broth (TSB) (Hardy Diagnos-tics) with 2.5 mg/L potassium tellurite (Invitrogen, Carlsbad, CA),50 mg/L cefixime (Invitrogen), and 40 mg/L vancomycin (Sigma, St.Louis, MO) (TSBþ T,C,V) (Catarame et al., 2003) and incubated at41.5 �C for 24 h. DNA was prepared from 1 mL of the 24 h cultureprepared by boiling as previously described (Feng and Weagant,2009). RT-PCR assays were also performed in multiplex format asmentioned earlier.

2.7. Stool specimen testing

Representative non-O157 strains from each O serogroup weregrown in BHI broth for 24 h at 35 �C and serially diluted in BPBS forinoculating human stool samples. Human stool samples wereacquired from a healthy donor and stools were aseptically placedinto a pre-sterilized container, and processed immediately. For eachof the representative bacterial strains, inoculated stool sampleswere prepared by adding approximately 1000 CFU to 0.5 g ofhuman stool. Uninoculated and inoculated stool samples wereenriched with 5 mL TSB for 24 h at 35 �C as previously described(Gilmour et al., 2009). Bacterial DNA was isolated from each inoc-ulated stool sample using QIAamp DNA Stool Mini Kit (Qiagen)according to the supplier’s protocol. As a negative control, DNAisolation was also performed on uninoculated stool. PCR amplifi-cation was performed in single assay format using primers andprobes specific to each non-O157 strain as previously described.

3. Results and discussion

Thewzx O antigen flippase andwzy O antigen polymerase genesencode for proteins that synthesize the unique and specificmolecular structures of the O antigen located on the surface of theE. coli outer cellular membrane. Consequently, these genes act asexcellent targets for the development of nucleic acid based

detection assays since O serogroups can be discriminated based onthe broad diversity of these gene sequences. Several conventionaland real time PCR assays have been developed using these genes forO serotyping specific O serogroup STECs (Durso et al., 2005; Fenget al., 2005; Fratamico et al., 2005, 2009; Paton and Paton, 1999;Perelle et al., 2002, 2005; Wang et al., 1998). Several other multi-plex assays have been developed for detecting multiple STECO serogroups. O’Hanlon et al. (2004) had developed a multiplexSybr Green assay to detect O26, O157 and O111. Perelle et al. (2007)had developed a multiplex assay for detecting O26, O103, O111,O145 and O157. These assays were performed with multiplexamplification but would require additional testing to identify thespecific O serogroup. Beutin et al. (2009) had developed RT-PCRassays on the gene disc platform that could detect O26, O103, O111,O145, and O157, but this method is specific for gene disc instru-mentation. Additionally Ballmer et al. (2007), had developeda microarray that can identify 24 STEC O serogroups, but thismethod was more suited to identifying the O serogroup of isolatesrather than detecting STECs in foods. We have developed TaqMan�RT-PCR assays for the nine most clinically relevant non-O157 STECsthat can be run as single assays, or multiplexed into three RT-PCRreactions. Although the assays were tested here on the CepheidSmartcycler II, the RT-PCR assays could be transferred to any plat-form that can accommodate Taqman� assays. Additionally, ourassays include an internal amplification control that is particularlyimportant for detection of pathogens in foods, since some foodmatrices that harbor pathogens can potentially elicit an inhibitoryeffect on PCR reactions. The internal amplification control wasconstructed from the gfp gene based on its wide availability andbecause gfp DNA is unlikely to naturally occur in original bacterialisolates (Murphy et al., 2007).

Each of the O serogroup specific RT-PCR TaqMan� assayswere tested for positive results among strains inclusive of theO serogroups, and tested for false positives among exclusive STECstrains and other non-E. coli bacteria (Table 4). Inclusive andexclusive tests were done using the strains described in Tables 1and 2. Inclusive tests gave identical results for single assay formatand multiplex format. Exclusive tests were done in single assayformat. Inclusive and exclusive accuracy of the assays wereconfirmed by agglutination with O specific antisera. gfp internalcontrols for all tests werewithin 5 cycle threshold values (CT) of thenegative controls, indicating no substantial PCR inhibition. In total,101 of 103 (98% STEC) isolates tested were correctly identified byusing the O serogroup RT-PCR assays. The O serogroup specificassayswere unsuccessful at identifying O antigen phenotype of O91strain 23/67 and O45 strain ED-371. It is postulated that poly-morphisms within the primer or probe binding sites of thewzx andwzy genes in these two strains may have led to negative RT-PCR

Table 5% Positive STEC detected in inoculated food samples by O serogroup specific RT-PCR

O serogroup Milk (n¼ 6) Beef (n¼ 3) Lettuce (n¼ 3) Apple juice (n¼ 3)

Inoculum(CFU/25 mL)

RT-PCRresults

Inoculum(CFU/25 mL)

RT-PCRresults

Inoculum(CFU/25 mL)

RT-PCRresults

Inoculum(CFU/25 mL)

RT-PCRresults

O26 15 100% 30 100% 12 100% 15 100%O45 11 100% 15 100% 15 100% 11 100%O91 20 100% 17 100% 17 100% 24 100%O103 2 100% 1 100% 24 100% 12 100%O111 14 100% 14 67% 12 100% 16 100%O113 14 100% 20 100% 20 100% 12 0%O121 20 100% 22 100% 22 100% 20 100%O128 16 100% 19 100% 19 100% 19 100%O145 9 100% 10 100% 10 100% 15 100%

A. Lin et al. / Food Microbiology 28 (2011) 478e483482

results, with no effect on the O antigen phenotype. The use of RT-PCR assays did allow for the identification of previously unknownand misclassified STEC strains (STEC Center, 2009). STEC strain 08-0017 was identified as serogroup O45 and STEC strain D55 asserogroup O103. There were no false positive results in any of thenine assays for either the exclusive E. coli strains or the non-E. colibacteria tested.

Detection of STECs in various food matrices was tested by theO serogroup specific RT-PCR assays (Table 5). Catarame et al.(2003), had demonstrated that enrichment with TSBþ T, C, V at41.5 �C was effective in detecting STECs O26 and O111 in foods.All STEC strains used in this study were found to grow in thisenrichment medium. Non-fat milk, ground beef, lettuce, and applejuice were inoculated with STEC strains representative of eachO serogroup and enriched overnight before DNA preparation andRT-PCR. All uninoculated controls were negative and gfp internalcontrols were all within 5 CT of the negative control, indicating nosubstantial PCR inhibition. All O serogroups were detected in thefoods tested using the RT-PCR assays, except for O113 in applejuice and one inoculated sample of O111 in ground beef. Twenty-four hour enrichment of O113 in apple juice failed to yieldturbidity, and produced negative results with FDA’s official stx1/2RT-PCR method (Feng and Weagant, 2009) for which the O113strain was positive (data not shown). Taken together, these resultssuggest that either O113 growth was inhibited in apple juice, orthere was reduced sensitivity of RT-PCR assays in this serogroup-food matrix combination. One sample of ground beef inoculatedwith O111 tested negative by RT-PCR. There was noticeableturbidity and the stx1/2 RT-PCR detected stx1 and 2 in that sampleenrichment. The O111 RT-PCR may not be as sensitive in beef andmay require a more selective enrichment or a more effective DNAextraction method to improve the recovery at the level of inoc-ulum tested.

In addition to detecting STECs in foods, the RT-PCR assaysmay aid in identifying O serogroups in clinical specimens. Recentguidelines issued by the Center for Disease Control state that allacute community acquired diarrhea should be screened for Shigatoxins, and Shiga toxin positive specimens should be sent toa reference laboratory for culturing of non-O157 STECs and sero-typing using a panel of O specific antisera (Gould et al., 2009). Thisprocess may take several days to complete, whereas RT-PCR assayscan be performed in a matter of hours. Early diagnosis of STECinfections is important for determining the proper treatment, andidentification of the STEC strain is essential for effective outbreakresponse (Gould et al., 2009). Stool samples that were inoculatedwith 1000 CFU non-O157 STECs/0.5 g were tested after 24 henrichments. The RT-PCR assays yielded positive results for eachspecific O serogroup (n¼ 3) after enrichment and DNA extractionwith the QiAMP stool kit. Use of this kit has been successful indetecting STECs in bovine stool (Gioffré et al., 2004) by reducing

the effect of PCR inhibitors found in stool specimens. We found thisstool kit also to be effective in extracting STEC DNA from humanstool specimens.

Here we have described RT-PCR assays for the detection of nineof the most commonly isolated and virulent non-O157 STECs in theU.S. though we have shown that these assays are effective atdetecting STEC’s in foods, there are still limitations to the assays.First, although we have developed assays for the nine most clini-cally relevant STECs, many more STECs can cause disease and coulddevelop into more serious problems in the future. Secondly, theRT-PCR assays do not result in a bacterial isolate that could beconfirmed with additional tests and enable food safety agencies totake regulatory action. Thirdly, the O serogroup RT-PCR assaysdetect the wzx flippase and wzy polymerase genes, not the viru-lence genes. These assays alone can not differentiate pathogenicSTECs and non pathogenic E. coli strains belonging to the sameO serogroup. However, when used in combination, RT-PCR assaysthat screen for the presence of Shiga toxin genes plus theO serogroup specific RT-PCR assays can facilitate the identificationof presumptive non-O157 STEC isolates. Results for current clinicaltests on non-O157 STECs utilizing O antisera agglutination testsrequire several days. RT-PCR assays bear an advantage over tradi-tional serological tests since RT-PCR results in faster identificationof non-O157 STECs in clinical specimens, reducing the timenecessary to begin treatment of infected patients. In addition, fasteridentification of O serogroups can reduce the impact of foodborneillness outbreaks by increasing the speed of matching clinical,environmental and food isolates to sources of contamination.

Acknowledgements

The authors would like to thank the STEC Center at MichiganState University’s National Food Safety and Toxicology Center, EastLansing, MI, Robert Mandrell, United States Department of Agri-culture-Agricultural Research Services (USDA-ARS), Albany, CA, theOhio Department of Agriculture, Reynoldsburg, OH, and Atin Datta,U.S. Food and Drug Administration (FDA), Center for Food Safetyand Applied Nutrition, College Park, MD for generously sharingtheir bacterial cultures. We would also like to thank David K. Lau,Roderick V. Asmundson, and Thomas H. Sidebottom of the FDA, andLaurie Clotilde and Mark J. Carter of the USDA-ARS. This work wasprimarily supported by the California State University Program forEducation and Research in Biotechnology.

References

Ballmer, K., Korczak, B.M., Kuhnert, P., Slickers, P., Ehricht, R., Hächler, H., 2007.Fast DNA serotyping of Escherichia coli by use of an oligonucleotide microarray.Journal of Clinical Microbiology 45, 370e379.

Bettelheim, K.A., 2003. Non-O157 verotoxin-producing Escherichia coli: a problem,paradox, and paradigm. Experimental Biology and Medicine 228, 333e344.

A. Lin et al. / Food Microbiology 28 (2011) 478e483 483

Bettelheim, K.A., 2007. The non-O157 Shiga-toxigenic (verocytotoxigenic)Escherichia coli; under-rated pathogens. Critical Reviews in Microbiology 33,67e87.

Beutin, L., 2006. Emerging enterohaemorrhagic Escherichia coli, causes and effectsof the rise of a human pathogen. Journal of Veterinary Medicine SeriesB-Infectious Diseases and Veterinary Public Health 53, 299e305.

Beutin, L., Jahn, S., Fach, F., 2009. Evaluation of the ‘GeneDisc’ real-time PCR systemfor detection of enterohaemorrhagic Escherichia coli (EHEC) O26, O103, O111,O145 and O157 strains according to their virulence markers and their O- andH-antigen-associated genes. Journal of Applied Microbiology 106, 1122e1132.

Brooks, J.T., Sowers, E.F., Wells, J.G., Greene, K.D., Griffin, P.M., Hoekstra, R.M.,Strockbine, N.A., 2005. Non-O157 Shiga toxin-producing Escherichia coli infectionsin the United States, 1983e2002. Journal of Infectious Diseases 192, 1422e1429.

Catarame, T.M., O’Hanlon, K.A., Duffy, G., Sheridan, J.J., Blair, I.S., McDowell, D.A.,2003. Optimization of enrichment and plating procedures for the recovery ofEscherichia coli O111 and O26 from minced beef. Journal of Applied Microbi-ology 95, 949e957.

Chart, H., Willshaw, G.A., Cheasty, T., 2001. Evaluation of a reversed passive latexagglutination test for the detection of verocytotoxin (VT) expressed by strains ofVT-producing Escherichia coli. Letters in Applied Microbiology 32, 370e374.

Cohen, N.J., Jones, R.C., Schermond, J.S., Gerber, S.I., 2008. Identifying Shiga toxin-producing Escherichia coli: results of a laboratory evaluation. Foodborne Path-ogens and Disease 5, 295e298.

Durso, L.M., Bono, J.L., Keen, J.E., 2005. Molecular serotyping of Escherichia coliO26:H11. Applied and Environmental Microbiology 71, 4941e4944.

Feng, L., Senchenkova, S.N., Tao, J., Shashkov, A.S., Liu, B., Shevelev, S.D., Reeves, P.R.,Xu, J., Knirel, Y.A., Wang, L., 2005. Structural and genetic characterization ofenterohemorrhagic Escherichia coli O145 O antigen and development of anO145 serogroup-specific PCR assay. Journal of Bacteriology 187, 758e764.

Feng, P., Weagant, S.D., 2009. Diarrheagenic Escherichia coli. In: BacteriologicalAnalytical Manual. http://www.fda.gov/Food/ScienceResearch/LaboratoryMethods/BacteriologicalAnalyticalManualBAM/UCM070080 Chapter 4a.

Fratamico, P.M., DebRoy, C., Strobaugh Jr., T.P., Chen, C.Y., 2005. DNA sequence of theEscherichia coli O103 O antigen gene cluster and detection of enter-ohemorrhagic E. coli O103 by PCR amplification of the wzx and wzy genes.Canadian Journal of Microbiology 51, 515e522.

Fratamico, P.M., DebRoy, C., Miyamoto, T., Liu, Y., 2009. PCR Detection of enter-ohemorrhagic Escherichia coli O145 in food by targeting genes in the E. coli O145O-antigen gene cluster and the Shiga toxin 1 and Shiga toxin 2 genes. Food-borne Pathogens and Disease 6, 605e611.

Gilmour, M.W., Chui, L., Chiu, T., Tracz, D.M., Hagedorn, K., Tschetter, L., Tabor, H.,Ng, L.K., Loui, M., 2009. Isolation and detection of Shiga toxin-producingEscherichia coli in clinical stool samples using conventional and molecularmethods. Journal of Medical Microbiology 58, 905e911.

Gioffré, A., Meichtri, L., Zumárraga, M., Rodríguez, R., Cataldi, A., 2004. Evaluation ofa QIAmp DNA stool purification kit for Shiga-toxigenic Escherichia coli detectionin bovine fecal swabs by PCR. Revista Argentina De Microbiologia 36, 1e5.

Gould, L.H., Bopp, C., Strockbine, N., Atkinson, R., Baselski, V., Body, B., Carey, R.,Crandall, C., Hurd, S., Kaplan, R., Neill, M., Shea, S., Somsel, P., Tobin-D’Angelo, M., Griffin, P.M., Gerner-Smidt, P., 2009. Recommendations for diag-nosis of Shiga toxin-producing Escherichia coli infections by clinical laboratories.

Morbidity and Mortality Weekly Report Recommendations and Reports 58,1e14.

Griffin, P.M., Tauxe, R.V., 1991. The epidemiology of infections caused by Escherichiacoli O157:H7, other enterohemorrhagic E. coli, and the associated hemolyticuremic syndrome. Epidemiologic Reviews 13, 60e98.

Gyles, C.L., 2007. Shiga toxin-producing Escherichia coli: an overview. Journal ofAnimal Science 85, E45eE62.

Hiramatsu, R., Matsumoto, M., Miwa, Y., Suzuki, Y., Saito, M., Miyazaki, Y., 2002.Characterization of Shiga toxin-producing Escherichia coli O26 strains andestablishment of selective isolation media for these strains. Journal of ClinicalMicrobiology 40, 922e925.

Johnson, K.E., Thorpe, C.M., Sears, C.L., 2006. The emerging clinical importance ofnon-O157 Shiga toxin-producing Escherichia coli. Clinical Infectious Diseases 43,1587e1595.

Karmali, M.A., 2004. Infection by Shiga toxin-producing Escherichia coli: an over-view. Molecular Biotechnology 26, 117e122.

Levine, M.M., Xu, J.G., Kaper, J.B., Lior, H., Prado, V., Tall, B., Nataro, J., Karch, H.,Wachsmuth, K., 1987. A DNA probe to identify enterohemorrhagic Escherichiacoli of O157:H7 and other serotypes that cause hemorrhagic colitis andhemolytic uremic syndrome. Journal of Infectious Diseases 156, 175e182.

March, S.B., Ratnam, S., 1986. Sorbitol-MacConkey medium for detection ofEscherichia coli O157:H7 associated with haemorrhagic colitis. Journal of ClinicalMicrobiology 23, 869e872.

Murphy, N.M., McLauchlin, J., Ohai, C., Grant, K.A., 2007. Construction and evalua-tion of a microbiological positive process internal control for PCR-basedexamination of food samples for Listeria monocytogenes and Salmonella enterica.International Journal of Food Microbiology 120, 110e119.

O’Hanlon, K.A., Catarame, T.M.G., Duffy, G., Blair, I.S., McDowell, D.A., 2004. RAPIDdetection and quantification of E. coli O157/O26/O111 in minced beef by real-time PCR. Journal of Applied Microbiology 96, 1013e1023.

Paton, A.W., Paton, J.C., 1999. Molecular characterization of the locus encodingbiosynthesis of the lipopolysaccharide O antigen of Escherichia coli serotypeO113. Infection and Immunity 67, 5930e5937.

Perelle, S., Dilasser, F., Grout, J., Fach, P., 2002. Identification of the O-antigenbiosynthesis genes of Escherichia coli O91 and development of a O91 PCRserotyping test. Journal of Applied Microbiology 93, 758e764.

Perelle, S., Dilasser, F., Grout, J., Fach, P., 2005. Detection of Escherichia coli serogroupO103 by real-time polymerase chain reaction. Journal of Applied Microbiology98, 1162e1168.

Perelle, S., Dilasser, F., Grout, J., Fach, P., 2007. Screening food raw materials for thepresence of the world’s most frequent clinical cases of Shiga toxin-encodingEscherichia coli O26, O103, O111, O145 and O157. International Journal of FoodMicrobiology 113, 284e288.

Shooter, R.A., Cooke, E.M., Rousseau, S.A., Breaden, A.L., 1970. Animal sources ofcommon serotypes of Escherichia coli in the food of hospital patients. Possiblesignificance in urinary-tract infections. Lancet 2, 226e228.

STEC Center, 2009. National Food Safety & Toxicology Center. Michigan StateUniversity. http://www.shigatox.net/stec/cgi-bin/index.

Wang, L., Curd, H., Qu, W., Reeves, P.R., 1998. Sequencing of Escherichia coli O111O-antigen gene cluster and identification of O111-specific genes. Journal ofClinical Microbiology 36, 3182e3187.