Pulsed-field gel electrophoresis typing, antibiotic resistance, and plasmid profiles of Escherichia coli strains isolated from foods Ahmet Uysal and Yusuf Durak Abstract: Bacterial contamination in foods and antimicrobial resistance levels of common pathogenic strains causing food- borne disease are important in human health. Thus, typing technologies are important tools to determine primary sources of bacterial contamination. In this study, 40 Escherichia coli strains isolated from 85 food samples were evaluated in terms of genetic diversity, susceptibility to certain antibiotics, and plasmid profiles. Pulsed-field gel electrophoresis was used to identify the genetic relations of E. coli isolates. It was determined that the 40 E. coli strains revealed 32 different pulsotypes represented by 6 subtypes. Antibiotic susceptibility tests conducted by using a disc diffusion method against 15 antibiotics showed that although the isolates revealed 14 different types of resistance profiles, the strains showed the greatest resistance to ampicillin (77.5%), followed by ticarcillin– clavulanic acid (30%), tetracycline (22.5%), and cephalothin (14.5%). Plasmid isolations studies of the strains conducted by the method of alkaline lysis revealed that 18 (45%) of 40 E. coli strains contain 31 different plasmid bands ranging between 64.4 and 1 kb. The results showed that PFGE was a powerful method in tracking sources of food contamination and that the antibiotic resistance levels of food isolates were high and should be monitored. Key words: Escherichia coli, pulsed-field gel electrophoresis, antibiotic resistance, genetic diversity, plasmid profiles. Résumé : La contamination bactérienne des aliments et les niveaux de résistance aux antimicrobiens des souches pathogènes courantes qui causent les maladies d’origine alimentaire sont importants pour la santé humaine. En conséquence, les technologies de typage sont devenues des outils importants pour déterminer les sources primaires de contamination bactérienne. Dans cette étude, 40 souches d’Escherichia coli isolées de 85 échantillons d’aliments ont été évaluées en termes de diversité génétique, de susceptibilité a ` certains antibiotiques et de profils de plasmides. L’électrophorèse sur gel a ` champ pulsé (PFGE) a été utilisée pour identifier les relations génétiques des isolats d’E. coli. Nous avons déterminé que les 40 souches d’E. coli présentaient 32 pulsotypes différents représentés par 6 sous-types. Une méthode de diffusion par disque utilisée pour déterminer la sensibilité envers 15 antibiotiques montrait que même si les isolats présentaient quatorze types de profils de résistance différents, les souches étaient principalement résistantes a ` l’ampicilline (77,5 %), a ` la combinaison ticarcilline–acide clavulinique (30 %), a ` la tétracycline (22,5 %) et a ` la céfalotine (14,5 %). L’isolement des plasmides par la méthode de lyse alcaline a révélé que 18 (45 %) souches d’E. coli contenaient 31 bandes de plasmides allant de 64,4 kb a ` 1 kb. Les résultats ont montré que la PFGE était une méthode puissante pour dépister les sources de contamination alimentaire, et que les niveaux de résistance aux antibiotiques des isolats alimentaires étaient élevés et devraient être surveillés. Mots-clés : Escherichia coli, électrophorèse sur gel en champ pulsé, résistance aux antibiotiques, diversité génétique, profils de plasmides. [Traduit par la Rédaction] Introduction Bacterial food-borne pathogens are a major cause of mor- bidity and mortality throughout the world. Because of their significant impact on human health and well-being, the ability to carry out epidemiological investigations to determine the primary sources of bacterial contamination is important to improve public health. Among the many potentially patho- genic microorganisms introduced into the environment, coli- form bacteria, particularly Escherichia coli, which are common bacteria in the intestinal flora of human and other warm-blooded animals, have been widely used as a fecal contamination indicator in environment. It can be easily spread through water, soil, and food. Some groups of E. coli are the causative agents of many enteric infections, including diar- rhea, dysentery, hemolytic uremic syndrome, and bladder and kidney infection worldwide. Animal and human commensal and environmental enteric E. coli are supposed to be the Received 2 July 2012. Revision received 7 September 2012. Accepted 9 September 2012. Published at www.nrcresearchpress.com/cjm on 6 November 2012. A. Uysal and Y. Durak. Selcuk University, Science Faculty, Department of Biology, Campus, 42031 Selçuklu, Konya, Turkey. Corresponding author: Ahmet Uysal (e-mail: [email protected] and [email protected]). 1278 Can. J. Microbiol. 58: 1278 –1287 (2012) Published by NRC Research Press doi:10.1139/w2012-108 Can. J. Microbiol. Downloaded from www.nrcresearchpress.com by GLASGOW UNIVERSITY LIBRARY on 03/13/13 For personal use only.
Transcript
Pulsed-field gel electrophoresis typing, antibioticresistance, and plasmid profiles of Escherichiacoli strains isolated from foods
Ahmet Uysal and Yusuf Durak
Abstract: Bacterial contamination in foods and antimicrobial resistance levels of common pathogenic strains causing food-borne disease are important in human health. Thus, typing technologies are important tools to determine primary sources ofbacterial contamination. In this study, 40 Escherichia coli strains isolated from 85 food samples were evaluated in terms ofgenetic diversity, susceptibility to certain antibiotics, and plasmid profiles. Pulsed-field gel electrophoresis was used toidentify the genetic relations of E. coli isolates. It was determined that the 40 E. coli strains revealed 32 differentpulsotypes represented by 6 subtypes. Antibiotic susceptibility tests conducted by using a disc diffusion method against 15antibiotics showed that although the isolates revealed 14 different types of resistance profiles, the strains showed thegreatest resistance to ampicillin (77.5%), followed by ticarcillin–clavulanic acid (30%), tetracycline (22.5%), andcephalothin (14.5%). Plasmid isolations studies of the strains conducted by the method of alkaline lysis revealed that 18(45%) of 40 E. coli strains contain 31 different plasmid bands ranging between 64.4 and 1 kb. The results showed thatPFGE was a powerful method in tracking sources of food contamination and that the antibiotic resistance levels of foodisolates were high and should be monitored.
Résumé : La contamination bactérienne des aliments et les niveaux de résistance aux antimicrobiens des souchespathogènes courantes qui causent les maladies d’origine alimentaire sont importants pour la santé humaine. Enconséquence, les technologies de typage sont devenues des outils importants pour déterminer les sources primaires decontamination bactérienne. Dans cette étude, 40 souches d’Escherichia coli isolées de 85 échantillons d’aliments ont étéévaluées en termes de diversité génétique, de susceptibilité a certains antibiotiques et de profils de plasmides.L’électrophorèse sur gel a champ pulsé (PFGE) a été utilisée pour identifier les relations génétiques des isolats d’E. coli.Nous avons déterminé que les 40 souches d’E. coli présentaient 32 pulsotypes différents représentés par 6 sous-types. Uneméthode de diffusion par disque utilisée pour déterminer la sensibilité envers 15 antibiotiques montrait que même si lesisolats présentaient quatorze types de profils de résistance différents, les souches étaient principalement résistantes al’ampicilline (77,5 %), a la combinaison ticarcilline–acide clavulinique (30 %), a la tétracycline (22,5 %) et a la céfalotine(14,5 %). L’isolement des plasmides par la méthode de lyse alcaline a révélé que 18 (45 %) souches d’E. coli contenaient31 bandes de plasmides allant de 64,4 kb a 1 kb. Les résultats ont montré que la PFGE était une méthode puissante pourdépister les sources de contamination alimentaire, et que les niveaux de résistance aux antibiotiques des isolats alimentairesétaient élevés et devraient être surveillés.
Mots-clés : Escherichia coli, électrophorèse sur gel en champ pulsé, résistance aux antibiotiques, diversité génétique,profils de plasmides.
[Traduit par la Rédaction]
Introduction
Bacterial food-borne pathogens are a major cause of mor-bidity and mortality throughout the world. Because of theirsignificant impact on human health and well-being, the abilityto carry out epidemiological investigations to determine theprimary sources of bacterial contamination is important toimprove public health. Among the many potentially patho-genic microorganisms introduced into the environment, coli-
form bacteria, particularly Escherichia coli, which arecommon bacteria in the intestinal flora of human and otherwarm-blooded animals, have been widely used as a fecalcontamination indicator in environment. It can be easily spreadthrough water, soil, and food. Some groups of E. coli are thecausative agents of many enteric infections, including diar-rhea, dysentery, hemolytic uremic syndrome, and bladder andkidney infection worldwide. Animal and human commensaland environmental enteric E. coli are supposed to be the
Received 2 July 2012. Revision received 7 September 2012. Accepted 9 September 2012. Published at www.nrcresearchpress.com/cjmon 6 November 2012.
A. Uysal and Y. Durak. Selcuk University, Science Faculty, Department of Biology, Campus, 42031 Selçuklu, Konya, Turkey.
natural reservoir of pathogenic strains (Falkow 1996; Manning2010). In addition to such problems, the emergence of aconstant resistance to antibiotics in E. coli isolates, as evi-denced by an increase in their number, clearly demonstratesthat the increase in antibiotic resistance levels has been aglobal problem (Patoli et al. 2010). Because of the indiscrim-inate use of antibiotics, E. coli has developed resistance tosome of these agents that has resulted in failures in the treat-ment of the infectious diseases caused by E. coli. Moreover,E. coli can serve as reservoirs of resistance genes (van denBogaard et al. 2001), which have been efficiently exchangednot only with each other but also with other enteric patho-gens of humans and animals. The spread of resistance, andespecially multiple antibiotic resistance, among differententerobacteria is generally caused by the presence of trans-ferrable plasmids and their transference among one another.Because the prevalence of resistance in commensal E. coliis a useful indicator of antibiotic resistance in bacterialisolates from humans and environment, routine monitoringof antibiotic resistance and plasmid profiles of E. coliprovide data for antibiotic therapy and resistance control(O’Brien 1997).
There are some published data on the availability and prev-alence of E. coli in foods, such as chickens, beef, pork, turkey,mutton, vegetables, fruit juice, ice cream, cheese, yoghurt, andmany more (Adzitey et al. 2012; Apun et al. 2011; Badri et al.2009; Rahimi et al. 2011; Senkel et al. 2003; Tambekar et al.2006), but little is known about the genetic diversity andplasmid profiles of E. coli in various marketed foods. Thisstudy was therefore designed to determine the genetic diver-sities and genetic relations, susceptibility to antibiotics, andplasmid profiles of 40 E. coli strains isolated from 85 foodsamples, such as ice cream, chicken meats, cheeses, sausages,and cakes in Konya (Turkey). The specific objectives of thepresent study were (i) to display similarities and differencesamong E. coli strains with the help of pulsed-filed gel elec-trophoresis (PFGE) and, thus, to achieve detailed informationfor clonal relations; (ii) to assess the correlation betweenantibiotic resistance patterns and plasmid profiles of isolates;and (iii) to clarify potential bacterial risks related to foodsources.
Materials and methods
Sample collectionEighty-five food samples, including ice cream, chicken
meat, sausage, raw milk, cheese, and cake, were collectedfrom different markets, local bazaars, and patisseries in Konyacity between March 2009 and July 2010. Sterile containerswere used for collection of food samples. All samples wereshipped to the Microbiology Research Laboratory (ScienceFaculty, University of Selcuk) in a cooling box and processedimmediately upon arrival.
Isolation of E. coli from food samplesAll the samples (10 g or 10 mL (grams for solid samples
and millilitres for liquids)) were homogenized and then10-fold serially diluted with saline solution using standardlaboratory practices. Raw milk and ice cream samples weredirectly analyzed. A 0.1 mL volume of the diluted foodsample was inoculated onto eosin methylene blue agar(EMB; Merck, Darmstadt, Germany) medium plates and
incubated at 37 °C for 24 – 48 h (Dogan-Halkman et al.2003). After incubation, metallic green colonies on EMB agarwere isolated for suspicion of possible E. coli and transferredto another EMB plate.
Identification of E. coli isolates by the API 20E testThe API 20E test was performed in accordance with the
manufacturer’s protocol (BioMérieux, Marcy l’Etoile, France).All cultures were transferred onto 5% sheep blood agar plates priorto the inoculation of the API 20E strips. A bacterial suspensionapproximating a 0.5 McFarland standard was used for inoculation.All strips were incubated at 35 °C for 24 h. The addition of reagentsand the interpretation of reactions were performed in accordancewith the manufacturer’s directions. The 20 biochemical test reactionson the strip were converted into an octal profile number. Each profilenumber was then decoded by using the Analytical Profile Index.Apiweb was used to identify species belonging to Enterobacteri-aceae and to identify percentages.
Finally, Chromocult TBX (Tryptone Bile X-glucuronide)agar medium (Merck, Darmstadt, Germany) was used to verifyof identification. Plates were incubated at 35 °C for 4 h. Afterthe resuscitation of strains had taken place, plates were trans-ferred and incubated at 44 °C for 22 h. Bluish green colonieswere determined as E. coli.
Genotyping of E. coli isolates by PFGEPFGE was conducted to assess the genetic diversity of
E. coli isolates to analyze genetic similarities between bacteriaisolated from food samples, i.e., PFGE was used to tracksources of food contamination caused by E. coli isolated fromdifferent locations. Isolates were subtyped based on PFGEpatterns of XbaI-digested genomic DNA fragments in accor-dance with the standard protocol established by the Centers forDisease Control and Prevention (PulseNet; Centers for Dis-ease Control and Prevention 2008) and by Durmaz et al.(2009), with some modification. Bacterial strains were grownovernight on tryptic soy agar plates at 37 °C. Bacterial colo-nies were suspended in cell suspension buffer (CSB;10 mmol/L Tris–HCl, 50 mmol/L EDTA, 20 mmol/L NaCl,pH 7.2) and adjusted to 0.7, 0.8, and 1.0 absorbance (or equalto McFarland 4 turbidity) at 590 nm using a UV/Vis spectro-photometer. In CSB, 2% (m/v) low-melting point agarose(GIBCO BRL, Paisley, UK) was prepared. After completelymelting the agarose in a microwave oven and cooling themelted agarose to 45–50 °C, sodium dodecyl sulfate wasadded to a final concentration of 1%. An equal volume of cellsuspension (200 �L) was added to the agarose tubes, and theagarose – cell suspension mixture was gently mixed by pi-petting 2–3 times. About 100 �L of this mixture was carefullydispensed into appropriate wells of a reusable plug mould(Bio-Rad Laboratories, Hercules, California). After solidifica-tion, the plugs were transferred individually to round-bottomtubes containing 500 �L of cell lysis solution 1 (10 mmol/LTris–HCl, 50 mmol/L EDTA, 20 mmol/L NaCl, pH 7.2, ly-sozyme (2.5 mg/mL), proteinase K (1.5 mg/mL)). Cells werelysed at 55 °C in a shaking water bath for 1 h. After lysis, theplugs were transferred into another tube containing 500 �L ofcell lysis solution 2 (0.5 mol/L EDTA (pH 8.0), 1% sarcosyl,and proteinase K (400 �g/mL)) and incubated at 55 °C inshaking water bath for 2 h. Then lysis suspension was care-fully removed from the plugs, which had been rinsed 5 times
Uysal and Durak 1279
Published by NRC Research Press
Can
. J. M
icro
biol
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.
com
by
GL
ASG
OW
UN
IVE
RSI
TY
LIB
RA
RY
on
03/1
3/13
For
pers
onal
use
onl
y.
with TE buffer (10 mmol/L Tris–HCl (pH 7.2), 1 mmol/LEDTA). Each of these rinsing steps was performed at 50 °C ina shaking water bath for 15 min. Plugs were then stored in2 mL of TE buffer at 4 °C until they were ready for DNArestriction enzyme digestion. The DNA in agarose plugs wasdigested with 20 U of XbaI for at least 2 h at 37 °C in a waterbath. During the restriction step, 1% (m/v) of pulsed-fieldcertified agarose (Bio-Rad Laboratories) was prepared in100 mL of 0.5� TBE buffer (44.5 mmol/L Tris base,44.5 mmol/L boric acid, and 1 mmol/L EDTA, pH 8.0). Theplugs were loaded into the wells. DNA restriction fragmentswere separated using a CHEF-DR II system (Bio-Rad Lab-oratories, Nazareth, Belgium) at 14 °C, with 6 V/cm2 for20 h. The initial and final switch times were 10 and 50 s,respectively. The gel was stained with ethidium bromide(5 �g/mL) for 20 min, visualized under UV light, andphotographed using a Gel logic 2200 imaging system(Kodak Co., Rochester, New York, USA).
The DNA band profiles were analyzed by GelCompar IIsoftware (version 6.5; Applied Maths, Sint-Martens-Latem,Belgium). DNA profiles for each gel were normalized usingthe external reference strains (E. coli ATCC 25922). Finger-prints were clustered by using the Dice coefficient evaluatedby the unweighted-pair group method (UPGMA). A tolerance
and optimization of 0.5% was allowed to account for gel-to-gel differences. Isolates that had �80% pattern similaritywere considered highly closely related. According to thecriteria of Tenover et al. (1997), the strains were categorizedas follows: indistinguishable, closely related, possibly related,or different.
Antibiotic susceptibility testingThe standard Kirby–Bauer disk diffusion method was used
to determine the antibiotic sensitivity profiles of the E. coliisolates (Clinical and Laboratory Standards Institute 2008) for15 antimicrobial agents (Table 1). A 12-cm-diameter Mueller–Hinton agar medium plate was swabbed with brain heartinfusion broth inoculated with E. coli and incubated to aturbidity of 0.5 McFarland standards. Fifteen commerciallyprepared antimicrobial agent disks were place on the inocu-lated plates. The plates were incubated at 35 °C for 18–20 h.The diameters (in millimetres) of the clear zones of growthinhibition around the antimicrobial agent disks, including the6 mm disk diameter, were measured by using a ruler. Zonediameters were interpreted by using guidelines from the Clin-ical and Laboratory Standards Institute (2008). Esche-richia coli ATCC 25922 was used for quality control. Inaddition, multiple antibiotic resistances (MAR) indexing of
Table 1. Concentrations and diffusion zone breakpoints for resistance for antimicrobial agentstested in this study, sorted by class of antimicrobial agent.
E. coli strains were determined according to Krumperman(1983).
Plasmid analysisExtraction of plasmid DNA was done using the alkaline
lysis method of Bimboim and Doly (1979). The samples wereprocessed using gel electrophoresis to identify the number ofplasmid copies present in different isolates. For this purpose,an agarose gel of 0.8% (m/v) was used. Crude DNA extractsolutions were subjected to electrophoresis using a horizontalapparatus the Submerged Agarose Gel Electrophoresis (AE-6125, Atto; Tokyo, Japan) and a constant voltage (100 V)power source for approx. 1.5 h, as described by Aladag et al.(2009). DNA bands were visualized using a 366 nm UVtransilluminator and photographed with the UVP, GelDoc ItImaging System. Plasmid molecular masses were estimated byelectrophoresis with plasmids of known molecular mass fromE. coli V517.
Results
Escherichia coli recovery from food samplesIn this study, 85 food samples were examined with regard
to the presence of E. coli and 40 E. coli strains in total wereisolated and identified. According to the identification re-sults, which were conducted with API 20E, microorganismsincluded in Enterobacteriaceae of different types and spe-cies were identified in food samples in addition to E. coli(Table 2). These microorganisms were identified as Entero-bacter cloacae, Enterobacter asburiae, Enterobacter gergov-iae, Lectercia adecarboxylata, Klebsiella pneumoniae subsp.pneumoniae, Klebsiella oxytoca, Kluyvera sp., and Citrobac-ter koseri. The sources of the microorganisms isolated andidentified from foods are shown in Table 2. After identifica-tion, TBX agar medium was used to verify the identifiedE. coli, and the 40 verified strains were used in the studies.
Determination of genetic diversities and relations ofE. coli isolates through PFGE method
Figure 1 shows PFGE band profiles produced by XbaIenzyme and dendrogram generated depending on Dice simi-larity coefficient through the UPGMA analysis conducted inthe obtained gel images of E. coli strains isolated from foodsamples.
Thirty-two different pulsotypes, from group A to group AG,and 6 different subtypes under these groups were identifiedamong the food isolates. Group Y has the most subtypes(Y1–Y4) and was identified as the largest group among thegroups. A 95% similarity was observed between numberedstrains DSG 14 (Y1) and DSG 15 (Y2) from members of thisgroup, and thus, it was believed that these 2 strains are closelyrelated. On the other hand, strain DSG 11 revealing the Y3subtype showed similarity to these 2 strains at a ratio of91.9%. Therefore, it was determined that strain DSG 11 maybe closely related to numbered strains DSG 14 and DSG 15.The similarity of DSG 13 included in the same pulsotype andrevealing the Y4 subtype with the other 3 strains in the groupwas determined as 85.4% (Fig. 1). According to this ratio,DSG 13 was believed to possibly be related to the other 3strains in group Y. All strains included in group Y originatedfrom ice cream samples collected from different patisseriesand markets at different times.
The other pulsotype generating 2 subtypes in the dendro-gram is Group E (Fig. 1). The similarity between the DSG 30(E1) and DSG 28 (E2) numbered strains was defined as 88.2%.Accordingly, it was concluded that the strains isolated fromdifferent ice cream samples were possibly related. Group Joriginating from ice cream samples and generating 2 sub-types has only 2 members. Based on an 80% similarity ratiobetween the DSG 4 (J1) and DSG 5 (J2), these 2 strainswere defined to be possible related strains. The status ofpotentially related isolate was observed between the DSG19 (P1) and DSG 20 (P2) numbered strains correspondinglybased on the similarity ratio of 88.2%. Furthermore, strainsDSG 20 and DSG 24 belonging to AA pulsotype weredetermined as possible related isolates due to the geneticrelationship ratio of 90.9% (Fig. 1). The last group generat-ing 2 subtypes in dendrogram is Group AF. DSG 31 (AF1) andDSG 32 (AF2) are clonally close-related strains at a ratio of96.6% (Fig. 1).
The strains outside the subtypes existing in the aforemen-tioned 6 groups were determined as nonrelated isolates. Thesimilarity between these strains was distributed within arange of 21.6%–76.4%. According to results of the PFGEanalysis, it was concluded that genetic diversities amongfood isolates are very high and, accordingly, clonal rela-tions are at a low ratio.
Table 2. Total isolate numbers and sources of microorganisms isolated and identified from various food samples.
Fig. 1. Pulsed-field gel electrophoresis (PFGE) XbaI digestion patterns and clonal analysis of 40 Escherichia coli isolates obtained fromfood samples. The dendrogram was constructed with the use of Dice similarity coefficient (0.5% tolerance) and UPGMA clustering methodby comparison of XbaI PFGE patterns. Numbers represent the distance values between the respective isolates.
1282 Can. J. Microbiol. Vol. 58, 2012
Published by NRC Research Press
Can
. J. M
icro
biol
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.
com
by
GL
ASG
OW
UN
IVE
RSI
TY
LIB
RA
RY
on
03/1
3/13
For
pers
onal
use
onl
y.
Antibiotic resistance phenotypes of E. coli isolatesA total of 40 E. coli isolates were screened for evaluation of
their antimicrobial sensitivities against 15 antimicrobial agents(Table 3). In addition, these isolates were grouped into 14categories based on the occurrence of different individualantibiotics resistances as well as co-occurrence of a combina-tion of different resistance phenotypes (Table 4). Thirty-two(80%) of the E. coli isolates were resistant to at least one of theantimicrobial agents tested. Overall, 17 (42.5%) of the isolateswere multidrug resistant (resistant to �2 antibiotics). Fifteenstrains were detected as resistant to only one antibiotic. Eight(20%) of the 40 isolates were not resistant to any of 15antimicrobial agents. Ampicillin is the antibiotic to whichlargest numbers of E. coli isolates were resistant, at a ratio of77.5%, followed by ticarcillin–clavulanic acid (30%), tetracy-cline (22.5%), cephalothin (14.5%), ciprofloxacin, cefotaximeand chloramphenicol (10%), ceftazidime (7.5%), cefuroximeand gentamicin (5%), and amoxicillin–clavulanic acid (2.5%)(Table 3). It was determined that piperacillin–tazobactam andmeropenem were the most effective antibiotics. No resistancesto these antibiotics were observed. Except for these, the strainswere highly sensitive to netilmicin and amikacin (95%), fol-lowed by gentamicin and ceftazidime (92.5%), cefotaximeand chloramphenicol (90%), ciprofloxacin and amoxicillin–clavulanic acid (87.5), cephalothin (67.5%), ticarcillin–clavulanic acid (65%), and tetracycline (62.5%).
According to the data obtained from the study, 14 different resis-tance profiles were recorded among the food isolates (Table 4). Themost prevalent resistance pattern among the resistant strainswas resisto-type I (ampicillin) with a distribution percentageof 46.88%, followed by resisto-type VI (ampicillin, ticarcillin–clavulanic acid, tetracycline) with a distribution percentage of15.63%. The profiles except for the 2 profiles were representedby one strain, and the distribution ratios of the profiles werefound to be 3.13%. The MAR index of the strains that wereresistant only to ampicillin was 0.07. The MAR index of DSG3 was found to be 0.47, and it was the highest value among thefood isolates. Considering MAR indexes, 8 strains (DSG 3, 4,8, 15, 16, 37, 38, and 39) revealing resisto-types VII to XIV
had an index higher than 0.2. According to Krumperman(1983), it is claimed that strains having an index higher than0.2 have come from a location in which antibiotics are inten-sively use, in other words, the DSG 3, 4, 8, 15, 16, 37, 38, and39 numbered strains were isolated from locations in whichantibiotics have been used intensively.
Plasmid analysis of E. coli isolatesFigure 2 shows gel image of plasmid bands extracted from
E. coli isolates. Plasmids were detected in 18 (45%) of thestrains isolated from foods while no plasmids were detected inthe other 22 strains (55%). A total of 31 different plasmidbands were detected and their plasmid sizes were observedwithin the range of 64.4 kb – 1 kb (Table 5). It was determinedthat the strain comprising the maximum number of plasmidwas DSG 1. Plasmid sizes of this strain ranged from 44.8 to2.3 kb. It was found that the DSG 3, 18, 40 numbered strainshave 4 plasmids while the DSG strains number 29, 31, and 36have 3 plasmids. The highest plasmid size was observed in theDSG 21 numbered strain as 64.4 kb, while the lowest one wasfound in the DSG 40 numbered strain to be 1 kb.
It was observed that 7 different plasmid profiles are presentamong the strains (Table 5); these profiles and plasmid sizesare 60.1 kb for Profile 1, 5.7 kb for Profile 2, 4.0 kb for Profile3, 3.5 kb for Profile 4, 5.0 and 2.3 kb for Profile 5, 5.3 and3.2 kb for Profile 6, and 45.4 kb for Profile 7. Strains DSG 1,3, and 9 have 2 different plasmid profiles while others haveonly 1 profile. Profiles 6 and 3 are the most prevalent profileamong the strains containing plasmid. In addition, it wasdetermined that certain plasmids did reveal a profile becausethey exist in only 1 strain.
DiscussionThe growth of coliform bacteria, especially E. coli, among
certain potentially pathogenic microorganisms present in theenvironment have been studied by researchers in determiningthe quality of food (Manning 2010). Escherichia coli strainshad not been considered a significant pathogen until it wasdetected as the cause of epidemics that occurred in certaincountries in the 1980s (Riley et al. 1983). The frequency ofnews items related to this pathogen in various foods and waterindicates a need for rapid, accurate, and cost-effective identi-fication systems so as to reduce the public’s exposure to E. coliinfection. The objective of the present study was to displaysimilarities and differences between E. coli strains with thehelp of PFGE and, thus, to achieve detailed information forclonal relations. According to the results obtained, high-degreegenetic diversity (80%, 32/40) was detected among E. colistrains. Thirty-two pulsotypes were obtained and it was ob-served that they were represented by 6 subtypes. It was ob-served that some of these strains, which were detected at thelevel of subtype, are closely related while some of them arepossibly related. Of 32 pulsotypes, 26 were generated byindividual strains and the ratio among food isolates was 65%(26/40). This situation increases the rate of genetic diversity.Similarly, Apun et al. (2006) used PFGE to compare the clonalrelationships of E. coli strains isolated from meat samples inMalaysia, and they observed a high degree of genetic diversityamong the isolates (88.2%, 45/51). Our findings coincide withthe high heterogeneity results obtained by researchers in thepast. A high degree of genetic diversity in E. coli strains
Table 3. Antibiotic susceptibility patterns of Escherichia coliisolates.
*See Table 1 for explanation of antibiotic abbreviations.
Uysal and Durak 1283
Published by NRC Research Press
Can
. J. M
icro
biol
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.
com
by
GL
ASG
OW
UN
IVE
RSI
TY
LIB
RA
RY
on
03/1
3/13
For
pers
onal
use
onl
y.
isolated from the pork slaughterhouse and pork carcasses wasalso found by Wu et al. (2009), and the results of the research-ers were supported by our findings. In a study conducted byBadri et al. (2009), 74 E. coli strains isolated from groundmeat, sausage, turkey, and well water samples were evaluatedin terms of genetic diversity by using PFGE. Forty-two pul-sotypes were defined among the isolates (56.7%, 42/74). In ourstudy, the genetic heterogeneity rate of food isolates wasdetermined to be 80%. The authors manifested that PFGE is avery useful and powerful technique for the discrimination offood-borne bacterial isolates, and this opinion was confirmedby our findings. Thorsteinsdottir et al. (2010) investigated the
genetic diversity and antibiotic resistance of 419 E. coli strainsisolated from healthy pork and chicken and from pork meat,chicken meat, and staff of slaughterhouse. These authors re-ported that the genetic diversity rate of isolates was deter-mined to be 78.6%, and the ratio for genetic diversity found inour study in food isolates has a strong consistency with theresults obtained by researchers in past studies. Also, Thor-steinsdottir et al. (2010) found that the antibiotic resistancerates of strains against enrofloxacin were observed within therange of 54.1%–34.1%. In our study, antibiotic resistance rateof food isolates against ciprofloxacin, an antibiotic similar tothe quinolone enrofloxacin, was determined to be 10%, and it
Table 4. Antibiotic resistance patterns of Escherichia coli isolates.
*See Table 1 for explanation of antibiotic abbreviations.†MAR, multiple antibiotic resistance.
Fig. 2. Image of plasmids extracted from Escherichia coli strains isolated from foods on agarose gel. Escherichia coli V517 plasmid DNAfragments were used as a size standard.
1284 Can. J. Microbiol. Vol. 58, 2012
Published by NRC Research Press
Can
. J. M
icro
biol
. Dow
nloa
ded
from
ww
w.n
rcre
sear
chpr
ess.
com
by
GL
ASG
OW
UN
IVE
RSI
TY
LIB
RA
RY
on
03/1
3/13
For
pers
onal
use
onl
y.
was concluded that this quinolone resistance was very lowcompared with the results of Thorsteinsdottir et al. (2010).Consequently, our findings revealed that E. coli strains iso-lated from different food samples were not closely related witheach other, except for some groups. While some of the strainsisolated from ice cream were close related with each other, themajority of strains isolated from chicken meat, cake, and rawmilk were categorized as different strains. For this reason,genetic relationship rates of strains were found at a low ratio,although the genetic diversity rate was very high. In our study,PFGE was verified as being a very useful method in deter-mining the clonal relationships of food isolates.
According to the results obtained from our study, a highdegree of resistance to ampicillin was observed in E. colistrains isolated from food samples (77.5%), followed by re-sistance to ticarcillin–clavulanic acid at a ratio of 30%, tetra-cycline at 22.5%, and cephalothin at 14.5%. Fourteenresistance profiles (resisto-type) were identified among foodisolates. In a study conducted by Lei et al. (2010), it wasreported that antibiotic resistance rates of E. coli strains iso-lated from food animal products were determined as 86.5%against tetracycline, 64% against ampicillin, and 47.8%against chloramphenicol. Sáenz et al. (2001) found that E. colistrains isolated from ground meat and turkey meats wereresistant to tetracycline at a ratio of 53%, ampicillin at 47%,gentamicin at 17%, and amoxicillin–clavulanic acid at 13%.Similarly, Aslam and Service (2006) determined that resis-tance rates of strains isolated from ground meat and carcasseswere 38% for tetracycline, 9% for ampicillin, and 6% forstreptomycin. According to our findings, resistance to ampi-cillin was found to be higher in the strains isolated from foodsamples than in those isolated from food animal products andmeat samples. However, our tetracycline and chloramphenicolresistances were observed to be lower than that found by Leiet al. (2010), Saenz et al. (2001), and Aslam and Service
(2006). In addition, Aslam and Service (2006) investigated thegenetic diversity of strains by using the random amplifiedpolymorphic DNA method and determined that the majority ofstrains revealed high degree of genetic diversity, as was foundin our study conducted by the PFGE method. Senkel et al.(2003) reported that the antibiotic resistance to tetracyclinewas 64%. This ratio was higher than our findings. In anotherstudy, ampicillin resistance rates of E. coli isolated fromcommercial chicken eggs were 42.9%, followed by tetracy-cline with a ratio of 40% (Arathy et al. 2011). Rahimi et al.(2011) found that E. coli strains isolated from traditionalcheese, ice cream, and yoghurt samples were resistant toampicillin and gentamicin at a ratio of 44.4%, erythromycin at33.3%, and amoxicillin and tetracycline at 11.1%. Accordingto our findings, resistance to ampicillin and tetracycline wasfound to be higher in the strains isolated from food samplesthan in those isolated from cheese, ice cream, and yoghurtsamples. Nabi et al. (2011) observed 100% resistance toamoxicillin and amoxiclav in E. coli strains isolated fromtraditional ice cream samples. This ratio was much higher thanthe 2.5% resistance ratio to amoxicillin–clavulanic acid thatwe detected in the isolates from food samples. Farzana et al.(2009) reported that E. coli strains isolated from milk productwere susceptible to amikin, amoxil, ceftran, and augmentin ata ratio of 100% and resistant to ampicillin at a ratio of 100%.In our study, food isolates were susceptible to amikacin andamoxicillin–clavulanic acid at a ratio of 100% and 97.5%,respectively. Our results obtained from food isolates correspond tothe results of Farzana et al. (2009) with respect to the ratio ofresistance to ampicillin and susceptibility to amoxicillin–clavulanic acid and amikacin. Ryu et al. (2012) reported thatthe antibiotic resistance rates of E. coli strains isolated fromcommercial fish and seafood against tetracycline, streptomy-cin, cephalothin, ampicillin, and ticarcillin were 30.7%,12.8%, 11.7%, 6.7%, and 6.1%, respectively. According to thefindings generated by our study, resistance to ampicillin ishigher than that found by Ryu et al. (2012), while resistancesto tetracycline and cephalothin coincide with the values re-ported by the authors.
Although resistance to ampicillin varies between locationsin similar studies in the literature, it is generally seen at lowratios. However, in our study ampicillin resistance was veryhigh. Accordingly, it may be said that the increase in synthesisof �-lactamase enzymes inhibiting �-lactam antibiotics, suchas ampicillin and penicillin, has also become prevalent in foodisolates, and it may be considered as a significant indicator forthe increasing resistance to antibiotics. The resistance devel-oped by bacteria to antibiotics is a result of evolution andbacterial genetics. The main reason for the development ofresistance to antibiotics is the unrestrained use of antibiotics(Kaya et al. 2006). Furthermore, antibiotic resistance may beacquired via horizontal transmission of genes from othersources, such as water contaminated with human sewage, soil,and animal feces. The spread of resistance genes from animalto human E. coli, and from commensal to pathogenicEnterobacteriaceae, has been documented (www.who.int/foodsafety/publications/micro/en/report.pdf). The role of foodin such transfer is not yet known and needs further investiga-tion (Newell et al. 2010).
In comparing strain antibiotic resistance and plasmid pro-files, resistance to ampicillin was observed for all 15 strains,
Table 5. Plasmid sizes and profiles of Escherichia coli isolates.
which have or do not have plasmids (Table 5). Althoughstrains DSG 7, 8, 13, 14, 15, 16, and 37 revealing resisto-typesIII, VIII, IV, II, XI, IX, and X, respectively, were resistant tosome antibiotics, they have no plasmids. Strains DSG 5, 9, 30,35, and 38 were resistant to the same antibiotics (ampicillin,ticarcillin–clavulanic acid, tetracycline). Some of the strains(DSG 9 and 30) have 2 plasmids, but the sizes of the plasmidswere not the same. While DSG 38 has only one plasmid, DSG5 and 35 have no plasmids. Thus, it may be concluded that theresistances of the isolates are not mediated by plasmids, andconsequently, this resistance may be chromosomal. Eight ofthe 40 food isolates (DSG 1, 2, 18, 23, 24, 29, 33, and 34) weresusceptible to all antimicrobial agents tested. Nevertheless,they comprise plasmids in different numbers and sizes. Nocorrelation was observed between resistance in E. coli strainsisolated from food samples to antibiotics and their plasmid orplasmids. As an alternative, it may be considered that theplasmid of the resistant strains may be lost or integrated withthe chromosome, in other words, an episome state may haveoccurred. For confirmation of this idea, conjugation assaysshould be performed. In their study, Alam et al. (2010) deter-mined 24 different plasmids from 54 isolated strains andinvestigated the resistance of these strains to antibiotics. Theresearchers emphasized the conclusion that there was no rela-tionship between resistance profiles and plasmid size, number,and profiles and that plasmids showed a random distributionamong the strains. Our findings support these findings. Simi-larly, Teophilo et al. (2002) determined plasmid profiles ofE. coli strains isolated from shrimp and fish. It was reportedthat only 2 resistant strains isolated from fish have plasmids,and plasmid bands have not been detected in other resistantstrains. Consequently, it was reported that there was no sig-nificant correlation between antibiotic resistance in E. colistrains isolated from fish and shrimp and their plasmid profiles.In our study, 31 different plasmid bands were identified infood isolates, and no correlation between plasmids and resis-tance profiles was observed. The results obtained coincidewith the results found by Teophilo et al. (2002). Contrary tothe data obtained from our plasmid study, Myaing et al. (2005)and Umolu et al. (2006) reported that there was a possiblecorrelation between plasmids and antibiotic resistance patternsof E. coli strains isolated from meat samples, but this was notcertain because a conjugation assays were not performed bythe researchers. Akter et al. (2011) performed conjugationassays for plasmids extracted from E. coli strains in freshvegetables. It was determined that there was a certain corre-lation between plasmids and amoxicillin resistance.
According to our findings, the presence of E. coli and certaincoliform microorganisms in food samples is an indicator of foodcontamination. Production and sales of foods in patisseries andmarkets are potentially the most important stage for bacterialcontamination. The determination that pathogenic microorgan-isms, such as Klebsiella pneumoniae, Klebsiella oxytoca, Entero-bacter cloacae, exist in foods indicates that public health is atstake. It has been observed that the production and sales staff ofmarkets and patisseries do not comply with hygiene rules enough,and as a result, resistant isolates from the human gut may readilycontaminate foods. Further epidemiological studies should beconducted to identify the primary sources of bacterial contami-nation to protect public health. Molecular typing methods such asPFGE, random amplified polymorphic DNA, amplified fragment
length polymorphism should be used together to rapidly andaccurately detect such harmful bacteria and their genetic relation-ships and detailed clonal sources. Also, we believe that munici-palities and public health laboratories should conduct the requiredcontrols and regulations in terms of food safety and quality inparticular.
Antimicrobial resistance is a rapidly changing challenge. Theincrease in resistance to antibiotics in food isolates tested in our studyis obvious upon comparison with the data in the literature. It has beensuggested that the consumption of contaminated food may be themain route of dissemination of resistance from food into humanpopulations. Contaminated food may serve as a vehicle to transportresistant bacteria and resistance genes between animals and humans.The evolution and adaptation of the bacterial agents and the changesin antimicrobial use, whether commercially or legislatively driven,are major influences on the impact of antimicrobial resistance onpublic health. Therefore, in the future, antimicrobial resistance levelsof E. coli strains in foods should be monitored frequently for foodsafety, and the monitoring of antimicrobial usage in human andanimals should be an integral part of the prevention and control ofantimicrobial resistance. Such data provide useful information ontrends in resistance and are necessary for risk assessment and riskmanagement and are a basis for choosing, implementing, and eval-uating interventions.
AcknowledgementWe would like to thank Selcuk University Scientific Re-
ReferencesAdzitey, F., Liew, C.Y., Aronal, A.P., and Huda, N. 2012. Isolation
of Escherichia coli from ducks and duck related samples. Asian J.Anim. Vet. Adv., 7(4): 351–355. doi:10.3923/ajava.2012.351.355.
Akter, S., Un-Nabi, R., Rupa, F.A., Bari, M.L., and Hossain, M.A.2011. Antibiotic resistance and plasmid profiles in bacteria isolatedfrom market-fresh vegetables. Agric., Food Anal. Bacteriol. 1:140–149.
Aladag, M.O., Durak, Y., and Uysal, A. 2009. Investigation ofimipenem and meropenem susceptibilities, plasmid profiles andESBL characteristic of Klebsiella pneumoniae isolated from uri-nary tract infections. World Appl. Sci. J. 7(3): 378–381.
Alam, M.J., Rahman, M.T., Siddique, M.P., Khan, M.F.R., andRahman, M.B. 2010. Antibiogram and plasmid profiling of E. coliisolates. Int. J. BioResearch, 1(3): 1–7.
Apun, K., Chang, P.P., Sim, E.U.H., and Micky, V. 2006. Clonaldiversity of Escherichia coli isolates from marketed beef in EastMalaysia. World J. Microbiol. Biotechnol. 22(7): 661–667. doi:10.1007/s11274-005-9086-0.
Apun, K., Kho, K.L., Chong, Y.L., Hashimatul, F.H., Abdullah,M.T., Rahman, M.A., et al. 2011. Detection of Escherichia coliO157:H7 in wildlife from disturbed habitats in Sarawak, Malaysia.Res. J. Microbiol. 6(2): 132–139. doi:10.3923/jm.2011.132.139.
Arathy, D.S., Vanpee, G., Belot, G., Mathew, V., De Allie, C., andSharma, R. 2011. Antimicrobial drug resistance in Escherichia coliisolated from commercial chicken eggs in Grenada, West Indies.West Indian Med. J. 60(1): 53–56. PMID:21809712.
Aslam, M., and Service, C. 2006. Antimicrobial resistance and ge-netic profiling of Escherichia coli from a commercial beef packingplant. J. Food Prot. 69(7): 1508–1513. PMID:16865878.
Badri, S., Fassouane, A., Filliol, I., Hassar, M., and Cohen, N. 2009.
Clonal analysis of Escherichia coli strains isolated from food bypulsed-field gel electrophoresis. Internet J. Food Safety, 11: 44–49.
Bimboim, H.C., and Doly, J. 1979. A rapid alkaline extraction pro-cedure for screening recombinant plasmid DNA. Nucleic AcidsRes. 7(6): 1513–1523. doi:10.1093/nar/7.6.1513. PMID:388356.
Centers for Disease Control and Prevention. 2008. Standardizedlaboratory protocol for molecular subtyping of Escherichia coliO157:H7, non-typhoidal Salmonella serotypes, and Shigella son-nei by pulsed field gel electrophoresis (PFGE) [online]. Availablefrom http://www.cdc.gov/pulsenet/protocols/ecoli_salmonella_shigella_protocols.pdf. [accessed 27 June 2012]
Clinical and Laboratory Standards Institute (CLSI). 2008. Perfor-mance standards for antimicrobial susceptibility testing, Eigh-teenth informational supplement. CLSI document M100-S18.Wayne, Penn., USA. pp. 34–38.
Dogan-Halkman, H.B., Çakır, ·I., Keven, F., Worobo, R.W., andHalkman, A.K. 2003. Relationship among fecal coliforms andEscherichia coli in various foods. Eur. Food Res. Technol. 216:331–334.
Durmaz, R., Otlu, B., Koksal, F., Hosoglu, S., Ozturk, R., Ersoy, Y.,et al. 2009. The optimization of a rapid pulsed-field gel electro-phoresis protocol for the typing of Acinetobacter baumannii, Esch-erichia coli and Klebsiella spp. Jpn. J. Infect. Dis. 62(5): 372–377.PMID:19762987.
Falkow, S. 1996. The evolution of pathogenicity in Escherichia coli, Shi-gella, and Salmonella. In Escherichia coli and Salmonella: cellular andmolecular biology. 2nd ed. Edited by F.C. Neidhardt, American Societyfor Microbiology, Washington, D.C., USA. pp. 2723–2769.
Farzana, K., Akhtar, S., and Jabeen, S. 2009. Prevalence and antibi-otic resistance of bacteria in two ethnic milk based products. Pak.J. Bot. 41(2): 935–943.
Kaya, O., Akçam, F.Z., Uyar, C., Demir, C., and Yaylı, G. 2006.2000–2004 yılları arasında izole edilen üropatojen Escherichiacoli suslarında artan antibiyotik direnci. S.D.Ü. Tıp Fak. Derg.13(4): 22–26.
Krumperman, P.H. 1983. Multiple antibiotic resistance indexing ofEscherichia coli to identify high-risk sources of fecal contamina-tion of foods. Appl. Environ. Microbiol. 46(1): 165–170. PMID:6351743.
Lei, T., Tian, W., He, L., Huang, X.H., Sun, Y.X., Deng, Y.T., et al.2010. Antimicrobial resistance in Escherichia coli isolates fromfood animals, animal food products and companion animals inChina. Vet. Microbiol. 146(1–2): 85–89. doi:10.1016/j.vet-mic.2010.04.025. PMID:20605690.
Myaing, T.T., Saleha, A.A., Arifah, A.K., and Raha, A.R. 2005.Antibiotic resistance and plasmid carriage among Escherichia coliisolates from chicken meat in Malaysia. In Applications of gene-based technologies for improving animal production and health indeveloping countries. Edited by H.P.S. Makkar and G.J. Viljoen.IAEA Press, Netherlands. pp. 521–527.
Nabi, S.F., Hossein, M.D., Mortaza, M.M., and Mohammad, G. 2011.Antimicrobial susceptibility of Escherichia coli isolated from tra-ditional ice cream in Gonabad. Ofogh-e-Danesh Journal, 16(4):1–4.
Newell, D.G., Koopmans, M., Verhoef, L., Duizer, E., Aidara-Kane,A., Sprong, H., et al. 2010. Food-borne diseases — the challengesof 20 years ago still persist while new ones continue to emerge. Int.J. Food Microbiol. 139(Suppl. 1): S3–S15. doi:10.1016/j.ijfood-micro.2010.01.021. PMID:20153070.
O’Brien, T.F. 1997. The global epidemic nature of antimicrobialresistance and the need to monitor and manage it locally. Clin.Infect. Dis. 24(Suppl. 1): S2–S8. doi:10.1093/clinids/24.Supplement_1.S2. PMID:8994775.
Patoli, A.A., Patoli, B.B., and Mehraj, V. 2010. High prevalence ofmulti-drug resistant Escherichia coli in drinking water samplesfrom Hyderabad. Gomal Journal of Medical Sciences, 8(1): 23–26.
Rahimi, E., Chaleshtori, S.S., and Parsaei, P. 2011. Prevalence andantimicrobial resistance of Escherichia coli O157 isolated fromtraditional cheese, ice cream and yoghurt in Iran. Afr. J. Microbiol.Res. 5(22): 3706–3710. doi:10.5897/AJMR11.497.
Riley, L.W., Remis, R.S., Helgerson, S.D., McGee, H.B., Wells, J.G.,Davis, B.R., et al. 1983. Hemorrhagic colitis associated with a rareEscherichia coli serotype. N. Engl. J. Med. 308(12): 681–685.doi:10.1056/NEJM198303243081203. PMID:6338386.
Ryu, S.H., Park, S.G., Choi, S.M., Hwang, Y.O., Ham, H.J., Kim,S.U., et al. 2012. Antimicrobial resistance and resistance genes inEscherichia coli strains isolated from commercial fish and seafood.Int. J. Food Microbiol. 152(1–2): 14–18. doi:10.1016/j.ijfoodmi-cro.2011.10.003. PMID:22071288.
Sáenz, Y., Zarazaga, M., Brinas, L., Lantero, M., Ruiz-Larrea, F.,and Torres, C. 2001. Antibiotic resistance in Escherichia coliisolates obtained from animals, foods and humans in Spain. Int.J. Antimicrob. Agents, 18(4): 353–358. doi:10.1016/S0924-8579(01)00422-8. PMID:11691568.
Senkel, I.A., Jr, Jolbitado, B., Zhang, Y., White, D.G., Ayers, S., andMeng, J. 2003. Isolation and characterization of Escherichia colirecovered from Maryland apple cider and the cider productionenvironment. J. Food Prot. 66(12): 2237–2244. PMID:14672219.
Tambekar, D.H., Hirulkar, N.B., Kalikar, M.V., Patil, Y.S., andGulhane, S.R. 2006. Prevalance of thermotolerant Escherichia coliin drinking water and its multidrug resistance. Res. J. Microbiol.1(5): 458–462. doi:10.3923/jm.2006.458.462.
Tenover, F.C., Arbeit, R.D., Goering, R.V., Molecular Typing Work-ing Group of the Society for Healthcare Epidemiology of America.1997. How to select and interpret molecular strain typing methodsfor epidemiological studies of bacterial infections: a review forhealthcare epidemiologists. Infect. Control Hosp. Epidemiol.18(6): 426–439. doi:10.1086/647644. PMID:9181401.
Teophilo, G.N., dos Fernandes Vieira, R.H., dos Prazeres Rodrigues,D., and Menezes, F.G.R. 2002. Eschericha coli isolated fromseafood: toxicity and plasmid profiles. Int. Microbiol. 5(1): 11–14.doi:10.1007/s10123-002-0052-5. PMID:12102230.
Thorsteinsdottir, T.R., Haraldsson, G., Fridriksdottir, V., Kristinsson,K.G., and Gunnarsson, E. 2010. Prevalence and genetic relatedness ofantimicrobial-resistant Escherichia coli isolated from animals, foods andhumans in Iceland. Zoonoses Public Health, 57(3): 189–196. doi:10.1111/j.1863-2378.2009.01256.x. PMID:19912612.
Umolu, P.I., Ohenhen, E.R., Okwu, I.G., and Ogiehor, I.S. 2006.Multiple antibiotic resistant index and plasmid of Escherichia coliin beef in Ekpoma. The Journal of American Science, 2(3): 1–3.
van den Bogaard, A.E., London, N., Driessen, C., and Stobberingh,E.E. 2001. Antibiotic resistance of fecal Escherichia coli in poultry,poultry farmers and poultry slaughterers. J. Antimicrob. Chemother.47(6): 763–771. doi:10.1093/jac/47.6.763. PMID:11389108.
Wu, S., Dalsgaard, A., Vieira, A.R., Emborg, H.D., and Jensen, L.B.2009. Prevalence of tetracycline resistance and genotypic analysis ofpopulations of Escherichia coli from animals, carcasses and cutsprocessed at a pig slaughterhouse. Int. J. Food Microbiol. 135(3):254–259. doi:10.1016/j.ijfoodmicro.2009.07.027. PMID:19717199.
