382

Journalof Food Protection, Vol. 66, No. 3, 2003, Pages 382– 389

Viabilityof Clostridiumperfringens, Escherichia coli, and Listeria monocytogenes SurvivingMild Heat or AqueousOzone Treatment onBeef Followedby Heat, Alkali,or SaltStress †

JOHN S.NOVAK 1* AND JAMES T.C.YUAN 2

1U.S.Department of Agriculture, Agricultural Research Service, EasternRegional Research Center,Microbial Food Safety Research Unit,600 East MermaidLane, Wyndmoor, Pennsylvania 19038; and 2AmericanAir Liquide, Chicago Research Center,5230 South East Avenue, Countryside,Illinois 60525, USA Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 MS02-200:Received 20June 2002/ Accepted 18October 2002

ABSTRACT

Thethreat of pathogen survival following ozone treatment of meatnecessitates careful evaluation of themicroorganisms survivingunder such circumstances. The objective of this study was to determine whether sublethal aqueous ozone treatment (3 ppm of O3 for5 min)of microorganisms on beef surfaces would result in increased or decreased sur vivalwith respect to subsequentheat, alkali, or NaCl stress. A mildheat treatment (55 8Cfor30 min) was used for comparison. Reductions in three-straincocktails of Clostridiumperfringens, Escherichia coli O157:H7,and Listeriamonocytogenes onbeef following theheat treatment were 0.14, 0.77, and 1.47 log 10 CFU/g,respectively, whereas reductions following ozone treatment were 1.28,0.85, and 1.09 log 10 CFU/g,respectively. C.perfringens cellsexhibited elevated heat resistance at 60 8C (D60 [time at 608Crequiredto reduce the viable cell population by 1 log 10 unitsor 90%] 5 17.76min) following heat treatment of beef (558Cfor30 min) but exhibited reduced viability at 60 8Cfollowingozone treatment ( D60 5 7.64min) compared with the viabilityof untreated control cells ( D60 5 13.84min). The D60-values for L.monocytogenes and E. coli O157:H7following heatand ozone exposures were not signiŽ cantly different ( P . 0.05). C.perfringens cellsthat sur vivedozone treatment did notexhibit increased resistance to pH (pH 6 to12) relative to non-ozone-treated cells when grown at 37 8Cfor24 h.Theheat treatmentalso resulted in decreased numbers of sur vivingcells above and below neutral pH valuesfor both E. coli O157:H7 and L.monocytogenes relativeto those of non-heat-treated cells grown at 37 8Cfor24 h. There were signiŽ cant differences (P , 0.05) in C.perfringens reductionswith increasing NaCl concentrations. The effects of NaCl were less apparent for E. coli and L. monocytogenes survivors.It is concluded that pathogens surviving ozone treatment of beef are less likely to endangerfood safety than are those surviving sublethal heat treatments.

Therecent U.S. Foodand Drug Administration ap- stituents,including proteins, unsaturated lipids, peptidogly- provalof ozone (O 3)for useas an antimicrobial agent in cans,enzymes, and nucleic acids (17). Ozone,a verypotent thestorage and processing of meats and produce under- oxidizingagent, readily inactivates microorganisms in scoresthe need for moreextensive research to evaluatethis aqueoussolutions; however ,infoods, ozone indiscrimi- technology (6). Over theyears, different studies have dem- natelyreacts with many organic components, resulting in onstratedthe success of ozonetreatment in the disinfection self-depletionand reduced biocidal effectiveness (5). The of shellŽ sh (34), theinhibition of the overripening and de- sameresults can be expected of most other chemical oxi- cayof fruits (8, 19), andthe inhibition of microbialgrowth dants,such as hypochlorite and hydrogen peroxide. It has duringthe storage of meats (29). Thereis no information, beensuggested that ozonated water might be more appli- however,onthe effect of ozone treatment on the cross- cableto the killing of microorganisms on fruits or vegeta- protectionor cross-sensitizationof survivingbacterial cells bles,where interference by organic components might be withregard to subsequent stresses. The food industry must minimal (27). Theuse of increased ozone concentrations beprepared to incorporate this technology safely without (.3 ppm O3 for 5min)to inactivate pathogens attached to therisk of increasedhazards arising from survivingorgan- foodscan become problematic, because ozone may cause isms. theoxidation of fats and discoloration (18). Asalientfea- Bacterialcell death following oxidation brought about tureof ozone reactions is that one atom from ozoneis in- byozone treatment has been attributed to changes in cel- corporatedin theoxidized product, while the unincorporat- lularpermeability followed by lysis (10). However,ozone edreactant (O 2)isreleased as innocuous, nontoxic oxygen hasalso been reported to damage numerous cellular con- (16). Thisis a majorfactor for considerationwith regard tothe use of ozonerather than commonly used chlorinated antimicrobialagents that may result in ef uents containing *Authorfor correspondence. Tel: 215-836-3762; Fax: 215-233-6581; carcinogenicchlorinated by-products. E-mail: [email protected]. †Mentionof brandor Žrm names doesnot constitute an endorsement by Severalstudies have highlighted the limited microbial theU.S. Department ofAgriculture over others of a similar naturenot reductionsresulting from ozonetreatment of meatsurfaces mentioned. (7,11, 14, 29). Typicalaerobic plate count reductions ob- J.FoodProt., Vol. 66, No. 3 SURVIVAL AFTEROZONE TREATMENT 383 tainedby rinsing beef carcasses with ozonated water (2 sterileloops to culture tubes containing 10 ml of BHI (for E. coli 2 O157:H7 and L. monocytogenes cultures)or 10 ml of  uid thio- ppm)at 28to 42 8Cwere 1to2log 10 CFU/cm ,reductions thatwere notany larger than those obtained with conven- glycolatemedium (Difco) (for C.perfringens cultures).Following tionalwashing procedures (26). Combinationtreatments in- 4to6 hofincubation at 37 8C,1-mlaliquots of eachof thethree E. cludingheat, irradiation, acidic pH, and competitive micro- strainsof each organism were transferred to 1 literof BHI (for coli O157:H7 and L.monocytogenes strainmixes) or 1 literof  oraenhance microbial inactivation in foods (32). There-  uidthioglycolate (for C.perfringens strainmixes). The cultures fore,a potentialuse for ozonemay be as an antimicrobial werethen incubated at 37 8Cfor16 h withoutaeration. treatmentin synergy with other treatments, such as heator irradiation.Others have reported that food preservation Beefinoculation and treatments. Londonbroil beef (top practicesinvolving heat, cold, irradiation, high pressure, os- roundcut from the steer ’slongissimusdorsi muscle) was pur- moticstress, and pH modi Ž cationshave the potential to chasedfrom a localsupermarket, cut into sections (7.5 by 10.0 increasethe resistance and virulence of microorganisms (1, by1.0 cm), and weighed, and these sections were separately vac- uumsealed to 2kPain avacuumpackager (Multi-V acInc.,Model

2,4). Escherichia coli O157:H7that had been prestressed Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 A300,Kansas City, Mo.). The meat was sterilized by irradiation ata lowpH for 60min at 37 8Cwas foundto be more witha self-containedgamma-radiation source of 137Cs (42 kGy, resistantto salt (28). Stressingof Listeriamonocytogenes 2308C)atthe ERRC facilityto eliminate contaminating micro- withthe addition of ethanol, hydrogen peroxide, hydro- organisms.Cocktails for a speci Ž cbacterialspecies contained chloricacid, or sodium chloride to the growth medium for threestrains per species as described above and were grown sep- 60 min at 358Cresultedin increasedtolerance to heat (21). aratelyto 10 9 CFU/ml.Meat samples were immersed in the bac- Resultsof variousstudies have shown that microorganisms terialsuspensions for 30 sandthen drip dried for 15 s.Treatments willsurvive ozone treatment of meats; thus, it is possible usedfor the samples are shown in Figure 1. thatsurviving organisms may be cross-protectedfrom other Beefinoculation and treatments:control samples. Con- antimicrobialor food preservation treatments. trolswere inoculated and placed in separate sterile Ž lter-lined Thepresent study is the Ž rst toevaluate the ability of Stomacherbags (Nasco, Fort Atkinson, Wis.), to which 10 ml of differentfoodborne pathogens surviving ozone treatment 0.1%(wt/ vol)peptone water was added. The bags were hand mas- onbeef surfaces to resist subsequent applied stresses such sagedfor 30 s andthen pummeled for 2 minwith a Stomacher asheat, pH, and salt. Heat (55 8Cfor 30min) and ozone (3 lab-blender(Model 400, Spiral Systems, Inc., Cincinnati, Ohio). ppmfor 5min)treatments that did not visually damage Thenumber of bacteria in the inoculated meat Ž ltratewas deter- meatwere usedto evaluate reductions in total viable path- minedby the spread plating of dilutions onto tryptic soy agar ogenson beef as well as thesusceptibility of thesurviving (TSA; Difco)(for E. coli O157:H7 and L.monocytogenes ) or bacteriato subsequent stresses. ShahidiFerguson perfringens (SFP; Difco)agar base medium withoverlay(for C.perfringens ) (Fig. 1). MATERIALS AND METHODS Beefinoculation and treatments:heat-treated samples. Microorganisms. Threedifferent strains were combined to Beefcuts were inoculated, placed individually into Stomacher Ž l- forma cocktailfor each bacterial species, and this species-speci Ž c terbags, vacuum sealed to 2 kPa,submerged in a waterbath at culturecocktail was then separately examined for its survival on 558Cfor30 min, and chilled on ice for 2 min.T enmilliliters of beefsurfaces. The three-strain cocktail of E. coli O157:H7used peptonewater was added, and samples were further evaluated as forthe inoculation of beefsamples consisted of strains380-94 (an describedfor the controls (Fig. 1). isolatefrom salami, obtained from the Food Safety and Inspection Beefinoculation and treatments:ozone-treated samples. Service[FSIS], Washington,D.C.), 932 (obtained from Michael Beefcuts were inoculated, washed in ozonated water (3 ppm) at P.Doyle,University of Georgia, Grif Ž n,Ga.), and 933 (a meat 48Cwithgentle agitation for 5 minwith the ozone system de- isolateobtained from the FSIS). Thethree-strain cocktail of Clos- scribedbelow, placed individually in Ž lter-linedStomacher bags tridiumperfringens consistedof enterotoxin-producingstrains H6 towhich 10 mlof peptonewater was added, and further evaluated (Hobbsserotype 6), obtained from the Centers for Disease Control asdescribed for the controls (Fig. 1). Ozone was generated on- andPrevention (CDC), Atlanta,Ga., and E13 and F5603, which siteand monitored with a systemdeveloped and installed by wereobtained from Bruce A. McClane,University of Pittsburgh AmericanAir Liquide (Countr yside,Ill.). Schoolof Medicine, Pittsburgh, Pa. The L.monocytogenes cock- tailconsisted of two strains from the American Type Culture Col- Stressresponse evaluations. Immediatelyfollowing the heat lection(ATTC) Rockville,Md.), A TCC 7644(strain Gibson), a orozonetreatment, the resistance of thesurviving microorganisms humanisolate, and A TCC 43256(CDCF2380, serotype 4B), an in the meat Ž ltratewas measured by producing serial dilutions isolatefrom a Mexican-stylecheese outbreak in California, and andspread plating these dilutions on TSA ata rangeof pHvalues ScottA (serotype4B), obtained from the culture collection at the (4to 12)and salt concentrations (0 to10%)(for E. coli O157:H7 U.S. Departmentof Agriculture –AgriculturalResearch Ser vice – and L. monocytogenes strains)or onSFP mediumwith SFP over- EasternRegional Research Center (USDA-ARS-ERRC), Wynd- layat a rangeof pH values (4 to 12) and salt concentrations (0 moor, Pa. to 10%) (for C.perfringens )(Fig.1). The pH of the recovery mediumwas adjusted with HCl toattain different acidic values, Inoculumpreparation. Stockcultures were maintained at NaOH wasadjusted to attaindifferent basic values, and NaCl was 2808Cinbrain heart infusion (BHI) broth (Difco Laboratories, adjustedto attain different salt concentrations. The TSA plates Detroit,Mich.) containing 15% (vol/ vol)glycerol (Sigma-Aldrich wereincubated aerobically at 37 8Cfor24 to48h.TheSFP plates ChemicalCo., St. Louis, Mo.). The BHI agarplates were streaked wereincubated at 37 8Cfor18 hunderanaerobic conditions (85% withcultures from a loopfulof partially thawed frozen stock to N2, 10% CO2, and 5% H2).Inaddition, for an evaluation of the producesingle colonies. Single colonies were transferred with thermaltolerance of bacteria surviving heat or ozone treatment, 384 NOVAK AND YUAN J.FoodProt., Vol. 66, No. 3

FIGURE 1. Aschematic owchart of the experimentaldesign used to examine the stressresponses of pathogenic bacteria survivingheat or ozone treatment on beef asdetailed in ‘ ‘Materialsand Methods.’ ’ Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021

1-mlsamples of meat Ž ltratewere vacuum sealed to 8 kPain againsttime at 60 8CwithMicrosoft Excel 2000 software (Micro- Stomacherbags. The samples were then placed into a waterbath softCorporation, Redmond, W ash.).The best- Ž tlinefor survivor set at 608Cforvarious periods (0 to 30 min) and cooled on ice. plotswas determined by regression analysis (25). Aregression Serialdilutions were spread plated on TSA orSFP agar(Fig. 1). equationfor the type y 5 a 1 bx wasderived, where b is the Allexperimental variables were tested in duplicate, and individual slopeof the best straight line that, when inverted and the sign experimentswere repeated three times. changedfrom 2 to 1, gives the D60-valuefor a speci Ž c temper- ature (608Cinthis study). Only survival curves with more than Aqueous ozonetreatment system. Theozone system was Ž vevaluesin thestraight portion, with correlation coef Ž cients (r2) customdesigned to ozonate a 6-gal(22.7-liter) reservoir of ice- of .0.90,and descending more than 5 log 10 cycleswere used. chilledwater (4 8C) to .3 ppm of O3 in15 min. Brie  y, the sys- temconsisted of an Ozonia corona discharge ozone generator, a Statisticalanalyses. Thedata from experiments run in trip- feedsupply of oxygengas, an ozone monitor and probe, a reserve licatefor each organism and variable parameter were analyzed watertank and pump for circulating freshly ozonated water ,an withthe SAS system(V ersion6, SAS Institute,Cary, N.C.) (31) ozonedestruct column for off-gassing, and a stainlesssteel treat- todetermine statistically signi Ž cantdifferences among treatments. mentvessel for the gentle agitation of beef samples. The system TheBonferroni mean separation test was used to determine sig- designenabled the continuous replacement and monitoring of niŽ cantdifferences ( P , 0.05)among means (23). freshlyozonated water in the reaction vessel during the sample treatmentphase. RESULTS

Beefsurface evaluation with SEM. Small (1-cm2) cubes of EfŽcacy of microbial inactivation treatments onbeef controland ozone-treated meat containing surface bacteria were surfaces. Theozone treatments did not macroscopically alter immersedin 2.5% gluteraldehyde –0.1M imidazolebuffer (pH thevisual appearance of the beef, although the mild heat 6.8)solution for 2 hat25 8Candthen stored at 4 8Covernight. treatmentcaused a slightbrowning. Aqueous ozone treat- Forscanning electron microscopy (SEM) processing,a thinslice mentdid, however ,changethe ultrastructural appearance of of Ž xedsample was cut with a stainlesssteel razor blade from an beefthat had been inoculated with foodborne pathogens (Fig. exposedsurface of each sample, and the slices were washed in 2Aand 2B). As shownin Figure 2A, the surfaces of inoc- 0.1M imidazolebuffer ,immersedin 2% osmium tetroxide –0.1 ulatedbeef samples contained collagen Ž bers,muscle Ž bers, Mimidazolebuffer (pH 6.8) for 2 h,washed in distilled water , fat,polysaccharides, unidenti Ž edparticulate matter ,andbac- dehydratedin a gradedseries of ethanol solutions, and critical terialcells. The ultrastructural analyses of meat samples pointdried with liquid CO 2 inamodelDCP-1 criticalpoint dryer (DentonV acuum,Inc., Cherr yHill,N.J.). Dried slices were at- washedwith ozone qualitatively showed fewer collagen Ž - tachedto aluminum stubs and were sputter-coated with a thin bers,with much of thesurface material being washed away, layerof gold with the use of a ScancoatSix Sputter Coater (Ed- exposingintact bacterial cells protected in crevices along wardsHigh V acuumInternational, Wilmington, Mass.). Digital withsome remaining debris (Fig. 2B). imageswere collected with a modelJSM840A scanning electron Thenumbers of bacteria on the control samples were microscope(Joel USA, Peabody,Mass.) operated in thesecondary comparedwith the numbers of organisms surviving the electronimaging mode and coupled with an IMIX digitalimage ozoneand heat treatments. A heattreatment of 55 8C for 30 workstation(Princeton Gamma-T ech,Princeton, N.J.). minwas chosenbecause it would cause minimal damage

Calculationof D60-values. D60-values(time [min] at 60 8C tothe meat tissue while potentially resulting in the induc- requiredto reduce the viable cell population by 1 log 10 unit or tionof protectiveheat shock proteins by thetarget bacteria 90%)were determined by plotting the log 10 numberof sur vivors (12, 13). Themean reductions of pathogens on beef re- J.FoodProt., Vol. 66, No. 3 SURVIVAL AFTEROZONE TREATMENT 385 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021

FIGURE 3. Theeffects of heat treatment (55 8Cfor30 min) or ozonetreatment (3 ppmofO 3 for5 min)on subsequentD 60-values (min) at 608Cforthree-strain cocktails of C.perfringens,L. monocytogenes, and E. coli O157:H7surviving on beefsurfaces. Treatmentswith different letters are signi Ž cantlydifferent ( P , 0.05)as determined by analysis of variance. Error bars depict standarddeviations of means for triplicate experiments.

ppmfor 5min)were 1.28,0.85, and 1.09 log 10 CFU/g for C.perfringens,E. coli O157:H7,and L.monocytogenes, respectively(T able1). Heattolerance of bacteriasurviving heat and ozone treatments. The D60-valuescalculated for surviving L. monocytogenes and E. coli O157:H7cells washed from beefsurfaces following heat or ozone treatment were not signiŽ cantlydifferent from thosefor bacteriawashed from controlsamples ( P . 0.05)(Fig. 3). Prior totreatment, the D60-value for L.monocytogenes washedfrom thebeef sur- FIGURE 2. Representativescanning electron micrographs de- faceswas 2.45min, whereas following heat treatment this pictingthe surfaces of Londonbroil beef inoculated with a three- valueexhibited a small,insigni Ž cantincrease to 2.86 min. straincocktail of E. coli O157:H7(A) priorto aqueous ozone Followingozone treatment, there was nosigni Ž cant in- treatmentand (B) followinga continuousgentle 5-min wash with creasein resistance to heating at 60 8C for L.monocyto- waterozonated to 3 ppm.Muscle tissue (M) and fatty deposits genes, asindicated by a D60-valueof 2.60 min, compared (F)are depicted. Fibrous material is mostly collagen (C) and with2.45 min for thecontrol ( P . 0.05). For E. coli O157: possiblydried polysaccharides. Boxed areas depict intact bacte- H7control cells washed from beefsurfaces, the D -value rialcells. Bar 5 10 mm. 60 obtained(1.70 min) was lowerthan that for L.monocyto- genes (2.45min). After thebeef had been heat treated (55 8C sultingfrom theheat treatment (55 8Cfor 30min) were for 30min), the D60-valuefor thesurviving E. coli O157: 0.14,0.77, and 1.47 log 10 CFU/g for C.perfringens,E. coli H7increased from 1.70to 2.77 min, but this increase was O157:H7,and L.monocytogenes, respectively(T able1). notstatistically signi Ž cant (P . 0.05).Ozone-treated E. Themean reductions resulting from theozone treatment (3 coli O157:H7survivors exhibited even smaller D60-value

TABLE 1. Bacterialsurvival on London broil beef surfaces following mild heat and ozone treatments a

Count (log10 CFU/g)forbacterium

Treatment C.perfringens L.monocytogenes E. coli O157:H7

Control 5.03 6 0.23 6.27 6 0.48 5.92 6 0.19 Heat (558C for 30 min) 4.98 6 0.28 4.80 6 0.61b 5.15 6 0.38 a Ozone(3 ppm O 3 for 5 min) 3.75 6 0.17 5.18 6 0.34 5.06 6 0.24 a Allexperiments were run in triplicate for each organism cocktail and variable parameter .Duplicatesamples were spread plated following dilutionin peptone water .Allstatistical means and standard deviations were derived from a samplenumber of 12 except for ozone- treated E. coli O157:H7,for which the sample number was 11. b Two heat-treated L.monocytogenes samplesexhibited signi Ž cantreductions (2.79 and 4.23 log 10 CFU/g)that were not exhibited by theother 10 samples but were included in the overall mean calculations. 386 NOVAK AND YUAN J.FoodProt., Vol. 66, No. 3 increases(from 1.70to 1.81 min) than the heat-treated cells did. Therewere signi Ž cantdifferences ( P , 0.05)between D60-valuesobtained for C.perfringens cellstreated with heatand with ozone and that obtained for controls(Fig. 3). At 608C, the D60-valuefor cellssurviving on beef at 558Cfor 30minwas 18.21min, compared with 13.87 min for thecontrols. The D60-value for C.perfringens cells resistingozone (3 ppm) treatments decreased from 13.87 minfor controlsto 7.64 min for ozone-treatedsamples. The D60-values for C.perfringens vegetativecells (for controland heat- and ozone-treate dsamples)were higher

thanthose for E. coli and L.monocytogenes. For C. per- Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 fringens, priorheat treatment increased resistance to a temperatureof 60 8C,whereasprior ozone treatment de- creasedsuch resistance. Alkalitolerance of bacteria surviving heat and ozonetreatments. Viablecell counts for controland heat- treated C.perfringens survivorson SFP mediumat pH 12 were comparedwith counts on the medium at pHs of 6,8, and10. Counts for controlbeef were ca.4.8 log 10 CFU/g atpHs of6, 8, and 10 and 4.0 log 10 CFU/gatpH 12, and countsfor heat-treatedbeef were 5.1log 10 CFU/g at pHs of6, 8, and 10 and 3.6 log 10 CFU/gatpH 12 (Fig. 4A). Followingozone treatment, a decreasefrom ca.3.8 log 10 CFU/g(atpHs of6, 8, and 10) to 3.5 log 10 CFU/g (at pH 12)was notsigni Ž cant (P . 0.05).The ozone-treated sur- vivorswere similarlypH resistant at pH levels of 6.0 to 12.0,whereas this was notthe case for thecontrol and heat- treatedsurvivors. However ,thenumbers of ozone-treated survivorsexposed to pHs of 6, 8, 10, and 12 were not signiŽ cantlydifferent from thenumbers of control and heat- treatedsurvivors exposed to pH 12 (Fig. 4A). Regardlessof thepH valueof theplating medium, the viablecell counts for L.monocytogenes were signiŽ cantly different (P , 0.05)for controls(6.8 log 10 CFU/g),heat- treatedsurvivors (2.5 to 3.8log 10 CFU/g),and ozone-treat- edsurvivors (5.6 log 10 CFU/g)(Fig. 4B). Cells surviving heattreatment showed increased susceptibility to all alka- linepH values (Fig. 4B). Viablecell counts for E. coli exhibitedsigni Ž cant de- creases (P , 0.05)following heat treatment in synergywith acidicpH values( 6.0)and basic pH values( 10.0) (Fig. 4C).Ozone treatment did not enhance the inactivation of FIGURE 4. Theeffects of prior treatment stress on recovery(log 10 E. coli cellssurviving on beef and subjected to pHs of 6 CFU/g)frominjury at differentpH levelsfor three-straincocktails to12(Fig.4C). Therefore, alkali stress in combinationwith of (A) C.perfringens, (B) L.monocytogenes, and (C) E. coli O157:H7surviving heat treatment at 55 8Cfor30 min or ozone heattreatment enhanced the inactivation of E. coli and L. treatmentat 3 ppmfor 5 minon beef surfaces. In each graph, monocytogenes ;however,thiseffect was lesspronounced treatmentswith different letters are signi Ž cantlydifferent ( P , for C.perfringens (Fig.4). Ozone treatments did not en- 0.05)as determined by analysis of variance. Error bars depict hancealkali stress effects for anyof the microorganisms standarddeviations of meansfor triplicate experiments. The limit tested. ofbacterial cell detection with spread plating was .10 cells per g,andthe absence of abarfor any given pH valueindicates that NaCltolerance of bacteriasurviving heat and ozone thelevel of survival was lower than this limit. treatments. Therewas asigni Ž cantreduction in the via- bility of C.perfringens withincreasing salt concentrations regardlessof prior treatment (Fig. 5A). Therewere signif- ozoneinactivated 1.64 log 10 CFU/gpriorto plating. Step- icantlyfewer ( P , 0.05)ozone-treated survivors (3.53 wisesequential reductions in viable C.perfringens cells log10 CFU/g)thancontrol (5.17 log 10 CFU/g)orheat-treat- resultedfrom increasedosmotic stress and heat or ozone edsurvivors(4.71 log 10 CFU/g)on0%NaClplates because treatment(Fig. 5A). J.FoodProt., Vol. 66, No. 3 SURVIVAL AFTEROZONE TREATMENT 387

log10 CFU/g)orozone-treated(5.20 log 10 CFU/g)samples, respectively(Fig. 5B). Salt concentrations of 6% were signiŽ cantlymore effective ( P , 0.05)in limiting the growthof bacteria from heat-treatedbeef surfaces than in limitingthe growth of those from ozone-treatedbeef sur- faces(Fig. 5B). No colonies were detectedon 6% NaCl platesfollowing heat treatment (55 8Cfor 30min),whereas 4.42 log10 CFU/gremainedfollowing ozone treatment (3 ppmfor 5min).An NaCl level of 6%inthe medium, along withmild heat, was effectivein limiting the growth of L. monocytogenes from heat-treatedsamples but not in limit- ingthe growth of L.monocytogenes from ozone-treated

samples(Fig. 5B). Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 For E. coli survivorswashed from treatedbeef sur- faces,control counts (6.05 log 10 CFU/g)were signi Ž cantly higher (P , 0.05)than counts for heat-or ozone-treated samples(4.74 and 4.79 log 10 CFU/g,respectively )(Fig. 5C);however ,therewas noevidence of aneffect of NaCl onsurvival at salt levels of up to 4% ( P . 0.05). No viablecounts were recordedfor saltconcentratio nsof 6%. DISCUSSION A Ž nebalance is needed in food-processing technolo- giesto ensure that the consumer is safe from pathogenic microorganismswhile preventing sensory and composition- alchanges in foods. The approval of ozone as an antimi- crobialtreatment for foodshas increased the likelihood that thistechnology will be used to complement current food- processingpractices. Although ozone is a highlyreactive chemicalwith short-term toxicity due to its unstable nature (9, 16), numerousreports have cited its low levels of an- timicrobialef Ž ciencyin the presence of complex organic foodsystems (high ozone demand) (5,7, 11,14, 29). Thus, itis unlikely that ozone application alone will ensure the microbiologicalsafety of foods. In combination with other preservativemeasures, however ,itis possible that greater biocidal efŽ cacymay be attained with less detriment to foodquality or freshness. Undoubtedly, a fractionof the microorganismswill survive initial ozone treatments and maybe cross-protected from orcross-sensitized to subse- quentstresses. The present study is the Ž rst toanalyzespe- ciŽ cstressresponses of foodborne pathogens surviving ozonetreatment of beef. FIGURE 5. Theeffects of prior treatment stress on recovery(log 10 Highconcent rationsof ozone would be expected to CFU/g)from injury at different NaCl concentrations for three- straincocktails of (A) C.perfringens, (B) L.monocytogenes, and alterbeef surfaces through the oxidatio noflipids and (C) E. coli O157:H7surviving heat treatment at 55 8C for 30 min proteins,resultinginnoticeab lemeatdiscolor ation (18). orozone treatment at 3 ppmfor 5 minon beef surfaces. In each Whilein the present study the meat appeared unaltere d graph,treatments with different letters are signi Ž cantlydifferent macroscopically,m icroscopicanalysesshowedthat (P , 0.05)as determined by analysis of variance. Error bars ozonetreatmen tresultedin theremoval of collagen Ž bers depictstandard deviations of means for triplicate experiments. andpolysacc haridesfrom thebeef ’ssurface,leaving a Thelimit of bacterialcell detection with spread plating was .10 few rod-shapedbacterialcellsintact. Total viable cell cellsper g, and the absence of a barfor any given pH value indicatesthat the level of survival was lower than this limit. countssupporte dtheinactiva tionof approxim ately90% ofthe attached cells (6 log 10 CFU/g)(Table 1). The ozoneand heat treatment sapplieddid not effective ly For L.monocytogenes survivorswashed from treated sterilizethe meat. However ,residualwash water usedfor beefsurfaces, the average count obtained with no added thebeef after ozone treatment was platedto verify that saltfor nontreatedcontrols (6.01 log 10 CFU/g)was signif- allnonattac hedmicroorga nismshad been inactiva ted icantlyhigher ( P , 0.05)than that for heat-treated(5.16 (datanot shown). 388 NOVAK AND YUAN J.FoodProt., Vol. 66, No. 3

The C.perfringens vegetativecells surviving a heat vorson platesadjusted to pHsof 6.0to 10.0compared with treatmentof 55 8Cfor 30min on beef showed increased thoseon pH12.0plates were notobserved; these cells may resistanceto subsequent heat at 60 8C(Fig.3). The D60- havebeen selectively inactivated by the ozone treatment. valuefor theheat-treated C.perfringens vegetativecells ThepH differenceswere observedfor thecontrol and heat- was higherthan that for controlcells by nearly4 min.How- treatedsurvivors, since the control treatment and the heat ever,vegetativecells of the three-strain C.perfringens treatmentdid not inactivate the subpopulation of cells that cocktailfrom beefthat underwent ozone treatment exhib- was thensusceptible to the higher pH (Fig.4A). Indeed, C. iteddecreased resistance to subsequent heat at 60 8C. The perfringens F5603,from thethree-strain cocktail mix, has D60-valuefor thesecells was lowerthan that for thecontrol previouslybeen shown to be less hardy and was expected cells by .6min.Thus, heat-treated C.perfringens cells tobe more susceptible than strain E13, also present in the were moreresistant to subsequent heat treatment but more inoculumcocktail (30). susceptibleto ozone treatment. Different mechanisms of re- Reportsof synergisticeffects of ozonein combination pairfollowing heat and ozone treatments may affect cellular withalkali stress against L.monocytogenes and E. coli Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 recoveryand responses to subsequent heat treatment. Heat O157:H7are lacking. There were synergisticeffects on stressadaptation of C.perfringens haspreviously been heat-treated L.monocytogenes onalkaline medium (pH shownto involve the production of protectiveproteins (12, 12)and on heat-treated E. coli O157:H7on mildly acidic 13, 24). medium(pH 6.0),resulting in counts below the detection Thestress responses for L.monocytogenes and E. coli limitof 10 CFU(Fig.4B and 4C). Previous reports have O157:H7were differentfrom thoseof C.perfringens. Both describedacid adaptation and increased thermotolera nce L.monocytogenes and E. coli were moresensitive to atem- for L.monocytogenes and E.coli(15, 21, 22, 28) ; how- peratureof 60 8C than C.perfringens was. Consequently, ever,toour knowledge, the present study is the Ž rst to small,statistically insigni Ž cantdifferences in D60-values examinethe effect of heat adaptation followed by alkali (D60-valuechanges of 10.4 min for L.monocytogenes and stress. 11.0 min for E. coli, respectively)were observedfollowing Anincrease in the NaCl concentration of the plating sublethalheat treatment, and no noticeableozone treatment mediumresulted in anincrease in osmoticstress for micro- effectswere observedrelative to controls. Because stresses bialcells surviving the heat or ozone pretreatment of beef were administeredsequentially, it is possible that inade- aswell as for cellsfrom thecontrol samples. There was a quatetime had elapsed from theinitial stress to the sub- signiŽ cantreduction in C.perfringens countswith increas- sequentstress for fulladaptation. In the present study, the ingsalt concentrations for allof thetreatments ( P , 0.05). temperatureused in the heat treatment to initiate the inac- Thereductions of survivors from controland heat-treated tivation of C.perfringens (558Cfor 30min) was higher sampleswere smallerthan those of survivors from ozone- thanthe temperature (46 8C)previouslyreported to bring treatedsamples (Fig. 5A). Synergisticeffects of NaClcon- aboutthe synthesis of protectiveproteins and resultant heat centrationwith heat treatment or ozone treatment on L. adaptationfor E. coli and L.monocytogenes(3, 20, 33). monocytogenes and E. coli survivorswere notreadily ap- Stress adaptationmight have been less evident than previ- parent(Fig. 5B and 5C). ouslyreported for thesemicroorganisms because the harsh- Theprimary aim of this study was toinvestigate the er treatmentconditions used in the present study resulted inthe inactivation of speci Ž cmetabolicenzymes and pro- stressresponses of microbial pathogens surviving ozone cesses. treatmenton beef surfaces. In summary, the examined Responsesto alkali stress following heat or ozone stressresponses of C.perfringens vegetativecells differed treatmentof beef also differed among the different bacterial from thoseof L.monocytogenes and E. coli O157:H7.Sur- genera.Viable counts of C.perfringens were signiŽ cantly viving C.perfringens cellswere moreresistant to a tem- reduced (P , 0.05)for controland heat-treated samples on peratureof 60 8Cthanwere untreatedcontrol cells following amediumof pH 12.0 relative to those on media with pHs pretreatmentwith 55 8Cfor 30min and were lessresistant of6.0 to 10.0. For theozone-treated survivors, all viable thancontrol cells following pretreatment with 3 ppmof countswere aslowas those obtained on themedium of pH aqueousozone for 5min. C.perfringens cellssurviving 12.0for thecontrol and heat-treated samples (Fig. 4A). heator ozone treatment were tolerantof pH variations, Therefore,for theozone-treated samples, no further popu- whereasheat treatment weakened the resistance of L. mono- lationreductions resulted from anincrease in pH. The cytogenes and E. coli O157:H7to alkalineand/ oracidicpH ozonetreatment might have disrupted proteins at the cell treatments.Ozone treatment did not signi Ž cantlyincrease membraneresponsible for hydrogenion gradient mainte- thealkali stress susceptibility of any of the test microor- nancevia proton motive forces and a lossof cellular A TP ganisms. C.perfringens vegetativecells were moresuscep- production.Another possible explanation is that a subpop- tibleto osmotic stress than either L.monocytogenes or E. ulation of C.perfringens cellsin the three-strain cocktail coli O157:H7cells were. Theseobservations highlight the was moresusceptible to pH 12.0 or ozone, and this situa- importanceof variation in the responses of different bac- tionmay have been responsible for thedecreases in total terialgenera to treatment with a novelantimicrobial agent viablecell growth observed at these pH values (Fig. 4A). likeozone. The results of this study indicate that ozone Acounterexplanationcould involve competitive strain in- treatmentresulted in fewer stressadaptations than sublethal teractions.Increased cell numbers for ozone-treatedsurvi- heattreatment did. Preservation technologies require care- J.FoodProt., Vol. 66, No. 3 SURVIVAL AFTEROZONE TREATMENT 389 fulevaluation to ensure that surviving cells do not exhibit cells andspores of Clostridiumperfringens typeA bysublethal heat increasedpotential to cause foodborne disease. shock. J.FoodProt. 60:998–1000. 13.Heredia, N.L.,R. G.Labbe,and J. S.Garcia-Alvarado. 1998.Al- Thepractical implications of this research are still in terationin sporulation, enterotoxin production, and protein synthesis development.Previously, aqueous ozone was approvedfor by Clostridiumperfringens typeA followingheat shock. J. Food useonly in the potable water industry. Currently, hot water Prot. 61:1143–1147. andacid washes are used to reducemicrobial contamination 14.Horvath, M., L.Bilitzky,and J. Huttner .1985.Fields of utilization ofcarcassesin meat-processingplants and slaughterhouses. ofozone, p. 257 –316. In R.J.H.Clark(ed.), Ozone. Elsevier Sci- ence PublishingCo., New York. Curedmeats include salts with other chemicals as preser- 15.Jorgensen, F .,T.B.Hansen,and S. Knochel.1999. Heat shock-in- vatives.This study has demonstrated that aqueous ozone ducedthermotolerance in Listeriamonocytogenes 13-249is depen- treatmentof meat is capable of injuring bacteria on meat denton growth phase, pH, and lactic acid. FoodMicrobiol. 16:185– surfacesso that these bacteria become more susceptible to 194. subsequentheat or recoveryin environmentsin which less- 16.Kanofsky, J. R.,and P .Sima.1991. Singlet oxygen production from thereactions of ozonewith biological molecules. J.Biol.Chem. 266: than-optimalpH or salt concentrations exist. None of the 9039–9042. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/66/3/382/1674080/0362-028x-66_3_382.pdf by guest on 01 October 2021 testedrecovery conditions resulted in increasedsurvival of 17.Khadre, M. A.,A.E.Yousef,and J.-G. Kim. 2001.Microbiological bacteriafollowing ozone exposure. This study on the sur- aspects ofozone applications in food: a review. J. Food Sci. 66:1– vivalpotential of different bacterial pathogens exposed to 11. aqueousozone provides evidence in favor of the use of 18.Kim, J.-G.,A. E.Yousef,and S. Dave. 1999.Application of ozone forenhancing the microbiological safety andquality of foods:a re- synergisticapproaches to reduce microbial loads in food- view. J.FoodProt. 62:1071–1087. processingoperations that rely heavily on large quantities 19.Kuprianoff, J. 1953.The use ofozone for the cold storage of fruit. ofwaterthat might be recycled,such as poultry-processin g Z.Kaltentech. 10:1–4. orfrankfurter-skinning operations. 20.Linton, R. H.,M. D.Pierson,and J. R.Bishop.1990. Increase in heat resistance of Listeriamonocytogenes ScottA bysublethal heat ACKNOWLEDGMENTS shock. J.FoodProt. 53:924–927. 21.Lou, Y .,andA. E.Yousef.1996. Resistance of Listeriamonocyto- We thankMr .EdwardSteiner and Ms. Sejal Thakkarfor their tech- genes toheat after adaptationto environmental stresses. J. Food nicaland engineering expertise, and American Air Liquide,Countryside, Prot. 59:465–471. Ill.,who installed the ozone system toour speci Ž cationsand made this 22.Mazzotta, A. S.2001.Thermal inactivationof stationary-phaseand projectpossible. Also, we acknowledgethe electron microscopic expertise acid-adapted Escherichiacoli O157:H7, Salmonella, and Listeria ofDr.Peter Cooke,USDA-ARS-ERRC, the statistical analyses ofdata by monocytogenes infruit juices. J. FoodProt. 64:315–320. Dr.JohnPhillips, ERRC, and the technical assistance ofMs. Kenyetta 23.Miller, R. G.1981.Simultaneous statistical inference, 2nded., p. Chaney,ERRC. We thankDrs. PinaFratamico andJohn Luchansky, 67–70.Springer-V erlag,New York. ERRC,for critical readingof the manuscript. 24.Novak, J. S.,M.H.Tunick,and V .K.Juneja.2001. Heat treatment adaptationsin Clostridiumperfringens vegetativecells. J.FoodProt. REFERENCES 64:1527–1534. 25.Ostle, B.,and L. C.Malone(ed.). 1987. Statistics in research: basic 1.Abee, T.,andJ. A.Wouters.1999. Microbial stress responsein min- conceptsand techniques for research workers,4th ed., p. 223 –255. imal processing. Int.J. FoodMicrobiol. 50:65–91. Iowa State Press, Ames. 2.Aldsworth, T .G.,C.Carrington,J. Flegg,G. S.A.B.Stewart,and 26.Reagan, J. O.,G.R.Acuff,D. R.Buege,M. J.Bucyk,J. S.Dickson, C.E.R.Dodd.1998. Bacterial adaptationsto environmental stress; C.L.Kastner,J.L.Marsden,J. B.Morgan,R. NickelsonII, G. C. theimplications for food processing. LeatherheadFood Res. Assoc. Smith,and J. N.Sofos.1996. Trimming and washing of beef car- Food Ind. J. 1:136–144. casses as amethodof improvingthe microbiological quality of meat. 3.Arsene, F.,T.Tomoyasu,and B. Bukau.2000. The heat shockre- J.FoodProt. 59:751–756. sponse of Eschericiacoli. Int. J. FoodMicrobiol. 55:3–9. 27.Restaino, L., E. W.Frampton,J. B.Hemphill,and P .Palnikar.1995. 4.Bower ,C.K.,andM. A.Daeschel. 1999.Resistance responsesof EfŽ cacy ofozonated water againstvarious food-related microorgan- microorganismsin food environments. Int.J. FoodMicrobiol. 50: isms. Appl.Environ. Microbiol. 61:3471–3475. 33–44. 28.Rowe, M. T.,andR. Kirk.1999. An investigation into the phenom- 5.Broadwater, W .T.,R.C.Hoehn,and P .H.King.1973. Sensitivity enonof cross-protectionin Escherichiacoli O157:H7. Food Micro- ofthree selected bacterial species toozone. Appl.Microbiol. 26:391– biol. 16:157–164. 393. 29.Rusch, A., and J. Kraemer. 1989.In  uence ofinstrumentsfor elim- 6.Federal Register.26June 2001. Treatment, storage,and processing inationof microorganisms on surface contaminationof fresh meat offoods; safe use ofozone in gaseous and aqueous phases as anti- andon airborne microorganisms in cold stores withincreased RH. microbialagent. Fed.Regist. 66:33829–33830. Arch.Lebensmittelhyg. 40:61–65. 7.Fournaud, J., andR. Lauret.1972. In  uence ofozone on the surface 30.Sarker ,M.R.,R.P.Shivers,S. G.Sparks,V .K.Juneja,and B. A. microbial  oraof frozen beef andduring thawing. Ind.Aliment. McClane.2000. Comparative experiments to examine theeffects of Agric. 89:585–589. heatingon vegetative cells andspores of Clostridiumperfringens 8.Gane, R.1936.The respiration of bananasin presence ofethylene. isolates carryingplasmid genes versus chromosomal enterotoxin New Phytol. 36:170–178. genes. Appl.Environ. Microbiol. 66:3234–3240. 9.Gordon, G. 1995.The chemistry and reactions of ozone in our en- 31.SAS Institute.1989. SAS/ STATuser ’sguide,version 6, IVed.,vol. vironment. Prog.Nucl. Energy 29:89–96. 2, p. 891–996.SAS Institute,Cary, N.C. 10.Greene, A.K.,B. K.Few,and J. C.Sera Ž ni.1993. A comparison 32.Scott, V .N.1989.Interaction of factors tocontrol microbial spoilage ofozonation and chlorination for the disinfection of stainless steel ofrefrigerated foods. J.FoodProt. 52:431–435. surfaces. J.Dairy Sci. 76:3617–3620. 33.Stephens, P .J.,and M. V.Jones.1993. Reduced ribosomal thermal 11.Greer ,G.G.,and S. D.M.Jones.1989. Effects ofozone on beef denaturationin Listeriamonocytogenes followingosmotic and heat carcass shrinkage,muscle qualityand bacterial spoilage. Can. Inst. shocks. FEMSMicrobiol. Lett. 106:177–182. FoodSci. Technol. J. 22:156–160. 34.Violle, H. 1929.The sterilization of sea water withozone: applica- 12.Heredia, N.L.,G.A.Garcia, R.Luevanos,R. G.Labbe,and J. S. tionof this method for puri Ž cationof contaminated shell Ž sh. Rev. Garcia-Alvarado. 1997.Elevation of theheat resistance ofvegetative Hyg.Med. Prev. 51:42–49.