813

Journalof Food Protection, Vol. 64, No. 6, 2001, Pages 813– 819

Bioluminescence ATPAssay for EstimatingT otalPlate Counts ofSurface Microora ofWhole Cantaloupe and Determining EfŽcacy ofW ashingTreatments †

DIKE O.UKUKU, * VLASTA PILIZOTA, AND GERALD M.SAPERS

U.S.Department of Agriculture,Agricultural Research Service, EasternRegional Research Center, 600East Mermaid Lane, Wyndmoor, Pennsylvania 19038, USA

MS00-226:Received 11July 2000/ Accepted 19December 2000

ABSTRACT

Thesurface micro ora of cantaloupes were estimated using a bioluminescenceA TPassay,and results were compared to platecount data. Cantaloupes were treated as follows: (i) water washed, or (ii) washed in solutions of sodium hypochlorite (1,000mg/ liter)or hydrogen peroxide (5%) for 5 min.Bioluminescence A TPassayresults showed differences in ATP level/ cm2 ofcantaloupes dipped in chlorine or hydrogen peroxide solution; A TPlevelsin these washed samples were lower than incontrols due to antimicrobial action of the treatments on the cantaloupe surface. Linear correlations were found between thebioluminescence A TPassayand aerobic plate counts of unwashed cantaloupe ( r 2 5 0.995)and those washed with water (r 2 5 0.990)determined before storage. Lower correlations between the bioluminescence ATP assayand the aerobic plate countswere obser vedon cantaloupes stored for 120 h at20 8C (r 2 5 0.751)than at 4 8C (r 2 5 0.980)without washing treatment.Lower correlation at 20 8Cmaybe the result of clusters or growth that occurred in chains. A TPlevelsof washed cantaloupescorrelated well with bacterial plate counts ( r 2 5 0.999).A reliableminimum detectable threshold using the 2 2 bioluminescenceA TPassaywas established at 3log 10 fg/cm correspondingto 4log 10 CFU/cm .BioluminescenceA TPassay isnot recommended for washed samples where the microbial load is near or below the threshold. Therefore, the biolumines- cenceATP assaywill be recommendedfor quick estimation of totalmicrobial load on cantaloupesurfaces where the population isexpected to exceed this threshold. The assay can save the industry time by eliminating the required incubation required by theconventional methods.

Themicro ora of all food items are of practical sig- ducemicrobial populations if done properly (7). Washing niŽcance to producers (1, 4, 25), processors (3,12, 15, 31), usuallyconsists of spraying with potable water or may in- andconsumers (16,20, 22). Fruitsand vegetables are fre- volveuse of disinfection by application of chlorine solu- quentlyin contact with soil, insects, animals, and humans tions.Chlorination of washwater has been reported to pre- duringgrowing, harvesting (32), andin theprocessing plant ventmicrobial contamination in produce processing lines (21). Thus,their surfaces are not free from naturalcontam- (36). However,widevarieties of organicmaterials are sub- inants,and by the time they reach the packing house, most jectto oxidationand chlorination reactions in aqueouschlo- fresh produceretain populations of 10 4 to 106 microorgan- rinesolutions. Chlorination reactions are of particular con- isms/g (6, 7). Thelevel of sanitation and the microbiolog- cernin foods because of formation of potentially toxic icalload are of primary importance to the quality, shelf chlororganiccompounds (37). Inthis study, the efŽ cacy of stability,and safety of fresh produce (7, 8). Therefore, hydrogenperoxide as analternative to chlorinein reducing knowledgeof the level of micro ora on fruit or vegetable surfacemicro ora of cantaloupe was investigated.While surfacesshould help processors in implementing hazard thereare numerous reports in the literature on the use of analysisand critical control point (HACCP) plansand good chlorineas a sanitizer,informationon the use of hydrogen manufacturingpractices. However ,estimationof themicro- peroxideon produce is limited. bialload of a foodstuffis problematic, particularly where Recoveryof bacteria or other microorganisms from thesurface of interest is uneven, as in the case of canta- cantaloupesurfaces may not represent a truevalue due to loupe.Accordingly, the problem of obtaining a represen- thesurface roughness of these fruits. The surface roughness tativesample for examinationis often difŽ cult (24). favorsmicrobial attachment and complicates detachment. Washingis one of the very Ž rst processingoperations Faveroet al. (11) reviewedmicrobiological sampling meth- towhich a fruitor vegetable is subjected. W ashingcan re- odsfor surfacesand summarized them into four basic meth- ods:(i) therinse, (ii) the swab rinse, (iii) agar contact, and *Authorfor correspondence. Tel: 215-233-6427; Fax: 215-233-6406; (iv)direct surface agar plating. Methods 2 through4 are E-mail: [email protected]. notsuitable for wholecantaloupe or other fruits with large †Mentionof brand or Ž rm name doesnot constitute an endorsement by theU.S. Department ofAgriculture over others of similar naturenot surfacearea for thefollowing reasons: (i) largeamounts of mentioned. mediawould be required;(ii) sampling would be time-con- 814 UKUKU ET AL. J.FoodProt., Vol. 64, No. 6

TABLE 1. Comparisonof extracting media for the biolumines- sitive.The purposes of thisstudy were toexaminea method cenceATP assayof cantaloupe surfaces a thatis suitable for extractingsurface microbial A TPof

Log10 Log10 cantaloupesand then use the method to monitor changes in Extractingmedia CFU/cm2b ATP (fg/cm2) thesurface A TPlevelof cantaloupesstored at differenttem- peratures.The correlation between log femtogram (fg) ATP Extralight 7.10 6 0.12 5.14 6 0.05 andlog CFU was investigatedto determine the reliability Chloroform-water(30% vol/ vol) 7.00 6 0.03 5.43 6 0.47 ofthe bioluminescent A TPassay.Finally, the assay was Tris-EDTAc 7.18 6 0.10 5.40 6 0.04 usedto compare the efŽ cacy of hydrogen peroxide and a Valuesare means of Žvetrials with duplicate determinations. All chlorinesolutions in reducing the surface micro ora of ATPextractionswith the exception of Extralight involved boil- cantaloupe. ing step. b Determinationswere performed immediately before A TPextrac- MATERIALS AND METHODS tionand determination. c 0.1M Trisin 2 mM EDTA. Preparationof sanitizers. Clorox,a commercialbleach con- taining5.25% sodium hypochlorite (NaOCl; Clorox Company, Oakland,Calif.), was diluted in sterile water to provide the desired suming;and (iii) these methods may not give an accurate concentration(1,000 mg/ liter)of chlorine in the wash solution. indicationof total microbial counts on the surface. The ThepH was adjusted downward to 6.4 6 0.1by adding citric rinsemethod involving the use of water would only be acid.Free chlorine in the solution was determined with a chlorine suitablefor cantaloupeif adhering bacteria were allde- testkit (Hach Co., Ames, Iowa) that has been approved by the tachedby the rinse and then enumerated in the rinse water . U.S. EnvironmentalProtection Agency. Hydrogen peroxide was Sincethis is unlikely, new methods that can take into ac- preparedfrom a 30%stock solution (Fisher ScientiŽ c, Suwannee, countsize, shape, and surface properties of cantaloupe are Ga.),which was diluted in sterilewater to providea concentration of5% in the wash solution. needed. Estimationof microbial numbers in foods by conven- Cantaloupes. Cantaloupespurchased from a localsupermar- tionalmicrobiological techniques takes at least2 to3days, ketand stored at roomtemperature ( ;208C) or 48Cfor0, 24,72, andthere is a needfor fastermethods that can give results or120 h weregrouped (six per treatment) into the following cat- inminutes. The bioluminescent A TPdeterminationwith egories:(i) unwashed; (ii) washed with tap water; and (iii) washed Žrey luciferaseprovides a sensitiveindicator for thepres- insodium hypochlorite (1,000 mg/ liter)or in hydrogen peroxide enceand growth of bacteria in biological material. The (5%)for 5 min.The concentrations of sodium hypochlorite and presenceof bacteria in various media can be estimated us- hydrogenperoxide were chosen based on the results of our pre- ingthe bioluminescent A TPtechnique (2,5, 13, 14, 18, viouslaboratory work (27). Chlorine-or hydrogen peroxide-treat- 34). Its usefor rapidmeasurement of microbial cell num- edsamples were rinsed by dipping in sterile water and agitating bersis based on the premise that all living things contain witha glove-coveredhand for 1 min,then air dried for 1 hat roomtemperature before sampling. ATPandthat the intracellular A TPlevelshave to be con- stantfor cellsto maintain normal physiological activities. Samplepreparation for microbiology. Asterilizedstainless Therefore,the level of ATPdeterminedin a sampleis pro- steelcork borer was used to cut randomly through the cantaloupe portionalto theactual cell number (30, 33). Theassay does surfaceto produce rind plugs of 22 mm diameterwith a surface notrequire incubation and is therefore very rapid and sen- area (pr2) of 3.80 cm2.Atotalof 152 rind plugs per cantaloupe

FIGURE 1. Relationshipbetween the plate counttechnique and the bioluminescence ATP assayof surface micro ora of canta- loupesstored at 20 8Cfor24 h. Determi- nationsof ATP wereperformed using 0.1 MTrisin 2 mM EDTA, 30%chloroform- water(vol/ vol),and Extralight, a commer- cialdetergent from Turner Design. Values aremeans of threedeterminations 6 stan- darddeviation. A correlation( r2 5 0.990) betweenplate count numbers and ATP val- ueswas noted. J.FoodProt., Vol. 64, No. 6 CANTALOUPE SURFACEMICROFLORA ESTIMATION USING ATPBIOLUMINESCENCE 815

TABLE 2. BioluminescenceATP estimationof the surface micro ora of cantaloupe enumerated from different media a Incubationtemperature

258C 358C

2 2 2 2 Media Log10 CFU/cm Log10 ATP (fg/cm ) Log10 CFU/cm Log10 ATP (fg/cm )

PCA 6.61 6 0.23 4.59 6 0.18 7.59 6 0.19 5.69 6 0.15 PDA 3.30 6 0.08 4.29 6 0.15 1.16 6 0.10 2.86 6 0.14 a Cellenumeration and A TPdeterminationwere performed after incubation at 25 8Cfor48 horat35 8Cfor24 h. Aftercell enumeration, plateswere  oodedwith 10 ml sterile dH 2Oandthen subjected to A TPextractionas stated in the ‘ ‘Materialsand Methods.’ ’ Values representmeans of Ž vetrials with duplicate determinations 6 standarddeviation. wereobtained. The  eshadhering to the rind plugs was trimmed oneof the extracting solution listed above were mixed, followed offusing a sterilizedstainless steel knife. byheating in boiling water for 5 minand cooling to room tem- peraturein an ice bath. A TPextractioninvolving the Extralight Microbiologicalanalysis. Cantalouperind plugs (70) weigh- wasnot subjected to theheating but was vortexed for 1 min.The ingapproximately 25 g wereblended (W aringcommercial blend- ATPcontentwas determined by the luciferin-luciferase reaction er;Dynamic Corp., New Hartford,Conn.; the speed set at level withan ATP bioluminescentassay kit (Sigma Chemical Co., St. 5,for1 min)with 75 ml of0.1%peptone water .Decimaldilutions Louis,Mo.). Bioluminescence A TPassayswere performed using ofthe sample were made with 0.1% peptone water ,andaliquots anML3000microtiter plate luminometer (Dynatech Laboratories, (0.1ml) were plated in duplicateon arangeof media. Plate count Chantilly,V a.).The generated light signal was measured after a agar(PCA; Difco,Detroit, Mich.) incubated at 25 8C for 48 h or 16-sdelay time and a 60-sintegration time. Assays of standard at 358Cfor24 hwasused for the enumeration of totalmesophilic amountsof A TPwereused to calculate A TPlevels,and values aerobes.Potato dextrose agar (PDA; Difco),acidiŽ ed with 10% 2 wereexpressed in log 10 fg per cm ofcantaloupesurface. A con- tartaricacid to pH 3.5 and incubated at 25 8Cfor5 days,or at trolwas run using 50 mlTris-EDTA bufferin place of the ATP 358Cfor2 days,was used for yeast and mold enumeration. For standard,and the control readings were subtracted as a measure- theA TPdetermination,the plates were  oodedwith 10 mlsterile mentof background luminescence. Possible inhibition of the lu- dH2O,andthe A TPwasextracted as stated below. ciferasereaction by extracts or residues from the washing agents Analyticalreagent. ATPextractionbuffers (0.1 M Tris- wascorrected by addingknown amounts of ATPstandardinto the EDTA, pH7.75, 30% [vol/ vol]chloroform-water, and Extralight reactionvial followed by additionof theluciferase enzyme. Based [TurnerDesign, Sunnyville, Calif.]) were used for the study. The ontheresults of theA TPextractionstudy, the T ris-EDTA solution Tris-EDTA wasprepared by mixing2 mM EDTA with0.1 M Tris waschosen and used throughout for the ATP determination. andadjusting the pH with 0.2 N HCl. Dataanalysis. Fivereplicate trials for each treatment were BioluminescenceATP assayfor surface micro ora. Sev- conducted.Data from each treatment were subjected to the SAS eralA TPextractants(0.1 M Tris-EDTA, pH7.75; 30% [vol/ vol] software (28) foranalysis of variance and for correlation of the chloroform-water;and Extralight [T urnerdesign]) were studied to twomethods. determinewhich solution would give the highest A TPyield.The RESULTS AND DISCUSSION methodof Anhsen and Nilsson (2), asmodiŽ ed by Ukuku and Shelef (33), wasused to determinelevel of bacterialA TP.Aliquots ATPextractionstudy. Eachof the A TPextraction (1ml) of samples taken for microbiological analysis plus 4 mlof methodsgave similar results with cantaloupe surface mi-

FIGURE 2. Relationshipbetween the plate counttechnique and the bioluminescence ATP assayof surface micro ora of canta- loupesstored at 20 or 4 8Cforup to 120 h.Values are means of three determina- tions 6 standarddeviation. 816 UKUKU ET AL. J.FoodProt., Vol. 64, No. 6

TABLE 3. Estimationof viablecount and ATP ofsurfacemicro ora of cantaloupeafter storage at 20or 4 8Cfor24 hbeforetreatment a Storagetemperature

20 6 18C 4 6 18C

Totalplate count ATP Totalplate count ATP Treatment log CFU/cm2 log (fg/cm2) log CFU/cm2 log (fg/cm2)

Control 7.00 6 0.09 4.98 6 0.02 7.10 6 0.10 4.95 6 0.04 Waterwashed 7.30 6 0.18 5.21 6 0.06 7.18 6 0.10 5.00 6 0.07 Chlorine(1,000 ppm) 4.41 6 0.15 2.83 6 0.02 3.80 6 0.16 2.48 6 0.04 (2.50) (2.68) Hydrogenperoxide (5%) 4.12 6 0.04 2.79 6 0.04 4.49 6 0.04 3.23 6 0.09 (2.70) (2.91) a Valuesrepresent means of Ž vetrials with duplicate determinations 6 standarddeviation. V aluesin parentheses represent predicted ATPvaluesif plate count numbers are set at 10 4 CFU. Calculationwas based on the least-square formula ( Y 5 a 1 bX), where predicted Y is ATP, (a) 5 theintercept, ( b) 5 slope, and (X) 5 setvalue for minimum plate count number for surface micro ora of cantaloupe. croora (T able1). On all samples, the bioluminescence modiŽcations to assay any surface for microbialload or ATPdeterminationgave an estimated A TPvalueof 5.30 sanitarycondition because of itsversatility. Bioluminescent log fg/cm2 correspondingto an average total plate count of ATPtechniquesfor estimatingsurface microbial load on 7.1log CFU/ cm 2 totalmicro ora on cantaloupe surfaces. cantaloupescan save the industry time by eliminating the requiredincubation applicable to theconventional methods. Sensitivitylevel of ATP extraction. Thecantaloupe samplesblended for microbiologicalassay of the surface Surfacemicro ora of cantaloupe. Tocomparethe microora were dilutedfurther to determine the sensitivity bioluminescenceA TPassaywith the total plate count reli- levelof the bioluminescent A TPassay.The results suggest ably,surface micro ora of cantaloupes were enumerated alinearrelationship between cell numbers determined by usingvarious media incubated at two growth temperatures theplate count method and the bioluminescence A TPvalue (Table2). T otalmicro ora of cantaloupesenumerated using 2 above2.0 log 10 fg/cm (Fig.1). Both the Tris-EDTA and PCAincubatedat 25 8Cfor 48h orat 35 8Cfor 24h were 2 chloroform-waterappeared to extract slightly more A TP 6.61and 7.59 log 10 CFU/cm ,respectively.T otalA TPval- thanExtralight. However ,thestandard deviation was higher uescorresponding to these organisms as determinedby the withchloroform-water than with T ris-EDTA orthe Extra- bioluminescenceA TPassaywere 4.59and 5.69 log 10 fg/ light(T able1). These results are in agreement with the re- cm2,respectively.The populations of yeast and mold of portof Prioli and Brown (26), whoindicated better intra- surfacecantaloupe enumerated using PDA at25 or 35 8C 2 cellularA TPextractionwhen boiling methods were uti- were 3.30or 1.16 log 10 CFU/cm ,respectively.The cor- lized.All extractants used in the study were inagreement respondingA TPvaluesat 25 or 35 8Cwere 4.29and 2.86 2 andcorrelated well with total micro ora of surface canta- log10 fg/cm ,respectively.T otalA TPvaluesdetermined 2 loupeabove 3.0 log 10 CFU/cm .Therefore,a sensitivity from thePDA were higherin contrast to those from PCA 2 levelof 3.0log 10 fg/cm was establishedfor thisstudy. For study,conŽ rming a studythat reported higher A TP/cellra- therest of the study, Tris-EDTA was usedas the extractant tioin yeast than in bacteria (2). toestimate bioluminescence A TPlevelof cantaloupe sur- Thebioluminescence A TPassaymeasures cellular facebecause of less variability, health reasons, and cost componentsand metabolic by-products or actual growth; whencompared to the chloroform or the Extralight. hence,it can be used to estimate bacterial numbers (9, 10, 2 Withthe bioluminescence assay, a 3.0log 10 fg/cm sur- 17). Therelative light unit value of any bacterial biolumi- facemicrobial load on cantaloupes can be measuredeffec- nescentis directly related to the A TPextractedand thus to tively.Below this threshold, the A TPdeterminationshowed thenumber of microbialcells from whichit originated (30). alargevariation and is therefore not reliable. This is in BioluminescenceA TPassaysfor estimationof viable agreementwith other reports (19, 23). Karl (17) reported countson six cantaloupes analyzed ranged from 5.29to 2 2 thatvacuum Ž ltrationcan be usedto increasethe sensitivity 5.80 log10 fg/cm ATPandcorrelated ( r 5 0.938) with ofthe A TPassayto less than 10 4 CFU/ml.GrifŽ ths (14) totalplate counts of surfacemicro ora of cantaloupe.How- reportedan approximate 70% agreement between the tra- ever, Greeg (13) reportedthat A TPestimationof cellnum- ditionalplating techniques and the A TPassayrapid method. bersby bioluminescence might not give good agreement However,thebioluminescence A TPassaycannot distin- withcolony count on an agar plate for gram-positivecocci guishA TPvaluesfrom individualclasses of microora rep- (staphylococciand streptococci). For example,the A TP resentedon the surface of cantaloupe. Y et,the assay pro- countmay show 10 5 CFU/ml,while the colony count videsan indication of the total microbial load, which may shows 104 CFU/mlor less. The reason for sucha discrep- helpestablish a thresholdfor goodmanufacturing processes ancy is that Staphyloccus and Streptoccus organismsgrow andHACCP guidelines.This technique can be utilizedwith inbunches and chains, respectively. Therefore, each CFU J.FoodProt., Vol. 64, No. 6 CANTALOUPE SURFACEMICROFLORA ESTIMATION USING ATPBIOLUMINESCENCE 817

FIGURE 3. Totalplate count and biolu- minescenceATP valuesof surface micro- oraof cantaloupes stored at 4 or20 8C andwashed with water, chlorine (1,000 mg/liter),or hydrogen peroxide (50,000 mg/liter).Determination of ATP wasper- formedusing 0.1 M Trisin 2 mM EDTA. Valuesare means of three determinations 6 standarddeviation. A correlation( r2 5 0.763)between plate count numbers and ATP valueswas noted.

oftheseorganisms may in fact represent 10 to 20bacteria. Effect ofsanitizer on bioluminescent technique. Thisphenomenon may not apply in our study since the Washingthe cantaloupes with water did not cause changes blendingprocess utilized in this study may have dispersed inthe total surface population count or the corresponding anybacterial clumps on thecantaloupe surfaces before A TP ATPvalue(T able3 andFig. 3). The chlorine treatment 2 extractionand determination. causeda 2.60log 10 CFU/cm reductionin plate count and 2 Effect ofstoragetemperature. Cellviability and total a 2.2 log10 reductionin ATP(fg/cm )ofcantaloupesstored ATPcontentof surfacemicro ora of cantaloupes stored at at 208Cfor 24h.Thehydrogen peroxide treatment resulted 2 20 or 48Cfor 120h were studied,and the results are shown in a 2.90 log10 CFU/cm reductionof total plate counts and 2 inFigure 2. The population of surfacemicro ora of canta- a 2.2 log10 ATP (fg/cm )reduction.The results at 120 h loupesstored at 20 8Cfor 120h andtheir corresponding (datanot shown) indicate the same trend as seen from the ATPlevelincreased slightly, probably due to an increase cantaloupesstored at 4 or20 8Cfor 24h. The linear rela- inyeast population. W ehavepreviously (T able2) demon- tionshipbetween the plate count method and the biolumi- stratedthat yeast has a higherA TPvaluein contrast to other nescenceA TPassaysof cantaloupes treated with chlorine surfacemicro ora of cantaloupes. T otalmesophilic aerobes orhydrogen peroxide showed a similarcorrelation ( r 2 5 oncantaloupes stored at 4 8Cdeclinedto 6.0 log CFU/ cm 2 0.999)between the two washing treatments in contrast to after24 h andremained the same throughout storage (Fig. theunwashed melons ( r 2 5 0.751)stored at 20 8C for 120 2).The same trend was observedusing the bioluminescence h(Table4). The slope ( a)andintercept ( b)ofall the cal- ATPassay.The decline in totalsurface micro ora and their culations,with the exception of the melons stored for 120 correspondingA TPlevelcould be attributed to the effect hwithoutwashing, showed the same trend, suggesting that ofcold temperature. Some of the aerobic mesophiles may thismethod can be usedto predictA TPlevelon cantaloupe havebeen injured by the cold storage temperature for 24 surfacesbefore and after washing treatment. Linear rela- h,resulting in their death. tionshipsbetween the plate count method and the biolu-

TABLE 4. Comparisonof linear regression and correlation coefŽ cient of bioluminescence ATP assayand plate count of cantaloupe surfaces

RegressioncoefŽ cients a

48C 208C

Treatment a b r a b r

Afterpurchase b — — — 21.585 1.126 0.995 Storedfor 120 h 28.927 2.167 0.980 3.280 0.283 0.751 Chlorine(1,000 mg/ liter) c 20.358 0.747 0.999 20.812 0.826 0.999 Hydrogenperoxide (5%) c 0.274 0.658 0.999 20.344 0.761 0.999 a Linearregression coefŽ cients were calculated based on Y 5 a 1 bx formula,where the slope ( b), the y-intercept( a),andthe correlation coefŽcient ( r)foreach plot are shown in the table. b Performedimmediately after purchase from the store. c Treatmentswere applied after storage for 24 h at4 or20 8C. 818 UKUKU ET AL. J.FoodProt., Vol. 64, No. 6 minescenceA TPassaysof cantaloupes determined imme- 6.Beuchat, L. R.1995.Pathogenic microorganisms associated with diatelyafter purchase ( r 2 5 0.995)and those stored at 4 8C fresh produce.J. FoodProt. 59:204 –216. 7.Brackett, R. E.1992.Shelf stability and safety offresh produceas r 2 for 120 h ( 5 0.980)were similar,suggestingconstant inuenced by sanitation and disinfection. J. FoodProt. 55:808– 814. microbialpopulation at this temperature. The lower linear 8.Brackett, R. E.1994.Microbiological spoilage and pathogens in correlationobserved between the two methods on melons minimallyprocessed refrigerated fruitsand vegetables, p. 269 –312. stored at 208Cfor 120h couldbe attributed to thepossible In R.C.Wiley(ed.), Minimally processed refrigerated fruitsand growthof aerophilic organisms that grow in bunches or vegetables.Chapman Hall, New York. chainsas suggested by Greeg (13). Itis also possible that 9.Chapman, A. G.,andD. E.Atkinson.1977. Adenine nucleotide con- centrationsand turnover rate. Theircorrelation with biological ac- thegrowth of yeast populations as observed in this study tivityin bacteria andyeast, p. 253– 306. In A.H.Rose andD. W. leadsto theproduction of metabolitescapable of producing Tempest (ed.),Advances in microbial physiology. Academic Press, quenchingsubstances (compounds that withdraw or reduce New York. lightsignal) since the melon surfaces were notwashed be- 10.Deming, J. W.,E.W.McGray,E. W.Chappelle,and G. L.Piciolo. fore ATPdetermination(Fig. 3). Selan et al. (29) demon- 1977.Procedures for the quantiŽ cation of bacteria in uid:Ž rey luciferase ATPassay usedwith concentration procedures, p. 465– stratedthe ability of chemicalextractant to contributeto the 490. In G.A.Borun(ed.), 2nd Bi-annual methodology symposium. quenchingof bioluminescence. V alazquezand Feritag (34) SAIT ech.Co., San Diego, Calif. observeda transitionfrom enhancementto quenching in 11.Favero, M. S.,J.J.McDade, J.A.Robertsen,R. K.Hoffman,and chemicalsanitizers like alkaline foam, acid foam, commer- R.W.Edwards.1968. Microbial sampling of surfaces. J.Appl.Bac- teriol.31:336. cialsodium hypochlorite, and D-limolene.They concluded 12.Frank, J. F.,andR. A.KofŽ. 1990.Surface adherentgrowth of Lis- thatthis effect was concentrationdependent and that en- teriamonocytogenes isassociated withincreased resistance tosur- hancementoccurred at lower concentration. factant sanitizer andheat. J. FoodProt. 53:550 –554. 13.Greeg, C.T.1991.Bioluminescence in clinical microbiology, p. 1– CONCLUSION 24. In W.H.Nelson(ed.), Physical methods for microorganism de- tection.CRC Press, Boca Raton,Fla. Theresults of the bioluminescent A TPestimationof 14.GrifŽ ths, M. W.1996.The role of ATPbioluminescence in the food surfacemicrobial load on cantaloupes correlated well with industry:new lighton old problems. Food T echnol.50:63– 72. theplate count data. However ,storagetemperature and 15.Hitchins, A. D.,P .Feng,W .D.Watkins,S. R.Rippey,and L. A. washingtreatments affected the linear correlation between Chandler.1992. Escherichiacoli andthe coliform bacteria, p.27– bioluminescentA TPestimationof surface microbial and to- 31.FDA bacteriologicalanalytical manual, 7th ed. FDA, Arlington, Va. talplate counts. This implies possible quenching activity of 16.International Commission on Microbiological SpeciŽ cations for thebioluminescence process by the chemical treatments. Foods(ICMS). 1980. Factors affecting life anddeath of microor- Storageat 20 8Cfor 120h followedby washing treatments ganisms,p. 102– 106. In J.H.Silliker,R.P.Elliott,A. C.Baird- resultedin lower correlations compared to samples stored Parker,F.L.Bryan,J. H.B.Christian,D. S.Clark,J. C. Olson,Jr ., at 48Cfor 24h beforewashing. Results of this study sug- andT .A.Roberts(ed.), Microbial ecology of foods.Academic Press, New York. gestthat hydrogen peroxide was similarin effectiveness to 17.Karl, D.M.1980.Cellular nucleotide measurements andapplications chlorinein reduction of surface bacterial population. Due inmicrobial ecology. Microbiol. Rev. 44:739 –796. tothe reduced bacterial population, the total A TPvalues 18.Kennedy, J. E.,Jr.,andJ. L.Oblinger.1985. Application of biolu- determinedshowed great variability and were belowthe minescence torapid determination of microbial levels inground minimumestablished threshold we reported. beef.J. FoodProt. 48:334 –340. 19.Ludwicka, A., L. M.Switalski,A. Lundin,G. Pulverer,andT .Wad- ACKNOWLEDGMENTS strom.1985. Bioluminescent assay formeasurement ofbacterial at- tachment topolyethylene. J. Microbiol.Methods 4:169– 177. Theauthors acknowledge the valuable assistance ofJohn G. Phillips 20.Madden, J. M.1992.Microbial pathogens in fresh produce—the reg- instatistical analysisof the data andthat of Ching-Hsing Liao, Vijay ulatoryperspective. J. FoodProt. 55:821– 823. Juneja,and Pina Fratamico fortheir valuable and critical review ofthis 21.Murdock, D. I.,andC. H.Brokaw.1957. Some speciŽ c sources of manuscript. contaminationin processing frozen concentrated orange juice-1. Handlingand preparing fruits for extraction. Proc. Fla. State Hortic. REFERENCES Soc.70:231– 237. 22.Murphy, J. 1999.FDA surveyof imported produce already Ž nding 1.Andrews, W .H.,V.R.Bruce,G. June,F .Satchell,and P .Sherrod. microbiologicalcontamination. Food Chem. News 41:3–21. 1992. Salmonella, p. 51–69. In FDAbacteriologicalanalytical man- 23.Ng, L. K.,D. E.Taylor,andM. E.Stiles.1985. Estimation of Ca- ual,7th ed. FDA, Arlington, Va. mylobacter spp.in broth culture by bioluminescence assay ofA TP. 2.Ansehn, S., and L. Nilsson.1984. Direct membrane-damagingeffect Appl.Environ. Microbiol. 49:730 –731. ofketoconazole and tioconazole on Candidaalbicans demonstrated 24.Patterson, J. T.1971.Microbiological assessment ofsurfaces. J.Food bybioluminescent assay ofA TP.Antimicrob.Agents Chemother. 26: Technol.6:63– 73. 22. 25.Poelma, P .L.,W.H.Andrew,and J. H.Silliker.1984. Salmonella, 3.Austin, J. W.,andG. Bergeron.1995. Development of bioŽ lms in p. 286–326. In M.L.Speck(ed.), Compendium of methods for the dairyprocessing lines. J. DairyRes. 62:509–519. microbiologicalexamination of foods. American PublicHealth As- 4.Baker, J. C.1926.Chlorine in sewage andwaste disposal.Can. Eng. sociation,Washington, D.C. Water Sew.50:127– 132. 26.Prioli, R. P.,andI. N.Brown.1984. A comparisonof methods for 5.Beckers, B.,H.R.M.Lang,and A. Beinke.1984. Determination of theextraction of ATPfrom mycobacteria, p.80–85. In L.J.Kricka, intracellularATP during growth of Escherichiacoli inthe presence P.E.Stanley,G. H.G.Thorpe,and T .P.Whitehead(ed.), Analytical ofampicillin, p. 67– 70. In J.L.Kricka,P .E.Stanley,G. H.G. applicationof bioluminescence and chemiluminescence. Academic Thorpe,and T .P.Whitehead(ed.), Analytical application of biolu- Press, New York. minescence andchemiluminescence. Academic Press, New York. 27.Sapers, G. M.,R.L.Miller,V .Pilizota,and A. M.Mattrazzo.2000. J.FoodProt., Vol. 64, No. 6 CANTALOUPE SURFACEMICROFLORA ESTIMATION USING ATPBIOLUMINESCENCE 819

Anti-microbialtreatments forminimally processed cantaloupe mel- 33.Ukuku, D. O.,and L. A.Shelef.1996. Bioluminescence measure- on.J. FoodSci. Accepted forpublication. ments oftheantilisterial activity of nisin:comparison with ampicillin 28.SAS Institute,Inc. 1999. SAS/ STATuser’ s guide,version 8. SAS andstreptomycin. J. Biolumin.Chemilumin. 11:169– 173. Institute,Inc., Cary, N.C. 34.V anhaecke, E.,and J. Pijck.1988. Bioluminescence assay formea- 29.Selan, L., F.Berlutti,C. Passariello,M. C.Thaller,andG. Renzini. suringthe number of bacteria adheringto the hydrocarbon phase in 1992.Reliability of a bioluminescenceA TPassay fordetection of thebath test. Appl. Environ. Microbiol. 54:1436– 1439. bacteria. J.Clin.Microbiol. 30:1739– 1742. 35.V elazquez, M.,andJ. M.Feritag.1997. Quenching and enhancement 30.Stanley, P .E.1989.A concise beginner’s guideto rapid microbiology effects ofATP extractants, cleansers, andsanitizers onthedetection usingadenosine triphosphate (ATP) and luminescence. In P. E. Stan- oftheA TPbioluminescencesignal. J. FoodProt. 60:799 –803. ley,B. J.McCarthy,and R. Smither(ed.), ATP luminescence. Rapid 36.Wei, C.-I., D. L.Cook,and J. R.Kirk.1985. Use ofchlorinecom- methodsin microbiology. Soc. Appl. Bacteriol. Tech. Ser. 26:1– 10. poundsin the Food Industry. Food Technol. 1:107– 115. 31.Steven, P .,andB. Eldon.1998. Reduction of microorganisms on 37.Yildiz, F .1994.Initial preparation, handling, and distribution of min- citrusfruit surfaces duringpackinghouse processing. J. FoodProt. imallyprocessed refrigerated fruitsand vegetables, Chapter 2. In R. 64:903–906. C.Wiley(ed.), Minimally processed refrigerated fruitsand vegeta- 32.Trueman, J. R.1971.The halogens, p. 137–183. In W.B.Hugo(ed.), bles.Chapman and Hall, New York. Inhibitionand destruction of microbial cells. Academic Press, Lon- don.