sign in sign up
<<
Home , Hepatitis A, Norovirus

85

Journal of Protection, Vol. 76, No. 1, 2013, Pages 85–92 doi:10.4315/0362-028X.JFP-12-311 Copyright G, International Association for Food Protection

Physical Removal and Transfer of Murine and A from Contaminated Produce by Scrubbing and Peeling

QING WANG, MARILYN C. ERICKSON, YNES ORTEGA, AND JENNIFER L. CANNON*

Center for , University of Georgia, Griffin, Georgia 30223, USA

MS 12-311: Received 14 July 2012/Accepted 18 September 2012 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021

ABSTRACT Human and virus are responsible for numerous outbreaks associated with handling fresh produce. In this study, physical removal of hepatitis A virus and , a human norovirus surrogate, from contaminated produce items (honeydew melons, cantaloupes, carrots, and celery) by scrubbing under running water with a nylon brush or scouring pad and by peeling (carrots and celery) with a peeler was investigated. The degree and extent of utensil contamination with during these operations in the presence and absence of food residue also was investigated. Scrubbing or peeling produce initially inoculated with ,5.5 log PFU of each virus resulted in significant levels of virus removal, ranging from 0.93 to 2.85 log PFU. However, utensil cross-contamination occurred, with .2 log PFU of virus transferred from a single produce item. After preparation of a contaminated produce item, utensil cross-contamination resulted in virus detection on seven successively prepared produce items. Produce residue accumulation on utensils variably impacted virus transfer to utensil surfaces. Results indicate that scrubbing and peeling produce can reduce levels of viruses on contaminated produce, but the importance of utensil to prevent cross-contamination is highlighted. Findings also provide important information for modeling virus cross- contamination during food preparation.

Human norovirus and hepatitis A virus (HAV) are contaminated water during irrigation or processing or con- important viral agents of foodborne disease. A large number tact with infected food workers via the fecal-oral route or of outbreaks are caused by norovirus (accounting for ,58% aerosolized vomit (8, 16, 17). According to data obtained of all foodborne illnesses with a known etiology), and high from 1927 to 2006, viruses caused 60% (491 of 816) of out- hospitalization and death rates are associated with HAV breaks in which food workers were implicated (14). Of these (32 and 2%, respectively) (29). Human norovirus outbreaks, norovirus was responsible for 33% (274 of 816), causes a self-limiting illness marked by and and HAV was responsible for 10% (84 of 816) (14). When . High levels of virus particles (up to 1011 virus food workers were implicated, 16% of the norovirus and particles per g of stool sample) can be shed by infected 12% of the HAV infection outbreaks involved produce (14). persons, and the infectious is very low (,18 virus Many studies have been conducted to assess virus particles) (1, 27, 32). Currently, human norovirus cannot be removal by water washing procedures (3–5, 7, 10, 15, 20). propagated in cell culture, so estimating its infectivity Soaking murine norovirus– or HAV-contaminated produce in requires the use of surrogate viruses, such as murine water generally results in ,1-log PFU reduction of viruses norovirus (2, 6). HAV causes and has a detected on produce surfaces. Baert et al. (3) found a 1.01-log long of ,28 days (range, 15 to 50 days), reduction in murine norovirus 1 (MNV-1) on inoculated after which some patients develop characteristic symptoms spinach leaves after one and a 1.26-log reduction after of (17). Virus shedding can begin 10 to 14 days three washes. Croci et al. (10) also obtained approximately a before the onset of symptoms (9) and can continue up to 1-log 50% tissue culture infective dose reduction in HAV 3 months after resolution of symptoms (28, 38). from inoculated fennel and carrots after washing with water. Both human norovirus and HAV are among the most Lukasik et al. (20) found that hand rubbing of produce in a common pathogens associated with produce, accounting for 22uC water bath for 5 s could lead to a 60 to 86% reduction of 40 and 4% of all produce-associated outbreaks, respectively bacteria (Escherichia coli O157:H7 and Mon- (11). Fresh fruit and contamination by foodborne tevideo), bacteriophage (MS2, WX174, and PRD1), and viruses can occur during production, harvesting, processing, 1 inoculated on the surface of strawberries. Feline transport, or preparation close to the point of consumption. calicivirus was reduced by 2 log PFU from strawberries and Contamination of produce can occur through contact with after these items were stirred in water at room temperature for 10 min (15). However, few studies have been * Author for correspondence. Tel: 770-467-6094; Fax: 770-229-3216; performed to determine virus removal by kitchen utensils E-mail: [email protected]. during preparation of fresh produce. 86 WANG ET AL. J. Food Prot., Vol. 76, No. 1

In the present study, kitchen utensils (nylon brushes, uptake. Plaques were then counted, and virus titers were expressed scouring pads, and peelers) were applied to artificially in PFU. Titers of stocks cultures were 5.63 to 7.40 log PFU/ml for contaminated produce items to determine the degree of virus MNV-1 and 5.76 to 7.56 log PFU/ml for HAV. decontamination that could be achieved during typical Inoculation of fresh produce with viruses. Produce items produce preparation. Produce (honeydew melons, canta- with visually consistent sizes and weights were purchased from loupes, carrots, and celery) were chosen based on their local grocery stores. Before inoculation, each produce item was variation in surface properties, softness, and compatibility surface disinfected by exposure to UV light (254 nm, 100 mW/cm2) with the various kitchen utensils chosen for the study. The for 30 min (15 min on each side). Western cantaloupes (Cucumis extent of virus transfer to these utensils and cross-contami- melo L. var. reticulates) and honeydew melons (C. melo) were cut nation in the kitchen also was investigated, as was the impact into halves, and six equally distributed dots were marked on the of produce residues accumulated on utensils on virus transfer. rind of each half, 6 cm away from the apex. A 25-ml inoculum of These investigations are important for determining the effec- MNV-1 or HAV was applied in a straight line between each set of tiveness of virus removal methods used in the kitchen dots for a total of 150 ml. For carrots (Daucus carota) and celery environment and for better assessing the potential risk of virus (Apium graveolens), a mark was made 8 cm from the stem or base Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 cross-contamination via utensils during preparation. and another was made 6 cm away. A 50-ml inoculum was dispensed in 12.5-ml portions by drawing four parallel lines around MATERIALS AND METHODS the circumferences of the produce item, between the marks. After inoculation, all the produce items were kept under the laminar flow Virus cultivation and plaque assay. MNV-1 (a gift from Dr. hood at room temperature until the liquid of each inoculum was Herbert W. Virgin, School of Medicine, Washington University, visibly dry (,1 h). To prevent moisture condensation on the St. Louis, MO) was cultured in RAW 264.7 cells (ATCC TIB-71, surface after drying, the produce items were sealed in a container American Type Culture Collection, Manassas, VA) in Dulbecco’s containing anhydrous calcium sulfate with a cobaltous chloride modified Eagles medium (DMEM) supplemented with 10% low- indicator (Drierite, Xenia, OH) for overnight storage at 4uC until endotoxin fetal bovine serum (FBS; HyClone, Logan, UT), 100 U/ they were scrubbed or peeled the next day. ml penicillin (Invitrogen, Carlsbad, CA), 100 U/ml streptomycin (Invitrogen), 10 mM 4-(2-hydroxy-ethyl)-1-piperazineethanesulfo- Inoculation of utensils with viruses. Utensils were inocu- nic acid (HEPES; Invitrogen) and 1 mM sodium pyruvate lated with 50 ml of virus stock and dried under the laminar flow (Invitrogen). MNV-1 was cultured by infecting 80 to 90% hood at room temperature until the liquid of the inoculum was confluent monolayers of RAW 264.7 cells until complete visibly dry (,1 h). Nylon brushes (10 cm long, 5 cm wide; OXO, cytopathic effect (CPE) was observed (typically 48 h for MNV- Chambersburg, PA) and scouring pads (6 cm long, 3 cm wide; 1). HAV strain HM175 (ATCC VR-1402) was propagated in fetal Scotch-brite, St. Paul, MN) were dotted with 35 to 40 spots equally rhesus monkey kidney cells (FRhK-4, ATCC CRL-1688) in distributed only on the surface that was to come into contact with DMEM containing 8% FBS, 100 U/ml penicillin, 100 U/ml the produce. Peelers (blade size 4 cm long, 0.2 cm in wide; streptomycin, 1% L-glutamine, and 1% nonessential amino acids. Hamilton Beach, Southern Pines, NC) were inoculated with 25 ml HAV was then infected into 80 to 90% confluent monolayers of of virus stock on both sides of the direct contact blade. Utensils FRhK-4 cells until complete CPE was observed (typically 7 days). were stored overnight at 4uC in a container containing anhydrous Viruses were released from cells by three cycles of freezing- calcium sulfate with a cobaltous chloride indicator to maintain a thawing. Cellular debris was precipitated by centrifugation for low relative humidity and prevent moisture condensation on the 15 min at 2,000 | g. The supernatant fluid was filtered through a utensil. 0.2-mm-pore-size membrane filter (Millipore, Billerica, MA) and then stored in 2-ml aliquots at 280uC until used. Protocol for virus removal and transfer during washing Plaque assay of MNV-1 and HAV was preformed similarly to and peeling. Melons were placed on a circular rotating pad and previous reports with some modifications (19, 37). RAW 264.7 stabbed with a pair of forceps at the apex so the melons could be and FRhK-4 cells maintained in complete DMEM were grown rotated and gripped. During the scrubbing procedures with nylon until 80 to 90% confluence was reached. Tenfold serial dilutions of brushes and scouring pads, melons were rotated two complete each virus-containing sample or stock prepared in phosphate- turns in 20 s under running sterile deionized water flowing at 10 ml/ buffered saline (PBS; pH 7.4) were dispensed (in duplicate) over s. During each turn, the inoculated area of each melon was monolayers grown in 60-mm-diameter cell culture plates. The scrubbed with a nylon brush or scouring pad. For carrots, the plates were incubated at 37uC with 5% CO2 for 1 h with gentle uncontaminated tip of each carrot was removed and the top (stem rocking every 15 min to evenly distribute the inoculum. MNV-1 side) of the carrot was held. Carrots were rotated in two turns in overlays consisted of 3 ml of 0.5% agarose (Lonza SeaKem LE, 10 s under running water, flowing at 10 ml/s, while scrubbing the Rockland, ME) with complete Eagle’s medium (Cellgro, Media- inoculated area. After the noninoculated tips were removed, carrots tech, Manassas, VA) supplemented with 5% low-endotoxin FBS, and celery were peeled from a location approximately 1 cm above 100 U/ml penicillin, 100 ml/ml streptomycin, 10 mM HEPES, and the mark next to stem to the end until all the inoculated area was 1 mM sodium pyruvate. HAV overlays consisted of 5 ml of 0.5% removed. After completing the washing (scrubbing) and peeling agarose with DMEM (HyClone, Logan, UT) supplemented with procedures, viruses were recovered and quantified from the 8% FBS (Atlanta Biological, Atlanta, GA), 100 U/ml penicillin, surfaces of produce items and/or utensils. 100 ml/ml streptomycin, 1% L-glutamine, and 1% nonessential amino acids. At the end of the virus incubation period (48 h for Protocol for virus transfer from contaminated utensils to MNV and 7 days for HAV), 3 ml of second overlay medium multiple uncontaminated produce items. Nylon brushes and containing 0.5% agarose and 1.0% 3.3 g/liter neutral red solution peelers were first used on carrots or celery, respectively, to obtain (Sigma) for MNV-1 and 0.5% agarose and 2.5% neutral red utensils contaminated with MNV-1 or HAV. Immediately after solution for HAV was applied for at least 5 h to allow neutral red contamination, the utensil was then applied successively to seven J. Food Prot., Vol. 76, No. 1 VIRUS REMOVAL AND TRANSFER DURING PRODUCE SCRUBBING AND PEELING 87

TABLE 1. MNV-1 and HAV reduction on inoculated produce dissolve PEG before centrifugation at 4uC. Pellets were obtained items after overnight storage at 4uC after centrifugation (9,000 | g, 30 min, 4uC) and resuspended in 1 ml of PBS. Samples were then either processed immediately or Virus reduction (log PFU) per produce itema frozen at 270uC for less than 2 weeks before plaque assay. When Produce item MNV-1 HAV the samples contained visible produce residue, extraction was conducted before the plaque assay at a of 1:1 ratio Carrot 1.05 ¡ 0.15 C 2.75 ¡ 0.70 A (vol/vol). After phase separation by centrifugation (17,000 | g, Celery 1.49 ¡ 0.00 A 0.98 ¡ 0.18 C 10 min, room temperature), the aqueous phase was retained for Cantaloupe 1.26 ¡ 0.65 B 1.42 ¡ 0.59 B plaque assay. Honeydew melon 1.07 ¡ 0.29 C 1.58 ¡ 0.40 B Calculations and statistical analysis. Log reductions in viral a Values are means ¡ standard deviations of three replicates with PFU were obtained by subtracting the amount of virus recovered five samples each. Within a column, means with the same letter (log PFU) after scrubbing or peeling from the amount of virus are not significantly different (P . 0.05). recovered (log PFU) from produce that did not undergo treatment.

Percent recoveries of viruses from nylon brushes, scouring pads, Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 noncontaminated produce items, and the extent of virus transfer to and peelers (with or without residues) were calculated to determine these produce items was determined. Results were reported as the effectiveness of virus recovery protocols. Because of the positive or negative for detection of viruses on each produce item differences in recovery efficiencies between utensil types and (unit). produce types, all virus concentration values obtained after transfer experiments were adjusted for the percent recoveries for each Artificial residue accumulation on utensils. All the utensils utensil type, so that direct comparisons could be made between were dried in a 37uC oven overnight to remove excess moisture transfer studies with different utensil plus produce combinations. and were weighed immediately after drying. Organic carrot juice Percent recovery and degree of virus transfer was calculated (Bolthouse, Bakersfield, CA) was purchased from a local grocery as follows: store, and 10 ml was transferred to a petri dish. Two-thirds of the % recovery bristles from each nylon brush were placed inside the dish to titer of viruses recovered after overnight storage absorb carrot juice for 5 min at room temperature. Brushes were ~ |100 then placed into a 37uC oven for 80 to 90 min. The nylon brushes titer of viruses initially inoculated were then flipped over until completely dried in the oven. For titer of virus transferred to utensil peelers, a more concentrated residue accumulation level was desired; 45 ml of juice was pipetted into 50-ml tubes and ~ titer of viruses recovered from utensil centrifuged (4,000 | g, 15 min, 21uC). The pellet was % recovery resuspended in supernatant to a final volume of 2 ml. Peelers Experiments were conducted with three or five samples in each of were then spiked with 100 ml of concentrated juice (50 ml on each three replicates. Data were analyzed using SAS software (version blade) and dried for 20 min in a 37uC oven. A second layer of 9.2, SAS Institute, Cary, NC). Values obtained from different residues was added to the blades using the same procedure. combinations of factors were analyzed by analysis of variance Utensils were weighed immediately after the residues on the (ANOVA) using a generalized linear model. Significant differenc- utensils were completely dried. The weight of residues on each es between the least square means of treatment groups were utensil was recorded by calculating the difference in weights before indicated at P , 0.05. In the studies of sequential virus transfer, and after the residue accumulation procedure. data were analyzed by ANOVA using a mixed model ANOVA for the binary response. Least-square means were separated using pair- Protocol for virus transfer from contaminated produce to wise t tests, and significant difference was indicated at P , 0.05. utensils with or without residue. Nylon brushes and peelers with and without residue were used to peel inoculated carrots. The RESULTS degree of virus transfer to utensils was then determined. Recovery and survival of viruses on the surface of Virus elution from produce items and utensils. Cantaloupe produce after overnight storage at 4uC. A study was and honeydew melon rinds were cut along the inoculation line into conducted to determine the survival rates of infectious virus six squares (2 by 2 cm). Inoculated rind squares from each melon particles (both MNV-1 and HAV) after overnight storage on and inoculated portions of carrots or celery were placed in a 50-ml produce at 4uC before the storage procedure was adopted in centrifuge tube containing 20 ml of elution buffer (PBS containing subsequent protocols. The log PFU reductions of both 1 M NaCl and 0.05% Tween 20) and vortexed for 30 s to elute the viruses on different produce items after overnight storage viruses. The pieces of produce were removed, and the virus- are listed in Table 1. Initial titers of viruses inoculated on containing elution buffer was retained for further analysis. produce items (cantaloupes, honeydew melons, carrots, and Nylon brushes, scouring pads, and peelers used for processing celery) were 5.91 ¡ 0.47, 5.80 ¡ 0.41, 4.69 ¡ 0.25, and contaminated produce items were placed into stomacher bags 4.74 ¡ 0.07 log PFU for MNV-1 and 6.42 ¡ 0.26, 6.13 ¡ containing 20 to 30 ml of elution buffer and massaged by hand for 0.24, 5.73 ¡ 0.59, and 4.91 ¡ 0.03 log PFU for HAV. The 30 s for virus elution. The utensils were removed, and the virus- percent recoveries from produce after overnight storage containing elution buffer was retained for further analysis. were 0.5 to 13% of the initial inoculum titers. After Concentration of eluted viruses. Polyethylene glycol (PEG; overnight storage, the infectivity of viruses on produce molecular weight 10,000 g/mol) was added to each virus- decreased by at least 1 log PFU regardless of differences in containing eluent to give a final concentration of 8%. Samples inoculation virus titers for melons, carrot, and celery. were shaken for at least 4 h (200 rpm on a shaking platform) to Reductions in virus titers were 1 to 2 log PFU for most 88 WANG ET AL. J. Food Prot., Vol. 76, No. 1

TABLE 2. MNV-1 and HAV reduction after physical removal from inoculated cantaloupes, honeydew melons, and carrots by scrubbing with nylon brushes and scouring pads and from carrots and celery by peeling with peelers MNV-1 HAV

Produce item Utensil Virus reduction (log PFU)a Transfer ratiob Virus reduction (log PFU) Transfer ratio

Cantaloupe Nylon brush 1.58 ¡ 0.46 BC 14/14 1.39 ¡ 0.32 BC 12/12 Scouring pad 1.91 ¡ 1.11 AB 13/14 1.71 ¡ 0.63 B 15/15 c Honeydew melon Nylon brush 2.26 ¡ 0.53 A 15/15 2.78 ¡ 0.32 A 15/15 c Scouring pad 2.23 ¡ 0.74 A 15/15 2.85 ¡ 0.56 A 14/15 Carrot Nylon brush 1.32 ¡ 0.76 CD 12/13 1.15 ¡ 0.87 C 15/15 c Scouring pad 2.30 ¡ 1.03 A 10/15 1.15 ¡ 0.73 C 13/15 Peeler 1.25 ¡ 0.45 CD 15/15 1.18 ¡ 0.92 C 15/15 Celery Peeler 0.93 ¡ 0.95 D 13/15 1.24 ¡ 1.01 BC 15/15 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 a Values are means ¡ standard deviations of three replicates with five samples each. Within a column, means with the same letter are not significantly different (P . 0.05) for produce-utensil combinations. b Transfer ratio is the number of positive samples/number of samples tested. c Significant difference (P , 0.05) when comparing removal of MNV-1 and HAV for this produce-utensil combination. produce types, with the exception of HAV on carrots, which (Table 3). Qualitative analysis (the ratio of positive samples was reduced by 2.75 log PFU. to the total number of samples tested) indicated that viruses frequently were transferred from contaminated produce to Physical removal of viruses from contaminated common kitchen utensils. The extent of virus transfer to produce items by scrubbing or peeling. Before scrubbing utensils also was determined quantitatively with an or peeling but after overnight storage at 4uC, titers recovered adjustment for loss occurring during the recovery procedure from cantaloupe, honeydew melon, carrot, and celery (adjusted by the percent recovery). surfaces were 4.66 ¡ 0.44, 4.77 ¡ 0.14, 3.65 ¡ 0.30, Comparing the transfer of different virus types across and 3.25 ¡ 0.09 log PFU for MNV-1 and 5.00 ¡ 0.39, produce items with the same utensil, significantly more 4.52 ¡ 0.18, 3.05 ¡ 0.72, and 3.93 ¡ 0.21 log PFU for HAV than MNV-1 particles transferred to nylon brushes HAV, respectively. After scrubbing or peeling, log titers and scouring pads from cantaloupes; however, more MNV- recovered from produce were again determined and 1 than HAV particles were transferred from honeydew subtracted from initial recovery values to calculate log melons to nylon brushes and from carrots to scouring pads reductions for MNV-1 and HAV. (P , 0.05). MNV-1 was more easily transferred from Significant reductions of both viruses were observed honeydew melons to nylon brushes than to scouring pads, after scrubbing under running water or removing the peel (P and HAV was more easily transferred from carrots to nylon , 0.05), with reductions of 0.93 to 2.85 log PFU (Table 2). brushes and peelers than to scouring pads (P , 0.05). For a few pieces of produce, virus was not detected (lower However, no significant difference was found for MNV-1 than the limit of detection: 1 log PFU per sample), but com- transfer from cantaloupes and carrots or for HAV transfer plete removal was not obtained for most of the produce items from honeydew melons and cantaloupes to any type of tested. When each virus type was examined separately, the utensil tested (P . 0.05). More MNV-1 particles were removal of MNV-1 was significantly greater than that of transferred to nylon brushes from honeydew melons than HAV for carrots that were scrubbed with scouring pads (P , from cantaloupes and carrots, whereas more HAV particles 0.05). Compared with nylon brushes and peelers, scouring were transferred to peelers from celery than from carrots (P pads achieved a significantly greater removal of MNV-1 from , 0.05). No significant difference was found in the amount carrots (P , 0.05), but MNV-1 and HAV removal was not of virus transferred when cantaloupes or honeydew melons significantly different when comparing the use of nylon were compared after scrubbing with scouring pads regard- brushes and scouring pads on both types of melons (P . less of the virus type (P . 0.05). 0.05). Comparing virus removal from different produce items when using the same utensil, more virus particles (both Extent of virus cross-contamination from contam- MNV-1 and HAV) were removed with a nylon brush (P . inated utensils to uncontaminated produce during 0.05). The scouring pad also removed significantly higher utensil application. After the first use of each utensil on numbers of HAV particles from honeydew melons than from an inoculated produce item, the newly contaminated utensils cantaloupes (P . 0.05). However, similar levels of virus were used successively to process seven uncontaminated reduction were achieved for carrots and celery with a peeler, produce units of the same type (units 1 through 7). The regardless of virus type (P . 0.05). frequency of virus transfer from contaminated nylon brushes and peelers to uncontaminated carrots and celery is Virus transferred to utensils during processing. The presented in Tables 4 and 5, respectively. In both scenarios, ability of viruses to transfer to utensils during scrubbing or contamination occurred in at least one sample in every unit peeling of contaminated produce also was examined examined. For unit 1 samples, viruses could be detected on J. Food Prot., Vol. 76, No. 1 VIRUS REMOVAL AND TRANSFER DURING PRODUCE SCRUBBING AND PEELING 89

TABLE 3. Transfer of MNV-1 and HAV to nylon brushes or scouring pads following scrubbing of contaminated carrots, cantaloupes, and honeydew melons or to peelers following peeling of contaminated carrots and celery MNV-1 HAV

Virus transferred (log Virus transferred (log Utensil Produce item PFU/utensil)a Transfer ratiob PFU/utensil) Transfer ratio

c Nylon brush Cantaloupe 3.59 ¡ 1.28 B 13/14 4.24 ¡ 0.69 A 15/15 c Honeydew melon 4.75 ¡ 0.78 A 14/14 3.26 ¡ 1.71 BCD 12/15 Carrot 2.17 ¡ 1.61 C 10/15 3.09 ¡ 1.20 CD 8/9 c Scouring pad Cantaloupe 3.59 ¡ 0.88 B 15/15 3.97 ¡ 0.73 A 15/15 Honeydew melon 3.52 ¡ 1.21 B 14/15 3.70 ¡ 0.65 ABC 15/15 c Carrot 2.50 ¡ 0.82 C 12/13 1.75 ¡ 0.98 E 15/15 Peeler Carrot 2.46 ¡ 1.17 C 10/12 2.94 ¡ 0.71 D 15/15

Celery 3.83 ¡ 0.37 B 15/15 3.81 ¡ 0.47 AB 15/15 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 a Values are means ¡ standard deviations of three replicates with five samples each and are adjusted by percent recovery. Within columns, means with the same letter are not significantly different (P . 0.05) when comparing virus transfer between produce-utensil combinations. b Transfer ratio is the number of positive samples/number of samples tested. c Significant difference (P , 0.05) when comparing transfer of MNV-1 and HAV within each produce-utensil combination.

87 to 93% of the celery samples that were peeled and 27 to and HAV inoculated on carrots were 5.30 ¡ 0.50 and 5.41 47% of the carrots that were scrubbed. For each subsequent ¡ 0.54 log PFU, respectively, which were higher than those use of the contaminated utensils on uncontaminated carrots previously used in virus removal investigations. No (units 2 through 7), there was a general trend of decreasing difference in MNV-1 and HAV transfer was found for number of virus-positive samples. Virus transfer from nylon nylon brushes with or without residues; however, signifi- brushes to carrots was generally lower than that from cantly more MNV-1 was transferred to peelers with residues contaminated peelers to celery. Within each experimental than to peelers without residues. In contrast, HAV trial, virus transfer was not always continuous; contamina- transferred more easily to peelers without residues than to tion sometimes occurred on a subsequent unit within the peelers with residues. same experimental trial after skipping one or two units. DISCUSSION Degree of virus transfer to utensils with or without In this study, both MNV-1 and HAV were capable of residues from contaminated carrots. The common persisting on the surface of produce items during overnight practice of reusing utensils in the kitchen without washing storage at 4uC. This finding is not surprising; MNV-1 and between uses can result in accumulation of produce residues HAV have been previously reported as relatively resistant to on utensil surfaces. Transfer of viruses to nylon brushes and dry conditions and low temperatures (6, 22, 23). Virus peelers after scrubbing or peeling of contaminated carrots survival rates differed by virus type and produce type, with and the impact of carrot residues on each utensil are shown the greatest reduction in survival observed for HAV on in Table 6. Carrot residues coated on nylon brushes and carrots. Croci et al. (10) reported that inactivation of HAV peelers weighed ,0.8 and ,0.03 g, respectively, before during produce storage (lettuce, fennel, and carrots) at 4uC transfer experiments were conducted. The titers of MNV-1 was greatest for HAV-contaminated carrots and postulated

TABLE 4. Ratio of positive samples to total number tested after contaminated nylon brushes were sequentially used on multiple units of carrots Transfer ratioa

Virus Trial no. Unit 1 Unit 2 Unit 3 Unit 4 Unit 5 Unit 6 Unit 7

MNV-1 1 3/5 2/5 0/5 1/5 0/5 0/5 1/5 2 2/5 3/5 0/5 1/5 1/5 0/5 0/5 3 2/5 1/5 2/5 1/5 1/5 2/5 2/5 Total 7/15 A 6/15 A 2/15 A 3/15 A 2/15 A 2/15 A 3/15 A HAV 1 0/5 0/5 0/5 0/5 0/5 0/5 0/5 2 1/5 0/5 0/5 0/5 0/5 0/5 0/5 3 3/5 1/5 1/5 0/5 0/5 0/5 0/5 Total 4/15 A 1/15 A 1/15 A 0/15 A 0/15 A 0/15 A 0/15 A a Transfer ratio is the number of positive samples/number of samples tested in three replicate trials with five samples per trial. Within a row, ratios with the same letter are not significantly different (P . 0.05). 90 WANG ET AL. J. Food Prot., Vol. 76, No. 1

TABLE 5. Ratio of positive samples to total number tested after contaminated peelers were sequentially used on multiple units of celery Transfer ratioa

Virus Trial no. Unit 1 Unit 2 Unit 3 Unit 4 Unit 5 Unit 6 Unit 7

MNV-1 1 3/5 3/5 4/5 3/5 5/5 4/5 3/5 2 5/5 2/5 3/5 3/5 1/5 1/5 2/5 3 5/5 2/5 3/5 3/5 2/5 2/5 0/5 Total 13/15 A 7/15 B 10/15 AB 9/15 AB 8/15 AB 7/15 B 5/15 B HAV 1 5/5 5/5 5/5 5/5 5/5 5/5 5/5 2 5/5 5/5 1/5 2/5 2/5 0/5 0/5 3 4/5 5/5 3/5 3/5 3/5 4/5 3/5 Total 14/15 A 15/15 A 9/15 A 10/15 A 10/15 A 9/15 A 8/15 A Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 a Transfer ratio is the number of positive samples/number of samples tested for each of three replicate trails with four or five samples per trial. Within a row, ratios with the same letter are not significantly different (P . 0.05). that the natural antimicrobial properties of carrots may be average log PFU reduction per milliliter were included in the responsible. results because only samples that were positive for viruses Rates of virus survival and recovery from the surface of were considered in the statistical analysis. produce can be affected by the unique surfaces properties of The number of virus particles detected on produce was each type of produce or by virus-specific properties. For significantly reduced after scrubbing or peeling of produce instance, carrots have relatively rough skins that are prone to items under running water (0.93 to 2.85 log PFU), but the drying out after overnight storage. Similarly, the virus produce items were not completely decontaminated after inoculum was more easily absorbed by the netted cantaloupe preparation. Similar conclusions have been reached in rinds, perhaps making it more difficult to elute viruses from previous studies of produce washed with water (4, 5, 12). these surfaces than from smooth surfaces such as those of Significantly more virus particles were removed with honeydew melons. To make direct comparisons of virus nylon brushes from honeydew melons than from cantaloupes, transfer between different utensils or produce items with possibly because of the weak attachment of viruses to the variable percent recoveries of virus, the percent recovery for smooth surfaces of the honeydew melons or because the each utensil and produce type was included in the physical features of the cantaloupe (roughness and absorptive calculations to determine the titers of viruses recovered from nature) made it difficult to elute viruses. Other researchers utensils and produce. However, in studies of changes in virus have found similar results; a 5-min water-wash treatment of populations due to treatment (i.e., virus survival during produce surfaces led to reduced removal of E. coli O157:H7 storage and virus removal after processing with a utensil), log from cantaloupe than from honeydew melons (26). PFU reductions of viable virus were determined by Similar to the effectiveness of brushes, peelers only comparison with an untreated control of the same produce partially removed viruses on the surface of carrots and type. Therefore, no adjustment for percent recovery was celery, but cross-contamination of the peelers and transfer of needed to determine log PFU reductions before and after viruses to underlying tissues of the produce likely occurred. treatment (i.e., storage, washing, or peeling). Both the ratio of Thus, utensils had a significant effect on virus removal but positive sample to the number of samples tested and the complete elimination of virus did not occur.

TABLE 6. Transfer of MNV-1 and HAV to nylon brushes and peelers following scrubbing or peeling contaminated carrots and the impact of carrot residues on each utensil surface Utensil Virus Carrot residue Residue wt (g)a Virus transfer/utensil (log PFU)b Transfer ratioc

Nylon brush MNV-1 With 0.84 ¡ 0.10 2.95 ¡ 1.77 A 7/9 Without 3.57 ¡ 0.96 A 9/9 HAV With 0.78 ¡ 0.10 2.74 ¡ 0.96 A 9/9 Without 2.80 ¡ 1.01 A 7/7 Peeler MNV-1 With 0.03 ¡ 0.01 3.63 ¡ 0.85 A 9/9 Without 3.20 ¡ 1.00 B 9/9 HAV With 0.03 ¡ 0.01 1.11 ¡ 1.22 A 7/9 Without 1.96 ¡ 1.68 B 6/9 a Values are means ¡ standard deviations of three replicates with three samples each. b Values are means ¡ standard deviations of three replicates with three samples each with adjustment for percent recovery. Within this column, means with the same letter are not significantly different (P . 0.05) when comparing transfer of MNV-1 and HAV between utensils with or without residue. c Transfer ratio is the number of positive samples/number of samples tested. J. Food Prot., Vol. 76, No. 1 VIRUS REMOVAL AND TRANSFER DURING PRODUCE SCRUBBING AND PEELING 91

The unique surface properties of viruses also could multiple items of uncontaminated meat. The fact that viruses affect the degree of MNV-1 and HAV attachment to a could be detected on the seventh uncontaminated produce produce item and in turn could explain some of the item in the present study was surprising, even when they differences in virus removal observed for a given utensil. were not always found on all intervening items. A reason for Virus attachment to produce is affected by various interac- this variation may be that the virus particles were unevenly tions (18, 31, 34–36). For example, both the electrostatic and distributed on the produce or the utensils or the viruses may hydrophobic interactions between viruses (e.g., feline have clustered at some sites on the utensils. Therefore, the calicivirus, echovirus 11, bacteriophage MS2, and WX174) contaminated sites might not have come into contact with and surfaces (e.g., membrane and lettuce) have been observed some of the sequentially prepared items, so virus transfer (31, 34, 35). More specifically, the ability of HAV to attach to could have skipped one or two units before being deposited stainless steel is dependent on surface energy parameters on a unit that was processed later. (e.g., total surface energy, nonpolar Lifshitz-Van der Waals The hypothesis was that residue accumulation provides reactions, and polar short-range hydrogen bonding compo- surface area that can facilitate virus transfer and lead to nents) and the total free energy of the environment (18). increased transferability. In the present study, more MNV-1 Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 Future studies should be conducted to measure virus-produce particles were transferred from contaminated carrots to peelers interactions and compare the strength of MNV-1 and HAV with residues than to peelers without residues. However, more attachment to produce surfaces. HAV particles were transferred to peelers without residues Contamination of kitchen utensils used to prepare than to peelers with residues. No difference in virus transfer artificially inoculated produce readily occurred, although was observed between nylon brushes regardless of residue not every time the utensil was used. The negative results may presence. The majority of residue applied to nylon brushes be due to the absorptive and/or adsorptive properties of the was absorbed into the bristles, and although the viruses may produce and the occasional sequestration of viruses within have briefly attached to surface residue, the residues and any tissues or on the surface of the produce item. Nylon brushes attached viruses may have been washed away, negating a and scouring pads were used under running water, which residue effect. Differences between MNV-1 and HAV in their could have removed some virus particles. This hypothesis adherence to peelers also may be due to unique viral was supported in preliminary studies when collected wash characteristics. For example, the isoelectric point of viral water was confirmed to be contaminated with infectious proteins affects the attachment of viruses to surfaces viruses (data not shown). The absence of contamination of (36), and the isoelectric points of noroviruses and HAV are the peelers, however, may be due to limited opportunities for different (5.9 to 6.0, and 2.8, respectively) (13, 24).Transfer the peeler blade to contact the contaminated surface because of human norovirus (GI and GII strains) and MNV-1 has most of the blade is cutting under the surface. occurred from contaminated fingertips (30). More virus particles (MNV-1 and HAV) were removed In summary, scrubbing and peeling can remove signif- from honeydew melons than from cantaloupes with nylon icant numbers of virus particles from the surfaces of produce brushes, and more MNV-1 particles were transferred from items. Because the infectious dose of human norovirus and honeydew melons to nylon brushes than to scouring pads. HAV is ,10 to 100 viral particles (27, 32, 33), scrubbing and MNV-1 may more readily attach to nylon brushes than to peeling could prevent illness when the level of contamination scouring pads. HAV did not appear to adhere to nylon is less than 2 to 3 log PFU. However, these methods would not brushes to the same extent as did MNV-1; little difference in be sufficient to decontaminating produce contaminated with HAV transfer to nylon brushes from both types of melons the higher titers of viruses, as in this study. was noted. This study provides a model for investigating the extent Cross-contamination occurred when nylon brushes and of virus removal and transfer during food preparation with peelers were applied to seven uncontaminated carrots kitchen utensils. The findings suggest that scrubbing and sequentially after being used on an inoculated carrot. These peeling of produce are important ways to improve the safety observations are in agreement with those from studies of of these items. Combined with other good hygienic bacterial cross-contamination during food preparation. practices and/or the use of and sanitizers on Salmonella, E. coli, and inoculated on the surfaces utensils or the exterior surfaces of produce, these physical of oranges contaminated cutting board, knives, and juice removal methods may benefit the public health. However, makers, and bacteria were found subsequently in the juice utensils play an important role in virus cross-contamination from the uncontaminated oranges that came into contact in the kitchen environment. Thus, food workers and con- with the utensils used on contaminated oranges (21). Zhao sumers must be equipped with knowledge of those et al. (39) and Montville et al. (25) found Enterobacter procedures that remove viruses from produce and those aerogenes in lettuce after the lettuce contacted cutting that limit contamination and cross-contamination. boards, gloves, or hands used to prepare chicken contam- inated with this bacterium. In the present study, contami- ACKNOWLEDGMENTS nation gradually decreased from unit 1 to unit 7, which can be explained by the decreasing number of virus particles on This research was funded by a U.S. Food and Drug Administration Food Safety Research grant (R01 FD003661). We thank Susan Downer, the utensils due to sequential transfers (dilution effect). Gill Amy Mann, Grishma Kotwal, Katie Roache, and Alison Payton for their and Jones (12) found a similar trend when E. coli was assistance in laboratory cell culture and virus plaque assays. We also thank transferred by handling inoculated meat and then handling Dr. Jerry Davis for his student training and support for the statistical analysis. 92 WANG ET AL. J. Food Prot., Vol. 76, No. 1

REFERENCES 20. Lukasik, J., M. L. Bradley, T. M. Scott, M. Dea, A. Koo, W. Y. Hsu, J. A. Bartz, and S. R. Farrah. 2003. Reduction of poliovirus 1, 1. Atmar, R. L., A. R. Opekun, M. A. Gilger, M. K. Estes, S. E. bacteriophages, Salmonella Montevideo, and Escherichia coli Crawford, F. H. Neill, and D. Y. Graham. 2008. Norwalk virus O157:H7 on strawberries by physical and washes. J. shedding after experimental human infection. 14: Emerg. Infect. Dis. Food Prot. 66:188–193. 1553–1557. 21. Martı´nez-Gonzales, N. E., A. Herna´ndez-Herrera, L. Martı´nez-Cha´vez, 2. Bae, J., and K. J. Schwab. 2008. Evaluation of murine norovirus, M. O. Rodrı´guez-Garcı´a, M. R. Torres-Vitela, L. M. De La Garza, and , poliovirus, and MS2 as surrogates for human A. Castillo. 2003. Spread of bacterial pathogens during preparation of norovirus in a model of viral persistence in surface water and freshly squeezed orange juice. J. Food Prot. 66:1490–1494. groundwater. Appl. Environ. Microbiol. 74:477–484. 22. Mbithi, J. N., V. S. Springthorpe, and S. A. Sattar. 1991. Effect of 3. Baert, L., M. Uyttendaele, M. Vermeersch, E. Van Coillie, and J. relative humidity and air temperature on survival of hepatitis A virus Debevere. 2008. Survival and transfer of murine norovirus 1, a on environmental surfaces. Appl. Environ. Microbiol. 57:1394–1399. surrogate for human noroviruses, during the production process of deep-frozen onions and spinach. J. Food Prot. 71:1590–1597. 23. Melnick, J. L. 1992. Properties and classification of hepatitis A virus. 4. Bidawid, S., J. M. Farber, and S. A. Sattar. 2000. Contamination of 10:S24–S26. by food handlers: experiments on hepatitis A virus transfer to 24. Michen, B., and T. Graule. 2010. Isoelectric points of viruses. food and is interruption. Appl. Environ. Microbiol. 66:2759–2763. J. Appl. Microbiol. 109:388–397. Downloaded from http://meridian.allenpress.com/jfp/article-pdf/76/1/85/1685466/0362-028x_jfp-12-311.pdf by guest on 29 September 2021 5. Bidawid, S., N. Malik, O. Adegbunrin, S. A. Sattar, and J. M. Farber. 25. Montville, R., Y. Chen, and D. W. Schaffner. 2001. Glove barriers to 2004. Norovirus cross-contamination during food handling and bacterial cross-contamination between hands to food. J. Food Prot. interruption of virus transfer by hand antisepsis: experiments with 64:845–849. feline calicivirus as a surrogate. J. Food Prot. 67:103–109. 26. Parnell, T. L., L. J. Harris, and T. V. Suslow. 2005. Reducing 6. Cannon, J. L., E. Papafragkou, G. W. Park, J. Osborne, L.-A. Jaykus, Salmonella on cantaloupes and honeydew melons using wash and J. Vinje. 2006. Surrogates for the study of norovirus stability and practices applicable to postharvest handling, , and inactivation in the environment: a comparison of murine norovirus consumer preparation. Int. J. Food Microbiol. 99:59–70. and feline calicivirus. J. Food Prot. 69:2761–2765. 27. Patel, M. M., A. J. Hall, J. Vinje, and U. D. Parashar. 2009. 7. Casteel, M. J., C. E. Schmidt, and M. D. Sobsey. 2009. Noroviruses: a comprehensive review. J. Clin. Virol. 44:1–8. inactivation of coliphage MS2 on strawberries by industrial-scale 28. Sair, A. I., D. H. D’Souza, and L. A. Jaykus. 2002. Human enteric water washing units. J. Water Health 7:244–250. viruses as causes of foodborne disease. Comp. Rev. Food Sci. Food 8. Cheong, S., C. Lee, S. W. Song, W. C. Choi, C. H. Lee, and S.-J. Saf. 1:73–89. Kim. 2009. Enteric viruses in raw and groundwater used 29. Scallan, E., R. M. Hoekstra, F. J. Angulo, R. V. Tauxe, M.-A. for irrigation in South Korea. Appl. Environ. Microbiol. 75: Widdowson, S. L. Roy, J. L. Jones, and P. M. Griffin. 2011. 7745–7751. acquired in the United States—major pathogens. 9. Cliver, D. O. 1997. Virus via food. Food Technol. 51: Emerg. Infect. Dis. 17:7–15. 71–78. 30. Sharps, C. P., G. Kotwal, and J. L. Cannon. 2012. Human norovirus 10. Croci, L., D. De Medici, C. Scalfaro, A. Fiore, and L. Toti. 2002. The transfer to stainless steel and small fruits during handling. J. Food survival of hepatitis A virus in fresh produce. Int. J. Food Microbiol. Prot. 75:1437–1446. 73:29–34. 31. Shields, P. A., and S. R. Farrah. 1983. Influence of on 11. DeWaal, C. S., and F. Bhuiya. 2007. Outbreaks by the numbers: fruits electrostatic interactions between poliovirus and membrane filters. and vegetables 1990–2005, poster P3-03. Presented at the 94th Annu. Appl. Environ. Microbiol. 45:526–531. Meet. IAFP, Orlando, FL, 8 to 11 July 2007. 32. Teunis, P. F. M., C. L. Moe, P. Liu, S. E. Miller, L. Lindesmith, R. S. 12. Gill, C. O., and T. Jones. 2002. Effects of wearing knitted or rubber Baric, J. L. Pendu, and R. L. Calderon. 2008. Norwalk virus: how gloves on the transfer of Escherichia coli between hands and meat. J. infectious is it? J. Med. Virol. 80:1468–1476. Food Prot. 65:1045–1048. 33. U.S. Food and Drug Administration. 2012. Bad bug book: foodborne 13. Goodridge, L., C. Goodridge, J. Wu, M. Griffiths, and J. Pawliszyn. pathogenic and natural handbook. Hepatitis A 2003. Isoelectric point determination of norovirus virus-like particles virus. Available at: http://www.fda.gov/food/foodsafety/foodborneill by capillary isoelectric focusing with whole column imaging ness/foodborneillnessfoodbornepathogensnaturaltoxins/badbugbook/ detection. Anal. Chem. 76:48–52. ucm071294.htm. Accessed 30 June 2012. 14. Greig, J. D., E. C. D. Todd, C. A. Bartleson, and B. S. Michaels. 34. van Voorthuizen, E. M., N. J. Ashbolt, and A. I. Schafer. 2001. Role 2007. Outbreaks where food workers have been implicated in the of hydrophobic and electrostatic interactions for initial enteric virus spread of foodborne disease. Part 1. Description of the problem, retention by MF membranes. J. Membr. Sci. 194:69–79. methods, and agents involved. J. Food Prot. 70:1752–1761. 35. Vega, E., J. Garland, and S. D. Pillai. 2008. Electrostatic forces 15. Gulati, B. R., P. B. Allwood, C. W. Hedberg, and S. M. Goyal. 2001. Efficacy of commonly used disinfectants for the inactivation of control nonspecific virus attachment to lettuce. J. Food Prot. 71: calicivirus on strawberry, lettuce, and a food-contact surface. J. Food 522–529. Prot. 64:1430–1434. 36. Vega, E., J. Smith, J. Garland, A. Matos, and S. D. Pillai. 2005. 16. Hedberg, C. W., and M. T. Osterholm. 1993. Outbreaks of food- Variability of virus attachment patterns to butterhead lettuce. J. Food borne and waterborne viral . Clin. Microbiol. Rev. 6: Prot. 68:2112–2117. 199–210. 37. Wobus, C. E., S. M. Karst, L. B. Thackray, K.-O. Chang, S. V. 17. Koopmans, M., C. H. von Bonsdorff, J. Vinje, D. de Medici, and S. Sosnovtsev, G. Belliot, A. Krug, J. M. Mackenzie, K. Y. Green, and Monroe. 2002. Foodborne viruses. FEMS Microbiol. Rev. 26:187–205. H. W. I. V. Virgin. 2004. Replication of norovirus in cell culture 18. Kukavica-Ibrulj, I., A. Darveau, J. Jean, and I. Fliss. 2004. Hepatitis reveals a tropism for dendritic cells and macrophages. PLoS Biol. 2: A virus attachment to agri-food surfaces using immunological, e432. virological and thermodynamic assays. J. Appl. Microbiol. 97: 38. Yotsuyanagi, H., L. Koike, K. Yasuda, K. Moriya, Y. Shintani, H. 923–934. Fujie, K. Kurokawa, and S. Iino. 1996. Prolonged fecal excretion of 19. Lemon, S. M., P. C. Murphy, P. A. Shields, L. H. Ping, S. M. hepatitis A virus in adult patients with hepatitis A as determined by Feinstone, T. Cromeans, and R. W. Jansen. 1991. Antigenic and polymerase chain reaction. Hepatology 24:10–13. genetic variation in cytopathic hepatitis A virus variants arising 39. Zhao, P., T. Zhao, M. P. Doyle, J. R. Rubino, and J. Meng. 1998. during persistent infection: evidence for genetic recombination. Development of a model for evaluation of microbial cross- J. Virol. 65:2056–2065. contamination in the kitchen. J. Food Prot. 61:960–963.