This item is only available as the following downloads:
1 EVALUATION OF OVERHEAD SPRAY APPLIED SANITIZERS FOR THE REDUCTION OF Salmonella ON TOMATO SURFACES By ALEXANDRA S. CHANG A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2011
2 2011 Alexandra S. Chang
3 To MF
4 ACKNOWLEDGMENTS I would like to express my gratitude to my major advisor, Dr. Keith Schneider, who first allowed me to join his lab in June 2009. I had just graduated with no plans or food science lab experience, but knew I wanted to study food microbiology I am grateful for the fascinating research project, financial assistance and guidance throughout the past 2 years. I w ould also like to thank my committee members, Dr. Steve Sargent and Dr. W eihua Yang for their time and support of this project. Thanks go out to the Center for Produce Safety who provided funding for the research, to Pacific Tomato Growers and DiMare Fres h for their generous tomato donations and shipments, and to Frank Kelsey of Highland Fresh Technologies for Selectrocide and advice. I would also like to thank my professors, Dr. Ren e Goodrich Schneider, Dr. Jesse Gregory, Dr. Susan Percival and Dr. Anita Wright, and the staff of the Food Science and Human Nutrition Department who taught and assisted me. Every class has been valuable to my experience at the University of Florida. M y lab mates and friends at the Aquatic Food Products Lab Building Mike Hu bbard, Dr. Oleksandr Tokarskyy, Marianne Fatica, Sweeya Gopidi and Xingyu Zhao have been very helpful I particularly want to thank Alina Balaguero for her enormous help including the long hours in the lab and on the road to pick up tomatoes. Finally, I would like to thank my family f or their never ending support.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ..................... 9 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 11 2 LITERATURE REVIEW ................................ ................................ .......................... 14 Fresh Produce Overview ................................ ................................ ........................ 14 Foodborne Disease ................................ ................................ ................................ 15 Sources of Contamination ................................ ................................ ....................... 17 Fresh Produce Outbreaks ................................ ................................ ....................... 19 Tomatoes ................................ ................................ ................................ ................ 20 Salmonella ................................ ................................ ................................ .............. 22 Salmonellosis ................................ ................................ ................................ ... 22 Salmonella Outbreaks ................................ ................................ ...................... 23 Salmonella and Tomatoes ................................ ................................ ................ 24 Prevention Methods ................................ ................................ ................................ 26 Washing and Sanitizing ................................ ................................ .................... 27 Sanitizers ................................ ................................ ................................ .......... 28 Sodium hypochlorite ................................ ................................ .................. 29 Chlorine dioxide ................................ ................................ ......................... 32 Peroxyacetic acid ................................ ................................ ....................... 33 Research Hypothesis and Objectives ................................ ................................ ..... 35 3 MATERIALS AND METHODS ................................ ................................ ................ 38 Bacterial Strains ................................ ................................ ................................ ...... 38 Inoculum Preparation ................................ ................................ .............................. 38 Growth Curves ................................ ................................ ................................ ........ 39 Tomato Inoculation and Plating ................................ ................................ .............. 39 Recovery Study ................................ ................................ ................................ ...... 40 Sanitizer Solution Preparation ................................ ................................ ................ 41 Sodium Hypochlorite (NaOCl) ................................ ................................ .......... 41 Chlorine Dioxide (ClO 2 ) ................................ ................................ .................... 41 Pero xyacetic Acid (PAA) ................................ ................................ .................. 42 Water Control ................................ ................................ ................................ ... 42 Overhead Spray System ................................ ................................ ......................... 43
6 Cross Contamination Study ................................ ................................ .................... 43 Sodium Hypochlorite Efficacy Study ................................ ................................ ....... 45 Sanitizer Efficacy Study ................................ ................................ .......................... 45 Flume vs. Overhead Spray Comparison Study ................................ ....................... 46 Natural Tomato Microflora Study ................................ ................................ ............ 47 4 RESULTS ................................ ................................ ................................ ............... 49 Growth Curves ................................ ................................ ................................ ........ 49 Recovery Study ................................ ................................ ................................ ...... 50 Cross Contamination Study ................................ ................................ .................... 51 Sodium Hypochlorite Efficacy Study ................................ ................................ ....... 52 Sanitizer Efficacy Study ................................ ................................ .......................... 5 3 F lume vs. Overhead Spray Comparison Study ................................ ....................... 54 Natural Tomato Microflora Study ................................ ................................ ............ 55 5 DISCUSSION AND CONCLUSION ................................ ................................ ........ 66 Growth Curves ................................ ................................ ................................ ........ 66 Recovery Study ................................ ................................ ................................ ...... 68 Cross Contamination Study ................................ ................................ .................... 69 Sodium Hypochlorite Efficacy Study ................................ ................................ ....... 72 Sanitizer Efficacy Study ................................ ................................ .......................... 75 Flume vs. Overhe ad Spray Comparison Study ................................ ....................... 77 Natural Tomato Microflora Study ................................ ................................ ............ 79 Conclusions and Future Work ................................ ................................ ................. 81 LIST OF REFERENCES ................................ ................................ ............................... 86 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 94
7 LIST OF TABLES Table page 2 1 Multistate tomato outbreaks caused by Salmonella in the United States ........... 37 4 1 Average log 10 concentration of rifampicin resistant Salmonella strains over 12 h at 37C ................................ ................................ ................................ ........ 57 4 2 Average log 10 recovery and loss of Salmonella from tomato surfaces after 2 h drying at room temperature ................................ ................................ ................ 59 4 3 Cross contamina tion of Salmonella from inoculated tomatoes to uninoculated tomatoes and brush rollers ................................ ................................ ................. 59 4 4 Average log 10 reduction of Salmonella after overhead spray treatment of 25, 50 and 100 mg/L sodium hypochlorite and water control ................................ ... 60 4 5 Average starting pH and measured concentration of sanitizers before and after sanitizer efficacy study experiments ................................ ........................... 61 4 6 Average log 10 reduction of Salmonella after overhead spray treatment of sanitizers and water control ................................ ................................ ................ 62 4 7 Average log 10 reduction of Salmonella after sodium hypochlorite and water control flume treatment, compared to sodium hypochlorite overhead spray treatment ................................ ................................ ................................ ............ 63 4 8 Average starting pH and measured concentration of sanitizers before and after na tural tomato microflora study experiments ................................ .............. 64 4 9 Average log 10 reduction of natural microflora after overhead spray treatment of sanitizers and water control ................................ ................................ ............ 65
8 LIST OF FIGURES Figure page 3 1 Overhead spray system. A) Entire overhead spray system, B) Brush rollers and spray nozzles. ................................ ................................ .............................. 48 4 1 Average log 10 concentration of rifampicin resistant Salmonella strains over 12 h at 37 C ................................ ................................ ................................ ........ 58 4 2 Average log 10 reduction of Salmonella after 25, 50 and 100 mg/L sodium hypochlo rite and water overhead spray treatment ................................ .............. 60 4 3 Average log 10 reduction of Salmonella after sanitizer and water control overhead spray treatment ................................ ................................ ................... 62 4 4 Average log 10 reduction of Salmonella after 100 mg/L sodium hypochlorite and water flume treatment compared to 100 mg/L sodium hypochlorite overhead spray treatment ................................ ................................ ................... 63 4 5 Average log 10 reduction of natural tomato microflora after sanitizer and water control overhead spray treatment ................................ ................................ ....... 65
9 Abstract of Thesis Presented to the Graduate School of the University of Florida in Part ial Fulfillment of the Requirements for the Degree of Master of Science EVALUATION OF OVERHEAD SPRAY APPLIED SANITIZERS FOR THE REDUCTION OF S almonella ON TOMATO SURFACE S By Alexandra S. Chang August 2011 Chair: Keith R. Schneider Major: Food Science a nd Human Nutrition Numerous foodborne disease outbreaks have been linked to fresh tomatoes. Salmonella is a leading cause of foodborne disease and ha s been implicated in all multistate fresh tomato outbreaks in the United States ( US ) Because the outbre aks occurred nationwide the source of contamination likely originated early in production, possibly in a packin ghouse Effective sanitation of tomatoes post harvest is one intervention method. Overhead spray and brush roller systems are used in commerci al packing lines for sanitizing tomatoes and have not been extensively studied. Compared to flumes, the traditional tomato sanitation system, a n overhead spray and brush roller system was hypothesized to achieve higher pathogen reduction on tomatoes becau se of increased physical removal of bacteria in conjunction with antimicrobial efficacy of sanitizers. The aim of this research was to examine the efficacy of sanitizers in the overhead spray and brush roller system for reducing Salmonella on unwaxed, mat ure green tomato es. S odium hypochlorite (NaOCl ; 25, 50 and 100 mg/L ) w as tested against a water control. A sanitizer study tested NaOCl (100 mg/L) chlorine dioxide (ClO 2 ; 5 mg/L ), peroxyacetic acid (PAA ; 80 mg/L ) and water. E fficacy of NaOCl (100 mg/L) was also compared between the overhead spray and brush roller system
10 and a scale model flume Surface inoculated tomatoes were tested for 5, 15, 30 and 60 s per treatment. Results of the sodium hypochlorite study showed that a ll NaOCl concentrations wer e significantly more effective at removing Salmonella than water and achieved at least a 3 log 10 CFU/ml reductio n at different treatment times (p<0.05). NaOCl (100 mg/L) in particular achiev ed an average reduction of 3.98 1.78 log 10 CFU/ml a t 15 s conta ct tim e In the sanitizer study, all sanitizers achieved at least a 3 log 10 CFU/ml reduction of Salmonella at 15 s. NaOCl (100 mg/L) in t he overhead spray and brush roller system significantly reduced more Salmonella than in the flume at 15 to 60 s T hi s research demonstrate d the ability of the overhead spray and brush roller system to reduce Salmonella on tomato surfaces. Compared to flumes, an overhead spray and brush roller system can achieve higher pathogen reduction with less water and sanitizer us e thereby lowering packing costs T his research has the potential to influence current industry practices by supporting the implementation of overhead spray systems to improve safety of tomatoes and keep the tomato industry a nomy
11 CHAPTER 1 INTRODUCTION Foodborne disease outbreaks related to fresh produce have increased significantly in recent decades B etween the 1970s and 1990s, reported produce associated outbreaks increased 5.3% and the median number of illnesses asso ciated with the outbreaks increased by 11% ( Sivapalasingam and others 2004 ) In 2007 there were 37 reported outbreaks attributed to vegetables that were responsible for over 800 illnesses (CDC 2010a). While many types of produce have been linked to out breaks, fresh large round tomatoes have been linked to numerous multistate outbreaks in the US since 1990 ( CDC 2002a, 2005, 2007, 2010c; Cummings and others 2001; Greene and others 2008; Hedberg and others 1999 ). Fresh tomatoes are of particular concern in Florida because the state is the number one prod ucer of fresh tomatoes in the U S, with the industry generating $630 million in 2010 (USDA 2011a). Additionally, m ore people are consuming fresh tomatoes. Annual p er capita consumption of fresh tomatoes g rew to an estimated 9.8 kg in 2011 compared to 5.8 kg in 1980 (USDA 2011b). To meet year round demand of fresh fruits and vegetables, produce distribution has expanded nationwide which may contribute to a higher proportion of consumers acquiring foodborn e disease. The increased handling throughout the supply chain leads to an in crease d risk for contamination with human pathogens. Salmonella is a leading cause of bacterial foodborne illness in humans causing over one million illnesses in the U S per y ear ( Scallan and others 2011a). Salmonella has also been implicated in all US multistate outbreaks of fresh tomatoes Because these outbreaks occur red throughout the US the source of contamination likely
12 originated earl y in production, possibly on the f arm or in a packinghouse S ources of contamination in the field include manu re, contaminated irrigation water, dust, animals, poor worker hygiene and dirty equipment and harvest containers In the packinghouse, u nsanitary conditions increase risk for con tamination such as presence of pests, dirty packing lines and trucks, poor worker hygiene and contaminated water or ice used to cool or wash produce (FDA 1998 ; Beuchat 1996 ) Prevention of contamination is the best method to minimize risk of a toma to foodb orne disease outbreak because once tomatoes are contaminated, there are no effective treatments to totally eliminate pathogens with the exception of cooking or irradiation Prevention methods are a part of Good Agricultural Practices (GAPs) that began in 1998 with the US Food and Drug Administration (FDA) Guide to Minimize Microbial Food Safety Hazards for Fresh Fruits and Vegetables Though GAPs are not yet mandatory for the entire fresh produce industry, the FDA Food Safety Modernization Act of 2011 requires food producers implement a prevention based food safety program of which GAPs are usually a prerequisite (PL 2011). In Florida, Tomato GAPs (T GAPs) and Tomato Best Management Practices (T BMPs) have been required since 2008 (FAC 2007). As part of the rule, tomatoes must be sanitized in flumes containing 150 mg/L free chlorine at pH 6.5 7.5 for a maximum of 2 min Water temperature must also be 5 C greater than pulp temperature of tomatoes. NaOCl is a common source of the active form of chlorin e, hypochlorous acid (HOCl), also referred to as free chlorine. Alternative approved sanitizers at their maximum allowed concentration are aqueous ClO 2 at 5 mg/L and PAA at 80 mg/L (CFR 2010c, d; EPA 2006). These sanitizers are strong oxidizers that disr upt cell permeability Efficacy of NaOCl ClO 2 and PAA against
1 3 pathogens on produce has been studied in flume systems ( Felkey and others 2006; Lopez Galvez and others 2010; Pao and others 2007; Shirron and others 2009; Walter and ot hers 2009; Yuk and other s 2005, 2006). Any other sanitizer or process used in tomato sanitation must be proven in a scientific study to reduce Salmonella or a similar organism by at least 3 log 10 units on tomato surfaces (FDACS 2007). An overhead spray and brush roller system, referred to as an overhead spray system for short, is not always used for tomato sanitation. It is aimed to replace part or all of flume systems traditionally used in tomato packinghouses and has been studied with ClO 2 with tomatoes (Pao and others 2009) Washing in a flume or water bath is effective at removing about 2 to 3 log 10 units of bacteria on produce surfaces (Gil and others 2009). Overhead spray systems are believe d to achieve higher pathogen reduction on tomatoe s because of increased physical removal of bacteria in conjunction with antimicrobial efficacy of sanitizers but they have not yet been extensively studied. O verall aim of this research was to determine the efficacy of sanitizers in a laboratory model overhead spray system for the red uction of inoculated Salmonella on tomato surfaces. NaOCl at three concentrations, ClO 2 PAA and water were evaluated for their ability to achieve at least a 3 log 10 unit reduction of Salmonella on tomato surfaces in order to support the overhead spray sy stem as a n effective method of reducing contamination and risk of foodborne disease outbreaks
14 CHAPTER 2 LITERATURE REVIEW F resh Produce Overview Fresh produce is defined as fruits and vegetables that are sold to consumers unprocessed. Fresh produce h arvested and sold in a whole form i nclu de s berries, melons and tomatoes whereas fresh produce that are cut during harvest but are still unprocessed include celery, lettuce and broccoli (FDA 1998) In contrast, p roduce that has been modified from its origi nal, raw state through methods like thermal processing, canning, freezing or dehydrating are defined as processed. Fresh cut produce is a sub category of fresh produce that includes produce that has been peeled or cut and not further processed. Fresh cut produce is often washed and packaged such as bagged salads and fresh cut melon prepared at retail locations All fresh cut produce are ready to eat though they are sometimes re washed or cooked by consumers (FDA 2001 ) Prepackaged fresh cut produce is a growing trend because of the added value of convenience for consumers. US D ietary G uidelines recommend individuals increase fruit and vegetable intake as part of a healthy eating pattern and to reduce the risk for chronic diseases such as coronary heart di sease, stroke, cancer and diabetes ( USDA and USDHHS 2010 ). Improving eating habits to include fresh produce is extremely important because the prevalence of overweight and obesity has expanded in recent decades to include over two thirds of adults and a g rowing percentage of children in the U S. Recommendations to increase consum ption of fruits and vegetables have ramifications for the produce industry to provide safe produce for consumers. Advance s in distribution and storage have led to the availability of fresh produce throughout the US year round
15 Unfortunately, fresh produce is a potential source of human pathogens and can cause foodborne disease (Beuchat 1996) Foodborne Disease In 1999, Mead and others estimated there to be 76 million illnesses an d 5 ,000 deaths a y ear due to food Recent data and improved method ology refined the numbers to 9.4 million illnesses and 1,351 deaths a year from known pathogens, and 38.4 million illnesses and 1,686 deaths a year from unknown agents (Scallan and others 2 011a, b ). Estimates of foodborne disease var ied greatly partly because diseases are often undiagnosed and are self limiting thus are not reported. Conversely severe cases of foodborne disease can require hospitalization especially if secondary complica tions or sequelae occur. People at higher risk of developing sequelae are individual s with weake r immune systems due to age or illness (Samuel and others 2007) Numerous outbreaks occur worldwide because of contaminated food. A foodborne disease outbre ak is defined by the Centers for Disease Control and Prevention (CDC) as the occurrence of two or more cases of a similar illness resulting from eating a common food ( CDC 20 10 a ). The CDC (2010 a ) reported 1,097 foodborne disease outbreaks in 200 7 that caus ed 21,244 cases and 1 8 deaths. Given the scope of outbreaks, foodborne disease create s a major heal th and economic burden on the U S. Costs to individuals and society include diagnosis, treatment, loss of work, public control efforts, pain and suffering, loss of revenue and perhaps legal actions. O utbreaks may be limited to a household or community while o ther outbreaks are widespread, traced around the country over several weeks. Multistate outbreaks are more likely caused by distribution of a food conta minated at the farm or manufacturing
16 level. Bacterial p athogens associated with outbreaks include Salmonella spp., Clostridium perfringens Escherichia coli O157:H7, Listeria monocytogenes and Camplyobacter spp. Other causes of outbreaks include viruses like n orovirus and h epatitis A, parasitic protozoa like Cryptosporidium and Giardia and toxins produced by microorganisms Norovirus is the most common cause of foodborne disease outbreaks causing 39% of confirmed outbreaks attributed to a single food i n 2007 (CDC 2010a). Scallan and others ( 2011 a ) estimate n orovirus causes more than five million illness es a year in the U S, though th e number of cases is compounded by non foodborne transmission. A variety of foods have caused foodborne disease. Poultry, beef and leafy vegetables were the cause of most outbreaks in 2007 (CDC 2010 a ). The largest outbreak in 2007 was Salmonella in hummus that caused 802 illnesses The second largest outbreak was norovirus at a hotel conference linked to several foods. Ot her outbreaks included poultry contaminated with C. perfringens leafy vegetables contaminated with norovirus, and beef contaminated with E. coli O157:H7 (CDC 2010a). While many food s commonly cause foodborne illness, f ruits and vegetables are especially p roblematic since they are often eaten raw. T here are several types of fruits and vegetables with different methods of growing and harvesting. Furthermore, different physical c haracteristics such as size, shape and surface texture may promote or inhibit m icrobial attachment Consequently, there are multiple ways produce can become contaminated in the field and /or after harvest and it is difficult to impose mandatory
17 produce, general GAPs established by the FDA are voluntarily adapted to specific operations. Proper implementation of GAPs helps reduce the risk of contamination. Sources of Contamination Fresh produce have the potential of becoming contaminated at any step durin g production Produce normally contain nonpathogenic microorganisms as part of their n atural microflora which can become spoilage organisms and plant pathogens. These organisms include the fungi, Botrytis Rhi z opus and Geotrichum and the bacteria Pseud omonas Xanthomonas and Erwinia carotovora ( Narayanasamy 2006). Some h uman p athogens can survive on the outside of produce, especially under ideal humidity and refrig eration temperatures (Harris and others 2002). Microorganisms can enter produce if there are bruises, punctures or abrasions that break open produce epidermis. In tomatoes, m icroorganisms usually cannot penetrate the epidermis and can only enter through wounds or at blossom end or stem scar end (Mahovic and others 2007). Decay organisms tha t enter produce can cause various rots that promote improved colonization of pathogens that can spread and infect an entire container of produce P roduce could also become contaminated pre harvest. Guo and others (2001) inoculated tomato plant stems and flowers with Salmonella and found 11 of 30 tomatoes harvested from the plants contained Salmonella six of which contained Salmonella in the pulp. Once internalized, pathogens are not affected by chemical surface treatments and more easily survive with ac cess to nutrients and moisture (Harris and others 2002). F ruits and vegetables are grown i n field s and can be exposed to many pathogens naturally found in the environment. The goal of GAPs is to minimize exposure to pathogens by preventing produce con tact with various hazards such as manu re,
18 contaminated irrigation water, dust, wild animals poor human handling and dirty equipment and harvest containers (FDA 1998 ; Beuchat 1996 ). One of the most basic sources of contamination is soil Soil can contain native pathogens such as Clostridium botulinum Bacillus cereus and L monocytogenes, or become contaminated from human or animal wastes (Santamaria and Toranzos 2003) Root vegetables such as carrots grow in soil while other produce are grown close enou gh to soil to be splashed when raining. If manure is used as fertilizer and is not properly aged, it can inoculate produce from t he close or d irect contact with soil Manure stored uphill to fields can contaminate plants and water supplies if erosion occ urs. Bacteria can also build up on harvest knives and containers if they are not washed and sanitized often (FDA 1998) Certain materials such as wood cannot be as easily cleaned as plastic and are also more likely to injure produce ( Thompson and others 2002 ). After harvest, produce can be transported to a facility for sorting and packing before being shipped to customers. Unsanitary conditions in a packinghouse such as presence of pests and dirty packing lines and trucks can contaminate produce (FDA 199 8). Other sources of contamination include contaminated ice used to cool produce wash water and dust (Beuchat 1996). Another important contributor to contamination is poor personal hygiene of workers who handle fresh produce. Infected employees can transfer pathogens through feces, blood and by skin anytime produce is handled, from harvest to the preparation of fresh cut produce in retail establishments (Todd and others 2008) F resh cut produce require s more stringent safety standards because in con trast to whole produce they are
19 stripped of their skins that provide natural protection against microorganism s (FDA 2001) Fresh Produce Outbreaks Outbreaks attributed to fresh produce are on the rise compared to foods that traditionally have caused foo db orne disease. Sivapalasingam and others (2004) report ed that from 1973 to 1997 there were 190 produce associated outbreaks associated with 16,05 8 illness, 598 hospitalizations and eight deaths. In the 1990s, 6% of reported outbreaks were caused by produc e compared to only 0.7% in the 1970s. Most outbreaks during this time period were caused by multiple produce it ems such as salads, mixed fruit and mixed vegetables. Single produce items implicated in outbreaks included lettuce, melons, sprouts, juice, be rrie s and tomatoes (Sivapalasingam and others 2004). Bacteria commonly isolated from produce include Aeromonas Shigella Salmonella E. coli O157:H7, Campylobacter Yersinia enterocolitica and L monocytogenes (Beuchat 1996). F resh produce outbreaks ca n be large and widespread Leafy green vegetables such as lettuce and spinach have been linked to several multistate E. coli O157:H7 outbreaks and recalls (Grant and others 2008; Rangel and others 2005). Between 1982 and 2002, while outbreaks associated with E. coli O157:H7 included 38 in produce and 75 in ground beef, t he median number of cases per outbreak was 20 and 8, respectively (Rangel and others 2005) Salmonella has been also associated with numerous large fresh produce outbreaks. In 2008, S S aintpaul caused an outbreak in jalapeno and serrano peppers that involved 43 states and Canada and caused 1, 500 illnesses, 308 hospitalizations and perhaps two deaths (CDC 2008 a ; Barton Behravesh and others 2011 ).
20 The increase in fresh produce outbreaks ma y be due to the availability of fresh produce for purchase year round. Demand for off season produce requires nationwide distribution which increases ri sk for contamination because of increased handling after harvest More p eople consuming produce c ould explain the i ncreased incidence of illness. Conversely i mproved epidemiological data and collection methods may increase reported numbers and may not reflect an actual significant increase in outbreaks. Tomatoes Fresh t omatoes are an important US comm odity but have been linked to numerous outbreaks ( Table 2 1 ) There are several varieties of tomat oes grown in the U S either in the field or in greenhouses. Field grown tomatoes include round, Roma (plum), cherry and grape tomatoes. G reenhouse tomatoes may be grown hydroponically and sold on the vine. T omato es are a large industry for the U S, second in the world after China (USDA 20 10 a ). In the US more than 100,000 acres of tomatoes were harvested in 20 10 with a value of almost $1. 4 billion (USD A 201 1 a ). Florida was ran ked number one in the nation in market share of fresh tomatoes, with 29,000 acres harvested in 20 10 The tomato plays a vital r $ 630 million in 20 10 Comparatively, California the second major producer of fresh t omatoes in the US harvested 3 8 ,000 acres of tomatoes with a value of $3 96 million in the same year. Other states that produce fresh tomatoes are Tennessee, Virginia, Ohio, New Jersey, Georgia, North Carolina and Michigan. In con trast to fresh tomatoes, California is the leader of tomatoes for processing, with 270,000 acres harvested with a value of $878 million in 2010 (USDA 2011 a ).
21 Though the US produced over 27 billion pounds of tomatoes in 2008, it still has import ed at least two billion pounds of tomatoes a year for the past decade (USDA 2010a supplement the Florida tomato supply during winter The majority of tomatoes are imported from Mexico (USDA 201 0b). Prevention of foodborne outbreaks caused by tomato contamination Tomatoes are fruit vegetables typically harvested at a mature green stage and further ripened post harvest Tom atoes are harvested by hand into small buckets. They are then transferred to large bins or a gondola attached to a trailer for transport to a packinghouse ( Thompson and others 2002). P otential contamination during harvest can occur from c ontaminated irri gation water, animals, unhygienic handling practices, unsanitary harvest bins or transport (FDA 1998; Beuchat 1996). At the pac kinghouse, tomatoes may undergo an initial cooling and storage period before processing. Processing begins with the unloading of tomatoes from the large bins. In large operations, the unloading is mechanized for controlled speed of tomato flow into a chlorinated water dump tank. The dump tank functions as a cushion to prevent tomato injury as well as remove debris and dirt. To matoes travel in a flume through pre sizing, which removes obviously defective fruit, and often a second, clean water tank or rinse for further washing The flume then carries tomatoes through hand sorting and electronic, belt or weight sizing. Tomatoes are packed by weight into 25 lb corrugated boxes or may be place packed if already ripe. Tomatoes are cooled by room or forced air cooling and stored at about 13 C. T emperatures below a pproximately 10 C will cause chilling injury. Postharvest ripening o ccurs in degreening
22 rooms by exposure to 100 mg/L ethylene for 3 to 4 d before being shipped to distribution and retail centers (Cantwell and others 2002) In packing facilities, tomatoes can become contaminated through dirty equipment in processing lines overh ead drips, wash water, pests and poor worker hygiene (FDA 1998; Beuchat 1996) Salmonella M ost tomato outbreaks have been linked to Salmonella Salmonella has been associated with disease since i t was first isolated in 1885 It is now divided in to two species, Salmonella enterica and Salmonella bongori and over 2, 500 serovars. Salmonella are facultative anaerob es, gram negative and straight rods of the family Enterobacteriaceae. Most strains are motile through flagella and grow optimally at 37 C, though growth can occur b etween 2 and 54C Salmonella is also able to survive in a large range of pH, from 4.5 to 9.5 Maurer 2007 ). Salmonellosis Most strains of Salmonella are nontyphiodal and can cause salmonellosis, a gastrointestinal infection. The strain of Salmonella that causes typhoid fever is S. Typhi Salmonellosis symptoms typically appear 8 to 72 hours after ingestion and include non bloody diarrhea and abdominal pain. The infection is usually self limiting S equelae of s a lmonellosis are reactive arthritis and ankylosing spondylitis ( Maurer 2007 ). Salmonella causes disease by invading intestinal ce lls via a type three secretion system and causing an influx of Ca 2+ into the intestinal tract Salmonella can pro duce an enterotoxin and cytotoxin that can induce apoptosis and cause diarrhea. Other virulence factors include the ability to acquire iron and the ability to avoid the complement system of innate immunity and antibacterial substances Since ingesting
23 fe wer than 100 Salmonella cells and perhaps as few as 1 to 10 cells could cause illness, it is critical to prevent any Salmonella contami nation in food ( Maurer 2007 ). Salmonella Outbreaks Salmonella is one of the leading cause s of foodborne disea se in humans It is estimated that Salmonella causes over one million foodborne diseases 19,336 hospitalizations and 378 deaths in the US per year (Scallan and others 2011 a ). After n orovirus, Salmonella caused the most foodborne disease outbreaks in 200 7 and caused the majority of confirmed o utbreaks attributed to bacteria S. enterica serovar Enterit idis caused the most Salmonella outbreaks (CDC 2010 a ). Salmonella is ubiquitous and has been linked to every category of food. Products that have caused outbreaks includ e egg salad, fish, cheese, chocolate, milk ice cream spices pork, cooked chicken, peanuts orange juice and a variety of fruits and vegetables ( and Maurer 2007 ). Food associated with Salmonella outbreaks in 2007 were frozen pot pies, processed vegetable snacks, eggs, spinach, tomatoes, tuna, groun d beef, cheese, alfalfa sprouts and fresh basil (CDC 2010a). Poultry is the main reservoir of Salmonella Studies that examined the prevalence of Salmonella in chicken have found up to 100% of samples testing positive (CAST 1994) Salmonella ha s also been associated with shell eggs. In 2010, a multistate outbreak of S. Enteritidis in shell eggs caused 1,939 illnesses (CDC 2010b ). Egg contamination results from the transmission of bact eria from the hen ovary to the interior of the egg prior to shell formation. Though traditionally associated with animals, i n the period 1973 to 1997, Salmonella was the most common bacteria agent that caused foodborne disease
24 outbreaks in produce. Wh ile there were 20 different serotypes involved, most common ones were Typhimurium, Montevideo, Javiana, Anatum, Enteritidis, Infantis, Newpo rt and Stanley (Sivapalasingam and others 2004). Salmonella outbreaks have been associated with imported produce. An outbreak of S. Saphra was reported in 1997 from cantaloupes imported from Mexic o (Mohle Boetani and others 1999). In 1999, mangos from a Brazilian farm caused an outbreak of S. Newport (Sivapa lasingam and others 2003). In each spring from 2000 to 2002 a multistate outbreak of S. Poona occurred from eating fresh cantaloupe imported from Mexico (CDC 2002 b ). More recently, imported cantaloupe was associated with an outbreak of S. Litchfield in 2008 and an outbreak of S. Panama in 2011 (CDC 2008b 2011 ). Because of the link between Salmonella and imported produce, t he prevalence of Salmonella in produce grown in Mexico was studied for 17 different vegetables. In total, 98 samples of 1 700 tested positive, including 12% of parsley samples, 11% of cilantr o samples, 9% of broccoli samples and 9% of cauliflower samples ( Quiroz Santiago and others 2009). Salmonella and Tomatoes Notably all fresh tomato multistate outbreaks in the U S have been caused by Salmonella (Table 2 1) S Baildon caused an outbreak of raw tomatoes in 1999 (Cummings and others 2001). In 2004, S. Braenderup caused an outbreak associated with Roma tomatoes (Gupta and others 2007). Of the four multistate outbreaks in 2006 attributed to Salmonella two were transmitted by tomatoes (CDC 2010). The fact that many outbreaks reach ed multiple states suggests that contamination often occurs early in production. Known sources of contamination occur at the farm or packinghouse level, often from contaminated water. Smaller tomato outbreaks ha ve occurred at
25 restaurants and households and were more likely to be caused by worker or consumer handling and temperature abuse. Outbreaks associated with a particular commodity can disrupt the entire industry The tomato industry suffered major financia l losses in 2008 when tomatoes were linked to S. Saintpaul an outbreak that was later attributed to peppers and caused 1, 5 00 illnesses ( Barton Behravesh and others 2011) The FDA issued a consumer advisory against eating Roma and round tomatoes for over one month. As a result, it was estimated that the Florida tomato industry lost $100 million (Taylor 2010). This outbreak illustrates the importance of safe imported tomatoes because the contamination occurred on a farm in Mexico that grew the peppers as well as Roma tomatoes. Samples of peppers and irrigation water were found positive for S. Saintpaul (CDC 2008 a ). Because of the frequent association of Salmonella with fresh tomatoes, it has been suggested that specific Salmonella serovars may be highly evolved or adapted to survival on and inside tomatoes. Zhuang and others (1995) found S Montevideo grew on tomato surfaces at 20 and 30 C and survive d at 10 to 30 C at 45 to 60% relative humidity for at least 18 d. S. Montevideo also survived at its initial 4.5 log 10 CFU/g inoculum level in chopped tomatoes for at least 9 d at 5 C. The pH of the ripe, chopped tomatoes was reported to be 4.1 0.1 (Zhuang and others 1995). The ability of Salmonella to survive in acidic tomatoes and at refrigeration t emperatures shows how difficult eliminating Salmonella can be and the importance of preventing the initial contamination.
26 Prevention Methods After produce is contaminated, it is nearly impossible to eliminate pathogens other th an thorough cooking or irradi ation Prevention of contamination is the most efficient way to ensure food safety and to prevent foodborne disease Prevention methods are a part of the GAPs program that Guide to Minimize Microbial Food Safety Hazards for F resh Fruits and Vegetables Though GAPs are not yet mandatory for the entire fresh produce industry, the FDA Food Safety Modernization Act of 2011 requires food producers to implement a prevention based food safety program of which GAPs are usually a prer equisite (PL 2011). Voluntary prevention measures by the produce industry include implementing GAPs hazard analysis critical control points (HACCP) microbiological testing and worker training programs (CAST 1994). In several tomato outbreaks, contamina ted tomatoes were distributed from farms to slicing facilities, and finally to restaurants, where consumers ultimately contracted salmonellosis. Using microbiological criteria in supplier contracts could help prevent pathogens from reaching the consumer. Additionally, implementing optional guidelines specific to certain commodities such as the FDA draft guide to minimize contamination of tomatoes may be helpful (FDA 2009) Interventions against field contamination have been developed for produce product ion. Examples include treated effluents from field irrigation, worker education, hygienic hand ling and washing produce with a sanitizer Proper s anitation can reduce risk of contamination. Sanitation standards are part of GAPs and are also found in the Code of Federal Regulations (FDA 1998 ; CFR 20 10 a).
27 Washing and Sanitizing Washing and sanitizing are designed to enhance the safety of raw fruits and vegetables and involves using a sanitizer with potable water. Washing is the removal of soil from produ ce surfaces Washing can remo ve harmful microorganisms, dirt and pesticide residues as well as pre cool produce after harvest (Zagory 1999). S anitizing is the use of a sanitizer to reduce the number of microorganisms on produce to a safe level. Reductio n of microorganisms is usually measured on a logarithmic scale. A sanitizer or a biocide, is a chemical agent that inactivates most microorganisms when it comes into contact with food (McDonnell and others 1999). Factors that should be considered in any washing operation include water quality, con tact time, application method targeted microorganisms, microbial load and type of produce (Gil and others 2009). One focus of GAPs is wash water quality. General water GAPs include performing water sampling an d microbial testing, developing standard operating procedures for changing water, cleaning and sanitizing water contact surfaces, installing backflow devices and performing regular inspections. Since wash water is reused throughout the day, sanitizer e ffi cacy must be maintained by monitoring variables like concentration and water temperature (FDA 1998). While some producers may have waited for the FDA to impose mandatory regulation s, the Florida tomato industry, faced with a history of l arge o utbreaks, d ecided to self regulate In 1998, T GAPs and T BMPs became rule with the goal of enhancing the safety of fresh tomatoes and preventing or minimizing contamination of tomatoes (FAC 2007) T GAPs and T BMPs created leadership and initiative within the indu stry and formalized safety practices.
28 Much of T GAPs and T BMPs are similar to general GAPs. The rule however, require s mandatory flume sanitation of tomatoes with 150 mg/L free chlorine at pH 6.5 7.5 or another approved sanitizer for a maximum of 2 min (FDACS 2007). Florida is currently the only state that requires tomatoes be sanitized and no longer allows field packing unless tomatoes are first sanitized with the approved flum ing method. T ypically after harvest, tomatoes are transported in large bins to a packinghouse The tomatoes are unloaded in to a chlorinated water dump tank to reduce injury and remove debris The flume system then conveys tomatoes through sorting and packing ( Cantwell and others 2002). Water quality is therefore a crucial factor in postharvest tomato handling. Flume washing is considered a critical control point of HACCP because it is a point where microbial hazards can be prevented or reduced via proper sanitation. If sanitizer efficacy is not controlled, there is a pote ntial for pathogens to spread from one contaminated tomato to an entire lot because of commingling and cross contamination in the flume Sanitizers While sanitizers inactivate microorganisms, t he efficacy is usually limited to a 2 to 3 log 10 unit reduction (Gil and others 2009 ). Several studies show simply washing with potable water removes 1 to 2 log 10 units of microorganisms ( Beuchat 1998). In a flume operation, t he purpose of a sanitizer is meant to maintain wash water safety and prevent cross contamin ation among produce that commingle in the flume Wash water without a sanitizer can cause a build up of bacteria and subsequent biofilms (Gil and others 2009). While the flume may act primarily to prevent injury of produce during unloading as well as coo l and carry produce through to packing sanitizers can also be applied via spray or dip methods ( Adaskaveg and others 2002). HOCl ClO 2 and PAA
29 are a few sanitizers used for produce and are the only ones currently approved for use with Florida tomatoes ( F DACS 2007 ). Sodium hypochlorite Chlorine has been used to disinfect water since the mid 1800s (Clair and others 2003). C hlorine gas (Cl 2 ) can react with water (H 2 O) to form HOCl and hydrochloric acid (HCl) at pH < 3. HOCl also known as free chlorine, is the active form of chlorine solutions. Since c hlorine gas is toxic, hypochlorites (ClO ), specifically NaOCl are more commonly used in large scale operations and also produce HOCl with water (Clair and others 2003). Traditionally, NaOCl is used to sani tize tomatoes as it is effective, inexpensive and acts rapidly and nonspecifically ( Asaskaveg and others 2002). HOCl i s a weak acid with a pKa of 7.5. Therefore, at pH > 7.5, HOCl can dissociate into its ions, OCl and H + the inactive forms of chlorine At pH < 7.5, the undissociated form predominates. The exact ratio of HOCl to OCl and H + depends on the pH. A pH of approximately 7.5 is ideal to maintain a bout a 50 / 50 ratio of HOCl and its ions. ompletely understood but HOCl is believed to oxidize thiol groups ( SH) to disulfides (S S), sulfoxides (S O), or disulfoxides; destroy cellular activity of proteins; inhibit DNA synthesis by forming chlorinated derivatives of nucleotide bases; and disrupt cell membrane activity (McDonnell and others 1999). Besides pH, t he efficacy of HOCl is dependent on t emperature. At higher temperature, there is shorter contact time necessary between produce and wash water, but H OCl becomes more volatile ( Asaskaveg and others 2002). For immersion treatments, w hen H OCl is at a much lower temperature than produce, the resulting decrease in internal gas pressure may pull wash water inside produce, increasing risk of
30 contamination (Harris and others 2002). In one study, 1 0 tomatoes were immersed for 10 min in a suspension of E. carotovora where water was 20 C lower than tomato temperature (Bartz and Showalter 1981) Tomatoes were then rinsed in 50 mg/L chlorine and stored for 2 d. Tomatoes gained an average of 44% in wei ght and all tomatoes developed soft rot. When temperature differential between water and tomatoes was zero or positive, no decay occurred, even if slight water infiltration occurred (Bartz and Showalter 1981). Ideal tempera ture of the water is now recomm ended to be 5.5C greater than the pulp temperature of produce to be washed (Mahovic and others 2007). Accordingly, Florida T BMPs requires water temperature to be 5 C higher than pulp temperature when using chlorine ( FDACS 2007 ). H OCl is also affected b y the presence of other compounds in water O rganic matter and other compounds interfere with oxidation reactions by reacting with chlorine and preventing its disinfection reactions This is called chlorine demand. Enough H OCl must be added to water to reach breakpoint chlorination, the point at which all extraneous reactions are completed and chlorine can be used for disinfection (Clair and others 2003). The maximum concentration of H OCl allowed for food contact is 200 mg/L though it is often used bet ween 50 to 200 mg/L and must be monitored regularly ( CFR 2010b ; Thompson and others 2002). If chlorine reacts with organic matter, trihalomethanes (THMs) and haloacetic acids can form and vaporize. These compounds are believed to be human carcinogens (Cl air and others 2003). At pH 3 6, free chlorine can also combine with nitrogenous compounds to form chloramines that have lower antimicrobial activity and are eye irritants (Asaskaveg and others 2002).
31 The effectiveness of H OCl in flume water has been stu died with tomatoes. Felkey and others (2006) inoculated tomato surfaces with Salmonella and placed the m in a scale model flume at 25C with 150 mg/L free chlorine for 0 to 120 s. Salmonella concentration before treatment was 6.52 log 10 CFU/ml compared to 3.49 3.29 and 0.16 log 10 CFU/ml after 30, 60 and 120 s, respectively All treatment times were significantly different than the control. Another simulated flume study found a 5 log 10 unit reduction of Salmonella on tomato surfaces at 200 mg/L H OCl afte r 60 s at 35 C (Yuk and others 2005). O ther studies also have shown that H OCl is effective against Salmonella on tomato surfaces in water bath systems compared to controls though populations were never eliminated (Wei and others 1995 ; Zhuang and others 1 995). HOCl in flumes can also be ef fective against decay organisms. Vigneault and others (2000) immersed 20 tomatoes in water contaminated with E. carotovora and Rhizopus stolonifer for 10 min at 20 C. Seventeen tomatoes developed decay during storage fo r 7 d at 26 C. When contaminated water was chlorinated at 200 mg/L however, only 1 tomato developed decay. At 400 mg/L HOCl, no decay occurred in 14 d. Researchers also hydrocooled tomatoes to 15 C in a flume or shower (1,000 L/min*m 2 flow rate) contain ing water contaminated with 6 log 10 CFU/ml E. carotovora or R stolonifer and between 50 and 200 mg/L chlorine. After 10 d storage at 20 C, decay caused by E. carotovora only occurred with tomatoes exposed to nonchlorinated water. Decay caused by R stol onifer primarily occurred with nonchlorinated water, but sporadic cases was also observed with chlorinated water in both hydrocooling methods ( Vigneault and others 2000).
32 Chlorine dioxide Because o f some limitations of HOCl other sanitizing solutions have recently been studied and used. ClO 2 is less affected by pH and organic matter than H OCl thus will not form THMs (Clair and others 2003) ClO 2 works against microorganisms by disrupting cell permeability through oxidation reactions. It has a strong ox idizing capacity, though the oxidation reactions occur at a slower rate than with HOCl (Gomez Lopez and others 2009). ClO 2 can be used as a gas or be dissolved in water and used as an aqueous s anitizing solution (Gomez Lopez and others 2009 ; Fatica and Sc hneider 2009). Gas must be used in a closed chamber and is reported to have more penetrability than liquid sanitizers (Gomez Lopez and others 2009). An aqueous ClO 2 solution is generated by reacting sodium chloride (NaCl) with Cl 2 gas or reacting an aci d, such as HCl with sodium chlorite (NaClO 2 ) and diluting with water (Mari and others 2003). Disadvantages of aqueous ClO 2 include that is must be generated on site, produces byproducts and is toxic at high concentrations ( Adaskaveg and others 2002). It is also more expensive than NaOCl (Clair and others 2003). Currently, fresh produce can only be treated with a concentration of ClO 2 that does not exceed 3 mg/L residual chlorine (CFR 2010c ). A maximum of 5 mg/L ClO 2 in wash water is allowed for fresh pr oduce wash water and rinses and must be followed by a cooking process (EPA 2006). ClO 2 has been shown to be as effective as NaOCl in sanitizing iceberg lettuce against natural microflora Though there were no significant differences between the chlori ne sanitizers and a water control, p opulations of Pseudomonas Enterobacteriaceae and yeasts and molds were reduced by an average of 1.2 0.1 log 10 CFU/g when washed in 20 L of each solution under constant agitation ( Lopez Galvez and others 2010).
33 Pao an d others (200 9 ) studied the efficacy of 5 mg/L ClO 2 and an overhead spray system Overhead sprays continuously spray water above tomatoes on revolving brushes and the used wash water is collected. Inoculated t omatoes were dr ied for 24 h and then subjecte d to ClO 2 or water via a spray wash over brushes or immersion in 2 L for 1 0 to 60 s. It was found that spray washing with just water significantly reduced Salmonella on tomato surfaces after 10 s, at 3.2 0.3 log 10 units Spray washing with ClO 2 at 10 s increased efficacy to 4.4 0.5 log 10 units. No significant reduction was observed for immersion treatment with water or ClO 2 While ClO 2 might not re move pathogen s on produce surfaces in a water bath system it is an effective sanitizer against pathoge ns in water that may cross contaminate produce. When sterile tap water was inoculated with Salmonella Pao and others (2007) found a 5 log 10 unit reduction with 5 mg/L ClO 2 after 6 s. Similar results were seen with E carotovora ClO 2 was found to be eff ective against contamination from inoculated brushes to uninoculated tomatoes in the overhead spray system When tomatoes were placed on brushes that were previously inoculated with about 6.9 log 10 CFU/cm 3 Salmonella 5.7 log 10 CFU/cm 2 was transferred to tomato surfaces without any spray. ClO 2 at 5 mg/L for 10 to 60 s reduced cross contamination by 4.5 0.3 to 5.0 0.3 log 10 units (Pao and others 2009) Peroxyacetic acid PAA also known as peracetic acid (CH 3 COOOH) is a non chlorine based potent bioci de that is not affected by organic matter or pH. PAA is an organic acid that works faster than ClO 2 by denaturing proteins and disrupting cell permeability through a drop in intracellular pH (McDonnell and others 1999 ; Mari and others 2003). A maximum co ncentration of 80 mg/L PAA is permitted for produce contact and is made by reacting
34 acetic acid ( CH 3 COOH) with hydrogen peroxide (H 2 O 2 ) and diluting with water (CFR 20 10 d ). When tomato surfaces were inoculated with Salmonella Yuk and others (2005) found a 4 log 10 unit reduction after treating with 87 mg/L PAA in a circulating water bath at 35 C for 60 s PAA was less effective against Salmonella inoculated in stem scars and puncture wounds, with an average reduction of 2.12 log 10 units and 1.17 log 10 unit s, respectively after 60 s. In a different flume study, PAA at 75 mg/L for 60 and 120 s achieve d a bout a 4 log 10 unit reduction of Salmonella on bel l pepper and cucumber surfaces. Comparatively, 200 mg/L NaOCl achieved a 4 log 10 unit reduction in cucumbe r but only a 2 log 10 unit reduction in bell pepper (Yuk and others 2006). The efficacy of PAA against L monocytogenes has also been studied in ripened green coconuts from Brazil. Coconuts were inoculated and dried for 24 h at 36 C at 81% relative humi dity. Coconuts were immersed for 2 min in 80 mg/L PAA, 200 mg/L NaOCl or sterile distilled water, which resulted in average reductions of 4.7, 2.7 and 1.7 log 10 CFU/coconut, respectively. PAA was determined to be significantly more effective than NaOCl i n this study (Walter and others 2009) In a study of natural mesophilic flora 80 mg/L PAA was found to only achieve an additional 2.7 0.2 log 10 CFU/g reduction on cut cucumber and a 0.5 0.4 log 10 CFU/g reduction on parsley compared to a water rinse for 3 min (Shirron and others 2009). PAA reduced inoculated Salmonella by an additional 0.4 0.3 and 0.2 0.1 log 10 CFU/g on cucumber and parsley, respectively, compared to water In this study, t he low antimicrobial effect of PAA compared to water may be because parsley ha d large
35 surface area, leaves and crevices where bacteria could hide, and the cucumber was cut, exposing a large wound that is easily colonized by bacteria. Research Hypothesis and Objectives I mplementing safety measures can help min imize risk of outbreaks. The goal of this research was to study if a non immersion tomato sanitation method would be as effective as traditional flume methods. If so, use of the non immersion method would likely decrease volume of water and sanitizer req uired for sanitation, thereby sav e on packing costs. An overhead spray system is not always used for tomato sanitation but has been shown to be more effective than flumes against Salmonella inoculated on tomato surfaces (Pao and others 2009) Numerous st udies have examined the efficacy of different sanitizers on reducing Salmonella from tomato es in flume systems. I t is believed that no studies have compared efficacy of NaOCl ClO 2 and PAA with tomatoes in an overhead spray system An o verhead spra y s ystem is believed to achieve higher pathogen reduction on tomatoes because of the increased physical removal of bacteria from the mechanical action of the brushes and pressure of the spray, in conjunction with antimicrobial efficacy of sanitizers. I t was therefore hypothesized that lower concentrations of NaOCl than typically used in flumes would be able to achieve a 3 log 10 reduction of Salmonella on tomato surfaces T herefore 25, 50 and 100 mg/L NaOCl were tested. It was also hypothesized that all san itizers tested would achieve at least a 3 log 10 CFU/ml reduction of Salmonella on tomato surfaces, though treatment time could vary Water could achieve up to a 3 log 10 CFU/ml reduction of Salmonella from physical removal Overall, the overhead spray sys tem would be more effective at removing Salmonella from
36 tomato surfaces than a scale model flume system but perhaps not be as effective at removing natural tomato microflora. This research was meant to determine optimum operating parameters of the overh ead spray system including type of sanitizer, concentration of sanitizer and contact time, and develop them into useful recommendations for the tomato industry. Doing so would potentially benefit consumers and the tomato industry by providing a scientific basis for using an overhead spray sanitation system The objectives of this research were to: 1. Establish growth curves for rifampicin resistant Salmonella strains, S Typhimurium, S Braenderup, S Enteri tidis S. Newport, and S Javiana. 2. Determine Salmo nella recovery from inoculated tomatoes 3. Determine the extent of cross contamination from inocu lated tomatoes to uninoculated tomatoes via brush rollers in the overhead spray system. 4. Examine the efficacy of NaOCl in the overhead spray system at different c oncentrations and treatment times against Salmonella on tomato surfaces 5. Examine the efficacy of other sanitizers in the overhead spray system against Salmonella on tomato surfaces 6. Compare a scale model flume against the overhead spray system for the redu ction of Salmonella from tomato surfaces. 7. Examine the efficacy of sanitizers in the overhead spray system against natural microflora on tomato surfaces.
37 Table 2 1. M ultistate tomato outbreaks caused by Salmonella in the United States Year Tomato type Se rovar Cases S ource of contamination a Reference 1990 round Javiana 176 Packinghouse water bath Hedberg and others 1999 1993 round Montevideo 100 Packinghouse water bath Hedberg and others 1999 1999 round Baildon 86 Packinghouse or farm Cummings and other s 2001 2000 Thompson 43 CDC 2010c 2002 Roma Javiana 159 CDC 2002 a 2002 round Newport 510 Irrigation pond water Greene and others 2008 2004 Roma Braenderup 125 Packinghouse or farm CDC 2005 Gupta and others 2007 2004 Roma Javiana 3 83 Tomato slicing facility CDC 2005 Typhimurium 27 Anatum 5 Thompson 4 Muenchen 4 2005 round Newport 72 Irrigation pond water CDC 2007; Greene and others 2008 2005 Roma Braenderup 82 Wild animals in fields CDC 2007 2006 Newport 115 CDC 2007 2006 r ound Typhimurium 190 Packinghouse CDC 2007 2007 Newport 65 CDC 2010c a Tomatoes were all domestically grown or of unconfirmed /unspecified source.
38 CHAPTER 3 MATERIALS AND METHOD S Bacterial Strains Salmonella enterica serovars used in this study w ere S Typhimurium (ATCC 13311), S. Braenderup (ATCC BAA 664), S. Enteri tid is (ATCC 4931), S. Newport (ATCC 6962) and S. Javiana (ATCC BAA 1593). Salmonella strains were selected based on their association with fresh tomato outbreaks. Bacterial strains were a dapted to be 200 mg/L rifampicin (Fisher, Fair Lawn, NJ) resistant. A 10,000 mg/L rifampicin stock was made by dissolving 0.4 g rifampicin in 40 ml methanol. The stock solution was filter sterili zed (0.20 m pore size, Fisher ) and stored at 2C until use d. Cultures were stored at 80 C in CryoCare beads (Key Scientific, Round Rock, TX) and on tryptic soy agar (TSA) ( Difc o Sparks, MD) slants at 2C. To prepare broth cultures each strain w as streaked for isolation on TSA supplemented with 200 mg/L rifa mpicin (denoted TSA/ rif 200) and incubated at 37 C After 24 h, an individual colony was transferred to 10 ml tryptic soy broth (TSB) (Difc o ) supplemented with 200 mg/L rifampicin (TSB/rif200) and incubated at 37 C After 24 h, 10 l of culture was transferred to fresh TSB/rif200 and incubated at 37 C. Cultures were subsequently transferred to fresh TSB/rif200 at least once every 48 h. Inoculum Preparation To prepare the inoculum, cultures underwent a successive three day transfer the first two in 10 ml TSB/rif200 as described above The final transfer was into 20 ml TSB/rif200 t o ensure a 9.0 log 10 CFU/ml or higher inoculum level T he five 20 ml cultures were combined as a 100 ml cocktail and centrifuged at 4000 x g for 10 min The culture was then washed by pouring o ff the supernatant and suspending the pellet
39 in 10 ml buff ered peptone water (BPW) (Difco ) The culture was centrifuged and washed two more times before adding a final 10 ml of BPW to complete the inoculum. TSB an d TSA used in experiments were supplemented with 80 mg/L of rifampicin (TSB/rif80 or TSA/rif80) to reduce stress on organisms yet still select for resistant strains. Inoculum was serially diluted and plated with TSA/rif80 to determine initial concentratio n. All serial dilutions in experiments were performed 1:10 in BPW. Growth Curves Growth of the five Salmonella s trains was individually measured each hour over 12 h Triplicate broth cultures were prepared individually and underwent a successive three day transfer in TSB/rif200. To begin the experiment, cultures were serially diluted to approximately 6 log 10 CFU/ml. One milliliter of culture was transferred to 99 ml TSB/rif200. The new c ultures were incubated at 37 C for 12 h. At each hour including hour 0, cultures were serially diluted and dilutions near the countable range (25 250 CFU /ml) were plated by pour plate with TSA/rif80. Original cultures were also plated. Plates were incubated at 37 C for 48 h and then counted. Average log 10 count at each hour was calculated for the three cultures per Salmonella strain Analysis of variance ( A NOVA ) using Statistica (Statsoft, Tulsa, OK) to determine differences among strain concentration at each hour. S ignificant difference was determined at p<0.05. Tomato Inoculation and Plating Unwashed, unwaxed, mature green t omatoes were obtained from Pacific Tomato Growers ( Palmetto, FL and Tracy, CA ) and DiMare Fresh ( Riverview, F L ) New i noculum was prepared fo r every experiment I noculum was approximately
40 9.5 log 10 CFU/ml. I noculum was plated to determine the exact concentration. For each experiment, tomatoes were spot inoculated in a circle around the blossom scar with 10 spots of 10 l inoculum each, for a total of 100 l inoculum per tomato, or about 8.5 log 10 CFU/tomato. Tomatoes were dried for 2 h in a fume hood To obtain bacterial counts tomatoes were placed in individual sterile Stomacher bags (Fisher ) containing 100 ml BPW supplemented with 0.1% sodium thiosulfate (Na 2 S 2 O 3 ) (Fisher ) to inactivate chlorine (Kemp and Schneider 2000) Chlorine was inactivated to be able to determine the effect of chlorine treatment at specific time points and not any residual effects. B ags were shaken and tomato su rfaces were rubbed for a total of 1 min per bag using a rub shake method (Zhuang and others 1995) Tomatoes were serially diluted (1:10) and pour plate d with TSA/rif80. P lates were incubated at 37 C and counted after 48 h. Negative controls of all media were pour plated. For the cross contamination and efficacy studies, f ive inoculated tomatoes ( the positive controls ) were plated to determine recovery of inoculum. The positive controls were inoculated and dried for 2 h but not treated. For each exper iment, an uninoculated tomato was u sed as a negative control and was plated to ensure no rifampicin resistant organisms were found on tomatoes. Recovery Study A recovery study w as conducted to determine how much Salmonella w ould be recovered from tomato su rfaces after inoculation and drying for 2 h. Three inoculum cultures were prepared One tomato was inoculated per culture. Tomatoes were dried for 2 h in a biological hood and then plated. Plates were incubated at 37 C and counted after 48 h. The expe riment was performed in triplicate and log 10 colony counts
41 using Statistica to determine differences between replicates. The amount of Salmonella inoculated on tomatoes was compared to amoun t of Salmonella recovered after 2 h drying Sanitizer S olution P reparation Sodium H ypochlorite (NaOCl) NaOCl solution was prepared at room temperature. NaOCl was prepared by adding 22 ml 5.65 to 6% NaOCl (Fisher ) to 10 L deionized (DI) water (University o f Florida) for a final concentration of 100 mg/L. Other concentrations were calculated from this ratio For experiments using the overhead spray system, additional NaOCl was added to adjust for the aerosoliz ation of NaOCl from the spray nozzles that lowe red concentration The NaOCl solution was adjusted to pH 6.5 with HCl (Fisher ) A Hach DR/890 colorimeter (Hach Co., Loveland, CO) was used to verify hypochlorous acid /chlorite ion (free chlorine) concentration using Hach method 8091 The DPD method use s Accu V ac free chlorine ampoules (Hach Co ) that contain the indicator, N, N diethyl p phenylenediamine sulfate which forms a pink color when reacted with chlorine. The range for the ampoules is 0 to 2.0 0 mg/L Cl To measure concentration, NaOCl soluti on was diluted 1:100 in chlorine demand free water or 18 MOhm water, from Barnstead Nanopure Diamond Lab Water System (Barnstead, Dubuque, IA). An ampoule was filled with the dilution and was measured by the colorimeter. Chlorine D ioxide (ClO 2 ) Aqueous ClO 2 was generated with Selectrocide 2L500 (Selective Micro Technologies, Canal Winchester, OH). Selectrocide 2L500 is a pouch containing
42 2 L of water was poured in to t he pouch, ClO 2 was activated at 500 mg/L. ClO 2 solution was only used in experiments in the overhead spray system. For each experiment, 180 ml ClO 2 was added to 10 L DI water to make a solution that delivered 5 mg/L ClO 2 from the spray nozzles. Concent ration of ClO 2 was measured with a Hach DR/890 colorimeter an d Hach m ethod 10126 using Accu V ac free chlorine ampoules (Hach Co.) The range of the ampoules in this test is 0 to 5.00 mg/L ClO 2 though the colorimeter display s c oncentrations up to 5.50 mg/L Peroxyacetic A cid (PAA) PAA was made by diluting a commercially available concentrate, Tsunami 100 solution (Ecolab Inc., St. Paul, MN) with DI water to obtain a working concentration of 80 mg/L Tsunami 100 contains 15% PAA and 11% H 2 O 2 PAA solution was only used in experiments in the overhead spray system. For each experiment, 5.24 ml Tsunami 100 was added to 10 L DI water to make a solution that delivered 80 mg/L PAA from the spray nozzles. PAA concentration was measured with an R Qflex 10 meter an d Reflectoquant test strips (EM D Chemicals, Gibbstown, NJ) with a range of 75 to 400 mg/L The meter measures PAA concentration spectrophotometrically by the color change of the strip when exposed to PAA. Water C ontrol D eionized water (University of Flori da) was used for water sprays in the studies testing the overhead spray system As a control, water was tested in addition to the other sanitizers to see if reductions of Salmonella were due to the sanitizer or the mechanical action of the spray and brush rollers Water was also used as a control in
43 the flume study to see if Salmonella reductions were due to the sanitizer or the mechanical action of the circulating water W ater was supplemented with 0.1% Na 2 S 2 O 3 to inactivate any chlorine that re mained i n the sanitation system. Overhead Spray System A custom built overhead spray and brush roller system was used in the cross contamination study and efficacy studies (Figure 3 1) The overhead spray system wa s located in a biological fume hood. Two nylon r ollers ( 4 6 cm long and 12 cm dia ) s at alongside in a box measuring 46 cm by 34 cm. The rollers rotated in the same direction at 180 rpm. The t omato to be tested s a t in the valley between the two rollers and revolved on its axis in a direction that depend ed on its shape The distance between the top of the brush rollers and spray nozzles wa s 1 3 cm. Th re e spray nozzles (Spraying Systems, Co., Wheaton IL) release d a cone shaped spray at a pressure of 12 psi and flow rate of 21.4 ml/s A 20 L bucket with a spigot fed sanitizer through piping to the nozzles. At least 6 L of solution had to be in the bucket at all times in order for the spigot to be submerged Cross Contamination Stud y The c ross contamination stud y w as conducted to determine if there wa s c ross contamination of Salmonella from inoculated tomatoes to uninoculated tomatoes via brush ro llers in the overhead spray system NaOCl ( 100 mg/L ) was tested for its ability to prevent cross contamination. Ten liters of 100 mg/L NaOCl was prepared and i ts c oncentration as released from the spray nozzles was verified. The experiment had four stages characterized by what was sprayed in the overhead spray system. Stage 1 was an initial 100 mg/L NaOCl spray for 15 s to wet the rollers. After the spray, ea ch roller was swabbed in a 10 cm by 10 cm square area to determine initial contamination,
44 if any The swabs were each placed in 10 ml BPW with 0.1% Na 2 S 2 O 3 In stage 2, five inoculated tomatoes were placed on the brush rollers and sprayed with 100 mg/L N aOCl for 15 s. Tomatoes were removed and placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 Each roller was also swabbed in the same way and area as in stage 1 to determine transfer of inoculum to brush rollers In stage 3, five uninocula ted tomatoes were placed on rollers and sprayed with 100 mg/L NaOCl for 15 s. Tomatoes were removed and placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 In stage 4, water with 0.1% Na 2 S 2 O 3 was sprayed for 15 s to neutralize residual chlo rine. Each stage was performed continuously. Five positive control tomatoes were run to determine initial inoculum level and calculate reduction of Salmonella from inoculated tomatoes Experiments were performed in triplicate with different areas of the rollers swabbed each time. Triplicate experiments were also conducted with water as a control. The same procedure was followed as above except water with 0.1% Na 2 S 2 O 3 was sprayed instead of NaOCl. In stage 4, 100 mg/L NaOCl was sprayed instead of water with 0.1% Na 2 S 2 O 3 to sanitize rollers between experiments. All tomatoes were plated with TSA/rif80, incubated at 37 C and counted after 48 h. Average log 10 recovery of Salmonella on uninoculated tomato surfaces and on brush rollers was calculated for 100 mg/L NaOCl and the water control. were performed with Statistica to determine the difference between the ability of NaOCl and water at prevent ing contamination of brush rollers, which then would prevent cross contam ination of uninoculated tomatoes
45 Sodium Hypochlorite Efficacy Study The efficacy of NaOCl for reducing Salmonella on tomato surfaces in the overhead spray system was determined at 25, 50 and 100 mg/L NaOCl Inoculated tomatoes were placed on brush roller s and Na OCl was released through the overhead spray Water was run as a control. NaOCl solution was prepared in 10 L batches and concentration released from the spray nozzles was verified. For each NaOCl concentration and the water control, five tomatoe s were tested for each of four treatment times of 5, 15, 30 and 60 s. After treatment, tomatoes were placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 Tomatoes were plated with TSA/rif80, incubated at 37 C and counted after 48 h Experim ents were run in triplicate Five positive controls were also run to calculate average log 10 reduction of Salmonella from tomato surfaces performed with Statistica to determine differences among treatment time and NaOCl concentration. Sanitizer Efficacy Study The efficacy of 100 mg/L NaOCl, 5 mg/L ClO 2 80 mg/L PAA and a water control for reducing Salmonella on tomato surfaces in the overhead spray system was determined for treatment times of 5, 15, 30 an d 60 s Five positive controls were run as time 0. Inoculated tomatoes were placed on brush rollers and sanitizer was released through the overhead spray For each sanitizer and the water control, five tomatoes were tested per treatment time After tre atment, tomatoes were placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 Tomatoes were plated with TSA/rif80, incubated at 37 C and counted after 48 h Experiments were run in triplicate and average log 10 reduction of Salmonella from tomat o surfaces was calculated
46 with Statistica to determine differences among treatment time and sanitizers. Flume vs. Overhead Spray Comparison Study A study examining the efficacy of a scale model flum e for reducing Salmonella from tomato surfaces was conducted to compare to the overhead spray system A Precision circulating water bath ( Jouan, Inc., Winchester, VA ) measuring 38.7 cm by 30.5 c m and 19.0 cm deep was used as the scale model flume. Inocul ated tomatoes were tested in 10 L of NaOCl (100 mg/L) and 10 L water at 25 C Five tomatoes were tested per sanitizer and treatment time of 5, 15, 30 and 60 s Five positive controls were also run as time 0. For a given treatment time, all five t omatoes were placed in the flume at once with the blossom scar end down so that the inoculum was submerged. After the specific time, tomatoes were removed with sanitized metal tongs and placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 Each trea tment time was tested sequentially with no change of NaOCl or water between times. Flume water w as tested at time 0 and after the 15 and 60 s treatment times to quantify the buildup of Salmonella that could cross contaminate subsequent tomatoes. Tomatoes and sanitizer samples were plated with TSA/rif80, incubated at 37 C and counted after 48 h Experiments were run in triplicate and average log 10 reduction of Salmonella from tomato surfaces was calculated Range test were per formed with Statistica to determine differences among treatment time and sanitizer. Flume data was also compared statistically with 100 mg/L NaOCl overhead spray data from the sodium hypochlorite efficacy study.
47 Natural Tomato Microflora Study The efficac y of sanitizers in the overhead spray system against natural tomato surface microflora was examined. Sanitizers tested were NaOCl (100 mg/L) ClO 2 (5 mg/L) PAA (80 mg/L) and a water control Five positive controls were run as time 0. Uni noculated tomat oes were placed on brush rollers and spray ed for 5, 15, 30 and 60 s For each sanitizer and the water control, five tomatoes were tested per treatment time After treatment, tomatoes were placed in Stomacher bags containing 100 ml BPW with 0.1% Na 2 S 2 O 3 Tomatoes were plated with TSA, incubated at 37 C and counted after 48 h Experiments were run in triplicate and average log 10 reduction of bacteria from tomato surfaces was calculated performed with Statistic a to determine differences among treatment time and sanitizers.
48 A B Figure 3 1. Overhead spray system A) Entire overhead spray system, B) Brush rollers and spray nozzles.
49 CHAPTER 4 RESULTS Growth Curves The growth of five rifampicin resistant Sa lmonella strains, S. Typhimurium S. Braenderup S. Enteritidis S. Newport and S. Javiana was measured over 12 h at 37 C in TSB/rif200 Average log 10 concentration of each strain at each hour was calculated from triplicate cultures ( Table 4 1 and Figure 4 1 ) Results s how that all five Salmonella strains had an initial average concentration of approximately 4 log 10 CFU/ml at hour 0 and grew to approximately 9 log 10 CFU/ml by hour 12 There was no statistically significant difference in average log 10 con centration between strain s at hour 0 or after 1 h of growth (p>0.05). S ignificant differences among strains beg an at hour 2 (p<0.05) Differences in average concentration were observed in at least one strain from hour 2 through hour 9 though the differe nce between the most concentrated and least concentrated strain was always less than 1.0 log 10 CFU/ml Average difference between the most concentrated and least concentrated strain was 0.70 0.15 CFU/ml. Between hour 2 and 9, S. Enteritidis and S Javi ana each exhibited the highest average log 10 concentrations for 4 of the 8 h. Conversely, S. Typhimurium exhibited the lowest average log 10 concentration for 6 of the same 8 h. At each of the last 3 h of the measured growth there were no significant dif ferences in a verage log 10 concentration among strains (p>0.05) Additionally, the concentration of each strain was not significantly different from each other during at least the last 3 h of growth but w as significantly different from the previous hours o f growth, indicating growth reached a plateau or stationary phase (not shown in Table 4 1)
50 Recovery Study Surface inoculated tomatoes were tested for recovery of Salmonella after dryin g for 2 h at room temperature. Average log 10 concentration of inoculum Salmonella recovery from tomatoes and Salmonella loss was calculated from triplicate experiments (Table 4 2). Average concentration of inoculum cultures was 9.67 0.11 log 10 CFU/ml. Since tomato es w ere inoculated with 100 l inoculum, each tomato was inoculated with an average of 8.67 0.11 log 10 CFU/tomato. Average concentration of Salmonella inoculated on tomatoes for each experiment replicate was 8.83, 8.60 and 8.57 log 10 CFU/ml There was no significant difference in inoculum concentration between replicates (p>0.05). After 2 h drying, tomatoes were placed in 100 ml BPW and plated to determine Salmonella recovery. Be cause tomatoes were diluted in 100 ml BPW, the limit of detection was 2 log 10 CFU/ ml Recovery data w as adjusted 2 log 10 CFU/m l because colony counts were 2 log 10 CFU/ml lower than what was actually on the tomato surface For example, average concentration re covered from tomatoes in experiment 1 was 6.09 0.79 log 10 CFU/ml but was reported as 8.09 0.7 9 log 10 CFU/ml Average recovery of Salmonella from tomato surfaces in each of the three experiment replicates was 8.09, 7.86 and 7.90 log 10 CFU/ml, with an overall average recovery of 7.95 0.30 log 10 CFU/ml. There was no significant difference in reco very between experiment replicates (p>0.05). Statistical significance between inoculum and recovery varied among replicates. There was no significant difference between inoculum and recovery in experiment 1 but there were differences in the last two exp eriments (p<0.05). Overall average log 10 inoculum and recovery was significantly different with p = 0.001, therefore there wa s a
51 statistically significant loss of Salmonella after the 2 h drying Average loss of Salmonella in the three replicates was 0.7 9, 0.75 and 0.67 log 10 CFU/ml respectively There was no significant difference in average loss between replicates (p>0.05). Overall average loss of Salmonella was 0.73 0.24 log 10 CFU/ml. Cross Contamination Study The cross contamination study test ed whether Salmonella would be transferred from inoculated tomatoes to uninoculated tomatoes via brush rollers in the overhead spray system during 100 mg/L NaOCl or water sanitation. A 10 L 100 mg/L NaOCl solution was prepared by adding 22 ml of 5 .65 to 6 % NaOCl to 10 L water. Aerosolization of chlorine in the overhead spray nozzles required an addition of 3 to 5 ml NaOCl to the 10 L solution to deliver 100 mg/L NaOCl from the nozzles. Concentration of free chlorine measured 95 mg/L before the experiment and 85 mg/L after. Ten liters of water supplemented with 0.1% sodium thiosulfate was also prepared. Five inoculated tomatoes and five uninoculated tomatoes were tested per sanitizer and each sanitizer was tested in triplicate. Inoculated tomatoes were sanitized for 15 s followed by 15 s sanitation of uninoculated tomatoes. Average reduction of Salmonella from inoculated tomatoes after the 15 s spray was 3.40 1.02 log 10 CFU/ml for NaOCl and 2.86 0.43 log 10 CFU/ml for water. There was no significant difference in average log 10 reduction of Salmonella from inoculated tomatoes between NaOCl and water (p>0.05). Spraying uninoculated tomatoes with water immediately after spraying inoculated tomatoes resulted in a transfer of 4.88 0.41 log 10 CFU/ml Sa lmonella (Table 4 3) Using 100 mg/L NaOCl in the spray significantly reduced cross contamination by 2.25 log 10 CFU/ml (p=0.0001) Average recovery of Salmonella from tomatoes sprayed
52 with NaOCl was 2.63 0.28 log 10 CFU/ml These values were adjusted 2 log 10 CFU/ml to because tomat oes were rinsed in 100 ml BPW, causing a dilution effect. C ontamination of brush rollers with Salmonella from inoculated tomatoes was also determined ( Table 4 3 ) Each of the two brush rollers was swabbed before and after con tact with inoculated tomatoes. Any initial contamination on brush rollers before contact with inoculated tomatoes was subtracted from colony counts after contact. Colony counts were averaged for the two rollers Average recovery of Salmonella from brush rollers was not statistically significant between NaOCl and water at 1.24 and 1.95 log 10 CFU/ cm 2 recovered, respectively (p>0.05) Sodium Hypochlorite Efficacy Study The efficacy of NaOCl at reducing Salmonella on tomato surfaces in the overhead spray s ystem was tested at concentratio ns of 25, 50 and 100 mg/L NaOCl and a water control. NaOCl solutions were prepared in 10 L buckets and used at room temperature. Five tomatoes were tested per treatment time of 5, 15, 30 and 60 s. Average log 10 reduction of Salmonella on tomato surfaces was calculated from triplicate experiments per NaOCl concentration and water by comparing recovery of Salmonella on treated tomatoes to f ive positive control tomatoes S tatistically significant differences in efficacy of NaOCl and water were observed depending on concentration and treatment time (Table 4 4 and Figure 4 2) There was no significant difference between NaOCl concentrations and water at 5 s, with an average reduction of 1.37 0.24 log 10 CFU/ml (p>0.05). At 15 s, 100 mg/L NaOCl was significantly different from 25 and 50 mg/L NaOCl and water by achieving a 3.98 1.78 log 10 CFU/ml reduction of Salmonella. A 3 log 10 unit reduction was also achieved by 50 mg/L NaOCl a t 30 s and 25 mg/L NaOCl a t 60 s. A 3 log 1 0 unit reduction was not
53 observed for the water control though it did achieve a 2.95 0.44 log 10 CFU/ml reduction a t 60 s. At 30 s, a 5 .55 0.37 log 10 CFU/ml reduction was observed for 100 mg/L NaOCl Increasing treatment time to 60 s did not signifi cantly increase efficacy for 100 or 50 mg/L, suggesting there is a limit in sanitizer efficacy. NaOCl at 25 mg/L was not more effective than water at 30 s and shorter treatment times, but it achieved a 4.23 1.11 log 10 CFU/ml reduction at 60 s. Efficacy of water did not increase much as treatment time increased. Highest reduction by water was seen at 60 s at 2.95 0.44 log 10 CFU/ml. Sanitizer Efficacy Study The sanitizer efficacy study examined the efficacy of NaOCl (100 mg/L) ClO 2 (5 mg/L), PAA (80 m g/L) and a water control in the overhead spray system for reducing Salmonella inoculated on tomato surfaces. All sanitizers were prepared in 10 L b atches and held in a bucket to feed into the overhead spray system All sanitizers required a higher concen tration in the bucket to offset the aerosolization and loss of concentration as it was released from the spray nozzles. Average pH and concentration of each sanitizer before and after triplicate experiments was verified ( Table 4 5 ) For each sanitizer, f ive tomatoes were tested per treatment time of 5, 15, 30 and 60 s. Average log 10 reduction was calculated from triplicate experiments by comparing recovery of Salmonella from treated tomato surfac es and five positive controls. Statistically significant d ifferences were observed in at least one sanitizer at all treatment times (Table 4 6 and Figure 4 3) After on ly 5 s, PAA nearly reached a 3 log 10 unit reduction at 2.79 0.94 log 10 CFU/ml. Conversely NaOCl, ClO 2 and water had a 1.94, 1.87 and 1.86 log 10 CFU/ml reduction, respectively. PAA consistently achieved about a 1 log 10 unit higher reduction than the other sanitizers for 5, 15 and
54 30 s treatment. A t 15 s, all sanitizers reached at least a 3 log 10 reduction of Salmonella including water. Incre asing treatment time to 30 s did not significantly increase reduction by ClO 2 or NaOCl but did for PAA to 5.50 0.12 log 10 CFU/ml At 60 s, average log 10 reductions by NaOCl, ClO 2 and PAA were all significantly higher than the water control. NaOCl, ClO 2 and PAA had a 5.51, 4.85 and 5.52 log 10 CFU/ml reduction, respectively and were not significantly different from each other (p<0.05) Compared to their efficacy at 30 s, efficacy of ClO 2 significantly increased at 60 s whereas PAA did not. Water only h ad a 3.75 log 10 CFU/ml reduction at 60 s Flume vs. Overhead Spray Comparison Study The efficacy of NaOCl (100 mg/L) and water at 25 C for reducing Salmonella on tomato surfaces was compared between a scale model flume and the overhead spray system Ten liters each of NaOCl and water supplemented with 0.1% sodium thiosulfate were prepared. Average measured starting pH for triplicate experiments was 6.50 for NaOCl and 7.16 for water. Average measured concentration of NaOCl before and after experim ents was 102 and 100 mg/L, respectively. F ive inoculated tomatoes were tested per treatment time of 5, 15, 30 and 60 s. Average log 10 reduction of Salmonella from tomato surfaces was calculated and compared to overhead spray 100 mg/L NaOCl data from the sodiu m hypochlorite efficacy study. The flume water control did not produce significantly different reductions in Salmonella among treatment times. Average reduction by the flume water control was 1.03 0.74 log 10 CFU/ml no matter how long tomatoes were treated. At 5 s, no significant differences were found among any sanitizer or sanitation system (p>0.05). Average reduction of Salmonella at 5 s was 0.97 0.56 log 10 CFU/ml.
55 Statistically significant differences were found between NaOCl and water, and overhead spray and flume starting at 15 s (Table 4 7 and Figure 4 4) O verhead spray NaOCl treatments of at least 15 s significantly reduced more Salmonella from tomatoes compared to flume treatments (p<0.05) At 15 s, a verage reduction by NaOCl in the overhead spray was 3.98 1.78 log 10 CFU/ml. Conversely, NaOCl and water in the flume had an average reduction of 1.25 and 0.98 log 10 CFU/ml respectively. Reduction by NaOCl was enhanced to 5.55 0.37 log 10 CFU/ml a t 30 s when treated with the overhead spray NaOCl in the flume achieved a 3.17 2.59 log 10 CFU/ml reduction at 30 s Increasing spray time to 60 s did not result in a significantly higher reduction in either the overhead spray or flume. Concentration of Salmonella in the flume was teste d at time 0 and after treating contaminated tomatoes for 15 and 60 s. At time 0, Salmonella was undetectable in flume water. Salmonella was recovered from flume water at an average of 4.54 0.37 log 10 CFU/ml after the 15 s treatment and 5.05 0.16 log 1 0 CFU/ml after the 60 s treatment, though these populations were not significantly different from each other (p>0.05). NaOCl (100 mg/L) effectively eliminated Salmonella in the flume as populations were undetectable throughout the study. Natural Tomato Microflora Study NaOCl (100 mg/L), ClO 2 (5 mg/L), PAA (80 mg/L) and a water control were tested for their ability to remove natural microflora from tomato surface s in the overhead spray system. Sanitizers were prepared in 10 L batches and required a highe r concentration feeding into the spray system to offset aerosolization as it was released from the spray nozzles. Starting pH and concentration of sanitizers before and after use were verified (Table 4 8). For each sanitizer, five tomatoes were tested pe r treatment time of 5, 15,
56 30 and 60 s. Five control tomatoes were also run to determine initial concentration of natural microflora. Average log 10 reduction was found from triplicate experiments by comparing recovery of bacteria from treated tomatoes an d the five control tomatoes Average initial population of natural microflora on tomato surfaces was 5.31 0.57 log 10 CFU/ml. No difference was observed among sanitizers a t the 5 s treatment (p>0.05) (Table 4 9 and Figure 4 5) A t 15 s, only NaOCl signi ficantly reduced more natural microflora than water with a 0.81 0.60 log 10 CFU/ml reduction (p<0.05). Similar efficacy was observed at 30 s of treatment. NaOCl reduced microflora by an average of 1.41 0.90 log 10 CFU/ml, which was significantly higher than all other sanitizers (p<0.05). ClO 2 PAA and water had a 0.55, 0.84 and 0.56 log 10 CFU/ml reduction, respectively at 30 s. Increasing treatment time to 60 s did not significantly affect e fficacy Generally, the efficacy of the sanitizers increased significantly with increased treatment time. The exception is ClO 2 whose efficacy was not affected by time of exposure. As mentioned, NaOCl achieved a significantly greater reduction of microflora after 30 s treatment compared to 5 and 15 s. While PAA did not significantly reduce more microflora compared to the water control throughout the study it did achieve a significantly higher reduction at 60 s of treatment compared to 5 s. The a verage r eduction of microflora by PAA was 1.19 0.28 log 10 CFU/ml at 60 s and 0.48 0.20 log 10 CFU/ml at 5 s. Similarly, water achieved a significantly higher reduction after 60 s of treatment compared to 15 s. Reduction of microflora by water was 0.88 0.12 log 10 CFU/ml at 60 s and 0.26 0.27 log 10 CFU/ml at 15 s.
57 Table 4 1. Average log 10 concentration of rifampicin resistant Salmonella strains over 12 h at 3 7 C Salmonella growth ( log CFU/ml) a Hour Typhimurium Braenderup Enteritidis Newport Javiana 0 3.91 0.04 a 4.09 0.04 a 4.09 0.02 a 3.98 0.11 a 4.18 0.10 a 1 4.01 0.09 a 4.08 0.04 a 4.10 0.15 a 4.04 0.13 a 4.14 0.05 a 2 4.28 0.00 b 4.29 0.03 b 4.65 0.08 a 4.22 0.13 b 4.06 0.43 b 3 4.93 0.16 b 4.98 0.15 ab 5.24 0.06 a 4.71 0.23 b 4.97 0.16 ab 4 5 .34 0.17 b 5.79 0.05 a 5.97 0.18 a 5.36 0.34 b 5.72 0.27 a 5 5.94 0.20 c 6.39 0.01 ab 6.63 0.03 a 6.23 0.27 b 6.64 0.11 a 6 6.69 0.17 b 7.08 0.04 a 7.12 0.07 a 7.08 0.37 a 7.29 0.29 a 7 7.21 0.07 c 7.85 0.11 b 7.89 0.07 b 7.68 0.38 b 8.18 0.12 a 8 7.84 0.07 c 8.53 0.07 a 8.68 0.13 a 8.25 0.39 b 8.71 0.06 a 9 8.37 0.11 c 9.01 0.02 ab 8.72 0.17 b 8.91 0.29 ab 9.14 0.08 a 10 8.90 0.07 a 9.08 0.00 a 8.93 0.06 a 9.11 0.07 a 9.22 0.09 a 11 8.91 0.08 a 8.93 0.02 a 9.14 0.02 a 9.04 0.09 a 9.03 0.09 a 12 9.08 0.15 a 8.91 0.05 a 9.15 0.03 a 9.06 0.07 a 9.17 0.26 a a Values are mean standard deviation of triplicate strains gro wn in tryptic soy broth supplemented with 200 mg/L rifampicin (n=3) Means with same letter in the same row (abc) are not statistically different ( p <0.05)
58 Figure 4 1. Average log 10 concentration of rifampicin resistant Salmonella strains over 12 h at 3 7 C Error bars represent standard deviation of triplicate strains
59 Table 4 2. Average log 10 recover y and loss of Salmonella from tomato surfaces after 2 h drying at room temperature Log CFU/ml Salmonella Experiment Inoculation a Recovery b Loss c 1 8. 83 0.17 a x 8.09 0.79 a x 0.79 0.70 x 2 8.60 0.30 a x 7.86 0.26 b x 0.75 0.26 x 3 8.57 0.08 a, x 7.90 0.30 b, x 0.67 0.30 x Average 8.67 0.11 a 7.95 0.30 b 0.73 0.24 a Values are mean standard deviation of 3 inoc ulum cultures inoculated on tomato surfaces. b Values are mean standard deviation of Salmonella recovered from 3 tomatoes, adjusted for the 2 log 10 CFU/ml loss from rinsing tomatoes in 100 ml BPW. c Values are mean standard deviation of difference betwee n inoculum and recovery of Salmonella from tomatoes. Means with same letter in the same row (ab) or in the same column (x) are not statistically different ( p <0.05) Table 4 3. Cross contamination of Salmonella from inoculated tomatoes to uninoculated t omatoes and brush rollers Salmonella log 10 recovery Treatment Uninoculated tomatoes a Brush rollers b Na OCl 100 mg/L 2 .63 0.28 y 1.24 1.12 x Water control 4 .88 0.41 x 1.95 0. 27 x a Values are mean standard deviation (log 10 CFU/ml) of trip licate experiments with 5 tomatoes each (n=15) adjusted for the 2 log 10 CFU/ml loss from rinsing tomatoes in 100 ml BPW b Values are mean standard deviation (log 10 CFU/cm 2 ) of triplicate experiments of 4 swab s each adjusted to 1 swab each for the subtra ction of initial contamination and averag e between 2 rollers (n=3). Means with same letter in the same column (xy) are not statistically different ( p <0.05)
60 Table 4 4 Average l og 10 reduction of Salmonella after overhead spray treatment of 25, 50 and 100 mg/L sodium hypochlorite and water control Salmonella log reduction (log CFU/ml) from tomatoes a Treatment (s) 100 mg/L 50 mg/L 25 mg/L Water control 5 1.66 0.51 a, z 1.43 0.79 a, z 1.02 0.24 a, z 1.36 0.36 a, y 15 3.98 1.78 a, y 2.84 1.10 b, y 1.96 0. 06 b, y z 2.29 0.36 b, xy 30 5.55 0.37 a, x 4.24 0.75 b, x 2.50 0.66 c, y 2.52 0.23 c, x 60 5.51 0.96 a, x 4.96 1.19 ab, x 4.23 1.11 b, x 2.95 0.44 c, x a Values are mean standard deviation of triplicate e xperiments of 5 tomatoes each (n=15) Means with same letter in the same row (abc) or in the same column (xyz) are not statistically different ( p <0.05) Figure 4 2 Average log 10 reduction of Salmonella after 25, 50 and 100 mg/L sodium hypochlorite a nd water overhead spray treatment Error bars represent standard deviation of triplicate experiments
61 Table 4 5 Average starting pH and measured concentration of sanitizers before and after sanitizer efficacy study experiments Sanitizer pH Before (mg/L) After (mg/L) NaOCl 6.48 94 97 ClO 2 7.50 4.48 4.82 PAA 3.62 77 76 Water 6.61 a a a Not tested
62 Table 4 6 Average l og 10 reduction of Salmonella after overhead spray treatment of sanitizers and water control Salmonella log reduction (log CFU/ml) from tomatoes a Treatment (s) NaOCl 100 mg/L ClO 2 5 mg/L PAA 80 mg/L Water control 5 1.94 0.59 b z 1.87 0.73 b z 2.79 0.94 a y 1.86 0.68 b y 15 3.49 1.06 b y 3.54 0.74 b y 4.73 0.53 a x 3.17 0.75 b x 30 4.07 1.22 b y 3.93 0.26 b y 5.50 0.12 a x 3.41 1.07 b x 60 5.51 0.17 a, x 4.85 0.26 a, x 5.52 0.12 a, x 3.75 0.74 b, x a Values are mean standard deviation of triplicate experiments of 5 tomatoes each (n=15). Means with same letter in the same ro w (ab) or in the same column (xyz) are not statistically different ( p <0.05). Figure 4 3 Average log 10 reduction of Salmonella after sanitizer and water control overhead spray treatment Error bars represent standard deviation of triplicate experiment s
63 Table 4 7 Average l og 10 reduction of Salmonella after sodium hypochlorite and water control flume treatment compared to sodium hypochlorite overhead spray treatment Salmonella log reduction (log CFU/ml) from tomatoes a Treatment (s) Overhead spray 1 00 mg/L NaOCl Flume 100 mg/L NaOCl Flume water control 5 1.66 0.51 a, z 0.79 0.91 a, y 0.46 0.27 a, x 15 3.98 1.78 a, y 1.25 1.14 b y 0.96 0.84 b, x 30 5.55 0.37 a, x 3.17 2.59 b, x 1.39 0.76 c, x 60 5.5 1 0.96 a, x 3.34 2.56 b, x 1.30 1.09 c, x a Values are mean standard deviation of triplicate experiments of 5 tomatoes each (n=15). Means with same letter in the same row (abc) or in the same column (xyz) are not statistically different ( p <0.05). Figure 4 4. Average log 10 reduction of Salmonella after 100 mg/L sodium hypochlorite and water flume treatment compared to 100 mg/L sodium hypochlorite overhead spray treatment Error bars represent standard deviation of triplicate experiments
64 Table 4 8 Average star ting pH and measured concentration of sanitizers before and after natural tomato microflora study experiments Sanitizer pH Before (mg/L) After (mg/L) NaOCl 6.50 100 102 ClO 2 7.55 5.09 4.80 PAA 3.58 81 84 Water 6.59 0 .00 a 0 .00 a a Values are concentrati on of measured free chlorine
65 Table 4 9 Average lo g 10 reduction of natural microflora after overhead spray treatment of sanitizers and water control Microflora log reduction (log CFU/ml) on tomatoes a Treatment (s) NaOCl 100 mg/L ClO 2 5 mg/L PAA 80 mg/ L Water control 5 0.50 0.36 a, y 0.58 0.26 a, x 0.48 0.20 a, y 0.56 0.38 a, xy 15 0.81 0.60 a, y 0.75 0.40 ab, x 0.74 0.72 ab, xy 0.26 0.27 b, y 30 1.41 0.90 a, x 0.55 0.20 b, x 0.84 0.72 b, xy 0.56 0.38 b, xy 60 1.5 8 0.46 a, x 1.06 0.46 b, x 1.19 0.28 ab, x 0.88 0.12 b, x a Values are mean standard deviation of triplicate experiments of 5 tomatoes each (n=15). Means with same letter in the same row (ab) or in the same column (xy) are not statistically d ifferent ( p <0.05). Figure 4 5. Average log 10 reduction of natural tomato microflora after sanitizer and water control overhead spray treatment Error bars represent standard deviation of triplicate experiments
66 CHAPTER 5 DISCUSSION AND CONCL USIO N Nume rous multistate outbreaks of Salmonella in tomatoes in past decades illustrate the need for effective food safety measures Risk of f oodborne disease outbreaks can be prevented or minimized on the farm and in the packinghouse by implementing GAPs and othe r intervention methods Flume washing in fresh tomato postharvest handling was targeted as a critical step in production where improper procedures could amplify potential contamination. Proper sanitizer use and monitoring of efficacy may reduce pathogen populations on tomatoes but more importantly, should prevent cross contamination from one tomato to another. While many studies have shown chlorinated flumes to achieve the 3 log 10 unit reduction of Salmonella required by Florida T GAPs and T BMPs, use of a non immersion sanitation system would decrease volum e of water and sanitizer needed and reduce costs. The goal of this research was to determine if a laboratory model overhead spray system would be as effective as a flume. E fficacy of s anitizers in th e overhead spray system was evaluated for the reduction of inoculated Salmonella from tomato surfaces. Growth Curves A five strain cocktail of rifampicin resistant Salmonella was used as the inoculum. The five serotypes used were S. Typhimurium S. Braend erup S. Enteritidis S. Newport and S. Javiana Salmonella serotypes were selected based on their association with multistate fresh tomato outbreaks and adapted to be rifampicin resistant in order to select for only the bacteria that was inoculated (Tabl e 2 1). T hough S. Enteritidis has not been the cause of a multistate fresh tomato outbreak in the US it is frequently the cause of outbreaks associated with other foods (CDC 2010 a ). A
67 growth curve study was meant to confirm that the five strains reached a stationary phase of about 9 log 10 CFU/ml at the same time. Therefore, all strains in a combined cocktail would initially be in equal concentrations and potentially be recovered after tomato inoculation The five strain cocktail represented a worst cas e scenario of Salmonella contamination that eliminated differences in overall Salmonella recovery due to differences between strain survival during sanitizer treatment. Growth of each strain was measured over 12 h (Table 4 1 and Figure 4 1) In a typica l bacterial growth curve, stationary phase follows a lag and logarithmic phase where growth stops, or rate cell division equals rate of cell death and concentration reaches a plateau. Within each strain, average concentrations during at least the last 3 h of growth were not statistically significant from each other but were significantly different from the previous hours of growth indicating that all strains reached stat ionary phase. Additionally, the same significance pattern was observed between strain s, indicating all strains reached stationary phase at the same time. Average concentration of the five strains at stationary phase (hour 12) was 9.07 0.11 log 10 CFU/ml. Because the selected strains exhibited very similar growth curves, they were used a s an inoculum cocktail in subsequent studies. This study was similar to one conducted by Felkey (2002) in which growth of five rifampicin resistant Salmonella strains were measured after 20 h. It was found that all strains grew to approximately the same concentration between 8.32 to 9.20 log 10 CFU/ml. The difference in concentration between studies could be because the incubation time was longer and different Salmonella strains were used.
68 Recovery Study A recovery study compare d the concentration of ino culum and the concentration of Salmonella recovered from inoculated tomatoes. Tomatoes were inoculated and dried for 2 h in a fume hood at room temperature. The 2 h drying period was meant to allow time for Salmonella to attach to tomato surfaces. Resul ts reveal ed an average difference of 0.73 0.24 log 10 CFU/ml Salmonella between inoculum and inoculated tomatoes (Table 4 2) The loss may have be en due to the 2 h drying period. After 2 h, inoculum wa s visually dry and appear ed like residue spots Fre shly inoculated tomatoes had wet spots. Though freshly inoculated tomatoes were not tested or statistically analyzed, a few single tomatoes were ino culated and immediately plated. These tomatoes produced similar colony counts to theoretical inoculation l evels. Additionally, t he fume hood could have increase d inoculum loss because the air flow increase d evaporation. Another possible explanation is that some Salmonella may have remained attached to the tomato surface during the rub and shake step of platin g. Because this research aim ed to determine the efficacy of various sanitizers in an overhead spray system, inoculated tomatoes needed to have a high enough initial level of Salmonella to show that sanitizers c ould achieve at least a 3 log 10 unit reducti on or even up to a 5 log 10 unit reduction The 3 log 10 unit reduction was a selected threshold of risk based on Florida T GAPs and T BMPs and represents a n estimated amount of contamina tion reasonably likely to occur (FDACS 2007). Any alternative sanitiz er or sanitation system must be validated for its ability to reduce Salmonella or like pathogens by a minimum of 3 log 10 units. The initial Salmonella level on tomatoes should also be high enough considering the limit of detection in the plating method wa s 2 log 10 CFU/ml. The refore, the purpose of the recovery study was to verify that any loss that occurred
69 was not too high to prevent the detection of Salmonella on treated tomatoes From the results, t he less than 1 log 10 unit loss was not considered too high because t omatoes still retained about 8 log 10 CFU/ml For that reason, all inoculated tomatoes were dried for 2 h in the fume hood and underwent the same plating procedure in subsequent studies. Cross Contamination Study The cross contamination, sod ium hypochlo rite efficacy, sanitizer efficacy, flume vs. overhead spray comparison and natural tomato microflora studies all examined the efficacy of a treatment for removing bacteria from tomato surfaces. Efficacy was measured by average log 10 reduction of bacteria from tomato surfaces. Any difference in average recovery between positive controls and treated tomatoes was considered an average reduction of bacteria attributed to the specific treatment applied. Accumulation of microorganisms c an occur in flume systems because tomatoes commingle and water is recirculated (Mahovic and others 2007) Cross contamination also presents a concern in an overhead spray system because tomatoes directly contact the brush rollers as a common surface. In the cross c ontamination study, NaOCl (100 mg/L) and water were tested for their ability to prevent cross contamination between inoculated tomatoes and uninoculated tomatoes via brush rollers in the overhead spray system. Inoculated tomatoes were treated for 15 s and removed, followed by 15 s treatment of uninoculated tomatoes. It was found that average log 10 reduction from inoculated tomatoes in the initial 15 s spray was not significantly different between NaOCl and water (p>0.05) Both NaOCl and water reduced Sal monella by about 3 log 10 units. As also seen in the sanitizer efficacy study discussed later water at times achieved similar efficacy as NaOCl and other sanitizers in the overhead spray
70 system depending on the treatment time. Generally, efficacy of wate r was comparable to other sanitizers at shorter treatment times less than 30 s when sanitizers may not have enough contact time on tomatoes It wa s proposed that the mechanical action of the brushes and pressure of the spray we re able to physically remove Salmonella from tomato surfaces to a certain extent. After a specific treatment time, the combined mechanical action of the overhead spray and antimicrobial power of sanitizers surpas se d mechanical action alone, resulting in higher efficacy of sanitizer s over water. Brush rollers were contaminated with an average of 1.60 0.70 log 10 CFU/cm 2 Salmonella after s praying inoculated tomatoes for 15 s (Table 4 3). There was no significant difference in recovery of Salmonella on brush rollers between NaOCl an d water (p>0.05). Cross contamination occurred at a n average of 4.88 0.41 log 10 CFU/ml on uninoculated tomatoes when sprayed with water. Spraying with NaOCl significantly reduced cross contamination to 2.63 0.28 log 10 CFU/ml (p<0.05). NaOCl was bett er able than water to prevent cross contamination from inoculated tomatoes to uninoculated tomatoes, but still resulted in greater than 2 log 10 unit transfer of Salmonella U ninoculated tomatoes had 15 s of contact time with rollers and contacted a larg e surface area that likely contributed in the transfer of Salmonella. It is believed than only one scientific study has examined contamination of brush rollers in an overhead spray system Pao and others (2009) found that w hen tomatoes were placed on brus hes that were previously inoculated with about 6.9 log 10 CFU/cm 3 Salmonella 5.7 log 10 CFU/cm 2 was transferred to tomato surfaces without any spray. A 5 mg/L ClO 2 spray significantly reduced contamination by 4.5 log 10 units after 10 s. Water reduced cont amination by 2.1 log 10 units only after 40 s. Even though ClO 2 was
71 significantly more effective than water at 10 s, there were still 1.2 log 10 units Salmonella recovered from tomato surfaces somewhat comparable to the 2.63 log 10 CFU/ml recovered in the c ross contamination study after 15 s NaOCl spray Increasing the ClO 2 spray to 60 s increased reduction by only 0.5 log 10 units, resulting in 0.7 log 10 units Salmonella remaining on tomatoes. Additional treatment times could be tested with the experimenta l protocol used to measure extent of cross contamination at longer spray times. Level of contamination from contaminated rollers to uninoculated tomatoes without any spray could also be tested. Though the methods were different, these results are similar to the Pao and others (2009) study. The researchers did not examine cross contamination but rather directly inoculated rollers with a known theoretical volume In this study the 3 log 10 CFU/ml of Salmonella removed from each inoculated tomato during th e 15 s spray w as hypothesized to ha ve transferred to brush rollers. T he overhead spray systems were also constructed differently with different sized brushes rotation speed and spray f low rate. The overall conclusion s are valuable however. U sing a san itizer instead of water in an overhead spray system effectively reduce s contamination of Salmonella from brush rollers to tomatoes Conversely, the overhead spray may not be more effective than chlorinated flumes at preventing cross contamination. Chlori nated water is known to be very effective at destroying microorganisms freely suspended in water At 2 min exposure, lowest lethal dose of chlorine shown to kill 7 log 10 units of E carotovora was 0.4 to 0.5 mg/L at pH 6 to 7 (Robbs and others 1995). Les s than 1 mg/L free chlorine has been shown to kill vegetative bacteria within 30 s ( Dychdala 2001). Geotrichum candidum mold has been inactivated in 20 to 25 mg/L free chlorine in 30 s (Bartz and
72 others 2001). While bacteria on tomato surfaces may only b e reduced by a few log 10 units in a flume, as seen in the flume vs. overhead spray comparison study discussed later, the bacteria become freely suspended and should be inactivated by properly sanitized water. Pao and others (2007) supported this conclusio n when they found that 5 mg/L ClO 2 effectively prevented cross contamination from inoculated tomatoes to uninoculated tomatoes placed in the same flume. Though about 2 log 10 units Salmonella w as transferred to uninoculated tomatoes in the overhead spray sy stem the extent of contamination that would likely occur outside a laboratory may be much lower. First, tomatoes, or brushes in the Pao and others (2009) study, were inoculated with a high concentration of Salmonella in order to be able to show a high lo g 10 reduction by specific treatments. The high concentration of inoculum also represents a worst case scenario of contamination and survival of at least one serovar of Salmonella after treatment Such a high level of contamination would likely only occur if GAP s were not properly followed. I mplement ation and adherence to GAPs should limit the extent of contamination that could occur in the field or packinghouse GAPs includ e using clean irrigation water, storing manure away and downhill of crops, practi cing good worker hygiene when handling fruit, using clean harvest bins and maintaining a validated effective concentration of sanitizer in wash water (FDA 1998) Sodium Hypochlorite Efficacy Study The sodium hypochlorite efficacy study examined efficacy of NaOCl (25, 50 and 100 mg/L) and water at reducing Salmonella on tomato surfaces in the overhead spray system. Significant reduction of Salmonella occurred at 15 s with NaOCl ( 100 mg/L ) (Table 4 4 and Figure 4 2). This time point had the highest relat ive standard deviation,
73 however. Average reduction by NaOCl (100 mg/L) at 15 s from 3 experiments ranged from 2.47 to 5.94 log 10 CFU/ml. Standard deviation was high likely because differences in morphology of individual tomatoes as illustrated by the fac t that r eduction on individual tomatoes ranged from 0.59 to 6.08 log 10 CFU/ml. Differences in reduction were likely due to treatment and not initial inoculum levels because positive control tomatoes were not significantly different. All tomatoes were pla ced in the same orientation on rollers with the inoculum area around the blossom scar adjacent to where tomatoes would initially contact the spray. If tomatoes remained in this orientation, they would spin on their axis in a way that the inoculated area w ould directly contact rollers and the spray. Throug hout the overhead spray studies, however, a few t omatoes were observed to roll on their axis in a direction that did not result in contact between the inoculated area and the brush rollers and/or direct c ontact with spray due to differences in tomato morphology, including surface characteristics, size and force from the characteristics of neighboring tomatoes. The inoculum on these tomatoes would not be subjected to as much physical removal or exposure to sanitizer and thus achieve a lower reduction. Including more replicates should lower standard deviation, but since statistical significance was observed, the n=15 sample size was thought to be sufficient. Little research examin ing efficacy of NaOCl in an overhead spray system is available Vigneault and others (2000) examined a chlorinated overhead spray shower system as a tomato hydrocooler. Because flume systems result in immersion of tomatoes, infiltration of water and pathogens into tomatoes is a co ncern, especially via the stem scar (Bartz and Showalter 1981) Potential for infiltration in t he hydrocooler shower system was investigated T omatoes were placed stem scar up or down and
74 sprayed with water contaminated with R. stolonifer and up to 200 m g/L chlorine. Because tomatoes were heated to 35 C and water was 10 C, all tomatoes increased in weight by the uptake of water, but tomatoes in the stem scar up orientation gained significantly more weight. Despite the water infiltration, c hlorine preven ted decay of all tomatoes after 10 d storage at 20 C (Vigneault and others 2000). This paper showed that there is likely little risk of water infiltration via tomato stem scars in overhead spray systems. The hydrocooler water flow rate was 1,000 L/min*m 2 and the exposure time was over 13 min, which are both much higher than the treatment parameters of the overhead spray system of this research. Therefore, lower flow rates and shorter contact times may present an even smaller risk of infiltration. In summ ary, t his study showed that all NaOCl concentrations achieved a 3 log 10 CFU/ml reduction or more depending on treatment time. The data of this study will be able to support the use of the specific concentration and treatment time combinations in an overhe ad spray system in accordance with T GAPs and T BMPs T BMPs require sanitation of tomatoes with an approved sanitizer or process. If not pre approved, the sanitizer or process must be shown in a reproducible scientific study to achieve at least a 3 log 1 0 unit reduction of Salmonella or like organisms (FDACS 2007). Beyond this requirement, a 5 log 10 CFU/ml reduction was achieved by 100 mg/L NaOCl at 30 s. Because T BMPs require flumes to have at least 150 mg/L free chlorine and flumes may be chlorinated up to 350 mg/L in practice to maintain sufficient free chlorine implementing an overhead spray system with just 100 mg/L NaOCl would reduce the amount of NaOCl currently used (FDACS 2007; Suslow 1997). A lower NaOCl concentration would reduce packinghou se operating costs.
75 Sanitizer Efficacy Study The sanitizer efficacy study examined efficacy of NaOCl (100 mg/L) ClO 2 (5 mg/L) PAA (80 mg/L) and water at reducing Salmonella on tomato surfaces in the overhead spray system. The overall conclusion gath ered from results was that all sanitizers achieved a 3 log 10 reduction at 15 s (Table 4 6 and Figure 4 3). Similar to the sodium hypochlorite efficacy study, NaOCl (100 mg/L) achieved a 3.49 1.06 log 10 CFU/ml reduction of Salmonella a t 15 s. Efficacy o f NaOCl at 30 s was 4.07 1.22 log 10 CFU/ml compared to 5.55 0.37 log 10 CFU/ml in the sodium hypochlorite efficacy study. Difference in efficacy may be explained by different tomato spins or unaccountable differences between experiments since average r eduction at 30 s per experiment ranged from 2.72 to 5.10 log 10 CFU/ml. Conducting additional experiment replicates should decrease standard deviation values Still, t his data can support the use of 100 mg/L NaOCl in overhead sp ray systems Pao and othe rs (2009) examined ClO 2 in an overhead spray system with inoculated tomatoes dried for 24 h at 40 to 50% relative humidity. Tomatoes were sprayed with 5 mg/L ClO 2 or water at a flow rate of 5.0 or 9.3 ml/s It was found that ClO 2 achieved a 4.4 to 5.2 l og 10 unit reduction of Salmonella from 10 to 60 s exposure, which was significantly better than water (p<0.01). The reductions were comparable to the 3.54 to 4.85 log 10 CFU/ml reductions from 15 to 60 s exposure seen in this study. Flow rate did not sign ificantly change efficacy. While PAA has not been previously studied in an overhead spray system, it has been studied in flumes. PAA has been shown to achieve a 4 log 10 unit reduction of Salmonella on surfaces of various produce including tomatoes when t reated for at least 60 s (Yuk and others 2005 2006). PAA in this study achieved greater than 4 log 10 CFU/ml reductions in just 15 s in the overhead
76 spray system. While ClO 2 and PAA are already approved in Florida for use in flumes, the results of this s tudy can support their use in an overhead spray system A s also seen in the cross contamination study discussed earlier, water sometimes achieved similar efficacy to the other sanitizers in the overhead spray system. T he physical removal of Salmonella f rom tomatoes by brush rollers and spray pressure must play an important role in efficacy in an overhead spray system. The efficacy of the brushes or water spray alone could be tested as a control in a future study. Brushing is known to aid sanitation eff orts because of physical removal of microbes. For example, Parnell and others (2005) inoculated rinds of whole melons with S Typhimurium. Melons were treated with a 200 mg/L total chlorine or distilled water soak, or wet with one of the above solutions and then scrubbed with a sterile brush for 60 s. Soaking resulted in about 1 and 2 log 10 CFU/sample reductions of Salmonella on melons in water and chlorine, respectively. Scrubbing with a brush significantly reduced Salmonella by about an additional 1 l og 10 CFU/sample. Overall, scrubbing in chlorine provided the highest reduction (Parnell and others 2005). I n contrast to the sodium hypochlorite efficacy study where water could not reach a 3 log 10 CFU/ml reduction a t 60 s, water achieved a 3 log 10 CFU/ ml reductio n at only 15 s in the sanitizer efficacy study. The discrepancy could be due to natural variation of tomatoes that cause d changes in spin direction because standard deviation is relatively high especially for 30 s Though sanitizer treatments were run sequentially on the same day with water controls run last it is unlikely that residual sanitizer remaining on rollers, if any, contributed to water efficacy because water was supplemented with sodium thiosulfate which inactivates chlorine. Wat er was also sprayed for about 1 min
77 to rinse lines and rollers after the PAA tests. Pao and others (2009) also observed a 3 log 10 unit reduction of Salmonella from tomato surfaces from a 5.0 ml/s water spray for 10 to 40 s. Reduction increased to 4.4 log 10 units at 60 s. Increasing flow rate to 9.3 ml/s generally increased reduction by water. This could be explained by the increased pressure of the spray being able to physically force more Salmonella from tomato surfaces. Flume vs. Overhead Spray Comp arison Study E fficacy of NaOCl (100 mg/L) and water against Salmonella on tomatoes were compared in a scale model flume. NaOCl flume data was also compared to NaOCl (100 mg/L) data from the sodium hypochlorite efficacy study. Results showed that t he over head spray system was significantly more effective than the chlorinated flume between 15 to 60 s contact time (Table 4 7 and Figure 4 4). Increasing treatment time from 30 s to 60 s did not significantly affect reduction by NaOCl in either system. The c hlorinated scale model flume (100 mg/L) d id achieve a 3 log 10 unit re duction at 30 s which supports the current use of flumes as a sanitation system in tomato packinghouses. Additionally, the c hlorinated flume was effective compared to the unchlorinated flume The flume water control achieved a maximum reduction of 1.30 1.09 log 10 CFU/ml at 60 s. The water control data shows that water was able to remove some Salmonella from tomato surfaces, perhaps from the mechanical action of the circulating water that force d tomatoes to flow, bump and rub surrounding tomatoes and sometimes roll over. Reduction by mechanical action was limited to an average of 1.03 0.74 log 10 CFU/ml, however. Rolling over so that the inoculated area on tomatoes was no longer sub merged may have lowered overall reduction. This phenomenon caused by chance or natural variation in tomatoes would help explain the
78 relatively high standard deviation of average reduction values throughout the comparison study. Flume water can cross con taminate commingled produce if water is not properly chlorinated. In this study, the a ddition of 100 mg/L NaOCl in the flume not only significantly increase d reduction of Salmonella compared to a plain water flume, but could also prevent cross contaminati on by eliminating freely suspended Salmonella Salmonella was recovered in unchlorinated water at 4.54 0.37 and 5.05 0.16 log 10 CFU/ml after inoculated tomatoes were tested for 15 and 60 s, respectively. The freely suspended Salmonella could contamin ate subsequent tomatoes placed in the flume. In contrast, Salmonella was undetect ed in the 100 mg/L NaOCl flume. Studies that examined the efficacy of chlorinated flumes report good levels of efficacy. Felkey and others (2006) observed about a 3 log 10 CFU/ml reduction of Salmonella in a 150 mg/L free chlorine flume at 25 C for 30 to 60 s, which is very similar to these results despite the higher chlorine concentration Salmonella population s were also reduced to 0.16 log 10 CFU/ml after treatment for 12 0 s, which was essentially a greater than 6 log 10 CFU/ml reduction. The maximum treatment time tested in this study was 60 s thus it is unknown whether similar results would be seen in the flume In a study by Zhuang and others ( 1995 ) entire surfaces o f tomatoes were inoculated with S Montevideo to about 4.81 log 10 CFU/cm 2 and were dr ied for 5 h. After submerging tomatoes in 60 and 110 mg/L NaOCl at 25 C for 2 min, populations were reduced to 4.17 and 3.59 log 10 CFU/cm 2 respectively. These levels we re significantly different from the control, equal to a 0.64 to 1.22 log 10 CFU/cm 2 reduction. Yuk and others (2005) found a 5 log 10 unit reduction of Salmonella on tomato surfaces at
79 200 mg/L NaOCl after 60 s at 35 C The higher concentration of NaOCl us ed and longer treatment times could explain the higher log 10 reductions achieved in these studies. Additionally, the different Salmonella serovars tested may have varying levels of survival during treatments. Felkey and others (2006) and Yuk and others ( 2005) both used a 5 strain cocktail of S. Agona, S Gaminara, S. Michigan, S. Montevideo and S Poona. In conclusion, while the chlorinated flume was able to achieve a 3 log 10 CFU/ml reduction in 30 s, the overhead spray system achieved a greater reduction in half the time Maximum reduction by the NaOCl flume was about 3 log 10 CFU/ml at 60 s whereas the overhead spray system reached a greater than 5 log 10 CFU/ml reduction in 30 s. The longer immersion times needed by the flume to achieve similar efficacy as the overhead spray system may increase risk of wate r infiltration. T his study demonstrates the superior ability of the overhead spray system to reduce Salmonella from tomato surfaces under laboratory conditions. Natural Tomato Microflora Study Efficac y of sanitizers was evaluated in the overhead spray system for the reduction of natural tomato microflora. The goal of this study was to determine how the overhead spray system would perform against non artificially inoculated bacteria. Many naturally oc curring organisms are actually, once internalized, pathogens to the tomato itself (Narayanasamy 2006). The uninoculated microflora represent ed both indigenous innocuous organisms an d potential spoilage organisms including fungi and bacteria that can affe ct tomato quality Generally, average log 10 reductions of natural microflora were lower than those seen in the efficacy studies with inoculated Salmonella (Table 4 9 and Figu re 4 5). The
80 low reductions of natural microflora by the sanitizers may be expl ained by the strong attachment of these organisms to the tomato surface that occurred over time. I n contrast, i n the other efficacy studies, Salmonella was inoculated and allowed to dry for only 2 h before testing. The mechanical action of the brush roll ers and pressure of the spray may not have been forceful enough to physically remove the microflora from tomatoes. Because bacteria remained attached, perhaps in a protective biofilm, they were less susceptible to the antimicrobial effects of the sanitize rs. This would explain why ClO 2 and PAA were not more effective than water in this study. A future study could examine the effect of adjusting the engineering of the overhead spray system to achieve more physical force to remove microflora yet still not injure the tomatoes Another study could examine the fate of non mesophilic bacteria in the overhead spray system, though they may be less likely to cause human disease. Because n atural tomato microflora usually do not affect safety of tomatoes efficacy the overhead spray system as it currently stands should be measured by pathogen reduction. A quality issue may arise if tomatoes are mechanically injured so that natural microflora can be internalized and spread disease throughout an entire box Opport unistic tomato spoilage fungi include Fusarium spp., Geotrichum spp. and yeasts ( Narayanasamy 2006). Naturally occurring bacteria found on and inside tomatoes include Bacillus spp., Cyanobacterium spp., Erwinia spp., Enterobacter spp., Pantoea spp. and Ps eudomonas putida (Shi and others 2009). While some of these genera contain human pathogenic species, most are nonpathogenic. Furthermore, some studies have shown Enterobacter and Bacillus to inhibit the growth of Salmonella in tomatoes and other plants a nd animals through competitive ex c lusion (Shi and others
81 2009). If natural microflora remain attached to tomato surfaces, Salmonella may not have available space or nutrients to survive. The presence of natural microflora of tomato surfaces could be a pr evention factor against Salmonella contamination. Conclusions and Future Work The overall objective of this research was to determine if a non immersion sanitation system could be as effective as a flume system and therefore decrease water and sanitizer re quirements. A laboratory model overhead spray system was evaluated for the reduction of inoculated Salmonella on tomato surfaces. Because sanitation of tomatoes is required in Florida as part of T GAPs and T BMPs, novel sanitation systems are researched and developed to improve sanitation and to reduce operating costs. S anitation system s must be evaluated for benefits and disadvantages. Flumes and overhead spray systems can be compared in terms of sanitation ability, sanitizer use, water use, ability to prevent cross contamination, required space, non sanitation uses and overall cost. Several studies were performed in this research to evaluate the sanitizer and treatment time combinations needed in the overhead spray system to achieve at least a 3 log 1 0 unit reduction of Salmonella S anitizers in the overhead spray system achieved higher log 10 reduction s of Salmonella compared to the scale model flume at treatment times of at least 15 s. Three log 10 unit reductions were achieved by all sanitizers and concentrations of NaOCl in the overhead spray system. Th e data collected can support the implementation of an overhead spray system because of the scientifically observed pathogen reduction F ew studies have examine d efficacy of sanitizers in an overhead spray system. One study compared a commercial overhead spray and dump tank method with an
82 alternative spray wash method in a cantaloupe packing facility in Mexico (Alvarado Casillas and others 2010). The typical sanitation method was a potable water spra y over polyvinyl chloride rollers followed by a chlorinated dump tank soak for 1 min. An alternative treatment was a 90 s manual water wash with a backpack sprayer followed by a 2% lactic acid spray for 15 s, both at about 17 m/s flow rate. Results showe d that the alternative spray method was significantly more effective against aerobic plate counts and coliforms than the dump tank (Alvarado Casillas and others 2010) Researchers suggest spray washing can help reduce cross contamination, though it is unl ikely that tomatoes can be manually spray washed given the large volumes entering packinghouses. NaOCl was the main sanitizer tested because of its established use in the produce industry. NaOCl concentrations tested were 25, 50 and 100 mg/L, which were s ubstantially lower than the minimum required and typically measured free chlorine levels in packinghouse flumes (FDACS 2007; Suslow 1997). Results of the sodium hypochlorite study could support the use of less NaOCl to achieve a similar or higher level of Salmonella reduction on tomato surfaces. The ability of 100 mg/L NaOCl to achieve a greater than 3 log 10 reduction of Salmonella in 15 s in the overhead spray system shows that NaOCl usage could be drastically reduced with use of overhead spray systems. NaOCl volume would also be reduced with the decreased amount of water needed for sanitation. F lumes hold tens of thousands of gallons of water that cannot be reused from day to day W ater must be hauled away by tanker trucks. Though water still cannot be recycled, overhead spray systems could possibly achieve a higher level of pathogen reduction with less water. Exact volume of water needed
83 would depend on each operation such as production speed and quantity of tomatoes produced a day Because wash wa ter is not reused in overhead spray systems, there is limited contact between chlorine and nitrogenous and organic compounds. Therefore, there would be reduced formation of chloramines and carcinogenic trihalomethanes The use of less water is both impor tant in reducing costs and environment impact While there may be no drawbacks in saving water, using a lower concentration of NaOCl has implications for the survival of other bacteria besides Salmonella While Salmonella is an important human pathogen a ssociated with tomatoes, optimization of the overhead spray system would require investigating its efficacy against decay organisms For example, mold spores have been shown to require 135 to 500 mg/L NaOCl for inactivation (Dychdala 2001). The small amount of cross contamination that occur red in the overhead spray system is its potential disadvantage compared to flumes. Additional experiments should be conducted to further examine the extent of cross contamination in the overhead spray system. In th is research, inoculated tomatoes were sprayed on rollers for 15 s followed by a 15 s spray on uninoculated tomatoes. Though 100 mg/L NaOCl significantly reduced cross contamination to uninoculated tomatoes compared to water, it did not prevent the transfe r of 2.63 0.28 log 10 CFU/ml Salmonella on to tomatoes. This level of contamination could represent the worst extent of contamination per tomato. A future study could examine the extent of cross contaminat ion to sets of subsequent tomatoes treated by ove rhead spray As replicates of uninoculated tomato sets contact rollers and are removed, the level of contamination per tomato set is hypothesized to decrease. This study would measure how long cross contamination
84 would persist in the overhead spray syste m in a worst case scenario of initial Salmonella contamination. Cross contamination of bacterial and fungal fruit pathogens are also of concern in terms of tomato quality. The same method used to examine cross contamination of Salmonella could be used to examine the fate of inoculated postharvest decay organisms like E carotovora and Geotrichum spp. on tomato es in the overhead spray system. In terms of required space inside a packinghouse, a brush roller bed engineered to achieve a 15 s tomato contact t ime may still be more compact than a flume. Flumes have other uses besides sanitation however, which would likely hinder the complete removal of flumes in packinghouses. The initial section of a flume is referred to as a dump tank because tomatoes from the field are first dumped into the water to cushion their fall and prevent injury The flume acts as a conveyor through the processing line. An overhead spray system c ould potentially be implemented after the initial dump tank and still conserve water. The water used in the overhead spray could be routed to the dump tank and save even more water. Scale up inevitably presents a challenge in terms of engineering and transfer of laboratory findings to commercial operations. Validation studies could be p erformed in individual operations that implement a full scale overhead spray system to ensure efficacy is reached. Because overhead spray systems can be constructed differently, it may be necessary to standardize systems to reproduce similar sanitizer eff icacy results. Overhead spray systems may need to be engineered in a way to optimize average pathogen reduction on tomato surfaces. Parameters that may affect efficacy include spray flow rate, water pressure, height of spray nozzles, number of nozzles, m aterial of
85 brush rollers and brush roller speed. A problem encountered in this research was the natural variation of tomatoes that caused them to spin in different directions. It was observed that if the inoculated spot on the tomato did not directly con tact the brush rollers and/or spray, it sometimes achieved a lower log 10 reduction. Outside a laboratory, tomatoes could become contaminated at a single spot or on their entire surface. Overhead spray systems may need to be engineered to maximize the sur face area of tomatoes that directly contact brush rollers and spray by forcing tomatoes to continuously spin in multiple directions so that any potential contamination is exposed. Similar efficacy studies could be performed with dip inoculated tomatoes. In conclusion, c ollected data show ed that the overhead spray system achieve d a 3 to 5 log 10 unit reduction of Salmonella from tomato surfaces under specific sanitizer and treatment times. An overhead spray system could provide benefits over conventional flumes including higher pathogen reduction, less sanitizer and less water all of which help to decrease tomato packing costs and keep the tomato industry a Additionally, because an overhead spray system does not require immersion of tomatoes in water, there may be no need to heat water to at least 5 C greater than pulp temperature of tomatoes. Using ambient temperature water would save on heating and energy costs. Improving sanitation as part of an effective food safet y program can minimize risk of contamination and potential of causing a foodborne disease outbreak. This research is only one of a few studies that have examined efficacy of sanitizer s in an overhead spray system and ha s the potential to directly influenc e current industry practices by supporting the implementation of overhead spray systems for tomatoes
86 LIST OF REFERENCES Adaskaveg JE, F rster H, Sommer NF. 2002. Principles of postharvest pathology and management of decays of edible horticultural crops In: Kader AA, editor Postharvest t echnology of h orticultural c rops. 3rd ed. Oakland, CA: University of California Agriculture and Natural Resources. p 163 9 5 Alvarado Casillas S, Ibarra Snchez LS, Martnez Gonzles NE, Rodrguez Garca MO, Castillo A 2010. Validation of a washing and sanitizing procedure for cantaloupes at a M exican packing facility. J Food Prot 73:362 5. Barton Behravesh C, Mody RK, Jungk J, Gaul L, Redd JT, Chen S, Cosgrove S, Hedican E, Sweat D, Chvez Hauser L, Snow SL, Hanson H, Nguyen TA, Sodha SV, Boore AL, Russo E, Mikoleit M, Theobald L, Gerner Smidt P, Hoekstra RM, Angulo FJ, Swerdlow DL, Tauxe RV, Griffin PM, Williams IT 2011. 2008 outbreak of Salmonella Saintpaul infections associated with raw produce. N Engl J Med 10:918 27 Bartz JA, Eayre CG, Mahovic MJ, Concelmo DE, Brecht JK, Sargent SA. 2001. Chlorine concentration and the inoculation of tomato fruit in packinghouse dump tanks. Plant Dis 85:885 9. Bartz JA, Showalter RK. 1981. Infiltration of tomatoes by aqueous bact erial suspensions. Phytopathology 71:515 8. B euchat LR. 1996. Pathogenic microorganisms associated with fresh produce. J Food Prot 59:204 16. Beuchat L. 1998. Surface decontamination of fruits and vegetables eaten raw: A review. World Health Organization, Food Safety Unit. Cantwell MI, Kasmire RF. 2002. Postharvest handling systems: Fruit vegetables. In: Kader AA, editor Postharvest t echnology of h orticultural c rops. 3rd ed. Oakland, CA: University of California Agriculture and Natural Resources. p 407 21 [CAST] Council for Agricultural Science and Technology. 1994. Foodborne pathogens: risks and consequences. Task Force Report No. 122. Ames, IA: Council for Agricultural Science and Technology. [CDC] Centers for Dise ase Control and Prevention 2002 a Outb reak of Salmonella serotype Javiana infections Orlando, Florida, June 2002. MMWR Weekly 51 : 683 4 [CDC] Centers for Dise ase Control and Prevention 2002 b Multistate outbreaks of Salmonella serotype Poona infections associated with eating cantaloupe from M exico -United States and Canada, 2 000 2002. MMWR Weekly 51 :1044 7.
87 [CDC] Centers for Dise ase Control and Prevention 2 005 Outbreaks of Salmonella infections associated with eating Roma tomatoes United States and Canada 2004 MMWR Weekly 5 4 : 325 8 [CDC] C enters for Dise ase Control and Prevention 2 007 Multistate outbreaks of Salmonella infections associated with raw tomatoes eaten in restaurants United States, 2005 2006 MMWR Weekly 5 6 : 909 11 [CDC] Centers for Dise ase Control and Prevention 2 008 a Outbr eak of Salmonella serotype Saintpaul infections associated with multiple raw produce items United States, 2008 MMWR Weekly 5 7 : 929 34 [CDC] Centers for Dise ase Control and Prevention 2 008b Investigation of o utbreak of i nfections c aused by Salmonella Lit chfield Atlanta, GA: US Department of Health and Human Services, CDC. Available from: http://www.cdc.gov/salmonella/litchfield/ Accessed May 5 2011. [CDC] Centers for Dise ase Control and Prevent ion 2 010 a Surveillance for foodborne disease outbreaks, United States, 2007 MMWR Weekly 5 9 : 973 9 [CDC] Centers for Dise ase Control and Prevention 2 010b Investigation u pdate: M ultistate o utbreak of h uman Salmonella Enteritidis i nfections a ssociated w ith s hell e ggs. Atlanta, GA: US Department of Health and Human Services, CDC. Available from: http://www.cdc.gov/salmonella/enteritidis/. Accessed May 5 2011. [CDC] Centers for Dise ase Contro l and Prevention 2 010c OutbreakNet: Foodborne Outbreak Online Database. Atlanta, GA: US Department of Health and Human Services, CDC. Available from: http://wwwn.cdc.gov/foodborneoutbreaks/ Access ed May 5 2011. [CDC] Centers for Dise ase Control and Prevention 2 011 Investigation update: Multistate o utbreak of Salmonella Panama i nfections l inked to c antaloupe Atlanta, GA: US Department of Health and Human Services, CDC. Available from: http://www.cdc.gov/salmonella/panama0311/032911/index.html Accessed May 5 2011. [CFR] Code of Federal Regulations 20 10 a. Current good manufacturing practice in manufacturing, packing, or holding human food. Title 21, Part 110. Washington, D.C.: US Food and Drug Administration, Office of the Federal Register. [CFR] Code of Federal Reg ulations 20 10 b. Indirect food additives: adjuvants, production aids, and sanitizers: s anitizing solutions Title 21, Part 17 8.101 0. Washington, D.C.: US Food and Drug Administration, Office of the Federal Register. [CFR] Code of Federal Reg ulations 20 10 c Secondary direct food additives permitted in food for human consumption: c hlorine dioxide. Title 21, Part 173.300. Washington, D.C.: US Food and Drug Administration, Office of the Federal Register.
88 [CFR] Code of Federa l Regulations 20 10 d Secondary direct food additives permitted in food for human consumption: c hemicals used in washing or to assist in the pee ling of fruits and vegetables. Title 21, Part 173.315. Washington, D.C.: US Food and Drug Administration, Office of the Federal Register. Cummings K, Barrett E, Mohle Boetani JC, Brooks JT, Farrar J, Hunt T, Fiore A, Komatsu K, Werner SB, Slutsker L. 2001. A multistate outbreak of Salmonella enterica serotype Baildon associated with domestic raw tomatoes. Emerg Infect Dis 7:1046 8. J Y, Maurer J 200 7. Salmonella s pecies. I n : Doyle MP, Beuchat LR, editors Food m icrobiology: f undamentals and f rontie rs. 3rd ed. Washington, D.C.: ASM Press. p 187 236. Dychdala GR. 2001. Chlorine and chlorine compounds. In: Block SS, editor. Disinfection, sterilization, and preservation. 5 th ed. Philadelphia, PA: Lippincott Williams & Wilkins. p 135 58. [EPA] Environmen tal Protection Agency 2006. Registration eligibility decision (RED) for chlorine dioxide and sodium chlorite (case 4023). Washington, D.C.: EPA. Available from : http://www.epa.gov/o ppsrrd1/REDs/chlorine_dioxide_red.pdf Accessed May 5 2011 [FAC] Florida Administrative Code. 2007. Rule 5G 6. Tomato inspection. Tallahassee, FL: Florida Department of State, FAC. Available from: https://www.flrules.org/gateway/chapterhome.as p?chapter=5g 6 Accessed May 5 2011. Fatica MK, Schneider KR. 2009. The use of chlorination and alternative sanitizers in the produce industry. CAB Rev 4 :1 10. [FDA] Food and Drug Adminis tration 1998. Guide to minimize microbial food safety hazards for fresh fruits and vegetables. Washington, D.C.: US Department of Health and Human Services, FDA. Available from: http://www.fda.gov/Food/GuidanceComplianceRegulatoryInformation/GuidanceDo cuments/ProduceandPlanProducts/UCM064574 Accessed May 5 2011 [FDA] Food and Drug Administration 200 1 Analysis and e valuati on of p reventive c ontrol m easures for the c ontrol and r eduction/ e limination of m icrobial h azards on f resh and f resh c ut p roduce Washington, D.C.: US Department of Health and Human Services, FDA. Available from: http://www.fda.gov/Food/ScienceResearch/ResearchAreas/SafePracticesforFood Processes/ucm090977.htm Accessed May 5 2011
89 [FDA] Food and Drug Administration 2009 Guide to minimize micro bial food safety hazards of tomatoes; draft guidance Washington, D.C.: US Department of Health and Human Services, FDA. Available from: http://www.fda.gov/Food/GuidanceComplianceRegulatoryInformation/GuidanceDo cuments/ProduceandPlanProducts/ucm173902.htm Accessed May 5 2011 [FDACS] Florida Department of Agricultu re and Consumer Services 2007. Tomato b est p ractices m anu al. Tallahassee, FL: Florida Department of Agriculture and Consumer Services, Division of Food Safety. Available from: http://www.doacs.state.fl.us/fs/TomatoBestPractices.pdf Accesse d May 5 2011 Felkey KD. 2002. Optimization of chlorine treatments and the effects on survival of Salmonella spp. on tomato surfaces [MSci thesis]. Gainesville, F L .: Univ. of Florida Felkey K, Archer DL, Bartz JA, Goodrich RM, Schneider KR. 2006. Chlor ine disinfection of tomato surface wounds contaminated with Salmonella spp. Hort Technol 16:253 6. Gil MI, Selma MV, Lpez Glvez F, Allende A. 2009. Fresh cut product sanitation and wash water disinfection: p roblems and solutions. Int J Food Microbiol 134 :37 45. Gomez Lopez VM, Rajkovic A, Ragaert P, Smigic N, Devlieghere F. 2009. Chlorine dioxide for minimally processed produce preservation: A review. Trends Food Sci Tech nol 20: 17 26. Grant J, Wendelboe AM, Wendel A, Jepson B, Torres P, Smelser C, Rolfs R T. 2008. Spinach associated Escherichia coli O157:H7 outbreak, Utah and New Mexico, 2006. Emerg Infect Dis 14:1633 6. Greene SK, Daly ER, Talbot EA, Demma LJ, Holzbauer S, Patel NJ, Hill TA, Walderhaug MO, Hoekstra RM, Lynch MF, Painter JA. Recurrent multi state outbreak of Salmonella Newport associated with tomatoes from contaminated fields, 2005. Epidemiol Infect 136:157 65. Guo X, Chen J, Brackett RE, Beuchat LR. 2001. Survival of salmonellae on and in tomato plants from the time of inoculation at floweri ng and early stages of fruit development through fruit ripening. Appl Environ Microbiol 67:4760 4. Gupta SK, Nalluswami K, Snider C, Perch M, Balasegaram M, Burmeister D, Lockett J, Sandt C, Hoekstra RM, Montgomery S. 2007. Outbreak of Salmonella Braender up infections associated with Roma tomatoes, northeastern United States, 2004: A useful method for subtyping exposures in field investigations. Epidemiol Infect 135:1165 73. Harris LJ, Zagory D, Gorny JR. 2002. Safety factors. In: Kader AA, editor Posthar vest t echnology of h orticultural c rops 3rd ed. Oakland, CA: University of California Agriculture and Natural Resources. p 301 14.
90 Hedberg CW, Angulo FJ, White KE, Langkop CW, Schell WL, Stobierski MG, Schuchat A, Besser JM, Dietrich S, Helsel L, Griffin PM, McFarland JW, Osterholm MT. 1999. Outbreaks of salmonellosis associated with eating uncooked tomatoes: I mplications for public health. Epidemiol Infect 122:385 93. Kemp GK, Schneider KR. 2000. Validation of thiosulfate for neutralization of acidified sodium chlorite in microbiological testing. Poultry Sci 79: 185 7 60. Lopez Galvez F, Allende A, Truchado P, Martinez Sanchez A, Tudela JA, Selma MV, Gil MI 2010. Suitability of aqueous chlorine dioxide versus sodium hypochlorite as an effective sanitizer f or preserving quality of fresh cut lettuce while avoiding by product formation Postharvest Biol Technol 55:53 60. Mahovic MJ, Bartz JA, Schneider KR. 2007. Controlling biotic factors that cause postharvest losses of fresh market tomatoes. Hort Rev 33:351 91. Mari M, Bertolini P, Pratella GC. 2003. Non conventional methods for the control of post harvest pear diseases. J Appl Microbiol 94:761 6. McDonnell G, Russell AD. 1999. Antiseptics and disinfectants: activity, action, and resistance. Clin Microbiol Re v 12 : 147 79. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, Griffin PM, Tauxe RV. 1999. Food related illness and death in the United States. Emerg Infect Dis 5:607 25. Mohle Boetani JC, Reporter R, Werner SB, Abbot A, Farrar J, Waterman SH, Vugia DJ. 1999. An outbreak of Salmonella serogroup Saphra due to cantaloupes from Mexico. J Infect Dis 180:1361 4. Narayanasamy P. 2006. Ecology of postharvest microbial pathogens. I n : Narayanasamy P. Postharvest pathogens and disease management. Hoboken NJ: John Wiley & Sons, Inc. p 79 116. Pao S, Kelsey DF, Khalid MF, Ettinger MR. 2007. Using aqueous chlorine dioxide to prevent contamination of tomatoes with Salmonella enterica and Erwinia carotovora during fruit washing. J Food Prot 70 : 629 34. Pao S, Kelsey DF, Long III W. 2009. Spray washing of tomatoes with chlorine dioxide to minimize Salmonella on inoculated fruit surfaces and cross contamination from revolving brushes. J Food Prot 72: 2448 52. Parnell TL, Harris LJ, Suslow TV. 2005. Reducing Salmon ella on cantaloupes and honeydew melons using wash practices applicable to postharvest handling, foodservice, and consumer preparation. Int J Food Microbiol 99:59 70. [PL] Public Law. 2011. FDA Food Safety Modernization Act. Pub L no. 111 353, 124 Stat 388 5.
91 Quiroz Santiago C Rodas Suarez OR Vazquez CR Fernandez FJ Quinones Ramirez EI Vazquez Salinas C Prevalence of Salmonella in vegetables from Mexico. J Food Prot 72:1279 82. Rangel JM, Sparling PH, Crowe C, Griffin PM, Swerdlow DL. 2005. Epidemiolog y of Escherichia coli O157:H7 outbreaks, United States, 1982 2002. Emerg Infect Dis 11:603 9. Robbs PG, Bartz JA, Brecht JK, Sargent SA. 1995. Oxidation reduction potential of chlorine solutions and their toxicity to Erwinia carotovora subsp. carotovora an d Geotrichum candidum Plant Dis 79:158 62. Samuel MC, Vugia DJ, Koehler KM, Marcus R, Deneen V, Damaske B, Shiferaw B, Hadler J, Henao OL, Angulo FJ 2007. Consumption of risky foods among adults at high risk for severe foodborne diseases: Room for improv ed targeted prevention messages. J Food Safety 27:219 32. Santamaria J, Toranzos GA. 2003. Enteric pathogens and soil: A short review. Int Microbiol 6:5 9. Sawyer CN, McCarty PL, Parkin GF. 2003. Residual chlorine and chlorine demand. In: Sawyer CN, McCart y PL, Parkin GF. Chemistry for environmental engineering and science. 5 th ed. Boston, MA: McGraw Hill. p 571 86. Scallan E, Hoekstra RM, Angulo FJ, Tauxe RV, Widdowson M A, Roy SL, Jones JL, Griffin PM. 2011 a Foodborne illness acquired in the United State s major pathogens. Emerg Infect Dis 17:7 15. Scallan E, Griffin PM, Angulo FJ, Tauxe RV, Hoekstra RM. 2011b. Foodborne illness acquired in the United States unspecified agents. Emerg Infect Dis 17:16 22. Shi X, Wu Z, Namvar A, Kostrzynska M, Dunfield K, Wa rriner K. 2009. Microbial population profiles of the microflora associated with pre and postharvest tomatoes contaminated with Salmonella Typhimurium or Salmonella Montevideo. J Appl Microbiol 107:329 38. Shirron N Kisluk G, Zelikovich Y Eivin I Shimon i E, Yaron S 2009. A comparative study assaying commonly used sanitizers for antimicrobial activity against indicator bacteria and a Salmonella Typhimurium strain on fresh produce. J Food Prot 72:2413 7. Sivapalasingam S, Barrett E, Kimura A, Van Duyne S, De Witt W, Ying M, Frisch A, Phan Q, Gould E, Shillam P, Reddy V, Cooper T, Hoekstra M, Higgins C, Sanders JP, Tauxe RV, Slutsker L. 2003. A multistate outbreak of Salmonella enterica serotype Newport infection linked to mango consumption: I mpact of water dip disinfestation technology. Clin Infect Dis 37:1585 90.
92 Sivapalasingam S, Friedman CR, Cohen L, Tauxe RV. 2004. Fresh produce: A growing cause of outbreaks of foodborne illness in the United States, 1973 through 1997. J Food Prot 67 : 2342 53. Suslow T. 1997. Postharvest chlorination. Publication 8003. Oakland, CA: University of California, Agriculture and Natural Resources Available from: http://ucgaps.ucdavis.edu/documents/W ater_Disinfection1890.pdf Accessed May 5 2011. Taylor E, Kastner J, Renter D. 2010. Challenges involved in the Salmonella Saintpaul outbreak and lessons learned. J Public Health Manag Pract 16:221 31. Thompson JF, Mitcham EJ, Mitchell FG. 2002. Preparat ion for fresh market. In: Kader AA, editor Postharvest t echnology of h orticultural c rops. 3rd ed. Oakland, CA: University of California Agriculture and Natural Resources. p 67 79. Todd ECD, Greig JD, Bartleson CA, Michaels BS 2008. Outbreaks where food workers have been implicated in the spread of foodborne disease. Part 5. Sources of contamination and pathogen excretion from infected persons. J Food Prot 71:2582 95. [USDA] US Department of Agriculture 20 10 a World tomatoes, all: P roduction by country, 1961 2008 Table 98. Washington, D C: USDA, Economic Research Service. Available from: http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=12 10 Accessed May 5 201 1 [USDA] US Department of Agriculture 20 10b Fresh tomatoes: US import volume and value by selected country, 1978 2009 Table 80. Washington, DC : USDA, Economic Research Service. Available from: http://usda.mannlib.cornell.edu/MannUsda/viewDocumentInfo.do?documentID=12 10 Accessed May 5 201 1 [USDA] US Department of Agriculture 201 1a Vegetables 20 10 s ummary. Washington, DC : USDA, National Agricu ltural Statistics Service Available from: http://usda.mannlib.cornell.edu/usda/current/VegeSumm/VegeSumm 01 27 2011.pdf Accessed May 5 201 1 [USDA] US D epartment of Agriculture 201 1 b Vegetables and melons outlook: Per capita use (consumption) Washington, DC : USDA, Economic Research Service. Available from: http://www.ers.usda.gov/pu blications/vgs/VGSTables.htm Accessed May 5 201 1 [USDA and USDHHS] US Departments of Agriculture and Health and Human Services 20 10 Dietary guidelines for Americans, 20 10 7 th Edition Washington, DC: US Government Printing Office. Available from: http://www.cnpp.usda.gov/Publications/DietaryGuidelines/2010/PolicyDoc/PolicyD oc.pdf Accessed May 5 201 1
93 Vigneault C, Bartz JA, Sargent SA. 2000. Postha rvest decay risk associated with hydrocooling tomatoes. Plant Dis 84:1314 8. Walter EHM, Nascimento MS, Kuaye AY. 2009. Efficacy of sodium hypochlorite and peracetic acid in sanitizing green coconuts. Letters Appl Microbiol 49:366 71. Wei CI, Huang TS, Kim JM, Lin WF, Tamplin ML, Bartz JA. 1995. Growth and survival of Salmonella Montevideo on tomatoes and disinfection with chlorinated water. J Food Prot 58:829 36. Yuk HG, Bartz JA, Schneider KR. 2005. Effectiveness of individual or combined sanitizer treatm ents for inactivating Salmonella spp. on smooth surface, stem scar, and wounds of tomatoes. J Food Sci 70:409 14. Yuk HG, Bartz JA, Schneider KR. 2006. The effectiveness of sanitizer treatments in inactivation of Salmonella spp. from bell pepper, cucumber, and strawberry. J Food Sci 71:95 9. Zagory D. 1999. Effects of post processing handling and packaging on microbial populations. Postharvest Biol Technol 1 4:313 21. Zhuang RY, Beuchat LR, Angulo FJ. 1995. Fate of Salmonella Montevideo on and in raw tomatoes as affected by temperature and treatment with chlorine. Appl Environ Microbiol 61:2127 31.
94 BIOGRAPHICAL SKETCH Alexandra S. Chang was born in Fresno, C alifornia and grew up in Rockville, M aryland In May 2009 she earned B achelor of S cience in b iology from the University o f North Carolina at Chapel Hill at the University of Florida in June 2009 as a lab assistant and then began the f ood s cience program in January 2010. She graduated in August 2011 and continue d h er professional interests in food safety and microbiology.