ScholarWorks @ UTRGV ScholarWorks @ UTRGV Effect of surrounding vegetation on microbial survival or die-off Effect of surrounding vegetation on microbial survival or die-off on watermelon surface in an agriculture setting on watermelon surface in an agriculture setting

Preharvest contamination of produce with food borne pathogens has been a major food safety issue. In this study, we investigated the effect of surrounding vegetation on the survival of natural and inoculated generic Escherichia coli on watermelon rinds in an agricultural field setting. There was no significant difference ( p > .05) on the populations of natural generic E. coli (1 – 1.46 log Most Probable Number (MPN)/sample) and coliforms (<3.99 log CFU/cm 2 ) on watermelons harvested from low, medium, and high levels of vegetation. However, the survival rate of generic E. coli inoculated on watermelon rind discs was variable with the level of vegetation. A significant reduction in generic E. coli count was observed within 12 hr at all vegetation levels. After 108 hr, discs placed at low vegetation level had a highest die-off reduction (3 log Colony Forming Units (CFU)/cm 2 ) compared to medium and high vegetation levels. in the agricultural field. The findings of this study emphasize the importance of considering the surrounding vegetation while making decisions for developing preharvest risk management strategies based on microbial die-off rate calculations. dislodge the microorganisms from the melon surface. In our preliminary study, we observed a low number of generic E. coli on watermelon surface. Thus, to reduce the dilution level, we used only 200 ml 0.1% peptone water but initiated a longer massaging time. The eluent was then used for the microbial analysis.


O R I G I N A L A R T I C L E
Effect of surrounding vegetation on microbial survival or die-off on watermelon surface in an agriculture setting Vijay Singh Chhetri 1 | Kathryn Fontenot 2 | Ronald Strahan 2 | Veerachandra K. Yemmireddy 1 | Katheryn J. Parraga Estrada 1 | Achyut Adhikari 1

Practical applications
To ensure preharvest produce safety, the Food Safety Modernization Act (FSMA) produce safety rule has suggested a time interval between last irrigation and harvest for potentially contaminating microorganisms to die-off. However, a knowledge gap exists regarding the influence of surrounding vegetation on microbial die-off rates on produce in the agricultural field. The findings of this study emphasize the importance of considering the surrounding vegetation while making decisions for developing preharvest risk management strategies based on microbial dieoff rate calculations.

| INTRODUCTION
Fruit and vegetable crops have the potential to be contaminated with pathogenic microorganisms in the field. As fresh produce is often consumed raw, a higher risk of foodborne illness is posed with this food group. The number of foodborne illness outbreaks associated with the consumption of fresh fruits and vegetables has noticeably increased in the last decade (Huang & Chen, 2011;Olaimat & Holley, 2012). During 1973During -2014, fruit and vegetable crops were the most commonly implicated commodities for several foodborne illness outbreaks (Crowe, Mahon, Vieira, & Gould, 2015;Herman, Hall, & Gould, 2015;Nguyen et al., 2015). Preharvest contamination of produce is commonly originated from the soil, inadequately composted manure, contaminated irrigation water, and improper human handling of produce (Annous, Solomon, Cooke, & Burke, 2005;Tomas-Callejas et al., 2011). The intrusion of crops by wild animals, birds, reptiles, and rodents, as well as insects and nematodes, act as a vector for transferring various pathogens (Brandl, 2006). The survival and growth of microorganisms is influenced by several environmental factors and agricultural practices such as, exposure to solar UV radiation, temperature changes, humidity, and poor fertilizer regimes (Bezanson et al., 2012;Brandl, 2006;Nyeleti, Cogan, & Humphrey, 2004;Tomas-Callejas et al., 2011;Weller et al., 2017).
Several studies reported that the sunlight of tropical latitudes (Davies & Evison, 1991;Nyeleti, Cogan, & Humphrey, 2004;Obiri-Danso, Paul, & Jones, 2001) and a concomitant increase in the surface temperature of produce (Tomas-Callejas et al., 2011) have an inhibitory effect against various microbial pathogens. Moreover, sunlight is found to reduce Salmonella levels in fresh water sources (Davies & Evison, 1991) and on food contact surfaces such as stainless steel (Nyeleti et al., 2004). Furthermore, microbial populations decline with decreasing nutrient availability because of failing to lower their metabolic rate to adopt the starvation condition (Fontaine, Mariotti, & Abbadie, 2003). All these factors contribute to the natural decline of microbial populations on produce surfaces in the field and should be considered for microbial die-off rate calculations (Davies & Evison, 1991;Reddy, Khaleel, & Overcash, 1981).
To minimize preharvest microbial food safety risk originated from contaminated irrigation water, the FDA Food Safety Modernization Act (FSMA) Produce Safety Rule requires agricultural water used for covered produce, must be of safe and adequate sanitary quality (USFDA, 2015). It should also meet the generic Escherichia coli requirements as proposed in the U.S. Environmental Protection Agency's 2012 Recreational Water Quality Criteria. However, the rule provides flexibility to the growers that are not able to meet the microbial water quality criteria initially by extending the time interval between the last irrigation and first harvest to allow for microbes to naturally die (USFDA, 2015).
Research on the surface of produce highlights the effect of environmental factors on the die-off rate of microorganisms (Wood, Bezanson, Gordon, & Jamieson, 2010). However, the level of exposure to environmental stressors, in particular, solar UV radiation on the surface of produce may be influenced by growing practices. Presence of weed and/or the surrounding vegetation from the plant may cover fruit surfaces, thus preventing the exposure of the contaminated surface to the natural sunlight. As per our knowledge, the effects of surrounding vegetation have not been considered while calculating microbial die-off rates on produce such as melons. This is especially imperative as melons grow in direct contact with the soil with surrounding vegetation making them at a higher risk of microbial contamination. Several recent outbreaks have been attributed to microbial contamination of melon crops (McCollum et al., 2013). During 1973During -2011loupe, and honeydew melons were responsible for 34 food borne disease outbreaks (Walsh, Bennett, Mahovic, & Gould, 2014).
Investigating the effect of surrounding vegetation on the microbial dieoff rate would help generate data for on-farm food safety risk assessments. Thus, the objective of this study was to determine the effect of surrounding vegetation (weeds and vines) on the survival or die-off rate of generic E. coli on watermelon rind surfaces in an agricultural setting.

| Experimental overview
The present study was conducted on July and August at the Louisiana State University Agriculture Center (LSU AgCenter) Botanic Gardens in Baton Rouge, LA. The test field was divided into three blocks, and each block contained six plots (12 × 30 ft 2 ). Eighteen total plots were tested. Plots were initially treated with one of the following five preemergent herbicides or untreated control. The herbicide treatments were applied 24 hr prior to transplanting. Preemergent herbicide treatments included Strategy 5 pts/acre (Ethalfluralin & Clomazone), Command 3ME 0.67 pts/acre (Clomazone), Strategy 5 pts/acre plus Sinbar 4 oz/acre, Valor 1 oz/acre (Flumioxazin), Sinbar 4 oz/acre (Terbacil), and an untreated control. Preemergent herbicides were applied using a CO 2 backpack sprayer delivering 15 gal/acre. Twentyfour hours after herbicide application, "Legacy' watermelon seedlings were transplanted. Additional herbicides, Sandea (0.67 oz/acre) and Prowl (1 qt/acre), were applied to the row middles as a lay-by application 14 days after planting. Overhead irrigation was applied to estab- and immediately transported to the food safety laboratory at 4 C.
After receiving the samples at the lab, a 200 ml of 0.1% peptone water was added to each watermelon bag and was hand massaged intensively for 5 min to dislodge the microorganisms from the melon surface. In our preliminary study, we observed a low number of generic E. coli on watermelon surface. Thus, to reduce the dilution level, we used only 200 ml 0.1% peptone water but initiated a longer massaging time. The eluent was then used for the microbial analysis.
2.2 | Testing natural coliform, generic E. coli levels and the bacterial pathogens on the watermelon surface Quanti-Tray 2000-Colilert (IDEXX Laboratories, Portland, ME) and Petrifilm EC plates (3 M Microbiology Products Co, St. Paul, MN) were used to enumerate generic E. coli and coliform levels on the watermelon rind surface, respectively. The Quanti-Tray method was used to enumerate the generic E. coli levels because of its lower detection limit (0.30 log MPN/sample) than petrifilm and VRBA plating methods. Briefly, each eluent sample (100 ml) obtained from hand massaging was poured into a sterile plastic container containing Quanti-Tray reagent powder. Contents were thoroughly mixed by gentle agitation and then poured into Quanti-Trays. The trays were sealed using a heat sealer (Quanti-Tray-2X, IDEXX Laboratories) and incubated at 35 AE 0.2 C for 24 hr. The colors of the wells were compared with a comparator provided by IDEXX laboratories, and the number of wells showing fluorescence under a UV lamp (WL200, Hanovia LTD, Aquionics, United Kingdom) was recorded as generic E. coli positive samples. The results were expressed in terms of MPN using a chart provided by IDEXX Laboratories. The enumeration of coliforms on the watermelon rind was done by using 3 M Petrifilm, and the results were expressed in CFU. This is because coliforms were present at higher numbers and the dilution used to detect generic E. coli resulted in all wells positive for coliforms. Each sample was also analyzed for bacterial pathogens (E. coli O157:H7 and Salmonella spp.) using immunomagnetic separation (BeadRetriever, Thermo Fisher Scientific, Waltham, MA) technique followed by spread plating on selective media. However, we did not detect the presence of E. coli O157: H7 or Salmonella spp. on the watermelon surface.

| Watermelon disc preparation and inoculation with generic E. coli
To bring uniform light exposure on the surface of watermelons and better understand the effect of surrounding vegetation on the die-off rate of bacteria, studies were also conducted by artificially inoculating watermelon discs with generic E. coli. Subsequently, these inoculated discs were exposed to natural sunlight in the field under different levels of surrounding vegetation. Briefly, fresh watermelons grown in our test plots were first washed with sterile deionized water and airdried inside a biological safety cabinet for 1 hr at room temperature. A total of 63 watermelon discs (surface area of 20 cm 2 and thickness of 0.5 cm) were prepared using sterile stainless-steel knives by coring off the edible portion. The watermelon discs were placed on sterile petridishes with the outer epidermal surface facing up.
A cocktail of three generic E. coli strains (ATCC 23716, 25922, & 11775) were used in this study. These strains are among the few wellcharacterized surrogates for use in field trials (Harris et al., 2012). Frozen cultures were activated in three successive passes by following the procedure described by Adhikari, Syamaladevi, Killinger, and Sablani (2015). The final inoculum size was 10 9 CFU/ml. The bacterial cocktail was agitated 25 times in a 30 cm arc to ensure thorough mixing. The surface of the discs placed in sterile petri-dishes was spot inoculated with 50 μl inoculum distributed into 15 small droplets. The inoculated discs were dried inside the biological safety cabinet for 12 hr.

| Enumeration of inoculated generic E. coli on watermelon rind disc
After harvesting all the watermelons, three plots were selected representing one for each level of surrounding vegetation (low, medium, and high). Inoculated watermelon discs kept on sterile petri plates (n = 21) were placed randomly around in each level of surrounding vegetation plot (Figure 4). Samples of three watermelon discs from each plot were collected at 0, 12, 36, 60, 84, and 108 hr (at 6 p.m.).
The discs were stored in an ice chest and transported to the laboratory maintaining 4 C. Each disc was placed into a sterile stomacher bag containing 90 ml of 0.1% peptone water and hand-massaged for 1 min followed by blending in a stomacher (BagMixer 400S, Interscience, Woburn, MA) for 5 min. The eluent was used for the enumeration of the generic E. coli by plating on Violet Red Bile Agar (Criterion, Bio-Rad Laboratories, Inc, Hercules, CA) (with detection limit: 1.65 log CFU/cm 2 ) and incubated at 37 C for 24 hr. Several studies reported a higher level (>5 log CFU/g) of microbial count on fresh produce such as cantaloupe, lettuce, and broccoli (Johnston et al., 2005;Liu, Tan, Yang, & Wang, 2017;Zhang & Yang, 2017). The aerobic mesophilic count was more than 3 log CFU/g in the fresh cut-honeydew melons (Chong, Lai, & Yang, 2015), while the total coliform level on the cantaloupe rinds was 2.4 log CFU/g (Gagliardi, Millner, Lester, & Ingram, 2003). The level of E. coli varied with the types of the fresh produce, with higher count (geometric mean of 1.5 log CFU/g) in cantaloupe (Johnston et al., 2005) Categorically, those plots with 0-3 were considered as low level of vegetation plots, 4-6 as medium level of vegetation plots, and 7-10 as high level of vegetation plots. The detectable limit was 0.30 log MPN/sample reduction was statistically not significant (p > .05) between medium and high vegetation levels. A relatively high reduction of generic E. coli on the surface of watermelon was expected in the first few hours of exposure to on-farm daylight conditions. Brandl (2006) reported that the direct exposure of produce surface to the sunlight has a detrimental effect on the survival of enteric bacteria. In our study, the watermelon discs at low surrounding vegetation were found to be fully exposed to sunlight. While the surface of watermelon was not fully exposed to sunlight at medium and high sur- radiation, and resume their growth when UVA light is very scarce or when the UVA irradiation is stopped. In our study, the night time (i.e., between 12 hr (6 p.m.) and 24 hr (6 a.m.) may have allowed adequate time for the bacterial cells to recover from damage and resume their growth. Studies reported that relative humidity and temperature could affect microbial survival or die-off rate on produce surface (Del Rosario & Beuchat, 1995;Stine, Song, Choi, & Gerba, 2005;Weller et al., 2017). During the current study, the relative humidity ranged from 62% to 74%, and the average temperature ranged from 25 C to 28.56 C (Table 1) Several studies reported that the sunlight and its ultraviolet component is a significant factor in inhibiting pathogens (Davies & Evison, 1991;Heaton & Jones, 2008;Nyeleti et al., 2004). Nyeleti et al. (2004 found that simulated sunlight decontaminated Salmonella on the stainless-steel surface under field conditions. In addition, longer In a similar study, Bezanson et al. (2012) observed an average daily dieoff rate of E. coli O157:H7 on lettuce by 0.56 log. It was interesting to notice that, in low vegetation level, the generic E. coli counts at 108 hr were similar to those of 12 hr. However, in medium and high vegetation level, the counts recovered to the initial level (~6 log CFU/cm 2 ).

| Statistical analysis
This indicates that the on-farm microbial die-off rate estimates based on environmental conditions such as exposure to sunlight, UV fraction of sunlight light, temperature and relative humidity variations, and bacterial recovery mechanisms seem plausible in low vegetation conditions.
Whereas, generic E. coli exhibited high survival especially when the melon surface was surrounded with medium and high vegetation.
Overall, this study investigated the effect of surrounding vegetation on the survival of generic E. coli on the surface of watermelon.
Controlling the common contamination sources such as soil and irrigation water, the levels of naturally present generic E. coli were very low