Evaluation of ultraviolet (UV‐C) light treatment for microbial inactivation in agricultural waters with different levels of turbidity

Abstract Produce growers using surface or well water to irrigate their crops may require an appropriate water treatment system in place to meet the water quality standard imposed by FSMA Produce Safety Rule. This study evaluated the potential of using ultraviolet (UV‐C) treatment in reducing the microbial population in agricultural water. Waters with turbidity levels ranging from 10.93 to 23.32 Nephelometric Turbidity Units (NTU) were prepared by mixing pond water and well water. The waters were inoculated with a cocktail of generic Escherichia coli (ATCC 23716, 25922, and 11775) and then treated with UV‐C light (20–60 mJ/cm2). All tested doses of the UV‐C treatment reduced the E. coli levels significantly (p < .05) in the water samples with the turbidity levels up to 23.32 NTU. The decrease in the turbidity from 23.32 to 10.93 NTU increased the level of reduction by more than 2.15 log most probable number (MPN)/100 ml). UV‐C treatment effectively reduces microbial load in agriculture water; however, turbidity of water may significantly affect the disinfection efficacy. The study also demonstrated that sprinkler system resulted in a higher level of contamination of cantaloupes compared with drip irrigation. The results indicated that UV‐C treatment could be a promising strategy in reducing the produce safety risks associated with irrigation water.

If the contaminated water is used for the irrigation of agricultural crops, it poses a high risk of foodborne outbreaks (Olaimat & Holley, 2012).
The Food Safety Modernization Act (FSMA) Produce Safety Rule (PSR) has identified agricultural water as an important source of microbial contamination to produce (Fan et al., 2008). The rule requires that the water used for irrigation and production/processing of fresh produce must be safe and of adequate sanitary quality for its intended use. The growers are recommended to regularly monitor the microbial quality of their water sources by testing generic E. coli (FDA, 2019;Chhetri, 2018). The easily distinguishable characteristics of generic E. coli make it a principal indicator organism to assess water contamination with human pathogens (FDA, 2019). It is a predictor of undesirable conditions such as ineffective treatment or fecal contamination (Gekenidis et al., 2018).
Several methods are currently available for the treatment of drinking and irrigation water such as chlorination (Beuchat, 1999;Chhetri, Janes, King, Doerrler, & Adhikari, 2019;Whan et al., 2001), chlorine dioxide (Carrillo, Puente, & Bashan, 1996), ozone (Kim, Yousef, & Dave, 1999), and filtration (Koivunen, Siitonen, & Heinonen-Tanski, 2003). Conventional chemical treatments may affect the quality of the crops and soil (Hua & Reckhow, 2007) while conventional filtration methods may not be suitable for surface water due to its complexity and variability in the effectiveness. The efficacy of filtrations may vary with raw water quality including turbidity, type of microorganisms, type of filtration material and pore size, and filtration rate (LeChevallier & Au, 2004). Regular cleaning for maintaining adequate water flow rate may be another limitation of filtration system (Burch & Thomas, 1998). The use of UV-C light may overcome these limitations (Hijnen, Beerendonk, & Medema, 2006). The mechanism of UV-C light disinfection is based on the formation of pyrimidine dimers in microbial DNA. The dimers interfere with the replication, transcription, and, thus, translation process of microorganisms (Koutchma, Forney, & Moraru, 2009).
In the United States, groundwater, surface water, and municipal water are the common sources of irrigation water for fresh produce (Jongman, Chidamba, & Korsten, 2017;Pedrero, Kalavrouziotis, Alarcón, Koukoulakis, & Asano, 2010). Although chlorine treatment is the most common practice for disinfecting municipal water, UV-C light treatment has also been applied by some municipalities (Hijnen et al., 2006). As the turbidity of groundwater and surface water is generally higher than municipal water (Topalcengiz, Strawn, & Danyluk, 2017), the efficacy of UV-C in these waters could be different. This study used a large volume of surface water and well water to simulate a water treatment system in agricultural settings (Jones, Worobo, & Smart, 2014). This study evaluated the efficacy of UV-C light treatment in reducing generic E. coli in surface water, well water, and their mixtures with different level of turbidity. Furthermore, the effect of UV-C-treated irrigation water on the generic E. coli levels on the cantaloupes was evaluated in an agriculture setting.

| Inoculum preparation
A cocktail of generic E. coli strains (ATCC 23716, 25922, and 11775) was used in this study. These strains were previously used in irrigation water treatment studies and are among the few well-characterized surrogates for use in field trials (Harris et al., 2012). Before activation, bacterial strains were stored at −80°C. Frozen cultures were activated in three successive passes by following the procedure described by . The final inoculum size of the generic E. coli was 10 8 CFU/mL.

| Water sample collection
Water samples were collected from a pond and a well located at the LSU AgCenter Botanic Gardens in Baton Rouge, Louisiana. To resemble natural variability that may occur in surface water irrigation sources, we used pond water (P), well water (W), and two mixtures of the pond and well water (PW, 1:1; WP, 4:1). The different sources and mixture of waters helped us to maintain a different level of turbidity and transmittance (Tables 1 and 2). The waters were collected in a tank (1,000 L) in the proportions mentioned above. The pond water was prefiltered using a cloth (Standard Test Sieve No. 25; W.S. Tyler) attached at the end of the pipe to remove larger particles. Each batch of water in the tank (1,000 L) was inoculated with the generic E. coli followed by agitation for 2 min using a sterile plastic pedal. Water samples (100 ml) were collected from the tank before and after inoculation. The final inoculum size in the water was maintained at 7-8 log MPN/100 ml. Experiments were replicated three times to capture the variability associated with the water and en-

| UV-C light treatment
A UV-C light treatment equipment (PMD 150C1/4, Aquionics, Slough) was used in this study. The PMD 150C1/4 uses a photon medium-pressure disinfection with an arc tube that has a medium-  (Table 1) and four best ranges of UV doses (20-30, 30-40, 40-50, and 50-60 mJ/cm 2 ) were selected for this study. Each batch of inoculated water was passed through the UV-C treatment equipment at different doses and was collected in another tank. Water samples (triplicate) were collected in a sterilized container (100 ml) before and after each UV treatment. All the experimental treatments were repeated three times.

| Generic E. coli microstructure by Scanning Electron Microscopy
Scanning electron microscopy (SEM) was used to examine the microstructure of generic E. coli in the pond water using the method previously described by Kenzaka and Tani (Kenzaka & Tani, 2012).
Briefly, the samples were centrifuged at 10,000g for 5 min, and the pellets were suspended in phosphate-buffered saline (PBS; pH 7.2).

| Irrigation of cantaloupes with UV-Ctreated water
"Hales Best Jumbo" cantaloupe seedlings (Cucumis melo reticulatus) were transplanted in a field at LSU AgCenter Botanic Garden, Baton Rouge, Louisiana. The field had plot sizes of 5′ × 10′ and 10′ × 15′ (3 plots each treatment) with 5 and 10 plants in each plot, respectively.
The study was conducted twice, in July and October. The treatment schemes were drip, and sprinkler irrigation with UV-C-treated and UV-C-untreated water. The irrigation water was prepared by mixing the pond water with well water in an equal proportion (final turbidity: 19.70 NTU). The total volume of the water was 1,000 L in a plastic tank with a capacity of 1,100 L. The water was then inoculated with three generic E. coli strain cocktail (ATCC 11775, 23716, and 25922) maintaining the final concentration of 7-8 log MPN/100 ml. The water was treated with the UV-C dose of 50-60 mJ/cm 2 (Figure 1) Abbreviations of treatments are as follows: Pond, pond water; PW, pond water + well water (1:1); Well, well water; WP, well water + pond water (4:1).

| Statistical analysis
Data were analyzed using the SAS ® program (SAS Institute) with an ANOVA analysis, using a Tukey's honestly significant separation difference with a confidence level of 95% (p ≤ .05) for every water mixture and each range of dose used.

| Water quality
The pH, turbidity, and transmittance in water samples are shown in Table 2. The pond water and the well water had a pH of 7.04 and 8.07, respectively. The mixtures of pond water and well water, PW (1:1), and WP (4:1) had the pHs of 7.76 and 8.01, respectively. The turbidity of the pond water, well water, PW, and WP was 23.32, 10.93, 19.70, and 13.16 NTU, respectively. The transmission value was lower for the pond water (29.16) compared with the well water (88.11).
The turbidity and pH of the water may vary with sources and season. The US EPA recommends the pH range of irrigation water to be in the range of 6.5-8.4 (USEPA, 2012). We observed that the pHs of the waters that were used in this study were within the EPA range. The

| UV-C light was effective in reducing the generic E. coli levels in all the water sources
The effect of UV-C light treatment in inactivating generic E. coli in water with different turbidity levels is shown in

TA B L E 3 Reduction of generic
Escherichia coli in the four water sources using different UV-C doses a reduction was by 3.75 due to UV-C dose of 20-30 mJ/cm 2 . An increase in the dose to 50-60 mJ/cm 2 resulted in a significantly (p < .05) higher level of reduction (by 5 log MPN/100 ml). In our preliminary study, we evaluated the UV-C dose up to 120-130 mJ/ cm 2 for the pond water and PW to see whether the bacterial reduction could further be increased. However, we did not see significant changes in the reduction levels (data not shown). Based on this result, we selected the UV-C doses of 20-60 mJ/cm 2 in this study.
We observed a strong negative correlation between the efficacy of UV-C treatment (

| Role of irrigation on the level of E. coli on cantaloupe surfaces
The level of E. coli on cantaloupe surfaces after irrigation for three consecutive days is shown in    et al., 2016). This seasonal variation in the microbial population may be attributed to the changes in weather conditions such as temperature, humidity, and day length (Ailes et al., 2008;Chhetri, Fontenot, et al., 2019;Chhetri et al., 2018). However, further study is needed to establish relationships between year-round weather conditions and microbial quality of produce.
Our results indicated that sprinkle irrigation resulted in higher microbial contamination on cantaloupe surfaces compared with drip irrigation. Other studies have reported similar results. Sprinkle irrigation increased the risk of microbial contamination of lettuce surfaces (Fonseca, Fallon, Sanchez, & Nolte, 2011;Van der Linden et al., 2013). Sprinkler irrigation delivers water directly to the surface of crops. If the irrigation water is contaminated, this method becomes an easy mechanism by which pathogens are disseminated on agricultural crops. Also, the sprinkler irrigation may produce injuries on the crop surfaces during harsh environmental conditions creating a favorable condition for the survival of E. coli (Harapas, Premier, Tomkins, Franz, & Ajlouni, 2010). Drip irrigation has been reported to be comparatively a safe method in terms of cross-contamination of agriculture crops (Song, Stine, Choi, & Gerba, 2006). A similar effect on the microbial load on cantaloupes between UV-C-treated and UV-C-untreated water indicated that there could be other potential sources of E. coli contamination, which greatly dominated the prevalence of E. coli level on cantaloupes. Contamination of fresh produce can also be attributed to other factors such as soil, via animals or insects, and subsequent human handling (Gutierrez-Rodriguez & Harriset al., 2003). Therefore, contamination of produce with E. coli in farm may be unavoidable, and the presence of generic E.coli may not necessarily confirm the presence of pathogens.
However, this study suggested that UV-C treatment can reduce the pathogen levels in irrigation water reducing the produce safety risks associated with irrigation water. Further study is needed using more specific bacterial strains that can be discriminated from the normal environmental flora for a better understanding of the effect of irrigation water quality on crop contamination.
Overall, our study demonstrated that UV-C treatment could be a promising tool to reduce the microbial load in agriculture water, including surface water. The UV-C disinfection efficacy was dependent on the turbidity of the water. Therefore, the efficacy of the treatment may be enhanced including a prefiltration step in the treatment system. Our on-farm study demonstrated that sprinkler system could result in a higher level of contamination of produce compared with drip irrigation. Further study is needed to understand the mechanism to produce contamination through irrigation. Slough, UK, for providing us with the UV-C treatment system for this study.

CO N FLI C T O F I NTE R E S T
The above-named authors of the submitted manuscript confirm no conflict of interest.

E TH I C A L S TATEM ENT
This study does not involve any human or animal testing. The protocol was reviewed and approved by the Inter-Institutional Biological and Recombinant DNA Safety Committee.