Tomatoes are an important agricultural fruit, placing nine per production volume among most popular agricultural produce in the Ukraine. Referencing FAOSTAT (2017), top ten tomato growing countries were China, India, Turkey, USA, Egypt, Iran, Italy, Spain, Mexico, and Brazil, with Ukraine present in the top twenty and producing as much as 2,267,460 tonnes in 2017 alone. The visual appearance of the tomato surface as “clean” may give a false feeling of safety of its consumption. However, two large groups of microorganisms, which are of concern on fresh produce, are spoilage and pathogenic, which may cause either spoilage or foodborne illnesses, are invisible to human eye (Jay, 1998). Enteric Gram-negative pathogens, including Salmonella and Escherichia coli O157:H7, may be present on fresh tomatoes as contamination from environment and may persist on the surfaces (Tokarskyy et al., 2018; Tokarskyy and Schneider, 2019). Common tomato spoilage microorganisms include Gram-negative rods (Erwinia carotowora), Gram-positive sporeformers (Bacillus spp.), yeasts and molds (Shi et al., 2009; Jay, 1998). Although novel methods to decontaminate surface of edible foodstuff are available (Tokarskyy and Marshall, 2010), they remain expensive comparing to the use of low-cost alternatives, such as chlorine sanitizers (Dychdala, 2001; Tokarskyy et al., 2015). One of the approaches to reduce microbial load and prevent cross-contamination on tomatoes before retail sale is through their washing with low concentration chlorine sanitizer (Chang and Schneider, 2012; Gereffi, Sreedharan and Schneider, 2015). For example, one of the most common tomato processing system in the United States is a flume tank with 150 ppm free chlorine (pH 6.5 to 7.5) and a maximum of a 2 minute treatment (Gereffi, Sreedharan and Schneider, 2015). Gereffi, Sreedharan and Schneider (2015) have shown that even 25 ppm of chlorine may be adequate to prevent cross-contamination of tomatoes with Salmonella if the concentration is properly maintained, chemical oxygen demand does not exceed 500 ppm, and tomatoes are treated for at least 120 seconds in a flume tank. Such tank may be terminally equipped with an additional overhead spray and brush roller system, where increased physical removal of bacteria with brushes in conjunction with antimicrobial efficacy of sanitizers may greatly improve decontamination step (Chang and Schneider, 2012). The primary purpose of chlorine sanitizer is to prevent cross-contamination and bacterial build-up (Gil et al., 2009). Though chlorine is believed to have non-specific mode of action, bacterial spores have innate resistance at concentrations used in food industry (Davidson and Harrison, 2002). Some of the naturally occurring bacteria found on tomatoes do include Bacillus spp., in addition to Cyanobacterium spp., Erwinia spp., Enterobacter spp., Pantoea spp., Pseudomonas putida, among others (Shi et al., 2009). Chlorine is capable to reduce natural microbial contamination level on produce, but never eliminate it completely (Allende et al., 2009; Rahman, Ding and Oh, 2010). Chang (2011) found that with initial population of natural microflora on tomato surfaces of 5.31 log units, 100 ppm chlorine significantly reduced more natural microflora than water with a 1.41 log units reduction after 30 seconds treatment (p <0.05), but never below detection limit. Increasing treatment time to 60 seconds did not significantly affect efficacy.

The purpose of the current study was to compare microbial loads of “visibly clean” and “visibly dirty” tomatoes, to evaluate influence of 100 ppm chlorine wash on different groups of natural microflora on tomato surfaces using overhead spray-brush roller system, as well as to evaluate resistance to chlorine of residual microflora in order to better understand surviving natural microorganisms after treatment.

There was a significant influence of analyzed factor with variations as “C”, “D”, “C-W”, and “C-CHL” on microbial counts of all three microbiological media plated (p <0.05).

Surprisingly, there was no significant difference in total plate count (6.01 ±0.50 vs. 6.33 ±0.52 log10 CFU.tomato_-1), Gram-negative counts (5.71 ±0.65 vs. 5.80 ±0.60 log10 CFU.tomato_-1), and yeast and mold counts (4.42 ±0.54 vs. 4.56 ±0.41 log10 CFU.tomato_-1) between clean and dirty tomatoes, respectively (p > 0.05, Figure 2). This can be attributed to smaller sample sizes, comparing to other studies (Schneider et al., 2017; De et al., 2018). Water alone decreased TPC by 0.40 log10 CFU.tomato_-1, GNC by 0.48 log10 CFU.tomato_-1, and YM by 1.24 log10 CFU.tomato_-1on water-treated tomatoes (Figure 2).

Chlorine wash was more effective with corresponding average reductions of 1.12, 1.19, and 1.66 log10 CFU.tomato_-1 for TPC, GNC, and YM, respectively (Figure 2, Table 1). Interestingly, the highest reduction was observed in YM counts, while Dychdala (2001) noted that higher chlorine, 135 to 500 ppm, is required to inactivate molds. Based on water alone data, it can be concluded that yeasts and molds might have been simply washed off tomatoes without kill step. Similarly, Schneider et al. (2017), while analyzing pre- and postprocessed tomatoes from Florida, New Jersey and Maryland packinghouses in spring, have found that average microbial TPC per untreated tomato was 6.25 log10 CFU.tomato_-1 and 5.31 log10 CFU.tomato_-1 for chlorine water flume tank treated tomatoes, corresponding to 0.94 log10 CFU.tomato_-1 reduction. Considering large number of analyzed samples, overall range for TPC for tomatoes collected year-round was 2.3 to 12.1 log10 CFU.tomato_-1 with median of 6.9 log10 CFU.tomato_-1 (Schneider et al., 2017).

MacConkey agar is selective and differential medium for bacteria, formulated to selectively isolate Gram-negative and enteric bacilli. Therefore, it may be argued that GNC is a subset of TPC, and though similar relative log reductions of microorganisms in each group of these two groups were found, absolute reductions in counts as CFU.tomato_-1 suggest that chlorine may have had an “all kill” approach, reducing not only Gram-negative bacteria counts, but also Gram positive (Table 1). Similarly, Schneider et al. (2017), while analyzing larger sets of preand post-processed tomatoes, have found total coliforms counts on CHROMagar™-ECC to be 5.13 log10 CFU.tomato_-1 and 4.70 log10 CFU.tomato_-1 for untreated and chlorine flume tank treated tomatoes, respectively.

This observation, together with no significant difference between inhibition zones by chlorine for untreated and chlorine-treated tomato residual microflora (p >0.05), 18.4 ±1.7 and 19.9 ±2.3 mm, respectively, suggested that chlorine may have had no preferential kill, but rather a shotgun approach (Figure 3).

However, concentrated circle patterns were observed on disc diffusion plates, suggesting that certain microorganisms on the tomato surface might be indeed more sensitive to chlorine (Figure 4).

High counts remaining after treatment with chlorine suggested resistant biofilm formation on the surface of tomatoes. Another suggested explanation by Fatica and Schneider (2009) is that natural microflora is hiding in crevices and pockets of the hydrophobic, waxy cuticles of the produce, where aqueous chlorine sanitizer cannot enter.

To summarize, cleanness may attribute to lower counts on the surface of tomatoes, irrespective of microbial group analyzed, though microbial counts were not significantly different. Larger sets of tomatoes are needed to fortify this statement. Although 100 ppm chlorine treatment reduced all microbial counts significantly better than water alone, it failed to bring them below detection levels, suggesting strong interaction such as biofilm formation, between natural microflora and tomatoes. Comparing reductions of microorganisms in each group, it was concluded that chlorine may have no preferential kill but rather a shotgun approach.