Reduction of Salmonella on alfalfa seeds using peroxyacetic acid and a commercial seed washer is as effective as treatment with 20 000 ppm of Ca(OCl)2
Karl R. Matthews, Department of Food Science, School of Environmental and Biological Sciences, Rutgers, The State University of New Jersey, 65 Dudley Road, New Brunswick, NJ 08901-8520, USA. E-mail: email@example.com
Aims: The efficacy of a commercial seed washer and 1 and 3% peroxyacetic acid or 20 000 ppm calcium hypochlorite for reducing Salmonella on alfalfa seeds was investigated.
Methods and Results: Alfalfa seeds were inoculated with Salmonella Stanley to achieve c. 5 log CFU g−1. Seeds were then treated with 1 or 3% peroxyacetic acid or 20 000 ppm calcium hypochlorite for 15 min in a commercial seed washer that uses air to enhance contact of the sanitizer with the seed. Experiments were also conducted using industry and laboratory methods. An c. 1-log reduction in number of Salm. Stanley was demonstrated regardless of the chemical treatment or method of treatment. Although this 1-log reduction was significant (P < 0·05), differences among the treatments were not significant. Treating the seed with 1 and 3% peroxyacetic acid resulted in similar Salm. Stanley reductions of 1·77 and 1·34 log, respectively, not being statistically significant (P > 0·05).
Conclusions: These results suggest that under conditions tested, 1 or 3% peroxyacetic acid solutions are equally effective as 20 000 ppm of Ca(OCl)2 in the reduction of Salm. Stanley on alfalfa seed when used in conjunction with a commercial seed washer.
Significance and Impact of the Study: A 1% peroxyacetic acid solution could potentially be used in place of 20 000 ppm of Ca(OCl)2 for treatment of seeds used for sprouting. The commercial seed washer did not enhance removal of Salm. Stanley from alfalfa seeds, but did facilitate removal of excess soil from seeds.
Nearly 30 outbreaks of foodborne illness during the past 15 years have been linked to the consumption of raw or lightly cooked seed sprouts contaminated with Escherichia coli O157:H7 or Salmonella (Fett et al. 2006). Bacterial communities associated with sprouts are diverse and can increase during storage, thereby decreasing shelf life and present a risk to human health (Doyle and Erickson 2008; Loui et al. 2008; Martinez-Villaluenga et al. 2008). Alfalfa sprouts that were grown from contaminated seeds had Salmonella populations as high as 106 CFU g−1 (Gandhi et al. 2001; Proctor et al. 2001; Gandhi and Matthews 2003). Growth of microbes during sprouting may be influenced by irrigation frequency, sprouting temperature, type of sprouting method and properties of the microbe (Barak et al. 2007; Fu et al. 2008).
The National Advisory Committee on Microbiological Criteria for Food (NACMCF) suggests sprouting conditions will only enhance the growth of the bacteria on seeds if present (NACMCF 1999). The NACMCF recommends that seeds intended for sprout production be exposed to at least one treatment that will effectively reduce the numbers of bacteria with the use of 20 000 ppm calcium hypochlorite specifically suggested. Unfortunately, the method is not 100% effective. An exhaustive list of chemical sanitizers and antimicrobials, exposure times and heat treatments has been tested independently and in combination with decontaminate alfalfa seed with reductions in E. coli O157:H7 and Salmonella ranging from c. 2 to >5 log CFU g−1 (Lang et al. 2002; Sharma et al. 2002; Fett and Cooke 2003; Kim et al. 2003; Fett et al. 2006; Feng et al. 2007). When a ‘hurdle-strategy’ was evaluated for inactivating Salmonella and E. coli O157:H7 on alfalfa seeds (Scouten and Beuchat 2002), ultrasound in combination with various chemicals [Tsunami® 200 (peroxyacetic acid), Ca(OH)2, Fit® (citric acid)] and heat yielded only a modest increase in the reduction of Salmonella and E. coli O157:H7 on alfalfa seed.
Research suggests that a range of aqueous chemicals are equally effective as 20 000 ppm chlorine at reducing populations of foodborne pathogens on artificially inoculated seeds used for sprouting (Fett et al. 2006). In the present study, we chose to use peroxyacetic acid because it has been shown to effectively reduce foodborne pathogen populations on seeds intended for sprout production (Fett et al. 2006). In addition, it is easy to handle, no pungent off odours are associated with its use, the efficacy is not influenced by pH, and determining working concentration is straightforward.
Alternative cost-effective methods that reduce the microbial levels on seeds used for sprouting must be developed. An air-driven washing apparatus has been developed to facilitate the contact of sanitizer with seeds. The system has been tested commercially (Rajkowski and Ashurst 2009), but no studies have been conducted to determine whether the apparatus is equally or more effective in reducing microbial populations on seeds than standard industry seed-sanitizing practices. This study was conducted to (i) compare a commercially available air-driven seed washer to an industry method and (ii) compare the efficacy of peroxyacetic acid to 20 000 ppm Ca(OCl)2 solution when used in conjunction with a commercial air-driven seed washer to reduce the levels of Salmonella Stanley associated with alfalfa seeds.
Materials and methods
Salmonella Stanley ATCC 7308 used in this study was transformed to express the green fluorescent protein. The isolate used was not associated with an outbreak. The isolate was used in previous studies investigating control of Salmonella on alfalfa seeds and sprouts (Gandhi et al. 2001; Gandhi and Matthews 2003; Liu and Schaffner 2007). Salmonella Stanley has been associated with several outbreaks linked to the consumption of contaminated alfalfa sprouts (Mahon et al. 1997; Fett et al. 2006). Stock cultures were maintained in tryptic soya broth (TSB; Difco, Becton Dickinson, Sparks, MD, USA) containing 30% glycerol at −70°C. Working cultures were maintained at 4°C on tryptic soya agar (TSA; Difco) plates.
Preparation of inocula and inoculation of seeds
Salmonella Stanley was subjected to two 24 h per 37°C transfers in 10 ml of TSB containing 100 μg ml−1 of ampicillin (Sigma, St. Louis, MO, USA). A 1- ml aliquot was then inoculated into 1 l of TSB containing 100 μg ml−1 of ampicillin and incubated at 37°C for 18 h. The culture was then held until reaching room temperature (22 ± 2°C) and then dispensed into a 10 -l screw cap vessel containing 9 l phosphate-buffered saline (PBS; Fisher Scientific, Pittsburg, PA, USA) to achieve a cell concentration of c. 8·0 log CFU ml−1. Mesh bags were filled with 10 kg alfalfa seeds and then immersed into the cell suspension and agitated for 5 min by moving bags using draw strings. The bags of seed were removed from the cell suspension and allowed to drain. Seeds were dried on screens within a biosafety cabinet for 24–48 h and were occasionally mixed to facilitate drying. Inoculated dry seed was stored at 5°C for no longer than 3 weeks before use. Two batches of seed were prepared to complete all experiments. After drying, the seed was assessed for microbial load. Inoculated and uninoculated alfalfa seeds (25 g) were added to 225 ml of sterile 0·1% peptone water in a sterile stomacher bag and pummelled in a stomacher for 2 min. Serial (1 : 10) dilutions in PBS were surface plated (100 μl) in duplicate on TSA supplemented with 100 μg ml−1 of ampicillin. All plates were incubated at 37°C for 24 h. Plates were exposed to ultraviolet illumination, and colonies fluorescing bright green were counted.
Chemical treatment of seeds
Industry method. Calcium hypochlorite [Ca(OCl)2] was added to 16 l of potable water in a 121- l polypropylene container to achieve an initial concentration of 20 000 ppm. Many commercial sprout growers sanitize seeds by placing mesh bags containing seeds into large polypropylene containers that contain 20 000 ppm Ca(OCl)2 solution (personal communication, Earl Hauserman, Brassica Protection Products, LLC). Two mesh bags containing 5 kg of inoculated seed were immersed in the Ca(OCl)2 solution for 15 min and agitated every 5 min. Following treatment, the bags of seed were removed and placed into a 121 -l polypropylene container containing 16 l of fresh potable water and soaked for 15 min, simulating commercial practices. The bags of seed were removed and placed onto a stainless steel screen and allowed to drain prior to microbiological testing. For all experiments regardless of method, treatment water and wash water temperature ranged from 21–23°C. Municipal tap water used in all experiments had a c. pH of 6·8.
Commercial air washer. The tank of a commercially available seed washer (Seed/Sanitizing Washer; Caudill Seed Co., Louisville, KY, USA) was filled with 333 l of 20 000 ppm calcium hypochlorite or 1 or 3% peroxyacetic acid. The washing unit consisted of a plastic tank (polypropylene) with an external pump delivering filtered air. Exact specifications are proprietary. The bottom of the tank contained a series of pipes fitted to deliver air into the sanitizing solution. An adjustable wire platform for holding bags of seed was held in place by chains. Vigorous bubbling of air into the washer facilitated the interaction of sanitizer with seed. Two mesh bags containing 5 kg of inoculated seed were then placed onto the adjustable platform after which the agitation cycle was run for 15 min. After treatment, the bags of seed were removed, the tank drained and filled with potable water. The process required 10 min. The bags of seed were placed back onto the adjustable platform after which the agitation cycle was again run for 15 min, simulating commercial operating practices. The bags of seed were removed and placed onto a stainless steel screen and allowed to drain prior to microbiological testing.
Laboratory treatment. Inoculated alfalfa seed (25 g) was added to 100 ml of the following treatments in 250- ml bottles: 20 000 ppm calcium hypochlorite solution or 1 or 3% peroxyacetic acid. After 15 min of rotary shaking (100 rev min−1), sterile cheesecloth was placed over the mouth of the bottle, and the treatment solution was decanted. Seed was held for c. 5 min before potable water (100 ml) was added to each bottle after which the bottles were again shaken for 15 min, and the wash water decanted. The 5 min dwell time was to account for dwell time associated with the processing of seed using the commercial seed washer. DE (Dey/Engley) broth (75 ml; Difco) was added to each bottle and held at room temperature for 10 min to effectively neutralize the sodium hypochlorite and peroxyacetic acid. Samples were then subjected to microbiological analysis.
Microbiological analysis of seed
Commercial air washer and industry method. Three 25-g seed samples collected from each 5-kg bag of treated seed were added to 75 ml of DE broth in a sterile stomacher bag and pummelled for 2 min using a Stomacher lab blender 400 (Dynatech Laboratories, Inc., Alexandria, VA, USA). Duplicate 100 μl samples were spread plated on TSA supplemented with 100 μg ml−1 ampicillin. Additionally, 1 ml was removed, serially diluted 1 : 10 in PBS, and plated on TSA supplemented with 100 μg ml−1 ampicillin. Population of Salm. Stanley on inoculated untreated seed was determined as described earlier.
Laboratory method. The contents of each jar (25 g seed, 75 ml DE broth) were transferred to sterile stomacher bags and pummelled for 2 min using a stomacher lab blender 400. Samples were processed as described earlier.
Chlorine and peroxyacetic acid analysis
Concentrations of total and free chlorine solutions were determined with a chlorine test kit (CHEMetrices, Inc., Calverton, VA, USA). The peroxyacetic acid concentration was determined using a peroxyacetic acid test kit (Ecolab, St Paul, MN, USA). Kits were used according to the manufacturer’s directions.
Moisture content of seeds was determined using the hot air oven method (Nijenstein et al. 2007). Triplicate 5-g seed samples were dried in a hot air oven at 130 ± 2°C for 1 h, and the per cent moisture determined from the weight loss after drying.
Seeds subjected to the chemical treatments were germinated to determine the impact of treatment on seed viability. Seeds (n = 100) were scattered onto water-saturated filter paper in a petri dish and covered with a second piece of water-saturated filter paper. Plates were held at 22°C for 24–48 h. Each seed was examined for development of a radicle, and the percentage of seeds that germinated calculated.
Microscopic analysis of seed
Approximately 10 g of seed from each treatment was placed in a petri dish viewed using a Leica MZ FL III stereo fluorescence microscope (Leica Microsystems Inc., Bannockburn, IL). At least five images from randomly selected areas of the plate were recorded. Changes in seed coat integrity were noted.
All experiments were conducted at least twice. For the commercial seed washer and industrial method scale experiments, data from each experiment were combined, thus each value reported represents the mean of 12 samples. For the laboratory scale experiments (conducted twice), data from each jar (n = 3) were combined with each reported value representing the mean of six samples. All data were analysed using the sas statistical package (by anova; SAS Institute, Cary, NC, USA).
There were no significant differences (P ≥ 0·05) in the population of Salm. Stanley associated with alfalfa seeds treated with 20 000 ppm Ca(OCl)2 using the commercial air-driven seed washer compared to the industry method (Table 1). Similar reduction (c. 1 log) in population of Salm. Stanley on seeds was achieved following treatment with 20 000 ppm Ca(OCl)2 and 1% peroxyacetic acid.
Table 1. Combined effect of chemical and agitation treatments in reducing Salmonella on alfalfa seeds in mesh bags using a commercial forced air seed washer or industry method
|Control, no treatment||5·11A|| ||96·0|
|Commercial seed washer: 20 000 ppm Ca(OCl)2||3·98B||1·13||90·0|
|Industry method: 20 000 ppm Ca(OCl)2||3·56B||1·55||92·0|
|Commercial seed washer: 1% peroxyacetic acid||3·97B||1·14||94·0|
Treating seeds with 1 or 3% peroxyacetic acid in the commercially available seed washer reduced Salm. Stanley populations 1·77 and 1·34 log, respectively (Table 2). The level of reduction between these two treatments was not significant (P > 0·05). The treatments had no impact on germination.
Table 2. Effect of 1% peroxyacetic acid or 3% peroxyacetic acid in combination with agitation against Salmonella on alfalfa seeds in mesh bags using a commercial seed washer
|Control, no treatment||5·29A|| ||93·0|
|1% Peroxyacetic acid||3·52B||1·77||91·0|
|3% Peroxyacetic acid||3·96B||1·34||88·0|
Experiments conducted using 250-ml-wide mouth media bottles represent ‘typical’ laboratory procedures for evaluating efficacy of aqueous chemical sanitizer for alfalfa seeds. The laboratory scale experiments were conducted to provide a comparison of treatments within the study and to those published in the literature. Similar reduction of viable Salm. Stanley associated with alfalfa seed was demonstrated regardless of the sanitizing treatment (Tables 3 and 4). Under conditions evaluated, treatment with 1% peroxyacetic acid was more effective in reducing the population of Salm. Stanley associated with seeds compared to 20 000 ppm Ca(OCl)2 (Table 3). However, there was no difference (P ≥ 0·05) in reduction of Salm. Stanley populations on seeds following treatment with 1 or 3% peroxyacetic acid. Germination was not notably impacted by the chemical sanitizing treatments.
Table 3. Combined effects of chemical and agitation treatments against Salmonella on alfalfa seed in 250-ml-wide mouth glass bottles (laboratory method)
|Control, no treatment||4·48A|| ||97·0|
|20 000 ppm Ca(OCl)2||4·24A||0·24||92·0|
|1% Peroxyacetic acid||3·17B||1·31||89·0|
Table 4. Combined effects of 1% peroxyacetic acid or 3% peroxyacetic acid and agitation treatments against Salmonella on alfalfa seed in 250-ml-wide mouth glass bottles (laboratory method)
|Control, no treatment||4·59A|| ||93·0|
|1% Peroxyacetic acid||4·03B||0·56||91·0|
|3% Peroxyacetic acid||3·93B||0·66||89·0|
Microscopic analysis of seeds was conducted to determine whether loss of integrity of the seed coat occurred (data not shown). Some seeds did exhibit abrasions; however, the seed coat remained intact. The average moisture content of the seeds before inoculation with Salm. Stanley was 6·71% and after inoculation 7·25%. Substantial change (>1%) in moisture content would suggest loss of integrity of the seed coat.
Treatment of seeds using the recommended treatment of 20 000 ppm Ca(OCl)2 can reportedly reduce populations of Salmonella and E. coli O157:H7 2–5 logs (USFDA 1999; Fett et al. 2006). The use of 20 000 ppm calcium hypochlorite is the standard against which the efficacy of other aqueous sanitizers has been compared (Fett et al. 2006). Research suggests that chemical sanitizers alone are not sufficient to achieve a consistent reduction of foodborne pathogens when associated with seeds used for sprouting (Fett et al. 2006). In the present study, the efficacy of 1% peroxyacetic acid and 20 000 ppm of Ca(OCl)2 to reduce the population of Salm. Stanley associated with alfalfa seeds was determined using a commercially available seed-washing apparatus. The commercial apparatus tested uses vigorously bubbled air to enhance the interaction of sanitizer with seeds. The unit was operated as it would be expected by a commercial sprout grower. Two batches of seed, in total 91 kg, were inoculated with Salm. Stanley, and reduction in population associated with seeds after treatment was determined. The 1-log reduction in Salm. Stanley associated with seeds following treatment using the commercially available seed washer and an industry method, although statistically significant, is likely of little practical relevance given the human health hazard associated with ingesting Salmonella.
Greater reductions in Salm. Stanley populations on alfalfa seeds were obtained using a commercial seed washer and an industry method compared to the laboratory treatment method (Tables 1–4). Regardless of the treatment/processing method used, the level of reduction was less than the c. 2–5-log reduction reported in some previous studies (Beuchat et al. 2001; Scouten and Beuchat 2002; Bari et al. 2003; Gandhi and Matthews 2003; Fett et al. 2006). The limited efficacy observed in our work associated with the 20 000 ppm Ca(OCl)2 treatment may be attributed to a number of factors including the Salmonella strain, exposure time or pH. The pH of the sanitizing solution in the present study was not adjusted to an optimal pH (c. pH 7), to more accurately reflect procedures followed by many sprout growers (personal communication, Earl Hauserman). As pH was not adjusted from c. pH 10 [associated with 20 000 ppm solution of Ca(OCl)2] to pH 7, most chlorine would be present as OCl−, the form least active against micro-organisms (Saper 2009).
In studies evaluating the efficacy of chlorine compounds, buffering agents were used in the preparation of the sanitizing solution, optimizing the activity of the sanitizer (Beuchat et al. 2001; Gandhi et al. 2001; Lang et al. 2002). The antimicrobial properties of chlorine are associated with hypochlorous acid that is most abundant at pH ≤5 with the higher levels of OCl− above pH 5 being relatively ineffective (Saper 2009). Indeed, a 20 000 ppm solution of Ca(OCl)2 has a pH of 10–11. As one of the goals of the present study was to simulate conditions used by many commercial sprout growers, a buffering agent was not used. Variability in reported results for 20 000 ppm Ca(OCl)2 within and between experiments may in part be associated with lack of buffering which would stabilize the Ca(OCl)2. A 5-log reduction in number of Salm. Stanley associated with alfalfa seed was reported following treatment of seed with 20 000 ppm Ca(OCl)2 prepared in buffer [pH 7·0] (Gandhi and Matthews 2003). A decided advantage to the use of peroxyacetic acid is that pH of the sanitizing solution does not need to be adjusted. However, other factors may influence the level of reduction associated with the 1% peroxyacetic acid treatment as results were not consistent among studies using the commercial seed washer (Tables 1 and 2) or using the laboratory method (Tables 3 and 4).
The commercial seed washer used bubbled air to enhance contact of the sanitizing wash with the seeds. The unit was set up and operated as would be expected by a commercial sprout grower. Few if any reports are available in which a direct comparison of commercially available seed-washing equipment to laboratory simulated processing methods have been conducted. A similar commercially available washing apparatus was referenced in a previous study (Rajkowski and Ashurst 2009). The system was used for a 2-year period by two commercial sprout growers, and no sprout water samples tested positive for Salmonella or E. coli O157:H7. However, no side-by-side comparison was made to methods commonly used by sprout growers. Whether the system provided equal or greater reduction in levels of foodborne pathogens that may have potentially been associated with seeds was not determined. Moreover, in that study, no testing was conducted to determine whether the apparatus reduced levels of commensal flora associated with seeds. This would at a minimum suggest that the apparatus enhanced contact of the sanitizer with the seeds. Under conditions evaluated in the present study, the commercial seed washer did not enhance the reduction of Salm. Stanley on inoculated seeds.
The action of the vigorously bubbled air used by the commercial apparatus appeared to remove residual soil (‘dirt’, inorganic material and organic material) from the seeds based on the appearance of a precipitate on the bottom of the tank. This vigorous action did not affect seed germination (Tables 1–4) or compromise seed coat integrity based on analysis using a microscope (data not shown). However, in some instances, abrasions were visible on seed coats following washing in the apparatus. This was likely the results of seeds rubbing against each other, or the interior of the nylon mesh bag. Cracks in the seed coat can provide a niche for microbes and reduce likelihood of the sanitizer to contact the microbe (Fett et al. 2006).
Results of this study suggest that 1% peroxyacetic acid compared favourably to 20 000 ppm of Ca(OCl)2 for reducing Salm. Stanley populations on inoculated alfalfa seed. A greater reduction in population of Salm. Stanley associated with seeds was not achieved using the commercially available seed washer compared to the industry method. A processing strategy was not established that resulted in reduction of Salm. Stanley populations >2 log on the seed. Alternative strategies (i.e. chemical and physical) must continue to be explored for elimination of human pathogens that may be associated with seeds used in sprout production.