The effect of urea and ammonia treatments on the survival of Salmonella spp. and Yersinia enterocolitica in pig slurry

Authors


Correspondence

Declan Bolton, Department of Food Safety, Ashtown Food Research Centre, Teagasc, Ashtown, Dublin 15, Ireland. E-mail: declan.bolton@teagasc.ie

Abstract

Aims

The objective of this study was to investigate the survival of Salmonella and Yersinia enterocolitica strains in pig slurry and evaluate urea and ammonia as disinfection strategies.

Methods and Results

Salmonella Anatum, Salmonella Derby, Salmonella Typhimurium DT19 and Y. enterocolitica bioserotypes 4, O:3, 2, O:5,27 and 1A, O:6,30 were selectively marked by insertion of the plasmid, pGLO encoding for green fluorescent protein and for ampicillin resistance. Strain cocktails were inoculated into fresh pig slurry (control), slurry treated with urea [final concentration 2% w/w, (0·33 mol l−1)] and slurry treated with ammonia [final concentration 0·5% w/w, (0·3 mol l−1)] and stored at 4, 14 and 25°C. Bacterial counts were determined at regular intervals on xylose lysine deoxycholate agar (XLD), and XLD supplemented with ampicillin (0·1 mg ml−1) and arabinose (0·6 mg ml−1) for Salmonella and cefsulodin-irgasan-novobiocin agar (CIN) and CIN supplemented with ampicillin and arabinose for Y. enterocolitica. The pH of the control-, urea- and ammonia-treated samples ranged from 7·1 to 7·7, 8·8 to 8·9 and 8·0 to 8·3, respectively. Salmonella D4 values ranged from 2·71 to 21·29 days, D14 values from 2·72 to 11·62 days and D25 values from 1·76 to 6·85 days. The equivalent D values ranges for the Y. enterocolitica strains were 3·7–19·23, 1·8–16·67 and 1·63–7·09 days, respectively. Treatment significantly (P < 0·01) affected D values with control > ammonia > urea, as did incubation temperature; 4 > 14 > 25°C.

Conclusions

Urea and to a lesser extent ammonia may be used to disinfect Salmonella- and/or Y. enterocolitica-contaminated pig slurry, decreasing the storage time required while increasing its fertilizer value.

Significance and Impact of the Study

This study presents data supporting the treatment of pig slurry to kill important zoonotic agents, thereby reducing environmental contamination, cross-infection of other animals and decreasing zoonotic disease in the food chain.

Introduction

Healthy pigs may carry human pathogenic Salmonella spp. and Yersinia enterocolitica. In 2009, there were 108 614 confirmed cases of salmonellosis in the European Union and fresh pork meat contamination rates of up to 13·7% were reported (Anon. 2011). The corresponding figures for Y. enterocolitica were 7597 confirmed cases with up to 48% of fresh pork products contaminated. Furthermore, as isolates from pork products and humans are often indistinguishable, pigs are considered to be a primary source of these pathogenic agents.

Salmonella enterica is a common cause of infectious food poisoning with up to 25% of human salmonellosis cases attributed to the consumption of contaminated pork or pork products (Berends et al. 1996; Van Pelt and Valkenburgh 2001). Salmonella carriage in pigs is seldom associated with clinical disease. Approximately 90% of confirmed yersiniosis cases are foodborne, with pork being the most important source (Mead et al. 1999; Nesbakken 2005). Yersinia enterocolitica is classified into six biotypes. Biotypes 1B, 2, 3, 4 and 5 cause disease in humans while biotype 1A is considered to be nonpathogenic (Nesbakken 2005). In addition, there are approximately 60 different Y. enterocolitica serotypes, 11 of which are associated with illness in humans (Bottone 1997). Pathogenic serotypes O:3, O:8, O:9 and O:5,27 are frequently found in porcine slurry. The most common Y. enterocolitica bioserotypes found in pigs are 4, O:3 and 2, O:5,27 (Poljak et al. 2010).

Slurry is composed of faeces, urine, uneaten food and bedding from intensively farmed animals. In Europe, it is usually stored in dedicated tanks located below or adjacent to the animal houses until field conditions are suitable for spreading on farm land. Pig slurry is a nutrient-rich soil improver and is commonly reused to fertilize agricultural land. However, the spread of Salmonella- and Y. enterocolitica-contaminated slurry on fields represents a hazard to other animals and ultimately humans (Cote et al. 2006). Treatment before application to land would decrease the risk of disease transmission to other animals as many bacterial zoonoses survive for extended periods of time in soil (Nicholson et al. 2005; Bolton et al. 2011).

Storage is currently the only common pig slurry disinfection practice. It is currently recommended that slurry be stored for 60 days in summer and 90 days in winter before application to grazing land (Park et al. 2002). Furthermore, a 30-day period before grazing should be observed and then used for adult or nonsusceptible animals. However, despite these recommendations, many farmers spread manure and slurry straight onto the land after removal from the animal shed or associated tank because of inadequate storage capacity or for convenience (Smith et al. 2000, 2001).

While several studies suggest that storage for a minimum of 90 days will render it free of pathogens (Guan and Holley 2003), the bactericidal effect is influenced by a range of parameters including initial level of contamination, storage temperature, pH, chemical composition of the slurry and the other microbes present. Thus, other studies have reported survival well in excess of 90 days (Strauch and Ballarini 1994; Hutchinson et al. 2005). Furthermore, minimum storage recommendations may not be observed and current practices are often not effective in eliminating bacterial zoonoses (Cardinale et al. 2004).

Chemical disinfection has been applied to stabilize, dry and disinfect animal waste (Himathongkham and Riemann 1999; Park and Diez-Gonzalez 2003; Vinneras et al. 2003). When selecting the chemical treatment, it is important to consider both the anti-microbial effectiveness and the agronomic value of the end product (Vinneras et al. 2003). Despite concerns about ammonia emissions and the environmental impact, low-level treatments with urea and aqueous ammonia have been suggested to be the most suitable animal waste disinfectants, and their use has been validated for bovine manure (Ottoson et al. 2008).

Although ammonia is toxic for most organisms, the exact mechanism of microbial destruction is unknown. It has been speculated that NH3 crosses the cell membrane causing rapid alkalization of the cytoplasm. Urea is quickly converted to ammonia and carbonate, both of which have antimicrobial activity (Park and Diez-Gonzalez 2003). The antimicrobial effect of inline image is probably the result of chelation of critical ions such as Ca2+ and Mg2+. Himathongkham et al. (1999a) reported ammonia to be effective in killing Escherichia coli O157:H7 in bovine manure and found the accumulation of free ammonia to be an important factor in the effectiveness of this treatment. Diez-Gonzalez et al. (2000) and Park and Diez-Gonzalez (2003) also reported that the carbonate anion had potent anti-E. coli and anti-Salmonella properties in cattle manure. However, there are no data currently available on their specific application for the removal of Salmonella and Y. enterocolitica form porcine slurry on pig farms. Indeed, survival data for Y. enterocolitica in livestock wastes are also lacking (Guan and Holley 2003). The aims of this study were therefore to investigate Salmonella and Y. enterocolitica survival and to evaluate urea and aqueous ammonia as chemical disinfectants for pig slurry specifically targeting the two most relevant human pathogens found in this animal waste.

Materials and methods

Pig slurry

Pig slurry was collected from finisher pens at a pig farm in County Cavan (Ireland) and stored at 2°C for no more than 24 h.

Bacterial cultures

Four cocktails (A, B, C and D), each containing one wild-type organism and one genetically modified organism, were used in these experiments. Cocktail A consisted of two Y. enterocolitica strains, bioserotypes 4, O:3 [National Collection of Type Cultures (NCTC), Porton Down, UK] and 1A, O:6,30 (NCTC). Cocktail B consisted of two Y. enterocolitica strains, Y. enterocolitica, bioserotype 2, O:5,27 (DSMZ, Braunschweig, Germany) and bioserotype 1A, O:6,30 (NCTC). Cocktail C consisted of two Salmonella spp., Salmonella Derby and Salmonella Typhimurium DT19. Cocktail D consisted of Salm. Derby and Salmonella Anatum. All isolates were previously isolated from pork sources.

Green fluorescent protein plasmid

To identify individual strains, one strain from each cocktail pair was genetically modified by insertion of the plasmid, pGLO (Bio-Rad Laboratories, Hercules, CA, USA) encoding for green fluorescent protein (GFP) and for resistance to ampicillin. The plasmid was inserted by transformation using heat shock in accordance with the pGLO Bacterial Transformation Kit (Bio-Rad Laboratories).

Inoculum preparation

One bead of each strain was aseptically transferred to 30-ml brain heart infusion broth (BHI; Merck, Darmstadt, Germany) and incubated at 30°C for 24 h. Following incubation, a 1-ml aliquot from each culture was transferred to 100-ml BHI and incubated for a further 18 h at 30°C. Each culture was then centrifuged at 3000 g for 10 min at 4°C. The recovered pellet was washed three times and resuspended in maximum recovery diluent. These suspensions were combined to form their respective cocktail inocula. Each cocktail inoculum had a final concentration of between 106–108 CFU ml−1.

Inoculation, disinfection, sampling and enumeration

Approximately 150 g of pig slurry was added to each of 27 stomacher bags 24 h prior to inoculation. All bags were then inoculated with cocktail A to a final concentration of 106–108 CFU ml−1. Eighteen bags were treated by adding 3 g of urea to nine bags to give a final concentration of 2% (w/w). The remaining nine bags were treated with 3-ml ammonia (aqueous, 28%) to a final concentration of 0·5% (w/w). Three urea- and three ammonia-treated bags were then incubated at each of 4, 14°C (corresponding to average winter and summer temperatures in Ireland) and 25°C (average temperature inside a pig finisher house). This procedure was repeated for cocktails B, C and D. Triplicate controls with no chemical treatment were also kept at each temperature.

Samples were removed periodically (days 0, 2, 6, 12, 19, 28 and 49) and analysed for the target organisms. Briefly, to enumerate surviving Salmonella cells, serial dilutions were prepared and plated onto xylose lysine deoxycholate agar (XLD; Merck) and also onto XLD supplemented with ampicillin (0·1 mg ml−1) and arabinose (0·6 mg ml−1) and incubated at 37°C for 24 h. To enumerate surviving Y. enterocolitica, serially diluted samples were plated onto cefsulodin-irgasan-novobiocin agar (CIN agar; Oxoid, Basingstoke, UK) and also onto CIN supplemented with ampicillin (0·1 mg ml−1) and arabinose (0·6 mg ml−1) and incubated at 30°C for 24 h.

Calculation of D and z values and statistical analysis

Each experiment was performed in duplicate and repeated three times. D values were obtained using regression analysis, and the t-test was used to compare means (Genstat; Rothamsted Experimental Station, Harpenden, UK).

Results

Prior to inoculation, the pig slurry was tested for the presence of naturally occurring Salmonella and Y. enterocolitica. All samples were negative.

At 4°C, the D values (D4 values) of Salm. Anatum, Salm. Derby and Salm. Typhimurium in the control slurry samples were 21·29, 16·66 and 17·54 days, respectively. The equivalent D values in urea- and ammonia-treated samples were 6·90, 4·02 and 2·71 days and 12·35, 12·50 and 11·11 days, respectively (Table 1). There was no significant difference (P > 0·01) between the different serotypes within a given treatment. However, for each serotype, the control D values were significantly higher (P < 0·01) than the ammonia-treated samples, which were in turn significantly greater (P < 0·01) than those obtained with urea-treated slurry.

Table 1. D4, D14 and D25 values (days) for Salmonella Anatum, Salmonella Derby and Salmonella Typhimurium DT19 in porcine slurry treated with urea and ammonia
Temperature (°C)OrganismnControlUreaAmmonia
R2SED valueaR2SED valueR2SED value
  1. a

    x/y: the first letter (x) indicates significant differences at the 1% level between different serotypes while the second letter (y) indicates significant differences at the 1% level between the different treatments.

4Salm. Anatum360·940·00421·29a/a0·870·0286·90a/c0·960·02012·35a/b
4Salm. Derby360·970·00516·66a/a0·970·0164·02a/c0·940·01012·50a/b
4Salm. Typhimurium DT19360·990·00317·54a/a0·920·0232·71a/c0·860·01611·11a/b
14Salm. Anatum360·990·00411·62a/a0·90·0483·23a/c0·910·0315·10a/b
14Salm. Derby360·980·0109·35a/a0·860·0322·72a/c0·910·0147·35a/b
14Salm. Typhimurium DT19360·930·01210·20a/a0·870·0152·80a/c0·930·0158·06a/b
25Salm. Anatum360·900·0186·71a/a0·880·0141·76a/b0·850·0703·16a/b
25Salm. Derby360·970·0193·57a/a0·780·0922·70a/b0·960·0272·73a/b
25Salm. Typhimurium DT19360·940·0156·85a/a0·880·0251·80a/c0·850·0312·85a/b

A similar pattern was observed at 14 and 25°C. D values at 14°C (D14 values) for control-, urea- and ammonia-treated samples ranged from 9·35 to 11·62, 2·72 to 3·23 and 5·10 to 8·06 days, respectively (Table 1). D values at 25°C (D25 values) for control-, urea- and ammonia-treated samples ranged from 3·57 to 6·85, 1·76 to 2·70 and 2·73 to 3·16 days, respectively (Table 1).

At 4°C, the D values for Y. enterocolitica O:3/biotype 4, O:5,27/biotype 2 and O:6,30/biotype 1A in the control slurry samples were 16·95, 18·52 and 19·23 days, respectively. The equivalent D values in urea- and ammonia-treated samples were 3·77, 4·02 and 3·7 and 14·92, 11·76 and 13·89 days, respectively (Table 2). There was no significant difference (P > 0·01) between the different strains within a given treatment. However, for each serotype/biotype, the control D values were significantly higher (P < 0·01) than the ammonia-treated samples, which were in turn significantly greater (P < 0·01) than those obtained with urea-treated slurry, except for Y. enterocolitica O:3/biotype 4 where there was no significant difference (P > 0·01) between the control- and ammonia-treated samples.

Table 2. D4, D14 and D25 values (days) for Yersinia enterocolitica 4, O:3, 2, O:5,27 and 1A, O:6,30 in porcine slurry treated with urea and ammonia
Temperature (°C)BioserotypenControlUreaAmmonia
R2SED valueaR2SED valueR2SED value
  1. a

    x/y: the first letter (x) indicates significant differences at the 1% level between different serotypes while the second letter (y) indicates significant differences at the 1% level between the different treatments.

44, O:3360·910·00816·95a/a0·950·023·77a/b0·930·01214·92a/a
42, O:5,27360·990·00218·52a/a0·970·0244·02a/c0·840·02311·76a/b
41A, O:6,30360·970·00419·23a/a0·960·0203·7a/c0·940·00813·89a/b
144, O:3360·940·01411·24a/a0·980·0312·92a/b0·900·0176·67a/b
142, O:5,27360·900·00916·67a/a0·920·0861·80a/b0·880·0613·06a/b
141A, O:6,30360·980·0109·62a/a0·920·0122·80a/b0·930·0314·41a/b
254, O:3360·870·0237·09a/a0·800·0761·69a/c0·930·0422·70a/b
252, O:5,27360·960·0332·56b/a0·760·1041·77a/b0·810·0941·95a/b
251A, O:6,30360·980·0183·97b/a0·770·1161·63a/b0·950·0422·81a/b

At 14°C, D values of 11·24, 16·67 and 9·62 days (control slurry samples), 2·92, 1·80 and 2·80 days (urea-treated samples) and 6·67, 3·06 and 4·41 days (ammonia-treated samples) were obtained for Y. enterocolitica O:3/biotype 4, O:5,27/biotype 2 and O:6,30/biotype, respectively (Table 2). As before, there was no significant difference (P > 0·01) between the different strains within a given treatment, and the control sample D values were significantly higher than the urea- and ammonia-treated samples, which were statistically similar.

The D values recorded at 25°C for Y. enterocolitica O:3/biotype 4, O:5,27/biotype 2 and O:6,30/biotype 1A were 7·09, 2·56 and 3·97 days (control slurry samples), 1·69, 1·77 and 1·63 days (urea-treated samples) and 2·70, 1·95 and 2·81 days (ammonia-treated samples), respectively (Table 2). There was no significant difference (P > 0·01) between the different strains within a given treatment except for the control samples where O:3/biotype 4 D values were significantly higher than the D values observed with the 2 other isolates. The control sample D values were significantly higher than the urea- and ammonia-treated samples, which were statistically similar, except for O:3/biotype 4 where the urea D value was significantly lower than the ammonia-treated samples.

Overall incubation temperature also significantly effected D values with 4 > 14 > 25°C, and the pH of the control-, urea- and ammonia-treated samples over the 49 days of the experiment ranged from 7·1 to 7·7, 8·8 to 8·9 and 8·0 to 8·3, respectively.

Discussion

The application of animal wastes such as pig slurry as natural fertilizers is an important risk factor in the occurrence and dissemination of pathogens in animal herds (Veling et al. 2002; Cardinale et al. 2004). Zoonotic agents such as Salmonella and Y. enterocolitica may be significantly reduced or eliminated by several different methods including treatment with lime, ammonia or urea, composting or the immediate ploughing and harrowing of amended soil (Boes et al. 2005). Lime treatment is widely used during outbreak situations but presents mixing as well as health and safety issues and has no fertilizer value. Composting is effective but requires a dedicated composting facility and time. Ploughing and harrowing may not be suitable if the slurry is spread on grassland. At low concentrations, urea and ammonia treatments are safe, economical (€2–4 per tonne), improve the agronomic value of the slurry and are environmentally sustainable (Ottoson et al. 2008).

This study demonstrated the decline of zoonotic agents in pig slurry over time as has been demonstrated in several other studies (Placha et al. 2001). In untreated samples, the decline was relatively slow, with D values of up to 21·29 and 19·23 days for Salmonella spp. and Y. enterocolitica, respectively. This may have been due to the reduced concentrations of carbonate anions and ammonia found in animal wastes at pH values below 8·5 (Park and Diez-Gonzalez 2003). In the treated samples, pathogen reduction was significantly faster, as expected, as the concentrations of urea (0·33 mol l−1) and ammonia (0·3 mol l−1) were well above the minimal antibacterial concentrations reported by Park and Diez-Gonzalez (2003).

The Salmonella D values ranged from 16·66 to 21·29 days at 4°C and from 3·57 to 6·85 days at 25°C in untreated pig slurry. These values were similar to those reported by Arrus et al. (2006); D4 values of 22 to 60 days and D25 values of 8–19 days and the winter (14·9 days) and summer (25·6 days) D values recorded by Hutchinson et al. (2005) also in pig slurry.

Our study demonstrated an average fourfold (urea) and 1·6-fold (ammonia) reduction in D values for Salmonella regardless on species or storage temperature. The corresponding decreases for Y. enterocolitica were 3·5- and 1·7-fold. In a similar study on bovine manure, Ottoson et al. (2008) reported an average 3·7-fold decrease for Salmonella with urea and a 26-fold decrease with ammonia treatment. The latter was attributed to the relatively high pH of 9·7 obtained in bovine manure treated with ammonia (0·5%, v/v). Ammonia has a pKa of 9·3, and the more alkaline the conditions the higher the concentration of free ammonia (NH3). This uncharged form of ammonia may cross bacterial membranes damaging cells by rapid alkalinization of the cytoplasm or through a decrease in intercellular K+ concentration (Kadam and Boone 1996; Park and Diez-Gonzalez 2003). In our studies, the pH of the pig slurry was approximately 8·9 and 8·2 after urea and ammonia treatments, respectively. The higher pH of the urea-treated samples was attributed to the production of ammonia and carbonate from urea hydrolysis. While significant D-value reductions were achieved with these treatments, a higher ammonium concentration could have been achieved by raising the pH using additional treatments such as KOH (Allievi et al. 1994).

Temperature significantly effected D values with 4 > 14 > 25°C. Previous studies have also reported temperature as an important factor influencing the decline of Salmonella in animal wastes (Placha et al. 2001; Guan and Holley 2003) including pig slurry (Nicholson et al. 2005; Arrus et al. 2006). In each case, the colder temperatures were more protective. This was attributed to several factors, including increased competition from indigenous micro-organisms (Himathongkham et al. 1999a,b), loss of nitrogen (Pratt et al. 2002) and the increased generation and accumulation of ammonia (Panetta et al. 2005) at the higher temperatures.

In conclusion, Salmonella (Hutchinson et al. 2005) and Y. enterocolitica (Letourneau et al. 2010) are normally present in pig slurry at levels of up to 103 and 105 CFU ml−1, respectively. Without treatment it would require 64-day storage at 4°C (worst case scenario) to eliminate Salmonella. This could be reduced to 37 and 21 days using ammonia and urea treatments, respectively. Similarly, the storage time required to eliminate Y. enterocolitica could be reduced from 96 to 70 days (ammonia) and 19 days (urea).

Although ammonia is a potent antimicrobial, its volatility and toxicity would suggest it is not suitable as a direct treatment of pig slurry. However, urea, which is naturally present in urine and approved for use as a fertilizer and feed supplement, could be readily applied as a source of antimicrobial ammonia and carbonate anions. Given the current limitations on storage capacity, urea treatment of pig slurry should be considered as part of food safety good farming practices (GFP) to reduce the pathogen recycling risks associated with using porcine waste as a fertilizer.

Acknowledgements

This project is funded by the Sixth Framework Programme for Research, Technological Development and Demonstration Activities, for the Integrated Project Q-PORKCHAINS FOOD-CT-2007-036245.

Ancillary