Sampling was carried out over an 18 month period. In all 405 samples were taken at the randomly selected single (out of seven) abattoir. For the oyster cut 25 - 50 g trimming was removed from the 50:50 product tray (50:50 is a half fat half pork trimming from the oyster cut) and transferred to a sterile jar (Trafalgar Scientific (CON7764) Leicester, UK). For faecal and caecal sampling the viscera was removed from the pig’s abdominal cavity and following separate incisions into the rectum and caecum approximately 30 - 50 g of sample was removed and placed into sterile jars. Carcass swabs were taken from four EU approved swabbing points identified on the pig’s carcass, jowl, belly, back and ham (EU directive 471/2001/EU) using EnviroSponge™ USDA approved cellulose sponges [Biotrace International (BP-133ES) Bridgend, UK]and placed in sterile bags provided with the sponge and labelled.

All samples were examined for the presence of Salmonella according to ISO 6579. Briefly 25 g of sample was transferred into 225 ml ISO Buffered peptone Water [BPW, Oxoid (CM1049) Basingstoke, UK], The sample and ISO BPW was transferred to a stomacher bag [Seward (BA6141) Worthing, West Sussex, UK] and stomached for 1 min (Seward Tekmar Stomacher 400, Seward, Worthing, West Sussex, UK). The sample was incubated at 37˚ for 18 - 20 hours. Controls comprised S. Nottingham NCTC 7832 (positive control) and Esche- richia coli NCTC10418 (negative control). There followed selective enrichment in Rappaport-Vassiliadis Soya (RVS) broth [Oxoid (CM0866) Basingstoke, UK], plating onto selective Xylose Lysine Decarboxylase [Oxoid (CM0469) Basingstoke, UK] and Brillant Green Agar [Oxoid (CM0263) Basingstoke, UK] followed by performance of confirmatory tests on suspect colonies.

All suspect Salmonella were sent to Agri-Food & Biosciences Institute (AFBNI), Newforge Lane, Belfast, Northern Ireland for serotyping. All resulting Salmonella were subsequently sent to The Laboratory of Enteric Pathogens, HPA, Collindale, UK to be phage typed using a standard set of the typing phage and following the Collindale scheme.

Resistance profiling used BSAC’s disc diffusion method (http://www.bsac.org.uk/wp-content/uploads/2012/02/Version-8-January-2009.pdf) according to established protocols [21] . All S. Rissen, were tested for their in vitro sensitivity to 16 different antibiotics. The antibiotics and their concentrations were as follows; Amikacin 30 µg (AK), Ampicillin 10 µg (AMP), Amoxycillin & Clavulanic Acid 30 µg (AMC), Apramycin 15 µg (APR), Cefotaxime 30 µg (CTX), Ceftazidime 30 µg (CAZ), Chloramphenicol 10 µg (C), Ciprofloxacin 1 µg (CIP), Compound Sulphonamides 300 µg (Su), Furazolidone 15 µg (FR), Gentamicin 10 µg (Cn), Nalidixic Acid 30 µg (NA), Neomycin 10 µg (N), Streptomycin 25 µg (S), Sulphamethoxazole/trimethoprim 25 µg (SXT) and Tetracycline 10 µg (T). Minimum inhibitory concentration (MIC) tests were carried using M.I.C. Evaluators (Oxoid, Basingstoke, UK) where resistance was identified. Multidrug resistance was defined as having a resistance pattern to five or more antibiotics.

PFGE was performed according to the CDC PulseNet protocol [22] with some minor modifications. Salmonella isolates were streaked onto Tryptone Soya Agar [TSA, Oxoid (CM0131) Basingstoke, UK] and incubated at 37˚C for 18 - 20 hours. Salmonella isolates from TSA were resuspended in 2 ml of cell suspension buffer (CSB) (100 mM Tris: 100mM EDTA, pH 8.0) creating at attenuance of approximately 1.35 at λ 610 using a spectrophotometer. 10 µl of Proteinase K (20 mg/ml stock) [Sigma-Aldrich (P6556) Pole, Dorset, UK] was added to 200 μl of Salmonella suspension and gently mixed. 200 μl of molten 2% SeaKem Gold agarose [Cambrex (BMA 50150) Wokingham, Berkshire, UK] was then added to this suspension. S. Branderup H9812 was used as the reference strain. The bacterial/agarose mixture was then dispensed into PFGE disposable moulds [BioRad (170-3713) Hemel Hempstead Hertfordshire, UK] and allowed to solidify for approximately 10 minutes. The freshly moulded plugs were then transferred to tubes containing 1.5 ml of CSB and stored at 4˚C. For cell lysis the plugs were placed in a 1.5 ml eppendorf containing 1 ml Cell Lysis Buffer-CLB (50 mM EDTA, pH 8, 1% sarcosyl) and 40 μl of Proteinase K solution (20 mg/ml) and incubated in a shaking incubator at 54˚C for 2 hours. The plugs were then washed in series of steps. Firstly, in a Polypropylene Super Clear™ 50 ml sterile tube, [Sarstedt (SLS8106) Leicester, UK] containing 10 - 15 ml of deionised waterin a shaking incubator at 50˚C for 10 minutes and following subsequent washing steps were placed in fresh TE buffer and stored at 4˚C until restriction digestion. The restriction enzyme Xba1 [Sigma-Aldrich (R7260) Poole, Dorset, UK] was first diluted 1:10 with dilute SuRE Cut buffer H [11417991001; Roche, UK] and 100 µl of this restriction mixture were added the micro-centrifuge tube containing the agarose plugs. Both sample and control plug slices were gently mixed and incubated for two hours at 37˚C.

A 1% agarose gel was prepared using SeaKem® Gold agarose [Cambrex (BMA50150) Woking, Berkshire, UK]. The gel plugs were then overlaid with a small quantity of molten SeaKem® Gold agarose [Cambrex (BMA50150) Woking, Berkshire, UK] and allowed to solidify. PFGE was conducted in a CHEF DR II chamber (BioRad, Hertfordshire, UK) containing fresh 0.5 M Tris Borate EDTA (TBE) running buffer [Gibco, (15581-044) Paisley, UK] under the following conditions: Initial switch time 2.2 s; Final switch time 63.8 s; Start Ratio 1.0; Voltage 200 V; Run time 19 - 20 hours at 14˚C. After staining for 20-30 min (in 40 µl ethidium bromide (EtBr) [Sigma-Aldrich (160539) Poole, Dorset, UK] and de-staining in sterile water for 60 min the gel image was then visualised and captured using a transillumination table (AlphaImagerTM 2200, Alpha Innotec, CA, USA). All PFGE gel images were converted to TIFF files prior to data analysis. Dendrograms and cluster analysis were created using Bionumerics 5.1 software (Applied Maths, Belgium). Similarity analysis was performed using Dice coefficient, and clustering was created using the unweighted pair group method with arithmetic means (UPGMA). When isolates were found to be indistinguishable upon XBa1 digestion and additional enzyme Spe1, was used to improve discrimination.

Plasmid profiling was carried out using a modified version of the Kado and Liu method [23] [24] . A single colony was inoculated into 3 ml of Brain Heart Infusion [Oxoid (CM1032) Basingstoke, UK] and following incubation at 37˚C for 18 - 20 hours 1ml of the culture centrifuged at 6600 G for two minutes. Cells were lysed by addition of 100 µl of lysis buffer (3% SDS in 50 mM Tris, pH 12.6). The sample was incubated at 56˚C for 30 minutes in a pre-heated heat block, followed by a chloroform extraction. The upper aqueous layer was transferred to a 1ml eppendorf. Loading dye (6 - 8 µl) was added to each sample and vortexed (sample stored at −20˚ for later analysis). Each sample (12 µl) was loaded onto a 0.7% agarose gel with a super-coiled ladder [Sigma-Aldrich (D5292) Poole, Dorset, UK] as molecular marker. Electrophoresis run-25 volts for four hours using a RunOne electrophoresis system (EmbiTec, CA, USA). The gel matrix was then stained using EtBr [Sigma-Aldrich (160539) Poole, Dorset, UK]. The resulting bands were viewed using a transilluminator (AlphaImagerTM 2200, Alpha Innotech Corp., USA).

All 29 S. Rissen isolates exhibited resistance to only a single antibiotic-tetracy- cline at a concentration of 10 µg/l. The MIC for Tetracycline was determined to be >256 µg/ml. No resistance to the fluoroquinolone antibiotic ciprofloxacin or to the third generation cephalosporin antibiotics cefotaxime and ceftazidime was observed.

Consistent and unambiguous Xba1-PFGE profiles were generated from 23 of the 29 isolates and analysis of these identified 19 that displayed a common profile consisting of 15 bands which ranged in size from 20 kb to 668 kb approximately. These 19 S. Rissen were designated cluster A and all presented a similarity higher than 96% - 97% using the DICE co-efficient. The remaining four S. Rissen displayed a common profile consisting of 16 bands which ranged in size from approximately 20 kb to 668 kb. These were designated cluster B and represented a similarity of approximately 97%. As both cluster A and B generated such similar Xba1 profiles further PGFE analysis using the restriction enzyme Spe1, as an alternative to Xba1, was conducted on representative S. Rissen from both cluster A and cluster B (Figure 1). Spe1 generated RFLP profiles that were identical for all S. Rissen typed. This analysis when combined with that of the Xba1 profiles and analysed using the BioNumerics 5.1 software (Applied Maths, Belgium) indicated that all the S. Rissen isolates from the pig abattoir were a clonal match. No further genetic differences were observed.

All 29 S. Rissen isolates were observed to have a low molecular weight band plasmid of approximately 4.650 kbp. A second low molecular weight plasmid of approximately 2.750 kb in size was found to be present in three out of the 29 isolates (Figure 2). All three of these S. Rissen that harboured the 2.750 kbp plasmid were recovered on a single sampling visit and were isolated from pigs originating in NI. Three other S. Rissen isolates were recovered that did not harbour the 2.750 kbp low molecular weight plasmid (Figure 2). The three S.

Rissen isolates displaying two plasmids were not the same isolates as the three S. Rissen observed to contain the extra band Xba1-PFGE in cluster B. The identification of two plasmid profile types significantly enhanced the differentiation of S. Rissen that were genetically dissimilar. When the genotyping data obtained through plasmid profiling is combined with that obtained by PFGE three different S. Rissen genetic profiles can be observed.