Slov Vet Res 2010; 47 (2): 45-56 UDC 579.62:576.851.49:616-078:577.2:631.854 Original Scientific Article DETECTION OF SALMONELLA IN pOULTRY FAECES BY MOLECULAR MEANS IN COMpARISON TO TRADITIONAL BACTERIOLOGICAL METHODS Darja Kušar, Mateja Pate, Jasna Micunovic, Vojka Bole - Hribovšek, Matjaž Ocepek* Institute of Microbiology and Parasitology, Veterinary Faculty, Gerbičeva 60, 1115 Ljubljana, Slovenia Corresponding author, E-mail: matjaz.ocepek@vf.uni-lj.si Summary: Comparison of traditional (cultivation-dependent) and molecular (nucleic acid-based) bacteriological methods was performed to detect Salmonella in reference capsules containing quantified amounts of Salmonella enterica subsp. enterica serovars Panama, Typhimurium or Enteritidis and in poultry faeces that was naturally contaminated with Salmonella or Salmonella-negative but spiked with reference materials. Traditional techniques were performed according to ISO 6579 using different enrichment (MSRV, MKTTn and RVS, respectively) and isolation plating media (XLD, BGA and Rambach agar, respectively). Molecular detection was preceded by the pre-enrichment step. Detection efficiency of two DNA isolation kits, namely High Pure foodproof I Kit (Roche Diagnostics, Germany) and QIAamp DNA Stool Mini Kit (Qiagen, Germany), in combination with classical and real-time PCR assay was compared. Results showed that traditional and molecular detection of Salmonella was unambiguous for reference control capsules, but was hindered for faecal samples. RVS medium was less appropriate than MSRV and MKTTn. Combination of MKTTn with Rambach agar plates generated the highest number of positive results with traditional approach. However, recommendation of using the semisolid MSRV medium was confirmed as it enabled detection of Salmonella in high proportion of samples, which was the least variable depending on the selection of isolation plating media. In contrast to culture-based methods, the molecular approach, especially a combination of High Pure foodproof I Kit and real-time PCR assay, enabled successful detection in all Salmonella-positive samples and should therefore be considered an important supplement to traditional protocol for Salmonella detection in foodstuffs. Key words: Salmonella; foodstuffs; faeces; detection; cultivation; molecular; PCR Introduction The routine microbiology laboratories for detection of different bacterial pathogens are complementing traditional diagnostic assays with continually evolving molecular techniques as they are not negatively affected by the presence of growth inhibitory compounds and enable rapid detection (1-3). Surveillance of alimentary zoonoses, diseases that are transmitted from animals to humans through food, and early detection of their causative agents in food producing animals and their environment are very important for the assurance of safe food. Food safety is a growing public health issue, since it was Received: 27 January 2010 Accepted for publication: 23 June 2010 estimated that up to 30 % of the population in industrialized countries is suffering from foodborne illnesses (4). Salmonelloses are the second most frequently reported human zoonoses in the European Union and can cause relatively vast economic damage due to chronic effects of the infections (5). The common reservoir of Salmonella is the intestinal tract of animals, however they can be detected in a wide variety of foodstuffs and food ingredients (5). Animal-to-human transmission occurs when bacteria are introduced into the food preparation process or through direct contact with infected animals and faecally contaminated environments. In-country laboratory-based monitoring of food-borne pathogens is being promoted (4). Traditional microbiological methods offer standardized proce- 46 D. Kusar, M. Pate, J. Micunovic, V. Bole - Hribovsek, M. Ocepek dures for their detection (e.g. ISO standards), but are time consuming and not always compatible with short-time-to-result demand. Therefore, food microbiology aims for supplementation of classical methods with molecular techniques based on detection of mi-crobial nucleic acids in foodstuffs, which shorten the analysis time and lower the limit of detection. It was shown previously that the polymerase chain reaction (PCR) has a great potential to speed-up the detection of Salmonella in food (6) and can be performed in a manner equivalent to the standard ISO 6579 culture method, which is set as the golden standard for Salmonella detection in food and feedstuffs (7,8). The main objective of our study was to evaluate the detection efficiency for different contamination levels of Salmonella spp. in the presence of competitive microorganisms. As a complex matrix, poultry faeces was selected for the starting material. The use of molecular methods polymerase chain reaction (PCR) and real-time PCR was compared to traditional, cultivation-dependent bacteriological methods. Materials and methods Reference materials (RMs) and poultry faeces were used. The RMs consisted of gelatin capsules containing a quantified amount of sub-lethally injured Salmonella strains of serovars Panama (SPan), Typhimu-rium (STM) or Enteritidis (SE) as spiked spray dried milk prepared by the Community Reference Laboratory (CRL) for Salmonella (9). The levels of contamination were SPan 5 (5 colony forming particles per capsule [cfp/caps]), STM 10 (10 cfp/caps), STM 100 (100 cfp/caps), SE 100 (100 cfp/caps) and SE 500 (500 cfp/ caps). Faeces, negative or positive for Salmonella spp., and reference capsules were stored at -20° C till use. We examined (i) 24 poultry faecal samples (numbered FC-1 to FC-24; 10 g each, negative for Salmonella spp.) in combination with a blank capsule (five samples) or a capsule containing STM (five samples STM 10 and four samples STM 100) or SE (five samples SE 100 and five samples SE 500), (ii) 20 faecal samples which were naturally contaminated with Salmonella and not spiked with capsules (numbered F-1 to F-20; 10 g each), and (iii) 10 control samples (numbered C-1 to C-10; no faeces added) consisting of two blank, two SPan 5, three STM 10 and three SE 100 capsules, respectively. In addition, two negative control samples were examined: procedure control (i.e. C-11; no faeces or capsule added) and negative faeces control (i.e. C-12; 10 g of faeces negative for Salmonella spp.). The sample outline is summarized in Table 1. Traditional bacteriological methods Cultivation-dependent detection of Salmonella was performed according to ISO 6579:2002 (10), including Amendment 1:2007 (11), and the instructions provided by CRL for Salmonella (9). In brief, detection involved the following stages: (i) overnight sample pre-enrichment in a nonselective broth medium BPW (Buffered Peptone Water), (ii) 24- (for the first isolation) and 48-hour (for the second isolation) enrichment in selective broth media MKTTn (Muller Kauffmann TetraThionate-novobiocin broth), RVS (Rappaport Vassiliadis medium with Soya) and MSRV (Modified Semi-solid Rappaport Vassiliadis medium; 11), (iii) isolation of colonies presumed to be Salmonellae on solid selective and differential plating media BGA (phenol red/Brilliant Green Agar), XLD (Xylose-Lysine-Deoxycholate agar) and R (Rambach agar; 12), and (iv) biochemical screening of Salmonella isolates on the confirmation media TSI (Triple Sugar/Iron agar), UA (Urea Agar) and LDC (1-Lysine decarboxylation medium). If colonies grown on the isolation media were not well separated, single colony isolation was performed on NA (Nutrient Agar) plates after 24-hour incubation at 37° C and followed by the aforementioned confirmation. For each of the samples from the three selective enrichment media, at least one individual colony, considered to be typical or suspect for Salmonella, was examined biochemically. If the selected colonies were not confirmed as Salmonella, maximum of five additional typical colonies were tested from the original isolation medium stored at 5° C. Sample was denoted with positive result if growth of Salmonella spp. was present at least on one of the isolation media. If not stated otherwise, media and reagents were prepared according to Annex B of ISO 6579:2002. Molecular methods Molecular detection of Salmonella involved the isolation of microbial DNA that was followed by Salmonella-specific PCR and real-time PCR assays. DNA was extracted from 1 mL of the pre-enrichment broths using two different commercial kits, namely the High Pure foodproof I Kit (Roche Diagnostics, Germany) and QIAamp DNA Stool Mini Kit (Qiagen, Germany) according to the manufacturers' instructions. The latter was not applied for samples devoid of faeces. Microbial DNA was subjected to PCR amplification using Salmonella genus-specific primers ST11 and ST15 (13) that were proved as appropriate Detection of Salmonella in poultry faeces by molecular means in comparison to traditional bacteriological methods 47 Table 1: Outline of the samples used for the study: control samples (C), naturally contaminated faecal samples (F), and faecal samples supplemented with capsules (FC). Sample name Faeces Capsule Sample name Faeces Capsule Sample name Faeces Capsule C-1 / blank F-1 pos / FC-1 neg SE 100 C-2 / SPan 5 F-2 pos / FC-2 neg blank C-3 / blank F-3 pos / FC-3 neg STM 100 C-4 / SE 100 F-4 pos / FC-4 neg SE 100 C-5 / STM 10 F-5 pos / FC-5 neg STM 100 C-6 / STM 10 F-6 pos / FC-6 neg SE 500 C-7 / STM 10 F-7 pos / FC-7 neg SE 500 C-8 / SE 100 F-8 pos / FC-8 neg blank C-9 / SE 100 F-9 pos / FC-9 neg SE 500 C-10 / SPan 5 F-10 pos / FC-10 neg SE 100 C-11 / / F-11 pos / FC-11 neg STM 10 C-12 neg / F-12 pos / FC-12 neg blank F-13 pos / FC-13 neg SE 100 F-14 pos / FC-14 neg STM 10 F-15 pos / FC-15 neg STM 100 F-16 pos / FC-16 neg blank F-17 pos / FC-17 neg STM 10 F-18 pos / FC-18 neg SE 500 F-19 pos / FC-19 neg STM 10 F-20 pos / FC-20 neg blank FC-21 neg STM 10 FC-22 neg SE 100 FC-23 neg SE 500 FC-24 neg STM 100 Legend: neg: faeces negative for Salmonella, pos: faeces positive for Salmonella, SPan: Salmonella Panama (SPan 5: 5 cfp/caps), STM: Salmonella Typhimurium (STM 10 and STM 100: 10 and 100 cfp/caps, respectively), SE: Salmonella Enteritidis (SE 100 and SE 500: 100 and 500 cfp/caps, respectively), blank: no cfp/caps for the confirmation of Salmonella-colonies obtained by the standard ISO 6579 culture method (14). Amplification was performed according to the optimized touchdown protocol as described previously (15). Real-time PCR was performed using the Light-Cycler foodproof Salmonella Detection Kit (Roche Diagnostics, Germany) according to the manufacturer's instructions. Briefly, a 20-^l reaction mixture was composed of foodproof Salmonella enzyme solution containing FastStart Taq DNA polymerase, internal amplification control (IC), master mix containing primers and hybridization probes specific for Salmonella DNA and Salmonella-specific IC, and 5 ^l of sample DNA, foodproof Salmonella positive control template or PCR-grade water as negative control. Amplification was performed by LightCy-cler 1.2 Real-Time PCR System (Roche Diagnostics, Germany). The inclusivity of foodproof Salmonella master mix for the Salmonella genus and exclusiv- ity for other genera was extensively tested by the manufacturer. Results Reference materials vs. faecal samples Detection of Salmonella in reference materials was unambiguous with no false negative or positive results regardless of the employed method (samples C in Tables 2 and 3). Both the traditional and the molecular methods in all the tested combinations were equally appropriate with detection limit of 5 cfp/ sample or lower. On the other hand, detection of Salmonella in samples containing poultry faeces was limited as it depended on the method type and the level of Salmonella contamination. Detection limit was impaired for traditional methods (above 10 cfp/ sample) in comparison to molecular methods (10 48 D. Kusar, M. Pate, J. Micunovic, V. Bole - Hribovsek, M. Ocepek cfp/sample or lower), since one sample of naturally contaminated faeces (F-2) and eight faecal samples containing reference capsules (all five samples supplemented with STM 10 [FC-11, FC-14, FC-17, FC-19 and FC-21], two samples with STM 100 [FC-3 and FC-5], and one sample with SE 100 [FC-1]) were denoted falsely negative but tested positive when molecular detection was performed (Table 3). To detail, Salmonellae from capsules STM 10 were detected by traditional methods only from the control samples, but not in samples containing faecal material. However, they were detected in all samples FC (100 %) when applying the molecular methods (particularly, real-time PCR in combination with High Pure food-proof I Kit) (Table 4). Salmonellae from samples FC supplied with higher cfp number of Salmonella Ty-phimurium (STM 100) or with Salmonella Enteri-tidis in equivalent cfp number (SE 100) were detected in marked proportions of samples (50 % or 80 %, respectively) by traditional bacteriological methods, but in all cases (100 %) when molecular approach was employed (Table 4). Salmonellae from SE 500 capsules with the highest Salmonella-contamination level were detected in all samples FC by both the traditional (with MSRV enrichment only) and the molecular approach (Table 4). Salmonellae from faecal samples F were detected in 95 % and in 100 % of cases applying traditional (particularly, MKTTn in combination with Rambach agar) and molecular (all combinations) methods, respectively (Table 4). To summarize, detection of Salmonella spp. in reference materials succeeded over the entire experimental range of contamination levels and did not depend on the method type, but was impaired in faecal samples when traditional approach was employed, enabling detection in 30 out of 39 Salmonella-positive faecal samples (19/20 for samples F and 11/19 for samples FC, respectively) (Table 3). No samples supplied with blank capsule and negative control samples tested falsely positive. Traditional bacteriological methods vs. molecular methods Results of cultivation-dependent detection of Salmonella in faecal samples after 24 and/or 48-hour incubation showed that MKTTn and MSRV selective enrichment media generated less falsely negative results than RVS (Tables 2 and 3). MKTTn enabled detection in 19 out of 20 Salmonella-positive samples F (19/20) and 10 out of 19 Salmonella-positive samples FC (10/19), MSRV in 17/20 and 9/19, and RVS in 14/20 and 3/19 samples, respectively. By traditional approach, 13 of 20 Salmonella-positive samples F and only 2 of 19 Salmonella-positive FC samples (FC-15 and FC-23) tested positive from all three enrichment media. All the rest tested positive from two (5/20 for samples F and 7/19 for samples FC) or only one (1/20 and 2/19, respectively) selective enrichment medium (Table 3). In the majority of faecal samples, detection failed from RVS enrichment regardless of the isolation medium, in particular with faecal samples FC that were contaminated with serovar Enteritidis (Tables 2 and 3). Although serovar Typhimurium was detected after RVS enrichment in 50 % of FC samples supplied with STM 100, detection of serovar Enteritidis of the same cfp number (SE 100) was completely absent (0 %) or markedly impaired (detection in 20 %) when supplied in higher cfp number SE 500 (Table 4). The highest number of true positive results (19/20 samples F and 9/19 samples FC, Table 2) was attributed to MKTTn when in combination with Rambach agar. However, the lowest number of positive results was also attributed to MKTTn, namely when it was combined with XLD or BGA isolation plating medium (6/20 or 0/20, respectively for faecal samples F; Table 2). Isolation of Salmonella colonies originating from different selective enrichment media was not affected by the selection of plating media with the above mentioned exception of MKTTn, which was likewise the only selective medium that generated higher number of positive cases when results obtained from all the three isolation plating media were combined in comparison to results obtained from individual isolation media (4 positive out of 5 samples FC supplied with SE 100 [4/5] for combined results vs. 3/5 for individual combinations of isolation media with MKTTn, respectively; Table 2). MKTTn and MSRV enrichments were comparably effective regarding the detection level, however only MSRV enabled detection of Salmonella in all samples FC supplied with SE 500 (Tables 2 and 3). That was the only case for faecal samples where detection by traditional approach was not limited. Results of molecular detection of Salmonella showed that it was successful, since the presence of Salmonella was confirmed not only in control samples but also in all faecal samples (detection level of 100 %; Table 4). However, differences were observed for samples FC regarding the procedure for DNA isolation (using the High Pure foodproof I Kit [protocol HP] or QIAamp DNA Stool Mini Kit [protocol S]) and Detection of Salmonella in poultry faeces by molecular means in comparison to traditional bacteriological methods 49 the type of PCR reaction (classical PCR or real-time PCR) (Tables 3 and 4). Protocol S enabled detection in more FC samples in comparison to protocol HP when using classical PCR detection (15/19 for S vs. 10/19 for HP), but protocol HP in more samples when performing real-time PCR (19/19 for HP vs. 15/19 for S) (Table 3). Protocol HP enabled detection of Salmonella in all samples FC, while protocol S generated false negative result for one sample supplemented with STM 10 capsule (FC-19) and for one with SE 100 (FC-1) (Table 3). Real-time PCR enabled detection in all samples FC, while classical PCR failed for three samples (FC-1, FC-19 and FC-22) (Table 3). The highest number of positive cases using the molecular approach for samples FC, where detection efficiency depended on the method type in con- trast to samples F, was obtained by protocol HP in combination with real-time PCR (19/19), followed by protocol S in combination with either classical or real-time PCR (15/15, however different samples were denoted as positive in four cases [FC-10, FC-11, FC-13 and FC-22] depending on PCR reaction type) (Table 3). The highest number of false negative results was obtained by protocol HP in combination with classical PCR that generated 10 positive cases out of 19 Salmonella-positive samples FC (10/19) (Table 3). When classical PCR reaction tested negative, in all but two such cases (samples FC-1 and FC-19 that failed to test positive when applying protocol S) real-time PCR tested positive (Table 3). When applying protocol S, two samples (FC-11 and FC-13) tested positive solely by classical PCR (Table 3). Table 2: Results of traditional detection of Salmonella: number of positive samples after 24 and/or 48-hour incubation in selective enrichment media MKTTn, RVS and MSRV as detected on individual isolation media BGA, XLD and R, respectively MKTTn RVS MSRV s e ° ^ BGA XLD R t» poses ° ^ BGA XLD R t» poses BGA XLD R t» poses o 5 Z m o 5 Z m o 5 Z m o 5 Z m 2 0 0 0 0 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 20 0 6 19 19 14 14 14 14 17 17 17 17 5 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 4 1 1 2 2 2 2 2 2 1 1 1 1 5 3 3 3 4 0 0 0 0 3 3 3 3 5 2 4 4 4 1 1 1 1 5 5 5 5