Fate of salmonellas and competing flora in meat sample enrichments in buffered peptone water and in Muller-Kauffmann's tetrathionate medium

Studies have been carried out in which growth patterns of a Salmonella sp. and competing micro-organisms, especially other Enterobacteriaceae, were followed during pre-enrichment in buffered peptone water (BPw) and subsequent selective enrichment in tetrathionate broth (TBB). Pre-enrichment cultures were inoculated with minced meat and three reference samples containing nalidixic acid-resistant salmonellas. Irrespective of their initial numbers in BPw, Enterobacteriaceae increased to 10⁸/ml or more. During incubation in TBB at 43C, numbers of lactose-positive Enterobacteriaceae decreased in most enrichments which resulted in a positive salmonella isolation, but remained constant in the majority of those that did not. Levels of lactose-negative Enterobacteriaceae did not decrease in most salmonella-positive tests, but did so in half of the salmonella-negative ones. In the salmonella-positive tests the numbers of salmonellas had increased to 10³–10⁷/ml in BPw and after transfer to TBB slowly reached 10⁴/ml or more. In all cases the numbers of salmonellas exceeded those of the competing flora on brilliant green agar (BGA). In the salmonella-negative tests the numbers of salmonellas had increased less in BPw and decreased in most of the TBB enrichments. In none of these negative tests did the numbers of salmonellas exceed those of the competing flora on BGA. Escherichia coli dominated in most of the salmonella-negative tests. The results suggest more influence of lactose-positive than lactose-negative Enterobacteriaceae on the detection of salmonellas. The effect of competing microorganisms seems to depend not only upon their initial numbers, but also upon the types that can interact with salmonellas during selective enrichment.     

A study of isolation procedures for multiple infections of Salmonella and Arizona in a wild marsupial, the quokka (Setonix brachyurus)

Rectal swabs and faeces were used in the regular sampling for salmonellas and Arizonas from a heavily-infected population of a marsupial, the quokka ( Setonix brachyurus ). The media used were strontium selenite A and strontium chloride B enrichment broths, with subculture onto modified bismuth sulphite agar and deoxycholate citrate agar. A study of sampling, enrichment, sub-culture and colony selection procedures produced an optimal scheme giving high yields but consistent with reasonable economy of time and materials. A three-swab sample was taken and inoculated into the two enrichment media, and with each enrichment subjected to three subcultures. The absolute efficiency of this procedure was greater than 80% (and confirmed by a serological method), compared with only 67% for a single swab in a single enrichment. Recovery of some serotypes depended on the media used; e.g. Arizonas could not be recovered satisfactorily from strontium chloride B enrichment. Faeces samples were found to be greatly superior to rectal swabs for detecting salmonellas and arizonas but they were less convenient in field studies. In a comparison of rectal swabs and faeces samples where the actual concentration of salmonellas was known, it was found that the efficiency of rectal swabs approached 100% if there were more than 10³ salmonellas/g faeces, but this declined to approximately 50% if there were 10²-10³ salmonellas/g faeces and only 25% if there were less than 10² salmonellas/g faeces. A new statistical procedure was introduced for comparing the number of isolations from two methods, and this should be of use in similar methodological studies.     

Experimental enzyme-linked amperometric immunosensors for the detection of salmonellas in foods

Two enzyme-linked amperometric immunosensors specific for salmonellas were developed as rapid methods for quantifying and detecting these organisms in pure cultures and foods. Both used alkaline phosphatase as the enzyme reporter molecule but one system used phenyl phosphate as the substrate followed by the electrochemical detection of phenol at a polarized platinum electrode. The other system incorporated an enzyme amplification step and relied on the electrochemical detection of a reduced mediator, ferrocyanide. Both assays were rapid (4 h) and specific and generated salmonella-dependent signals above 10⁴ cfu/ml (phenyl phosphate system) or 10⁵ cfu/ml (enzyme amplified system) in pure cultures and samples of several foods, although the results with beef samples showed considerable variation. Both systems were able to detect low (1–5 cfu/g or /ml) numbers of salmonellas in foods after non-selective (18 h) and selective (22 h) enrichment steps but four samples, out of 147, gave false positive results. False positive results were eliminated by reducing the enrichment steps to 6 h and 18 h respectively (90 samples).     

An improved ELISA method for the detection of Salmonella typhimurium

The applicability of enzyme-linked immunosorbent assay (ELISA) for the detection of salmonellas in foodstuffs was investigated. Several factors affecting the sensitivity of the ELISA, such as the type of protein used for plate post-coating, the method of antibody labelling, and accelerators for antigen-antibody and enzyme-substrate reactions, were studied. Labelling of the antibody with horseradish peroxidase and the use of o -phenylenediamine as substrate in the detection system were demonstrated to be most suitable for the enzyme assay.
Based on these findings, an improved ELISA method was developed for the detection of Salmonella typhimurium. The improved technique was able to detect as few as 5×10⁴-10⁵ cell/ml of salmonellas, and about 24 h were required to enrich the bacteria in food samples and to perform the test. With some modifications, the ELISA assay could reach a very high level of sensitivity and provide excellent repro-ducibility.     

Rapid and definitive detection of Salmonella in foods by PCR

The BAX^™ system for screening Salmonella is one of the first commercial PCR-based systems for the detection of food-borne pathogens. It is able to give a confirmed result within 28 h. There was 98·6% and 95·8% agreement between the BAX^™ system and conventional cultural analysis for the detection of Salmonella in artificially inoculated and uninoculated food samples, respectively. In both cases, the BAX^™ system generated more positive detections than cultural analysis. The speed of assay, ease of use and high specificity and sensitivity of the BAX^™ system for the detection of food-borne Salmonella make it an attractive method for routine food microbiology laboratories.     

Development of a 24 h screen to detect viable salmonellas in faeces

A 24 h screen to detect viable salmonellas in faeces was developed by studying growth dynamics of salmonellas and competing flora in combinations of enrichment media and artificially-inoculated pig faeces. Muller-Kauffmann tetrathionate (MK) broth, incubated overnight at 42°C, maintained the lowest ratio of salmonella: competing flora and identified all inoculated samples. A 4 h postenrichment in M broth plus novobiocin reduced the number of false-positive results in subsequent ELISAs. Adjusting the negative cut-off values and incubation time of the chromogenic substrate from that recommended in the ELISA instructions reduced the rate of false-positive results further and allowed the detection of 10³ salmonellas per ml in the presence of up to 10⁷ ml⁻¹ aerobic-competing cells. Suspension of faeces diluted 1 in 2 and 1 in 5, rather than 1 in 10 in MK broth did not necessitate further adjustments to the ELISA baseline values. The proposed screen protocol is an overnight incubation of faeces suspended 1 in 10 in MK broth, a 1 in 100 subculture into M broth plus 10 μg ml⁻¹ novobiocin (MbN) for 4 h, steam inactivation of MbN cultures and testing by ELISA, and can detect three salmonella cells per g faeces.     

DETECTION OF SALMONELLA SPP. CONTAMINATION IN RAW MATERIALS AND COSMETICS/PHARMACEUTICAL PRODUCTS USING THE BAXTM SYSTEM, A PCR-BASED ASSAY

Abstract A system utilizing the polymerase chain reaction (PCR), the BAX ^TM , was compared and validated against standard selective/enrichment assays to detect the presence of Salmonella spp. in artificially contaminated samples of raw materials and cosmetic/pharmaceutical products. After a 24 h incubation in lactose broth or lactose broth with Tween 20, the inoculated samples were analyzed both by the BAX ^TM system and by standard enrichment/selective methods. Standard enrichment assays required 5–7 days to confirm the presence and identification of Salmonella typhimurium , while the BAX ^TM system reduced the detection time to 30 h. The BAX ^TM system allowed a faster quality control evaluation of those raw materials and cosmetic/pharmaceutical formulations that require Salmonella spp. screening.     

Detection of low levels of specific Salmonella species by fluorescent antibodies and flow cytometry

The use of fluorescently-labelled monoclonal antibodies, with detection by multi-parameter flow cytometry, was investigated for the rapid detection of salmonellas in pure cultures. Accurate detection of specific Salmonella serotypes was demonstrated down to levels of below 10⁴ cells ml^-1 (within 30 min) and 1 cell ml^-1 (after 6 h non-selective pre-enrichment). This level of sensitivity was attainedeven in the presence of high levels of other bacterial species that would otherwise have interfered with the results. With combinations of different antibodies, each with a unique fluorescent label, simultaneous analysis for two species was possible.     

Sandwich capture ELISA by a murine monoclonal antibody against a genus-specific LPS epitope for the detection of different common serotypes of salmonellas

D.CHOI, R.S.W. TSANG AND M.H. NG. 1992. A sandwich capture ELISA based on a murine monoclonal antibody against a genus-specific epitope in the outer core region of the Salmonella lipopolysaccharide is described for the detection of different common serotypes of salmonellas. Four h broth cultures of seven standard and 24 wild strains of salmonellas were all detected by the capture ELISA while overnight broth cultures of 21 non-salmonella standard strains were all negative. The capture ELISA detected 1 ng/ml of Ra lipopolysaccharide, 10⁶/ml of a smooth wild strain of Salm. typhimurium , and 1120 cells of Salm. heidelberg after enrichment culture for 4 h.     

A note on the effect of the lactoperoxidase systems on salmonellas in vitro and in vivo

The lactoperoxidase system (LPS), a natural bactericidal system in milk, was investigated for its activity against salmonellas in vivo and in vitro. In acidified raw milk, in which the LPS was supplemented with an exogenous supply of H₂O₂, the numbers of salmonellas decreased rapidly. Different salmonella serotypes were affected to the same extent; rough strains, however, were more susceptible than smooth strains. When calves were fed on fresh milk, containing the LPS, and challenged with Salmonella typhimurium in doses of either 10⁹ or 10¹⁰, the clinical findings and salmonella excretion patterns were similar to those of control calves fed on heated milk. It was concluded that further studies, perhaps in the field, are necessary to evaluate LPS as a possible non-antibiotic system to control salmonellosis.     

Microbiological composition of raw milk from selected farms in the Camembert region of Normandy

Raw milk from 27 farms was sampled over 6 months for listerias, salmonellas, Yersinia enterocolitica and campylobacters. Total bacterial counts and somatic cell counts were measured. Lactococci, lactobacilli, dextran-producing leuconostocs, Brevibacterium linens , yeasts and moulds, Staphylococcus aureus and other Micrococcaceae, Pseudomonas , coliforms, Escherichia coli , enterococci, Clostridium perfringens and spores of anaerobic lactate-fermenting bacteria were also counted. Pseudomonas (2000 cfu ml⁻¹), lactococci (760 cfu ml⁻¹) and Micrococcaceae (720 cfu ml⁻¹) were the most numerous groups. Lactic acid bacteria were detected in all samples. Coliforms were present in most samples, but 84% of samples had counts <100 cfu ml⁻¹. Staphylococcus aureus was detected in 62% of milks, the average count was 410 cfu ml⁻¹. About 80% of supplies had ≤10 E. coli cfu ml⁻¹ and all samples had 1 Cl. perfringens cfu ml⁻¹. Two of the tested milks were positive for salmonellas (2·9%), four were positive for Listeria monocytogenes (5·8%), 25 for Yersinia enterocolitica (36%) and one for campylobacters (1·4%).     

RAPID DETECTION OF SALMONELLA TYPHIMURIUM IN FOOD SAMPLES USING A BIENZYME ELECTROCHEMICAL BIOSENSOR WITH FLOW INJECTION

A bienzyme (tyrosinase and horseradish peroxidase) electrochemical biosensor was developed for detection of Salmonella typhimurium, and evaluated for application in a flow injection system coupled with immunomagnetic separation for food samples. Parameters for immunomagnetic separation, enzymatic reaction, flow injection and electrochemical detection were determined using pure culture samples. The selectivity was tested in the presence of Listeria monocytogenes, Campylobacter jejuni and E. coli 0157:H7. The results showed a linear relationship for logarithmic values between peak current ratio and the cell number of S. typhimurium in the range of 10³ 10⁵ cfu/mL, with R²= 0.99. The detection limit of this method was 1.09 × 10³ cfu/mL for S. typhimurium and the detection time was 2.5 h. Samples of chicken carcass wash water and ground beef were used to evaluate the biosensor. The results demonstrated that this biosensor has a potential for rapid detection of different pathogens in various food samples.     

Comparison of classical isolation protocols with a 24 h screen to detect viable salmonellas in faeces

A 24 h screen which detects three viable salmonella cells per g of faeces was compared with classical isolation procedures for their ability to identify salmonella-positive samples from a pig rearing unit. The screen involved an overnight enrichment in Muller-Kauffmann tetrathionate (MK) broth, subculture for 4 h in M broth containing 10 μg ml⁻¹ novobiocin, followed by detection of the presence of salmonellas by BacTrace and Salmonella-tek ELISAs. The classical protocols were: (1) an overnight and 48 h incubation in MK or selenite cysteine broth; or (2) overnight incubation in buffered peptone water and 24 h subculture in Rappaport-Vassiliadis broth (BPW-RV). Salmonellas were isolated from the broth cultures on xylose lysine deoxycholate and brilliant green agars. Thirty four of 100 samples were positive for salmonellas but no single isolation protocol identified all of them. The best of the classical isolation protocols, 48 h incubation in MK broth, identified 27 (79%) of the 34 positive samples whilst the screen identified 26 (76%) of the 34 positive samples. False-positive results were obtained from all isolation protocols except BPW-RV.     

Viable but non-culturable salmonellas in soil

P.E. TURPIN, K.A. MAYCROFT, C.L. ROWLANDS AND E.M.H. WELLINGTON. 1993. An enzyme-linked immunosorbent assay (ELISA) and a microwell fluorescent antibody (FA) direct count method have been developed for the monitoring of salmonellas in soil. Both methods have a minimum detection level of ca 10⁶ cells per gram of soil. The FA direct count method gave a linear recovery for the inoculum range 10⁶–10⁹ cells per gram of soil. When monitored by plate counts the survival of salmonellas was greater in a sterile than in a non-sterile soil. Evidence was found for the production of viable but non-culturable salmonellas in non-sterile soil; plate counts dropped rapidly with time, but FA direct counts and ELISA remained level. The salmonella cells became progressively smaller and rounder with time. Dead salmonella cells introduced into soil rapidly disappeared.     

Use of impedance measurements to estimate numbers of antibiotic resistant Salmonella strains

Impedance detection times were compared with traditional plating methods for enumerating antibiotic-resistant strains of Salmonella stanley, Salm. thompson and Salm. infantis grown in laboratory medium and pork slurry. The correleation of log₁₀ counts of salmonellas with detection times was highly significant ( r = -0·96 for broth and r = -0·94 for slurries. The confidence limits (± og₁₀ 1·0 for broth and ± log₁₀ 1·65 for slurry) indicated that detection times could reliably be used as a rapid means of enumerating salmonellas when large numbers of counts of known strains are required for growth studies. Use of antibiotic-resistant strains also permitted their selective detection by impedance from the natural spoilage flora of pork slurry when the same antibiotics were incorporated in the detection medium.     

The use of a surface adhesion immunofluorescent (SAIF) method for the rapid detection of Yersinia enterocolitica serotype O:3 in meat

This study examined the attachment kinetics of Yersinia enterocolitica serotype O:3 to determine the optimum conditions for its isolation from meat enrichment systems using a novel surface adhesion technique. Minced beef was inoculated with Y. enterocolitica at an initial level of 10 cfu g⁻¹ and incubated at 25 °C in an enrichment broth. Yersinia was recovered from enriched samples on polycarbonate membranes by surface adhesion and enumerated using immunofluorescence microscopy. The surface adhesion immunofluorescence technique (SAIF) had a minimum detection limit of approximately 4·0–4·5 log₁₀ cfu ml⁻¹ and provided good correlation between the estimation of the numbers of Yersinia in the enrichment broth derived from plate counts on Yersinia Selective agar (CIN) and those determined by SAIF ( r ² ₌ 0·94; rsd ₌ ± 0·21). A derived regression equation of the SAIF count vs plate counts was used to predict Y. enterocolitica numbers in spiked meat samples stored at 0 °C for up to 20 d. The numbers as predicted by the SAIF method showed good correlation with counts derived by plating techniques ( r ² ₌ 0·78; rsd ₌ ± 0·42). The application of the SAIF technique for the rapid detection of Y. enterocolitica serotype O:3 from meat is discussed.     

Efficiency of different enrichment and isolation procedures for the detection of Salmonella serotypes in edible offal

Rapid detection systems for Salmonella in foodstuffs are currently being developed. However, existing standards still call for application of traditional methods employing pre-enrichment followed by selective enrichment and isolation. The efficacy of various methods was tested using 264 chicken and lamb organ meats. Pre-enrichment was carried out in Tryptone Soy Broth (TSB) and enrichment in Tetrathionate Brilliant Green Broth (TTB) at 37°C, Selenite Broth with Brilliant Green and Sulphapyridine at 37°C and 43°C, and Rappaport-Vassiliadis Broth (RV 10) at 42°C. The isolation media were Brilliant Green Agar (BGA), Deoxycholate Citrate Agar, Hektoen Enteric Agar (HEA) and Salmonella-Shigella Agar.
Enrichment in RV/42°C followed by isolation on BGA as recommended by ISO standard no. 6579 and enrichment in TTB/37°C followed by isolation in HEA, no longer recommended by that standard, produced the best results. Low percentages of positive samples and difficulties in detecting Salmonella are the result of interference by competing organisms (Enterobacteriaceae) and the number of salmonellas present after enrichment.
A total of 528 samples (TSB, eggs, lamb liver and chicken liver) were inoculated with Salm. enteritidis, Salm. kapemba and Salm. virchow , and the preceding experiment was repeated. All the TSB and egg samples tested positive, but the percentage of positive samples from the lamb and chicken liver was only 81–92%. Recovery of the salmonellas did not depend upon the method employed or the serotype inoculated but instead on interference by competing flora and the numbers of Salmonella present in the samples.     

Fate of salmonellas in the alimentary tract of chicks pre-treated with a mature caecal microflora to increase colonization resistance

Oral challenge with a strain of Salmonella kedougou , resistant to nalidixic acid, gave a time-course for salmonella survival in the the alimentary tract of chicks pre-treated with a caecal culture from an eight-week-old bird to increase colonization resistance. In untreated, control chicks, salmonella colonization of the caeca resulted in counts of > 10⁶/g within 48 h and a mean generation time of 1·6 h. With treated birds, however, the salmonellas failed to multiply in the caeca and decreased to a low level over a 48-h period, thus suggesting a mainly bacteriostatic effect. Pretreatment of chicks with the caecal culture also reduced the proportion of salmonella-positive crop samples. Growth of the salmonella in a feed-slurry system resembling moistened crop contents occurred rapidly at 37°C but was prevented by addition of the treatment culture, a decline in numbers coinciding with Lactobacillus populations of ca 10⁹/g and a pH value of 5·5. Incorporation of a feed decontaminant, 1% formic acid (pH 4·0), rapidly eliminated both salmonellas and organisms added from the treatment culture.     

Comparison of the VIDAS system, FTA filter-based PCR and culture on SM ID for detecting Salmonella in Dermanyssus gallinae

AIMS: To compare different analytical methods for detecting Salmonella in Dermanyssus gallinae. METHODS AND RESULTS: The detection limit of three Salmonella detection methods [Vitek immunodiagnostic assay (VIDAS) Salmonella immuno-concentration/immunoassay, FTA filter-based PCR, and Salmonella detection and identification medium (SM ID) preceded by a pre-enrichment step] was evaluated by crushing mites in serial dilutions of pure cultures of Salmonella enterica ssp. Enterica serotype Enteritidis. Each method was then compared for its ability to detect Salmonella in artificially contaminated mites. In 105 mites artificially engorged with Salm. Enteritidis-contaminated blood, Salmonella was isolated from 68 samples of the samples cultured on SM ID and tests were positive for Salmonella using FTA filter-based PCR and VIDAS in 77 and 65 samples, respectively. Using SM ID as our reference method, specificities and sensitivities were 97% and 94% and 73% and 98.5% for VIDAS and PCR, respectively. CONCLUSIONS: Each method allowed the detection of Salmonella in contaminated mites and is usable for screening mites. PCR is more sensitive but less specific than VIDAS for detecting Salmonella. SIGNIFICANCE AND IMPACT OF THE STUDY: This is the first time that the VIDAS has been used to detect pathogens in vectors. The development of analytical methods for Salmonella detection in mites is a necessary step in the study of the role of D. gallinae as a vector of salmonellae and to check the contamination of D. gallinae in poultry facilities.     

Detection of Clostridium botulinum types A, B, E and F in foods by PCR and DNA probe

A PCR procedure was developed for the detection of Clostridium botulinum in foods. PCR products were detected in agarose gels and by Southern hybridization. The sensitivity of PCR was tested in broth cultures and in canned asparagus, dry cured ham and honey. The sensitivity of the method in broth was high (2·1–8·1 cfu ml⁻¹) for types A and B, but rather low (10⁴ cfu ml⁻¹) for types E and F. However, after enrichment at 37°C for 18 h, it was possible to detect Cl. botulinum types A, B, E and F in food samples at initial levels of about 1 cfu 10 g⁻¹ of food. This PCR detection protocol provides a sensitive and relatively rapid technique for the routine detection of Cl. botulinum in foods.