A chromogenic plating medium for the isolation and identification of Enterobacter sakazakii from foods, food ingredients, and environmental sources.

A chromogenic agar, R&F Enterobacter sakazakii chromogenic plating medium (ESPM), was developed for isolating presumptive colonies of E. sakazakii from foods and environmental sources. ESPM contains two chromogenic substrates (5-bromo-4-chloro-3-indoxyl-alpha-D-glucopyranoside and 5-bromo-4-chloro-3-indoxyl-beta-D-cellobioside), three sugars (sorbitol, D-arabitol, and adonitol), a pH indicator, and inhibitors (bile salts, vancomycin, and cefsulodin), which all contribute to its selectivity and differential properties. On ESPM, 79 pure culture strains of E. sakazakii (10 clinical isolates and others from food and environmental sources) yielded blue-black (three strains were blue-gray) raised colonies, 1 to 2 mm in diameter with and without halos after 24 h at 35 degrees C. Other enteric organisms plus Pseudomonas aeruginosa yielded white, yellow, green, or clear colonies with and without clear halos. Of these genera, only Shigella sonnei and one Pantoea strain produced blue-black to blue-gray colonies. ESPM was used to isolate E. sakazakii from a variety of foods: corn, wheat, and rice flours; powdered infant formula; dairy products (dried milk, whey, and caseinates); cereals; and environmental sources. Most false-positive results on ESPM were eliminated by observing acid production on either sucrose or melibiose after 6 h at 35 degrees C on a R&F E. sakazakii screening medium (ESSM) biplate. In an analysis of 240 samples, the number of samples positive for E. sakazakii by the ESPM-ESSM method and the U.S. Food and Drug Administration protocols (violet red bile glucose agar and tryptic soy agar) were 27 and 16, respectively, with sensitivity and specificity values of 100.0 and 96.9% versus 59.3 and 43.7%, respectively. These data support the fact that E. sakazakii confirmation should be based on more than one confirmation system. Both the API 20E and Biolog Microlog3 4.20 systems should be used for confirmation of E. sakazakii isolates.

Efficacy of enrichment broths in the recovery of freeze-injured Escherichia coli O157:H7 in inoculated ground beef by PCR.

Escherichia coli O157:H7 strains ATCC 35150 and ATCC 43894 and five pooled isolates from beef and pork freeze injured at -25 degrees C in beef infusion were used to inoculate ground beef. Samples (25 g each) were added to 225 ml of buffered peptone water with vancomycin, cefsulodin, and cefixime (BPW-VCC), 225 ml of modified EC broth plus novobiocin (mEC+n), and 225 ml of R&F enrichment broth (R&F-EB) and aerobically incubated at 41 to 42 degrees C. After 6, 7, 8, and 24 h of incubation, levels of E. coli O157:H7 recovered from each broth by a PCR assay with the BAX automated system as well as by conventional enrichment with the use of nonaerated mEC+n incubated at 35 degrees C for 24 h were compared with levels recovered by cultural isolation with immunomagnetic separation and plating on BCM E. coli O157:H7 chromogenic agar. For ground beef inoculated with a mean of 4.23 +/- 1.00 total cells (74% freeze injured) per 25 g, after 6 h the PCR assay identified 72.7, 57.6, and 66% of the samples for R&F-EB, BPW-VCC, and mEC+n, respectively, as presumptive positive, whereas the recovery rates after 7 and 8 h exceeded 90%, with the rate for R&F-EB being 100%. For ground beef inoculated with a mean of 1.50 +/- 0.56 total cells (80% freeze injured) per 25 g, after 6 h the PCR assay identified 47.6, 19.1, and 9.5% of the samples for R&F-EB, BPW-VCC, and mEC+n, respectively, as presumptive positive. These values increased to 81.0, 61.9, and 52.4% after 7 h and to 95.2, 61.9, and 71.4% after 8 h. After 24 h, only 55 to 60% of the samples at both inoculum levels tested positive by PCR with conventional enrichment and incubation, whereas >95% of the samples tested positive with R&F-EB aerated at 41 to 42 degrees C. Culture results for R&F-EB and mEC+n after 7 and 8 h of incubation were closely correlated with presumptive positive PCR results.

Reveal 8-Hour Test System for detection of Escherichia coli O157:H7 in raw ground beef, raw beef cubes, and iceberg lettuce rinse: collaborative study.

Five different food types were analyzed by the Reveal for E. coli O157:H7 8-Hour Test System (Reveal 8) and either the U.S. Food and Drug Administration's Bacteriological Analytical Manual (BAM) culture method or the U.S. Department of Agriculture Food Safety Inspection Service (FSIS) culture method for the presence of E. coli O157:H7. A total of 27 laboratories representing academia and private industry in the United States and Canada participated. Food types were inoculated with E. coli O157:H7 at 2 different levels: a high level where predominantly positive results were expected, and a low level where fractional recovery was anticipated. During this study, 1,110 samples and controls were analyzed by both the Reveal 8 and by BAM or FSIS by each of the collaborators (2,220 samples in total). For each set of samples, 740 were artificially inoculated with E. coli O157:H7, and 370 were uninoculated controls. The Reveal 8 detected 528 presumptive positives of which 487 were confirmed positive by the BAM culture method. In comparison, BAM and FSIS detected 489 of the 740 artificially contaminated samples as positive. In an additional in-house study performed only on chilled and frozen raw ground beef, 240 artificially inoculated samples were analyzed by both the Reveal 8 and by FSIS. The Reveal 8 detected and confirmed 104 samples as positive compared to 79 confirmed positive by FSIS.

Comparison of the Reveal 20-hour method and the BAM culture method for the detection of Escherichia coli O157:H7 in selected foods and environmental swabs: collaborative study.

Four different food types along with environmental swabs were analyzed by the Reveal for E. coli O157:H7 test (Reveal) and the Bacteriological Analytical Manual (BAM) culture method for the presence of Escherichia coli O157:H7. Twenty-seven laboratories representing academia and private industry in the United States and Canada participated. Sample types were inoculated with E. coli O157:H7 at 2 different levels. Of the 1,095 samples and controls analyzed and confirmed, 459 were positive and 557 were negative by both methods. No statistical differences (p <0.05) were observed between the Reveal and BAM methods.

Efficacy of chromocult coliform agar for coliform and Escherichia coil detection in foods.

Chromocult coliform agar (CCA) was compared with Petrifilm Escherichia coli count plate (PEC) for identifying coliforms and E. coli in a variety of meat products. Products examined included 45 raw beef samples, 12 sausage emulsion samples, 11 samples of meat-based ready-to-eat appetizers, and 8 pork trimming samples. Coliforms from CCA and PEC were confirmed by gassing in brilliant green lactose broth plus a positive reaction on purple broth agar plus lactose after incubation at 35 degrees C for 48 h. Lauryl sulfate tryptose plus methylumbelliferyl-beta-glucuronide and tryptophan broth were used to confirm E. coli from CCA and PEC with 48-h incubations at 35 and 42.5 degrees C, respectively. API 20E test strips were inoculated for final confirmation. The overall respective confirmation percentages (CFU/g) for the PEC and the CCA methods were 93.1 and 93.7% for coliforms and 99.8 and 98.1% for E. coli, although the CCA method yielded significantly (P < 0.001) higher mean CFU/g values for both coliforms and E. coli. Regression analyses of these data indicated a strong positive linear relationship existed between the two methods over a wide CFU/g range for both coliforms and E. coli. The respective correlation coefficients obtained for coliforms and E. coli of 0.89 and 0.86 indicate that the CCA method provides a reliable optional method for these determinations in meat products.

A chromogenic plating medium for isolating Escherichia coli O157:H7 from beef.

There was no significant difference (P > 0.05) in the percentages of Escherichia coli O157:H7 cells recovered on BCM O157:H7 (+) agar (69.7%) and MacConkey sorbitol agar containing 5-bromo-4-chloro-3-indoxyl-beta-D-glucuronic acid (MSA-BCIG) (76.8%) vs Tryptic soy agar. Three E. coli O157:H7 strains (ATCC 35150, 43890 and 43894) were separately inoculated into raw ground beef at low (mean 0.32 cfu g-1) and high (mean 3.12 cfu g-1) levels. Using the United States Department of Agriculture (USDA) m-EC + novobiocin enrichment broth, BCM O157:H7 (+) medium surpassed MSA-BCIG agar with overall percentage sensitivities for BCM O157:H7 (+) of 92.1 and 94.4 compared with 52.6 and 84.7 for MSA-BCIG at low and high levels, respectively. A comparison of BCM O157:H7 (+) and MSA-BCIG agars using naturally contaminated beef samples was made utilizing presumptively positive enrichment broths previously identified by rapid methods. The E. coli O157:H7 cells in these broths were concentrated with Dynabeads anti-E. coli O157 before inoculating the agars. The respective percentage sensitivity and specificity values were 90.0 and 78.5 for BCM O157:H7 (+) and 70.0 and 46.4 for MSA-BCIG. Thus, under identical pre-plating conditions, BCM O157:H7 (+) medium displayed a greater sensitivity than MSA-BCIG for detecting E. coli O157:H7 in artificially inoculated beef, and both greater sensitivity and specificity upon examining naturally contaminated beef samples.

Isolation and detection of Listeria monocytogenes using fluorogenic and chromogenic substrates for phosphatidylinositol-specific phospholipase C.

The BCM Listeria monocytogenes detection system (LMDS) consists of a selective preenrichment broth (LMPEB), selective enrichment broth (LMSEB), selective/differential plating medium (LMPM), and identification on a confirmatory plating medium (LMCM). The efficacy of the BCM LMDS was determined using pure cultures and naturally and artificially contaminated environmental sponges. The BCM LMPEB allowed the growth of Listeria and resuscitation of heat-injured L. monocytogenes. The BCM LMSEB, which contains the fluorogenic substrate 4-methylumbelliferyl-myo-inositol-1-phosphate and detects phosphatidylinositol phospholipase C (PI-PLC) activity, provided a presumptive positive test for the presence of pathogenic Listeria (L. monocytogenes and L. ivanovii) after 24 h at 35 degrees C. An initial inoculum of 10 to 100 CFU/ml of L. monocytogenes in BCM LMSEB yielded a fluorogenic response after 24 h. On BCM LMPM, L. monocytogenes and L. ivanovii were the two Listeria species forming turquoise convex colonies (1.0 to 2.5 mm in diameter) from PI-PLC activity on the chromogenic substrate, 5-bromo-4-chloro-3-indoxyl-myo-inositol-1-phosphate. L. monocytogenes was distinguished from L. ivanovii by either its fluorescence on BCM LMCM or acid production from rhamnose. False-positive organisms (Bacillus cereus, Staphylococcus aureus, Bacillus thuringiensis, and yeasts) were eliminated by at least one of the media in the BCM LMDS. Using a pure culture system, the BCM LMDS detected one to two L. monocytogenes cells from a sponge rehydrated in 10 ml of DE neutralizing broth. In an analysis of 162 environmental sponges from facilities inspected by the U.S. Department of Agriculture (USDA), the values for identification of L. monocytogenes by BCM LMDS and the USDA method were 30 and 14 sites, respectively, with sensitivity and specificity values of 85.7 and 100.0% versus 40.0 and 66.1%, respectively. No false-positive organisms were isolated by BCM LMDS, whereas 26.5% of the sponges tested by the USDA method produced false-positive results.

A 5-h screening and 24-h confirmation procedure for detecting Escherichia coli O157:H7 in beef using direct epifluorescent microscopy and immunomagnetic separation.

An antibody-direct epifluorescent filter technique (Ab-DEFT) detected 100% of the raw ground beef samples inoculated with Escherichia coli O157:H7 cells (0.15 cells g-1) and incubated in a prewarmed, modified buffered peptone water (mBPW) non-selective enrichment broth for 5 h at 42 degrees C in an orbital shaking water bath (200 rev min-1). Over 50% of the microscopic fields viewed were positive (1-10 fluorescent cells field-1) in the Ab-DEFT. All positive screening results were confirmed within 24 h by subjecting 1 ml of the mBPW to the Dynabeads anti-E. coli O157 immunomagnetic separation procedure, followed by plating on MacConkey sorbitol agar containing 5-bromo-4-chloro-3-indolyl-beta-D-glucuronide. At this cell concentration, 41.7% of the inoculated samples were detected by the conventional method involving a 24-h selective enrichment. Exposure to viable cells before filtration was minimized by using a 0.58% formaldehyde concentration for 5 min at 50 degrees C (killed > 4.00 logs of E. coli O157:H7 cells) without affecting cell fluorescence.

Efficacy of ozonated water against various food-related microorganisms.

The antimicrobial effects of ozonated water in a recirculating concurrent reactor were evaluated against four gram-positive and four gram-negative bacteria, two yeasts, and spores of Aspergillus niger. More than 5 log units each of Salmonella typhimurium and Escherichia coli cells were killed instantaneously in ozonated water with or without addition of 20 ppm of soluble starch (SS). In ozonated water, death rates among the gram-negative bacteria--S. typhimurium, E. coli, Pseudomonas aeruginosa, and Yersinia enterocolitica--were not significantly different (P > 0.05). Among gram-positive bacteria, Listeria monocytogenes was significantly P < 0.05) more sensitive than either Staphylococcus aureus or Enterococcus faecalis. In the presence of organic material, death rates of S. aureus compared with L. monocytogenes and E. coli compared with S. typhimurium in ozonated water were not significantly (P > 0.05) affected by SS addition but were significantly reduced (P < 0.05) by addition of 20 ppm of bovine serum albumin (BSA). More than 4.5 log units each of Candida albicans and Zygosaccharomyces bailii cells were killed instantaneously in ozonated water, whereas less than 1 log unit of Aspergillus niger spores was killed after a 5-min exposure. The average ozone output levels in the deionized water (0.188 mg/ml) or water with SS (0.198 mg/ml) did not differ significantly (P < 0.05) but were significantly lower in water containing BSA (0.149 mg/ml).

Growth Response of an Osmotolerant, Sorbate-Resistant Saccharomyces rouxii Strain: Evaluation of Plating Media.

Standard Potato Dextrose Agar (PDA) was inferior to PDA/60% sucrose plating medium for enumerating osmotolerant yeasts from a high-sugar food product. The largest disparity in the number of Saccharomyces rouxii cells (>2.00 log number cells/g) recovery by PDA/60% sucrose and standard PDA was during the lag and early log phase. The optimum growth temperatures of S. rouxii propagated in double strength (2X) Potato Dextrose Broth (PDB) were 28°C at 0.995 aw level and 35°C for 0.92 and 0.90 aw levels. No significant (P>0.05) difference was determined between shaking and stomaching for enumeration of S. rouxii , an osmotolerant yeast, from chocolate syrup. Therefore, use of PDA/60% sucrose incubated at 35°C for 5 d is recommended for enumeration of osmotolerant yeasts from a liquid high-sugar product.

Antimicrobial effects of ionizing radiation on artificially and naturally contaminated cacao beans.

With an initial microbial level of ca. 10 microorganisms per g of Ivory Coast cacao beans, 5 kGy of gamma radiation under an atmosphere of air reduced the microflora per g by 2.49 and 3.03 logs at temperatures of 35 and 50 degrees C, respectively. Bahia cacao beans were artificially contaminated with dried spores of Aspergillus flavus and Penicillium citrinum, giving initial fungal levels of 1.9 x 10 and 1.4 x 10 spores per g of whole Bahia cacao beans, respectively. The average D(10) values for A. flavus and P. citrinum spores on Bahia cacao beans were 0.66 and 0.88 kGy, respectively.

Growth characteristics of Saccharomyces rouxii isolated from chocolate syrup.

We investigated the growth parameters of Saccharomyces rouxii isolated from spoiled chocolate syrup. The optimum pH range for S. rouxii was 3.5 to 5.5, whereas the minimum and maximum pH values that permitted growth were 1.5 and 10.5, respectively. For cells grown in 0 and 60% sucrose the optimum water activity (aw) values were 0.97 and 0.96, respectively. The optimum temperature for S. rouxii increased with a decreasing aw regardless of whether glucose or sucrose was used as the humectant. The optimum temperatures for S. rouxii were 28 degrees C at an aw of greater than 0.995 and 35 degrees C at an aw of 0.96 to 0.90 in 2 X potato dextrose broth with sucrose. Increasing the sorbate concentration (from 0.03 to 0.10%) caused the growth of S. rouxii to become more inhibited between aws of greater than 0.995 and 0.82. S. rouxii did not grow when the sorbate level was 0.12% (wt/vol). At lower sorbate levels, the effect of sorbate on the growth of S. rouxii depended on the aw level. Lowering the aw enhanced the resistance of S. rouxii to increasing concentrations of potassium sorbate. Permeability and polyol production are discussed with respect to sorbate tolerance of S. rouxii at different aw levels.

Growth Response of an Osmotolerant Sorbate-Resistant Yeast, Saccharomyces rouxii , at Different Sucrose and Sorbate Levels.

Effects of preconditioning Saccharomyces rouxii to various sucrose and sorbate concentrations were investigated. Cells were preconditioned to 0 and 60% (wt/vol) sucrose without and with 0. 1% (wt/vol) sorbate by a series of four transfers in potato dextrose broth (PDB). The preconditioned cells were inoculated in growth medium containing 0, 30, or 60% (wt/vol) sucrose with 0. 0.05, or 0.1% (wt/vol) sorbate and in chocolate syrup without and with 0.1% (wt/vol) sorbate. Preconditioning of S. rouxii cells to 0 and 60% sucrose offered no advantage for growth in a medium containing 0, 30 or 60% sucrose. When S. rouxii cells were preconditioned in 0.1% sorbate, an increased resistance to sorbate occurred. S. rouxii cells preconditioned in 60% sucrose/0.1% sorbate were more sensitive to sorbate than cells preconditioned in 0% sucrose/0.1% sorbate. Regardless of the sucrose level, S. rouxii cells, preconditioned in 0. 1% sorbate showed an increased resistance to sorbate when inoculated into the growth medium and the chocolate syrup.

Antibiotic Susceptibility Patterns of Yersinia enterocolitica.

Antibiotic susceptibility patterns for Yersinia enterocolitica strains involving 10 different serotypes were analyzed and compared. All Y. enterocolitica were susceptible to colistin, gentamicin, kanamycin, neomycin and doxycycline, whereas all isolates displayed resistance to penicillin G, methicillin (derivative of penicillin), novobiocin, and clindamycin. The antibiograms for the Y. enterocolitica isolates were in some instances related to the somatic serotypes, especially serotype 0:8 for which the antimicrobial susceptibility pattern displayed the greatest disparity. By eliminating the antibiograms for the four serotype 0:8 strains, antimicrobial susceptibility patterns for atypical and typical strains were similar.

Microbiology of Meats in a Hypobaric Environment.

Growth of total bacteria on fresh meat and heat-processed commodities was analyzed at refrigerated hypobaric and conventional cooler storage. When using the time required for bacterial levels to reach one million/in.2 as the estimate of the shelf life, the refrigerated hypobaric storage system substantially extended the shelf-life of broiler chickens, pork loins, bone-in heat-processed hams, and lamb and beef carcasses, as compared to storage in conventional coolers. The dominant microorganism isolated from the surface of the bone-in heat-processed hams stored in hypobaric and conventional coolers was a Streptococcus which was resistant to the heat process and tolerant to salt and sodium nitrite. For the fresh meat products, Pseudomonas was the dominant microorganism isolated from these products stored in either hypobaric or conventional coolers.

Thermal injury of Yersinia enterocolitica.

Procedures were developed to evaluate thermal injury to three strains of Yersinia enterocolitica (serotypes 0:3, 0:8, and 0:17). Serotype 0:17 (atypical strain) was more sensitive to bile salts no. 3 (BS) and to sublethal heat treatment than the typical strains, 0:3 and 0:8. When the 0:3, 0:8, and 0:17 serotypes were thermally stressed in 0.1 M PO4 buffer, pH 7.0, at 47 degrees C for 70, 60, and 12 min, respectively, greater than 99% of the total viable cell population was injured. Injury was determined by the ability of cells to form colonies on brain heart infusion (BHI) agar, but not on Trypticase soy agar (TSA) plus 0.6% BS for serotypes 0:3 and 0:8 and TSA plus 0.16% BS for 0:17. Heat injury of serotype 0:17 cells for 15 min in 0.1 M PO4 buffer caused an approximate 1,000-fold reduction in cell numbers on selective media as compared with cells heated in pork infusion (PI), BHI broth, and 10% nonfat dry milk (NFDM). The extended lag and resuscitation period in BHI broth was 2.5 times greater for 0:17 cells injured in 0.1 M PO4 than for cells injured in BHI or PI. The rate and extent of repair of Y. enterocolitica 0:17 cells in three recovery media were directly related to the heating menstruum used for injury. The use of metabolic inhibitors demonstrated that ribonucleic acid synthesis was required for repair, whereas deoxyribonucleic, cell wall, and protein synthesis were not necessary for recovery of 0:17 cells injured in 0.1 M PO4 buffer, BHI, or PI. Inhibition of respiration by 2,4-dinitrophenol slowed repair only for 0:17 cells injured in 0.1 M PO4 buffer, not for cells injured in PI or BHI.

Evaluation of the Minitek and API 20E Systems for Identification of Yersinia enterocolitica.

The basic 12-disk Minitek system (14 tests) and the API 20E system were evaluated for identification of Yersinia enterocolitica strains isolated at various locations in the world. The API 20E system correctly identified all the Y. enterocolitica strains while the Minitek system identified 80% (20/25) of the strains. Four of the five cultures misidentified by Minitek were atypical strains isolated from vacuum-packaged meats. These four cultures were identified as Citrobacter freundii by Minitek due to a rhamnose-positive reaction. Minitek and API 20E are rapid and convenient systems which could prove to be beneficial for identifying Y. enterocolitica strains.

Effects of varying concentrations of novobiocin incorporated into two salmonella plating media on the recovery of four enterobacteriaceae.

Hydrogen sulfide-producing strains of salmonellae, Escherichia coli, Citrobacter freundii, and Proteus mirabilis were isolated from fresh pork sausage. All the strains produced black-centered colonies on Hektoen enteric agar (HE). On xylose lysine deoxycholate agar (XLD), C. freundii produced yellow colonies, and the strains of the other three genera formed black-centered colonies. The selectivity of HE and XLD for salmonellae was improved by the addition of novobiocin to both media. With increasing concentrations of novobiocin, the degree of growth inhibition for the four genera was less on HE than on XLD. Novobiocin concentrations of 80 mug/ml in HE and 5 mug/ml in XLD did not affect the growth or colonial morphology of salmonellae. When 80 mug of novobiocin per ml was incorporated into HE, P. mirabilis strains were not recovered, 40% of C. freundii strains failed to form black-centered colonies, and growth of E. coli strains was not affected but colonies were altered without eliminating the black centers. When novobiocin at 5 mug/ml was incorporated into XLD, colonies of P. mirabilis strains were not recovered, C. freundii formed yellow colonies, and the colonies of the H(2)S-producing E. coli strains were unaffected.