Coliforms and Enterobacteriaceae Isolates from Selected Foods.

We obtained coliform counts and Enterobacteriaceae counts using violet red bile agar (VRB) and VRB + 1% glucose (VRBG), respectively, of samples of five food products. From each set of VRB and VRBG plates, 28 to 40 "typical" colonies were randomly selected and identified by use of the R-B Enteric Differential System. A pure culture of each isolate was also subjected to the sequential tests for gas production in LST and BGLB broths (confirmed coliforms) and in EC broth at 45.5 C (fecal coliforms). IMViC reaction patterns of EC-positive cultures were also determined. Approximately 80% of the VRB isolates from broiler skin and from mechanically deboned poultry meat (MDPM) met all the criteria for fecal coliforms, whereas only 62.5% and 36.5%, respectively, of the VRBG isolates from these two products met these criteria. Fewer than 10% of the VRB and VRBG isolates from chicken pot pie, ground beef, or pork sausage produced gas in LST broth. The percentages of fecal coliforms and Escherichia coli (Type I or II) among the 179 VRB isolates were 34.1 and 33.5, respectively. Corresponding percentages for the 193 VRBG isolates were 20.7 and 19.7. E. coli was the predominant species isolated on both media from broiler skin and MDPM. Enterobacter agglomerans was the principal species isolated from chicken pot pie and pork sausage; Serratia liquefaciens predominated in ground beef.

Comparative Study of Micro-ID, Minitek and Conventional Methods with Enterobacteriaceae Freshly Isolated from Foods.

A total of 400 Enterobacteriaceae isolates freshly taken from broiler carcasses, ground beef, pork sausage, raw shrimp, pre-wrapped sandwiches, raw carrots, lettuce and fresh strawberries was inoculated into the 15 biochemical tests of the Micro-ID (4-h) system and into the 15 corresponding tests in the Minitek (24-h) and conventional systems. For each food there were 750 biochemical test comparisons (50 isolates × 15 tests). The overall agreement between Micro-ID and conventional tests was 96.8%, whereas the agreement between Minitek and conventional tests was 93.6%. Three laboratory technicians who independently recorded results of 6000 biochemical tests from each of the three systems were in complete agreement for 99.3%, 98.9% and 99.7% of the Micro-ID. Minitek and conventional tests, respectively. Thus results obtained with the miniaturized systems were as easy to read and interpret as conventional tests in tubes. The most frequently encountered Enterobacteriaceae from these foods were Escherichia coli (broiler carcasses, pork sausage). Enterobacter agglomerans (carrots, lettuce, shrimp, strawberries), Enterobacter cloacae (pre-wrapped sandwiches), and Serratia liquefaciens (ground beef).

Accuracy of Micro-ID for Identification of Salmonella and Other Enterobacteriaceae from Clinical and Food Sources.

The accuracy of a new 4-h Enterobacteriaceae identification kit, not previously tested with food isolates, was evaluated by using 265 known bacterial cultures. These clinical and food isolates represented 11 genera of Enterobacteriaceae ; about 10% were atypical strains. Micro-ID test strips were inoculated from 24-h brain heart infusion agar slants and then incubated at 37 C for 4 h. With initial tests, 251 of the 265 cultures (95%) were correctly identified to species. When cultures initially misidentified were re-inoculated into Micro-ID, accuracy of identification increased to 98%. In tests to determine whether this multitest system would perform efficiently if incubated overnight at a lower temperature. 100 cultures of Enterobacteriaceae were inoculated into Micro-ID strips and incubated at 22 C for 16 h. Under these conditions, the system correctly identified 86 cultures to genus and 78 to species. Most of the misidentifications involved Enterobacter agglomerans . The degree of accuracy of Micro-ID at 22 C for 16 h was comparable to that at 37 C for 4 h, i.e., 95%, for identification to genus of the following: Arizona , Citrobacter , Edwardsiella , Escherichia , Klebsiella , Proteus , Salmonella , Serratia , Shigella and Yersinia .

Recovering Low Levels of Various Salmonella Serotypes from Deep-frozen Broiler Carcasses by Direct Enrichment.

Two hundred and forty broiler carcass halves were each inoculated with either 14 or 180 cells of Salmonella typhimurium . Each carcass half was then placed in a plastic bag, blast-frozen (-40 C) for 6 h, and stored at -23 C. After 1, 7 and 30 days of frozen storage, 80 of these samples were removed and allowed to thaw; then each carcass-half was shaken in its bag with 150 ml of added sterile water. Lactose broth was used to preenrich 40 of these rinse-fluid samples and selenite cystine broth was used for direct enrichment of the remaining 40 samples. S. typhimurium was successfully recovered from all 240 samples. Other serotypes successfully recovered by direct enrichment on similarly frozen carcass-halves stored for 30 days were Salmonella california , Salmonella derby , Salmonella heidelberg , Salmonella montevideo , Salmonella newport and Salmonella senftenberg . These data suggest that a preenrichment medium such as lactose broth may not be necessary for detection of salmonella on frozen broiler carcasses.

Recovery of Salmonellae from Broiler Carcasses by Direct Enrichment.

Each of four serotypes of Salmonella ( S. anatum , S. montevideo , S. saint-paul , S. typhimurium ), inoculated at low levels on broiler carcasses (ca. 20 cells/carcass) was detected by direct enrichment of the whole carcass rinse fluid with either Selenite Cystine Broth (SC) or Selenite Brilliant Green Broth (SGB). Neither Selenite Brilliant Green Sulfa Broth (SBGS) nor TT Broth was effective in detecting the serotypes unless the entire broiler carcass with the rinse fluid was incubated with either of these enrichment broths. SBG and SC were effective as direct enrichment broths for recovering pure cultures of the four seroytpes subjected to sublethal heat treatment (53 C for 1 min) approximating that to which broiler carcasses are subjected during the commercial scalding process.

Effectiveness of Sampling Methods for Salmonella Detection on Processed Broilers.

A total of 240 processed broiler carcasses (water-chilled and unfrozen) were each sampled by three methods (whole-carcass rinse, neck-skin rinse, and macerated neck skin) for detection of Salmonella . In addition to this, various procedures were compared: destructive (incubating the entire carcass with the rinse fluid) versus non-destructive (incubating the rinse water with concentrated lactose or selenite cystine broth added after removal of the carcass) sampling and pre-enrichment versus no pre-enrichment during Salmonella detection procedures. There was no significant difference (p < 0.05) between the percentage of Salmonella -positive carcasses obtained by destructive sampling and the percentage obtained by non-destructive samples of whole carcasses. There was also no significant difference (p < 0.05) in results obtained by rinsing and blending excised neck-skin samples. There was highly significant difference (p = 0.001), however, between whole carcass and neck-skin analyses. With whole-carcass sampling, 45% of the carcasses were positive for the presence of Salmonella while with rinsing or blending the neck skin of these same carcasses, only 11% and 12%, respectively, were positive for the organism. Pre-enrichment of the whole carcass, of the whole-carcass rinse, or of the neck-skin samples did not result in significantly greater percentages of positive results than did direct enrichment of these samples.

Relation of Enterobacteriaceae Counts to Salmonella Contamination of Market Broilers.

We determined the Enterobacteriaceae counts and the Salmonella status (positive or negative) of 20 individual birds in each of 12 groups of broiler carcasses. The overall logarithmic mean Enterobacteriaceae count for the 240 carcasses was 2.7 with group means ranging from 1.8 to 3.6. One hundred and twenty three (51.2%) of the 240 carcasses were Salmonella -positive. The number of Salmonella - positive carcasses within groups ranged from 0 to 18. No relationship was found between Enterobacteriaceae counts and presence of Salmonella in broiler carcasses.

Comparison of Two Minikits (API and R-B) for Identification of Enterobacteriaceae Isolated from Poultry and Meat Products.

Two miniaturized commercial kits (API and R-B) were evaluated for identification of 373 strains of Enterobacteriaceae isolated from selected poultry and meat and meat products, such as frozen chicken pot pie, frozen comminuted chicken, processed chicken, ground beef, and pork sausage. The taxonomic classification by these two systems was identical for 59% (221/373) of the isolates. Whenever the two systems disagreed, cultures were identified by conventional methods. The API correctly classified 82% (306/373) of the isolates, while R-B correctly classified 72% (267/373). Most of the disagreements involved organisms of the Klebsiella-Enterobacter-Serratia group of the Enterobacteriaceae family.

Growth and Production of Enterotoxin by Staphylococcus aureus S-6 in Soy Proteins and Soy-Supplemented Beef and Pork Sausage.

The effect of three commercial soy proteins on growth and production of enterotoxin by Staphylococcus aureus S-6 was determined. Sterile isolated soy protein (ISP), soy protein concentrate (SPC), and textured soy protein (TSP) were adjusted to 20% protein by diluting with sterile nutrient medium, inoculated with S. aureus S-6 and incubated at 37 C. Generation times of S. aureus S-6 in ISP, SPC, and TSP were 41, 38, and 33 min, respectively. At 48 h, log 8.4 - 8.5 organisms/g were found in the soy products, and staphylococcal enterotoxin A (SEA) and staphylococcal enterotoxin B (SEB) were produced. Each of three hydrated soy protein products was added to ground beef or pork sausage to attain 20% (wt/wt) and cooked to 71 C. Each product was then inoculated and incubated at 37 C. Generation times of S. aureus S-6 did not differ significantly for any beef-soy or pork-soy samples compared to beef or pork sausage controls. Production of SEB (12-72 h) was similar in most beef-soy and pork-soy samples compared to controls, but was significantly lower in the beef-SPC product and beef-TSP and significantly higher in pork-ISP. Small quantities of SEA were produced in beef, beef-soy, pork, and pork-soy samples. Possibly reduction of enterotoxin production in some meat-soy samples was due to outgrowth and competition by spore contaminants of the soy proteins that survived cooking. Production of SEB in all beef-soy and beef control samples was not significantly different when raw samples were autoclaved before inoculation.

Total Aerobic and Coliform Counts in Beef-Soy and Chicken-Soy Patties During Refrigerated Storage.

Three commercial texturized soy proteins (TSP-A, TSP-B, TSP-C) and one commercial soy protein concentrate (SPC) were each added to raw ground beef or chicken at a 10% or 30% level to form patties which were stored at 4 C. During storage up to 8 days, aerobic plate counts of beef were significantly higher with than without the TSP's. Coliform counts were higher with the beef containing TSP-B and TSP-C. During storage up to 10 days chicken patties containing TSP-A or TSP-C exhibited higher aerobic plate counts than controls, whereas patties with TSP-B or SPC did not. No increases occurred in coliform counts. In beef-soy mixtures, the counts were higher for the 30% than for the 10% level but the reverse was true in chicken-soy mixtures. Coliforms were identified by the Analytab Products, Inc. System. In the beef control and beef-soy mixtures, Serratia was the predominant genus at 0 and 8 days of storage. In the chicken control, Escherichia was the predominant genus at 0 day of storage, but Enterobacter was predominant in the control and in chicken-soy mixtures after 10 days of storage.

Evaluation of Four ONPG Tests for Enterobacteriaceae from Human, Animal and Selected Food Sources.

An ortho-nitrophenyl-ß-D-galactopyranoside (ONPG) test was done on 250 Enterobacteriaceae from human, poultry, and selected food sources by use of four commercially available miniaturized systems and by the conventional test for ONPG. For the 102 cultures from human and poultry sources, all four systems agreed with the conventional test as follows: API (98%), Difco (94.1%) Minitek (98%), and Pathotec (98%). For the 148 food isolates, the percent agreement between the conventional and these four systems was significantly lower, API (87.2%), Difco (85.8%), Minitek (88.5%), and Pathotec (85.8%).

Swab and excised tissue sampling for total and Enterobacteriaceae counts of fresh and surface-frozen broiler skin.

Excised broiler skin tissue (12.3 cm 2 area) was (a) blended for various times up to 2 min. or (b) shaken 25, 50, or 75 times with or without sterile glass beads. Blending for 10 sec. or shaking 75 times with glass beads yielded the highest total plate count (TPC) and Enterobacteriaceae count (ENT). Counts did not significantly differ between these methods when freshly processed, unfrozen broiler skin was sampled. However, blending yielded significantly higher TPC and ENT than the conventional sampling procedure or swabbing a 12.3 cm. 2 area of the intact skin for 30 sec. with a calcium alginate swab. When sampling hard chilled (surface frozen) broiler skin, there was no significant difference between shaking and blending for TPC, but shaking gave a significantly higher ENT with less variation than blending. Blending, however, yielded significantly higher TPC and ENT counts than swabbing.

Bacteriological examination of commercial precooked Eastern-type turkey rolls.

Studies were conducted to ascertain the bacteriological condition of commercially cooked Eastern-type (foil-wrapped-oven roasted) turkey rolls during processing and storage. After 2 weeks at 5 C, numbers of aerobes on the surface of rolls, in slices, and in whole rolls reached levels of from 1 to 10 million per cm(2) or per g. In stored whole rolls, coliform and enterococcus counts ranged, respectively, from about 10,000 to more than 1 million per g and from < 100 to more than 1 million per g. Postcooking processing operations in two plants did not significantly affect the total count of turkey rolls. Eight of 28 rolls obtained after handling and packaging contained coagulase-positive staphylococci.