Stanley M. Harmon

Bacillus cereus Contamination of Seeds and Vegetable Sprouts Grown in a Home Sprouting Kit

Sprouting seeds (alfalfa, mung bean and wheat) were purchased at local health food stores and examined for Bacillus cereus by the official AOAC method. Of 98 units collected, 56 (57%) were positive for B. cereus at levels ranging from 3 to >500 per g. Population levels of B. cereus on sprouts grown from naturally contaminated seeds in a home sprouting kit ranged from a mean of log10 3.72 for alfalfa to 5.39 for wheat; the log10 mean for mung bean sprouts was 4.52. Washing contaminated sprouts for 10 min with warm tap water as recommended by the manufacturer of the sprouting kits reduced the B. cereus count for mung bean sprouts by approximately one log unit but was less effective for wheat sprouts. B. cereus populations large enough to cause food poisoning (>105/g) frequently remained on wheat sprouts even after three wash cycles, and significant numbers of viable B. cereus remained on wheat sprouts even after cooking for 20 min.

Incidence and growth potential of Bacillus cereus in ready-to-serve foods.

To simulate temperature abuse, 106 test portions of ready-to-serve moist foods, 12 test portions of rehydrated powdered infant formula, and 18 test portions of nonfat dry milk were incubated for 20 and 24 h at 26°C, and then examined for Bacillus cereus . Of the ready-to-serve moist foods, 88 of 106 were positive for B. cereus at levels ranging from 0.25 to 8.5 × 106/g after 20 h of incubation and from 0.1 to 58 × 106/g after 24 h. All of the powdered milk and 12 of the 15 units of infant formula, representing five brands, were positive, with counts ranging from 0.15 to 5.0 × 106/g in 20 h and 5.0 to 49 × 106 after 24 h. B. cereus counts in the powdered products were low, ranging from 0.09/g for one of two soy-based products to an average of 0.29/g for milk-based products. However, these levels were sufficient to initiate growth of B. cereus in almost every 2-oz serving. Similar results were obtained for rehydrated nonfat milk, with initial B. cereus counts ranging from 0.29 to 1.5/g; at 26°C the counts averaged 3.3 × 107 after 20 h and 5.5 × 107 after 24 h. Counts ranged from 2.0 × 104 to 1.1 × 105 after 9 h in milk and were in excess of 106/g after 10.5 h.

Evaluation of a Reversed Passive Latex Agglutination Test Kit for Clostridium perfringens Enterotoxin

A reversed passive latex agglutination (RPLA) test kit for Clostridium perfringens enterotoxin (CPE) marketed by the Denka-Seiken Co., Tokyo, Japan, was evaluated by using culture supernatant fluids and extracts from feces of food poisoning patients. Nanograms of CPE were detectable with the assay and the reaction was specific, as shown by parallel activity in a double antibody enzyme-linked immunosorbent assay (ELISA). Although less sensitive, the RPLA method is easier to perform than the ELISA and counterimmunoelectrophoresis, both of which require special test reagents and equipment not generally available.

Improved Media for Sporulation and Enterotoxin Production by Clostridium perfringens

Sporulation of 24 strains of Clostridium perfringens isolated from stools of food poisoning patients and normal controls was improved by adding sodium carbonate to Duncan-Strong (DS) sporulation medium and replacing starch with raffinose in Tanigutis AEA medium. Viable counts of heat-tolerant spores (75°C for 20 min) were 2 to 186 times greater in modified AEA medium and 2 to 169 times greater in modified DS than in DS medium. Reversed passive latex agglutination assays revealed a corresponding increase in enterotoxin titers in supernatant fluids of the 12 enterotoxigenic strains grown in modified AEA medium and in modified DS medium.

AN IMPROVED MEDIUM FOR DETECTION OF CLOSTRIDIUM BOTULINUM TYPE E

An enrichment medium containing trypticase, peptone, glucose, yeast extract, and 1 mg of trypsin/ml (TPGYT) has been developed for detection of Clostridium botulinum type E. It was designed to potentiate toxin as produced and destroy the boticins of competing type E variants in cultures of foods and environmental materials. Samples of 227 sediments and 283 shellfish from different areas of the continental United States yielded 74 more positive cultures of type E in TPGYT than in the same medium free of trypsin. Type E toxin was detected in all smoked fish inoculated with 4 to 100 type E spores per fish in TPGYT. Incorporation of trypsin into the medium caused no reduction in type A or proteolytic type B toxin production from spore inocula. Toxins of nonproteolytic types B and F in pure culture were fully potentiated in the trypsin-containing medium.

Estimating Population Levels of Clostridium perfringens in Foods Based on Alpha-toxin

A method for estimating population levels of Clostridium perfringens in foods based on the titer of α-toxin in food extracts was evaluated. Twenty-two samples of food associated with 15 food poisoning outbreaks and 12 ready-to-serve foods inoculated with approximately 100 C. perfringens spores/g and incubated to simulate improper holding conditions were examined. Alpha-toxin was present in extracts from most of the foods which could be correlated with the viable population as determined by plate counts in sulfite-polymyxin-sulfadiazine (SPS) agar. The amount of α-toxin present could be correlated with previous growth of C. perfringens in food regardless of whether the organisms were viable when the examination was made. This is important because storage and shipment of the food specimens for brief periods at low temperature results in a decrease of 2 to 4 log cycles in the viable plate count in SPS agar, but does not affect the accuracy of estimates of the population level based on the presence of α-toxin.

Enumeration and Characterization of Clostridium perfringens Spores in the Feces of Food Poisoning Patients and Normal Controls

The fecal spore enumeration method for confirming Clostridium perfringens as the cause of food poisoning was evaluated using strains implicated in nine outbreaks in the United States. Confirmed spore counts from 66 stool specimens were made on tryptose-sulfite-cycloserine (TSC) without egg yolk and trypticase soy-sheep blood (TSB) agars. Counts from outbreak stools on TSC agar ranged from 2.0 × 104 to 3.5 × 108 (mean ≥ 1.4 × 106/g) as compared with <103 to 5.0 × 105/g (overall mean 9.5 × 103/g) from normal stools. Similar results were obtained with TSB agar. Isolates from seven of the nine outbreaks were nonhemolytic and produced ≥100 ng enterotoxin/ml in spore broth, as measured by an enzyme-linked immunosorbent assay. Spores in stools from six of the outbreaks were heat-resistant and survived heating for 30 to 60 min at 100°C in cooked meat medium. Strains from the three remaining outbreaks were heat-sensitive and survived heating for only 15 min at 100°C. Enterotoxigenic isolates from all but one of the outbreaks were serotyped. In all instances, the predominant strain in specimens from an outbreak was of the same serotype, indicating that it was the causative strain. Reexamination of five specimens from each of three outbreaks after storage at -20°C for 6 months showed only a minimal reduction in the spore counts.