Rapid Determination of Salmonella in Samples of

APPLIED MICROBIOLOGY, Nov. 969, p. 88-84 Vol. 8, No. 5 Copyright 969 American Society for Microbiology Printed in U.S.A. Rapid Determination of Salmonella in Samples of Egg Noodles, Cake Mixes, and Candies GEORGE J. BANWART AND MADELEINE J. KREITZER Market Quality Research Division, U.S. Department of Agriculture, Beltsville, Maryland 75 Received for publication 4 July 969 A glass apparatus system was compared with a standard enrichment broth-selective agar method to test samples of egg noodles, cake mixes, and candy for the presence or absence of salmonellae. The glass apparatus system used fermentation of mannitol, production of HS, or motility, in conjunction with a serological test of flagellar antigens, to detect salmonellae. No salmonellae were detected in 7 samples of food products. Of these samples, 7 were found to be Salmonella-negative after 48 hr with the glass apparatus system. After 7 hr, the standard Salmonella procedure yielded 8 samples which produced Salmonella false-positive results on selective agars. Inoculation of samples with cultures of Salmonella showed that approximately one inoculated cell could be detected after 48 hr of incubation with the glass apparatus. The standard Salmonella test requires a minimum of 7 hr for completion. Compared with the standard Salmonella test, the glass apparatus system is a more rapid and simple system that can be used to determine the presence or absence of Salmonella in these food products. The use of a motility system to detect salmonellae in mixed cultures was reported by Stuart and Pivnick (9) and Harper (7). Banwart et al. (5) and Banwart (4) not only utilized a motility system but also included biochemical tests in the system to hasten the determination of salmonellae. In the glass apparatus described by Banwart (4), the organisms were grown in the system so that their serological reactions could also be determined. The presence of Salmonella in cake mixes has been reported by Adinarayan et al. (), Butler and Josephine (6), and Skoll and Dillenberg (8). Adinarayan et al. () analyzed 5 samples of egg noodles, but found no salmonellae. This report describes the use of a glass apparatus system to detect Salmonella in samples of egg noodles, cake mixes, and candies. MATERIALS AND METHODS The glass apparatus described by Banwart (4), consisting of a central chamber from which three U tubes project, was used in this study (Fig. ). Semisolid SIM and mannitol purple (MP) agars were prepared as described by Banwart (4) and were used in two of the side tubes. Into the third side arm was placed a selenite-brilliant enrichment (SBG) broth (Difco) solidified with 8 g of agar (Difco) per liter. To prepare this medium, double-strength SBG broth was made and then heat-treated according to the manufacturer's directions. A mixture containing 6 g of agar (Difco) per liter of water was prepared and sterilized at C for 5 min. Just prior to use, equal 88 portions of the melted agar and SBG broth were mixed so that the final medium consisted of single-strength SBG solidified with 8 g of agar per liter. After the agar plugs had solidified in the side tubes, 5 g of food sample was mixed with 5 ml of sterile lactose broth (Difco) in the center chamber of the apparatus. Then to 5 ml of sterile Brain Heart Infusion (BHI) was added to the top of agar plugs in each of the outside side tubes. The prepared flasks plus samples were then incubated at 7 C. After 4 and 48 hr, the fermentation of mannitol and production of hydrogen sulfide were noted. Growth from turbid BHI was tested with Salmonella-H polyvalent antiserum (Difco). To simulate the standard Salmonella test, after the 4-hr incubation period, ml of the lactose brothsample mixture was transferred to 9 ml of selenitecystine (SC) broth. The inoculated SC broth was incubated at 7 C for 4 hr and then loopfuls of the mixture were streaked onto surfaces of BG Sulfa agar (Difco) and bismuth-sulfite agar (Difco). These were incubated for 4 hr at 7 C and observed for typical Salmonella colonies. Negative bismuth-sulfite surfaces were incubated a total of 48 hr. Salmonella-like colonies on the agar surfaces were picked and transferred to triple sugar-iron (TSI) and lysine-iron slants. All organisms giving typical reactions on the agar slants were tested serologically with Salmonella-O and Salmonella-H antisera. Food products tested were egg noodles, cake mixes (mainly angel cake mixes which contained dried egg albumen), and candies (mostly those containing egg albumen). These foods were purchased in grocery stores in the Beltsville, Md., area. Downloaded from http://aem.asm.org/ on June, 8 by guest

VOL. 8, 969 DETERMINA4ION OF SALMONELLA IN FOODS FIG.. Glass apparatus with three side U tubes proecting into a central chamber. To determine the effectiveness of the glass apparatus system in detecting salmonellae in the presence of these foods, samples of the egg noodles and cake mixes were prepared as described and inoculated with pure cultures of S. anatum and S. tennessee, respectively. A most probable numbers (MPN) technique () was used to determine the number of bacteria inoculated and the number that were recovered. The cultures were grown in tubes of lactose broth at 7 C for 4 hr and were diluted in sterile.% peptone water. Three dilutions, referred to as, -', and " dilutions, were added to the food sample-lactose broth mixture for the MPN tests. Three tubes of sterile lactose broth and three flasks containing the food and lactose broth were inoculated with each of the dilutions. These were incubated at 7 C. The lactose broth tubes were observed for turbidity after 4 hr, and the number of turbid tubes was used to determine the number of cells in the inoculum (). The presence of the inoculated Salmonella in the tubes of lactose broth was confirmed by streaking growth from the incubated broth onto BG Sulfa agar and testing the resultant colonies with the corresponding "O"-group antiserum. RESULTS AND DISCUSSION With this glass apparatus system, a sample was considered to be Salmonella-negative if, after incubation for 48 hr, there was no fermentation of the semisolid MP or no motility through this 89 medium, no HS production or no motility through the semisolid SIM, or no motility through the semisolid SBG. If there was mannitol fermentation and motility through the MP agar, HS production and motility through the SIM, or motility through the SBG, growth from the BHI was tested with Salmonella-H polyvalent antiserum. If no visible agglutination occurred, the sample was considered to be Salmonellanegative. In practice, only those samples showing an agglutination with the Salmonella-H antiserum would need to be tested with the selective agars for isolation and further identification. None of the 7 samples of purchased foods was found to contain any salmonellae (Table ). However, the use of the glass apparatus to determine quickly the Salmonella-negative character of food products could be advantageous. After 48 hr, 48 of the 64 samples of egg noodles contained motile organisms causing a fermentation of mannitol (Table ). An additional eight samples contained organisms which migrated through the MP agar but did not cause a fermentation of the mannitol. Only four samples contained organisms which produced HS in the SIM agar. An additional 9 samples had organisms which migrated through the SIM but did not produce HS. Thus, with the use of the biochemical aspect of SIM, i.e., the detection of HS production, only four of the samples needed to be tested further for Salmonella when this medium was used in the side tube. Since no biochemical change could be discerned in the SBG medium, all samples which contained motile organisms migrating through this medium had to be considered possibly positive for Salmonella. When the organisms in the turbid BHI above the semisolid agars were tested with Salmonella-H polyvalent antiserum, none showed any agglutination for the egg noodle samples (Table ). Thus, after 48 hr of incubation it could be stated that all egg noodle samples were negative for Salmonella when the glass apparatus system was used. In contrast with this, 4 of the 64 samples were possibly Salmonella-positive on selective agars when the standard Salmonella-test was used. Since a minimal time period of 7 hr is needed to observe organisms on the selective agars, the glass apparatus was a distinct benefit in detecting Salmonella-negative samples of egg noodles. Salmonellae were not isolated from any of the 64 egg noodle samples. Many of the egg noodle samples contained mold. After pre-enrichment incubation of the samples in lactose broth, there was considerable mold growth on the surface of the broth. For the Downloaded from http://aem.asm.org/ on June, 8 by guest

84 standard Salmonella test, it was necessary to open the container to pipette ml of the lactose broth into the SC broth for further enrichment. This operation exposed the laboratory to possible mold contamination. Using the glass apparatus system, the mold problem was not important since these organisms were not capable of growing through the semisolid agar plugs, and the turbid BHI, which was sampled, contained no mold growth. After 48 hr of incubation in the glass apparatus system, 47 of 58 cake mix samples were shown to contain bacteria that migrated through the MP agar and fermented mannitol (Table ). Two BANWART AND KREITZER additional samples had MP motile bacteria, but there was no evidence of mannitol fermentation. There were only nine samples which showed microbial movement through SIM agar with HS production. An additional samples showed motility of bacteria through SIM, but no blackening of the medium was observed. Thus, all but nine samples could be considered Salmonellanegative with this medium since only 6 of the 58 cake mix samples showed bacterial motility through the SBG. Of the three semisolid used, the SIM eliminated the most samples from the possible Salmonella category. TABLE. Reactions of bacteria in various food samples determined with the glass apparatus system No. of samples showing Salmonella-positive reactions Egg noodles (64 samples) Cake mix (58 samples) Candy (5 samples) Total (7 samples) Semisolid medium in side tube Motility and Motility and Motility and Motility and biochemical "H" testb biochemical "H" testb biochemical "H" testb biochemical "H" testb reactions reactionsa reactions" reactions" 4r 48r 4hr 48 hr 4 br 48 hr 4 hr 48 hr 4 hr 48 hr 4 hr 48 hr 4 hr 48 hr 4 hr 48 hr Mannitol-purple... 7 48 47 8 6 8 SIM... 4 9 Selenite-Brilliant enrichment... 8 8 6 7 4 8 a Biochemical tests were fermentation of mannitol with acid production and production of hydrogen sulfide in SIM. No particular change could be determined in the semisolid selenite-brilliant green. Organisms were considered motile if they grew through the agar plug and caused turbidity in the Brain Heart Infusion. b The "H" test consisted of testing the organisms in the turbid Brain Heart Infusion with Salmonella-H polyvalent antiserum. Lack of an observable agglutination was considered to be Salmonella-negative. TABLE. Recovery of Salmonella from egg noodles inoculated with S. anatum in the glass apparatus (number of positive results with a three-"tube" or "flask" MPN) Dilution - - MPNd Limits, 9..5-8 4..7- Mannitol.9.-.6 SIM.9.-.6 Selenite-Brilliant 4..7- APPL. MICROBIOL. Total of semisolid 4..7- " The number inoculated was determined by a three-tube MPN with tubes of sterile lactose broth. ' Organisms in the turbid BHI broth were tested with Salmonella-H polyvalent antiserum. Lack of visible agglutination was considered to be Salmonella-negative. d Values obtained from reference. The value indicates the number of organisms present in the e The 95% confidence limits are from reference. Downloaded from http://aem.asm.org/ on June, 8 by guest

VOL. 8, 969 DETERMINATION OF SALMONELLA IN FOODS 84 TABLE. Recovery of Salmonella from egg noodles inoculated with S. derby in the glass apparatus (number of positive results with a three-"tube" or "flask" MPN) Dilution Inoculuma Standard testb Mannitol Si Selenite-Brilliant Total of semisolid - - MPNd..5..9..5 Limits.4-.-4.4.5-..-.6 -.-4.4 a The number inoculated was determined by a three-tube MPN with tubes of sterile lactose broth. c Organisms in the turbid BHI were tested with Salmonella-H polyvalent antiserum. Lack of visible agglutination was considered to be Salmonella-negative. d Values obtained from reference. The value indicates the number of organisms present in the e The 95% confidence limits are from reference. TABLE 4. Recovery of Salmonella from cake mix samples when inoculated with pure cultures ofs. tennessee in the glass apparatus (number of positive results with a three-"tube" or "flask" MPN) Dilution Inoculuma Standard testb Mannitol Sim Selenite-Brilliant Total of semisolid - W MPNd 4. 9..5.5.9 4. Limitse.7-.5-8.-4.4.-4.4.-.6.7- a The number inoculated was determined by a three-tube MPN with tubes of sterile lactose broth. c Organisms in the turbid BHI were tested with Salmonella-H polyvalent antiserum. Lack of visible agglutination was considered to be Salmonella-negative. d Values obtained from reference. The value indicates the number of organisms present in the e The 95% confidence limits are from reference. Testing of the microbial growth in the BHI with Salmonella-H polyvalent antiserum revealed two possible Salmonella-positive samples. However, 56 of the 58 cake mix samples could be considered as Salmonella-negative after 48 hr of incubation. With the standard test of pre-enrichment, enrichment, and streaking on selective agars, 4 of the 58 samples contained suspect Salmonella organisms after 7 hr. Further biochemical testing revealed no salmonellae in any of the cake mix samples. The reactions observed in the glass apparatus during the analysis of 5 candy samples are also shown in Table. There was no HS production in the SIM agar with any of the candy samples. Only one-third of the samples showed microbial migration through the SBG, and about two-thirds were possibly Salmonella-positive in the MP medium, as shown by the movement through this medium and the fermentation of mannitol. Testing the growth in the BHI with Salmonella-H polyvalent antiserum showed no agglutination. Thus, all 5 candy samples were considered to be Salmonella-negative after 48 hr with the glass apparatus. With the standard Salmonella test, 5 of the 5 candy samples were found to be Salmonellanegative after 7 hr. With further testing, the Downloaded from http://aem.asm.org/ on June, 8 by guest

84 BANWART AND KREITZER APPL. MICROBIOL. remaining sample was also found to be Salmonella-negative. A summary of the reactions in the glass apparatus of all 7 samples is shown in Table. Of the three semisolid used in the side tubes, SBG was the most effective, since, of the 8 samples which showed turbidity in the BHI above the SBG, none gave any reactions in the Salmonella-H polyvalent antiserum. Therefore, all 7 could have been declared Salmonella-negative after 48 hr with this medium. SIM also was a very satisfactory medium; only samples produced HS, and only of these gave a false-positive reaction in the Salmonella-H polyvalent antiserum. Since there were only samples that needed to be tested, less time and effort were expended with the SIM medium than with either the SBG with 8 tests or the MP with 8 tests. Besides requiring the greatest number of agglutination tests, the semisolid MP agar also revealed one sample that was Salmonella-suspect after 48 hr. Considering all three semisolid, 7 of the 7 food samples were determined to be Salmonella-negative after only 48 hr. The two Salmonella-suspect samples were found to be Salmonella-negative with further testing. With the standard Salmonella test procedure, a minimum of 7 hr was needed before any determination of the Salmonella character of the sample could be made. Even after 7 hr, there were 8 possible Salmonella-positive samples when the standard Salmonella test was used. If Salmonella could be detected in 4 hr, further time could be saved. The sensitivity of the glass apparatus system in determining salmonellae in egg noodle samples was tested by inoculating a pure culture of S. anatum and S. derby. The MPN results for S. anatum are shown in Table. When lactose broth in test tubes was used, the MPN of cells inoculated into the flasks in each series was 9.,.9, and.9 cells. The SBG semisolid medium revealed as many S. anatum as did the standard test, and the results with MP and SIM were within the 95 % confidence limit of the result of the standard test. The results obtained when S. derby was added to egg noodle-lactose broth mixtures are shown in Table. No one semisolid medium revealed as many S. derby as did the standard test, but of the three SIM gave results within the 95% confidence limit of the result of the standard test. It is noted that no S. derby cells were detected when SBG was used, which is a complete reversal of the data obtained when S. anatum was the test organism. To compare with the standard test, it was necessary to add the results obtained with both MP and SIM agars. It is evident from these data that the use of more than one medium is necessary in this type of test. The results obtained when S. tennessee was inoculated into flasks containing cake mix samples are shown in Table 4. The three dilutions used contained approximately 4.,.4, and.4 cells. Inoculation of the cake mix samples with these three dilutions revealed an MPN of 4. cells detected by means of the Salmonella-H polyvalent agglutination of the BHI in the side tubes. By the standard system of enrichin& and streaking on selective agars, a recovery 9. cells was obtained. The flask in which Tennessee was not detected by the agglutinatio system was apparently inoculated with only -one cell. As with S. anatum or S. derby in egg noodles, the S. tennessee added to cake mix samples was not readily detected after only 4 hr and in four flasks after 48 hr. The poor recovery of S. tennessee with any one semisolid is being investigated. It is evident that more than one semisolid medium is needed to detect all possible salmonellae. This is not unusual, as the presently recommended Salmonella test () suggests the use of two enrichment broths and three selective agars. With the glass apparatus, three agars can be used with a onepiece system. Both pre-enrichment and enrichment can be accomplished without physical transfer of organisms from one system to another. The data from these experiments show that some Salmonella-positive samples can be detected after incubation for 4 hr; however, the detection of all positive samples requires incubation for 48 hr. LITERATURE CITED. Adinarayan, N., V. D. Foltz, and F. McKinley. 965. Incidence of salmonellae in prepared and packaged foods. J. Infec. Dis. 5:9-6.. American Public Health Association. 965. Standard method for the examination of water and waste water, th ed., p. 68. American Public Health Association, Inc., New York.. Anonymous. 967. Microbiological methods. J. Ass. Offic. Anal. Chem. 5:-9. 4. Banwart, G. J. 968. Glassware apparatus for determining motile bacteria.. Salmonella. Poultry Sci. 47:9-. 5. Banwart, G. J., A. J. Mercuri, and T. R. Ryan. 968. Screening method for determining Salmonella-negative samples of pasteurized dried whole egg. Poultry Sci. 47:598-6. 6. Butler, R. W., and J. E. Josephine. 96. Egg-containing cake mixes as a source of Salmonella. Can. J. Public Health 5: 478-48. 7. Harper, J. 968. Semi-solid agar as a selective medium for Salmonella. J. Pathol. Bacteriol. 95:55-554. 8. Skoll, S. L., and H.. Dillenberg. 96. Salmonella thompson in cake mix. Can. J. Public Health 54:5-9. 9. Stuart, P. F., and H. Pivnick. 965. Isolation of salmonellae by selective motility systems. Appl. Microbiol. :65-7. Downloaded from http://aem.asm.org/ on June, 8 by guest