Submitted to: Analytical and Bioanalytical Chemistry
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/8/2008
Publication Date: 5/1/2008
Citation: Irwin, P.L., Nguyen, L.T., Chen, C. 2008. Binding of non-target microorganisms from food washes to anti-Salmonella and anti-E. coli O157 immuno-magnetic beads: minimizing the errors of random sampling in extreme dilute systems. Analytical and Bioanalytical Chemistry. 391:515-524.
Interpretive Summary: The contamination of foods with pathogenic bacteria (e.g., Salmonella or E. coli O157:H7) may lead to substantial food poisoning epidemics. In order to sensitively detect food poisoning bacteria certain isolation methods, such as the well-known immuno-magnetic bead (IMB) technique for separating and concentrating these bacteria, have been developed. Unfortunately, the level of pathogenic bacteria in foods is very low relative to the concentration of benign bacteria. Thus, IMB methods for extracting pathogenic bacteria from foods also capture some of the good bacteria thereby limiting the functionality of these methods. In this report we have developed a sampling protocol to determine accurately the most probably occurring nontarget organisms in several foods which bind nonspecifically to commercial IMBs. This information is useful to microbiologists in the detection of pathogenic bacteria from foods using IMBs protocols.
Technical Abstract: For most applications, 3-5 observations (n) are utilized for estimating total aerobic plate count in an average population greater than about 50 (mu) cells or colony forming units per sampled volume. We have chosen to utilize the 6x6 drop plate method because it offers the means to rapidly perform all requisite dilutions for colony selection in a 96-well format and plate these dilutions using a minimum of materials. Besides traditional quantitative purposes, we also need to select colonies which are well-separated from each other for 16S ribosomal DNA isolation, sequencing, and identification. To achieve this goal using the 6x6 drop plate format, it requires the utilization of very dilute solutions (mu < 10 counts in 7-10 microL). Using either observed or computer-generated data, we assigned a large sample of individual counts randomly into N subsamples each of n = 2-24 replicates (N x n = 360). From these data we calculated the average total mean-normalized deviation of the total standard deviation (Stot) from each jth subsample's estimate (Sj) which we call Delta. When either observed or computer-generated Delta values were analyzed as a function of the total average counts, a set of relationships proportional to the inverse square root of the total average count were generated which appeared to converge at n = 12-18. This observation was verified analytically since the first derivative of Delta with respect to n approached 0 as n approached 12 which argues that this is the nominal value for n which we seek for both enumeration and colony selection. Test sampling experiments were performed on three chicken harvests testing these this drop plate sampling protocol (using n = 18) as compared with MPN methods and were found not to be significantly different.