Reader Comments

Post a new comment on this article

Referee Comments: Referee 2 (Brian D Wade)

Posted by PLOS_ONE_Group on 30 Nov 2007 at 18:12 GMT

Reviewer 2's Review


The manuscript submitted by Cooley et al. for publication in PLoS ONE entitled, "Molecular typing of Escherichia coli O157:H7 isolated from a major produce production region and human strains causing outbreak and sporadic illness", compares the efficacy of two molecular genotyping techniques, PFGE and MLVA, to differentiate subpopulations of Escherichia coli O157:H7 ("EcO157"). Food safety scientists employ these two techniques to determine relatedness among subpopulations of food-borne pathogens. Currently, they use PFGE more often than MLVA for such purposes, due primarily to the availability of a database of pathogen PFGE profiles which they can query for matches with ones of their environmental and clinical isolates. Here, the authors isolated EcO157 subpopulations from various types of samples collected from agricultural regions implicated in previous leafy green vegetable-borne EcO157 outbreaks. They then utilized PFGE and MLVA to genotypically profile these environmental isolates and human clinical EcO157 isolates cultivated by others during various food-borne EcO157 outbreaks. The authors used these profiles to compare the two techniques and to examine the relatedness of subpopulations isolated from different locales and from the same locales sampled during different months of the year. The latter was done in an attempt to identify some potential environmental factors underlying the aforementioned leafy green vegetable-borne EcO157 outbreaks and to narrow the possible sources of these outbreaks. The authors conclude, as have others in previous publications, that MLVA is an effective technique for genotyping environmental and clinical isolates of EcO157. Regarding the discriminatory power of each technique, they report that MLVA profiles generally permitted them to better discriminate isolates than did PFGE ones. The ability of MLVA to discern closely related EcO157 isolates has also been documented by others in previous publications. This apparent advantage of MLVA over PFGE when genotyping EcO157 isolates should be noted by the food safety community. The use of these techniques by the authors also aided them in narrowing the possible sources of the aforementioned leafy green vegetable-borne EcO157 outbreaks to certain farm/ranch sites. They also conclude that such outbreaks are dynamic and complex processes involving temporal and spatial aspects yet to be fully understood. Results presented by the authors here will likely appeal to food safety scientists and perhaps to professionals in other public health fields. As stated by the authors, MLVA "is being evaluated by [the] CDC as a potential next generation typing system".

Below are methodological suggestions and an idea for possible future work, they are not intended for inclusion in the manuscript reviewed here:

The following are methodological suggestions for the authors regarding two parts of their isolation procedure, that, if implemented, could potentially surmount the specific problems and limitations they mention. First, concerning the growth of non-EcO157 microbes on the selective plates (page 26, line 1), consider including tellurite and novobiocin in the TSB during the enrichment step, as this may reduce the number of non-target microbes on the these plates. Second, regarding the preclusion of determining the concentration of EcO157 in water samples due to the enrichment step (page 26, line 22), consider passing the samples through a 0.2 micron filter to collect the cells, resuspend them in a greatly reduced volume, and then plate portions of this concentrate directly onto the selective plates. Chemical (e.g., addition of a nonionic detergent such as Tween 80) and/or physical (e.g., brief sonication) treatment of the samples before filtering them may aid in detaching cells adhering to sediments. Alternatively, Neogen Corporation ( offers an EcO157-specific kit (ISO-GRID) which essentially does this (Payne and Kroll, 1991, Trends in Food Science and Technology 2: 315-319), and probably more easily and rapidly than the just described method. This procedure may eliminate the need to enrich the dilute water samples and thus allow the authors to determine the concentration of EcO157 in the water samples. For a review of methods for separating bacteria from food matrices, of which the authors could possibly adapt to their sample types, see Stevens and Jaykus (2004, Critical Reviews in Microbiology 30: 7-24).

The following is an idea for possible future work. If the authors have not already done so, perhaps they may consider PCR amplifying directly from the environmental samples (whole community DNA extracts), if frozen portions of them still exist, the EcO157 virulence genes that they amplified from the environmental isolates (rfbE, fliC, stx1, stx2, eae, hly). There is at least one pitfall to doing this, most of these genes can be found in other strains of E. coli and thus are not unique to EcO157, so PCR products from environmental samples will likely contain some non-EcO157 amplicons. To overcome this, they may consider using the recently developed E. coli O157:H7-specific RQ3 primer set (Lin and Lin, 2007, Molecular and Cellular Probes 21: 182-189). Another possible primer set is RJD3/SF6 (Perelle et al., 2002, Journal of Applied Microbiology 93: 250-260). If the authors amplify these genes (former) and/or regions (latter two) from clinical isolates, environmental isolates, and directly from the environmental samples, phylogenetic comparisons of alleles from the former two may clarify relationships among the isolates, while analyses of the latter may further elucidate the temporal and spatial dynamics of EcO157 subpopulations in the environment, perhaps enhancing traceback investigations. Analyses of gene alleles comprising the PCR products from the environmental samples could include fingerprinting techniques such as Denaturing Gradient Gel Electrophoresis (DGGE), and subsequent sequencing of the separated alleles, and other allele separating and sequencing methods such as clone library construction. If the PCR products of the virulence genes they used to screen environmental isolates or the DNA extracts from these isolates still exist, part of what has just been outlined, multigene phylogenetic comparisons between the environmental isolates and clinical isolates, could be done with little further effort. The MLVA and PFGE derived relationships the authors established among the EcO157 isolates could serve as a guide to confine the proposed analyses. Doing so may afford the authors even better phylogenetic resolution of their isolates than MLVA profiles provided, a need they mention on page 30, line 19 and again on page 31, line 7.