An Enhanced Discriminatory Pulsed-Field Gel Electrophoresis Scheme for Subtyping Salmonella Serotypes Heidelberg, Kentucky, SaintPaul, and Hadar

Transcription

1 2067 Journal of Food Protection, Vol. 71, No. 10, 2008, Pages Copyright, International Association for Food Protection Research Note An Enhanced Discriminatory Pulsed-Field Gel Electrophoresis Scheme for Subtyping Salmonella Serotypes Heidelberg, Kentucky, SaintPaul, and Hadar MEILI XI, 1 JIE ZHENG, 2 SHAOHUA ZHAO, 3 ERIC W. BROWN, 4 AND JIANGHONG MENG 1,2 * 1 College of Food Science and Engineering, Northwest A&F University, Shaanxi, China; 2 Department of Nutrition and Food Science, University of Maryland, College Park, Maryland 20742, USA; 3 Division of Animal and Food Microbiology, Office of Research, Center for Veterinary Medicine, Food and Drug Administration, Laurel, Maryland 20708, USA; and 4 Division of Microbiology, Center for Food Safety and Applied Nutrition, Food and Drug Administration, College Park, Maryland 20740, USA MS : Received 20 March 2008/Accepted 31 May 2008 ABSTRACT Conventional pulsed-field gel electrophoresis (PFGE) protocols, used extensively as a successful approach for subtyping many salmonellae, may be inadequate for discriminating strains sharing levels of homogeneity within the same serotype. Four additional restriction enzymes (SpeI, PacI, SfiI, and NotI), in addition to XbaI and BlnI, were used in PFGE typing of 33 Salmonella Heidelberg, 27 Salmonella Kentucky, 27 Salmonella SaintPaul, and 27 Salmonella Hadar isolates that were recovered from poultry and porcine retail meats from different states of the United States. A dendrogram derived from the combined analysis of six enzymes was highly discriminatory with a Simpson index of diversity value of over The ratio of nodes to isolates was more than 0.75 with an average of fewer than three isolates in each polytomy for all four serotypes. Two three-enzyme combinations, SpeI/NotI/SfiI for Salmonella Heidelberg and Salmonella Hadar, and SpeI/BlnI/SfiI for Salmonella Kentucky and Salmonella SaintPaul, were found to have comparable discriminatory abilities of differentiating isolates of these Salmonella serotypes with the six-enzyme combination. The enhanced discriminatory PFGE-based subtyping scheme can be used effectively for the differentiation of strains of the four Salmonella serotypes. The findings also highlight PFGE analysis as a continued essential and informative subtyping method for source tracking and outbreak investigations of these and other Salmonella pathogens. In recent years, Salmonella Heidelberg, Salmonella Kentucky, Salmonella SaintPaul, and Salmonella Hadar have been recovered frequently from humans with gastroenteritis (2 4, 6, 9, 15). Since 2002, these four serotypes remained on the list of the 10 most common Salmonella serovars from retail meat source, humans, and animals in the United States ( pg. html#data). These serotypes are often associated with foodborne illness derived from multiple food vehicles (6, 13, 14, 16). The use of DNA fingerprinting techniques such as multilocus sequence typing, ribotyping, and pulsed-field gel electrophoresis (PFGE) have proved to be useful in discriminating isolates of Salmonella serovars. For example, PFGE has become a useful tool for typing and differentiating strains of Salmonella serotypes Typhimurium, Kentucky, and Heidelberg. However, a conventional PFGE protocol (using only XbaI and BlnI) had some difficulties in differentiating strains that shared high levels of genetic ho- * Author for correspondence. Tel: ; Fax: ; jmeng@umd.edu. mogeneity within the same serotype. Most recently, an enhanced PFGE approach with six restriction enzymes was found to be capable of resolving a group of highly homogeneous Salmonella Enteritidis isolates from different poultry sources into meaningful phylogenetic lineages (18). Here we report the use of a six-enzyme based PFGE scheme to assess the relatedness of 114 isolates of Salmonella serotypes Heidelberg, Kentucky, SaintPaul, and Hadar and provide an evaluation of the usefulness of this scheme in differentiating closely genetically related isolates of Salmonella serotypes. MATERIALS AND METHODS Bacteria. Thirty-three isolates of Salmonella Heidelberg and 27 isolates each from serotypes Kentucky, SaintPaul, and Hadar were selected in the study based on conventional PFGE using two enzymes, XbaI and BlnI. The bacteria were isolated from poultry and porcine retail meats from different states of the United States and were obtained from the collection at the Food and Drug Administration Center for Veterinary Medicine, Laurel, Md., and the Department of Nutrition and Food Science, University of Maryland, College Park. They were all cultured on Luria-Bertani agar plates (BD, Franklin Lakes, N.J.).

2 2068 XI ET AL. J. Food Prot., Vol. 71, No. 10 TABLE 1. Pulsed-field gel electrophoresis run conditions for enzymes used to subtype Salmonella Heidelberg, Salmonella Kentucky, Salmonella SaintPaul, and Salmonella Hadar Enzyme Digestion temp ( C) Digestion time (h) Enzyme unit (plug) Run time (h) Initial switch time (s) Final switch time (s) BlnI NotI PacI SfiI SpeI XbaI Restriction enzymes. After screening numerous enzymes, the six enzymes studied in-depth here included XbaI, BlnI, and SpeI, all of which are employed in the PulseNet protocol; SfiI and PacI, applied previously in PFGE studies of Escherichia coli O157:H7 (7); and NotI, found to yield an optimal number of cut sites among the Salmonella genomes tested by REBASE at PFGE analysis. The standard CDC-PulseNet protocol for PFGE of nontyphoidal Salmonella was performed as described previously (5), using a CHEF-DRIII (Bio-Rad, Hercules, Calif.) with individual run conditions as listed in Table 1. Salmonella Branderup H9812 was used as the standard. Resultant PFGE gel images were analyzed using the molecular analyst DST version 1.6 software (Bio-Rad). The gels were normalized against the molecular weight marker system used, with the same settings as those of PulseNet CDC, and bands were assigned automatically. By using unweighted pair group method with arithmetic mean (UPGMA), dendrograms were obtained based on Dice coefficient and the optimization set at 0.5%, with a position tolerance of 1.5% for band comparison. Enzyme diversity measures. Four distinct diversity measures were applied to assess the relative discriminatory potential of each of the restriction enzymes and combinations thereof. The majority of these diversity values were extracted from the dendrograms that resulted from cluster analysis of a specific enzyme or enzyme combination. First, the mean number of isolates per polytomy (an unresolved cluster comprised of genotypically identical isolates) was calculated as the total number of polytomous isolates divided by the total number of polytomies in the tree. Second, the total percentage of polytomous isolates (of 33 Salmonella Heidelberg isolates and 27 isolates each of serotypes Kentucky, SaintPaul, and Hadar) was calculated. Third, the nodeto-isolate ratio was calculated as the total number of nodes (bifurcating forks in a dendrogram) divided by the total number of isolates of each serotype. A relative node-to-isolate value closer to 1 is indicative of a more resolved tree, while a value closer to 0 indicates a more poorly resolved dendrogram (18). Finally, Simpson s index of diversity (SID) was calculated as a numerical assessment of the relative discriminatory potential of each enzyme and enzyme combination (11), with approximate 95% confidence intervals (CI) determined as described previously (8). RESULTS Genetic homogeneity among isolates of Salmonella serotypes Heidelberg, Kentucky, SaintPaul, and Hadar. PFGE using XbaI orblni two enzymes commonly used in the subtyping of Salmonella isolates was performed on all isolates of Salmonella serotypes Heidelberg, Kentucky, SaintPaul, and Hadar. The resultant dendrogram derived from the single enzyme analysis of XbaI and BlnI displayed a very poor resolution with isolates in each serotype, especially in the XbaI data set. Over 90% of isolates were captured by unresolved clusters (data not shown). Further, the resultant dendrogram from the combined XbaI/BlnI analysis was slightly improved, revealing only marginal discriminatory power (Figs. 1A, 2A, 3A, and 4A). As an example, 79% of Salmonella Heidelberg, 78% of Salmonella Kentucky, 67% of Salmonella SaintPaul, and 67% of Salmonella Hadar isolates were found in unresolved clusters (i.e., polytomies) (Table 2). Additionally, the ratio (1: 2) of nodes to isolates signaled a poorly bifurcated tree presented by the XbaI and BlnI two-enzyme PFGE scheme (Table 2). Discrimination of genetically homogeneous Salmonella by using six selected restriction enzymes. Previous studies employing PFGE have noted the combining independent restriction enzyme data into a single analysis as an approach for improving strain differentiation (7, 18). Here, the dendrogram that resulted from simultaneous analysis of binary band data from all six enzymes yielded several striking observations (Figs. 1B, 2B, 3B, and 4B). For example, six-enzyme trees for each of the four serotypes were highly resolved (SID 0.950) (Table 2). Consistent with this statistic, the ratio of nodes to isolates was more than 0.75 with an average of fewer than 3 isolates in each polytomy for all serotypes (Table 2). Relative discriminatory power among three-way combinations of restriction enzyme analyses. The simultaneous analysis of six enzymes demonstrated that discrimination of some Salmonella serotypes could be improved given a set of enzymes capable of providing a more diverse set of band classes for these isolates. A previous study has revealed a three-enzyme combination that retains, for the most part, the diversity present in the six-enzyme analysis (18). An assessment of the discriminatory power of selected three-enzyme combinations was undertaken for effective differentiation of clustered isolates within each serotype. Different sets of combinations were explored, and the combinations with the most sensitive discrimination for each serotype are presented in Table 2. The SpeI/NotI/SfiI combination for Salmonella Heidelberg and Salmonella Hadar and the SpeI/BlnI/SfiI combination for Salmonella Kentucky and Salmonella SaintPaul were found to out perform the other three-enzyme combinations (not shown) in all diversity categories. As an example, the SpeI/NotI/SfiI combination for Salmonella Heidelberg yielded an average of

3 J. Food Prot., Vol. 71, No. 10 PFGE SCHEME FOR SUBTYPING SALMONELLA 2069 FIGURE 1. Dendrograms from cluster analyses of 33 Salmonella Heidelberg isolates, using (A) a standard XbaI/BlnI combined PFGE protocol (5); and (B) a six-enzyme (XbaI, BlnI, SpeI, SfiI, PacI, and NotI) PFGE-based scheme. The dendrograms shown were derived using a UPGMA clustering algorithm and a Dice similarity coefficient. Shaded cones to the right of terminal tree branches denote polytomies within the dendrogram, while adjacent n numbers reveal the total number of strains composing that polytomy. A scale depicting percent divergence is presented above the dendrogram. 2.3 isolates per unresolved cluster, a node-to-isolate ratio of 0.86, and a SID of 0.981, with a CI of to (the CI values overlapped with CI values of SID from six-enzyme analysis of Salmonella Heidelberg). These measures are close to corresponding values in the six-enzyme analysis, indicating that the diversity present in the six-enzyme FIGURE 2. Dendrograms from cluster analyses of 27 Salmonella Kentucky isolates, using (A) a standard XbaI/BlnI combined PFGE protocol (5); and (B) a six-enzyme (XbaI, BlnI, SpeI, SfiI, PacI, and NotI) PFGE-based scheme. The dendrograms shown were derived using a UPGMA clustering algorithm and a Dice similarity coefficient. Shaded cones to the right of terminal tree branches denote polytomies within the dendrogram, while adjacent n numbers reveal the total number of strains composing that polytomy. A scale depicting percent divergence is presented above the dendrogram.

4 2070 XI ET AL. J. Food Prot., Vol. 71, No. 10 FIGURE 3. Dendrograms from cluster analyses of 27 Salmonella SaintPaul isolates, using (A) a standard XbaI/BlnI combined PFGE protocol (5); and (B) a six-enzyme (XbaI, BlnI, SpeI, SfiI, PacI, and NotI) PFGE-based scheme. The dendrograms shown were derived using a UPGMA clustering algorithm and a Dice similarity coefficient. Shaded cones to the right of terminal tree branches denote polytomies within the dendrogram, while adjacent n numbers reveal the total number of strains composing that polytomy. A scale depicting percent divergence is presented above the dendrogram. FIGURE 4. Dendrograms from cluster analyses of 27 Salmonella Hadar isolates, using (A) a standard XbaI/BlnI combined PFGE protocol (5); and (B) a six-enzyme (XbaI, BlnI, SpeI, SfiI, PacI, and NotI) PFGE-based scheme. The dendrograms shown were derived using a UPGMA clustering algorithm and a Dice similarity coefficient. Shaded cones to the right of terminal tree branches denote polytomies within the dendrogram, while adjacent n numbers reveal the total number of strains composing that polytomy. A scale depicting percent divergence is presented above the dendrogram.

5 J. Food Prot., Vol. 71, No. 10 PFGE SCHEME FOR SUBTYPING SALMONELLA 2071 TABLE 2. PFGE diversity indicators for various combinations of restriction enzymes used for Salmonella Heidelberg, Salmonella Kentucky, Salmonella SaintPaul, and Salmonella Hadar Enzyme combination Mean no. of strains/ polytomy a % Polytomous strains b Node-strain ratio c SID (CI) d Salmonella Heidelberg XbaI/BlnI (0.832, 0.956) SpeI/NotI/SfiI (0.965, 0.997) XbaI/BlnI/PacI/SpeI/NotI/SfiI (0.989, 1.00) Salmonella Kentucky XbaI/BlnI (0.735, 0.923) SpeI/BlnI/SfiI (0.933, 0.993) XbaI/BlnI/PacI/SpeI/NotI/SfiI (0.972, 1.00) Salmonella SaintPaul XbaI/BlnI (0.711, 0.941) SpeI/BlnI/SfiI (0.918, 0.996) XbaI/BlnI/PacI/SpeI/NotI/SfiI (0.943, 0.995) Salmonella Hadar XbaI/BlnI (0.602, 0.856) SpeI/NotI/SfiI (0.885, 0.961) XbaI/BlnI/PacI/SpeI/NotI/SfiI (0.923, 0.987) a Total number of polytomous isolates/number of polytomies in the dendrogram. b Number of polytomous isolates/total isolates of the same serotype analyzed. c Total number of nodes in the dendrogram/total isolates of the same serotype analyzed. d Calculated as follows: D 1 [1/N(N 1)] n j (n j 1), where j is 1 s (the total number of types described), N is the total number of isolates in the sample population, and n j is the number of isolates belonging to the jth type. CI [D 2, D 2 ]. analysis was captured for the most part by this particular three-enzyme combination for differentiating Salmonella Heidelberg. Similar results were obtained by optimal threeenzyme combinations in the other three serotypes studied here (dendrogram not shown). DISCUSSION Conventional PFGE using the restriction enzymes XbaI and BlnI has been reported widely as a successful approach for differentiating strains of Salmonella Heidelberg and Salmonella Kentucky (1, 12). Conversely Salmonella Saint- Paul and Salmonella Hadar are known to be less diverse than other Salmonella serotypes such as Salmonella Typhimurium and those mentioned above (10, 17). It is notable, however, that many isolates in each serotype selected in this study all shared levels of genetic homogeneity which confounded conventional PFGE analysis. The combined enzyme PFGE scheme presented here may serve as a useful tool for retrospective investigation, specifically in cases where Salmonella isolates are tightly associated. Similar combined analyses of PFGE fragment data have been exploited in previous studies with E. coli O157: H7 and Salmonella Enteritidis (7, 18). The simultaneous analysis of six restriction enzymes was found to allow for an epidemiologically useful partitioning of a group of homogeneous Salmonella Enteritidis isolates, potentially resulting from clonal expansion of a particular isolate. This approach was capable of uniquely partitioning most of the strains from the same locality into disparate places in the dendrogram (Figs. 1 through 4), therefore greatly improving strain discrimination. Although highly effective for strain discrimination, six-enzyme simultaneous analysis would increase the time and resources needed to complete investigations. Thus, possible three-enzyme combinations provide an alternative without any substantial loss of diversity among strains. Despite evident topological differences in the three- and six-enzyme trees, it is worth noting that the three-enzyme combination provided comparable strain differentiation capacity. However, the combination may vary for different Salmonella serotypes due to the genetic variation that uniquely delineates serotypes of Salmonella isolates. The simultaneous analysis of a three- and six-enzyme combination proved to be a beneficial scheme for the PFGE-based differentiation of closely related isolates of the four Salmonella serotypes. This approach may allow for an epidemiologically useful partitioning of Salmonella strains tightly linked both ecologically and temporally. Moreover, these findings highlight PFGE analysis as a continued essential and informative subtyping method for the molecular epidemiological investigation of these and other Salmonella pathogens. ACKNOWLEDGMENTS The authors thank Jason Abbott for assistance in data analysis. This study was supported in part by the Joint Institute for Food Safety and Applied Nutrition (JIFSAN) of the University of Maryland and the U.S. Food and Drug Administration. REFERENCES 1. Amavisit, P., P. F. Markham, D. Lightfoot, K. G. Whithear, and G. F. Browning Molecular epidemiology of Salmonella Heidelberg in an equine hospital. Vet. Microbiol. 80:85 98.

6 2072 XI ET AL. J. Food Prot., Vol. 71, No Beatty, M. E., T. N. LaPorte, Q. Phan, S. V. Van Duyne, and C. Braden A multistate outbreak of Salmonella enterica serotype Saintpaul infections linked to mango consumption: a recurrent theme. Clin. Infect. Dis. 38: Centers for Disease Control Salmonella heidelberg outbreak at a convention New Mexico. Morb. Mortal. Wkly. Rep. 35: Centers for Disease Control and Prevention Outbreak of salmonellosis associated with beef jerky New Mexico, Morb. Mortal. Wkly. Rep. 44: Centers for Disease Control and Prevention Standardized molecular subtyping of foodborne bacterial pathogens by pulsed-field gel electrophoresis. CDC training manual. Centers for Disease Control and Prevention, Atlanta. 6. Chittick, P., A. Sulka, R. V. Tauxe, and A. M. Fry A summary of national reports of foodborne outbreaks of Salmonella Heidelberg infections in the United States: clues for disease prevention. J. Food Prot. 69: Davis, M. A., D. D. Hancock, T. E. Besser, and D. R. Call Evaluation of pulsed-field gel electrophoresis as a tool for determining the degree of genetic relatedness between strains of Escherichia coli O157:H7. J. Clin. Microbiol. 41: Grundmann, H., S. Hori, and G. Tanner Determining confidence intervals when measuring genetic diversity and the discriminatory abilities of typing methods for microorganisms. J. Clin. Microbiol. 39: Hargrett-Bean, N. T., A. T. Pavia, and R. V. Tauxe Salmonella isolates from humans in the United States, MMWR CDC Surveill. Summ. 37: Hata, M., M. Suzuki, M. Matsumoto, M. Takahashi, M. Yamazaki, R. Hiramatsu, H. Matsui, K. Sakae, Y. Suzuki, and Y. Miyazaki A new clonal line of Salmonella Saintpaul having emerged and prevailed since 1999 in Aichi, Japan. Jpn. J. Infect. Dis. 56: Hunter, P. R., and M. A. Gaston Numerical index of the discriminatory ability of typing systems: an application of Simpson s index of diversity. J. Clin. Microbiol. 26: Nesse, L. L., K. Nordby, E. Heir, B. Bergsjoe, T. Vardund, H. Nygaard, and G. Holstad Molecular analyses of Salmonella enterica isolates from fish feed factories and fish feed ingredients. Appl. Environ. Microbiol. 69: O Hare, C., G. Doran, N. Delappe, D. Morris, V. Buckley, G. Corbett-Feeney, P. McKeown, W. Anderson, and M. Cormican Antimicrobial resistance and phage types of human and non-human Salmonella enterica isolates in Ireland, Commun. Dis. Public Health 7: Olsen, S. J., R. Bishop, F. W. Brenner, T. H. Roels, N. Bean, R. V. Tauxe, and L. Slutsker The changing epidemiology of Salmonella: trends in serotypes isolated from humans in the United States, J. Infect. Dis. 183: O Mahony, M., J. Cowden, B. Smyth, D. Lynch, M. Hall, B. Rowe, E. L. Teare, R. E. Tettmar, A. M. Rampling, M. Coles, R. J. Gilbert, E. Kingcott, and C. L. R. Bartlett An outbreak of Salmonella saint-paul infection associated with beansprouts. Epidemiol. Infect. 104: Tadesse, W. M., and A. Cizek The isolation of salmonellae from poultry carcasses and equipment in the poultry processing plant by means of two procedures. Vet. Med. (Praha) 39: Zhao, S., P. F. McDermott, S. Friedman, J. Abbott, S. Ayers, A. Glenn, E. Hall-Robinson, S. K. Hubert, H. Harbottle, R. D. Walker, T. M. Chiller, and D. G. White Antimicrobial resistance and genetic relatedness among Salmonella from retail foods of animal origin: NARMS retail meat surveillance. Foodborne Pathog. Dis. 3: Zheng, J., C. E. Keys, S. Zhao, J. Meng, and E. W. Brown Enhanced subtyping scheme for Salmonella Enteritidis. Emerg. Infect. Dis. 13: