Previous Article | Next Article ![]()
Journal of Clinical Microbiology, July 2006, p. 2449-2457, Vol. 44, No. 7
0095-1137/06/$08.00+0 doi:10.1128/JCM.00019-06
Division of Animal and Food Microbiology, Office of Research, Center for Veterinary Medicine, U.S. Food and Drug Administration, Laurel, Maryland 20708
Received 4 January 2006/ Returned for modification 16 February 2006/ Accepted 22 April 2006
|
|
|---|
|
|
|---|
DNA-based strain-typing methods for bacterial pathogens play a crucial role in understanding infectious-disease transmission, tracking, and response and have been used widely to distinguish Salmonella clinical isolates recovered from animals, food-borne disease, and nosocomial infections (21, 30, 53, 58). Pulsed-field gel electrophoresis (PFGE) and antimicrobial susceptibility typing (AST) are two commonly used methods for studying microbial epidemiology and trends in the antibiotic resistance of bacteria. PFGE is currently used by the CDC PulseNet surveillance program and is generally accepted as the "gold standard" for molecular typing of Salmonella (7, 18, 34, 41, 57). Due to its limitations, many studies have compared PFGE to other genetic typing methods in attempts to identify more powerful tools for epidemiological investigations and evolutionary analyses. Comparisons have been made with various typing schemes, including serotyping and phage typing, repetitive-element PCR, multilocus variable-number tandem repeat analysis, amplified fragment length polymorphism, antibiotic susceptibility typing, and DNA sequence typing (4, 14, 19, 21, 22, 29, 32, 44, 45, 59).
Multilocus sequence typing (MLST) is based on allelic differences in the nucleotide sequences of housekeeping or virulence genes among bacterial strains (31). MLST methods and databases have been developed for a growing number of clinically important bacterial pathogens, including Staphylococcus aureus, Streptococcus pneumoniae, Campylobacter jejuni, Escherichia coli O157:H7, and Salmonella serotypes (1, 10, 14, 26, 29, 39, 52, 54, 55). MLST has shown various degrees of utility as a discriminatory typing method for several bacterial pathogens, including E. coli O157:H7 and several Salmonella serotypes (14, 29, 39, 40, 52, 54). Fewer MLST data are available for S. enterica serovar Newport, as only three publications have reported on a limited number of isolates (29, 52, 54). In this study, 81 Salmonella enterica serotype Newport isolates from clinically ill animals, animal-derived foods, and human infections were analyzed by antimicrobial susceptibility typing, PFGE, and MLST, in order to compare the discriminatory powers of the methods.
|
|
|---|
Antimicrobial susceptibility testing. Antimicrobial MICs were determined using the Sensititre automated antimicrobial susceptibility system (Trek Diagnostic Systems, Westlake, Ohio) and interpreted using the CLSI (formerly NCCLS) standards (37, 38). The antimicrobials tested included amikacin, amoxicillin-clavulanic acid, ampicillin, cefoxitin, ceftiofur, ceftriaxone, cephalothin, chloramphenicol, ciprofloxacin, gentamicin, kanamycin, nalidixic acid, streptomycin, sulfamethoxazole, tetracycline, and trimethoprim-sulfamethoxazole. The quality control organisms used included E. coli ATCC 35218, Enterococcus faecalis ATCC 29212, Staphylococcus aureus ATCC 29213, and Pseudomonas aeruginosa ATCC 27853 to ensure that all antimicrobial agents were appropriately quality controlled, except for streptomycin, for which official quality control standards have not been set (37, 38). Chi-square analysis and logistical-regression analysis were performed to indicate significant differences.
PFGE.
Pulsed-field gel electrophoresis was performed according to the procedures developed by the CDC (35) and as previously described (59). Briefly, agarose-embedded DNA was digested with 50 U of XbaI (Boehringer Mannheim, Indianapolis, IN) overnight in a water bath at 37°C. The restriction fragments were separated by electrophoresis in 0.5x Tris-borate-EDTA buffer at 14°C for 18 h using a Chef Mapper electrophoresis system (Bio-Rad, Hercules, CA) with pulse times of 2.16 to 63.8 s. The gels were stained with ethidium bromide, and DNA bands were visualized with UV transillumination (Bio-Rad). Salmonella enterica serotype Newport am01144 was used as the control strain (59). Isolates presenting DNA smear patterns were retested using plugs digested with XbaI and subjected to electrophoresis in buffer containing 50 µM of thiourea in 0.5x Tris-borate-EDTA buffer. Interpretation of DNA fingerprint patterns was accomplished using Bionumerics 4.0 software (Applied Maths, Austin, TX). The banding patterns were compared using Dice coefficients with a 1.5% band position tolerance. Isolate relatedness was determined using the unweighted pair group method using arithmetic averages (UPGMA). Simpson's index of diversity (D) was used as an indicator of the discriminatory power of each method and is calculated according to the following formula: D = 1 (
n(n 1)/N(N 1)), where D is the diversity, N is the total number of strains in the sample, and n is the number of strains in each type (25).
MLST. Genomic-DNA templates were prepared using the UltraClean Microbial DNA Kit (MoBio Laboratories, Inc., Carlsbad, CA) according to the manufacturer's instructions. Seven genes were chosen for MLST: aroC, dnaN, hemD, hisD, purE, sucA, and thrA, to allow comparison to an existing Salmonella MLST database (http://web.mpiib-berlin.mpg.de/mlst). Amplification protocols detailed in the database were used in this study, including primer sequences and annealing temperatures. Amplifications for all genes were carried out with approximately 0.2 µg DNA template, 250 µM (each) deoxynucleoside triphosphates, 2.5 mM MgCl2, 25 pmol of primers, and 1 U of Gold Taq polymerase (Perkin-Elmer, Foster City, Calif.) in 50-µl reaction mixtures. PCR cycling conditions were a 10-min hold at 94°C, followed by 34 cycles of 94°C for 1 min, 55°C for 1 min, and 72°C for 1 min, and a final extension at 72°C for 5 min. Products were separated by 1.5% agarose gel electrophoresis and visualized with ethidium bromide staining and UV illumination with a gel documentation system (Gel Doc 2000; Bio-Rad, Hercules, Calif).
Amplification products were purified using a 96-well Millipore Multi-screen Filter plate according to the manufacturer's recommendations (Millipore, Billerica, MA). Amplicons were resuspended in 50 µl of nuclease-free water, and DNA sequences were determined using the BigDye Terminator v3.1 cycle-sequencing kit (Applied Biosystems, Foster City, Calif.) according to the manufacturer's instructions. Reaction mixtures contained 20 mM of primer, approximately 20 ng of DNA template, 1x BigDye Terminator v3.1 sequencing buffer, and 1 µl of terminator ready-reaction mixture in a 5-µl total volume. Cycle-sequencing conditions were a 96°C hold for 1 min and 25 cycles of 96°C for 10 s, 50°C for 5 s, and 60°C for 4 min. After the cycling was completed, the sequenced products were precipitated with 20 µl 75% ethanol for 30 min and centrifuged for 45 min in a Centra CL3 centrifuge (Thermo Electron Corporation, Waltham, Mass.) at 4,000 x g. Dried DNA pellets were resuspended in 7 µl of Hi-Di Formamide (Applied Biosystems, Foster City, Calif.), and the products were analyzed on an ABI PRISM DNA analyzer 3700 (Applied Biosystems, Foster City, Calif.).
MLST data analysis. Sequences were edited, and complementary sense and antisense fragments were aligned using Bionumerics 4.0 software (Applied Maths, Austin, TX). The sequences were submitted to the MLST database website (http://web.mpiib-berlin.mpg.de/mlst) and assigned existing or novel allele type numbers and sequence type numbers defined by the database. This multimicroorganism database defines a novel allele type if it contains one or more nucleotide changes from existing allele sequences. Composite sequence types (STs) are assigned based on the set of allele types derived from each of the seven loci. STs were analyzed for relatedness using the eBURST v3 program (http://eburst.mlst.net; 15).
|
|
|---|
0.10). |
View this table: [in a new window] |
TABLE 1. Antimicrobial resistance phenotypes of Salmonella Newport isolates from different animal and food types
|
16 µg/ml). These isolates were comprised of 16 strains from cattle (41%), 10 strains from human infections (26%), 6 strains from swine (15%), 3 strains from chickens (8%), and 1 strain from a ground turkey meat sample. None of the strains obtained from clinically ill turkeys exhibited the MDR-AmpC phenotype. The high rate of resistance among the cattle isolates and the high rate of susceptibility among the turkey isolates were significantly different at a level of 0.05 (since the confidence interval at 95% did not contain 1). The most common resistance profile was represented by 22% of the collection (n = 18). The indicator of discrimination as computed using Simpson's index of diversity was 0.78, where a score of 0.90 or greater is considered a high level of diversity.
![]() View larger version (55K): [in a new window] |
FIG. 1. UPGMA analysis of PFGE profiles of S. enterica serovar Newport isolates (identified by unique CVM numbers) showing PFGE fingerprints (62% similar), state of origin (STATE), source of the isolate (SOURCE), antimicrobial susceptibility type, and ST. Major clusters are marked A, B, and C. Resistance in the AST is denoted by a black box. Intermediate resistance in the AST is denoted by a gray box, and susceptibility is denoted by blank space. Antimicrobial abbreviations are as follows: amoxicillin-clavulanic acid, AUG; ampicillin, AMP; cefoxitin, FOX; ceftiofur, TIO; ceftriaxone, AXO; cephalothin, CEP; chloramphenicol, CHL; gentamicin, GEN; kanamycin, KAN; streptomycin, STR; sulfamethoxazole, SMX; tetracycline, TET; and trimethoprim-sulfamethoxazole, COT.
|
Certain PFGE pattern clusters correlated well with antimicrobial resistance phenotypes. For example, cluster A was almost exclusively comprised of isolates exhibiting the MDR-AmpC phenotype (34/39 isolates) from all animal origins, with only one isolate originating from turkey (CVM 17015) (Fig. 1). This cluster differed markedly from cluster C, containing 22 S. enterica serovar Newport isolates, 19 (86%) of which were susceptible to all tested antimicrobials. With regard to specific PFGE patterns associated with S. enterica serovar Newport isolates recovered from different animals and retail meats, 13 patterns were generated from 20 human isolates, 14 patterns from 20 cattle isolates, 11 patterns from 15 turkey/ground-turkey isolates, 10 patterns from 16 swine/pork chop isolates, and 6 patterns from 10 chicken isolates.
When PFGE profile and antimicrobial susceptibility phenotypes were analyzed together by UPGMA, two major clusters were identified that displayed 31% profile and pattern similarity (data not shown). Simpson's index of diversity, when calculated for the PFGE method, resulted in a score of 0.97, which was the highest diversity observed in this study when one method was analyzed. The index of diversity was slightly increased when antimicrobial susceptibility typing was combined with PFGE in the diversity equation, resulting in a score of 0.978.
MLST profiles. In order to compare single nucleotide polymorphisms against the whole genome profile provided by PFGE, MLST was conducted on all 81 isolates comparing a partial DNA sequence of seven genes (aroC, dnaN, hemD, hisD, purE, sucA, and thrA). Sequence data from both strands were assembled using Bionumerics 4.0 software (Applied Maths, Austin, TX) and entered into the Salmonella enterica MLST database (http://web.mpiib-berlin.mpg.de/mlst) for comparison to existing allele types. Between two and four alleles were identified among the 81 S. enterica serovar Newport isolates (Table 2). One novel allele type was identified in the sucA allele set. Overall, the seven-gene MLST scheme defined 12 sequence types, with one ST (ST 45) encompassing 61.7% of the S. enterica serovar Newport collection in this study (Table 2). The second most common sequence type was ST 118, which included 12.3% of the isolates, followed by STs 115 and 116 (7.4% each). eBURST v3 analysis separated the STs into two groups and one singleton (ST 123), with ST 45 as the founder of complex 1 and ST 118 as the founder of complex 2 (Fig. 2). The founders of both complexes were heptalocus variants and therefore were complexed separately and indicate evolutionary distance. The singleton, ST 123, was not grouped with either complex, as it shared only three loci with ST 45 and four loci with ST 118, so it was not closely enough related to either complex to belong. This ST appears to represent a combination of the two complexes; however, it was not represented by enough strains within the sequence type to substantiate this suggestion.
|
View this table: [in a new window] |
TABLE 2. ST definitions based on allele type for each of seven loci sequenced and assigned by the Salmonella enterica databasea
|
![]() View larger version (9K): [in a new window] |
FIG. 2. eBURST v3 diagram of 81 S. enterica serovar Newport sequence types. The size of the circle marking each ST indicates the relative number of isolates belonging to that ST. Single-locus variants of each ST are connected by one line. The founder of each complex is defined as the ST that is related to the greatest number of STs in the population that differs at a single locus. The founder of complex 1 is ST 45, and the founder of complex 2 is ST 118. Complex 2 has a subgroup founder (ST 115), or an ST that is not the founder that has at least two single-locus variants. AST information for ST 45, ST 118, ST 116, and ST 115 is listed.
|
Concordance of antimicrobial susceptibility typing, PFGE, and MLST. Composite analysis of all three methods revealed a relationship between the pulsed-field fingerprint, the antimicrobial susceptibility phenotype, and the ST (Fig. 1). For example, the largest cluster of isolates (cluster A) was made up largely of MDR-AmpC-positive strains and was comprised mainly of strains with ST 45 and ST 116 (Fig. 1). In fact, eight of nine isolates comprising the largest PFGE pattern in cluster A were identified as ST 45, with the remaining human isolate being ST 116. These two STs are single-locus variants at the sucA locus.
Cluster B consisted of a subcluster of 14 isolates from cluster A originating from clinically ill turkey, swine, humans, and retail ground-turkey samples. All cluster B S. enterica serovar Newport isolates were ST 45, with the exception of one human isolate from California (ST 116). Cluster C (Fig. 1) was made up almost entirely of pansusceptible isolates and contained mostly ST 118 and ST 115. However, the second largest PFGE pattern in cluster C contained six isolates recovered from either clinically ill chickens (five) or cattle (one) but comprised four different sequence types (ST 118, ST 123, ST 120, and ST 115). STs 118 and 115 are single-locus variants at the sucA locus, just as STs 45 and 116 are single-locus variants at that locus (Fig. 2). However, the two groups (STs 45 and 116 versus STs 118 and 115) are widely different by MLST type and differ from each other at six of the seven loci, reinforcing the evolutionary distance between the isolates in clusters A and B and those in cluster C.
|
|
|---|
MLST is increasingly being used as an evolutionary and epidemiological tool, with schemes being developed for a number of bacterial pathogens (1, 52, 54, 55). MLST has been used for numerous evolutionary analyses of large and small populations, such as identifying highly clonal lineages of bacteria, as in the case of Mycobacterium tuberculosis (51). Epidemiological analyses have also been conducted using large and small bacterial populations and have often been focused on the emergence of virulent phenotypes as identified by MLST (27) and/or the emergence of antibiotic-resistant phenotypes of a particular bacterial population (23). Although it has been reported that MLST of several housekeeping genes provides a satisfactory level of discrimination among diverse Salmonella isolates (29), recent studies suggest that it may not be suitable for distinguishing closely related strains within a particular serovar, due to high sequence identity and slow accumulation of variations of their housekeeping genes (8, 13, 14, 47, 52). We therefore characterized a diverse collection of 81 S. enterica serovar Newport isolates originating from a variety of clinically ill animals, humans, and retail meats via MLST and compared the results to those of two classic typing schemes, antimicrobial susceptibility typing and PFGE.
The majority of S. enterica serovar Newport isolates were either susceptible to all tested antimicrobials (n = 33) or characterized as the MDR-AmpC phenotype, exhibiting resistance to at least 9 of the 16 antimicrobials tested (n = 39). No correlation could be determined between specific antimicrobial susceptibility phenotypes and S. enterica serovar Newport origin. The majority of both pansusceptible and MDR-AmpC isolates were recovered from all of the different foods and animal types, including humans. However, two interesting observations were noted. S. enterica serovar Newport isolates recovered from either clinically ill turkeys (n = 8) or retail ground turkey (n = 7) were considerably more susceptible than isolates from other animals. In contrast, 85% (17/20) of S. enterica serovar Newport isolates from ill cattle displayed the MDR-AmpC phenotype. These data support previous studies of the epidemiology of S. enterica serovar Newport MDR-AmpC infections in humans and links with cattle or beef products (3, 22, 45, 50, 59).
Pulsed-field gel electrophoresis was used to assess genetic relatedness and revealed 43 distinct genetic patterns and three major clusters (A to C) among the 81 S. enterica serovar Newport isolates. Forty-eight percent (39/81) of all isolates grouped in cluster A, which was also genetically diverse, comprising 19 different PFGE patterns. Certain PFGE clusters showed good correlation with the antimicrobial susceptibility phenotypes. For example, the majority of MDR-AmpC strains grouped together in cluster A (n = 34/39), whereas the majority of pansusceptible strains grouped in cluster C (n = 22/33). The demarcation of PFGE patterns and clusters between the MDR-AmpC isolates and pansusceptible isolates has been previously noted (3, 22, 59) and suggests that the recent emergence of MDR-AmpC S. enterica serovar Newport is due to the clonal expansion of a limited number of genetically related strains that have acquired plasmid-mediated resistance genes. A similar finding was recently reported by Alcaine et al., who postulated that plasmid-mediated ceftiofur-resistant Salmonella evolved by independent emergence and clonal spread (2).
Twelve sequence types were defined with MLST, with the majority of isolates (61.7%) grouped in ST 45. The next most common sequence types included ST 118 (12.3%) and STs 115 and 116 (7.4% each). Ninety-seven percent of MDR-AmpC S. enterica serovar Newport isolates were characterized as either ST 45 (n = 32/39) or ST 116 (n = 6/39) and were found only in PFGE clusters A and B. These two sequence types are single-locus variants at the sucA locus, and because they differ in only one of the seven loci sequenced, are therefore closely related. MLST also provided more discriminatory power among S. enterica serovar Newport isolates recovered from retail meats on two occasions than did PFGE. S. enterica serovar Newport CVM 33000, isolated from a retail ground-turkey sample from California, was indistinguishable by PFGE from two S. enterica serovar Newport isolates recovered from clinically ill cattle from Missouri but differed by MLST (ST 125 versus ST 45, respectively). S. enterica serovar Newport CVM 22697, isolated from a retail ground-turkey sample from Maryland, was indistinguishable by PFGE from two S. enterica serovar Newport isolates recovered from retail pork chops from the same grocery store and collection time, suggesting contamination of meats, possibly at the retail establishment. Nevertheless, MLST resolved the ground-turkey S. enterica serovar Newport isolate as ST 45, whereas the two pork chop isolates were characterized as ST 116. ST 45 is different from ST 116 at the sucA locus, and sequence analysis showed that the different alleles at this locus (sucA12 and sucA39) differ at 7 bp throughout the allele. Therefore, a spontaneous mutation event causing the change in allele type and ST is unlikely. On the other hand, MLST did not resolve the 40% of clinically ill cattle isolates that were resistant to kanamycin in addition to four or more antimicrobials. All eight were identified as ST 45, and all but three isolates varied in their AST and PFGE types. Therefore, AST phenotype and PFGE resolved these isolates into widely scattered subtypes within cluster A, while MLST grouped them in the same complex.
In contrast to early work by Kotetishvili et al. (29), the seven-gene MLST scheme did not further discriminate either animal origins or geographic locations among our S. enterica serovar Newport collection compared with PFGE results. This lack of discrimination versus PFGE has been recently reported with other Salmonella serotypes as well (13, 14, 40, 52). Recently, Torpdahl et al. (54) presented MLST data on 25 serotypes of S. enterica using the same seven-gene MLST scheme used in this study and found that overall, discrimination was not improved within a serotype compared with PFGE and amplified fragment length polymorphism data. This study included only three S. enterica serovar Newport isolates from veterinary and human sources, and upon MLST analysis, they belonged to sequence types 31, 45, and 46, in contrast to the current study, where over 60% of the S. enterica serovar Newport isolates belonged to ST 45. It is important to remember, even in light of the recent study analyzing 110 isolates of 25 different serotypes (54), that only a limited number of the over 2,500 Salmonella serovars and isolates within them have been characterized using MLST, and most published reports have used a different set of housekeeping genes in their respective MLST schemes (2, 14, 29, 47). As this is the first report characterizing a particular Salmonella serovar using the seven genes selected for the global MLST Salmonella database housed at the Max-Planck Institut für Infektionsbiologie (http://web.mpiib-berlin.mpg.de/mlst/), further study using this scheme with larger numbers of isolates and additional serovars is warranted to fully explore the utility of MLST for epidemiological purposes.
Combining the results obtained from all three methods yielded more information than any of the methods alone and elicited a higher index of diversity (0.986, or 98.6%), indicating that two randomly selected samples would be identified as different types more often when all three typing methods were employed and that identical isolates would be identified as the same type. The partitioning of S. enterica serovar Newport isolates into two major PFGE clusters showed an interesting association with antimicrobial susceptibility and MLST sequence type, as shown in Fig. 1. Based on UPGMA clustering of PFGE fingerprints, sequence types were separated into two groups, which were very closely related intracluster (STs 45 and 116, single-locus variants) but more distantly related intercluster (ST 45 versus ST 118, different at every locus, or heptalocus variants). This study supports previous findings that PFGE can predict clonal-complex designations and that these two DNA-based methods give rise to very similar clustering results (20). MLST revealed that the separation of this population into two major groups was characterized by the relatedness of these isolates and could be identified using seven highly conserved housekeeping genes. This partnership of MLST and PFGE or antimicrobial susceptibility type has not been extensively applied to food-borne pathogens, although it has been applied to methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus (12). Here, it was shown that antimicrobial susceptibility phenotypes were associated with specific MLST types, and particular sequence types were proposed as predictors of whether S. aureus isolates would exhibit resistance to methicillin. Similar associations between the antimicrobial susceptibility phenotype and MLST types were also observed in our current study, where particular sequence types displayed a multidrug resistance phenotype and others were represented by an almost completely pansusceptible phenotype. This observation showed the relationship between these two methods, which previously had been used as separate components instead of as tiers in the levels of epidemiological classification. Although MLST did not add substantial discriminatory power to this study (97.8% typeability using PFGE and AST versus 98.6% typeability using all three methods), the relationship between the three methods provides a phylogenetic aspect to a pure discriminatory-power study.
MLST separated 81 diverse strains from 27 states and five sources into two discrete complexes and one "intermediate" singleton, which shared three or four alleles with each founder. In this study, where 88% of the isolates were recovered from clinically ill animals, polygenetic data that separate related versus distantly related isolates based on conserved sequences give more information about the further expansion of those pathogens than simple banding-pattern differences alone. This, coupled with the fact that the two complexes were associated with antimicrobial resistance or almost complete susceptibility, shows that MLST is a useful polyphyletic and epidemiological tool for tracking pathogens of veterinary or human importance. Due to the moderate to slow accumulation of mutations within the chosen seven housekeeping genes, discrimination between very closely related isolates, such as those within a PFGE cluster, has been shown to be low, but it can provide the information for reliable evolutionary relationships on a more global scale (8). However, for applications in local epidemiological outbreaks, where tracking of particular isolates is imperative, the higher discriminatory power of PFGE may be more useful.
This study is part of our long-range goal to identify strain-typing schemes that provide the analytical power needed to help ascertain the origins of strains for which the vehicle or vector is unknown. While MLST does provide unambiguous data that are easily comparable among different laboratories and provides phylogenetic-relationship inferences that PFGE data cannot provide, the discriminatory ability of MLST may be somewhat problematic due to the high sequence conservation of the genes used to characterize S. enterica serovar Newport isolates in this study.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»