(PDF) STM JJID

Jpn. J. Infect. Dis., 66, 180-188, 2013 Original Article Molecular Characterization of Salmonella enterica Serovar Typhimurium Isolated from Human, Food, and Animal Sources in Malaysia Soo Tein Ngoi1,2, Bjørn-Arne Lindstedt3, Haruo Watanabe4, and Kwai Lin Thong1,2* 1Institute of Biological Sciences, Faculty of Science, and 2Laboratory of Biomedical Science and Molecular Microbiology, Institute of Graduate Studies, University of Malaya, Kuala Lumpur, Malaysia; 3Division of Infectious Diseases Control, Norwegian Institute of Public Health, Oslo, Norway; and 4National Institute of Infectious Diseases, Tokyo 162-8640, Japan (Received November 6, 2012. Accepted February 7, 2013) SUMMARY: Salmonella Typhimurium is an important nontyphoidal Salmonella serovar associated with foodborne diseases in many parts of the world. This organism is the major causative agent of non- typhoidal salmonellosis in Malaysia. We aimed to investigate the genetic profiles of the strains isolated from clinical, zoonotic, and dietary sources in Malaysia using multilocus variable number tandem repeat analysis (MLVA) and pulsed-field gel electrophoresis (PFGE). By focusing on the 5 common variable number tandem repeat (VNTR) loci, we found that PFGE (D ＝ 0.99) was more discriminative than MLVA (D ＝ 0.76). The low MLVA score might be because of a lack of VNTR loci STTR6 (81.0z) and STTR10pl (76.2z). Both subtyping methods suggested that our S. Typhimurium strains were largely endemic with limited genetic variation. Furthermore, we observed that biphasic S. Typhimurium strains were dominant (99z) and multidrug resistance was prevalent (50z) within our sample pool. The most frequently observed phenotypes were resistance to compound sulfonamides (49z), tetracycline (51z), and streptomycin (52z). In this study, we documented the genetic relation- ship, antimicrobial resistance characteristics, and flagellar-phase dominance among S. Typhimurium strains found in Malaysia. animals were previously reported (5,6); presence of INTRODUCTION MDR strains in animals can often result in human infec- Salmonella enterica serovar Typhimurium is one of tion via the consumption of contaminated processed the important nontyphoidal Salmonella (NTS) serovars meats (7). The prevalence of these MDR strains raised associated with foodborne diseases. Notably, the con- clinical issues because these strains could complicate the tinuous rise in the number of outbreaks of foodborne currently available therapeutic options. illnesses is associated with consumption of Salmonella- S. Typhimurium (antigenic profile 4,[5],12:i:1,2) contaminated raw vegetables and fruits and poorly shows motility by means of peritrichous flagella. These cooked meats (1,2). In Malaysia, S. Typhimurium is the flagella are made up of either of the 2 types of flagellar most common causative agent of nontyphoidal sal- antigens (H:i and H:1,2), which are encoded by flagellin monellosis (3) and is frequently found in infected genes fliC and fljB, respectively (8). Phase transition of patients, contaminated food, and animal sources (2,4). biphasic S. Typhimurium is achieved by switching be- S. Typhimurium is the dominant NTS serovar (12.7z) tween the expressions of the above-mentioned flagellin isolated from poultry and livestock in this region (4). genes (9). However, since the mid-1990s, a worldwide The prevalence of S. Typhimurium poses a threat to increase has been observed in the prevalence of public health. In developing countries, the spread of the Salmonella 4,[5],12:i:–, a monophasic variant of S. pathogenic S. Typhimurium is mainly attributed to un- Typhimurium (8,10). Similar to their biphasic counter- hygienic practices during food preparation. Evidence parts, many of the monophasic variants also show mul- supporting this notion can be found in a study in which tidrug resistance (10), and therefore, pose an additional S. Typhimurium was isolated from ready-to-eat (RTE) threat to public health. Unfortunately, Salmonella food (2). 4,[5],12:i:– is antigenically similar to S. Typhimurium The emergence of multidrug-resistant (MDR) pheno- and can thus be easily misclassified as S. Typhimurium types of S. Typhimurium has been a major public health during conventional serotyping. In this regard, PCR concern since the 1990s. Detection of MDR strains in serotyping is gaining increased popularity in recent years because it affords better precision than that af- *Corresponding author: Mailing address: Microbiology forded by traditional serotyping (11,12). Division, Institute of Biological Science, Faculty of Detailed strain identification or strain typing is essen- Science, University of Malaya, 50603, Kuala Lumpur, tial for successful epidemiological investigation of S. Malaysia. Tel: ＋603-79674437, Fax: ＋603-79675908, Typhimurium outbreaks. Currently, pulsed-field gel E-mail: thongkl＠um.edu.my electrophoresis (PFGE) is considered the gold standard 180 technique for molecular subtyping of Salmonella and in 50 ml of deionized distilled water and boiling at 999C has been widely used in Malaysian studies (2,3). for 5 min. After quick centrifugation, 5 ml of the However, the PFGE method has been less successful in supernatant (¿150 ng of DNA template) was used for discriminating between genetically homogeneous strains PCR. Fluorescence-labeled primers (STTR3-F-NED, (13). The introduction of the multilocus variable num- STTR5-F-FAM, STTR6-F-PET, STTR9-F-PET, and ber tandem repeat analysis (MLVA) method has ad- STTR10pl-F-VIC) were adapted from the study by dressed this problem (14,15). Lindstedt et al. (15), with minor modifications on the In 2003, Lindstedt et al. first proposed the use of 8 dye labels so as to not affect the interpretation of variable number tandem repeat (VNTR) loci (STTR1 to results. This set of primers has been widely applied and STTR8) as the molecular markers, and subsequently, 2 was sufficiently discriminative for typing S. Typhimuri- new loci (STTR9 and STTR10pl) were used together um in several European countries (17). with 3 previously reported loci (STTR3, STTR5, and Two multiplex PCR reactions were performed: mul- STTR6) to develop a universal MLVA typing scheme tiplex-1 (M1), consisting of primer pairs for STTR3 and (14,15). In another report, Witonski et al. proposed 10 STTR6, and multiplex-2 (M2), consisting of primers for VNTR loci that could simultaneously characterize and STTR5, STTR9, and STTR10pl. The M1 PCR master discriminate between the S. Typhimurium and S. New- mix consisted of 1 × PCR buffer, 2 mmol･l－1 MgCl2, port strains (16); 3 of these loci were used in the Lind- 200 mmol･l－1 dNTP mix, 0.3 mmol･l－1 of each primer stedt study. PulseNet USA has also developed a 7-locus pair, and 2 U of Taq DNA polymerase (Promega, MLVA protocol for S. Typhimurium typing (17), 5 of Madison, Wis., USA), while the M2 PCR master mix which overlapped with the Lindstedt study (15). The 5- consisted of 1 × PCR buffer, 3 mmol･l－1 MgCl2, 200 locus MLVA scheme developed by Lindstedt et al. (15) mmol･l－1 dNTP mix, 0.3 mmol･l－1 of each primer pair, is the most widely adopted protocol in both European and 1.5 U of Taq DNA polymerase. The PCR ther- and Australian laboratories (17,18). mocycling conditions were performed as described by Although studies have been conducted on Salmonella Lindstedt et al. (15). The PCR products were diluted in in other Southeast Asian countries (19–21), currently a ratio of 1:10, and 1 ml of the diluent was mixed with 10 there is a lack of comprehensive information on the ml of Hi-DiTM formamide (Applied Biosystems, Foster genetic background of S. Typhimurium in Malaysia. City, Calif., USA) and 0.3 ml of GeneScanTM 600 LIZ} This study aims to investigate the molecular characteris- as an internal size standard (Applied Biosystems). The tics of S. Typhimurium strains from clinical, zoonotic, samples were then denatured for 5 min at 959C and and food sources collected from multiple locations in cooled to room temperature before being subjected to Malaysia. The genetic relationship among the strains capillary electrophoresis using the ABI Prism} 3130 xl was determined by MLVA and PFGE. In addition, the Genetic Analyzer (Applied Biosystems). An elec- antimicrobial resistance patterns and flagellar phases of tropherogram was generated for each sample; the S. Typhimurium strains were also determined. The data presence of a VNTR locus appeared as a colored peak. generated provide better understanding of the charac- Fragment size was determined by comparing with the teristics of the S. Typhimurium population and their size standard. The peak table generated by GeneMapper transmission dynamics in the region, emphasizing the v4.0 software (Applied Biosystems) was imported into need for implementation of efficient control measures BioNumerics v6.0 software (Applied-Maths, Kortrijk, for disease prevention. Belgium). A minimum spanning tree (MST) was con- structed using the categorical coefficient and the un- weighted pair group method with arithmetic mean MATERIALS AND METHODS (UPGMA) algorithm. The categorical coefficient gives Bacterial strains: All S. Typhimurium strains isolated an equivalent weight to a multistate character at any from 1969 to 2009 were retrieved from glycerol stocks. locus regardless of the number of repeats (23). The strains were obtained from the Laboratory of PFGE analysis: PFGE analysis was performed ac- Biomedical Science and Molecular Microbiology, cording to a previously described protocol (24). Genom- Institute of Graduate Studies, University of Malaya. ic DNA was digested with XbaI restriction enzyme (12 However, only 84 viable S. Typhimurium strains from U per plug) (Promega). XbaI-digested Salmonella clinical (n ＝ 48), zoonotic (n ＝ 16), and food (n ＝ 20) Braenderup (H9812) was used as the DNA size marker. sources were recovered. These strains were isolated Pulsotypes were analyzed using BioNumerics v6.0 from 11 states in Malaysia, namely, Perlis, Kedah, software. The variability of the strains was determined Penang, Perak, Kelantan, Selangor, Pahang, Melaka, on the basis of the Dice coefficient of similarity (F ) and Negeri Sembilan, Johor, and Sabah, and the capital UPGMA clustering analysis at 1.5z position tolerance. city, Kuala Lumpur. Eighty strains were isolated be- Statistical analysis: The discriminatory power of tween 2002 and 2009. Four older strains isolated in 1970 PFGE and MLVA was determined by Simpson's index (n ＝ 3) and 1998 (n ＝ 1) were included. All strains were of diversity (D) (25). The genetic (allelic) diversity of serotyped as S. Typhimurium according to the each VNTR locus was determined by Nei's diversity in- Kauffman-White scheme by the Salmonella Reference dex using the formula 1 － S (allele frequency)2. The Laboratory at the Institute for Medical Research, and ``null'' allele was included in the calculation of allelic Veterinary Research Institute in Malaysia. PCR target- diversity. ing a serotype Typhimurium-specific region was used to DNA hybridization: The absence of a VNTR locus confirm the genus and serotype (22). was confirmed by DNA hybridization. Approximately 5 MLVA analysis: Crude DNA for each strain was mg of genomic DNA was loaded onto a nylon membrane prepared by suspending a loopful of bacterial colonies using a PR600 24-slot blot filtration manifold (Hoefer} 181 Inc., Holliston, Mass., USA) attached to a vacuum Table 1. VNTR loci characteristics pump R-300 (Boeco, Hamburg, Germany). The digox- No. of Allelic No. of Typability igenin (DIG) nonradioactive system (Roche Applied Locus alleles diversity1) strains (z) Science, Madison, Wis., USA) for DNA hybridization STTR3 5 0.41 84 100.0 was used. The purified PCR products of the VNTR loci STTR5 12 0.70 84 100.0 were used as DNA templates to synthesize probes with STTR6 9 0.34 16 19.0 the PCR DIG Probe Synthesis Kit (Roche Applied STTR9 7 0.26 84 100.0 Science). The probes targeted the VNTR loci and their STTR10pl 9 0.41 20 23.8 flanking regions. The hybridization temperatures for each locus were optimized (549C for STTR6 and STTR9 1): Allelic diversity was calculated as 1 －∑(allele frequency)2 and 409 C for STTR10pl). Subsequently, the probe- (Nei's diversity index). target hybrids were visualized by a chemiluminescence assay using a chemiluminescent alkaline phosphatase substrate. The nylon membrane was then exposed to an The STTR3 locus contained 2 repeat units (27 and 33 X-ray film, which was subsequently developed and fix- bp, respectively); therefore, the actual amplicon length ed. was used to construct an allele string according to the Antimicrobial susceptibility test: The antimicrobial system adopted by Australian laboratories (18). susceptibility of the strains was tested using the Kirby- Altogether, 28 different allelic profiles (MLVA types) Bauer disk-diffusion method (26), and the zones of were identified (Fig. 1), and were arbitrarily designated inhibition obtained were interpreted according to the as M0001 to M0028. The discriminatory power of Clinical and Laboratory Standards Institute (CLSI) MLVA was 0.76 (Simpson's diversity index). M0003 guidelines (27). The strains were screened for resistance (allele string 03-14-00-00-510) was the most commonly to ampicillin (AMP, 10 mg), amoxicillin-clavulanic acid found MLVA type (48z) among clinical (n ＝ 23), zoo- (AMC, 30 mg), cephalothin (CEF, 30 mg), cefotaxime notic (n ＝ 6), and food (n ＝ 11) isolates (Fig. 1). (CTX, 30 mg), ceftriaxone (CRO, 30 mg), ceftazidime MLVA types M0001 (03-15-00-00-510), M0002 (CAZ, 30 mg), gentamicin (GEN, 10 mg), kanamycin (03-13-00-00-510), and M0003 (03-14-00-00-510) were (KAN, 30 mg), streptomycin (STR, 10 mg), tetracycline single-locus variants (SLVs, varied in the STTR5 locus) (TET, 30 mg), ciprofloxacin (CIP, 5 mg), nalidixic acid that predominated (62z) in our strain collection (Fig. (NAL, 30 mg), trimethoprim-sulfamethoxazole (SXT, 1). Surprisingly, we observed little genetic variation 25 mg), compound sulfonamide (SUL, 300 mg), trime- among strains from different sources isolated years thoprim (TMP, 5 mg), and chloramphenicol (CHL, 30 apart. For instance, a clinical strain (STM002/70) from mg). 1970 shared identical MLVA type with strains as recent PCR serotyping: The primers used were adapted from as 2008 (STM084/08, STM087/08, and STM088/08). the study by Cardona-Castro et al. in 2009 (12). The fliC When we studied the genetic relatedness of our strains primer pairs targeted the phase 1 H:i flagellin gene, using the MST method based on the allelic diversity of while the fljB primer pairs targeted the phase 2 H:1,2 all the 5 VNTR loci, the strains were distributed flagellin gene. PCR was performed according to a previ- throughout the tree with no clear branching dominated ously described protocol (28). The PCR products were by strains from a single source, location, or year of iso- then subjected to gel electrophoresis on a 1.5z agarose lation (Fig. 2). gel (Promega) and visualized using the Gel DocTM XR Both STTR6 and STTR10pl loci were largely absent imaging system (Bio-Rad, Berkeley, Calif., USA) after in the S. Typhimurium strains in this study. Therefore, being stained by ethidium bromide solution (Sigma DNA hybridization was performed on a selected subset Aldrich, St. Louis, Mo., USA). of the strains to confirm the presence or absence of these 2 VNTR loci. A total of 23 S. Typhimurium strains with 20 strains positive for the STTR10pl locus RESULTS were hybridized with a DIG-labeled STTR10pl probe. MLVA analyses of S. Typhimurium strains: A total The result of hybridization concurred with the MLVA of 84 S. Typhimurium strains were examined by MLVA results, i.e., probe-target hybrids were detected for using 5 VNTR loci (STTR9, STTR5, STTR6, strains positive for the STTR10pl locus, while no signal STTR10pl, and STTR3). Among the 5 commonly tested was detected for strains negative for the STTR10pl locus VNTR loci, STTR3, STTR5, and STTR9 were present (data not shown). Similarly, DNA hybridization using in all strains. Analysis of loci diversity revealed that the the DIG-labeled STTR6 probe for a selected subset of STTR5 locus had the highest number of alleles (n ＝ 23 strains with 6 strains positive for the STTR6 locus 12), whereas STTR6 and STTR9 showed the lowest showed positive signals only for strains with STTR6. diversity (Nei's diversity index of 0.34 and 0.26, respec- PFGE analysis of S. Typhimurium strains: PFGE tively). VNTR loci STTR6 and STTR10pl were absent in analysis of XbaI-digested chromosomal DNA from 84 most of the strains tested, with a typability of only S. Typhimurium strains yielded 67 pulsotypes, which 19.0z and 23.8z, respectively (Table 1). were arbitrarily designated as X0001 to X0067 (Fig. 1). Next, an allelic profile was constructed for each Each pulsotype contained 12 to 20 DNA fragments, and strain. The allelic profile was a string of 5 numbers, their sizes ranged from 20.5 to 1,135 kb. The genetic each indicating the number of repeat units in an individ- diversity (F value) of the strains ranged from 0.6 to 1.0, ual locus (except STTR3), arranged in the order of as determined by the PFGE pattern similarity. X0009 (n STTR9-STTR5-STTR6-STTR10pl-STTR3. The num- ＝ 2), X0012 (n ＝ 4), X0013 (n ＝ 5), X0016 (n ＝ 3), ber 00 indicates the absence of a specific VNTR locus. X0046 (n ＝ 2), X0050 (n ＝ 4), X0052 (n ＝ 3), and 182 Fig. 1. Dendrogram showing a cluster analysis of 84 Salmonella Typhimurium strains based on the PFGE profiles of the XbaI-digested chromosomal DNA of the bacterial strains. The dendrogram was constructed using the Dice coefficient (F ) and UPGMA clustering parameters at 1.5z position tolerance. Roman numerals I to VII denote strain clusters with a cutoff value of F ＝ 0.80. Strains marked with an asterisk (*) are multidrug resistant. The year of isolation is indicated as the last 2 digits of the strain code (e.g., STM001/70 was isolated in 1970). The alle- lic profile indicates the allele numbers of the VNTR loci (except STTR3, where the number indicates the amplicon length), arranged in the order STTR9-STTR5-STTR6-STTR10pl-STTR3. The number 00 indicates that no ampli- con was obtained. The resistotypes of the strains correspond with data shown in Table 2. The sources of the strains are indicated as letters, where ``C'' represents clinical strains, ``Z'' represents zoonotic strains, and ``F'' represents food strains. JHR, Johor; KDH, Kedah; KLT, Kelantan; KUL, Kuala Lumpur; MLK, Melaka; NGS, Negeri Sembilan; PHG, Pahang; PLS, Perlis; PNG, Penang; PRK, Perak; SBH, Sabah; SLG, Selangor. 183 Fig. 2. Minimum spanning tree for MLVA of 84 S. Typhimurium strains. A circle denotes an MLVA type. The thickness and dotting of the lines indicated the allelic differences between the MLVA types; thus, a thick line indi- cates a single-locus variant, a thin line indicates allelic differences at 2 different VNTR loci, and a dotted line denotes allelic differences at 3 or more VNTR loci. The sources of the strains are indicated as letters in the circles, where ``C'' represents clinical strains, ``Z'' represents zoonotic strains, ``F'' represents food strains, ``C, Z'' represents a mix of clinical and zoonotic strains, ``C, F'' represents a mix of clinical and food strains, and ``C, Z, F''' represents a mix of strains from all 3 sources. 184 X0058 (n ＝ 2) were common pulsotypes that appeared resistance to AMP (25z), CEF (26z), SUL (49z), in more than 1 strain. The remaining 60 pulsotypes were TET (51z), and STR (52z). The presence of MDR S. unique. The Simpson's index of diversity for PFGE was Typhimurium strains (defined as strains resistant to 3 or 0.99. more different classes of antimicrobial agents) in our A dendrogram based on 67 pulsotypes revealed 7 strain collection was as high as 50z. These included clusters (clusters I to VII) of closely related strains, with strains that were isolated from 1970 to 2008 from all a Dice coefficient of similarity of 0.80 (Fig. 1). Cluster I sources. The MDR strains mostly originated from urban (F ＝ 0.80) was the major group, consisting 57z (n ＝ and densely populated areas such as Kuala Lumpur and 48) of the studied strains. The majority of the strains in Selangor. this cluster were from clinical sources (n ＝ 30), fol- A total of 34 resistotypes were identified from our lowed by food (n ＝ 11) and zoonotic (n ＝ 7) sources, screening and were arbitrarily designated as R01 to R34 and were isolated between 1970 and 2008. Within this (Table 2). Resistotype R25 was the most dominant, cluster, 44z of the strains were isolated from the capital showing simultaneous resistance to STR, SUL, and city Kuala Lumpur (n ＝ 21). Clusters II (F ＝ 0.83), III TET (R-type SSuT). Overall, the S. Typhimurium (F ＝ 0.87), IV (F ＝ 0.81), VI (F ＝ 0.83), and VII (F ＝ strains in our study showed high susceptibility to third- 0.96) were smaller and contained 2 to 5 strains. generation cephalosporins. Less than 5z of the strains Antimicrobial drug susceptibility of S. Typhimurium tested were resistant to CTX, CAZ, and CRO (Table 2). strains: Of the 84 S. Typhimurium strains screened, 27 Quinolone resistance was observed; specifically, 20z of strains were susceptible to all the 16 antimicrobial our samples were resistant to NAL, CIP, or both. agents tested (Table 2). We observed high rates of Determination of S. Typhimurium flagellar phase: Table 2. Antimicrobial susceptibility profiles of S. Typhimurium strains No. of Resistotype Resistance profile1) isolates R01 AMC AMP CEF — CAZ — CHL — NAL GEN KAN STR TET SXT SUL — 1 R02 AMC AMP CEF — — — CHL — NAL — KAN STR TET SXT SUL TMP 1 R03 — AMP — — — — CHL CIP NAL GEN KAN STR TET SXT SUL TMP 1 R04 AMC AMP CEF CTX CAZ CRO — CIP — — — STR TET — SUL — 1 R05 — AMP CEF — — — CHL — NAL — KAN STR TET SXT SUL TMP 1 R06 — AMP CEF — — — CHL — NAL GEN KAN STR TET SXT SUL — 2 R07 AMC AMP CEF CTX CAZ CRO CHL — — — — — — — SUL — 1 R08 AMC AMP CEF — — — CHL — — — KAN STR TET — SUL — 1 R09 — AMP CEF CTX CAZ CRO — — — — — STR TET — SUL — 1 R10 — AMP CEF — — — CHL — — — KAN STR TET SXT SUL — 1 R11 — AMP — — — — CHL — — — KAN STR TET SXT SUL TMP 2 R12 AMC AMP CEF — — — CHL — — — KAN — TET — SUL — 1 R13 AMC — CEF — — — — — — — KAN STR TET — SUL TMP 1 R14 — AMP — — — — — — NAL — — STR TET SXT SUL TMP 1 R15 — AMP CEF — — — — — — — — STR TET SXT SUL TMP 1 R16 — AMP — — — — CHL — — — — STR — SXT SUL TMP 1 R17 — — CEF — — — — — — — — STR TET SXT SUL TMP 1 R18 AMC AMP CEF — — — — — NAL — — — — — SUL — 1 R19 — — CEF — — — — — — — — STR TET — SUL TMP 1 R20 — AMP — — — — — — NAL GEN — — TET — — — 1 R21 — — — — — — — — NAL GEN — STR TET — — — 1 R22 — — CEF — — — — — NAL — — STR TET — — — 1 R23 — — CEF — — — — — — — — STR TET — SUL — 1 R24 — — CEF — — — — — — — — STR TET — — — 2 R25 — — — — — — — — — — — STR TET — SUL — 15 R26 — — CEF — — — — — NAL — — — — — — — 1 R27 — — — — — — — — — — — STR — — SUL — 4 R28 — — — — — — — — NAL — — — TET — — — 1 R29 — — — — — — — — — — — — TET — SUL — 1 R30 — AMP — — — — — — — — — — — — — — 1 R31 — — CEF — — — — — — — — — — — — — 1 R32 — — — — — — — — NAL — — — — — — — 2 R33 — — — — — — — — — — — STR — — — — 2 R34 — — — — — — — — — — — — TET — — — 2 $ — — — — — — — — — — — — — — — — 27 1): The resistance profiles excluded antimicrobial agents that exhibited intermediate resistance. AMC, amoxicillin-clavulanic acid; AMP, ampicillin; CEF, cephalothin; CTX, cefotaxime; CIP, ciprofloxacin; CAZ, ceftazidime; CHL, chloramphenicol; CRO, ceftriaxone; GEN, gentamicin; KAN, kanamycin; STR, streptomycin; TET, tetracycline; SXT, trimethoprim-sul- famethoxazole; NAL, nalidixic acid; SUL, compound sulfonamide; TMP, trimethoprim; $, sensitive to all antimicrobial agents tested. 185 The majority of our samples (99z, n ＝ 83) were types among the strains. This suggested that the S. biphasic S. Typhimurium. Only 1 strain (STM032/04) Typhimurium from farm animals might infect humans, of clinical origin isolated in 2004 was identified as a with contaminated food serving as the vehicle. monophasic variant of S. Typhimurium, lacking the Overall, in this study, we observed a high percentage fljB H:1,2 allele. (50z) of MDR S. Typhimurium strains isolated from various sources and locations. All earlier strains (isolat- ed in 1970 and 1998) showed multidrug resistance. DISCUSSION However, because of the 30-year gap in the years of iso- S. Typhimurium infection is a common foodborne in- lation, we could not conclude whether the MDR pheno- fection in developing countries; however, there is a lack types had persisted since 1970 or whether there was a of systematic characterization of existing strains in relapse between 1970 and 2002. More strains isolated many parts of Southeast Asia. We report here the first between these 2 years should be included to provide a MLVA-based subtyping of S. Typhimurium strains in better understanding of the prevalence of MDR pheno- Malaysia. Nei's diversity index indicated that the diver- types in Malaysia over the years. sity indices for STTR5, STTR6, STTR9, and STTR10pl Previous reports on the occurrence and characteriza- were lower in our sample population as compared to the tion of MDR S. Typhimurium were source-specific values reported by Lindstedt et al. (15). The absence of (2,3). Therefore, the present study provides a more ex- STTR6 and STTR10pl in most strains (81.0z and tensive analysis, by encompassing strains from clinical, 76.2z, respectively) further contributed to the low zoonotic, and food sources. The food-derived and diversity. To rule out the possibility of false-negative clinically-derived MDR strains mostly originated from results for these 2 loci, we performed DNA hybridiza- developed and densely populated areas such as Kuala tion, and confirmed that STTR6 and STTR10pl were in- Lumpur and Selangor. The high genetic proximity deed absent (data not shown). The low prevalence of among these strains suggested the probable spread of these loci may be attributed to the association of STTR6 the pathogen from food to humans. This phenomenon with Gifsy–1 prophage and of STTR10pl with the plas- is not uncommon in Malaysia, because the isolation of mid pSLT (15). We postulated that most Malaysian S. pathogenic Salmonella serovars from RTE food was Typhimurium strains lacked these 2 genetic entities; fu- frequently reported (2). Additionally, half of the zoo- ture studies are needed to elucidate this finding. There notic strains screened in this study showed MDR pheno- seems to be a difference between Asian and European types, suggesting that the occurrence of MDR strains in strains. Notably, 30z of the strains obtained from food, especially meat and poultry products, might be Southeast Asia (Thailand, Malaysia, and Indonesia) due to the use of antimicrobial agents in animal feeds at were negative for STTR6 and 97z were negative for livestock farms. Previous studies showed that food STTR10pl (B. A. Lindstedt, personal communication). animals indeed served as a reservoir of MDR Salmonella In contrast, only 16z and 48z of the strains from Nor- (29–31). way were negative for STTR6 and STTR10pl, respec- The most common type of MDR S. Typhimurium tively (15). On the contrary, STTR3 showed higher strain in Malaysia was R-type SSuT (27.4z), followed diversity in our samples. Nevertheless, the overall low by a hexa-resistant pattern R-type ACKSSuT (11.9z; diversity indices (0.26–0.70) for all 5 VNTR loci suggest- strains showed simultaneous resistance to AMP, CHL, ed that these loci are less discriminative if used as KAN, STR, SUL, and TET). The resistance to these molecular markers for subtyping S. Typhimurium in conventional antimicrobial agents has also been report- this geographical region. The inclusion of other VNTR ed previously (32–34). Mixed results were obtained for loci that are yet to be examined may prove useful for resistance to newer antimicrobial agents such as quino- more precise strain mapping. lone; resistance to NAL was shown by a relatively high By comparing the MLVA and PFGE methods using number of strains (18z); however, resistance to CIP the same strains, we found that MLVA subtyping was was rare (2 z ). Resistance to third-generation less discriminative (D ＝ 0.76) than PFGE (D ＝ 0.99); cephalosporins was very low in our study (CTX, n ＝ 3; nevertheless, we still found a strong correlation. For ex- CAZ, n ＝ 4; and CRO, n ＝ 3). These findings indicat- ample, the PFGE cluster I coincided with the SLVs ed that the newer drugs remained effective in this region M0001, M0002, M0003, and M0004, while M0005 and during the study period. However, we should also be M0006 (SLVs) corresponded with the PFGE cluster V. wary of the emergence of resistant strains, although the Both typing methods suggested a rather homogeneous frequency is low. More recent strains should be included S. Typhimurium population circulated within Malaysia in future studies so as to accurately reflect the an- during this period of study. The MST constructed (Fig. timicrobial resistance of currently circulating S. 2) did not show clear branching of strains exclusively by Typhimurium populations in Malaysia. a single source, location, or year of isolation in the over- The worldwide emergence and spreading of all distribution of the MLVA types. Because MLVA is monophasic Salmonella 4,[5],12:i:– has been reported reported to be sensitive to the origins of the strains (15), since 1990s (8). In Asia, the isolation of this clinically this observation further supports the inference regard- important NTS from clinical samples has been reported ing the genetic homogeneity of local S. Typhimurium in Thailand and Taiwan (35,36) but not in Malaysia. In strains in Malaysia. This should probably be expected, this study, we found only 1 monophasic variant of S. as nontyphoidal salmonellosis caused by S. Typhimuri- Typhimurium (STM032/04) from a hospitalized patient um is endemic in this region. The close genetic relation- in 2004. It has been known that the majority of ship between clinical, food, and zoonotic strains was Salmonella 4,[5],12:i:– strains reported in European seen in the sharing of similar allelic profiles and pulso- and Asian countries showed MDR phenotypes (8). We 186 found that the monophasic STM032/04 strain in this minced beef, Norway, October–November 2005. Euro Surveill., study was resistant to AMP, GEN, and TET, in addi- 10, 45. 8. Moreno Switt, A.I., Soyer, Y., Warnick, L.D., et al. (2009): tion to NAL; this finding was similar to those obtained Emergence, distribution, and molecular and phenotypic charac- for monophasic strains in Thailand (37). The PCR de- teristics of Salmonella enterica serotype 4, 5, 12: i:–. Foodborne tection of a serotype Typhimurium-specific sequence Pathog. Dis., 6, 407–415. (22) adopted in this study failed to distinguish between 9. Aldridge, P.D., Wu, C., Gnerer, J., et al. (2006): Regulatory pro- monophasic strains and their biphasic counterparts. In a tein that inhibits both synthesis and use of the target protein con- trols flagellar phase variation in Salmonella enterica. Proc. Natl. nutshell, these findings indicated that the monophasic Acad. Sci. USA, 103, 11340–11345. Salmonella 4,[5],12:i:– is relatively uncommon in this 10. Hopkins, K., Kirchner, M., Guerra, B., et al. (2010): Mul- region, and so far, there have been no reports of an out- tiresistant Salmonella enterica serovar 4,[5],12:i:– in Europe: a break caused by this serovar in Malaysia. However, we new pandemic strain? Euro Surveill., 15, 1–9. 11. Barco, L., Lettini, A.A., Ramon, E., et al. (2011): A rapid and acknowledge a caveat here, which is that Salmonella sensitive method to identify and differentiate Salmonella enterica 4,[5],12:i:– is commonly misclassified as S. Typhimuri- serotype Typhimurium and Salmonella enterica serotype um by both conventional serotyping and PCR detection 4,[5],12:i:– by combining traditional serotyping and multiplex methods, and this could lead to underreporting of cases. polymerase chain reaction. Foodborne Pathog. Dis., 8, 741–743. A larger sample size and more careful analysis in the fu- 12. Cardona-Castro, N., S áanchez-Jim áenez, M., Lavalett, L., et al. (2009): Development and evaluation of a multiplex polymerase ture are needed to determine and monitor the spreading chain reaction assay to identify Salmonella serogroups and sero- of this strain. types. Diagn. Microbiol. Infect. Dis., 65, 327–330. To our knowledge, this study represents the first de- 13. Murphy, T.M., McNamara, E., Hill, M., et al. (2001): Epidemio- tailed report of the molecular characteristics of S. logical studies of human and animal Salmonella Typhimurium DT 104 and DT 104b isolates in Ireland. Epidemiol. Infect., 126, Typhimurium strains in Malaysia, studying the genotyp- 3–9. ic and phenotypic relationships among the strains 14. Lindstedt, B.A., Heir, E., Gjernes, E., et al. (2003): DNA finger- through MLVA and PFGE subtyping and determining printing of Salmonella enterica subsp. enterica serovar the antimicrobial resistance patterns and flagellar Typhimurium with emphasis on phage type DT104 based on vari- phases of the strains. We conclude that the MLVA and able number of tandem repeat loci. J. Clin. Microbiol., 41, 1469–1479. PFGE typing of the S. Typhimurium strains showed 15. Lindstedt, B.A., Vardund, T., Aas, L., et al. (2004): Multiple- moderate congruency in strain clustering, but PFGE (D locus variable-number tandem-repeats analysis of Salmonella en- ＝ 0.99) was more discriminative than MLVA (0.76) on terica subsp. enterica serovar Typhimurium using PCR multiplex- the basis of the 5 VNTR loci screened. A multiple-typ- ing and multicolor capillary electrophoresis. J. Microbiol. Methods, 59, 163–172. ing approach encompassing the gold standard PFGE 16. Witonski, D., Stefanova, R., Ranganathan, A., et al. (2006): and the high-throughput MLVA assays for detailed Variable-number tandem repeats that are useful in genotyping strain differentiation is suggested for future epidemio- isolates of Salmonella enterica subsp. enterica serovars logical study of the pathogen. Finally, in this study, the Typhimurium and Newport. J. Clin. Microbiol., 44, 3849–3854. biphasic S. Typhimurium strains were dominant, and 17. Larsson, J.T., Torpdahl, M., Petersen, R.F., et al. (2009): De- velopment of a new nomenclature for Salmonella typhimurium most of these strains showed MDR phenotypes. multilocus variable number of tandem repeats analysis (MLVA). Euro Surveill., 14, 1–5. 18. Gilbert, G.L. (2008): Using MLVA to type strains of Salmonella Acknowledgments This study was funded by University of Typhimurium in New South Wales. N.S.W. Public Health Bull., Malaya Post Graduate Research Fund (PPP) (PS319/2010B), Univer- 19, 29–31. sity of Malaya Research Grant (RG017-09BIO), and Japanese Nation- 19. Amavisit, P., Boonyawiwat, W. and Bangtrakulnont, A. (2005): al Institute of Infectious Diseases Grant (57-02-03-1009). Characterization of Salmonella enterica serovar Typhimurium We thank Dr Ser Huy Teh and Dr Cindy Shuan Ju Teh for their and monophasic Salmonella serovar 1,4,[5],12:i:– isolates in technical assistance. Thailand. J. Clin. Microbiol., 43, 2736–2740. 20. Kusumaningrum, H. and Dewanti-Hariyadi, R. (2012): Mul- Conflict of interest None to declare. tidrug resistance among different serotypes of Salmonella isolates from fresh products in Indonesia. Int. Food Res. J., 19, 57–63. REFERENCES 21. Ling, M., Goh, K., Wang, G., et al. (2002): An outbreak of mul- tidrug-resistant Salmonella enterica subsp. enterica serotype 1. Brandl, M.T. (2006): Fitness of human enteric pathogens on Typhimurium, DT104L linked to dried anchovy in Singapore. plants and implications for food safety 1. Annu. Rev. Epidemiol. Infect., 128, 1–5. Phytopathol., 44, 367–392. 22. Alvarez, J., Sota, M., Vivanco, A.B., et al. (2004): Development 2. Modarressi, S. and Thong, K.L. (2010): Isolation and molecular of a multiplex PCR technique for detection and epidemiological sub-typing of Salmonella enterica from chicken, beef and street typing of Salmonella in human clinical samples. J. Clin. foods in Malaysia. Sci. Res. Essays, 5, 2713–2720. Microbiol., 42, 1734–1738. 3. Douadi, B., Thong, K.L., Watanabe, H., et al. (2010): Charac- 23. Ramisse, V., Houssu, P., Hernandez, E., et al. (2004): Variable terization of drug-resistant Salmonella enterica serotype number of tandem repeats in Salmonella enterica subsp. enterica Typhimurium by antibiograms, plasmids, integrons, resistance for typing purposes. J. Clin. Microbiol., 42, 5722–5730. genes, and PFGE. J. Microbiol. Biotech., 20, 1042–1052. 24. Thong, K.L., Lai, K.S. and Goh, Y.L. (2003): Technical im- 4. Roseliza, R., Maswati, M.A., Hasnah, Y., et al. (2011): Identifi- provement in DNA preparation for pulsed-field gel electrophore- cation of Salmonella serotypes isolated from meat samples in sis. Malaysian J. Sci., 22, 55–59. Malaysia. Malaysian J. Vet. Res., 2, 59–64. 25. Hunter, P.R. and Gaston, M.A. (1988): Numerical index of the 5. Wright, J.G., Tengelsen, L.A., Smith, K.E., et al. (2005): Mul- discriminatory ability of typing systems: an application of tidrug-resistant Salmonella Typhimurium in four animal facili- Simpson's index of diversity. J. Clin. Microbiol., 26, 2465–2466. ties. Emerg. Infect. Dis., 11, 1235–1241. 26. Cavalieri, S.J., Harbeck, R.J., McCarter, Y.S., et al. (2005): 6. Gebreyes, W.A. and Altier, C. (2002): Molecular characterization Manual of Antimicrobial Susceptibility Testing. American Soci- of multidrug-resistant Salmonella enterica subsp. enterica serovar ety for Microbiology, Washington, D.C. Typhimurium isolates from swine. J. Clin. Microbiol., 40, 27. Clinical and Laboratory Standards Institute (CLSI) (2010): Per- 2813–2822. formance standards for antimicrobial susceptibility testing. 20th 7. Isakbaeva, E., Lindstedt, B.A., Schimmer, B., et al. (2005): Informational Supplement, 30, M100-S20. CLSI, Wayne, Pa. Salmonella Typhimurium DT104 outbreak linked to imported 28. Nori, E. and Thong, K.L. (2010): Differentiation of Salmonella 187 enterica based on PCR detection of selected somatic and flagellar Typhimurium among patients from several European countries. antigen. Afr. J. Microbiol. Res., 4, 871–879. Clin. Microbiol. Infect., 10, 904–910. 29. Angulo, F.J. and Griffin, P.M. (2000): Changes in antimicrobial 34. Heir, E., Lindstedt, B.A., Nygård, I., et al. (2002): Molecular resistance in Salmonella enterica serovar Typhimurium. Emerg. epidemiology of Salmonella Typhimurium isolates from human Infect. Dis., 6, 436–438. sporadic and outbreak cases. Epidemiol. Infect., 128, 373–382. 30. Angulo, F.J., Nargund, V.N. and Chiller, T.C. (2004): Evidence 35. Chiu, C.H., Su, L.H., Chu, C.H., et al. (2006): Detection of mul- of an association between use of anti-microbial agents in food tidrug-resistant Salmonella enterica serovar Typhimurium phage animals and anti-microbial resistance among bacteria isolated types DT102, DT104, and U302 by multiplex PCR. J. Clin. from humans and the human health consequences of such Microbiol., 44, 2354–2358. resistance. J. Vet. Med. B, 51, 374–379. 36. Pornruangwong, S., Sriyapai, T., Pulsrikarn, C., et al. (2008): 31. McEwen, S.A. and Fedorka-Cray, P.J. (2002): Antimicrobial use The epidemiological relationship between Salmonella enterica and resistance in animals. Clin. Infect. Dis., 34, 93–106. serovar Typhimurium and Salmonella enterica serovar 32. Antunes, P., Machado, J. and Peixe, L. (2006): Characterization 4,[5],12:i:– isolates from humans and swine in Thailand. of antimicrobial resistance and class 1 and 2 integrons in Southeast Asian J. Trop. Med. Public Health, 39, 288–296. Salmonella enterica isolates from different sources in Portugal. J. 37. Boonmar, S., Bangtrakulnonth, A., Pornruangwong, S., et al. Antimicrob. Chemother., 58, 297–304. (1998): Significant increase in antibiotic resistance of Salmonella 33. Ethelberg, S., Lisby, M., Torpdahl, M., et al. (2004): Prolonged isolates from human beings and chicken meat in Thailand. Vet. restaurant-associated outbreak of multidrug-resistant Salmonella Microbiol., 62, 73–80. 188