- Email: [email protected]
Molecular analysis of Salmonella enterica subsp. enterica serovar Agona isolated from slaughter pigs
Veterinary Microbiology 112 (2006) 43–52 www.elsevier.com/locate/vetmic
Molecular analysis of Salmonella enterica subsp. enterica serovar Agona isolated from slaughter pigs Geovana Brenner Michael a,b, Marisa Cardoso b, Stefan Schwarz a,* b
a Institut fu¨r Tierzucht, Bundesforschungsanstalt Fu¨r Landwirtschaft (FAL), Ho¨ltystr. 10, 31535 Neustadt-Mariensee, Germany Departamento de Medicina Veterina´ria Preventiva, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
Received 24 August 2005; received in revised form 30 September 2005; accepted 4 October 2005
Abstract Salmonella enterica subsp. enterica (S.) serovar Agona plays an important role in Brazil as causative agent of salmonellosis in food-producing animals – in particular, pigs and poultry – as well as in humans. A total of 45 S. Agona isolates collected from slaughter pigs at three different slaughterhouses in Southern Brazil was investigated in this study for their phenotypic and genotypic relatedness. For this, the antimicrobial susceptibility patterns and the phage types were determined. Molecular analysis included the determination of plasmid profiles as well as the analysis of XbaI- and BlnI-generated macrorestriction patterns. Moreover, a novel typing method called subtracted restriction fingerprinting (SRF) was successfully applied to the S. Agona isolates. Based on all properties determined, a dominant clonal group comprising 33 of the 45 isolates was identified. Members of this group were susceptible to all antimicrobials tested, did not carry plasmids, shared the same phage type and were closely related or even indistinguishable by their EcoRI–PauI SRF patterns as well as their XbaI and BlnI macrorestriction patterns. Members of this clonal group were identified at all 3 slaughterhouses at variable frequencies and originated from pig herds raised in 15 different cities in Southern Brazil which were located up to 450 km apart from each other. Since the S. Agonacarrying slaughter pigs were from various integrated production lines, the results of this study suggest that a specific clonal group of S. Agona had entered numerous pig production lines. This observation supports the requirement for the establishment of monitoring and control programmes in Brazil which should also include molecular techniques to better trace the dissemination of S. Agona and other Salmonella serovars in pigs and other food-producing animals. # 2005 Elsevier B.V. All rights reserved. Keywords: Macrorestriction analysis; Subtracted restriction fingerprinting; Plasmid profiles; Antimicrobial susceptibility patterns; Phage typing
1. Introduction
* Corresponding author. Tel.: +49 5034 871 241; fax: +49 5034 871 246. E-mail address: [email protected] (S. Schwarz).
Salmonella enterica subsp. enterica serovar Agona (S. Agona) is considered a zoonotic bacterium, which has been associated with various infections in humans and animals worldwide. Transmission of S. Agona
0378-1135/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2005.10.011
44
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
from animals to humans commonly occurs by ingestion of contaminated beef (Sorensen et al., 2002), pork (Escartin et al., 2000), chicken and turkey meat (Beli et al., 2001a,b; Synnott et al., 1998), although contaminated vegetarian food (Killalea et al., 1996), cereals (CDC, 1998) or tea (Koch et al., 2005) have also been identified as sources of infections. Human infections due to S. Agona have been described mostly as local outbreaks, although nation-wide outbreaks and international outbreaks have been reported. In Brazil, S. Agona isolates seem to play a particular role since they have been found not only among the 10 serovars most commonly isolated from cases of human salmonellosis, but also represented the fourth most common serovar among the non-human isolates (Tavechio et al., 2002). Slaughter-age pigs (Bessa et al., 2001; Oliveira et al., 2002; Michael et al., 2005) as well as poultry (Fuzihara et al., 2000) from Brazil have been found to be infected with S. Agona. Most of the Brazilian S. Agona isolates from humans and animals showed extended antimicrobial resistance patterns (Oliveira et al., 2002; Mulvey et al., 2004; Michael et al., 2005). These facts stress the hazard that S. Agona of animal origin represents for human health and underlines the importance to achieve a better understanding of the epidemiology of S. Agona infections (Davies and Funk, 1999). In the present study, we investigated S. Agona isolates obtained from slaughter pigs at three different slaughterhouses in Southern Brazil for their genomic relatedness. Epidemiological background data of the isolates were put in context with the typing data to gain better insight into the geographical and temporal distribution of particular isolates.
2. Materials and methods
terhouses A and B were located 96 km away from each other and received animals from farms located in almost the same geographic area. In contrast, the slaughterhouse C was located about 300 km away from the others and received animals from a different region. During an 11-month period from 09/1999 to 07/2000, 600 samples (mesenteric lymph nodes and intestinal content) from 300 individual animals representing different herds of different farms were collected at four distinct dates at each of the three slaughterhouses. A total of 41 (6.83%) S. Agona isolates was identified from all 3 slaughterhouses during this period. The presence of S. Agona in both, mesenteric lymph nodes and intestinal content, was detected in only 4 of the 300 animals sampled. Of the 41 S. Agona isolates, 18 originated from slaughterhouse A, 7 from slaughterhouse B and 16 from slaughterhouse C. Another 3 (0.56%) S. Agona isolates were obtained from 535 samples of mesenteric lymph nodes, head lymph nodes/tonsils and/or intestinal content collected from 230 individual animals at slaughterhouse B between 09/2000 and 10/2001. During that period, 165 samples from meat products collected at slaughterhouse B were analyzed and a single sample (0.61%) from minced meat was found to be contaminated with S. Agona. For various reasons, most of which were far beyond our control, it was not possible to obtain samples from slaughterhouses A and C during the second sampling period. Of the 45 isolates included in this study, 22 were from mesenteric lymph nodes, 21 from intestinal content, 1 from head lymph nodes/tonsils and 1 from minced meat. All isolates were identified as S. Agona by routine laboratory procedures (Michael et al., 2003). Serological tests were performed at the National Reference Centre, Instituto Oswaldo Cruz, Rio de Janeiro, Brazil, by standard slide agglutination using commercially available antisera.
2.1. Bacterial isolates 2.2. Antimicrobial susceptibility testing S. Agona isolates, all exhibiting the antigenic formula 4,12:f,g,s:, were obtained from apparently healthy slaughter pigs and in one case from a meat product (minced meat) at three slaughterhouses (A–C) in Southern Brazil (Bessa et al., 2001). Each slaughterhouse was supplied by finishing farms in a vertically integrated production system. The slaugh-
All S. Agona strains were investigated for their antimicrobial resistance by the agar disk diffusion test using the following disks (Oxoid, Wesel, Germany; Becton Dickinson, Heidelberg, Germany): ampicillin (AMP, 10 mg), amikacin (AMI, 30 mg), gentamicin (GEN, 10 mg), kanamycin (KAN, 30 mg), neomycin
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
(NEO, 30 mg), streptomycin (STR, 10 mg), tobramycin (TOB, 10 mg), tetracycline (TET, 30 mg), compound sulfonamides (SUL, 300 mg), sulfamethoxazole/trimethoprim (SXT, 23.75/1.25 mg), trimethoprim (TMP, 5 mg), chloramphenicol (CHL, 30 mg), nalidixic acid (NAL, 30 mg) and ciprofloxacin (CIP, 5 mg). In addition, spectinomycin (SPT, 100 mg), minocycline (MIN, 30 mg) or florfenicol (FFC, 30 mg) were also tested with strains resistant to streptomycin, tetracycline or chloramphenicol, respectively. In vitro susceptibility testing was performed and the results evaluated according to the document M31-A2 of the Clinical and Laboratory Standards Institute (NCCLS, 2002). 2.3. Isolation of plasmid DNA and whole cell DNA Whole cell DNA of the S. Agona isolates were extracted as previously described (Schwarz and Liebisch, 1994). For plasmid preparation, a modification of the alkaline lysis procedure (Kado and Liu, 1981) was applied which proved to be particularly suitable for the detection of large plasmids (Schwarz and Liebisch, 1994). Gel electrophoretic presentation of the plasmid profiles followed previously described protocols. The plasmids of Escherichia coli V517 (Macrina et al., 1978), the virulence plasmid of S. Typhimurium LT2 (McClelland et al., 2001) and the 150 kb resistance plasmid R55 from Klebsiella pneumoniae (Cloeckaert et al., 2001) served as size standards for the determination of plasmid sizes. 2.4. Phage typing All 45 S. Agona isolates were sent to the National Reference Center for Salmonella at the Robert Koch Institute in Wernigerode, Germany, for phage typing. A recently established S. Agona-specific phage typing system that allowed the detection of 52 phage types was applied to the Brazilian S. Agona isolates (Rabsch et al., 2005). 2.5. Macrorestriction analysis Whole cell DNA for pulsed-field gel electrophoresis was prepared as described earlier (Schwarz and Liebisch, 1994; Liebisch and Schwarz, 1996). DNAcontaining slices of the agarose plugs were incubated
45
for 4 h in the presence of 20 units of XbaI (Roche, Mannheim, Germany) or 10 units of BlnI (AmershamBuchler, Braunschweig, Germany). The respective DNA fragments were separated by pulsed-field gel electrophoresis (SeaKem GTG, 1%, w/v, Biozym, Hess, Oldendorf, Germany) in a CHEF DR III system (BioRad, Mu¨nchen, Germany) at 5.6 V/cm with 0.5 TBE as running buffer. The pulse times for XbaI digests were increased from 10 to 30 s during the first 11 h and subsequently from 30 to 50 s during the next 13 h; those for the BlnI digests were increased from 7 to 12 s for the first 11 h and from 20 to 65 s for the following 13 h. The XbaI or BlnI fragments of S. Typhimurium LT2 served as size standards (Liu et al., 1993). The gel was stained with ethidium bromide (2 mg/ml, Sigma, Deisenhofen, Germany) for 4 min, destained in distilled water for 12 min and photographed under UV-illumination. 2.6. Subtracted restriction fingerprinting analysis (SRF) Extraction of genomic DNA for SRF analysis followed a previously described protocol (Terletski et al., 2003a) with slight modifications. In brief, double digestion of 1.8 mg of the whole cell DNA was performed in y+Tango restriction buffer (Fermentas, St. Leon-Rot, Germany) using 3 U EcoRI (Fermentas) and 7 U PauI (Fermentas) in a total reaction volume of 40 ml. The digestion reaction was incubated at 37 8C for 2 h. The subsequent fill-in reaction of the sticky ends was performed by using 2 ml of a dNTP mixture to result in 40 mM dATP (Sigma–Aldrich, Taufkirchen, Germany), 40 mM dGTP (Sigma–Aldrich), 2 mM DIG-11-dUTP (Roche, Mannheim, Germany) and 2 mM biotin-14-dCTP (Invitrogen, Karlsruhe, Germany). Then, 1 U of Klenow enzyme (Fermentas) was added and the mixture was incubated for 10 min at room temperature. The unincorporated dNTPs were removed with spin micro columns (Quantum PREP1 PCR Kleen Spin Columns; Bio-Rad Laboratories GmbH, Munich, Germany) filled with 1 ml prehydrated Sephadex1 G-50 (Amersham Bioscience, Freiburg, Germany) by centrifugation at 3000 rpm for 2 min at room temperature. The biotinylated fragments were captured by adding 40 ml (20 mg) of the pre-washed streptavidin-coated magnetic particles (Roche, Mannheim, Germany) to each tube. The
46
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
tubes were incubated with continuous agitation for 2 h at room temperature. Separation was performed in a magnetic rack (Dynal MPC-P-12; Dynal GmbH, Hamburg, Germany) and the resulting suspensions containing the Dig-labeled fragments were transferred to spin micro columns and centrifuged as described above. The final suspensions were analyzed by agarose gel electrophoresis with 1% (w/v) gels at 100 V/cm for 11 h with 0.5 TBE as running buffer. The digoxigenin-labeled DNA molecular weight marker III (Roche) was used as size standard. After gel electrophoretic separation, the DNA fragments were blotted and the digoxigenin-labeled fragments immunochemically detected as previously described (Terletski et al., 2003a,b). 2.7. Analysis of the fragment patterns The macrorestriction patterns obtained after XbaI or BlnI digestion as well as the SRF patterns were analyzed using the GelCompar software package (Applied Maths, Kortrijk, Belgium). For this, the images were saved as TIFF files and imported into the GelCompar software. The similarities between the
profiles were calculated using the Dice coefficient, with a maximum position tolerance of 1.2%. The patterns were clustered by using the unweighted pair group method with arithmetic averages (UPGMA). Since the large plasmids present in five S. Agona isolates were found to be cut into fragments smaller than 90 kb by XbaI as well as BlnI (data not shown), only macrorestriction fragments larger than 90 kb were used for the comparison of the patterns (Olsen et al., 1997). A macrorestriction or SRF pattern was considered as unique when it differed from the next closely related one by three or more bands. Patterns that differed from others by only one or two bands were considered as variants of pre-existing patterns. The most common pattern obtained with each method received the number 0 0 1 and according to the similarity detected by the cluster analysis the next closely related patterns received the following numbers. Lower case letters after the numbers indicated variants of a pattern. The alphabetical order was used to address the level of similarity of these variants to the corresponding pattern. Thus, SAX001 indicated the most common restriction pattern (0 0 1) of S. Agona obtained by
Table 1 Phenotypic and genotypic characteristics of the S. Agona isolates Total no. of isolates
Isolates obtained from slaughterhouse
Plasmid profiles
Phage types
Resistance patternsa
BlnI
SRF patterns EcoRI–PauI
XbaI 29 1 3
A (14)b, B (1), C (14)b A A (1), B (2)
SAX001 SAX001a SAX001
SAB001 SAB001 SAB003
SAS001 SAS001 SAS001
PP0 PP0 PP0
PT07 PT07 PT07
S S S
C
SAX002
SAB001b
SAS001
PP0
PT18
S
B A (2), C (1)
SAX003 SAX003
SAB001a SAB001b
SAS004 SAS001
PP1 PP0
PT40 PT40
R1 S
1c
B
SAX003a
SAB001b
SAS002
PP2
ut/O1-
S
c
B B
SAX003b SAX004
SAB001b SAB002
SAS003 SAS001a
PP0 PP4
PT40 PT40
S S
1 1d 1d
B B B
SAX004a SAX005 SAX005a
SAB002 SAB004 SAB005
SAS005 SAS003a SAS006
PP3 PP0 PP5
PT16 PT16 PT16
R2 S R3
1 2 3
1 1
c
Macrorestriction patterns
a S: Susceptible to all antimicrobial tested; R1: resistant to AMP-CHL-KAN-MIN/TET-STR-SUL-TMP; R2: resistant to CHL-MIN/TETSTR-SPT-SUL-TMP; R3: resistant to CHL-MIN/TET-SUL-TMP. b Include three matched pairs of isolates from mesenteric lymph nodes and intestinal content, two pairs from slaughterhouse A and one pair from slaughterhouse C (29 isolates from 26 individual animals). c Isolates from the second sampling period. d From the same animal, the susceptible isolate is from intestinal content and the multiresistant one from mesenteric lymph nodes.
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
XbaI digestion, whereas SAX001a indicated a variant of this pattern. In this way, SAB indicated the patterns obtained with BlnI digestion and SAS the patterns obtained by the SRF analysis. The plasmid profiles (PP) were indicated by numbers 0–5 and phage types (PT) by the respective number in the phage typing system used.
47
isolate from minced meat proved to be non-typeable (ut/O1) by this phage typing system. PT07 isolates were found in all three slaughterhouses, while PT40 isolates were obtained from slaughterhouses A and B. The only PT18 isolate was from slaughterhouse C, whereas the three PT16 isolates were detected in samples from slaughterhouse B (Table 1). 3.4. Macrorestriction patterns
3. Results 3.1. Antimicrobial resistance patterns Of the 45 S. Agona isolates included in this study, 41 (91.1%) were susceptible to all antimicrobials tested, whereas the remaining 4 isolates (Table 1) were multiresistant. Among the latter four isolates, the following resistance patterns were observed: CHLMIN/TET-SUL-TMP (one isolate), CHL-MIN/TETSTR-SPT-SUL-TMP (one isolate) and AMP-CHLKAN-MIN/TET-STR-SUL-TMP (two isolates). All multiresistant isolates were obtained from mesenteric lymph nodes or intestinal content samples collected at different time periods in slaughterhouse B. 3.2. Plasmid profiles A total of 39 S. Agona isolates (86.6%) were plasmid-free (PP0). The six plasmid-bearing strains exhibited five different plasmid profiles (PP1–PP5). The four multiresistant strains harbored a ca. 150 kb plasmid. While two of these isolates carried only this large plasmid (PP1), the other two isolates harboured additional plasmids of ca. 10 kb (PP3) or 40 kb (PP5), respectively. A single plasmid of approximately 10 kb (PP4) and a plasmid of approximately 90 kb (PP2) were identified in another two isolates. The plasmidbearing S. Agona isolates originated from mesenteric lymph nodes, intestinal content and minced meat and were obtained during different times only at slaughterhouse B (Table 1). 3.3. Phage types Among the 45 S. Agona isolates included in this study, 33 isolates (73.3%) were assigned to PT07. The types PT16, PT18 and PT40 were found in three, one and seven isolates, respectively. Only the S. Agona
XbaI digestion showed 5 major patterns and 5 variant patterns each consisting of 10–14 fragments in the size range approximately between 90 and 830 kb (Fig. 1a). All isolates shared at least seven common fragments. The most common pattern, SAX001, was shared by 32 isolates (71.1%), while 5 isolates exhibited pattern SAX003. All other XbaI-generated fragment patterns were represented by single isolates. Cluster analysis revealed that 41 isolates were grouped in a main cluster with an overall similarity of 77.8% (Fig. 1b). All strains of the most common pattern were found in this major cluster. The remaining four isolates formed a second cluster with an overall similarity of 83.0% (Fig. 1b). The largest difference of 10 fragments was seen between the patterns SAX001 and SAX005. BlnI digestion resulted in 5 major patterns and 2 variant patterns comprising 9–14 bands in the same size range as seen for XbaI (Fig. 1c). Four common bands were seen in the BlnI patterns of all isolates. Only the most common pattern SAB001, which was represented by 30 isolates, showed variants. The variant patterns SAB001a and SAB001b were seen with two and six isolates, respectively. Two isolates exhibited pattern SAB002, three isolates SAB003, whereas single isolates represented patterns SAB004 and SAB005. Cluster analysis grouped 43 isolates in a main cluster with an overall similarity of 78.5% (Fig. 1d). The remaining two isolates (SAB004 and SAB005) shared 88.9% similarity and formed a minor cluster (Fig. 1d). The maximum difference of 13 fragments was seen between the pattern SAB005 and pattern SAB003. A total of 29 S. Agona isolates (64.4%) was indistinguishable by their XbaI (SAX001) and BlnI (SAB001) macrorestriction patterns. These isolates were found at variable frequencies in animals slaughtered at each of the three slaughterhouses and
48
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
Fig. 1. Macrorestriction patterns of S. Agona isolates obtained with the restriction enzymes XbaI (a) and BlnI (c). Lane letters and numbers correspond to the different patterns in Table 1. Lane M contains XbaI-digested (a) or BlnI-digested (c) DNA of S. Typhimurium LT2. Dendograms showing the relationships of the different XbaI (b) and BlnI (d) patterns. Similarity analysis was performed using the Dice coefficient and the clustering was generated by UPGMA.
were recovered during the entire sampling period 09/ 1999–07/2000. 3.5. Subtracted restriction fingerprinting patterns EcoRI/PauI double digestion produced six major SRF fragment patterns and two variant patterns (Fig. 2). Each of these SRF patterns consisted of 33–50 bands ranging in size between ca. 1 and 19 kb. A total of 37 S. Agona strains (82.2%) were
indistinguishable by their SRF patterns. In addition to the most common pattern SAS001, only the pattern SAS004 was shared by two isolates. All other SRF patterns were represented by single isolates. All SRF patterns had approximately 30 bands in common. As seen with macrorestriction analysis, two main clusters were found with the SRF analysis (Fig. 2). Forty-one isolates were grouped in the major cluster (91.4% similarity), while the remaining four isolates representing patterns SAS004, SAS005 and SAS006
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
49
Fig. 2. (a) EcoRI–PauI SRF patterns of S. Agona isolates. Lane letters and numbers correspond to the patterns in Table 1. Lane M contains a digoxigenin-labeled DNA molecular weight marker. (b) Dendogram showing the relationships of the different SRF patterns. Similarity analysis was performed using the Dice coefficient and the clustering was generated by UPGMA.
formed a minor cluster with 85.6% of similarity. The largest difference of 19 bands was detected between the patterns SAS001 and SAS006. 3.6. Correlation between the characteristics and the origin of the isolates All phenotypic and genotypic properties of the 45 S. Agona isolates are summarized in Table 1. It became obvious that 29 S. Agona isolates were indistinguishable on the basis of all properties determined: susceptible to all antimicrobials tested, plasmid-free, indistinguishable by their SRF and XbaI and BlnI macrorestriction patterns and by the phage type. Another four isolates showed only slight variations in either their XbaI or their BlnI patterns,
but were otherwise indistinguishable from the aforementioned 29 S. Agona isolates. These 33 isolates were considered to represent members of the same clonal group. Members of this dominant clonal group were obtained mainly from animals slaughtered at the slaughterhouses A (16 isolates) and C (14 isolates), whereas only three isolates of this clonal group were detected at slaughterhouse B. A comparison with the origin of the herds from which these isolates were obtained revealed that the 16 isolates from slaughterhouse A originated from herds of 5 different cities, each of the 3 isolates from slaughterhouse B was from a different city and the 14 isolates from slaughterhouse C were from 7 different cities. In three individual animals (two from slaughterhouse A and one from slaughterhouse C), S. Agona isolates representing this
50
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
clonal group were detected in mesenteric lymph nodes and intestinal content. Among the remaining 12 S. Agona isolates, 6 of the 7 PT40 isolates as well as the single PT18 isolate and the non-phage typeable isolate were closely related or even indistinguishable in their XbaI, BlnI or SRF patterns. These isolates originated from herds of individual cities and were isolated at all three slaughterhouses between 09/1999 and 07/2000, but also at slaughterhouse B between 09/2000 and 10/ 2001. The three PT16 isolates, two of which originated from mesenteric lymph nodes or intestinal content of the same animal, showed considerable variability in all phenotypic and genotypic traits.
4. Discussion In this retrospective study, porcine S. Agona isolates from three different slaughterhouses located in Southern Brazil were characterized by macrorestriction and SRF analysis, phage typing, plasmid profile analysis and antimicrobial resistance patterns. Macrorestriction analysis has already been used successfully for the analysis of S. Agona outbreaks (Threlfall et al., 1996; Taylor et al., 1998; Lindqvist et al., 2002). Previous studies showed that other molecular methods used for the differentiation of Salmonella isolates, such as ribotyping, 16S–23S intergenic spacer PCR or IS200 typing proved to be unsuitable for S. Agona (Threlfall et al., 1996; Lagotolla et al., 1996). To complement our macrorestriction data, we, therefore, used SRF analysis (Terletski et al., 2004). This technique has previously proved to be suitable for the differentiation of S. Choleraesuis, S. Typhimurium and S. Dublin (Terletski et al., 2003b). Macrorestriction analysis with two enzymes as well as SRF analysis produced a number of different fragment patterns among the S. Agona isolates, thus, confirming that these methods are able to discriminate between the S. Agona isolates. The observation, that 33 isolates were either indistinguishable or closely related by their genomic fragment patterns obtained by macrorestriction and SRF analyses, strongly suggests that these isolates represent members of the same clonal group. Additional tests included plasmid analysis and antimicrobial susceptibility patterns. A previous study
which focused on the resistance genes present in the four multiresistant isolates revealed that all resistance properties detected in these isolates were associated with conjugative ca. 150 kb plasmids (Michael et al., 2005). The use of a recently established phage typing system for S. Agona supplemented the other strain characteristics determined in this study. So far, no data have been published that show how common the phage types observed among the Brazilian porcine S. Agona isolates are and whether these phage types have been found in isolates from other animal sources or from humans. Members of the dominant clonal group were found in pigs from herds of 15 different cities, which were located up to 450 km apart from each other. This observation confirmed that such S. Agona isolates were widely distributed among slaughter-age pigs in Southern Brazil. The animals, from which members of this S. Agona clonal group were isolated, originated from integrated production systems in which piglets from one producer were distributed to different grower farms and subsequently to finisher farms, which then sent their slaughter-age pigs to the same slaughterhouse. If piglets are infected at the production stage, this might explain why animals from different herds in the same or different cities, carry members of the same S. Agona clonal group. However, since contracts exist between a specific slaughterhouse and the farms within the respective integrated production line, it might not explain why members of the same clonal group are found at different slaughterhouses. The exchange of pigs between the farms finally delivering their pigs to different slaughterhouses could be excluded. Thus, other factors, such as contaminated animal feed (Clark et al., 1973; Lindqvist et al., 2002) or living and non-living vectors, might play a role in the dissemination of these S. Agona isolates. The observation that members of this clone have been isolated in the majority of the cases from mesenteric lymph nodes (17 out of 22 samples) suggested that the detection of these S. Agona isolates resulted from an infection of the respective animals which might have occurred either at the farm level or during the transport of the animals to the slaughterhouse or in the holding pens at the slaughterhouse rather than from crosscontamination of the carcasses during the slaughter process (Morgan et al., 1987; Hurd et al., 2002). Transport and holding pens have been considered as
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
major points of Salmonella infection (Hurd et al., 2002), particularly since it is known that infected animals shed salmonellae under stress conditions and that even 2 h after oral ingestion, Salmonella isolates can be found in the mesenteric lymph nodes (Hurd et al., 2001). A comparison of the isolates obtained during both periods at slaughterhouse B revealed that isolates found during the first period, were not detectable during the second period. Although all three S. Agona isolates obtained from animals during the second period at slaughterhouse B were of PT40, they differed considerably in their genotypic traits from the single PT40 isolate identified during the first sampling period at the same slaughterhouse. This observation points towards the temporal occurrence and disappearance of specific S. Agona isolates rather than the persistence of these isolates for longer periods. Similar observations were also made in the follow-up study on S. Agona from cattle and human infections by Lindqvist et al. (2002). In conclusion, the results of this study suggest that – by using molecular typing methods, such as XbaI and BlnI macrorestriction analysis and SRF analysis in addition to phage typing – a specific clonal group of S. Agona was detectable among slaughter-age pigs from independent pig production lines at two slaughterhouses (A and C) and a variety of genomic patterns at the third slaughterhouse (B). This observation supports the requirement for the establishment of molecular-based surveillance programmes in Brazil which should also include techniques to better trace the dissemination of Salmonella isolates in foodproducing animals and to identify sources of S. Agona infections.
Acknowledgements The S. Agona strains used in this study represent part of the strain collection of the Departamento de Medicina Veterina´ria Preventiva, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil and were originally isolated by Marjo Cado´ Bessa and Sandra Maria Ferraz Castagna. We thank Wolfgang Rabsch, National Reference Center for Salmonella, Wernigerode, Germany, for phage typing of the S. Agona isolates, Roswitha Becker and Vera No¨ding, for excellent technical assistance, as well as
51
Marjo C. Bessa and Sandra M.F. Castagna for helpful discussions. Geovana Brenner Michael is supported by a scholarship of the German Academic Exchange Service (DAAD).
References Beli, E., Duraku, E., Telo, A., 2001a. Salmonella serotypes isolated from chicken meat in Albania. Int. J. Food Microbiol. 71, 263– 266. Beli, E., Telo, A., Duraku, E., 2001b. Salmonella serotypes isolated from turkey meat in Albania. Int. J. Food Microbiol. 63, 165– 167. Bessa, M.C., Costa, M., Cardoso, M., 2001. Prevalence of Salmonella sp. in slaughtered pigs in Rio Grande do Sul, Brazil. In: Proceedings Salinpork 2001—Fourth International Symposium on Epidemiology and Control of Salmonella and Other Foodborne Pathogens in Pork, Leipzig, Germany, pp. 189–191. Center for Disease Control and Prevention (CDC), 1998. Multistate outbreak of Salmonella serotype Agona infections linked to toasted oats cereal, United States. MMWR Morb. Mortal. Wkly. Rep. 47 (April–May), 462–464. Clark, G.M., Kaufmann, A.F., Gangarosa, E.J., Thompson, M.A., 1973. Epidemiology of an international outbreak of Salmonella Agona. Lancet ii, 490–493. Cloeckaert, A., Baucheron, S., Chaslus-Dancla, E., 2001. Nonenzymatic chloramphenicol resistance mediated by IncC plasmid R55 is encoded by a floR gene variant. Antimicrob. Agents Chemother. 45, 2381–2382. Davies, P.R., Funk, J.A., 1999. Epidemiology and control of Salmonella in pork—some of the questions. In: Second International Symposium on Epidemiology and Control of Salmonella in Pork, Washington, USA, pp. 1–11. Escartin, E.F., Lozano, J.S., Garcia, O.R., 2000. Quantitative survival of native Salmonella serovars during storage of frozen pork. Int. J. Food Microbiol. 54, 19–25. Fuzihara, T.O., Fernandes, S.A., Franco, B.D., 2000. Prevalence and dissemination of Salmonella serotypes along the slaughtering process in Brazilian small poultry slaughterhouses. J. Food Prot. 63, 1749–1753. Hurd, H.S., Gailey, J.K., McKean, J.D., Rostagno, M.H., 2001. Rapid infection in market-weight swine following exposure to a Salmonella Typhimurium-contaminated environment. Am. J. Vet. Res. 62, 1194–1197. Hurd, H.S., McKean, J.D., Griffith, R.W., Wesley, I.V., Rostagno, M.H., 2002. Salmonella enterica infections in market swine with and without transport and holding. Appl. Environ. Microbiol. 68, 2376–2381. Kado, C.I., Liu, S.T., 1981. Rapid procedure for detection and isolation of large and small plasmids. J. Bacteriol. 145, 1365– 1373. Killalea, D., Ward, L.R., Roberts, D., de Louvois, J., Sufi, F., Stuart, J.M., Wall, P.G., Susman, M., Schwieger, M., Sanderson, P.J., Fisher, I.S.T., Mead, P.S., Gill, O.N., Bartlett, C.L.R., Rowe, B., 1996. International epidemiological and microbiological study
52
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
of outbreak of Salmonella Agona infection from a ready to eat savoury snack. Part I: England and Wales and the United States. Br. Med. J. 313, 1105–1107. Koch, J., Schrauder, A., Alpers, K., Werber, D., Frank, C., Prager, R., Rabsch, W., Broll, S., Feil, F., Roggentin, P., Bockemu¨hl, J., Tscha¨pe, H., Ammon, A., Stark, K., 2005. Salmonella Agona outbreak from contaminated aniseed, Germany. Emerg. Infect. Dis. 11, 1124–1127. Lagotolla, C., Dolzani, L., Tonin, E., Lavenia, A., di Michele, M., Tommasini, T., Monti-Bragadin, C., 1996. PCR ribotyping for characterizing Salmonella isolates of different serotypes. J. Clin. Microbiol. 34, 2440–2443. Liebisch, B., Schwarz, S., 1996. Evaluation and comparison of molecular techniques for epidemiological typing of Salmonella enterica subsp. enterica serovar dublin. J. Clin. Microbiol. 34, 641–646. Lindqvist, N., Siitonen, A., Pelkonen, S., 2002. Molecular followup of Salmonella enterica subsp. enterica serovar Agona infection in cattle and humans. J. Clin. Microbiol. 40, 3648– 3653. Liu, S.L., Hessel, A., Sanderson, K.E., 1993. The XbaI–BlnI–CeuI genomic cleavage map of Salmonella typhimurium LT2 determined by double digestion, end labelling and pulsed-field gel electrophoresis. J. Bacteriol. 175, 4104–4120. Macrina, F.L., Kopecko, D.J., Jones, K.R., Ayers, D.J., McCowan, S.M., 1978. A multiple plasmid-containing Escherichia coli strain: convenient source of size reference plasmid molecules. Plasmid 1, 417–420. McClelland, M., Sanderson, K.E., Spieth, J., Clifton, S.W., Latreille, P., Courtney, L., Porwollik, S., Ali, J., Dante, M., Du, F., Hou, S., Layman, D., Leonard, S., Nguyen, C., Scott, K., Holmes, A., Grewal, N., Mulvaney, E., Ryan, E., Sun, H., Florea, L., Miller, W., Stoneking, T., Nhan, M., Waterston, R., Wilson, R.K., 2001. Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature 213, 852–856. Michael, G.B., Simonetti, R., Costa, M., Cardoso, M.R.I., 2003. Comparison of different selective enrichment steps to isolate Salmonella sp. from faeces of finishing swine. Braz. J. Microbiol. 34, 138–142. Michael, G.B., Cardoso, M., Schwarz, S., 2005. Class 1 integronassociated gene cassettes in Salmonella enterica subsp. enterica serovar Agona isolated from pig carcasses in Brazil. J. Antimicrob. Chemother. 55, 776–779. Morgan, I.R., Krautil, F.L., Craven, J.A., 1987. Effect of time in lairage on caecal and carcass Salmonella contamination of slaughter pigs. Epidemiol. Infect. 98, 323–330. Mulvey, M.R., Boyd, D.A., Baker, L., Mykytczuk, O., Reis, E.M., Asensi, M.D., Rodrigues, D.P., Ng, L.K., 2004. Characterization of a Salmonella enterica serovar Agona strain harbouring a class 1 integron containing novel OXA-type b-lactamase (blaOXA-53) and 60 -N-aminoglycoside acetyltransferase genes [aac(60 )-I30]. J. Antimicrob. Chemother. 54, 354–359.
National Committee for Clinical Laboratory Standards (NCCLS), 2002. Performance standards for antimicrobial disk and dilution susceptibility test for bacteria isolated from animals, second ed. Approved Standard M31-A2, NCCLS, Wayne, PA, USA. Oliveira, C.J., Carvalho, L.F., Fernandes, S.A., Tavechio, A.T., Menezes, C.C., Domingues Jr., F.J., 2002. Antimicrobial resistance of Salmonella serotypes isolated from slaughter-age pigs and environmental samples. Microb. Drug Resist. 8, 407–411. Olsen, J.E., Skov, M.N., Angen, O., Threlfall, E.J., Bisgaard, M., 1997. Genomic relationships between selected phage types of Salmonella enterica subsp. enterica serotype Typhimurium defined by ribotyping, IS200 typing and PFGE. Microbiology 143, 1471–1479. Rabsch, W., Prager, R., Koch, J., Stark, K., Roggentin, P., Bockemu¨hl, J., Beckmann, G., Stark, R., Siegl, W., Ammon, A., Tscha¨pe, H., 2005. Molecular epidemiology of Salmonella enterica serovar Agona: characterization of a diffuse outbreak caused by aniseed-fennel-caraway infusion, Epidemiol. Infect. (cited from reference Koch et al. 2005), 133, 837–844. Schwarz, S., Liebisch, B., 1994. Pulsed-field gel electrophoretic identification of Salmonella enterica serovar Typhimurium live vaccine strain Zoosaloral H. Lett. Appl. Microbiol. 19, 469–472. Sorensen, O., van Donkersgoed, J., McFall, M., Manninen, K., Gensler, G., Ollis, G., 2002. Salmonella spp. shedding by Alberta beef cattle and the detection of Salmonella spp. in ground beef. J. Food Prot. 65, 484–491. Synnott, M.B., Brindley, M., Gray, J., Dawson, J.K., 1998. An outbreak of Salmonella Agona infection associated with precooked turkey meat. Commun. Dis. Public Health 1, 176–179. Tavechio, A.T., Ghilardi, A.C.R., Peresi, J.T.M., Fuzihara, T.O., Yonamine, E.K., Jakabi, M., Fernandes, S.A., 2002. Salmonella serotypes isolated from nonhuman sources in Sa˜o Paulo, Brazil, from 1996 through 2000. J. Food Prot. 65, 1041–1044. Taylor, J.P., Barnett, B.J., del Rosario, L., Williams, K., Barth, S.S., 1998. Prospective investigation of cryptic outbreaks of Salmonella Agona salmonellosis. J. Clin. Microbiol. 36, 2861–2864. Terletski, V., Schwarz, S., Carnwath, J., Niemann, H., 2003a. Subtracted restriction fingerprinting—a tool for bacterial genome typing. BioTechniques 34, 304–313. Terletski, V., Schwarz, S., Carnwath, J., Niemann, H., 2003b. Typing of Salmonella enterica subsp. enterica serovars Choleraesuis, Typhimurium, Dublin and laboratory strains of Escherichia coli using subtracted restriction fingerprinting (SRF) Microbiol. Res. 158, 135–142. Terletski, V., Michael, G.B., Schwarz, S., 2004. Subtracted restriction fingerprinting—a new typing technique using magnetic capture of tagged restriction fragments. FEMS Immunol. Med. Microbiol. 41, 1–8. Threlfall, E.J., Hampton, M.D., Ward, L.R., Rowe, B., 1996. Application of pulsed-field gel electrophoresis to an international outbreak of Salmonella Agona. Emerg. Infect. Dis. 2, 130–132.
Molecular analysis of Salmonella enterica subsp. enterica serovar Agona isolated from slaughter pigs Geovana Brenner Michael a,b, Marisa Cardoso b, Stefan Schwarz a,* b
a Institut fu¨r Tierzucht, Bundesforschungsanstalt Fu¨r Landwirtschaft (FAL), Ho¨ltystr. 10, 31535 Neustadt-Mariensee, Germany Departamento de Medicina Veterina´ria Preventiva, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil
Received 24 August 2005; received in revised form 30 September 2005; accepted 4 October 2005
Abstract Salmonella enterica subsp. enterica (S.) serovar Agona plays an important role in Brazil as causative agent of salmonellosis in food-producing animals – in particular, pigs and poultry – as well as in humans. A total of 45 S. Agona isolates collected from slaughter pigs at three different slaughterhouses in Southern Brazil was investigated in this study for their phenotypic and genotypic relatedness. For this, the antimicrobial susceptibility patterns and the phage types were determined. Molecular analysis included the determination of plasmid profiles as well as the analysis of XbaI- and BlnI-generated macrorestriction patterns. Moreover, a novel typing method called subtracted restriction fingerprinting (SRF) was successfully applied to the S. Agona isolates. Based on all properties determined, a dominant clonal group comprising 33 of the 45 isolates was identified. Members of this group were susceptible to all antimicrobials tested, did not carry plasmids, shared the same phage type and were closely related or even indistinguishable by their EcoRI–PauI SRF patterns as well as their XbaI and BlnI macrorestriction patterns. Members of this clonal group were identified at all 3 slaughterhouses at variable frequencies and originated from pig herds raised in 15 different cities in Southern Brazil which were located up to 450 km apart from each other. Since the S. Agonacarrying slaughter pigs were from various integrated production lines, the results of this study suggest that a specific clonal group of S. Agona had entered numerous pig production lines. This observation supports the requirement for the establishment of monitoring and control programmes in Brazil which should also include molecular techniques to better trace the dissemination of S. Agona and other Salmonella serovars in pigs and other food-producing animals. # 2005 Elsevier B.V. All rights reserved. Keywords: Macrorestriction analysis; Subtracted restriction fingerprinting; Plasmid profiles; Antimicrobial susceptibility patterns; Phage typing
1. Introduction
* Corresponding author. Tel.: +49 5034 871 241; fax: +49 5034 871 246. E-mail address: [email protected] (S. Schwarz).
Salmonella enterica subsp. enterica serovar Agona (S. Agona) is considered a zoonotic bacterium, which has been associated with various infections in humans and animals worldwide. Transmission of S. Agona
0378-1135/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.vetmic.2005.10.011
44
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
from animals to humans commonly occurs by ingestion of contaminated beef (Sorensen et al., 2002), pork (Escartin et al., 2000), chicken and turkey meat (Beli et al., 2001a,b; Synnott et al., 1998), although contaminated vegetarian food (Killalea et al., 1996), cereals (CDC, 1998) or tea (Koch et al., 2005) have also been identified as sources of infections. Human infections due to S. Agona have been described mostly as local outbreaks, although nation-wide outbreaks and international outbreaks have been reported. In Brazil, S. Agona isolates seem to play a particular role since they have been found not only among the 10 serovars most commonly isolated from cases of human salmonellosis, but also represented the fourth most common serovar among the non-human isolates (Tavechio et al., 2002). Slaughter-age pigs (Bessa et al., 2001; Oliveira et al., 2002; Michael et al., 2005) as well as poultry (Fuzihara et al., 2000) from Brazil have been found to be infected with S. Agona. Most of the Brazilian S. Agona isolates from humans and animals showed extended antimicrobial resistance patterns (Oliveira et al., 2002; Mulvey et al., 2004; Michael et al., 2005). These facts stress the hazard that S. Agona of animal origin represents for human health and underlines the importance to achieve a better understanding of the epidemiology of S. Agona infections (Davies and Funk, 1999). In the present study, we investigated S. Agona isolates obtained from slaughter pigs at three different slaughterhouses in Southern Brazil for their genomic relatedness. Epidemiological background data of the isolates were put in context with the typing data to gain better insight into the geographical and temporal distribution of particular isolates.
2. Materials and methods
terhouses A and B were located 96 km away from each other and received animals from farms located in almost the same geographic area. In contrast, the slaughterhouse C was located about 300 km away from the others and received animals from a different region. During an 11-month period from 09/1999 to 07/2000, 600 samples (mesenteric lymph nodes and intestinal content) from 300 individual animals representing different herds of different farms were collected at four distinct dates at each of the three slaughterhouses. A total of 41 (6.83%) S. Agona isolates was identified from all 3 slaughterhouses during this period. The presence of S. Agona in both, mesenteric lymph nodes and intestinal content, was detected in only 4 of the 300 animals sampled. Of the 41 S. Agona isolates, 18 originated from slaughterhouse A, 7 from slaughterhouse B and 16 from slaughterhouse C. Another 3 (0.56%) S. Agona isolates were obtained from 535 samples of mesenteric lymph nodes, head lymph nodes/tonsils and/or intestinal content collected from 230 individual animals at slaughterhouse B between 09/2000 and 10/2001. During that period, 165 samples from meat products collected at slaughterhouse B were analyzed and a single sample (0.61%) from minced meat was found to be contaminated with S. Agona. For various reasons, most of which were far beyond our control, it was not possible to obtain samples from slaughterhouses A and C during the second sampling period. Of the 45 isolates included in this study, 22 were from mesenteric lymph nodes, 21 from intestinal content, 1 from head lymph nodes/tonsils and 1 from minced meat. All isolates were identified as S. Agona by routine laboratory procedures (Michael et al., 2003). Serological tests were performed at the National Reference Centre, Instituto Oswaldo Cruz, Rio de Janeiro, Brazil, by standard slide agglutination using commercially available antisera.
2.1. Bacterial isolates 2.2. Antimicrobial susceptibility testing S. Agona isolates, all exhibiting the antigenic formula 4,12:f,g,s:, were obtained from apparently healthy slaughter pigs and in one case from a meat product (minced meat) at three slaughterhouses (A–C) in Southern Brazil (Bessa et al., 2001). Each slaughterhouse was supplied by finishing farms in a vertically integrated production system. The slaugh-
All S. Agona strains were investigated for their antimicrobial resistance by the agar disk diffusion test using the following disks (Oxoid, Wesel, Germany; Becton Dickinson, Heidelberg, Germany): ampicillin (AMP, 10 mg), amikacin (AMI, 30 mg), gentamicin (GEN, 10 mg), kanamycin (KAN, 30 mg), neomycin
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
(NEO, 30 mg), streptomycin (STR, 10 mg), tobramycin (TOB, 10 mg), tetracycline (TET, 30 mg), compound sulfonamides (SUL, 300 mg), sulfamethoxazole/trimethoprim (SXT, 23.75/1.25 mg), trimethoprim (TMP, 5 mg), chloramphenicol (CHL, 30 mg), nalidixic acid (NAL, 30 mg) and ciprofloxacin (CIP, 5 mg). In addition, spectinomycin (SPT, 100 mg), minocycline (MIN, 30 mg) or florfenicol (FFC, 30 mg) were also tested with strains resistant to streptomycin, tetracycline or chloramphenicol, respectively. In vitro susceptibility testing was performed and the results evaluated according to the document M31-A2 of the Clinical and Laboratory Standards Institute (NCCLS, 2002). 2.3. Isolation of plasmid DNA and whole cell DNA Whole cell DNA of the S. Agona isolates were extracted as previously described (Schwarz and Liebisch, 1994). For plasmid preparation, a modification of the alkaline lysis procedure (Kado and Liu, 1981) was applied which proved to be particularly suitable for the detection of large plasmids (Schwarz and Liebisch, 1994). Gel electrophoretic presentation of the plasmid profiles followed previously described protocols. The plasmids of Escherichia coli V517 (Macrina et al., 1978), the virulence plasmid of S. Typhimurium LT2 (McClelland et al., 2001) and the 150 kb resistance plasmid R55 from Klebsiella pneumoniae (Cloeckaert et al., 2001) served as size standards for the determination of plasmid sizes. 2.4. Phage typing All 45 S. Agona isolates were sent to the National Reference Center for Salmonella at the Robert Koch Institute in Wernigerode, Germany, for phage typing. A recently established S. Agona-specific phage typing system that allowed the detection of 52 phage types was applied to the Brazilian S. Agona isolates (Rabsch et al., 2005). 2.5. Macrorestriction analysis Whole cell DNA for pulsed-field gel electrophoresis was prepared as described earlier (Schwarz and Liebisch, 1994; Liebisch and Schwarz, 1996). DNAcontaining slices of the agarose plugs were incubated
45
for 4 h in the presence of 20 units of XbaI (Roche, Mannheim, Germany) or 10 units of BlnI (AmershamBuchler, Braunschweig, Germany). The respective DNA fragments were separated by pulsed-field gel electrophoresis (SeaKem GTG, 1%, w/v, Biozym, Hess, Oldendorf, Germany) in a CHEF DR III system (BioRad, Mu¨nchen, Germany) at 5.6 V/cm with 0.5 TBE as running buffer. The pulse times for XbaI digests were increased from 10 to 30 s during the first 11 h and subsequently from 30 to 50 s during the next 13 h; those for the BlnI digests were increased from 7 to 12 s for the first 11 h and from 20 to 65 s for the following 13 h. The XbaI or BlnI fragments of S. Typhimurium LT2 served as size standards (Liu et al., 1993). The gel was stained with ethidium bromide (2 mg/ml, Sigma, Deisenhofen, Germany) for 4 min, destained in distilled water for 12 min and photographed under UV-illumination. 2.6. Subtracted restriction fingerprinting analysis (SRF) Extraction of genomic DNA for SRF analysis followed a previously described protocol (Terletski et al., 2003a) with slight modifications. In brief, double digestion of 1.8 mg of the whole cell DNA was performed in y+Tango restriction buffer (Fermentas, St. Leon-Rot, Germany) using 3 U EcoRI (Fermentas) and 7 U PauI (Fermentas) in a total reaction volume of 40 ml. The digestion reaction was incubated at 37 8C for 2 h. The subsequent fill-in reaction of the sticky ends was performed by using 2 ml of a dNTP mixture to result in 40 mM dATP (Sigma–Aldrich, Taufkirchen, Germany), 40 mM dGTP (Sigma–Aldrich), 2 mM DIG-11-dUTP (Roche, Mannheim, Germany) and 2 mM biotin-14-dCTP (Invitrogen, Karlsruhe, Germany). Then, 1 U of Klenow enzyme (Fermentas) was added and the mixture was incubated for 10 min at room temperature. The unincorporated dNTPs were removed with spin micro columns (Quantum PREP1 PCR Kleen Spin Columns; Bio-Rad Laboratories GmbH, Munich, Germany) filled with 1 ml prehydrated Sephadex1 G-50 (Amersham Bioscience, Freiburg, Germany) by centrifugation at 3000 rpm for 2 min at room temperature. The biotinylated fragments were captured by adding 40 ml (20 mg) of the pre-washed streptavidin-coated magnetic particles (Roche, Mannheim, Germany) to each tube. The
46
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
tubes were incubated with continuous agitation for 2 h at room temperature. Separation was performed in a magnetic rack (Dynal MPC-P-12; Dynal GmbH, Hamburg, Germany) and the resulting suspensions containing the Dig-labeled fragments were transferred to spin micro columns and centrifuged as described above. The final suspensions were analyzed by agarose gel electrophoresis with 1% (w/v) gels at 100 V/cm for 11 h with 0.5 TBE as running buffer. The digoxigenin-labeled DNA molecular weight marker III (Roche) was used as size standard. After gel electrophoretic separation, the DNA fragments were blotted and the digoxigenin-labeled fragments immunochemically detected as previously described (Terletski et al., 2003a,b). 2.7. Analysis of the fragment patterns The macrorestriction patterns obtained after XbaI or BlnI digestion as well as the SRF patterns were analyzed using the GelCompar software package (Applied Maths, Kortrijk, Belgium). For this, the images were saved as TIFF files and imported into the GelCompar software. The similarities between the
profiles were calculated using the Dice coefficient, with a maximum position tolerance of 1.2%. The patterns were clustered by using the unweighted pair group method with arithmetic averages (UPGMA). Since the large plasmids present in five S. Agona isolates were found to be cut into fragments smaller than 90 kb by XbaI as well as BlnI (data not shown), only macrorestriction fragments larger than 90 kb were used for the comparison of the patterns (Olsen et al., 1997). A macrorestriction or SRF pattern was considered as unique when it differed from the next closely related one by three or more bands. Patterns that differed from others by only one or two bands were considered as variants of pre-existing patterns. The most common pattern obtained with each method received the number 0 0 1 and according to the similarity detected by the cluster analysis the next closely related patterns received the following numbers. Lower case letters after the numbers indicated variants of a pattern. The alphabetical order was used to address the level of similarity of these variants to the corresponding pattern. Thus, SAX001 indicated the most common restriction pattern (0 0 1) of S. Agona obtained by
Table 1 Phenotypic and genotypic characteristics of the S. Agona isolates Total no. of isolates
Isolates obtained from slaughterhouse
Plasmid profiles
Phage types
Resistance patternsa
BlnI
SRF patterns EcoRI–PauI
XbaI 29 1 3
A (14)b, B (1), C (14)b A A (1), B (2)
SAX001 SAX001a SAX001
SAB001 SAB001 SAB003
SAS001 SAS001 SAS001
PP0 PP0 PP0
PT07 PT07 PT07
S S S
C
SAX002
SAB001b
SAS001
PP0
PT18
S
B A (2), C (1)
SAX003 SAX003
SAB001a SAB001b
SAS004 SAS001
PP1 PP0
PT40 PT40
R1 S
1c
B
SAX003a
SAB001b
SAS002
PP2
ut/O1-
S
c
B B
SAX003b SAX004
SAB001b SAB002
SAS003 SAS001a
PP0 PP4
PT40 PT40
S S
1 1d 1d
B B B
SAX004a SAX005 SAX005a
SAB002 SAB004 SAB005
SAS005 SAS003a SAS006
PP3 PP0 PP5
PT16 PT16 PT16
R2 S R3
1 2 3
1 1
c
Macrorestriction patterns
a S: Susceptible to all antimicrobial tested; R1: resistant to AMP-CHL-KAN-MIN/TET-STR-SUL-TMP; R2: resistant to CHL-MIN/TETSTR-SPT-SUL-TMP; R3: resistant to CHL-MIN/TET-SUL-TMP. b Include three matched pairs of isolates from mesenteric lymph nodes and intestinal content, two pairs from slaughterhouse A and one pair from slaughterhouse C (29 isolates from 26 individual animals). c Isolates from the second sampling period. d From the same animal, the susceptible isolate is from intestinal content and the multiresistant one from mesenteric lymph nodes.
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
XbaI digestion, whereas SAX001a indicated a variant of this pattern. In this way, SAB indicated the patterns obtained with BlnI digestion and SAS the patterns obtained by the SRF analysis. The plasmid profiles (PP) were indicated by numbers 0–5 and phage types (PT) by the respective number in the phage typing system used.
47
isolate from minced meat proved to be non-typeable (ut/O1) by this phage typing system. PT07 isolates were found in all three slaughterhouses, while PT40 isolates were obtained from slaughterhouses A and B. The only PT18 isolate was from slaughterhouse C, whereas the three PT16 isolates were detected in samples from slaughterhouse B (Table 1). 3.4. Macrorestriction patterns
3. Results 3.1. Antimicrobial resistance patterns Of the 45 S. Agona isolates included in this study, 41 (91.1%) were susceptible to all antimicrobials tested, whereas the remaining 4 isolates (Table 1) were multiresistant. Among the latter four isolates, the following resistance patterns were observed: CHLMIN/TET-SUL-TMP (one isolate), CHL-MIN/TETSTR-SPT-SUL-TMP (one isolate) and AMP-CHLKAN-MIN/TET-STR-SUL-TMP (two isolates). All multiresistant isolates were obtained from mesenteric lymph nodes or intestinal content samples collected at different time periods in slaughterhouse B. 3.2. Plasmid profiles A total of 39 S. Agona isolates (86.6%) were plasmid-free (PP0). The six plasmid-bearing strains exhibited five different plasmid profiles (PP1–PP5). The four multiresistant strains harbored a ca. 150 kb plasmid. While two of these isolates carried only this large plasmid (PP1), the other two isolates harboured additional plasmids of ca. 10 kb (PP3) or 40 kb (PP5), respectively. A single plasmid of approximately 10 kb (PP4) and a plasmid of approximately 90 kb (PP2) were identified in another two isolates. The plasmidbearing S. Agona isolates originated from mesenteric lymph nodes, intestinal content and minced meat and were obtained during different times only at slaughterhouse B (Table 1). 3.3. Phage types Among the 45 S. Agona isolates included in this study, 33 isolates (73.3%) were assigned to PT07. The types PT16, PT18 and PT40 were found in three, one and seven isolates, respectively. Only the S. Agona
XbaI digestion showed 5 major patterns and 5 variant patterns each consisting of 10–14 fragments in the size range approximately between 90 and 830 kb (Fig. 1a). All isolates shared at least seven common fragments. The most common pattern, SAX001, was shared by 32 isolates (71.1%), while 5 isolates exhibited pattern SAX003. All other XbaI-generated fragment patterns were represented by single isolates. Cluster analysis revealed that 41 isolates were grouped in a main cluster with an overall similarity of 77.8% (Fig. 1b). All strains of the most common pattern were found in this major cluster. The remaining four isolates formed a second cluster with an overall similarity of 83.0% (Fig. 1b). The largest difference of 10 fragments was seen between the patterns SAX001 and SAX005. BlnI digestion resulted in 5 major patterns and 2 variant patterns comprising 9–14 bands in the same size range as seen for XbaI (Fig. 1c). Four common bands were seen in the BlnI patterns of all isolates. Only the most common pattern SAB001, which was represented by 30 isolates, showed variants. The variant patterns SAB001a and SAB001b were seen with two and six isolates, respectively. Two isolates exhibited pattern SAB002, three isolates SAB003, whereas single isolates represented patterns SAB004 and SAB005. Cluster analysis grouped 43 isolates in a main cluster with an overall similarity of 78.5% (Fig. 1d). The remaining two isolates (SAB004 and SAB005) shared 88.9% similarity and formed a minor cluster (Fig. 1d). The maximum difference of 13 fragments was seen between the pattern SAB005 and pattern SAB003. A total of 29 S. Agona isolates (64.4%) was indistinguishable by their XbaI (SAX001) and BlnI (SAB001) macrorestriction patterns. These isolates were found at variable frequencies in animals slaughtered at each of the three slaughterhouses and
48
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
Fig. 1. Macrorestriction patterns of S. Agona isolates obtained with the restriction enzymes XbaI (a) and BlnI (c). Lane letters and numbers correspond to the different patterns in Table 1. Lane M contains XbaI-digested (a) or BlnI-digested (c) DNA of S. Typhimurium LT2. Dendograms showing the relationships of the different XbaI (b) and BlnI (d) patterns. Similarity analysis was performed using the Dice coefficient and the clustering was generated by UPGMA.
were recovered during the entire sampling period 09/ 1999–07/2000. 3.5. Subtracted restriction fingerprinting patterns EcoRI/PauI double digestion produced six major SRF fragment patterns and two variant patterns (Fig. 2). Each of these SRF patterns consisted of 33–50 bands ranging in size between ca. 1 and 19 kb. A total of 37 S. Agona strains (82.2%) were
indistinguishable by their SRF patterns. In addition to the most common pattern SAS001, only the pattern SAS004 was shared by two isolates. All other SRF patterns were represented by single isolates. All SRF patterns had approximately 30 bands in common. As seen with macrorestriction analysis, two main clusters were found with the SRF analysis (Fig. 2). Forty-one isolates were grouped in the major cluster (91.4% similarity), while the remaining four isolates representing patterns SAS004, SAS005 and SAS006
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
49
Fig. 2. (a) EcoRI–PauI SRF patterns of S. Agona isolates. Lane letters and numbers correspond to the patterns in Table 1. Lane M contains a digoxigenin-labeled DNA molecular weight marker. (b) Dendogram showing the relationships of the different SRF patterns. Similarity analysis was performed using the Dice coefficient and the clustering was generated by UPGMA.
formed a minor cluster with 85.6% of similarity. The largest difference of 19 bands was detected between the patterns SAS001 and SAS006. 3.6. Correlation between the characteristics and the origin of the isolates All phenotypic and genotypic properties of the 45 S. Agona isolates are summarized in Table 1. It became obvious that 29 S. Agona isolates were indistinguishable on the basis of all properties determined: susceptible to all antimicrobials tested, plasmid-free, indistinguishable by their SRF and XbaI and BlnI macrorestriction patterns and by the phage type. Another four isolates showed only slight variations in either their XbaI or their BlnI patterns,
but were otherwise indistinguishable from the aforementioned 29 S. Agona isolates. These 33 isolates were considered to represent members of the same clonal group. Members of this dominant clonal group were obtained mainly from animals slaughtered at the slaughterhouses A (16 isolates) and C (14 isolates), whereas only three isolates of this clonal group were detected at slaughterhouse B. A comparison with the origin of the herds from which these isolates were obtained revealed that the 16 isolates from slaughterhouse A originated from herds of 5 different cities, each of the 3 isolates from slaughterhouse B was from a different city and the 14 isolates from slaughterhouse C were from 7 different cities. In three individual animals (two from slaughterhouse A and one from slaughterhouse C), S. Agona isolates representing this
50
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
clonal group were detected in mesenteric lymph nodes and intestinal content. Among the remaining 12 S. Agona isolates, 6 of the 7 PT40 isolates as well as the single PT18 isolate and the non-phage typeable isolate were closely related or even indistinguishable in their XbaI, BlnI or SRF patterns. These isolates originated from herds of individual cities and were isolated at all three slaughterhouses between 09/1999 and 07/2000, but also at slaughterhouse B between 09/2000 and 10/ 2001. The three PT16 isolates, two of which originated from mesenteric lymph nodes or intestinal content of the same animal, showed considerable variability in all phenotypic and genotypic traits.
4. Discussion In this retrospective study, porcine S. Agona isolates from three different slaughterhouses located in Southern Brazil were characterized by macrorestriction and SRF analysis, phage typing, plasmid profile analysis and antimicrobial resistance patterns. Macrorestriction analysis has already been used successfully for the analysis of S. Agona outbreaks (Threlfall et al., 1996; Taylor et al., 1998; Lindqvist et al., 2002). Previous studies showed that other molecular methods used for the differentiation of Salmonella isolates, such as ribotyping, 16S–23S intergenic spacer PCR or IS200 typing proved to be unsuitable for S. Agona (Threlfall et al., 1996; Lagotolla et al., 1996). To complement our macrorestriction data, we, therefore, used SRF analysis (Terletski et al., 2004). This technique has previously proved to be suitable for the differentiation of S. Choleraesuis, S. Typhimurium and S. Dublin (Terletski et al., 2003b). Macrorestriction analysis with two enzymes as well as SRF analysis produced a number of different fragment patterns among the S. Agona isolates, thus, confirming that these methods are able to discriminate between the S. Agona isolates. The observation, that 33 isolates were either indistinguishable or closely related by their genomic fragment patterns obtained by macrorestriction and SRF analyses, strongly suggests that these isolates represent members of the same clonal group. Additional tests included plasmid analysis and antimicrobial susceptibility patterns. A previous study
which focused on the resistance genes present in the four multiresistant isolates revealed that all resistance properties detected in these isolates were associated with conjugative ca. 150 kb plasmids (Michael et al., 2005). The use of a recently established phage typing system for S. Agona supplemented the other strain characteristics determined in this study. So far, no data have been published that show how common the phage types observed among the Brazilian porcine S. Agona isolates are and whether these phage types have been found in isolates from other animal sources or from humans. Members of the dominant clonal group were found in pigs from herds of 15 different cities, which were located up to 450 km apart from each other. This observation confirmed that such S. Agona isolates were widely distributed among slaughter-age pigs in Southern Brazil. The animals, from which members of this S. Agona clonal group were isolated, originated from integrated production systems in which piglets from one producer were distributed to different grower farms and subsequently to finisher farms, which then sent their slaughter-age pigs to the same slaughterhouse. If piglets are infected at the production stage, this might explain why animals from different herds in the same or different cities, carry members of the same S. Agona clonal group. However, since contracts exist between a specific slaughterhouse and the farms within the respective integrated production line, it might not explain why members of the same clonal group are found at different slaughterhouses. The exchange of pigs between the farms finally delivering their pigs to different slaughterhouses could be excluded. Thus, other factors, such as contaminated animal feed (Clark et al., 1973; Lindqvist et al., 2002) or living and non-living vectors, might play a role in the dissemination of these S. Agona isolates. The observation that members of this clone have been isolated in the majority of the cases from mesenteric lymph nodes (17 out of 22 samples) suggested that the detection of these S. Agona isolates resulted from an infection of the respective animals which might have occurred either at the farm level or during the transport of the animals to the slaughterhouse or in the holding pens at the slaughterhouse rather than from crosscontamination of the carcasses during the slaughter process (Morgan et al., 1987; Hurd et al., 2002). Transport and holding pens have been considered as
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
major points of Salmonella infection (Hurd et al., 2002), particularly since it is known that infected animals shed salmonellae under stress conditions and that even 2 h after oral ingestion, Salmonella isolates can be found in the mesenteric lymph nodes (Hurd et al., 2001). A comparison of the isolates obtained during both periods at slaughterhouse B revealed that isolates found during the first period, were not detectable during the second period. Although all three S. Agona isolates obtained from animals during the second period at slaughterhouse B were of PT40, they differed considerably in their genotypic traits from the single PT40 isolate identified during the first sampling period at the same slaughterhouse. This observation points towards the temporal occurrence and disappearance of specific S. Agona isolates rather than the persistence of these isolates for longer periods. Similar observations were also made in the follow-up study on S. Agona from cattle and human infections by Lindqvist et al. (2002). In conclusion, the results of this study suggest that – by using molecular typing methods, such as XbaI and BlnI macrorestriction analysis and SRF analysis in addition to phage typing – a specific clonal group of S. Agona was detectable among slaughter-age pigs from independent pig production lines at two slaughterhouses (A and C) and a variety of genomic patterns at the third slaughterhouse (B). This observation supports the requirement for the establishment of molecular-based surveillance programmes in Brazil which should also include techniques to better trace the dissemination of Salmonella isolates in foodproducing animals and to identify sources of S. Agona infections.
Acknowledgements The S. Agona strains used in this study represent part of the strain collection of the Departamento de Medicina Veterina´ria Preventiva, Universidade Federal do Rio Grande do Sul (UFRGS), Porto Alegre, Brazil and were originally isolated by Marjo Cado´ Bessa and Sandra Maria Ferraz Castagna. We thank Wolfgang Rabsch, National Reference Center for Salmonella, Wernigerode, Germany, for phage typing of the S. Agona isolates, Roswitha Becker and Vera No¨ding, for excellent technical assistance, as well as
51
Marjo C. Bessa and Sandra M.F. Castagna for helpful discussions. Geovana Brenner Michael is supported by a scholarship of the German Academic Exchange Service (DAAD).
References Beli, E., Duraku, E., Telo, A., 2001a. Salmonella serotypes isolated from chicken meat in Albania. Int. J. Food Microbiol. 71, 263– 266. Beli, E., Telo, A., Duraku, E., 2001b. Salmonella serotypes isolated from turkey meat in Albania. Int. J. Food Microbiol. 63, 165– 167. Bessa, M.C., Costa, M., Cardoso, M., 2001. Prevalence of Salmonella sp. in slaughtered pigs in Rio Grande do Sul, Brazil. In: Proceedings Salinpork 2001—Fourth International Symposium on Epidemiology and Control of Salmonella and Other Foodborne Pathogens in Pork, Leipzig, Germany, pp. 189–191. Center for Disease Control and Prevention (CDC), 1998. Multistate outbreak of Salmonella serotype Agona infections linked to toasted oats cereal, United States. MMWR Morb. Mortal. Wkly. Rep. 47 (April–May), 462–464. Clark, G.M., Kaufmann, A.F., Gangarosa, E.J., Thompson, M.A., 1973. Epidemiology of an international outbreak of Salmonella Agona. Lancet ii, 490–493. Cloeckaert, A., Baucheron, S., Chaslus-Dancla, E., 2001. Nonenzymatic chloramphenicol resistance mediated by IncC plasmid R55 is encoded by a floR gene variant. Antimicrob. Agents Chemother. 45, 2381–2382. Davies, P.R., Funk, J.A., 1999. Epidemiology and control of Salmonella in pork—some of the questions. In: Second International Symposium on Epidemiology and Control of Salmonella in Pork, Washington, USA, pp. 1–11. Escartin, E.F., Lozano, J.S., Garcia, O.R., 2000. Quantitative survival of native Salmonella serovars during storage of frozen pork. Int. J. Food Microbiol. 54, 19–25. Fuzihara, T.O., Fernandes, S.A., Franco, B.D., 2000. Prevalence and dissemination of Salmonella serotypes along the slaughtering process in Brazilian small poultry slaughterhouses. J. Food Prot. 63, 1749–1753. Hurd, H.S., Gailey, J.K., McKean, J.D., Rostagno, M.H., 2001. Rapid infection in market-weight swine following exposure to a Salmonella Typhimurium-contaminated environment. Am. J. Vet. Res. 62, 1194–1197. Hurd, H.S., McKean, J.D., Griffith, R.W., Wesley, I.V., Rostagno, M.H., 2002. Salmonella enterica infections in market swine with and without transport and holding. Appl. Environ. Microbiol. 68, 2376–2381. Kado, C.I., Liu, S.T., 1981. Rapid procedure for detection and isolation of large and small plasmids. J. Bacteriol. 145, 1365– 1373. Killalea, D., Ward, L.R., Roberts, D., de Louvois, J., Sufi, F., Stuart, J.M., Wall, P.G., Susman, M., Schwieger, M., Sanderson, P.J., Fisher, I.S.T., Mead, P.S., Gill, O.N., Bartlett, C.L.R., Rowe, B., 1996. International epidemiological and microbiological study
52
G.B. Michael et al. / Veterinary Microbiology 112 (2006) 43–52
of outbreak of Salmonella Agona infection from a ready to eat savoury snack. Part I: England and Wales and the United States. Br. Med. J. 313, 1105–1107. Koch, J., Schrauder, A., Alpers, K., Werber, D., Frank, C., Prager, R., Rabsch, W., Broll, S., Feil, F., Roggentin, P., Bockemu¨hl, J., Tscha¨pe, H., Ammon, A., Stark, K., 2005. Salmonella Agona outbreak from contaminated aniseed, Germany. Emerg. Infect. Dis. 11, 1124–1127. Lagotolla, C., Dolzani, L., Tonin, E., Lavenia, A., di Michele, M., Tommasini, T., Monti-Bragadin, C., 1996. PCR ribotyping for characterizing Salmonella isolates of different serotypes. J. Clin. Microbiol. 34, 2440–2443. Liebisch, B., Schwarz, S., 1996. Evaluation and comparison of molecular techniques for epidemiological typing of Salmonella enterica subsp. enterica serovar dublin. J. Clin. Microbiol. 34, 641–646. Lindqvist, N., Siitonen, A., Pelkonen, S., 2002. Molecular followup of Salmonella enterica subsp. enterica serovar Agona infection in cattle and humans. J. Clin. Microbiol. 40, 3648– 3653. Liu, S.L., Hessel, A., Sanderson, K.E., 1993. The XbaI–BlnI–CeuI genomic cleavage map of Salmonella typhimurium LT2 determined by double digestion, end labelling and pulsed-field gel electrophoresis. J. Bacteriol. 175, 4104–4120. Macrina, F.L., Kopecko, D.J., Jones, K.R., Ayers, D.J., McCowan, S.M., 1978. A multiple plasmid-containing Escherichia coli strain: convenient source of size reference plasmid molecules. Plasmid 1, 417–420. McClelland, M., Sanderson, K.E., Spieth, J., Clifton, S.W., Latreille, P., Courtney, L., Porwollik, S., Ali, J., Dante, M., Du, F., Hou, S., Layman, D., Leonard, S., Nguyen, C., Scott, K., Holmes, A., Grewal, N., Mulvaney, E., Ryan, E., Sun, H., Florea, L., Miller, W., Stoneking, T., Nhan, M., Waterston, R., Wilson, R.K., 2001. Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature 213, 852–856. Michael, G.B., Simonetti, R., Costa, M., Cardoso, M.R.I., 2003. Comparison of different selective enrichment steps to isolate Salmonella sp. from faeces of finishing swine. Braz. J. Microbiol. 34, 138–142. Michael, G.B., Cardoso, M., Schwarz, S., 2005. Class 1 integronassociated gene cassettes in Salmonella enterica subsp. enterica serovar Agona isolated from pig carcasses in Brazil. J. Antimicrob. Chemother. 55, 776–779. Morgan, I.R., Krautil, F.L., Craven, J.A., 1987. Effect of time in lairage on caecal and carcass Salmonella contamination of slaughter pigs. Epidemiol. Infect. 98, 323–330. Mulvey, M.R., Boyd, D.A., Baker, L., Mykytczuk, O., Reis, E.M., Asensi, M.D., Rodrigues, D.P., Ng, L.K., 2004. Characterization of a Salmonella enterica serovar Agona strain harbouring a class 1 integron containing novel OXA-type b-lactamase (blaOXA-53) and 60 -N-aminoglycoside acetyltransferase genes [aac(60 )-I30]. J. Antimicrob. Chemother. 54, 354–359.
National Committee for Clinical Laboratory Standards (NCCLS), 2002. Performance standards for antimicrobial disk and dilution susceptibility test for bacteria isolated from animals, second ed. Approved Standard M31-A2, NCCLS, Wayne, PA, USA. Oliveira, C.J., Carvalho, L.F., Fernandes, S.A., Tavechio, A.T., Menezes, C.C., Domingues Jr., F.J., 2002. Antimicrobial resistance of Salmonella serotypes isolated from slaughter-age pigs and environmental samples. Microb. Drug Resist. 8, 407–411. Olsen, J.E., Skov, M.N., Angen, O., Threlfall, E.J., Bisgaard, M., 1997. Genomic relationships between selected phage types of Salmonella enterica subsp. enterica serotype Typhimurium defined by ribotyping, IS200 typing and PFGE. Microbiology 143, 1471–1479. Rabsch, W., Prager, R., Koch, J., Stark, K., Roggentin, P., Bockemu¨hl, J., Beckmann, G., Stark, R., Siegl, W., Ammon, A., Tscha¨pe, H., 2005. Molecular epidemiology of Salmonella enterica serovar Agona: characterization of a diffuse outbreak caused by aniseed-fennel-caraway infusion, Epidemiol. Infect. (cited from reference Koch et al. 2005), 133, 837–844. Schwarz, S., Liebisch, B., 1994. Pulsed-field gel electrophoretic identification of Salmonella enterica serovar Typhimurium live vaccine strain Zoosaloral H. Lett. Appl. Microbiol. 19, 469–472. Sorensen, O., van Donkersgoed, J., McFall, M., Manninen, K., Gensler, G., Ollis, G., 2002. Salmonella spp. shedding by Alberta beef cattle and the detection of Salmonella spp. in ground beef. J. Food Prot. 65, 484–491. Synnott, M.B., Brindley, M., Gray, J., Dawson, J.K., 1998. An outbreak of Salmonella Agona infection associated with precooked turkey meat. Commun. Dis. Public Health 1, 176–179. Tavechio, A.T., Ghilardi, A.C.R., Peresi, J.T.M., Fuzihara, T.O., Yonamine, E.K., Jakabi, M., Fernandes, S.A., 2002. Salmonella serotypes isolated from nonhuman sources in Sa˜o Paulo, Brazil, from 1996 through 2000. J. Food Prot. 65, 1041–1044. Taylor, J.P., Barnett, B.J., del Rosario, L., Williams, K., Barth, S.S., 1998. Prospective investigation of cryptic outbreaks of Salmonella Agona salmonellosis. J. Clin. Microbiol. 36, 2861–2864. Terletski, V., Schwarz, S., Carnwath, J., Niemann, H., 2003a. Subtracted restriction fingerprinting—a tool for bacterial genome typing. BioTechniques 34, 304–313. Terletski, V., Schwarz, S., Carnwath, J., Niemann, H., 2003b. Typing of Salmonella enterica subsp. enterica serovars Choleraesuis, Typhimurium, Dublin and laboratory strains of Escherichia coli using subtracted restriction fingerprinting (SRF) Microbiol. Res. 158, 135–142. Terletski, V., Michael, G.B., Schwarz, S., 2004. Subtracted restriction fingerprinting—a new typing technique using magnetic capture of tagged restriction fragments. FEMS Immunol. Med. Microbiol. 41, 1–8. Threlfall, E.J., Hampton, M.D., Ward, L.R., Rowe, B., 1996. Application of pulsed-field gel electrophoresis to an international outbreak of Salmonella Agona. Emerg. Infect. Dis. 2, 130–132.