General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: research.competitions@defra.gsi.gov.uk

SID 5 Research Project Final Report

SID 5 (2/05) Page 1 of 21

NoteIn line with the Freedom of Information Act 2000, Defra aims to place the results of its completed research projects in the public domain wherever possible. The SID 5 (Research Project Final Report) is designed to capture the information on the results and outputs of Defra-funded research in a format that is easily publishable through the Defra website. A SID 5 must be completed for all projects.

A SID 5A form must be completed where a project is paid on a monthly basis or against quarterly invoices. No SID 5A is required where payments are made at milestone points. When a SID 5A is required, no SID 5 form will be accepted without the accompanying SID 5A.

This form is in Word format and the boxes may be expanded or reduced, as appropriate.

ACCESS TO INFORMATIONThe information collected on this form will be stored electronically and may be sent to any part of Defra, or to individual researchers or organisations outside Defra for the purposes of reviewing the project. Defra may also disclose the information to any outside organisation acting as an agent authorised by Defra to process final research reports on its behalf. Defra intends to publish this form on its website, unless there are strong reasons not to, which fully comply with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.Defra may be required to release information, including personal data and commercial information, on request under the Environmental Information Regulations or the Freedom of Information Act 2000. However, Defra will not permit any unwarranted breach of confidentiality or act in contravention of its obligations under the Data Protection Act 1998. Defra or its appointed agents may use the name, address or other details on your form to contact you in connection with occasional customer research aimed at improving the processes through which Defra works with its contractors.

Project identification

1. Defra Project code VM2205

2. Project title

Genoprofiling of mulitiresistant Salmonella enterica

3. Contractororganisation(s)

Professor John Threlfall,Health Protection Agency,Centre for Infections,61, Colindale Avenue,LondonNW9 5EQ

54. Total Defra project costs £ 128,423

5. Project: start date................ 01 June 2006

end date................. 31 May 2007

SID 5 (2/05) Page 2 of 21

6. It is Defra’s intention to publish this form. Please confirm your agreement to do so...................................................................................YES NO (a) When preparing SID 5s contractors should bear in mind that Defra intends that they be made public. They

should be written in a clear and concise manner and represent a full account of the research project which someone not closely associated with the project can follow.Defra recognises that in a small minority of cases there may be information, such as intellectual property or commercially confidential data, used in or generated by the research project, which should not be disclosed. In these cases, such information should be detailed in a separate annex (not to be published) so that the SID 5 can be placed in the public domain. Where it is impossible to complete the Final Report without including references to any sensitive or confidential data, the information should be included and section (b) completed. NB: only in exceptional circumstances will Defra expect contractors to give a "No" answer.In all cases, reasons for withholding information must be fully in line with exemptions under the Environmental Information Regulations or the Freedom of Information Act 2000.

(b) If you have answered NO, please explain why the Final report should not be released into public domain

Executive Summary7. The executive summary must not exceed 2 sides in total of A4 and should be understandable to the

intelligent non-scientist. It should cover the main objectives, methods and findings of the research, together with any other significant events and options for new work.

Main project objectives1. A panel of strains of Salmonella Typhimurium phage type U288 with resistance to ampicillin,

chloramphenicol, streptomycin, sulphonamides, tetracyclines and trimethoprim (ACSSuTTm), isolated from cases of human infection, pigs and poultry or poultry meat in England and Wales in 2005 will be selected from the strain collections of the HPA and the VLA.

2. Drug resistance genes, plasmids and integron structures will be characterised by methods developed under VM02136.

3. Isolates will be characterised by PFGE and by VNTR fingerprinting, using in the latter case a consensus set of primers shared between the two organisations. Animal isolates will be characterised at the VLA and human and food isolates at the HPA. Strains will be exchanged to assess the validity and comparability of results.

4. Similar studies will be instigated for multiresistant (MR) S. Typhimurium DTs 104 (human, bovine, porcine and poultry isolates), and 193 (human and pig isolates) and MR S. Paratyphi B variant Java (human, cattle and poultry isolates). It is possible that the primers used, which have been developed specifically for S. Typhimurium, may not be appropriate for S. Java. In this eventuality new VNTR primers specific for S. Java will be designed and field-tested.

5. Where appropriate, all isolates of S. enterica exhibiting resistance to third-generation cephalosporins and studied under VM02136 will be further subtyped by genoprofiling, using as comparison strains of Findings

A literature search identified potential VNTR schemes for Salmonella, which were then tested using five strains of Salmonella from the top ten human serotypes plus S. Montevideo and S. Ajiobo. These primer sets gave inconsistent amplification and poor discrimination between serotypes.

A new VNTR database was searched to identify novel repeat regions within the sequenced genomes of salmonellae that could be used for typing. Nine loci were identified and primers tested with isolates of the top ten human serotypes plus S. Montevideo and S. Ajiobo. The new primers were able to amplify in most of the serotypes tested. However, there were insufficient loci to provide the level of discrimination required for a VNTR typing scheme.

A recently published VNTR scheme for S. Enteritidis was tested. This could distinguish between phage types and the data suggest it may prove useful as a future additional typing strategy for S. Enteritidis.

Salmonella isolates belonging to eight distinct outbreaks, and also multiple isolates from the same patient were analysed by PFGE and VNTR typing to determine concordance between the methods. VNTR provided better discrimination than PFGE for some phage types and could therefore be used in

SID 5 (2/05) Page 3 of 21

definition of outbreak-associated cases of infection during parallel outbreaks of common types with the same phage type/PFGE type. VNTR was also able to distinguish between concurrent, unrelated strains that may be misclassified as part of an outbreak by initial epidemiological information or PFGE. Small changes in loci were noted within both the course of a salmonella outbreak and even salmonella carriage by the patient, but overall the data suggest that VNTR-typing is stable enough to be used in outbreak characterization.

67 isolates of Typhimurium U288 were characterised by PFGE, VNTR and plasmid profiling: 27 of human origin (10 of which were believed to be associated with an outbreak and counted as only one isolate for the purposes of this study), 10 isolates from food (5 from raw chicken, 5 from raw pork sausage meat), 27 from pigs, two from turkeys and one environmental isolate. VNTR typing was more discriminatory than PFGE for typing U288 isolates, identifying 37 profiles compared to 17 by PFGE. The most common PFGE profile could be subdivided into 20 different VNTR types. Outbreak isolates were easily differentiated from other U288 by VNTR, but not by PFGE.

The panel of DT104 isolates consisted of 18 isolates from cattle, 21 from pigs, 22 from food (raw beef, pork and chicken products, milk), 17 from cases of human infection and 22 isolates from chickens. Isolates were chosen from between 2002 and 2006, and had a diverse range of resistance profiles. VNTR typing was much more discriminatory than PFGE for typing DT104 isolates, identifying 80 profiles compared to 15 by PFGE. The most common PFGE profile could be subdivided into 50 different VNTR types.

A panel of S. Typhimurium DT193 consisting of 28 isolates from human, 61 from animals and 10 from food were typed by VNTR and PFGE. VNTR fingerprinting of 99 isolates of this heterogeneous phage type DT 193 has demonstrated 75 allelic profiles. Fifteen clusters were identified, 2 of the clusters were food specific, 5 were animal specific, 2 were human specific and 6 contained both animal and human isolates. These results indicate that VNTR fingerprinting can provide a rapid and robust method of subtyping within this phage type, which give slightly better discrimination than PFGE.

Anticipated sequence data from S. Paratyphi B was not available at the time of study, thereby preventing the construction of sequence-valid primers for the VNTR fingerprinting of S. Paratyphi B variant Java. This organism was therefore excluded from the investigation. VNTR data for S. Enteritidis became available in the latter stages of the study and preliminary studies indicate that it would provide a higher level of discrimination for this serotype than current PFGE methods.

16 third generation cephalosporin resistant S. Typhimurium were subtyped by VNTR. Isolates produced very different VNTR profiles, regardless of mechanism of resistance to third-generation cephalosporins. This shows that emergence of resistance to these antimicrobials is still a somewhat random event and occurs in unrelated strains.

CTX-M and AmpC plasmids isolated under VM02136 were tested using the PCR-based inc/rep incompatibility group assay. There was a high prevalence of some replicons such as repI1, repA/C and repFII in association with relevant extended-spectrum cephalosporin resistance genes, such as the blaCMY-2 and blaCTX-M-15 genes.

The inc/rep incompatibility group assay has proved useful in characterisation of plasmids that are associated with emerging resistance genes.

63 enterobacterial isolates isolated between October 2006 and April 2007 exhibiting reduced susceptibility to ciprofloxacin (MIC = 0.125 – 1.0 mg/L) but were fully susceptible to nalidixic acid (MIC <16 mg/L), or exhibiting high-level resistance to ciprofloxacin (MIC > 1.0 mg/L) were screened for plasmid-mediated quinolone resistance genes qnrA, qnrB and qnrS by multiplex PCR.

No isolates were positive for qnrA. One isolate of Citrobacter freundii and two isolates of S. enterica (one Schwarzengrund, one Typhimurium DT49b) were positive for qnrB. Thirty-seven salmonellae and two Shigella flexneri were positive for qnrS; six isolates were serotype Corvallis, 21 Typhimurium and 10 Virchow. Nineteen of the 21 Typhimurium isolates were either phage type DT120 or DT193, and 7/10 Virchow isolates were phage type DT43.

VNTR typing of qnr-positive S. Typhimurium showed that they could be differentiated from other Typhimurium isolates due to their unusual VNTR profile.

90 enterobacterial isolates from 2006 to April 2007 expressing resistance to ceftriaxone and/or cefotaxime were screened for the presence of CTX-M, AmpC, TEM, SHV and OXA-1 group genes by PCR.

16.9% salmonellae were positive for CTX-M genes, 20.8% were positive for AmpC genes, 26% were positive for TEM genes, 15.6% were positive for SHV genes and 1.3% were positive for OXA-1 group

SID 5 (2/05) Page 4 of 21

genes.

The resistance gene nanoarray developed under VM02136 was used to identify the resistance genes in 6 isolates of Typhimurium U288.

Ampicillin resistance was encoded by blaTEM, chloramphenicol by a cmlA1-like gene, streptomycin by aadA1-like and aadA2-like genes, sulphonamides by sul2 and sul3, tetracyclines by tet(A) and trimethoprim by dfr12. One isolate had an additional aac(3’)-IVa gentamicin resistance gene. All isolates were also positive for the intI1 gene, which encodes the class 1 integrase.

Primers specific for these resistance genes were then used to screen the entire collection of U288 to determine whether the same genes were present in all isolates. All isolates of R-type ACSSuTTm carried the same genes as identified by nanoarray.

30 S. Typhimurium DT193 with R-type ASSuT were tested by PCR for resistance genes. blaTEM

encoded resistance to ampicillin, strA resistance to streptomycin and sul2 resistance to sulphonamides. Tetracycline resistance was encoded by tetB in four isolates and tetA in the remainder. Class 1 integron gene cassettes were not amplified from any of the isolates.

Genoprofiling data of the panel of strains described above was integrated with other data, including antimicrobial susceptibility and PFGE profiles, to provide a fully comprehensive microbiological and epidemiological profile of the strains in the test panel.

Project Report to Defra8. As a guide this report should be no longer than 20 sides of A4. This report is to provide Defra with

details of the outputs of the research project for internal purposes; to meet the terms of the contract; and to allow Defra to publish details of the outputs to meet Environmental Information Regulation or Freedom of Information obligations. This short report to Defra does not preclude contractors from also seeking to publish a full, formal scientific report/paper in an appropriate scientific or other journal/publication. Indeed, Defra actively encourages such publications as part of the contract terms. The report to Defra should include: the scientific objectives as set out in the contract; the extent to which the objectives set out in the contract have been met; details of methods used and the results obtained, including statistical analysis (if appropriate); a discussion of the results and their reliability; the main implications of the findings; possible future work; and any action resulting from the research (e.g. IP, Knowledge Transfer).

INTRODUCTIONMultiple drug resistance (MR) in strains of Salmonella enterica, which have their primary reservoirs in food-production animals and are spread to humans through the food chain are now a major cause of concern to animal welfare and to human health in many countries throughout the world. This has been particularly apparent in the UK, where resistant strains, particularly of S. Typhimurium but also of other serotypes, have caused outbreaks of infection in both humans and food production animal since the early 1990s. A further concern in the last five years is the appearance of resistance to ‘new’ antimicrobials, particularly fluoroquinolones and third-generation cephalosporins, in such strains. To control the spread of MR strains through the food chain it is important to be able to fully characterise isolates from both food animals, not only in terms of drug resistance but also in relation to their phenotypic and molecular profile. The establishment of rapid response resistance gene profiling under the auspices of VM02136 has resulted in the identification of a range of genes coding for resistance to third-generation cephalosporins and their host plasmids, and also of mutations conferring resistance to quinolone antimicrobials. A key area that now needs addressing is the precise identification, at the molecular level, of such resistant strains from humans and food animals. Pulsed-field gel electrophoresis (PFGE) is currently the internationally-accepted ‘gold standard’ for molecular sub-typing and databases of PFGE subtypes have been established world-wide (Peters et al. 2003; Swaminathan et al. 2001). Regrettably certain key strains of MR S. enterica – e.g S. Typhimurium DT 104 – fall into relatively few PFGE profiles and further subdivision may be required for detailed outbreak investigations and for tracing sources of infection through the food chain. Such subdivision may be achieved by sequence-based molecular sub-typing as may be provided by a method such as Variable Number of Tandem Repeats (VNTR) fingerprinting. Preliminary results have indicated that common PFGE types can be further subdivided using this method (Lindstedt et al. 2004; Lindstedt et al. 2003; Liu et al. 2003; Ramisse et al. 2004)

SID 5 (2/05) Page 5 of 21

We have extended the work done under VM 02136 by combining VNTR with PFGE and resistance gene profiling (genoprofiling) to the study of a few defined serotype and phage types of MR Salmonella enterica with different animal reservoirs, which have caused infections in humans. The adoption of such genoprofiling could facilitate the development of intervention strategies to limit or control the spread of such strains and provide a framework for subsequent risk assessment and formulation of policy.

The following questions were addressed:

1. Can genoprofiling be used to further characterise MR S. enterica and MR S. enterica with additional

resistance to third-generation cephalosporins and / or fluoroquinolone antimicrobials?

2. Is genoprofiling suitably robust for inter-laboratory comparisons?

3. Are the results amenable for incorporation into existing databases of strain types?

4. Can results be achieved in real-time?

5. Can the results be used to assist in the recognition of newly emerged drug-resistant strains?

6. Can the information be incorporated into existing and proposed international networks?

OBJECTIVES1. Construction of primers for the amplification of selected VNTR loci.

2. Selection of strains – HPA, human and food isolates; VLA, veterinary isolates.

3. Characterisation of drug resistance genes, integron structures and plasmids.

4. Genoprofiling.

5. Database entry (HPA).

ACHIEVEMENT OF OBJECTIVES1. Construction of primers for the amplification of selected VNTR loci.In the first instance the choice of VNTR loci was based on the sequences published by Lindstedt et al (2004). In parallel with this a literature- and GenBank search was initiated to identify primer sequences which may be suitable for amplification of selected VNTR loci in other S. enterica serovars, particularly S. Paratyphi B var. Java.

Milestones completed within this section: Complete literature search, Complete assembly of agreed panel of strains - HPA and VLA. (1) Complete construction of primers for PCR amplification of selected S. Typhimurium VNTR loci. (2) Complete validation of primers using S. Typhimurium as validation serotype. (3) Complete construction of primers for other serotypes wherever possible. Complete field trial of new primers. Complete transfer of technology and materials between collaborating institutes.

1.1 Summary of current primers available for such schemesA literature search was carried out to identify published papers describing VNTR strategies with potential additional VNTR loci for Salmonella. Four main papers were identified, these being two papers by Lindstedt et al (2003, 2004), a paper by Ramisse et al (2004) and a paper by Liu et al (2003). The papers by Lindstedt et al (2003, 2004) described the VNTR system already in use and five additional loci. These additional loci were designed for S. Typhimurium and were considered suitable for testing for S. Typhimurium as well as other serotypes of Salmonella. The Ramisse et al paper described seven primer sets for S. enterica and the Liu et al paper described a number of different primers for S. Typhi, but two of these (TR1 and TR5) had been described as suitable for testing on other serotypes of Salmonella (See Appendix 1 for list of primers).

1.2 Searches carried out to identify new VNTR loci which could be usedNew VNTR regions were searched for “in silico” by the use of a new VNTR database “The Variable Number Tandem Repeats Database” (VNTRDB; http://www.hpa-bionum.org.uk/VNTRUK/). This had the added advantage of also searching incomplete genomes as well as fully sequenced ones, giving a greater coverage of serotypes. Multiple Salmonella genomes were searched, which included S. Typhimurium, S. Hadar, S. Infantis, S.

SID 5 (2/05) Page 6 of 21

Paratyphi, S. Typhi, S. Gallinarum and S. Choleraesuis. Genomes were searched for repeats by applying specific criteria for the types of repeats required. Fourteen potential new VNTR regions were identified, of which 9 appeared to be present in all strains including Gallinarum, Hadar, Infantis, Paratyphi, Typhimurium and Typhi and were investigated further for potential use as new VNTR loci. These appeared to have favourable characteristics, e.g. small repeat size in the range (2-9bp), a high % match (85-100%) and a high number of copy number differences (4-12) between strains. The sequences of the flanking regions and repeat regions were concatenated then aligned in the program Bioedit. Primers were successfully designed using Primer Express to amplify the repeat regions (see Appendix 1).

1.3 Testing of a panel of primers with the top 10 serotypesPrimers were tested using a panel of strains, which included the top ten serotypes of Salmonella (see Appendix 1) as well as S. Montevideo and S. Ajiobo. Testing involved using primers labelled with fluorescent dyes so they could be accurately sized using the Beckman capillary electrophoresis system and multiplexed. For all the primer sets tested there was little variability detected between serotypes and in some cases inconsistent amplification. The study highlighted a number of problems with the primer sets tested in relation to repeat size and primer specificity. For future studies the use of VNTR loci with small repeat sizes and more specific primers may provide better discrimination (for further information see Appendix 1).1.4 Testing of new primer panelsNewly designed primers were tested on the panel of eleven different serotypes of Salmonella (Appendix 1). The new primer sets successfully amplified in most serotypes although they were more successful for serotypes S. Enteritidis, S. Virchow, S. Newport, S. Hadar, S. Infantis, S. Goldcoast, S. Kentucky, S. Java. In order to establish a new VNTR scheme a number of highly discriminatory VNTR loci needed to be assembled and unfortunately from this panel of primers tested there were insufficient loci determined to provide a sufficient level of discrimination. However there were a number of loci, which could be investigated further in conjunction with novel loci, which may be identified in the future (Appendix 1).

During the course of this study a VNTR scheme for S. Enteritidis was described which utilised a combination of ten novel loci (see Appendix 1 and Boxrud et al. 2007). As part of this study the new scheme was tested on a variable panel of S. Enteritidis to determine how successful it could be as a future VNTR scheme. The primers amplified the loci successfully and the fragment sizes detected matched those described by the author suggesting that the scheme could be used to provide comparable results. Interestingly from the small panel of isolates tested a number of novel loci, not described by the author were identified, suggesting that the method would provide a high degree of strain discrimination. Future studies aim to further test and develop this scheme.

1.5 Investigation into the stability of VNTRs in S. TyphimuriumThere is scarce data on the stability of VNTR markers in bacteria, and on the correlation between pulsed-field gel electrophoresis (PFGE) and VNTR-typing of salmonella in the course of outbreak investigations. Potentially VNTRs may evolve so rapidly that multiple types could emerge during an outbreak initially caused by a single clone, thereby diminishing its utility for outbreak detection (Noller et al. 2003). Salmonella isolates (see Table 1) belonging to eight epidemiologically-distinct outbreaks, and also multiple isolates from the same patient were analysed by PFGE and the VNTR typing scheme of Lindstedt et al. (2004), and the data compared with epidemiological data to determine concordance between the methods (Appendix 1). The study showed VNTR provided better discrimination than PFGE for some phage types. VNTR-typing was able to distinguish between isolates from three outbreaks and isolates from a patient that were all designated definitive phage type (DT) 104 or 104b, and indistinguishable by PFGE. All isolates produced PFGE pattern STYMXB.0061, which is the predominant DT104 profile in England and Wales. Two outbreaks were both caused by Typhimurium DT104 resistant to streptomycin, spectinomycin and sulphonamides and receipt of samples in LEP overlapped by six days. VNTR-typing was able to distinguish between the two outbreaks based on differences at three loci. This demonstrates how VNTR-typing can be used in definition of outbreak-associated cases of infection during parallel outbreaks of common types with the same phage type/PFGE type. VNTR-typing was also able to distinguish between concurrent, unrelated strains that may be misclassified as part of an outbreak by initial epidemiological information or PFGE. Small changes in loci as a result of loss or gain of a single 6-bp repeat (single locus variants) were noted within both the course of a salmonella outbreak and even salmonella carriage by the patient. Application of a cut-off defined by Noller et al. (2003) for analysis of Escherichia coli O157 VNTR data as a difference of two repeats at one or fewer loci allowed correct classification of these Typhimurium isolates as part of an outbreak. Analysis of further outbreaks is needed to determine whether this model is widely applicable to the S. Typhimurium VNTR scheme, although the decision to include these isolates as part of an outbreak was confirmed by PFGE data since each single locus variant had a PFGE profile identical to the main outbreak profile. Isolates identical by VNTR-typing but with one to three band differences in PFGE profile were also occasionally noted in outbreaks. Our data suggest that VNTR-typing is stable enough to be used in outbreak characterization. Further information is available in Hopkins et al. 2007. J Clin Microbiol. 45:3058-61.

Table 1:Outbreak or patient Phage type No. of isolates*

SID 5 (2/05) Page 7 of 21

A U288 7B DT104 8C DT104 71D DT208 2E DT135 3F DT15 7G DT104 90H DT94 2patient I DT104 2patient II DT104b 2patient III DT41 2patient IV RDNC 3patient V untypable 2patient VI unknown 2patient VII DT8 2

* isolates were all those available in the HPA culture collection and are not a selected subset from each outbreak/patient.

APPENDIX 1. List of current and new primers for VNTR analysis of S. Typhimurium.

Results of testing current STM primers with top ten serotypes.

Additional loci identified and construction and testing of primer panel.

VNTR method.

2. Selection of strains - – HPA, human and food isolates; VLA, veterinary isolates and 4. Genoprofiling

S. Typhimurium U288 isolates were characterised by PFGE and by VNTR fingerprinting, using in the latter case a consensus set of primers shared between the two organisations. Animal isolates were characterised at the VLA and human and food isolates at the HPA.

Similar studies were performed for MR S. Typhimurium DT 104 and 193 isolates. Because anticipated sequence data from S. Paratyphi B was not available at the time of study, thereby preventing the construction of sequence-valid primers for the VNTR fingerprinting of S. Paratyphi B variant Java, this organism was excluded from the investigation.

Milestones completed within this section: Complete VNTR fingerprinting of panel of S. Typhimurium phage types U288, 104 and 193 isolates.

2.1 Genoprofiling of isolates of S. Typhimurium from animals and humans by VNTR, PFGE and plasmid profiling.

2.1.1 S. Typhimurium phage type U288 (MK KH)

2.1.1.1 Criteria for selection. A panel of 60 strains of S. Typhimurium phage type U288 of R-type ACSSuTTm from cases of human infection (20 strains), pigs (20 strains) and poultry or poultry meat (20 strains), isolated in England and Wales in 2005 were to be selected from the culture collections of the HPA and the VLA. There were insufficient isolates from poultry meat in the HPA collection (only 5) and poultry in the VLA collection (only two) despite expanding the search criteria to include strains isolates from 2004 and 2006, therefore additional isolates from other animal sources were selected. Isolates were selected based on epidemiological and/or patient data to ensure that apparently related isolates were not chosen, and with most having the ACSSuTTm phenotype or variations of this. Isolates from animals were also selected based on the resistance phenotype.

In total we tested 67 isolates: 27 of human origin (10 of which were believed to be associated with an outbreak and counted as only one isolate for the purposes of this study), 10 isolates from food (5 from raw chicken, 5 from raw pork sausage meat), 27 from pigs, two from turkeys and one environmental isolate. Numbers of S. Typhimurium isolations in England and Wales are increasing, from 73 in 2004 to 151 isolates in 2006.

2.1.1.2 PFGE analysis of isolates. PFGE blocks were prepared as described in the SalmGene protocol (Appendix 2). Digested blocks were run on 1.2% agarose gels in buffer prechilled to 14C at 6V/cm for 22hr with switch times of 2-64 secs. The gels were photographed and the electronic images imported into the PulseNet Europe PFGE database assignment of profile types by comparison to those already in the database. Dendrograms were constructed using the Dice similarity coefficient and the unweighted pair group method with

SID 5 (2/05) Page 8 of 21

arithmetic averages (UPGMA); position tolerance settings were Optimization = 1.5% and Position tolerance = 1.5%.

All but one U288 isolate were typable by PFGE (see Figure 1, Appendix 3). Overall the U288 isolates shared approx. 88% genetic similarity when isolate H061460168 was discounted, as this appears to be a mixed culture from the PFGE profile (Figure 1, Appendix 3). Seventeen different profiles were identified, although many of these are very similar. Variation between profiles was often due to one fragment that varied in size from approx. 150 to 170 kb between isolates and was considered by Bionumerics to be the same band. A number of fragments at the bottom of the profiles were less than 20.5 kb in size and therefore quite faint. These small fragments are probably plasmids as a number were identified (see 2.1.1.4). Variation in gel quality between runs and the difficulty in deciding whether these faint bands are present or not may have led to some artificial variation between the U288 profiles. The most commonly identified profile, STYMXB.0179 represented 27 isolates; these were from humans, food and animals. The outbreak-associated isolates did not cluster together, but still shared >88% similarity. This is more likely due to the artificial variation discussed above than true variation between the profiles.

There are some discrepancies between clustering as shown in the dendrogram and the profile name, for example a STYMXB.0253 profile clustering together with STYMXB.0213 profiles. This is because the profile naming and clustering of isolates to produce a dendrogram are two separate processes, both of which have a margin of error. This highlights one of the problems of comparing PFGE data, which is not present when comparing VNTR data.

There were some problems experienced with degradation of DNA if blocks were not digested and analysed straight away. PFGE profiles for some isolates were really faint despite using a photometer to ensure the same cell density was used per strain and adding thiourea to the gel running buffer. However, VNTR profiles could be obtained from these isolates.

2.1.1.3 VNTR analysis of isolates. Boiled cell template was prepared by resuspending one colony in 50l of distilled water and adding 50l of 15% Chelex 100 (BioRad). The cell suspension was heated at 100 C for 10 min and then centrifuged at 13,000 rpm for 10 min. The supernatant was transferred to a fresh tube and stored at -20C until required. VNTR analysis was carried out as described in Appendix 1. Fragment sizes were imported into Bionumerics and the allele assigned to each locus using a script. Dendrograms were constructed using categorical coefficients and the Ward algorithm. Polymorphism indexes were calculated for each locus as 1-(allele frequency)2 (see Appendix 3).

66 isolates were typable by VNTR; one pig isolate was untypable by PFGE and VNTR. VNTR analysis was more discriminatory than PFGE, identifying 37 different profiles, 10 of which were shared by more than one isolate (see Figure 2, Appendix 3). Three of these clusters were a mixture of human and animal strains, three were specific to human or human and food isolates, and four were specific to pigs or pork food products. This suggests that certain strains may be more frequent in some hosts, such as pigs than from other sources, but it is not possible to say whether these isolates are host-specific due to the relatively small panel of strains analysed. Other strains isolated from pigs appear able to cause salmonellosis in humans as strains with identical VNTR profiles were isolated from cases of human infection. PFGE profile STYMXB.0179 could be subdivided into 20 different VNTR profiles (Figure 3, Appendix 3). In comparison to PFGE data, the outbreak isolates were easily distinguished from the other isolates by VNTR analysis (see Figure 2, Appendix 3).

The loci varied in their degree of polymorphism, as indicated by the polymorphism index (see Table 2 below). STTR10pl was the most polymorphic locus among U288s of human origin, while in isolates from animals loci STTR5, STTR6 and STTR10pl all had high polymorphism indexes >0.7. Human isolates had lower polymorphism indexes for each locus than strains from pigs and turkeys. Amongst all U288 isolates, all loci but STTR3 were considerably less polymorphic than noted by Lindstedt et al. (2004). This highlights the clonal nature of the U288 phage type. Drahovská et al. (2007) noted that VNTR grouping of S. enterica strains corresponded well with the phage type of the strains.

Table 2:Source STTR9 STTR5 STTR6 STTR10pl STTR3human (n=18*) 0 0.54 0.53 0.75 0.28animals 0.13 0.72 0.78 0.81 0.35all strains 0.07 0.66 0.74 0.82 0.3Lindstedt scheme 0.51 0.87 0.9 0.92 0.25*Only one of the ten U288 outbreak isolates were included in the comparison

The most frequent alleles at each locus for isolates from humans and pigs and the percentage of isolates with this allele are shown below in Table 3. The profiles were unusual in that the most frequently occurring alleles seen in the Lindstedt et al. (2004) scheme (see Appendix 1) did not feature significantly in any of the isolates; only allele 8 at STTR5 (in two isolates) and allele 3 at STTR3 (in one isolate) were identified.

Table 3: Humans Pigs

Loci Allele % Allele %

SID 5 (2/05) Page 9 of 21

STTR9 4 100 4 100STTR5 15 44.4 15 48.1STTR6 5 44.4 5 29.6STTR10 21 37 1 29.6STTR3 2 59.2 2 77.8

2.1.1.4 Plasmid profiling of isolates. Plasmid DNA was isolated using the method of Kado and Liu (1981) (see Appendix 2), and run on a 0.6% agarose gel with Escherichia coli K12 strain 39R861 as the plasmid molecular mass marker. Plasmid sizes were determined with reference to plasmids of 148, 63, 36 and 6.9 kilobases (kb) (98, 42, 23.9 and 4.6 megadaltons (mDa) respectively) carried in E. coli 39R861 and a supercoiled DNA ladder (Invitrogen).

Twenty-seven different plasmid profiles were identified among the 67 U288 isolates (see Appendix 3). Seventeen profiles were unique to a single isolate; only four of these were of human origin. The most common profile among human isolates contained two plasmids of 1.5 and 98 kb, while the most common profile in animal isolates contained three plasmids of 5, 7.3 and 98 kb. A plasmid of 98 kb was found alone or in combination with other plasmids in 53 (87%) of isolates.

APPENDIX 2 PFGE and plasmid profiling methods used. APPENDIX 3 Results for U288 from humans, food and animals.

2.1.2 S. Typhimurium DT 104

2.1.2.1 Criteria for selection. A panel of 100 isolates previously characterised by PFGE, including historical isolates, outbreak strains, isolates from foods, and strains from cattle, pigs and poultry, were to be selected. Our panel of isolates consisted of 18 isolates from cattle, 21 from pigs, 22 from food (raw beef, pork and chicken products and milk; these represent all the food strains in the HPA strain collection isolated between 2002 and 2006 that had no apparent epidemiological link), 17 from cases of human infection (of which 16 were outbreak-associated and shown during the study described in section 1.5. to have variable VNTR profiles) and 22 isolates from chickens. Isolates were chosen from between 2002 and 2006, were selected based on epidemiological and/or patient data to ensure that apparently related isolates were not chosen, and also to reflect a diverse range of resistance profiles. Animal isolates were also selected to give a wide diversity in resistance phenotype and only a single isolate was selected from a farm submission.

and had a diverse range of resistance profiles (Appendix 4).

2.1.2.2 PFGE analysis of isolates. PFGE analysis was performed as described in section 2.1.1.2. Ninety-eight of the DT104 panel were subtyped by PFGE; two were omitted because the strains looked contaminated on culture. Overall the DT104 isolates shared approx. 77% genetic similarity (Figure 1, Appendix 4). Fifteen different profiles were identified; the largest being of profile STYMXB.0061, which represented 59% of isolates. Isolates in this cluster were from chickens, cattle, pigs, food and humans. Analysis of Typhimurium DT104 PFGE profiles submitted to the SalmGene database from nine European countries shows that profile STYMXB.0061 was the most common DT104 profile in Denmark, Spain, England and Wales, and The Netherlands, representing between 36 – 54% of DT104 profiles in these countries (Gatto et al. 2006). With the exception of profiles STYMXB.0051, 0310, 0241, 0199 and 0309 (see bottom of Figure 1) profiles differed by presence/absence of only one of two bands. In contrast to the U288 isolates it was easy to differentiate between these very similar profiles.

2.1.2.3 VNTR analysis of isolates. VNTR analysis was performed as described in section 2.1.1.3. Ninety-nine isolates were typable by DT104; one pig isolate (L00077-05) did not amplify from any loci. Cattle isolate S04853-05 only produced an amplicon for locus STTR10pl therefore was excluded from the dendrogram (Figure 2, Appendix 4). VNTR typing is believed to be the most discriminatory DNA-based typing method for fingerprinting DT104 isolates. As shown previously by Lindstedt et al. (2003) VNTR typing was more discriminatory than PFGE, with 80 profiles being identified. Only 11 profiles were shared by more than one isolate, two such clusters being associated with outbreaks of human infection. Clusters of food, pig, chicken and cattle isolates were also noted. Five clusters of animals isolates were detected, of these, three clusters contained isolates linked by a farm or farmer. PFGE profile STYMXB.0061 could be subdivided into 50 VNTR profiles, only six of which were shared by more than one isolate (Figure 3, Appendix 4). In comparison to PFGE data, the outbreak isolates were easily distinguished from the other isolates by VNTR analysis.

The loci varied in their degree of polymorphism, as indicated by the polymorphism index (see Table 4 below). The most frequent alleles at each locus and the percentage of isolates with this allele are shown below in Table 5. Loci STTR6 and STTR10pl were the most polymorphic among all categories of isolates except human isolates. The human isolates show the least polymorphism as they were associated with two outbreaks. These loci were also shown to be the most polymorphic by Lindstedt et al. (2004). Both loci are situated on mobile genetic elements: locus STTR6 within the Gifsy-1 bacteriophage of S. Typhimurium LT2, while STTR10pl is

SID 5 (2/05) Page 10 of 21

located on the S. Typhimurium serotype-specific plasmid pSLT. Comparing the polymorphism indexes for each locus within isolates from cattle, pigs and chickens reveals some differences. Chicken isolates were considerably more polymorphic at locus STTR9 and less polymorphic at locus STTR6 than isolates from pigs or cattle, while cattle isolates were more polymorphic at loci STTR5 and STTR3 than isolates from pigs or chickens. Overall, DT104 isolates were less polymorphic at loci STTR9 and STTR5, but more polymorphic at STTR3 than observed by Lindstedt et al. (2004).

Table 4:Source STTR9 STTR5 STTR6 STTR10pl STTR3Cattle 0.2 0.75 0.81 0.84 0.57Pig 0.1 0.65 0.87 0.89 0.34Chickens 0.37 0.67 0.74 0.81 0.38All animal isolates 0.24 0.75 0.87 0.90 0.45Food 0 0.73 0.86 0.83 0.24Animal + Food 0.18 0.75 0.89 0.91 0.41Human 0 0.54 0.65 0.59 0All isolates 0.13 0.73 0.89 0.92 0.35Lindstedt scheme 0.51 0.87 0.9 0.92 0.25

Table 5:Cattle % Pig % Chicken % All animal %

STTR9 2 88.9 2 95 2 77.3 2 86.7STTR5 7 38.9 7, 8 35* 7 54.5 7 43.3STTR6 8 33.3 9 25 10 40.9 10 20STTR10pl 4, 26 22.2* 5 20 0 36.4 0 20STTR3 3 55.6 3 80 3 77.3 3 71.7

food % animal + food % human % All isolates %STTR9 2 100 2 90.2 2 100 2 92.9STTR5 7 45.5 7 43.9 7 64.7 7 47.5STTR6 0, 6 18.2* 8 17.1 10 52.9 10 22.2STTR10pl 9 27.3 0, 5, 26 13.4 4 58.8 4, 5 15.2*STTR3 3 86.4 3 75.6 3 100 3 79.8* percentage for each allele

2.1.2.4 Plasmid profiling of isolates. Plasmid profiling was performed as described in section 2.1.1.4. In a study of 600 DT104 isolates from humans, food and animals with resistance patterns ACSSuT, ACSSuTCp and ACSSuTTm, the most common plasmid profile identified contained just one plasmid of 60 mDa (corresponding to the Typhimurium-specific mouse virulence plasmid) and was characteristic of the epidemic strain of DT104 (Threlfall et al. 1996). Plasmid profiling of isolates from this study found a single isolate of 60 mDa was the most common profile.

APPENDIX 4 Results for DT104 from humans, food and animals.

2.1.3 S. Typhimurium DT 193 DT 193 consistently falls within the top five phage types in isolations of S. Typhimurium from cases of human infection and from food production animals, including both cattle and pigs, in England and Wales. DT 193 is a heterogeneous type, which can be developed from other existing common phage types by acquisition of bacteriophages and plasmids. Thus DT 193 per se is a relatively meaningless designation and identification of subtypes within this phage type is an essential prerequisite for epidemiological investigations. In the past such subtypes have been identified on the basis of antibiogram, plasmid profile and/or PFGE profile. These methods can be subjective, and more robust typing is very important for tracing of strains through the food chain.

2.1.3.1 Criteria for selectionA panel of 102 isolates of Salmonella Typhimurium phage-type DT193 from humans (28), animals (63) and food (11), isolated in England and Wales between 2002 and 2007 were selected from the strain collections of the HPA and VLA. Animal isolates originated from pigs (39), cattle (12), chicken (4), turkey (1), sheep (1), goat (1), dog (1), cat (1), horse (2) and fox (1). Due to the potential heterogeneity in PFGE profiles, twenty human isolates were pairs of isolates which shared identical PFGE profiles and were selected to determine whether VNTR could further distinguish between them. Isolates were selected based on epidemiological and/or patient data to ensure that apparently related isolates were not chosen, and also to reflect a diverse range of resistance profiles (appendix 5). Animal isolates were selected based on resistance phenotype. For chicken, all isolates with resistance information were selected as DT193 is not common in this species. For isolates from cattle and pigs, these were selected to give a wide variety of resistance phenotypes from different farm submissions.

SID 5 (2/05) Page 11 of 21

2.1.3.2 PFGE analysis of isolates. PFGE analysis was performed as described in section 2.1.1.2All isolates in the panel were also typed by PFGE. Ninety-six isolates were typable by PFGE. Five isolates gave profiles that were not typable due to degradation of the DNA and 2 isolates were unnamed as the profiles were very different they may be mixed cultures or not Typhimurium. Sixty-four different PFGE profiles were identified among the isolates tested. The DT193 isolates overall sharing 51.06% similarity (Figure 1, Appendix 5). The most common PFGE profiles detected in this study were STYMXB.0131 which was identified in 4 human isolates and 1 cattle isolate, and STYMXB.0179 which was shared by 5 pig isolates. Among the 10 pairs of human isolates selected with identical PFGE profiles, only 3 pairs also had identical VNTR types (2 of these pairs had PFGE profile STYMXB.0131). Also 2 isolates had profile STYMXB.0061, which is the most common DT104 profile.

2.1.3.3 VNTR analysis of isolates. VNTR analysis was performed as described in section 2.1.1.3.Ninety-nine isolates were typable by VNTR, one food isolate that was selected (62160268) was found not to be S. Typhimurium when tested and two animal isolates (L1412-05 and L0801-05) could not be typed. VNTR identified 75 profiles among all isolates tested. Therefore VNTR was slightly more discriminatory than PFGE for isolates of DT193 in this study. 15 VNTR profiles were shared by more than one isolate (see Figure 2, Appendix 5). Two of the clusters were food specific, 5 were animal specific, 2 were human specific and 6 contained both animal and human isolates. Polymorphism indexes demonstrated that the loci differed in their degree of polymorphism (see table 6). The most polymorphic loci were STTTR5 and STTR6. STTR10 was less polymorphic than was demonstrated by Lindstedt et al. (2004). Animal and human isolates were more polymorphic than food isolates at all loci, however the number of food isolates tested was small. Among pig isolates, which were the largest group of animal isolates tested, the polymorphism indexes were lower slightly lower then the overall values. The most frequent alleles at each locus and the percentage of isolates with this allele are shown below in Table 7. The most frequent allele at STTR10pl was 0 in all isolates tested. This indicates that no amplicon was observed for this locus in these isolates. STTR10pl is located on the S. Typhimurium serotype-specific plasmid pSLT and was found to be absent in 48% of the isolates tested by Lindstedt et al. (2004). The most frequent amplified allele at STTR10pl was 14, which was detected in 20% of isolates. Among the animal isolates 5 VNTR allele strings were observed in more than one isolate. In two cases the isolates originated from the same farm (2-4-5-0-2 and 4-19-4-21-2).

SID 5 (2/05) Page 12 of 21

Table 6:

STTR9 STTR5 STTR6 STTR10pl STTR3Human 0.806 0.832 0.893 0.658 0.446Food 0.340 0.820 0.740 0.480 0.420Animal 0.649 0.898 0.859 0.778 0.372Pig 0.548 0.869 0.836 0.752 0.228All strains 0.654 0.890 0.879 0.730 0.430Lindstedt scheme 0.51 0.87 0.9 0.92 0.25

Table 7:

2.1.3.4 Plasmid profiling of isolates. Plasmid profiling was performed as described in section 2.1.1.4. Plasmid profiling was performed on 82 isolates of S. Typhimurium DT193 from this study. The plasmid content of the isolates was variable. For 17 isolates were found to be plasmid-free and the size of the most frequently detected plasmids were between 70 and 85MDa.

APPENDIX 5 Results for DT193 from humans, food and animals.

2.2. S. Paratyphi B var. Java S. Paratyphi B var. Java was among the top 10 serotypes from cases of human infections in England and Wales during 2005, therefore a small number of isolates were included in the test panel described under section 1.3. These five isolates were tested using the published primers of Lindstedt et al. (2003), Lindstedt et al. (2004), Ramisse et al. (2004) and Liu et al. (2003). These primers either failed to amplify fragments or fragment sizes showed very little variability. The exception was primer set STTR1, which targeted a VNTR locus present in a putative membrane-spanning protein. This was used by Lindstedt et al. (2003) but discarded from the final VNTR scheme (Lindstedt et al. 2004) with no given reason. A further 14 potential VNTR loci were identified, primers designed and tested against the panel (see section 1.2).

Primer sets specifically targeted to the Paratyphi B var. Java genome could not be designed as the genome sequence has still not been completed. During the timescale of this project a VNTR scheme was published for S. Enteritidis (Boxrud et al., 2007). This serotype is the most common serotype isolated from cases of human infection in England and Wales, accounting for nearly half of all infections therefore we tested this scheme against a panel of Enteritidis isolates (see section 1.4).

2.3. Other S. Typhimurium A look back at the database of third generation cephalosporin resistant salmonellae identified under project VM2136 and carrying CTX-M or AmpC enzymes identified three isolates of serotype Typhimurium. These isolates were typed by VNTR along with isolates of the same phage-type which lacked CTX-M or AmpC enzymes. Isolates produced very different VNTR profiles and there was no connection between the presence of a CTX-M or AmpC enzyme and a particular VNTR profile. Therefore the emergence of third-generation cephalosporin in these isolates has occurred in unrelated isolates.

Sixteen isolates of third generation cephalosporin resistant Typhimurium were identified during the screen of salmonellae in section 3.3 were typed by VNTR typing using the Lindstedt scheme (see Appendix 1). Isolates produced very different VNTR profiles, regardless of whether they produced a CTX-M or AmpC -lactamase, or resistance to third-generation cephalosporins was due to some other unidentified mechanism. This shows that emergence of resistance to these antimicrobials is still a somewhat random event and has occurred in unrelated strains. The factors influencing this emergence are as yet unknown. It is possible that specific clones may emerge in the future, as appears to be happening with isolates of serovar Stanley producing CIT-enzymes. Due

SID 5 (2/05) Page 13 of 21

Human (n=28) Food (n=10) All animals (n=61) Pigs (n=38)Loci Allele % Allele % Allele % Allele %STTR9 2 42.8 2 80.0 4 50.8 4 63.1STTR5 2 28.5 5 30.0 2 20.0 15 23.7STTR6 4, 5 or 8 14.3 4 40.0 7 21.6 4 23.7STTR10pl 0 50.0 0 70.0 0 40.0 0 42.1STTR3 2 71.4 3 70.0 2 77.0 2 86.8

to its increased discrimination in comparison to PFGE, VNTR typing would be an excellent method for monitoring this emergence.

3. Characterisation of drug resistance genes, integron structures and plasmids

3.1. Screening of β-lactamase plasmids using inc/rep typing methodThe PCR-based inc/rep typing scheme developed in collaboration with the Istituto Superiore di Sanità in Rome under VM2136 was applied to the collection of plasmids carrying blaCTX-M and blaCMY genes that were obtained from 29 Salmonella and 38 E. coli epidemiologically unrelated strains by conjugation or transformation during studies carried out under project VM2136. The PCR-based inc/rep typing scheme can identify 18 major plasmid incompatibility groups circulating among Enterobacteriaceae (FIA, FIB, FIC, HI1, HI2, I1-Iγ, L/M, N, P, W, T, A/C, K, B/O, X, Y, F and FIIA) using five multiplex and three simplex PCRs (Carattoli et al. 2005) (Appendix 6). Thirty-one plasmids carrying AmpC genes were analysed: 16 plasmids carried CMY-2, 11 CMY-7, two DHA-1 and one each of CMY-4 and CMY-21. Twenty-seven were from the UK and 6 from patients reporting recent travel abroad.

The data indicate a high prevalence of some replicons such as repI1, repA/C and repFII in association with relevant extended-spectrum cephalosporin resistance genes, such as the blaCMY-2 and blaCTX-M-15 genes. This in turn suggests a large diffusion of particular plasmids, prevailing in the UK but also identified in other studies in bacterial populations from other continents. These epidemic plasmids seem to prevail in different environments and might spread across different bacterial species, in humans as well in animals. PCR-based replicon typing was demonstrated to be a good method for detecting the replicons in large collections of plasmids. The origin of replication can be considered as an additional marker for the constant backbone of the plasmid. The association of replicons with specific plasmid-borne resistance genes opens the possibility to easily detect and trace the diffusion of successful plasmids as well as to detect the mobilization capability of a resistance gene among different plasmids. Further information is available in Hopkins et al. 2006 Antimicrob Agents Chemother. 50:3203-3206.

Since introduction into LEP during project VM2136, the inc/rep typing method has been used for characterisation of a number of plasmids that are associated with emerging resistances and has proved very useful in studying their epidemiology, for example previously isolated qnr plasmids and a plasmid encoding the 16S rRNA methylase armA, which confers resistance to aminoglycosides. The method will be applied to qnr, CTX-M and AmpC plasmids when they are isolated from the strains described in sections 3.2 and 3.3. Inc/rep typing data of previous qnr plasmids suggests that the incompatibility group cannot be determined by this method, therefore new primers will be ordered.

3.2. Screening of enterobacteria for plasmid-mediated quinolone resistance genesA previous study carried out under VM2136 to look for qnr genes in enterobacteria suggested that

isolates exhibiting reduced susceptibility to ciprofloxacin (MIC = 0.125 – 1.0 mg/L) but were fully susceptible to nalidixic acid (MIC <16 mg/L) may be carriers of qnr genes (Hopkins et al. 2007). Between January and April 2007 isolates received by the HPA that have this resistance phenotype or exhibit high-level resistance to ciprofloxacin (MIC > 1.0 mg/L) were screened for qnr genes. Isolates were tested for qnr genes using a multiplex PCR to detect qnrA, qnrB and qnrS (Robicsek et al. 2006). For primer sequences see Appendix 7.

No isolates were positive for qnrA. One isolate of Citrobacter freundii and two isolates of S. enterica (one Schwarzengrund, one Typhimurium DT49b) were positive for qnrB. The Typhimurium DT49b isolate from this study was an equine isolate originally from South Africa, while the Schwarzengrund was from a patient reporting recent travel abroad to an unidentified destination. qnr genes have been previously identified in an avian salmonella isolate (Kehrenberg et al. 2006) and from a young child living on a UK, therefore identification of a qnr-positive Salmonella of animal origin in this study again suggests that the occurrence of qnr genes in isolates of animal origin warrants investigation.

Thirty-seven salmonellae and two Shigella flexneri were positive for qnrS; six isolates were serotype Corvallis, 21 Typhimurium and 10 Virchow. 19/21 Typhimurium isolates were either phage type DT120 or DT193, and 7/10 Virchow isolates were phage type DT43. Twelve isolates were from patients who had not travelled abroad, therefore are assumed to be from UK-acquired infections or from consumption of imported food contaminated with a qnrS-positive isolate. This study further confirms that qnrS-positive salmonellae are present in the UK.

All Salmonella qnr-positive isolates were subtyped by PFGE; Typhimurium isolates were also typed by VNTR. Plasmid profiling was also carried out for each isolate. Twenty-two qnrS-positive S. Typhimurium were typed by VNTR. The VNTR data suggest that the DT120 and DT193 isolates may all originate from Asia and but since introduction into the UK via infected travellers or contaminated imported food have become established in the UK. For further information see Appendix 7.

3.3. Screening of enterobacteria for -lactamase resistance genesWithin the HPA Laboratory of Enteric Pathogens antimicrobial susceptibility is determined by a breakpoint

method and the final concentrations of ceftriaxone and cefotaxime used are 1mg/L. Ninety isolates from 2006 and early 2007, expressing resistance to ceftriaxone and/or cefotaxime, were available for screening to determine whether there has been an increase in prevalence of these genes since the studies carried out under VM2136.

SID 5 (2/05) Page 14 of 21

The 90 isolates consisted of five Citrobacter sp., four Escherichia coli, one Hafnia alvei, 77 S. enterica and 3 unknown enterobacterial species. Details of isolates tested are in Appendix 8.

3.3.1 Screening for CTX-M genes. Isolates were initially screened by PCR using universal primers CTX-M-F and –R previously used under VM2136. PCR-positive isolates were then further screened using a multiplex PCR to differentiate between the different groups of CTX-M gene (Woodford et al., 2006). For primer sequences see Appendix 8.

Thirteen isolates of S. enterica were positive for a CTX-M (Appendix 8). These belonged to serotypes Enteritidis (1 x phage type [PT] 4), Typhimurium (1 x definite type [DT] 136, 1 x DT193, 1 x untypable), Concord, Virchow (2 x PT25, 1 x PT37), Saint-Paul, Tel-el-Kebir and Java (3 x var. Colindale). After doing a PubMed literature search it appears this is the first report of CTX-M genes in Salmonella serotypes Concord, Saint-Paul and Tel-el-kebir. PCR and sequencing identified 11 strains carrying a group 1 CTX-M, one Typhimurium a group 2 CTX-M and one Virchow a group 9 CTX-M. The group 1 CTX-M found in the Typhimurium DT193 isolate was fully sequenced as this isolate was sent by a hospital laboratory to confirm ESBL production. A novel CTX-M sequence, CTX-M-57 was identified and has been confirmed by the Lahey Clinic who maintain a database of -lactamase genes. This isolate is further described in Hopkins et al. 2008. Int J Antimicrob Agents. 31:85-6. Further details in Appendix 8.

3.3.2 Screening for AmpC -lactamases. Isolates were screened using a multiplex PCR to detect groups MOX, CIT, DHA, ACC, EBC and FOX (Pérez-Pérez and Hanson, 2002). For primer sequences see Appendix 8.

Eighteen isolates produced an amplicon consistent with the CIT primers (detect LAT-1 to LAT-4, CMY-2 to CMY-7 and BIL-1); these were one Citrobacter freundii, one S. Enteritidis PT1, one S. Anatum, six S. Stanley, four S. Typhimurium (phage types DT193, untypable, RDNC (reacts but does not conform to a recognised pattern) and DT49b), four S. Virchow (all RDNC) and one culture submitted as a Shigella, but from which Shigella was not isolated (Appendix 8). The Typhimurium DT49b isolate was also PCR-positive for qnrB (see section 3.2).

One isolate of Hafnia alvei was positive for the ACC group; ACC plasmid-mediated genes are derived from the chromosomal ampC of Hafnia alvei (Girlich et al. 2000). Two cultures that were submitted as Salmonella isolates but failed to yield Salmonella sp. were positive for FOX genes. FOX AmpC genes have an unknown origin.

3.3.3 Molecular typing of third-generation cephalosporin-resistant Salmonella. Isolates that were found to carry a CTX-M or AmpC -lactamase were subtyped using PFGE, and for Typhimurium isolates by VNTR. Plasmid profiling was carried out using the method of Kado and Liu (1981).

The six CIT-positive Stanley isolates produced profiles that were closely related, sharing 81.3% genetic similarity. These six isolates all had similar R-types and carried a single plasmid of ~98 mDa. At least two of these isolates were associated with travel to Thailand. Isolates were received at the HPA between April 2006 and March 2007, so it appears that these CIT-positive Stanley isolates have persisted over time. A ~98 mDa plasmid was also identified in the CIT-positive S. Anatum isolate from a patient reporting travel to Thailand. Further characterisation of the CIT genes and plasmids encoding them will determine whether this plasmid is related to the ~98 mDa blaCMY-2 plasmids previously identified in serovars Heidelberg and Anatum with similar resistance patterns (Batchelor et al. 2005). Similarly the Virchow RDNC isolates clustered together with 96.6% genetic similarity.

Apart from the three Java var. Colindale group 1 CTX-M-positive isolates having identical PFGE profiles, no significant clustering was seen among CTX-M-positive isolates. Group 1 CTX-M-positive isolates also had quite different plasmid profiles, suggesting that the occurrence of group 1 CTX-Ms in UK salmonellae is not due to diffusion of one plasmid. CTX-M genes appear to be spreading among a diverse range of Salmonella serovars within the UK and throughout the world on a variety of plasmids.

3.3.4 Screening for TEM, SHV and OXA-1 -lactamases. Isolates were screened using a multiplex PCR to detect TEM, SHV and OXA-1 group -lactamases (Colom et al. 2003). For primer sequences see Appendix 8.

Some problems were experienced with faint bands of 516 bp, which corresponds to the TEM amplicon. This could be due to the presence of exogenous DNA encoding blaTEM-1 in the Taq DNA polymerase (Chiang et al. 2005). This problem was overcome in VM2136 by using primers TEM-hpa-F and –R together with Invitrogen Taq polymerase, so isolates were retested with these primers to confirm the results of the multiplex PCR. Twenty-two isolates were positive for blaTEM genes (Appendix 8); 20 of these were salmonellae and two were E. coli. In six isolates a blaTEM gene was identified in an isolate that was also carrying a blaCTX-M gene, suggesting that they could be present on the same plasmid as has been shown for other CTX-M plasmids (Boyd et al. 2004). TEM genes were also identified in four isolates that were positive for blaSHV genes. Twelve salmonellae were positive for blaSHV genes and three isolates (one Salmonella and two E. coli) were positive for a blaOXA-1 gene.

3.4 Identification of resistance genes in U288, DT193 and DT104 panel of isolates3.4.1. U288 isolates. Six isolates of U288 from the panel (see section 2.1.1) were selected for testing

with the antimicrobial resistance gene nanoarray developed under project VM2136. These were from humans

SID 5 (2/05) Page 15 of 21

(H052680159 and H052060288), food (raw pork sausage: H062040389) and animals (H050400172: porcine origin, S02647-05: pig, S0574905: turkey). Results are shown in Appendix 9. Nanoarray results suggest that in each isolate ampicillin resistance was encoded by blaTEM, chloramphenicol by a cmlA1-like gene, streptomycin by aadA1-like and aadA2-like genes, sulphonamides by sul2 and sul3, tetracyclines by tet(A) and trimethoprim by dfr12. Human isolate H050400172 had an additional aac(3’)-IVa gene, which would contribute to the gentamicin resistance expressed by this isolate. All isolates were also positive for the intI1 gene, which encodes the class 1 integrase responsible for catalysing site-specific recombination of gene cassettes into the adjacent insertion site of class 1 integrons. Primers specific for these resistance genes were then used to screen the entire collection of U288 to determine whether the same genes were present in all isolates. PCR was also used to amplify any gene cassettes inserted into the class 1 integron structure. Primers sequences are detailed in Appendix 9.

Results (Appendix 9) showed that all isolates of R-type ACSSuTTm carried the same genes as identified by nanoarray. The resistance genes detected matched the phenotypic resistance profile of each isolate. Although all isolates were positive for intI1, resistance gene cassettes were only amplified from two isolates, which were both associated with a foodborne outbreak in 2005. Presence of intI1 alone suggests that no resistance gene cassettes have yet been integrated into the class 1 integron structure and these isolates have the potential to acquire more resistance genes.

3.4.2 DT193 isolates of R-type ASSuT. Approx. one-third of DT193 have R-type ASSuT. Isolates of this R-type are usually associated with pigs. Thirty isolates of DT193 with R-type ASSuT from section 2.1.3 were screened by PCR for the resistance genes encoding this phenotype. Primer sequences are detailed in Appendix 9.

Results indicated that blaTEM encoded resistance to ampicillin, strA resistance to streptomycin and sul2 resistance to sulphonamides (Appendix 9). Tetracycline resistance was encoded by tetB in four isolates and tetA in the remainder. Class 1 integron gene cassettes were not amplified from any of the isolates. A previous study showed that in ASSuT DT193 the complete spectrum of resistance is encoded by a self-conjugative plasmid of ~80 mDa (120 kb) (Threlfall et al. 1994); the precise resistance genes were not identified. There was a plasmid of ~80 mDa within the plasmid profiles of thirteen isolates from this study. However, the remainder carry plasmids much larger or smaller than 80 mDa or no plasmids at all. This suggests that the genes may be localised on a different mass plasmid or are integrated into the chromosome.

3.4.3 DT104 isolates. Resistance to antimicrobials ampicillin, chloramphenicol, streptomycin, spectinomycin, sulphonamides and tetracyclines in DT104 is conferred by chromosomally located resistance genes. Integration into the chromosome means that these genes may persist in the absence of antimicrobial selection pressure. The genes are organised into two class 1 integron structures, both with a single gene cassette aadA2 (encodes streptomycin resistance), and blaPSE-1 (ampicillin resistance) and both conferring additional resistance to sulphonamides due to sul1. Between these integrons are genes floR and tet(G), which encode resistance to chloramphenicol/florfenicol and tetracyclines respectively.

DT104 isolates were screened using primers to amplify the resistance gene cassette(s). The majority of isolates produced amplicons of 1.0 and 1.2 kb, corresponding to the two integron structures described above. A single amplicon of 1.2 kb, encoding blaPSE-1 was detected in isolates that were resistant to ampicillin and sulphonamides, but sensitive to streptomycin. A single amplicon of 1.0 kb, encoding aadA2 was found in isolates resistant to streptomycin, but sensitive to ampicillin (Appendix 4).

APPENDIX 6 Inc/rep typing methodAPPENDIX 7 Isolates tested for qnr genesAPPENDIX 8 Isolates tested for -lactamase genesAPPENDIX 9 Identification of resistance genes in U288 and DT193 isolates

5. Database entry (HPA)Bionumerics software is used within the HPA LEP and the VLA for the recording and analysis of microbiological and epidemiological strain data. Genoprofiling data generated in this study for the panel of strains described above was integrated with other data, including antimicrobial susceptibility and PFGE profiles, to provide a fully comprehensive microbiological and epidemiological profile of the strains in the test panel. This will serve as the basis for the development of an integrated database of genoprofile data in conjunction with relevant epidemiological and molecular information.

DISCUSSIONVNTR schemes for several serovars of S. enterica have now been published. Where PFGE and VNTR typing were compared each study concluded that VNTR analysis is more discriminatory than PFGE and may be a useful tool for detection and investigation of outbreaks. For each of the Typhimurium phage types characterised by PFGE and VNTR analysis in this study, VNTR typing was consistently more discriminatory than PFGE. This was most pronounced when typing U288 and DT104 isolates; both phage types were clonal and it was difficult to discriminate between different strains on the basis of PFGE alone.

SID 5 (2/05) Page 16 of 21

Concern has been raised by researchers that VNTRs may evolve so rapidly that multiple profiles could potentially emerge during the course of an outbreak (Noller et al. 2003). The investigation into stability of VNTR profiles by typing of epidemiologically-distinct outbreaks and multiple isolates of S. Typhimurium from the same patient using PFGE and VNTR analysis demonstrates how VNTR typing can be a good method for outbreak detection. Similar conclusions were made by Torpdahl et al. (2007) when they included VNTR analysis in routine surveillance over a two-year period. Both studies encountered outbreak isolates that had different alleles at one locus, suggesting that small changes in loci are possible within the course of an outbreak and even salmonella carriage by the patient, thereby highlighting the problems of relying on a single subtyping method and reinstating the importance of combining laboratory typing data with accurate and meaningful epidemiological information. Guidelines for interpretation of Typhimurium VNTR data are lacking and application of VNTR typing to further Typhimurium outbreaks is required to develop specific guidelines and international consensus for interpretation of VNTR data. Studies to evaluate the long-term stability of each allele and to determine the mutational rate for each locus are required before VNTR typing can be proposed for long-term monitoring and phylogenetic studies of S. Typhimurium.

Interpretation of VNTR data may be easier if the rationale for assigning different length amplicons to alleles was reconsidered. The current Lindstedt (2004) scheme assigned new alleles arbitrarily as new repeat lengths were discovered so, for example allele 5 at locus STTR9 has a repeat length of 194-200 base pairs, while allele 6 has a repeat length of 149-155 base pairs. The significance of changes at loci would be easier to determine if alleles were assigned based on the number of repeat units at each locus, therefore making comparison of VNTR data a more straight-forward process.

Comparison of PFGE and VNTR analysis as typing methods did raise some technical issues with both techniques. A number of U288 and DT193 isolates were virtually untypable by PFGE due to degradation of DNA during the PFGE process leading to banding profiles that would appear smeary without defined bands. The appearance can be improved by addition of thiourea or HEPES to the running buffer, however thiourea is a suspected carcinogen and HEPES buffer must be freshly prepared before each gel run. No problems were encountered when typing these isolates by VNTR analysis. VNTR typing also had an advantage over PFGE for typing of contaminated cultures. When typing such cultures by PFGE a banding profile is usually obtained that is a mixture of bands from the original isolate and the contaminant and it can be difficult to determine which bands derive from the isolate of interest. In contrast, VNTR analysis was still able to type mixed cultures of Typhimurium if they were contaminated with another serovar or species; contamination with another Typhimurium could be inferred by multiple peaks for each locus. When using VNTR typing it was not immediately obvious why some reactions failed, i.e. no amplicons were detected. This could be due to incorrect typing of the isolate, i.e. misclassified as Typhimurium, a problem with DNA template preparation or storage, or failure of the VNTR reaction. As PFGE is not serovar-specific it was easier to identify isolates that had been misclassified as Typhimurium.

Application of the antimicrobial resistance (AMR) gene nanoarray developed under project VM02136 to the U288 panel of isolates significantly reduced the amount of work required to identify the genes responsible for the ACSSuTTm phenotype. Resistance gene profiling provided an additional level of discrimination and was able to further demonstrate the clonal nature of the U288 panel.

Can genoprofiling be used to further characterise MR S. enterica and MR S. enterica with additional resistance to third-generation cephalosporins and / or fluoroquinolone antimicrobials?

Genoprofiling was used very successfully for identification and characterisation of qnrS-positive S. enterica isolates (see section 3.2). These isolates were shown by VNTR analysis to differ from many Typhimurium isolates because of the simultaneous absence of amplicons at loci STTR6 and STTR10pl. STTR6 is located within the Gifsy-1 bacteriophage of S. Typhimurium LT2 and STTR10pl is located on the S. Typhimurium serotype-specific plasmid pSLT (Lindstedt et al. 2003; Lindstedt et al. 2004); both of which are reported to play important roles in S. Typhimurium virulence in mice, and in extra-intestinal infections in humans. However, as these are clinical isolates their virulence does not appear to be attenuated as they are still capable of causing infections in humans.

VNTR typing of -lactamase-resistant salmonellae identified profiles that were quite diverse (see section 3.3). This suggests that so far the spread of -lactamase resistance in S. Typhimurium is due to isolated strains acquiring resistance genes and not due to spread of an established clone. Third-generation cephalosporin resistance does not appear to be emerging among serotype Typhimurium isolates despite indications that it might, and most isolates expressing -lactamase activity have been restricted to less common serotypes. VNTR typing will be useful in monitoring emergence of such strains should this occur.

SID 5 (2/05) Page 17 of 21

Previous work has shown that resistance genes for other classes of antimicrobial are often co-transferred on qnr and -lactamase plasmids. The AMR nanoarray will prove useful in determining the other resistance genes in these isolates.

Is genoprofiling suitably robust for inter-laboratory comparisons?

VNTR analysis has been proposed as a technique that can be easier implemented in a new laboratory (Lindstedt et al. 2007). Use of only a string of numbers of strain identification should be an easy way of sharing results with other laboratories if the technique is standardised. Use of different sequencing machines and running conditions has been noted to lead to differences of 3-bp between the sizes of allele fragments (Pasqualotto et al. 2007). The smallest repeat unit detected in the Typhimurium VNTR scheme is 6-bp therefore this should not greatly affect inter-laboratory comparison where different sequencers are used, but is worth bearing in mind if new VNTR loci are designed.

VNTR typing within this project has been carried out using the same protocol, molecular size standard and a Beckman capillary sequencer at both institutions, which has been reported previously to give results within each laboratory that were inside the sequencer’s uncertainty range (Lindstedt et al. 2007), so our results were comparable. Inter-laboratory comparison of VNTR data occurs routinely between Denmark and Norway (Lindstedt et al. 2007), and VNTR data has been exchanged between the HPA and the VLA, and between the HPA and Denmark. The Danish VNTR database contains Typhimurium isolates with the same VNTR profile as the qnrS-positive isolated characterised in this project; as with UK isolates these are associated with tourists returning from Malaysia and Thailand. This exchange of VNTR data has prompted the Danish laboratory to screen their isolates for qnr genes. A further ring-trial needs to be arranged between laboratories in the UK, Europe and the United States to determine whether VNTR typing is truly amenable to inter-laboratory comparison.

Are the results amenable for incorporation into existing databases of strain types?

At present the HPA and VLA use Bionumerics software for maintaining databases of strain types. VNTR data has been incorporated into these databases by creating new experiment types. Data can either be entered manually or using a script developed in collaboration with the HPA Bioinformatics Unit, which enables automated upload and assignment of VNTR alleles based on output from the sequencer. Incorporation of VNTR data into Bionumerics has the added advantage of allowing full comparison of PFGE and VNTR data, which is essential before VNTR analysis is considered as a subtyping tool in its own right.

Can results be achieved in real-time?

In November 2006 a cluster of Salmonella Typhimurium DT120 in the North East of England was putatively associated with the consumption of pork. At the same time cases of illness in Denmark were associated with this Salmonella type, and a EU alert was issued to determine the type of S. Typhimurium DT120 identified. Isolates from the UK and Denmark were compared on the basis of antibiogram, PFGE and VNTR to identify the S. Typhimurium DT 120 type and results were compared electronically. Isolates from England had the resistance profile ApSSuT (ampicillin, streptomycin, sulfamethoxazole and tetracycline), VNTR profile (171-244-316-0-487) and with the distinct PFGE type (STYMXB.0083). Isolates from Denmark were resistant to Ap (ampicillin) only, had the VNTR type (171-270-324-0-490) and a PFGE type distinct from England (STYMXB.0010). It was therefore possible to confirm that the isolates from England and Denmark were not identical. These results have verified the significance of VNTR in outbreak investigations for S. Typhimurium and have demonstrated how new molecular strategies may be used to supplement existing methods such as PFGE to enable the accurate and rapid comparison of isolates from different countries.

Can the results be used to assist in the recognition of newly emerged drug-resistant strains?

VNTR typing was useful in characterising emergent qnrS-positive isolates of S. Typhimurium DT120 and DT193. VNTR profiles for these isolates were unusual in that amplicons were absent for loci STTR6 and STTR10pl.

Can the information be incorporated into existing and proposed international networks?

Within this project data was incorporated into and exchanged via the PulseNet Europe and Enternet networks. Using a standardised VNTR protocol and equipment it should be possible to set up an international database where the alleles and VNTR profiles for isolates from many laboratories can be entered and compared; similar microbial identification and typing databases are already hosted by the HPA Centre for Infections (see http://www.hpa.org.uk/cfi/bioinformatics/dbases.htm).

SID 5 (2/05) Page 18 of 21

FUTURE WORK

qnr-positive salmonellae. The data from plasmid-mediated quinolone resistance gene (qnr) study represents only basic characterisation of these isolates and further work needs to be done to characterise the plasmids that are encoding the qnr genes. A collaborative effort together with Professor Dik Mevius (Central Institute for Animal Disease Control, Lelystad, The Netherlands) and Dr Alessandra Carattoli (Istituto Superiore di Sanità, Rome) will enable comparison of our qnrS-positive S. Corvallis and those being identified in The Netherlands. Further funding to enable this work to be carried out will be sought initially from the HPA Pump Priming & Small Initiatives Fund.

One of the qnrS1-plasmids from our previous study was sent to Stefan Schwarz (Institut fur Tierzucht, Bundesforschungsanstalt fur Landwirtschaft (FAL), Germany) for sequencing. Preliminary data have shown the plasmid to be only a bit larger than 10 kb and consists of a qnrS region similar to that in S. Infantis (except that the Tn3 is missing) and a mobABC gene region similar to that of plasmid RSF1010 and other plasmids. There were no other reading frames that showed significant homology to any proteins deposited in the databases. When the sequence is complete this will help to determine the incompatibility group of this, and possibly other qnrS-plasmids, as the majority of them are untypable using the current inc/rep PCR scheme. A paper will be co-authored describing the plasmid and comparing it to other qnrS plasmid partial sequences.VNTR scheme for Salmonella EnteritidisPreliminary testing of a recently described VNTR scheme for S. Enteritidis confirmed findings of the author and yielded additional novel loci worthy of study. Further testing and development of this scheme has the potential to enhance the degree of strain discrimination possible for this important serovar.Ring-trial. For S. Typhimurium the results of this investigation have demonstrated that VNTR fingerprinting can be standardised within the UK between the HPA and the VLA. Similarly the results of a real-time investigation involving a putative outbreak of MR S. Typhimurium DT 120 has demonstrated that the method can be used for international investigations. Standardisation of the method internationally is now imperative and to take this forward a ring-trial involving several countries within the EU has been proposed. Included in this would be the standardisation of allele naming. As currently the scheme devised by Lindstedt et al. (2004) is used for the naming of alleles. However, provision needs to be made for the addition of new allele types as they are identified. Also the current system does not sequentially number allele based on size. Therefore alleles which differ by one or two repeats may have very different allele numbers, making comparison difficult. Standardisation and a ring trial will be discussed in detail at the forthcoming Annual meeting of the EU-funded Enternet project, to be held in July 2007. Already five countries (England and Wales, Scotland, Denmark, Norway, Sweden, the Netherlands and Germany have expressed interest in participating in such a trial. If successful the ring trial will also be extended to S. Enteritidis.

OUTPUTS

Milestone completed within this section:

Hold progress update meeting with all project staff and with Defra. (Minutes of meeting Appendix 10)

Hold final progress meeting with collaborators and Defra officials. (Minutes of meeting Appendix 10).

Submit final account of project activities and findings to Defra

Project outputs Minutes of meetings

Shared database between HPA and VLA.

Liebana, E., Best, E.L., Lindstedt, B-A., Clifton-Hadley, F.A., Threlfall, J. and Cook, A. (2006). Application of variable-number of tandem-repeats (VNTR) analysis to Salmonella Typhimurium: a longitudinal investigation on pig farms, and assessment of its value for food attribution of human cases. International Symposium Salmonella and Salmonellosis Proceedings, St. Malo, France 10-12 May 2006.

Best, E. Variable-number Tandem Repeat for Salmonella Typhimurium. A Comparison of isolates obtained from poultry, pigs and cases of human gastroenteritis at Health Protection 2006. University of Warwick, 13th September 2006.

Best, E. Joint activities between the Health Protection Agency and the Veterinary Laboratories Agency in Salmonella Research at VLA Conference, University of Cambridge, 20th September 2006.

Best, E. Joint activities between the Health Protection Agency and the Veterinary Laboratories Agency in Salmonella Research at HPA Division of Gastrointestinal Infections Post Warwick Seminar, 25th September 2006.

SID 5 (2/05) Page 19 of 21

Best, E. Variable-number Tandem Repeat for Salmonella Typhimurium at Gastrointestinal Disease and Food, Water and Environmental Specialists network meeting at HPA Centre for Infections, 4th December 2006.

Best, E. Variable Number Tandem Repeats for analysis of Salmonella Typhimurium from animals and humans in Microbiologist 2007 (Society for Applied Microbiology quarterly publication).

Best E, Hampton M, Ethelberg S, Davies R., Threlfall J. Porcine associated Salmonella Typhimuirum DT120: use of PFGE and MLVA in a putative outbreak investigation. ‘Safepork’ 7th

International Symposium on the Epidemiology and Control of Foodborne Pathogens in Pork, Verona, Italy, 9-11 May 2007.

Hopkins, K.L. ‘Plasmid-mediated quinolone resistance (qnr) genes in enterobacteria isolated in England and Wales’. Oral presentation given at MEDVETNET WP21 1st Plenary meeting. Held at HPA Centre for Infections, 18th - 19th April 2007.

Hopkins, K. L., E. Liebana, L. Villa, M. Batchelor, E.J. Threlfall and A. Carattoli. 2006. Replicon typing of plasmids carrying CTX-M or CMY β-lactamases circulating among Salmonella and Escherichia coli isolates. Antimicrob. Agents Chemother. 50:3203-6.

Hopkins, K.L., C. Maguire, E. Best, E. Liebana and E.J. Threlfall. 2007. Investigation into the Stability of Multiple-Locus Variable-Number Tandem-Repeats in Salmonella enterica serovar Typhimurium. J Clin Microbiol. 45:3058-61.

Hopkins, K.L., M. Day, E.J. Threlfall. Increasing incidence of plasmid-mediated quinolone resistance in Salmonella enterica in England and Wales. Emerg. Infect. Dis. 2008 Feb; [Epub ahead of print].

Hopkins, K.L., E. Karisik, J.K. Wardle and E.J. Threlfall. 2008. Identification of novel plasmid-mediated extended-spectrum -lactamase CTX-M-57 in Salmonella enterica serovar Typhimurium. Int J Antimicrob Agents. 31:85-6.

Lucarelli, C., I. Luzzi, L. Villa, A.M. Dionisi, K.L. Hopkins, E. De Pinna, E.J. Threlfall. Salmonella enterica serotype Typhimurium phage type U311: a multidrug resistant clone emerging in Italy and England. Submitted to MEDVETNET 3rd Annual Scientific meeting, Lucca Italy 27th to 30th June 2007.

Hopkins, K.L., C. Maguire, E. Best, E. Liebana, E.J. Threlfall. Investigation into the Stability of Multiple-Locus Variable-Number Tandem-Repeats in Salmonella enterica serovar Typhimurium. Submitted to the Health Protection 2007, Warwick, 17th to 19th September, 2007.

Kehrenberg, C., Hopkins, K.L., Threlfall, E.J. and Schwarz, S. 2007. Complete nucleotide sequence of a small qnrS1-plasmid from Salmonella enterica subsp. enterica Typhimurium DT193. J Antimicro Chemother 60:903-5.

In preparation Hopkins, K.L., M. Day, C. Maguire, E.J. Threlfall. Full-length paper describing the

characterisation of qnr-positive isolates of enterobacteria. Best E., Hampton M., Ethelberg S., Liebana E. Clifton-Hadley F.A., Davies R., Threlfall J.

Porcine associated Salmonella Typhimurium DT120: use of PFGE and MLVA in a putative outbreak investigation. In preparation.

References to published material9. This section should be used to record links (hypertext links where possible) or references to other

published material generated by, or relating to this project.Batchelor, M., K. L. Hopkins, E. J. Threlfall, F. A. Clifton-Hadley, A. D. Stallwood, R. H. Davies, and E.

Liebana. 2005. Characterization of AmpC-mediated resistance in clinical Salmonella isolates recovered from humans during the period 1992 to 2003 in England and Wales. J.Clin.Microbiol. 43:2261-2265.

Boxrud, D., K. Pederson-Gulrud, J. Wotton, C. Medus, E. Lyszkowicz, J. Besser, and J. M. Bartkus. 2007. Comparison of multiple-locus variable-number tandem repeat analysis, pulsed-field gel electrophoresis, and phage typing for subtype analysis of Salmonella enterica serotype Enteritidis. J.Clin.Microbiol. 45:536-543.

Boyd, D. A., S. Tyler, S. Christianson, A. McGeer, M. P. Muller, B. M. Willey, E. Bryce, M. Gardam, P. Nordmann, and M. R. Mulvey. 2004. Complete nucleotide sequence of a 92-kilobase plasmid harboring the CTX-M-15 extended-spectrum -lactamase involved in an outbreak in long-term-care facilities in Toronto, Canada. Antimicrob.Agents Chemother. 48:3758-3764.

Carattoli, A., A. Bertini, L. Villa, V. Falbo, K. L. Hopkins, and E. J. Threlfall. 2005. Identification of plasmids by PCR-based replicon typing. J.Microbiol.Methods 63:219-228.

Chiang, C. S., C. P. Liu, L. C. Weng, N. Y. Wang, and G. J. Liaw. 2005. Presence of beta-lactamase gene TEM-1 DNA sequence in commercial Taq DNA polymerase. J.Clin.Microbiol. 43:530-531.

SID 5 (2/05) Page 20 of 21

Colom, K., J. Pérez, R. Alonso, A. Fernández-Aranguiz, E. Lariño, and R. Cisterna. 2003. Simple and reliable multiplex PCR assay for detection of blaTEM, blaSHV and blaOXA-1 genes in Enterobacteriaceae. FEMS Microbiol.Lett. 223:147-151.

Girlich, D., T. Naas, S. Bellais, L. Poirel, A. Karim, and P. Nordmann. 2000. Biochemical-genetic characterization and regulation of expression of an ACC-1-like chromosome-borne cephalosporinase from Hafnia alvei. Antimicrob.Agents Chemother. 44:1470-1478.

Hopkins, K. L., L. Wootton, M. Day, and E. J. Threlfall. 2007. Plasmid-mediated quinolone resistance determinant qnrS1 found in Salmonella enterica strains isolated in the UK. J Antimicrob.Chemother. doi:10.1093/jac/dkm081.

Drahovská, H., E. Mikasová, T. Szemes, A. Ficek, M. Sásik, V. Majtán, and J. Turna. 2007. Variability in occurrence of multiple prophage genes in Salmonella Typhimurium strains isolated in Slovak Republic. FEMS Microbiol.Lett. 270:237-244.

Kado, C. I. and S. T. Liu. 1981. Rapid procedure for detection and isolation of large and small plasmids. J.Bacteriol. 145:1365-1373.

Threlfall, E. J., Hampton, M. D., Schofield, S. L., Ward, L. R., Frost, J. A. and Rowe, B. 1996. Epidemiological application of differentiating multiresistant Salmonella typhimurium DT104 by plasmid profile. Commun.Dis.Rep.CDR Rev. 6:R155-159.

Kehrenberg, C., S. Friederichs, A. de Jong, G. Brenner Michael, and S. Schwarz. 2006. Identification of plasmid-borne quinolone resistance gene qnrS in Salmonella enterica serovar Infantis. J.Antimicrob.Chemother. 58:18-22.

Lindstedt, B. A., E. Heir, E. Gjernes, and G. Kapperud. 2003. DNA fingerprinting of Salmonella enterica subsp. enterica serovar Typhimurium with emphasis on phage type DT104 based on variable number of tandem repeat loci. J.Clin.Microbiol. 41:1469-1479.

Lindstedt, B. A., M. Torpdahl, E. M. Nielsen, T. Vardund, L. Aas, and G. Kapperud. 2007. Harmonization of the multiple-locus variable-number tandem repeat analysis method between Denmark and Norway for typing Salmonella Typhimurium isolates and closer examination of the VNTR loci. J.Appl.Microbiol. 102:728-735.

Lindstedt, B. A., T. Vardund, L. Aas, and G. Kapperud. 2004. Multiple-locus variable-number tandem-repeats analysis of Salmonella enterica subsp. enterica serovar Typhimurium using PCR multiplexing and multicolor capillary electrophoresis. J.Microbiol.Methods 59:163-172.

Liu, Y., M. A. Lee, E. E. Ooi, Y. Mavis, A. L. Tan, and H. H. Quek. 2003. Molecular typing of Salmonella enterica serovar typhi isolates from various countries in Asia by a multiplex PCR assay on variable-number tandem repeats. J.Clin.Microbiol. 41:4388-4394.

Noller, A. C., M. C. McEllistrem, A. G. Pacheco, D. J. Boxrud, and L. H. Harrison. 2003. Multilocus variable-number tandem repeat analysis distinguishes outbreak and sporadic Escherichia coli O157:H7 isolates. J.Clin.Microbiol. 41:5389-5397.

Pasqualotto, A. C., D. W. Denning, and M. J. Anderson. 2007. A cautionary tale: the lack of consistency in allele sizes between two laboratories for a published Multi-Locus Microsatellite Typing (MLMT) system. J.Clin.Microbiol. 45:522-528.

Pérez-Pérez, F. J. and N. D. Hanson. 2002. Detection of plasmid-mediated AmpC -lactamase genes in clinical isolates by using multiplex PCR. J Clin.Microbiol. 40:2153-2162.

Peters, T. M., C. Maguire, E. J. Threlfall, I. S. Fisher, N. Gill, A. J. Gatto, and (on behalf of the Salm-gene project participants). 2003. The Salm-gene project - a European collaboration for DNA fingerprinting for food-related salmonellosis. Euro.Surveill 8:46-50.

Ramisse, V., P. Houssu, E. Hernandez, F. Denoeud, V. Hilaire, O. Lisanti, F. Ramisse, J. D. Cavallo, and G. Vergnaud. 2004. Variable number of tandem repeats in Salmonella enterica subsp. enterica for typing purposes. J.Clin.Microbiol. 42:5722-5730.

Swaminathan, B., T. J. Barrett, S. B. Hunter, and R. V. Tauxe. 2001. PulseNet: the molecular subtyping network for foodborne bacterial disease surveillance, United States. Emerg.Infect.Dis. 7:382-389.

Threlfall, E. J., M. Hampton, H. Chart, and B. Rowe. 1994. Identification of a conjugative plasmid carrying antibiotic resistance and salmonella plasmid virulence (spv) genes in epidemic strains of Salmonella typhimurium phage type 193. Lett.Appl.Microbiol. 18:82-85.

Torpdahl, M., G. Sørensen, B. A. Lindstedt, and E. M. Nielsen. 2007. Tandem repeat analysis for surveillance of human Salmonella Typhimurium infections. Emerg.Infect.Dis. 13:388-395.

Woodford, N., E. J. Fagan, and M. J. Ellington. 2006. Multiplex PCR for rapid detection of genes encoding CTX-M extended-spectrum -lactamases. J.Antimicrob.Chemother. 57:154-155.

SID 5 (2/05) Page 21 of 21