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 26

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 OD2013

2. Project title

The potential for veterinary antibiotic use to cause glycopeptide resistance in human Staphylococcus aureus.

3. Contractororganisation(s)

Dr Jodi LindsayDepartment of Cellular & Molecular Medicine

St George's, University of LondonCranmer TceLondon SW17 0RE          

54. Total Defra project costs £ 124,751

5. Project: start date................ 01 April 2004

end date................. 31 March 2006

SID 5 (2/05) Page 2 of 26

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

We request a delay in publication of the full report in order to submit manuscripts to peer-reviewed journals who have policies on prior disclosure. We are happy for the executive summary to be published immediately.

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.

Staphylococcus aureus is a major human pathogen that is becoming increasingly resistant to antibiotics, and includes methicillin-resistant S. aureus (MRSA). We are now dependent on one class of antibiotics to treat all suspected MRSA infections in hospitals - the glycopeptides, which include vancomycin. At the same time, S. aureus is a pathogen of animals, particularly dairy cows where it causes mastitis and leads to significant economic losses. Until 1997, the antibiotic avoparcin, which is a glycopeptide, was used in the agricultural setting. Although no vancomycin-resistant S. aureus (VRSA) were identified, avoparcin use probably contributed to the emergence of vancomycin-resistant enterococci (VRE). VRE are now widespread in hospitals and still reported in the agricultural setting. VRE are not as pathogenic as S. aureus, but they are a serious risk for the horizontal transfer of the vancomycin resistance gene, vanA, to S. aureus. Indeed, four cases of vanA transfer from VRE to MRSA have now been reported in US hospitals.

This project was designed to 1) test how easily vanA may spread from VRE to animal S. aureus, 2) characterise animal S. aureus isolates and compare their relatedness to human S. aureus using the most comprehensive technology available, comparative genomics using microarrays, and 3) investigate barriers to the spread of genes between animal and human S. aureus. We could then make better estimates about the risks of VRSA developing, if this was likely in the agricultural setting, and what strategies might reduce this risk.

Comparison of animal S. aureus versus human S. aureusWe compared 56 UK S. aureus causing infection in cows, horses, goats, sheep and a camel to over 200 human isolates from UK healthy carriers, community and hospital acquired-infection and including MRSA. We used a seven-strain S. aureus microarray which allowed us to investigate the presence or absence of 3623 genes in each isolate. We had previously shown that human isolates can be clustered into about ten dominant lineages and a few minor lineages. These lineages carry different combinations of surface proteins that bind to human tissue, and presumably affect each lineage's ability to interact with their specific mammalian hosts. In this study we found the bovine (cow) isolates were predominantly from three major lineages that were different to the human lineages, and are called ST151, ST771 and ST130. In contrast, isolates from horses were from either bovine lineages or human lineages, perhaps due to their close contact with humans. However, the various lineages did not cluster as an "animal branch" and a "human branch", and our attempts to identify host-specific markers were not conclusive. We therefore suggest that animal and human isolates are perhaps more closely related than originally predicted. We also noted that animal isolates from other countries are generally from different lineages to those in our

SID 5 (2/05) Page 3 of 26

UK study. It also became clear that the identification of S. aureus in the veterinary laboratory is more complex than in human laboratories, and there is a need for improved methods.

Model of resistance gene transfer from VRE to animal S. aureusIn the laboratory, it is very difficult to move foreign DNA such as resistance genes into S. aureus using standard methods. However, the model of DNA transfer between enterococci and S. aureus was surprisingly efficient. It was so effective, that we plan to develop laboratory tools based on this model. So why has vanA moved from enterococci to S. aureus so infrequently in nature? We think the answer might be because our model uses a transposon that is more mobile than those currently carrying vanA in nature, and we need to investigate this further.A second very important and surprising result was that about one third of the bovine isolates accepted DNA from enterococci at an extremely high rate. We call these isolates "hyper-recipient". They were found to belong to one lineage, ST151. We have recently shown that Sau1 restriction modification (R-M) is the major mechanism in S. aureus that controls foreign DNA acquisition, and we hypothesised that Sau1 might be compromised in this lineage. The ST151 isolates were found to have a mutation in each of their sau1hsdS1 and sau1hsdS2 genes, encoding the specificity subunits of the Sau1 R-M system. R-M is a mechanism used by many bacteria to protect themselves from foreign DNA. They produce a restriction enzyme that recognises specific DNA sequences and cuts the DNA. They also produce a modification enzyme that recognises the same DNA sequences and modifes them which protects them from the restriction enzyme, thus protecting the cell's own DNA. If both specificity units are mutated, the bacteria will be unable to modify or digest DNA. The data from this study shows all ST151 isolates have both mutations, and this is the likely reason for their "hyper-recipient" phenotype. No other naturally occurring S. aureus have been identified with this ability.

Gene transfer from animal S. aureus to human S. aureusThe most efficient transfer mechanism between S. aureus is via transduction using bacteriophage. We have shown that Sau1 R-M controls this transfer, and that variation in the specificity subunits is correlated with lineage. Thus isolates from the same lineage should exchange DNA at higher frequency than between lineages. We have showns the animal isolates of different lineages to human isolates have unique Sau1 specificity and this would block transfer. However, those isolates from animals that are similar to human isolates are the greatest risk for DNA exchange. Bacteriophage models of gene transfer confirmed that transfer can occur but at low frequency. They also showed Sau1 R-M plays a role in preventing bacteriophage to deliver DNA to S. aureus. However, a second mechanism or mechansims were identified that could also effectively block gene transfer. These mechanisms need to be investigated further.What is the evidence that animal and human S. aureus are actually exchanging DNA? Resistance patterns of the animal S. aureus suggest they are mostly sensitive to antibiotics, while those equine isolates from human lineages tended to be multi-resistant. Microarray data also suggested some exchange of mobile genetic elements, but at relatively low frequency. Therefore, our data suggests that genetic exchange is possible, but frequency is low due to multiple genetic barriers. It is also possible that decreasing use of antibiotics in the agricultural sector is reducing opportunities for resistance gene spread. However, antibiotic use in horses where both human and animal S. aureus isolates are found may be the most likely setting for resistance gene transfer to occur. At the present time, the high risk ST151 isolates have not been found on horses.

Main Conclusions1. In a small study of UK animal isolates, three distinct lineages were predominant. These lineages are different to animal isolates from other countries.2. Bovine isolates are different to human isolates, while equine isolates can be from either bovine or human lineages. Human lineages are generally more multi-resistant.3. One third of bovine isolates belong to the lineage ST151 in this small UK study, and they have mutations in the major R-M pathway. This means they are "hyper-recipients" and capable of accepting resistance genes (such as vanA) from enterococci at high frequency. No other naturally occuring S. aureus "hyper-recipients" have been described.4. R-M effectively blocks the transfer of resistance genes between certain bovine S. aureus lineages and human S. aureus. Further research is required to characterise a second pathway that blocks transfer between S. aureus isolates. 5. There is evidence that animal S. aureus can exchange MGE with human S. aureus, but also evidence that this occurs at low frequency. This is likely to be due to robust genetic barriers, but may also be due to limited interaction between these strain populations, and limited selective pressure. Main implications1. One third of bovine S. aureus in this small study in the UK belonged to a hyper-recipient lineage of S. aureus that is susceptible to resistance gene transfer from enterococci. No other strains of S. aureus have this ability.

SID 5 (2/05) Page 4 of 26

This unexpected result suggests the agricultural ban of avoparcin use in 1997 was justified, as it would be a risk for the transfer of vancomycin resistance genes from VRE to animal S. aureus. It suggests some sentinel screening for VRSA in bovine strains is warranted. This is particularly relevant in areas where VRE is still found. 2. The transfer of resistance genes between different S. aureus lineages is occuring at low frequency and this is likely due to effective genetic barriers. There may also be limited opportunities for mixing of isolates, and limited selective pressure due to decreasing use of growth promotors and antibiotics. The most likely setting for gene transfer may be with horses exposed to animal and human S. aureus and antibiotics.3. Veterinary diagnostic laboratories need assistance with identifying S. aureus.

Future research is necessary to build a more representative model of vanA transfer from enterococci to S. aureus, to investigate the second genetic barrier preventing S. aureus to S. aureus spread, to identify epidemiological barriers to the spread of animal S. aureus, and to improve diagnosis of S. aureus in veterinary laboratories.

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).

Background

Original abstract and justification for research project.

The horizontal transfer of vancomycin resistance genes from enterococci to Staphylococcus aureus would be a human medical catastrophe, as we are dependent on this antibiotic to treat hundreds of thousands of human S. aureus (including MRSA) infections in the UK every year. For this reason, avoparcin, a glycopeptide antibiotic related to vancomycin, was banned in the EU. This decision was based on the assumption that horizontal transfer of glycopeptide resistance would be accelerated by the use of this antibiiotic. These new super-resistant S. aureus would then spread to humans and to hospitals. However, there are several confounding issues with this hypothesis. First, S. aureus, unlike most bacteria, actively blocks uptake of foreign genes from other bacteria. Secondly, animal strains of S. aureus are thought to be quite different to human strains, and it is unlikely that they will colonise, and therefore infect, humans in the same way that human strains do. Thirdly, we do not know if resistance can easily be spread from animal S. aureus to human S. aureus strains. This project will address these issues, providing baseline information that can be used to assess the true risks of antibiotic resistance spreading to clinical S. aureus, and to identify the "red-flag" strains indicating resistance is spreading.

Scientific Context

SID 5 (2/05) Page 5 of 26

S. aureus is carried in the nose by 30-70% of the human population, and is a significant cause of minor skin infection. In hospitals, however, S. aureus is a serious threat and is the most common cause of hospital acquired infection. For example, in England, it accounts for 18 000 cases of bacteremia per year [1] with a 20% mortality rate [2], and also infects 1.6% of all surgical interventions [3]. Hospital S. aureus are frequently resistant to antibiotics and infections can be difficult to treat. Presently, 40% of all UK hospital S. aureus are resistant to methicillin (MRSA), the antibiotic of choice. Hospital MRSA comprise only a few clones that are highly successful, called epidemic MRSA [4]. Infections causes by MRSA are treated with vancomycin, a glycopeptide antibiotic and the only remaining reliable class of antibiotics for treating these infections. Despite widespread use of vancomycin and the relatively high rates of vancomycin-resistant enterococci (VRE) in hospitals, only four strains of high level vancomycin resistant S. aureus (VRSA) have been identified, all in the US, and all successfully contained [5,6]. However, the potential impact of epidemic S. aureus strains resistant to methicillin and vancomycin would be catastrophic.

S. aureus is also an important cause of dairy cow mastitis which is extremely difficult to eradicate, as well as causing infections in a range of other livestock. Because of the risk of vancomycin resistance arising in animal strains of S. aureus and enterococci, avoparcin, a glycopeptide antibiotic, was banned in the EU for agricultural use in 1997. Although vancomycin resistance has not been seen in bovine isolates of S. aureus, resistance in enterococci is relatively common and thought to be due to avoparcin use [7]. Rates of VRE have been decreasing in animals since the ban [7], but there is still a significant fear that resistance genes could spread from animal VRE to S. aureus, leading to unprecedented rates of morbidity and mortality in hospitals.

This fear is based on a number of assumptions. Firstly, that vancomycin resistance will spread from enterococci to S. aureus. Secondly, that animal strains of VRSA will colonise and infect humans, and thirdly that vancomycin resistance will spread from animal strains of S. aureus to human strains including epidemic MRSA. At the time of the grant submission (2003), the evidence for these assumptions was flawed, and accurate data is required for making predictions about the likelihood of antibiotic resistance transfer.

1. S. aureus takes up DNA from other bacteria poorly.When antibiotics were introduced into animal feed in the 1940s, Gram negative bacteria quickly developed resistance, leading to the first antibiotic bans in the 1960s-70s, but antibiotic resistance in Gram positive bacteria was not seen until the 1990s [7]. Even now, VRSA have not been seen in the agricultural setting. In hospitals, incredible amounts of glycopeptides are used, often in patients co-infected with VRE and S. aureus, yet the horizontal transfer of resistance has been detected only four times [5,6].

Noble et al. [8] demonstrated the transfer of vancomycin resistance from enterococci into S. aureus in the laboratory. For ethical reasons his experiments have not been repeated. Unfortunately, we know nothing about the mechanisms involved and the VRSA strains have been destroyed. However, Clewell et al. established the horizontal transfer mechanism that was predicted to be responsible for transfer of vanA (a gene responsible for high level vancomycin resistance) from enterococci to S. aureus [9,10], and this was subsequently shown to be the mechanism in the first case of VRSA in 2002 [11], and the likely mechanism for the remaining VRSA cases. S. aureus actually signal to enterococci carrying pheromone susceptible conjugative plasmids such as pAM373. Subsequently, enterococci aggregate with the S. aureus and the plasmid is transferred. Although the plasmid cannot replicate in S. aureus, if it carries a transposon, such as Tn918 carrying tetR, the transposon will jump to the S. aureus chromosome before the plasmid is lost. Naturally occuring plasmids related to pAM373 and carrying vanA on transposons have been identified in hospital strains of E.

SID 5 (2/05) Page 6 of 26

faecalis [12]. It is very important to note that prior to this project, transfer had only been tested in laboratory generated S. aureus strains that are "restriction deficient", and had not been attempted in typical wild type clinical strains [9].

It is not known how common restriction deficient strains are in nature. Restriction-modification (R-M) is a system a bacterium employs to protect itself from invading foreign DNA [13]. Restriction enzymes recognise specific sequences on the incoming DNA and digest them; protection of its own DNA is via a modification of these specific sequences. During the course of this grant application, our laboratory was also funded by BBSRC to investigate barriers to horizontal transfer in S. aureus [14]. This project discovered that the standard "restriction deficient" laboratory strain RN4220 was in fact defective in a R-M system that we call Sau1. This R-M consists of a restriction gene (sau1hsdR), two modification genes (sau1hsdM1 and sau1hsdM2) and two specificity genes (sau1hsdS1 and sau1hsdS2). RN4220 has a mutation in sau1hsdR, meaning it can accept foreign DNA at high frequency because it is not recognised as foreign or digested. RN4220 can accept DNA via conjugation from enterococci, by electroporation from E. coli, by transduction from other S. aureus, and is more susceptible to S. aureus bacteriophage. Since Sau1 is found in all strains of S. aureus, we predicted it would be the major barrier controlling foreign DNA transfer into S. aureus [14]. However, the distribution of R-M deficient strains in clinical and animal S. aureus strains was unknown.

A further discovery during the course of this project is the distribution of sau1hsdS variants. These genes control the specific sequence targetted by the Sau1 system. Thus, two isolates with identical sau1hsdS genes are predicted to exchange DNA more easily than two isolated with different sau1hsdS genes. We have shown there is significant variation in sau1hsdS genes in S. aureus, and this variation correlates with the ten dominant lineages of S. aureus [14]. The dominant lineages have each evolved from a common ancestor, but each lineage has a unique combination of hundreds of variant genes, called core variable (CV) genes [15]. CV genes are predominantly surface expressed or control expression of surface structures. Therefore, we now suggest that antibiotic resistance genes acquired by an isolate of S. aureus will preferentially transfer to S. aureus isolates of the same lineage, rather than S. aureus of other lineages. However, proof of this phenomena was still lacking.

2. Animal and human strains of S. aureus are different.Many studies have shown that S. aureus carried by humans including those working with dairy cows are different to the strains causing mastitis [16]. Differences have been measured using pulse field gel electrophoresis and other methods show that patterns between strains are distinct. However, it was not clear exactly how strains were different.

The recent sequencing of seven human strains of S. aureus has allowed us to investigate how strains differ [17]. We have designed and built a seven strain S. aureus microarray in collaboration with the BG@S group at St George's, University of London (www.bugs.sgul.ac.uk) [18]. Using the microarray to analyse a large population of human community acquired invasive and nasal carriage isolates allowed us to identify the S. aureus core genome common to all strains. It also allowed us to discover the CV genes that have evolved independently in the ten dominant lineages, the majority of which are surface expressed or regulators of surface structures and interact with host [15]. Here we had an opportunity to investigate the similarities and differences between bovine and human isolates of S. aureus using the most powerful technology available. This could potentially identify specific surface receptors found in bovine isolates that may interact specifically with bovine hosts.

The microarray also allows the third component of the S. aureus genome to be visualised - the mobile genetic elements (MGE), including plasmids, transposons, staphylococcal cassette chromosomes (SCC) and other elements that encode antibiotic resistance genes. The

SID 5 (2/05) Page 7 of 26

horizontal transfer of MGE into and out of S. aureus is suspected to occur at high frequency, including during the course of infection [19]. However, this is likely to be controlled by R-M. Using the microarray we had an opportunity to investigate sau1hsdS gene variation in bovine and human strains to create models of how DNA exchange occurs between isolates, and how these strains evolve. It also allowed us to visualise if bovine strains carried typical human S. aureus MGE allowing frequent transfer between these isolates.

3. Can animal S. aureus genes spread to human S. aureus strains?In 1980, Lacey produced convincing evidence that it was exceedingly difficult to transfer DNA from animal strains of S. aureus to human strains of S. aureus [20]. There have been few subsequent studies, and the mechanisms controlling this were unknown. Similarly, methicillin resistance was first introduced into S. aureus strains that appeared to not be as pathogenic or epidemic, and it took 30 years for methicillin resistance to transfer into the epidemic backgrounds that cause significant rates of hospital acquired infection [21]. There are still many S. aureus backgrounds that have not acquired methicillin resistance genes. Again, it seems likely that restriction plays a role in this lack of transfer.

We suspected that S. aureus can be classed into subgroups [14]. Within each subgroup, genes could be horizontally transferred. Between these subgroups, genes would be transferred at much lower frequency. The implication is that vanA in bovine S. aureus background may not be particularly pathogenic for humans or carried by humans, and transfer of vanA to human S. aureus, particularly epidemic MRSA may not be straight-forward.

In summary, we proposed there were likely to be several steps necessary for the use of agricultural glycopeptides to lead to vancomycin resistance in human S. aureus.The steps are:a) transfer of vancomycin resistance genes from enterococci to animal S. aureusb) selection of animal VRSA by agricultural antibiotic use leading to increased incidence and disseminationc) transfer of animal VRSA to humansd) horizontal gene transfer of vanA from animal VRSA to human epidemic MRSA.If this hypothesis is true, the detection and containment of these intermediate strains can be addressed.

Scientific Objectives

1. Establish the frequency of transfer of antibiotic resistance markers from enterococci to animal strains of S. aureus. This includes the screening for restriction deficient strains, and preliminary investigation of restriction mechansims.

2. Use the seven-strain S. aureus microarray to establish molecular differences between animal and human S. aureus with emphasis on whether they share accessory genes, which genes they may carry that could be involved in restriction of gene transfer, and which genes might be associated with bovine pathogenesis.

3. Establish frequency of transfer of antibiotic resistance markers from animal S. aureus to human S. aureus, with preliminary investigation into the nature of transfer barriers.

Achievement of Scientific Objectives

1. Frequency of transfer of antibiotic resistance markers from enterococci to 50 strains of animal S. aureus and another 20 strains of human S. aureus has been completed. All strains accepted markers at a suprisingly high rate. Screening for restriction deficient strains

SID 5 (2/05) Page 8 of 26

successfully identified a group of animal strains (lineage ST151) with a particularly high rate of acceptace. These isolates were shown to have mutations in R-M specificity genes.

2. The 50 isolates of animal S. aureus were microarrayed, and compared to over 200 human isolates of S. aureus from separate studies. Most of the bovine isolates belonged to three dominant and unique lineages, compared to the ten human lineages already described. Lineages differ predominantly in surface proteins that mediate interaction with their mammalian host. Equine isolates were from bovine or human lineages. The animal lineages were suprisingly similar to human lineages, and there were only a very limited number of "host specific" genes. Known accessory genes (now referred to as being those carried on mobile genetic elements (MGE)) were found in animal strains, but were less common than in human strains. This suggests some genetic exchange with human isolates occurs but only at low level. Animal S. aureus were shown to carry R-M genes, and unique R-M specificity variants correlating with lineage.

3. The frequency of transfer of antibiotic resistance markers between S. aureus of animal lineages and human lineages via transduction was low. We believe this is due to effective R-M blocking transfer between isolates of different lineages. ST151 isolates do not preferentially donate to other lineages. We found evidence of a second pathway that also effectively blocks transduction.

SUMMARY

All the objectives were achieved. Animal S. aureus belong to a few discrete S. aureus lineages that are different to human lineages. Enterococci can donate resistance genes (potentially vancomycin resistance) to most S. aureus, but one animal lineage is particularly susceptible due to mutations in R-M. Transfer between S. aureus of different lineages is efficiently blocked by R-M. Therefore, glycopeptide use in animals may promote the transfer of vanA from enterococci to animal S. aureus of lineage ST151, but the transfer from ST151 to human lineages will be delayed. Some screening for animal vancomycin-resistant S. aureus of lineage ST151 would be prudent, and a continued ban on avoparcin use is highly recommended.

Methods

Bacterial strains and growth conditions. Strains of S. aureus are listed in Table 1 (Appendix). 43 were kindly collected by Prof. David Lloyd and colleagues at the Royal Veterinary College, South Mimms. They were from a variety of large animals, but predominantly cows (mostly mastitis) and horses. During the course of the study, we realised that many farmers who suspect their dairy cows have mastitis are unlikely to want to pay for laboratory diagnosis. Therefore, we were not able to collect as many S. aureus mastitis isolates as we had hoped. A further 20 isolates of bovine mastitis S. aureus were kindly donated by CJ Teale, Veterinary Laboratories Agency. During the course of this project, a S. aureus isolated from bovine mastitis in Ireland was sequenced by Herron-Olsen et al. and deposited in GenBank, accession number AJ938182. This strain, RF122, was kindly sent to us by Dr Ross Fitzgerald, University of Edinburgh.

A reference strain collection included 8325-4, a standard laboratory strain, RN4220 which is a proven restriction modification deficient mutant of 8325-4 [14], and 879R4RF which has a restriction deficient phenotype [7], kindly donated by Prof Don Clewell, University of Michigan. In addition, a series of human disease isolates were used as controls. These isolates were from the Oxford collection of Nick Day that have been microarrayed [15]. Several human isolates from [19] were also compared.

SID 5 (2/05) Page 9 of 26

Strains were identified as S. aureus by growth on mannitol salt agar with yellow pigmentation, catalase positive, staphylase positive (erythrocyte agglutination for the detection of clumping factor) and positive by PCR for nuc and fem genes [22].

Enterococci were obtained from RVC, but they were mostly from non-bovine mammals. Enterococci are commensals of the bovine intestine, so we screened five healthy bovine pats from Boltons Park Farm at the RVC. Samples were plated onto Slatetz and Bartley agar (Oxoid, manufacturer's instructions), and 15 red or moroon colonies were chosen for further evaluation. Enterococci were confirmed using bile aesculin agar and PYR [23]. A PCR reaction was used to identify if the isolates were E. faecalis or E. faecium [24]. Species was also confirmed using tellurite tolerance, as E. faecalis can grow [23]. Three faecium and two faecalis were chosen for further study.

Unless otherwise stated, S aureus were grown in brain heart infusion (BHI) broth or on BHI agar plates. E. faecalis were routinely grown in Todd Hewitt broth (THB) or on THB agar. Media were supplemented where appropriate with chloramphenicol (10g/ml), ampicillin (50g/ml), tetracycline (20g/ml), fusidic acid (25g/ml) or erythromycin (20g/ml). Isolates were stored in BHI plus 20% glycerol at -20ºC or -80ºC.

Antibiotic resistance was determined using standard guidelines for disc diffusion testing (Oxoid). Isolates were grown on Mueller-Hinton agar, and the antibiotics were rifampicin, erythromycin, penicillin, ciprofloxacin, oxacillin, fusidic acid, neomycin, trimethoprim, tetracycline, gentamicin, ampicillin, kanamycin, tobramycin, clindamycin, mupiricin.

SequencingPCR products were amplified using Platinum Pfx DNA polymerase (Invitrogen) with conditions of 94º for 5 min followed by 25 cycles of 94º for 30s, 52º for 30s, 68º for 210s. PCR products were cleaned on QiaQuick spin columns (Qiagen), and both strands of the products sequenced by Lark Technologies. PCR primers to amplify and sequence the hsd genes and their promotors were as follows : sau1hsdR 5'-CGTTTGCGTTGATTGTATTCGG-3' and 5'- ATGCGGATCCTACACTAATCTAGCGAGG-3', sau1hsdS 5'-AATGCATACCTGAAAGAACTTGG-3' and 5'-GACACTGCGCTTTCACAGTGCC-3', and sau1hsdS 5'-ATGCATACCTGAAAGAACTTGG-3' and 5'-CAATTAATAGGTTGTTATCAGG-3'. Subsequently, all the sau1hsdS genes from each of the lineages were amplified and sequenced using 5'-CATACCGAGATATGTCGATAC-3' and 5'-CACTGTGCTATCACAGTGCC-3'. Multilocus sequencing typing (MLST) was performed according to www.mlst.net [25].

Horizontal gene transfer assays and bacteriophage susceptibility

For conjugation, the method of Clewell et al. [9] was used. The donor bacteria was E. faecalis JH2-2 carrying the pheromone-responsive conjugative plasmid pAM378. pAM378 carries a copy of Tn918 encoding a tet selectable marker. pAM378 encodes a pheromone receptor expressed on the E. faecalis surface that responds to pheromone produced by S. aureus. Once triggered, the E. faecalis produces aggregation substance and binds to the S. aureus, allowing conjugative transfer of the pAM378 plasmid. The plasmid does not have an functional replication mechanism in S. aureus and is lost. However, the Tn918 carried on pAM373 can jump to the S. aureus chromosome prior to loss and detected by resistance to tetracycline [9]. 0.5ml of an overnight culture of E. faecalis donor was mixed thoroughly with 0.05ml of overnight S aureus recipient strain culture and pipetted onto the surface of a Millipore filter (0.2mm) placed on Todd Hewitt (TH) agar. The matings were incubated at 37C for 18 hours. The cells were then resuspended in 1.0 ml of BHI by vortexing. 100l aliquots were plated onto Mannitol Salt agar (MSA) supplemented with tetracycline to select for S. aureus transconjugants. MSA contains salt which inhibits the growth of E. faecalis, and the S. aureus

SID 5 (2/05) Page 10 of 26

ferment mannitol causing their colonies to appear yellow. To allow estimation of transfer frequency per donor, aliquots were also plates on THB supplemented with fusidic acid, which is selective for the E. faecalis. All plates were incubated at 37C for 48 hours and colonies counted.

For bacteriophage susceptibility assays, recipient cells were lysed with phage 80, grown on either RN4220 (lineage CC8) or S. aureus 879R4RF (lineage CC51). Recipient bacteria were grown in BHI until mid-log phase, centrifuged, and the pellets resuspended in 7 ml phage buffer (1mM MgSO4; 4mMCaCl2; 0.1M NaCl; 50mM Tris-HCl, 0.1% gelatin; pH 7.8) plus 7 ml BHI. 100l of phage stock were added and the mixture was incubated at room temperature for 10 min. The tubes were transferred to a shaking waterbath and incubated at 32 C and 70 r.p.m. for 4 hours and/or overnight until the cells had visibly lysed. To prepare a phage stock, lysed cultures were centrifuged to pellet any cellular debris, filter sterilized through a 0.2m filter and stored at 4C. Phage stocks were counted by mixing a dilution of stock with recipient bacteria (0.5ml of mid-log phase culture) with additional CaCl2 to 4mM, and left at room temperature for 10 mins. 5ml of molten LK top agar (1% tryptone, 0.5% yeast extract, 0.7% KCl, 0.3% agar) at 45ºC was added, and poured immediately onto a LK bottom agar plate (with 1% agar and selective antibiotic if appropriate). Plates were incubated at 30ºC overnight, and plaques counted.

For transduction between isolates of S. aureus, the marker transferred was the integrated Tn918 carrying tet, or the small class I plasmid pT181 carried in the sequenced strain COL, and carrying a tet selectable marker. The phage stock used was 80, and the method of [14]. Briefly, recipient S. aureus were grown in 50ml LK broth (1% tryptone, 0.5% yeast extract, 0.7% KCl) overnight and resuspended in 3.5ml of LK broth. 50l of 1M CaCl2 and 1ml of phage stock were added. The mixture was incubated at 37ºC statically for 25 mins and then for 15 mins in a shaking waterbath at 37ºC. 1ml of ice-cold 0.02M sodium citrate was added, centrifuged and the pellet resuspended in 1ml ice-cold 0.02M sodium citrate and left on ice for 2 hours. The entire mixture was spread plated (100ml per plate) onto LK with 1% agar plates containing 0.2mg per ml tetracycline, and incubated for 1 hour at 37ºC. An overlay agar (LK with 0.3% agar base and 20mg/ml tetracycline) was carefully poured onto the plates, which were incubated for 48 hours, and transductants counted.

All bacteriophage susceptibility and horizontal transfer assays were performed in triplicate, and data expressed as mean and standard deviation. Statistical significance was determined using Students t test.

For transfer of pAM378 from E. faecalis JH2-2 to bovine E. faecalis or E. faecium, we chose recipient bovine enterococci that were naturally resistant to erythromycin to make identification of transconjugates easier. To distinguish JH2-2 donors from bovine E. faecalis recipient, we chose RAPD using the methods of Harakeh et al. [26] and Monstein et al. [27]. The most successful was RAPD5 [27]. The conjugative transfer reaction was similar to that used with a S. aureus recipient. Transconjugants were plated onto BHI with tetracycline and erythromycin and colonies were screened using the RAPD5.

MicroarraysMicroarrays were performed as reported previously [15]. Briefly, whole genome DNA was extracted from each S. aureus isolates using Qiagen midi-prep columns and labelled with Cy-5 dye. Control DNA from sequenced isolate MRSA-252 was labelled with Cy-3 dye. The two preparations were mixed and hybridised to a seven-strain S. aureus PCR product microarray containing spots for every gene from the seven S. aureus sequencing projects [18]. Microarrays were then washed and scanned. The presence or absence of a gene or gene variant region in a test isolate is determined by its hybridisation to the relevant spot (detected by Cy3 fluorescence) in comparison to hybridisation of the reference strain MRSA252

SID 5 (2/05) Page 11 of 26

(detected by Cy5 fluorescence), and is expressed as a ratio. The ratios can be visualised in GeneSpring by their colour. Spots on the microarray that fluoresce "green" are genes found in the test strain only, spots fluorescing "red" are genes found in the control strain only, spots fluorescing "yellow" are genes found in both strains, and spots that do not fluoresce represent genes found in neither strain. Fully annotated microarray data will shortly be deposited in the public databases BG@Sbase http://bugs.sgul.ac.uk/E-BUGS-33 and ArrayExpress.

When designing the microarray, we specifically included spots corresponding to putative target recognition domain (TRD) regions of sau1hsdS that show the major sequence variation. They were (with forward and reverse primers) : N315sau1hsdS2TDR1 (5'-TGTTTTAACATCGTCAAGACAAG-3' and 5'-ATCGTCCGACTTTTGAAGATTG-3'); MW2sau1hsdS1TDR1 5'-CAAAGGCATACCATTTTTAAGGA-3' and 5'-AGACCTTCTCGACTACCTCCA-3'); MW2sau1hsdS1TDR2 (5'-TTGAGAATAAGGGTGGCACTG-3' and 5'-AAAACGCCTTTTTCCTCTCTTT-3'); MRSA252sau1hsdS2TDR2 (5'-CAAGTATATGGAGGCGGAACAC-3' and 5'-TTTTGAAGTAATCCTTGTTTGAGA-3'); MRSA252sau1hsdS1TDR1 (5'-GGAAAAGAATATTTTGGCTCAGG-3' and 5'-GAAACAGGGTATATGACCTTCA-3'); and MRSA252sau1hsdS1TDR2 (5'-AAAAGGCTATATGCAGAAAATC-3' and 5'-TTAAATTTATGAATCATTCTATGTG-3'). The prefix refers to the sequenced strain used as the template in the PCR reaction. Data analysis was performed using GeneSpring version 7.0 (Silicon Genetics) analysis and there were a series of aims to address.The first aim was to identify animal S. aureus lineages. A gene list containing 728 core variable genes (all genes subtract core genes subtract MGE genes) [15], was used to cluster the isolates using Spearman correlation. The human isolates fall into ten dominant and some minor groups, and these groups correlate with clonal complexes determined by multi-locus sequence typing [15]. For this study, we wished to determine if animal isolates belonged to the same lineages as human isolates, or if they belonged to different lineages, or whether they belonged to one or multiple lineages, or whether animal and human lineages were clearly distinct. In addition, we searched for animal specific genes that may shed light on the animal strain host specificity, especially because many core variable genes are surface expressed and thought to interact with host.The second aim was to determine if the microarray spots corresponding to the sau1hsd genes were present, and/or if sau1hsdS gene variants varied, as this might indicate whether animal and human S. aureus could exchange DNA under the R-M control. The third aim was to look at the distribution of mobile genetic elements (MGE) as an indicator of whether animal and human S. aureus exchange DNA frequently.

Results

1. Establish frequency of transfer of antibiotic resistance markers from enterococci to animal strains of S. aureus. This includes screening for restriction deficient strains, and preliminary investigation of restriction mechanisms.

Eight of the "S. aureus" sent to us by RVC or VLA were not S. aureus. S. aureus and the common canine pathogen S. intermedius are both coagulase positive. This makes discrimination more difficult than in human diagnostic labs, where virtually all coagulase positive staphylococci are S. aureus. We noted that 6 of the "S. aureus" sent to us by RVC and 2 of the "S. aureus" sent to us by VLA were not S. aureus. This was confirmed for three strains by microarray. Correct species identification is usually confirmed using a series of biochemical tests, but for unusual isolates they can be difficult to interpret. We believe all 3 isolates are S. intermedius, but have not examined all options exhaustively. There is clearly a need to identify which species these isolates belong to, to develop simple and more reliable speciation tests for the veterinary setting, and to distribute this information widely.

SID 5 (2/05) Page 12 of 26

Our final strain collection consists of 56 S. aureus isolates from animal sources, including 38 from bovine sources, 13 from equine, 2 ovine, 2 caprine, and 1 camel. They are described in Appendix, Table 2. Three animal strains were naturally resistant to tetracycline and were therefore excluded from conjugation experiments.

The frequency of transfer of tet from enterococci to S. aureus via conjugation is illustrated in Fig 1, and compared to human isolates. Isolates have been grouped according to source (laboratory, bovine, equine, other animal, human). The results show that virtually all strains could accept the tet marker. This was surprising, since the genetic manipulation of environmental S. aureus is notoriously difficult, and we had originally predicted we would have difficulty moving the marker in to many strains, if any. Note that RN4220 has a mutation in sau1hsdR and had a high transconjugate rate, as did the uncharacterised 879RFR4.

1

10

100

1000

10000

100000

tran

scon

juga

nts

ST151ST771ST130ST188 animalST1 animal other animalhuman

bovine equine other human animal

Fig 1. Conjugation frequency of 78 isolates, plotted as log transconjugants per 108 donors, grouped by source and coloured by lineage. Isolates are grouped by source of : laboratory (originally human), bovine, equine, other animal, and human. The laboratory isolates are 8325-4, the next is RN4220 (known R-M mutant of 8325-4) and the next is 879RFR4 (putative R-M mutant). Colours refer lineage (determined by microarray as described in section 2) : red is ST151, blue is ST771, green is ST130, turquoise is ST188 (animal variant), brown is ST1 (animal variant), orange are other animal lineages, and violet are human. There is a strong association between high transfer frequency and bovine S. aureus lineage ST151.

The bovine mastitis isolate RF122 was sequenced during the course of this project (GenBank AJ938182). We MLST typed this isolate using the published sequence, and it is ST151. By microarray it clustered with a group of our ST151 bovine isolates (see section 2), and it also accepted the tet marker by conjugation at high frequency. RF122 has a stop codon at the 5' end of both sau1hsdS1 and sau1hsdS2, predicted to lead to truncated and non-functional proteins. The predicted phenotype for this strain is that it will not be able to recognise foreign DNA because it has no sequence specificity subunits for R-M complexed. Therefore it will not

SID 5 (2/05) Page 13 of 26

digest foreign DNA, and will be more susceptible to bacteriophage lysis and more susceptible to horizontal transfer of DNA. It will also fail to modify its own DNA, and will be a relatively poor donor of DNA to strains of other lineages. To determine if other ST151 had similar mutations, we sequenced the sau1hsdS1 and sau1hsdS2 genes from five ST151 isolates from our collection. They all had identical mutations, suggesting the entire lineage has the same genotype. This explains why our ST151 accepted DNA from enterococci at elevated frequency.

We were somewhat surprised that nearly all isolates could accept DNA from enterococci. A possible explanation was that the assay may be particularly successful at donating DNA, possibly due to factors carried on Tn918. Tn918 carries a gene that is homologous to ardA. This protein encodes a system that blocks R-M. Such genes are not found on the naturally occuring transposons associated with transfer of vanA from enterococci to S. aureus [8a, 9]. We have tried to determine whether the ardA homolog on Tn918 is functional and plays a role in transfer, but so far we have been unsuccessful.

Another explanation for high rates of transfer may have been the donor strain of enterococcus. E. faecalis JH2-2 is a laboratory strain that may not be representative of animal enterococci in nature. In particular, E. faecium is thought to harbour vancomycin resistance genes at higher frequency than E. faecalis. We therefore investigated enterococci from healthy bovine "pats" on farms nearby to the RVC. The pAM378 plasmid was successfully transferred from JH2-2 to five enterococci. Surprisingly, most of the "transconjugants" represented reverse transfer of the recipient erythromycin marker back to the donor JH2-2 strain. Nevertheless, multiple transconjugants of bovine E. faecium and E. faecalis were isolated. Then we used Southern blotting to confirm that the intact pAM378 plasmid was present. All of the E. faecium were not carrying the intact plasmid, but did have the tet marker, suggesting the plasmid was unable to replicate in E. faecium, and the transposon has jumped to the bacterial chromosome. In contrast, three out of nine E. faecalis were carrying an inact pAM378.

The conjugation experiments from enterococci to animal S. aureus were repeated using the naturally occuring enterococcal donors. Transfer frequency to several ST151 isolates was high (10-6), indicating JH2-2 is typical of naturally occuring enterococci. We also showed that conjugation from bovine enterococci that did not have the intact plasmid and likely had an integrated copy of Tn918 in the chromosome, could also transfer to S. aureus but at a lower frequency (10-8), proving that Tn918 is conjugative, as originally suggested by Clewell. The data showed that the assay worked with environmental strains of enterococci as donors just as well as with the laboratory strain JH2-2.

In summary, in naturally occurring animal isolates of enterococci and S. aureus, the transfer of antibiotic resistance genes by conjugation occurs at much higher frequency than predicted. In particular, one bovine lineage ST151 which accounts for one third of bovine isolates is "hyper-recipient" and accepts foreign DNA at 10-10000 times frequency than other isolates. This is due to a mutation in both sau1hsdS1 and sau1hsdS2 genes, such that foreign DNA is taken up by the bacterial cell but not digested.

2. Use the seven-strain S. aureus microarray to establish molecular differences between animal and human S. aureus, with emphasis on whether they share accessory genes, which genes they may carry that could be involved in restriction of gene transfer, and which genes might be associted with bovine pathogenesis.

All 56 strains were hybridised to the seven-strain S. aureus microarray according to [15]. Profiles were compared to well characterised collections of S. aureus, including 61 community acquired invasive S. aureus from the Oxford region, 100 nasal carriage isolates from healthy volunteers in the Oxford region, and MRSA from patients at St George's Hospital and St

SID 5 (2/05) Page 14 of 26

Thomas' Hospital (South London), and a range of "reference" S. aureus that have been sequenced and used widely in laboratory studies.

Data analysis was split into two parts, firstly the investigation of lineage, and then the investigation of mobile genetic elements (MGE). The major lineages of the animal S. aureus were identified by clustering the isolates using 728 CV genes (non-core, non-MGE). The majority of isolates clustered into three unique animal lineages. Representatives of each cluster were subjected to multi-locus sequence typing (MLST; www.mlst.net) [25] to assign a sequence type (ST) to each group. They are ST151, ST771 (novel) and ST130.

The ST151 cluster includes the sequenced isolate RF122, and another nine bovine mastitis isolates. Four more bovine isolates are clearly of the same lineage but show slight variation, they are all from the Essex region, and labelled ST151a. The ST771 cluster included 9 bovine isolates, 2 equine isolates, 1 ovine isolate, 1 caprine isolate, and 1 camel isolate. The ST130 cluster included 10 bovine isolates and 3 equine isolates.

A further two smaller clusters are seen. Three bovine isolates clustered together, and MLST assigned them to group ST188, a relatively uncommon lineage in humans. However, by microarray the bovine ST188 and human ST188 did not cluster together. This could represent the fact that MLST is based on data from seven genes, while microarray lineage clustering is based on 728 genes. Thus we refer to the bovine isolates as ST188a. The second cluster include 1 bovine isolate, 1 caprine and 4 equine isolates and are related to the common human lineage CC1, although they are clear variants and we call them ST1a.

A few unusual isolates are noted. One equine isolate is an MRSA-15, a UK hospital-associated MRSA, but is phenotypically methicillin sensitive and may have a mutation in a regulatory gene (ST22). Two equine isolates are CC8, a common human lineage, and are also multi-drug resistant. One bovine and one equine isolate were unusual types, and are called orphan.

SID 5 (2/05) Page 15 of 26

1 188 130 771 151

Fig 2. Microarray data from 217 isolates (161 human carriage and invasive isolates [15] and 56 animal isolates), clustered by Spearman correlation using 728 CV genes. Vertical lines represent each isolate and are coloured in the tree and blocks below by their lineage; all animal isolates have been coloured red with the corresponding ST types written below. ST1 isolates cluster with human CC1 isolates (pink), ST188a, ST130, ST771 and ST151 form unique clusters. The horizontal lines within the main picture represent genes. A yellow or red signal indicates a gene is present, a blue, white or grey signal indicates absent.

Although the majority of animal isolates clustered into lineages that were not found in humans, the animal clusters were not not on a separate branch of the S. aureus tree. Our current hypothesis is that there was a common ancestor of S. aureus, which then branched into multiple lineages that have evolved separately, but undergone some recombination between lineages. The clustering suggests recombination between animal and human lineages has occurred at a similar rate to recombination between human lineages.

Are there unique human-specific or animal-specific genes? Our microarray can be used to address human-specific genes, but not animal-specific as it does not include genes from a sequenced animal isolate (S. aureus RF122 sequence has only been released in the last few months). Therefore we asked if there are unique genes found in most human isolates (core genes, found in >95%) that are not found in animal isolates (ie. human specific genes). A visual inspection of 2013 core genes found several that were under-represented (but not uniformly missing) in animal strains. They were : SAR0179 (putative transporter, similar to LmrP), SAR0744 (putative DNA photolyase), SAR0745 (putative membrane protein), SAR0746 (putative membrane protein), SAR2561 (conserved hypothetical protein), SAR2633 (putative gene regulator of the tetR family). The sequenced S. aureus genomes carry many homologs of transporters, membrane proteins and regulators. Thus, the list was small, with few obvious markers of human isolates.

A further way to ask this question is to compare the CV genes found in human isolates and associated with specific lineages, and determine if any are relatively more or less common in animal isolates. The 728 human CV genes were investigated. After visual inspection, no CV genes appeared to be clearly associated or not-associated with the animal isolates. We then extended this to the genes associated with genomic islands - originally thought to be MGE, but from microarray evidence suspected to be CV [15]. Two genes were associated with animal isolates at higher frequency than human, MW0378 (conserved hypothetical protein), MW0379 (conserved hypothetical protein). In summary, we are unable to explain why animal isolates generally affect animals, and human isolates affect humans, and raises the possibility that host-specific barriers may not be robust.

SID 5 (2/05) Page 16 of 26

The next step was to investigate MGE, which includes plasmids, bacteriophage, pathogenicity islands (SaPI), transposons and integrative elements. Firstly we asked if there was evidence of genetic exchange between human and animal isolates. It is important to understand that MGE vary significantly between human S. aureus, and this occurs in specific ways. Some MGE are common in all isolates or in some lineages, and this is evidence of stability. Others vary a lot, even between isolates of the same lineage, and these are probably unstable and/or highly mobile. Each MGE has a mosaic structure, and is made up of a combination of fragments that may be found in other MGE, indicating recombination between MGE is high. Since microarrays do not give information about the position of genes in the chromosome, we use "composite genomes" to estimate their location based on sequenced MGE [18].

The data for the animal S. aureus suggests that lineages ST151 and ST771 have evidence of relatively stable phage that are similar to those found in human isolates. ST151 and ST130 have evidence of relatively stable SaPI that are similar to those found in human isolates. However, both phage and SaPI show evidence of different distribution of genes in human versus animal S. aureus. SCC elements were not found, or were sufficiently unrelated to those found in human isolates, suggesting no transfer of SCC. Human plasmids were rare in animal isolates, although the transposon encoding b-lactamase (and commonly found on plasmids) was found in some ST130 isolates. A transposon similar to Tn554 encoding erythromycin resistance was found in some strains, but other transposons were rare. This data suggests that bacteriophage and SaPI can transfer between animal and human isolates, but relatively infrequently, while plasmids, SCC and transposons transfer very infrequently. Bacteriophage are necessary for most MGE transfer via generalised transduction, but not for conjugation. Its interesting to note that most resistance genes are encoded on plasmids, SCC and transposons.

SID 5 (2/05) Page 17 of 26

Fig 3. Microarray data comparing human (top) versus animal (bottom) S. aureus, and the distribution of genes from the bacteriophage 3-252. Each horizontal line represents a strain, and each vertical line represents a gene (opposite to Fig 2). int is phage integrase, sep is enterotoxin P, sak is staphylokinase, chp is chemotaxis inhibitory protein. yellow or red = gene present; blue, grey or black = gene absent. While some of the 3-252 phage genes are found in the animal isolates, they are clearly less common than in the human isolates.

We have recently proposed that lineages evolve independently due to variation in the specificity subunit of their R-M systems, and that this would also extend to the transfer of MGE. We therefore investigated the distribution of hsdS genes in animal isolates, and the sequence of sau1hsdS1 and sau1hsdS2 genes. By microarray, the sau1hsdS gene variant regions were distributed according to lineage, and this fits our hypothesis. Sequencing revealed animal lineages have some variable regions matching those found in human lineages, as well as some unique regions. The exception was CC1a which had sequences matching the human CC1 lineage, and suggesting greater potential for MGE transfer between these populations.

Antibiotic resistance patterns were determined for all of the isolates. We noted that the isolates from human lineages were more likely to be multi-drug resistant, whereas most animal isolates were resistant to few antibiotics. Penicillin and kanamycin resistance were common in isolates of ST130, indicating this lineage has a typical MGE common to all strains, and indeed a region corresponding to the bla (b-lactamase) lit up on the microarray for these isolates.

In summary, S. aureus from cows are predominantly from three lineages that are distinct from human lineages. Two minor lineages are also found, one of which is similar to a human lineage. Isolates from horses are from either bovine or human lineages, and the isolates from human lineages tend to be multi-resistant suggesting they have come directly from humans or interacted with human isolates. Isolates from sheep, goats and a camel were from the animal lineages. Animal lineages are different to humans, but have not evolved as a distinct "animal" branch. This suggests cross-infection to new hosts may be possible. Genetic exchange between animal and human S. aureus appears to be possible, but occurs at low frequency.

3. Establish frequency of transfer of antibiotic resistance markers from animal S. aureus to human S. aureus, with preliminary investigation into the nature of transfer barriers

The generalised transducing phage 80 was grown on animal strains of lineage ST151 carrying Tn918, and transduced to 4220, 4220 + pSK, 4220 + pSKhsdR, 8325-4 (Table 1). On some occasions, one or two transconjugants were seen. This indicates transfer can occur but at very low frequency.

donor lineage Recipient for transduction Plaques on4220 4220pSK 4220hsdR 8325-4 RF122 4220 8325-4 RF122

RF122 151 0 0 1 0 106 3.0 x 106 4.0 x 102 3.0 x 108

982BL 151 0 0 0 0 2 1.7 x 107 4.9 x 103 1.7 x 107

AN22 151 0 2 0 0 0 2.5 x 107 1.0 x 108 8.0 x 104

29905 771 0 0 0 0 0 1.3 x 103 2.2 x 103 032959 130 0 0 0 0 0 6.4 x1010 5.9 x 109 3.0 x 102

818 188a 0 0 0 0 0 1.4 x 101 5.1 x 101 0

Table 1. 80 was grown on the S. aureus donors listed, each carrying the Tn918 and tet marker. The lineage of these strains is also listed. The resultant phage lysates were tested for their ability to transduce to various recipients, and also for their ability to plaque on various recipients. The data show that most phage can plaque on human S. aureus recipients, suggesting the phage can deliver infectious phage DNA, but there is a major block to the successful transfer of the marker DNA.

SID 5 (2/05) Page 18 of 26

Why is transfer inefficient? One explanation could be that once the phage was grown on the animal isolates, it was altered in some way and was then unable to deliver DNA to the 8325-4 derived strains. We therefore checked the phage lysates by plating them out on RN4220 and 8325-4 and counting the numbers of plaques. We noticed that plaque titres were high, indicating the phage can successfully deliver infectious phage DNA, but there is a block delivering the Tn918 DNA. We noted that phage generally plaqued on 4220 at higher frequency than 8325-4 (102 to 104 times) showing that functional R-M in 8325-4 blocks the ability of phage to deliver DNA. However, an additional factor or factors prevents the transfer. In combination, the factors reduce the efficiency of transfer between S. aureus so that it cannot be detected in laboratory models.

To investigate this second factor further, we tested the ability of animal isolates of ST151 to transduce the Tn918 markers to RF122, also of ST151 and with a disabled R-M. Higher rates of transfer were seen with the RF122 recipient than other isolates of the lineage. Therefore, the second factor(s) is sufficient to block transduction of markers between S. aureus of the same lineage, but not as efficiently as between the same isolates. The identity of these factors is unknown. There is new evidence that additional R-M systems are found on some MGE [30] which could explain our data, but it is also possible there are specific blocks to phage-mediated transfer. This needs further investigation.

In summary, transduction of antibiotic resistance genes between isolates of different lineages occurs at very low frequency. Frequency is low in part due to Sau1 R-M, but there is also a second factor or factors that block transfer between S. aureus.

Discussion

This project has lead to some key new information that is important for understanding what animal S. aureus isolates are, and how they acquire antibiotic resistance genes.

The majority of animal isolates clustered into three major lineages, ST151, ST771 and ST130 and these are different to human lineages. ST151 isolates have recently been reported in the UK by Smith et al. [28], where they accounted for 9 of 11 bovine isolates. The recently sequenced bovine isolate RF122 from Ireland is also ST151. ST771 are novel and have not been reported before in the MLST database of over 1400 S. aureus isolates. ST130 isolates have been reported previously in bovine samples from Norway. Interestingly, Smith et al. [28] reported that the majority of USA and Chilean bovine isolates belonged to different lineages than those found in our study. This suggests that different countries have different lineages of animal S. aureus. This somewhat mimics the situation with human S. aureus, where each country has unique combinations of MRSA lineages.

The majority of bovine isolates came from the three dominant lineages, and are distinct from human lineages. This confirms and extends earlier studies that suggested bovine and human strains are distinct [16]. However, while the animal isolates were from lineages that were different from human isolate lineages, the animal lineages did not cluster as a group independently of human lineages. We were unable to identify strong "human markers" or "animal markers". Since the main difference between lineages is their combination of surface proteins that are presumed to interact with host receptors, it is still not clear why animal strains interact with animals and human strains with humans. The data raise the possibility that the host-specific interaction may not be robust, and perhaps lineages are associated with specific hosts due to geography and/or epidemiological boundaries. Further work in this area is required.

SID 5 (2/05) Page 19 of 26

In contrast to the bovine isolates, many of the equine isolates came from lineages associated with humans, including a hospital MRSA. Human MRSA in horses have recently been reported in Canada [29], and perhaps are the result of close human-horse contact, particularly with horses in the racing industry. S. aureus is not considered to be a commensal of horses, although there are limited recent studies on the subject.

Horizontal transfer of antibiotic resistance genes into S. aureus occurs via two main pathways, conjugation and transduction. Conjugation from enterococci is the pathway for vancomycin resistance (vanA) transfer. We used a model of this transfer described by Clewell et al., [9] that uses a pheromone-sensitive plasmid carrying a conjugative transposon Tn918 encoding a tet marker to deliver to S. aureus. We were extremely surprised to see that the tet marker transferred to nearly all S. aureus that we tested, including animal and human clinical isolates. We know of no other reliable transfer mechanism for delivery of markers into clinical isolates. This model was so successful that we hope to exploit it in the future to make tools for genetic manipulation of clinical isolates. The reason for the transfer is currently not clear, although we suspect that the conjugative transposon Tn918 plays a role. This transposon may not be typical of transposons that carry vanA genes, and further research into the characteristics of these transposons is warranted.

The second surprise was that some animal isolates accepted the tet marker from enterococci at very high frequencies, even greater than those of known laboratory mutant isolates that are exploited for their susceptibility to foreign DNA transfer. Further investigation showed that the "hyper-recipient" isolates belonged to one of the bovine lineages, ST151. Fortuitously, during the course of this project, the whole genome sequence of an ST151 isolate (RF122) was released on a public database (GenBank AJ938182). This genome sequence carried mutations in both sau1hsdS1 and sau1hsdS2 genes, and this is highly likely to explain its enhanced ability to accept the tet marker. We have recently shown [10a] that the Sau1 restriction modification system in S. aureus is the major pathway blocking horizontal gene transfer into S. aureus. This system uses five genes: sau1hsdR, sau1hsdM1, sau1hsdM2, sau1hsdS1 and sau1hsdS2. The gene products encode subunits that are predicted to form two enzyme complexes, M2S (modification) and R2M2S (restriction). Both complexes recognise specific short sequences on DNA via the sau1hsdS encoded specificity subunit. The modification enzyme will modify the DNA at this sequence, while the restriction enzyme will digest unmodified DNA. Thus a bacterial cell will protect its own DNA from digestion, but any foreign DNA that is unmodified will be recognised as foreign and digested, protecting the cell. If both sau1hsdS subunits are mutated, then the Sau1 system will be non-functional; no DNA will be modifed or digested. This would explain why ST151 isolates can accept the tet marker by conjugation from enterococci at high frequency.

We further sequenced the sau1hsdS1 and sau1hsdS2 genes from 5 other ST151 isolates from our collection and they all had identical stop mutations in these genes compared to RF122. Therefore, we propose that this lineage has evolved with these mutations and they are probably common to all isolates of this lineage. ST151 isolates are therefore "hyper-recipient" for receiving DNA via any mechanism, as they have two mutant genes in the major pathway for blocking foreign DNA acquisition.

The next issue is can resistance genes spread between S. aureus of animal and human origin? This would occur via generalised transduction, that is, using bacteriophage (bacterial viruses) to package bacterial DNA and deliver it to other bacteria. In S. aureus this is the most common way of exchanging DNA, and is highly species specific. Our laboratory models show that Sau1 is capable of blocking generalised transduction of MGE between isolates of different lineages, and since bovine and human isolates generally come from different lineages, transfer is blocked. However, ST151 are an exception since their Sau1 pathway is compromised. In

SID 5 (2/05) Page 20 of 26

addition, a few bovine isolates and many equine isolates were from lineages closely related to human lineages, suggesting this barrier would not be sufficient to block transfer.

Further experiments identified a second pathway that blocks transfer of DNA between S. aureus via transduction. Further research is required to identify what this pathway is. This is important, as if the second pathway is compromised in some strains, and this is combined with the ST151 genotype, there is potential for "hyper-recipient and hyper-donor" strains to arise, and these would be the source of rapid antibiotic resistance spread in S. aureus.

Although transfer of MGE between animal and human S. aureus is possible, the next issue is whether antibiotic resistance gene transfer is actually occuring in animal S. aureus. Our antibiotic resistance studies suggest that transfer is possible, but when we screened animal isolates for resistance levels they were generally low. Furthermore, our microarray data suggests animal S. aureus are capable of exchanging MGE with human S. aureus, but they are not exchanging frequently. Both of these data suggest that the mechanisms of transfer are intact, but so are some barriers to transfer. It also suggests opportunity for transfer may be limited. Perhaps this is due to minimal contact between animal and human S. aureus. It may also be due to minimal selective pressure, particularly since avoparcin has been banned since 1997 in the EU, and due to the decreasing use of growth promoters of the antibiotic classes. However, the situation in horses may need close monitoring, as they can be exposed to antibiotics, animal lineages and human lineages of S. aureus. At the present time, the high risk lineage ST151 was not identified on horses.

Main Conclusions

1. In a small study of UK animal isolates, three distinct lineages were predominant. These lineages are different to animal isolates from other countries.2. Bovine isolates are different to human isolates, while equine isolates can be from either bovine or human lineages. Human lineages are generally more multi-resistant.3. One third of bovine isolates belonged to the lineage ST151 in this small UK study, and they have mutations in the major R-M pathway. This means they are "hyper-recipients" and capable of accepting resistance genes (such as vanA) from enterococci at high frequency. No other naturally occuring S. aureus "hyper-recipients" have been described.4. R-M effectively blocks the transfer of resistance genes between certain bovine S. aureus lineages and human S. aureus. Further research is required to characterise a second pathway that blocks transfer between S. aureus isolates. 5. There is evidence that animal S. aureus can exchange MGE with human S. aureus, but also evidence that this occurs at low frequency, perhaps indicating limited interaction between these strain populations, and limited selective pressure. Main implications

One third of bovine S. aureus in this small study in the UK belonged to a hyper-recipient lineage of S. aureus that is susceptible to resistance gene transfer from enterococci. No other strains of S. aureus have this ability. This unexpected result suggests the agricultural ban of avoparcin use in 1997 was justified, as it would be a risk for the transfer of vancomycin resistance genes from VRE to animal S. aureus. It suggests some sentinel screening for VRSA in bovine strains is warranted. This is particularly relevant in areas where VRE is still found.

The transfer of resistance genes between different S. aureus lineages is occuring at low frequency and this is likely due to effective genetic barriers, limited opportunities for mixing of isolates, and limited selective pressure due to decreasing use of growth promotors and

SID 5 (2/05) Page 21 of 26

antibiotics. Periodically, lineage and antibiotic resistance rates should be monitored in animals, particularly horses.

Veterinary diagnostic laboratories need support to correctly identify S. aureus.

Future Work

While we have used a model of vanA transfer from enterococci to S. aureus in the laboratory, the model could be improved. Our current model involves donation on a conjugative transposon (Tn918) carried on a conjugative plasmid. The likely donor in the four cases of vanA transfer described in nature are of a non-conjugative transposon on a conjugative plasmid. In addition there is evidence Tn918 carries a mechanism that blocks R-M. To refine our model, these factors should be investigated.

We have also identified a second pathway that blocks transfer of resistance genes between animal and human isolates via transduction, but it remains to be characterised. It could be that this pathway has so far prevented resistance gene spread throughout S. aureus. If some isolates are deficient in this pathway, and this deficiency was spread to S. aureus with deficient Sau1 (such as ST151), these isolates could be hyper-recipient and hyper-donors of antibiotic resistance transfer, leading to rapid evolution of fully multi-resistant S. aureus. This possibility needs to be investigated further.

The reasons for the geographical variation in animal S. aureus from different countries are unknown. It could be due to lack of opportunity for spread of strains, or due to specific interactions between certain lineages and certain types of animals. This mimics a dilemma in human S. aureus, in that we still don't know why some countries or geographical regions have different S. aureus lineages. Further investigation could identify the key factors that control host-specificity, which would be targets for anti-staphylococcal therapies and prophylaxis.

Rapid and simple methods for discriminating S. aureus from S. intermedius and other animal staphylococci are required to assist veterinary laboratories in screening for antibiotic resistant S. aureus, including MRSA. Information about suitable tests needs to be widely publicised to relevant laboratories.

Appendix

Table 2. S. aureus isolates, source, lineage, resistance profiles and conjugation frequencies.

Isolate source region lab supp-lier

Conjugationexp 1

Conjugationexp 2

conjug per 108 donors

lineage resistance

RF122 bovine mastitis

Ireland Fitz-gerald

2.34 x 10-4 2.47 x 10-4 24050 151 not done

38963 Bovine milk Essex rvc 4.55 x 10-5 5.44 x 10-5 4995 151

982BL Bovine milk Essex rvc 1.04 x 10-4 7.16 x 10-5 8780 151

C00759 bovine milk Dorset vla 1.40 x 10-4 1.04 x 10-4 12200 151

C01262 bovine milk Cumbria vla 1.01 x 10-4 7.14 x 10-5 8620 151

C01756 Bovine milk Cambridgeshire

vla 3.10 x 10-4 5.17 x 10-4 41350 151

C01835 bovine milk Wiltshire vla 5.05 x 10-5 9.53 x 10-5 7290 151 Cli

C01893 bovine milk Shropshire vla 1.12 x 10-4 1.26 x 10-4 11900 151

C0422 bovine milk Carmarthen

vla 2.35 x 10-4 2.35 x 10-4 23500 151

C123/5/05-29

bovine milk Essex rvc 8.44 x 10-5 8.33 x 10-5 8385 151 Cli

38690 Bovine milk Essex rvc 8.44 x 10-5 1.18 x 10-5 4810 151a

C123/5/0 Bovine milk Essex rvc 2.62 x 10-5 2.76 x 10-5 2690 151a

SID 5 (2/05) Page 22 of 26

5-22C123/5/05-31

Bovine milk Essex rvc 1.09 x 10-4 8.75 x 10-5 9825 151a

C123/5/05-25

Bovine milk Essex rvc 6.14 x 10-5 3.84 x 10-5 4990 151a Kan

29905 Bovine milk Essex rvc 3.42 x 10-8 2.62 x 10-8 3.02 771

30424 Bovine milk Essex rvc 2.50 x 10-9 5.21 x 10-8 2.705 771 Cip

30425 Bovine milk Essex rvc 2.88 x 10-8 6.20 x 10-8 4.54 771 Cip

32320 Bovine milk Essex rvc 1.30 x 10-8 1.59 x 10-9 0.75 771

32932 Bovine milk Essex rvc 0 0 0 771

32933 Bovine milk Essex rvc 0 7.56 x 10-8 3.78 771 Cip

38162 Bovine milk Herts rvc 3.77 x 10-7 1.06 x 10-6 71.85 771

C01719 bovine milk Staffordshire

vla 3.67 x 10-7 1.08 x 10-8 18.89 771

C01771 bovine milk Somerset vla 6.72 x 10-7 8.74 x 10-7 77.3 771 CipFos

459J Bovine milk Herts rvc 2.44 x 10-8 3.10 x 10-8 2.77 130 Pen

C00595 bovine milk Powys vla 1.30 x 10-7 2.29 x 10-7 17.95 130 Pen

C00704 bovine milk Oxon vla 1.88 x 10-7 1.58 x 10-7 17.3 130 PenKan

C01122 bovine milk Sussex vla 2.21 x 10-6 3.65 x 10-6 293 130 Pen

C01123 bovine milk Derbyshire vla 6.64 x 10-7 9.44 x 10-7 80.4 130 PenCipKan

C01312 bovine milk Cornwall vla 5.47 x 10-7 2.60 x 10-6 157.35 130 Pen

C01611 bovine milk Nottinghamshire

vla 3.30 x 10-7 1.37 x 10-7 23.35 130 PenCipCli

C01791 bovine milk Monmouthshire

vla 9.88 x 10-7 1.92 x 10-6 145.4 130 PenGenAmpKanTobFos

C01801 bovine milk Cheshire vla 5.13 x 10-6 3.38 x 10-6 425.5 130 PenKan

C01899 Bovine milk vla 5.54 x 10-6 9.91 x 10-7 326.55 130 PenKan

818 Bovine milk Herts rvc 1.52 x 10-5 1.14 x 10 -5 1330 188a30296 Bovine milk Herts rvc 1.26 x 10-7 5.80 x 10-6 296.3 188a

30375 Bovine milk Kent rvc 3.70 x 10-6 2.45 x 10-5 1410 188a

C01865 bovine milk Denbighshire

vla 1.35 x 10-4 1.20 x 10-4 12750 1a Kan

38428 Bovine milk Essex rvc 2.70 x 10-8 2.44 x 10-7 13.55 orphan

34060 Equine blood Herts rvc 1.75 x 10-8 7.60 x 10-9 1.275 1a EryPenCipOxAmp

34095 Equine hoof swab

Berkshire rvc 2.26 x 10-8 9.45 x 10-8 5.855 1a

30067 Equine skin swab

Hants rvc 3.08 x 10-8 3.33 x 10-8 3.205 1a Pen

30725 Equine submandibular

Herts rvc 1.62 x 10-7 4.47 x 10-7 30.45 1a Cip

29957 Equine catheter tip

Dorset rvc Naturally tetR Naturally tetR 8 PenTriTetGenAmpKanTobCli

39528 Equine swab Herts rvc Naturally tetR Naturally tetR 8 PenTetAmpKan

31235 Equine nasal swab

Herts rvc 7.41 x 10-8 1.67 x 10-8 4.54 22 Kan

32942 Equine nasopharyngeal swab

Middlesex rvc 2.79 x 10-7 1.27 x 10-6 77.45 130

32326 Equine swab Herts rvc 5.6 x 10-8 1.60 x 10-6 82.8 130 Cli

32959 Equine tail swab

Herts rvc 1.02 x 10-8 1.02 x 10-8 1.02 130 Cli

36364 Equine nasopharynx

Middlesex rvc 9.88 x 10-6 4.28 x 10-5 2634 771

37695 Equine shoulder wound

Herts rvc 2.35 x 10-7 2.97 x 10-7 26.6 orphan

37967 Equine uterine swab

Herts rvc 1.37 x 10-5 3.95 x 10-5 2660 771

32451 Caprine milk Herts rvc Naturally tetR Naturally tetR 1a RifPenFusNeoTetCipKan

SID 5 (2/05) Page 23 of 26

39587 Caprine milk Herts rvc 4.00 x 10-7 4.26 x 10-7 41.3 771

28766 Ovine lung tissue

Herts rvc 2.43 x 10-7 1.13 x 10-6 68.65 771 Kan

38474 Ovine milk Herts rvc 5.51 x 10-7 1.32 x 10-6 93.55 771

37609 Camel dermatitis

London rvc 5.82 x 10-7 NA 58.2 771 Mup

Resistance : Rif = rifampicin, Ery = erythromycin, Pen = penicillin, Cip = ciprofloxacin, Ox = oxacillin, Fus = fusidic acid, Neo = neomycin, Tri = trimethoprim, Tet = tetracycline, Gen = gentamicin, Amp = ampicillin, Kan = kanamycin, Tob = tobramycin, Cli = clindamycin, Mup = mupiricin.

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.

SID 5 (2/05) Page 24 of 26

1. PHLS (2003) Staphylococcus aureus bacteremia: England, Wales, Northern Ireland, January to December 2002. CDR Weekly 13(12):2. Gould IM (2005). The clinical significance of methicillin-resistant Staphylococcus aureus. J Hosp Infect. 61(4):277-82.3. PHLS Nosocomial Infection Surveillance Unit "Surgical Site Infection, 1997-8". Currently not available on-line, but a copy can be emailed on-request. See hpa.org.uk/infections/topics_az/hai/SSIreport.pdf for most recent document (although the incidence of S. aureus infection cannot be calculated from data in the newer document).4. Moore PCL, Lindsay JA (2002) Molecular characterisation of the dominant UK methicillin-resistant Staphylococcus aureus strains, EMRSA-15 and EMRSA-16. J. Med. Microbiol. 51:516-21.5. Chang S, Sievert DM, Hageman JC, Boulton ML, Tenover FC, Downes FP, Shah S, Rudrik JT, Pupp GR, Brown WJ, Cardo D, Fridkin SK (2003). Infection with vancomycin-resistant Staphylococcus aureus containing the vanA resistance gene. N Eng J Med 348:1342-7.6. Centers for Disease Control and Prevention (CDC). (2004) Vancomycin-resistant Staphylococcus aureus--New York, 2004. MMWR Morb Mortal Wkly Rep. 53(15):322-3.7. Witte W (2000) Selective pressure by antibiotic use in livestock. Int. J. Antimicrob. Agent 16: S19-24.8. Noble WC, Virani Z, Cree RG. (1992) Co-transfer of vancomycin and other resistance genes from Enterococcus faecalis NCTC 12201 to Staphylococcus aureus. FEMS Microbiol. Lett. 72(2):195-8.9. Clewell DB, An FY, White BA, Gawon-Burke C (1985) Streptococcus faecalis sex pheromone (cAM373) also produced by Staphylococcus aureus and identification of a conjugative transposon (Tn918). J. Bacteriol. 162(3):1212-20.10. Clewell DB, An FY, Flannagan SE, Antiporta M, Dunny GM (2000) Enterococcal sex pheromone precursors are part of signal sequences for surface lipoproteins. Mol. Microbiol. 35(1):246-7.11. Weigel LM, Clewell DB, Gill SR, Clark NC, McDougal LK, Flannagan SE, Kolonay JF, Shetty J, Killgore GE, Tenover FC (2003). Genetic analysis of a high-level vancomycin-resistant isolate of Staphylococcus aureus. Science. 302(5650):1569-71.12. Showsh SA, De Boever EH, Clewell DB (2001). Vancomycin resistance plasmid in Enterococcus faecalis that encodes sensitivity to a sex pheromone also produced by Staphylococcus aureus. Antimicrob Agents Chemother 45(7):2177-8.13. Bickle TA, Kruger DH (1993) Biology of DNA restriction. Microbiol Rev 57(2):434-50.14. Waldron DE, Lindsay JA. (2006) Sau1: a novel lineage-specific Type I Restriction-Modification system that blocks horizontal gene transfer into Staphylococcus aureus, and between S. aureus isolates of different lineages. J. Bact. in press. 15. Lindsay, JA, Moore CE, Day NP, Peacock SJ, Witney AA, Stabler RA, Hussain SE, Butcher PD, Hinds J. (2006). Microarrays reveal each of the ten dominant lineages of Staphylococcus aureus have unique combinations of surface-associated and regulatory genes. J. Bact. 188: 669-76.16. Zadoks RN, van Leeuwen WB, Kreft D, Fox LK, Barkema HW, Schukken YH, van Belkum A (2003) Comparison of Staphylococcus aureus isolates from bovine and human skin, milking equipment, and bovine milk by phage typing, pulsed-field gel electrophoresis, and binary typing. J. Clin. Microbiol. 40:3894-902.17. Lindsay JA, Holden MTG. (2004) Staphylococcus aureus : SuperBug, SuperGenome? Trends Microbiology 12(8): 378-85.18. Witney AA, Marsden GL, Holden MTG, Stabler RA, Husain SE, Vass JK, Butcher PD, Hinds J, Lindsay JA. (2005) Design, validation and application of a seven-strain Staphylococcus aureus PCR product microarray for comparative genomics. Appl. Environ. Microbiol. 71:7504-7514.19. Moore PC, Lindsay JA (2001). Genetic variation among hospital isolates of methicillin-sensitive Staphylococcus aureus: evidence for horizontal transfer of virulence genes. J. Clin. Microbiol 39(8):2760-7.20. Lacey RW (1980) Rarity of gene transfer between animal and human isolates of Staphylococcus aureus in vitro. J Gen Microbiol 119:437-42.21. Enright MC, Robinson DA, Randle G, Feil EJ, Grundmann H, Spratt BG (2002) The evolutionary history of methicillin-resistant Staphylococcus aureus (MRSA). Proc Natl Acad Sci USA 99:7687-92.22 Becker K, von Eiff C, Keller B, Bruck M, Etienne J, Peters G. (2005) Thermonuclease gene as a target for specific identification of Staphylococcus intermedius isolates: use of a PCR-DNA enzyme immunoassay. Diagn Microbiol Infect Dis. 51(4):237-44.23. Teixeira LM, Facklam RR. "Enterococcus" in Murray PR (ed) Manual of Clinical Microbiology (2003), ASM Press.24. Harwood VJ, Delahoya NC, Ulrich RM, Kramer MF, Whitlock JE, Garey JR, Lim DV (2004). Molecular confirmation of Enterococcus faecalis and E. faecium from clinical, faecal and environmental sources. Lett. Appl. Microbiol. 38(6):476-8225. Enright MC, Day NP, Davies CE, Peacock SJ, Spratt BG (2000). Multilocus sequence typing for characterization of methicillin-resistant and methicillin-susceptible clones of Staphylococcus aureus. J Clin Microbiol. 38(3):1008-15.26. Harakeh HS, Uwaydah M, Matar GM. 2004. Random amplified polymorphic DNA typing of Enterococcus faecalis isolated from Lebanese individuals. Eastern J Medicine 5(1):18-2027. Monstein HJ, Quednau M, Samuelsson A, Ahrne S, Isaksson B, Jonasson J (1998). Division of the genus Enterococcus into species groups using PCR-based molecular typing methods. Microbiology. 144 (5):1171-9.28. Smith EM, Green LE, Medley GF, Bird HE, Fox LK, Schukken YH, Kruze JV, Bradley AJ, Zadoks RN, Dowson CG (2005). Multilocus sequence typing of intercontinental bovine Staphylococcus aureus isolates. J Clin Microbiol. 43(9):4737-43.29. Weese JS, Rousseau J, Willey BM, Archambault M, McGeer A, Low DE (2006). Methicillin-resistant Staphylococcus aureus in horses at a veterinary teaching hospital: frequency, characterization, and association with clinical disease. J Vet Intern Med. 20(1):182-6.30. Dempsey RM, Carroll D, Kong H, Higgins L, Keane CT, Coleman DC. (2005). Sau42I, a BcgI-like restriction-modification system encoded by the Staphylococcus aureus quadruple-converting phage Phi42. Microbiol. 151:1301-11.

SID 5 (2/05) Page 25 of 26

SID 5 (2/05) Page 26 of 26