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 22

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 SE3020

2. Project title

An integrated approach to the application of M.bovis genotyping for the control of bovine tuberculosis in GB

3. Contractororganisation(s)

Veterinary Laboratories Agency Woodham LaneAddlestoneSurreyKT15 3NB     

54. Total Defra project costs £ 927801

5. Project: start date................ 01 April 2001

end date................. 30 September 2004

SID 5 (2/05) Page 2 of 22

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.

Genome variation studied by molecular epidemiology has made new insights into disease transmission and evolution possible. The information gathered from studying strain variation can be used for modelling disease dynamics, predicting epidemics, policy planning, monitoring interventions, gaining insight into the life processes of the organism and the nature of the selective pressures driving strain variation.

At VLA the current methods of choice for the molecular typing of M. bovis isolates are spacer-oligonucleotide typing (spoligotyping) and variable number of tandem repeat (VNTR) typing. The VLA spoligotype database that currently holds typing information on ~32,000 M. bovis strains isolated from 1975-2005 (with ~95% of data for strains isolated since 1997). The VNTR method uses 7 ETR targets (A-F) originally described by Frothingham and Meeker-O'Connell (1998) and the VNTR database currently holds data for ~8,000 isolates.

The aim of this project was to answer the question posed in the MAFF research requirement R4 – ‘How can new molecular typing tests and reagents be used to address questions on the epidemiology of TB in cattle and wildlife?’ The M. bovis genome sequence was used to develop complementary strain typing tests to VNTR and spoligotyping i.e. SNP (SNM) typing and whole genome deletion typing which helped to clarify the genetic relationships between strains found in GB. The tests were targeted to strains of epidemiological importance including archival material collected over the past 20-30 years. The data generated was used to analyse transmission of TB types among farm and wildlife species using a combination of population genetic analysis, spatial statistical analysis, molecular sequencing, phenotypic and epidemiological analyses.

Analysis of the population structure of M. bovis in GB revealed Spoligotype 9 is the progenitor strain of approximately 94% of all strains of M. bovis in GB and that different daughter strains have emerged in different geographical areas. Thus the epidemic of bovine tuberculosis in GB may be seen as a series of local epidemics caused by different strains emerging in different areas of the country.

The most plausible explanation for the population structure of M. bovis in GB is that M. bovis has undergone a series of 'clonal expansions' either caused by the spread of a favourable mutation, or as a result of a 'founder effect' as a clone invades a previously uninfected host or geographical area. In reality both selection and population sampling effects are probably involved in generating clonal expansion and geographical localization of M. bovis in GB. It is both possible and worthwhile to tease apart the relative contributions of these population level forces to the geographical localization of M. bovis in GB in order to

SID 5 (2/05) Page 3 of 22

target disease control strategies more effectively. This remains an important challenge for future research so that control strategies can be targeted most effectively.

As part of this project, we demonstrated that the geographical localization of different strains could be exploited to determine the source of M. bovis as a result of farmers purchasing infected animals. For example, new breakdowns occurring in geographically separate and previously TB-free regions of the country could be linked by combining genotyping and cattle movement data to the movement of cattle from TB hotspot areas as a result of restocking herds following the FMD outbreak in 2002. This approach highlighted that the restocking of herds had resulted in the introduction of M. bovis isolates with spoligotypes that had not been seen in these areas of the country before. These data were used to underpin a policy of pre-movement testing of cattle to prevent movement of infected animals, especially those that are to be moved from areas of the country that are at high risk of M. bovis infection to areas of low risk for the disease.

These studies also demonstrated that the combined use of the State Veterinary Service’s TB99 epidemiological questionnaire and Vet Net (disease management database) TB recording data, the British Cattle Movement Service’s Cattle Tracing System (CTS) data, and the VLA’s molecular typing database provides a powerful method for the investigating the source of confirmed bovine tuberculosis breakdowns. However, detailed discussions with SVS on uses of molecular epidemiology highlighted that the development of a “point and click” Intranet application that could display all these data along with probability estimates of this spoligotype being detected – on the basis of past data – on this farm, would have a significant impact on implementing this approach. The development of statistical analysis for the complex spoligotyping dataset as part of this project along with other VLA initiatives including introduction of a web-based database system for project decision support and the use of XML-based technology to successfully display maps over the DEFRA intranet will facilitate the development of such a decision support system and following the development of a prototype system, a concept note for the development of this system was submitted to Defra in 2004.

In general, the Spoligotype and VNTR patterns obtained from badger isolates between 1972-1976 were the same as those observed in badgers and cattle in the same geographical areas today. This suggests that the geographical clustering of strains has not changed since the first isolation of M. bovis from badgers over thirty years ago. This data is in sharp contrast with the rapid movement of strains to geographical areas outside their normal range that was observed as a consequence of the restocking of cattle herds after the FMD epidemic of 2002.

Typing data was also used to support the ISG in interpreting findings from the Randomised Badger Culling Trial (RBCT). In one such study, M. bovis infections in badgers and cattle were shown to be spatially associated on a scale of 1-2km. Badgers and cattle infected with the same strain type of M. bovis were particularly closely correlated. These observational data support the hypothesis that transmission occurs between the two host species; however they could not be used to evaluate the relative importance of badger-cattle and cattle-badger transmission.

Epidemiological analysis designed to look at associations between spoligotypes and the epidemiological features of an outbreak was performed. Preliminary analysis revealed some intriguing differences across different molecular types of M. bovis. This analysis showed differences in numbers of inconclusive reactors across spoligotypes, and revealed that particular spoligotypes are more frequently detected on repeat testing. This suggests that clonal groups of M. bovis have distinct phenotypes that may be relevant to the control strategy.

Therefore the phenotypic differences between spoligotypes was investigated further. Fourier transform infra-red (FT-IR) spectroscopy was used to examine 100 blinded strains of M. bovis of diverse spoligotype. FT-IR is a rapid whole-organism fingerprinting method that generates a biochemical signature of the bacteria. Cluster analyses of the resulting spectra generated strain-groupings that closely mirrored the phylogeny generated from a combination of spoligotype and single nucleotide mutations. Hence, fingerprinting methods based on phenotype or genotype grouped the strains into similar clusters. These results indicate that clonal groupings of M. bovis share distinct phenotypic characteristics, possibly cell wall differences, that may result in differences in virulence, transmission or ability to evade the tuberculin skin test.

M. bovis AN5, a strain that was originally isolated in England circa 1948 and is used worldwide for bovine PPD production. However, the spoligotype and VNTR profile of this strain is not shared by any extant strains in the database. This raises the possibility that the AN5 strain may not be optimal for the detection infection by M. bovis strains currently prevalent in GB and highlights that further work should be done to resolve this matter.

SID 5 (2/05) Page 4 of 22

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 and Overview

Objective 01Type archival M. bovis isolates held at VLA-Weybridge by spoligotyping and VNTR to permit detection of rate of genetic change for these techniques over the past 20-30 years.

Molecular typing from Archival Samples– (01/01), Stock taking of all freeze dried and paraffin-embedded tissue samples was completed. Spoligotyping was performed by the method of Kamerbeek et al. (1997) and VNTR typing of these samples was performed using the standard VNTR (ETR) assays used at VLA (see Frothingham and Meeker-O’Connell, 1996 and Smith et al., 2003).

(01/04, 01/05) Spoligotyping was successfully carried out on 590 samples from between 1988 and 1990, including 134 badger isolates, 328 cattle isolates, and 116 isolates from deer. A further set of freeze-dried samples of M. bovis from badgers (1972-1976) was also subjected to spoligotyping. A total of 170 spoligotypes were obtained from this latter collection of isolates.(01/03, 01/06) A feasibility study of the suitability of archival material for VNTR analysis suggested that much of the material was not suitable for VNTR analysis. Further work was unable to overcome this technical limitation. Complete VNTR profiles are therefore only available for 113 of the 590 samples from 1988 to 90 and 73 VNTR profiles are available for the 170 isolates from badgers (1972-1976).

ResultsIn general, the Spoligotype and VNTR patterns obtained from badger isolates between 1972-1976 were the same as those observed in the same geographical areas today. This suggests that the geographical clustering of strains has not changed since the first isolation of M. bovis from badgers over thirty years ago. This strongly supports our population genetic analysis of M. bovis in GB (see Objective 05) that suggests that M. bovis in GB has undergone a series of clonal expansions in different areas of the country. This data is in sharp contrast with the rapid movement of strains to geographical areas outside their normal range that was observed as a consequence of the restocking of cattle herds after the FMD epidemic of 2002.

02.Development of spatial statistical procedures appropriate for M. bovis type clustering and multivariate diffusion models.

Currently there is no standard methodology to handle the statistical problem presented by the spoligotyping data-set, which essentially is multivariate count data with both first order (trend surface) and second order (spatial autocorrelation) effects. The aim of the spatial component of this project was to develop the techniques to enable such analyses to be undertaken. The progress made is described in the attached pdf.

See pdf: P.Diggle report.pdf.

03Undertake epidemiological analyses to identify isolates for further investigation and for validation of the use of SNP and phenotype analysis for local, ‘on farm’ epidemiological analysis.

SID 5 (2/05) Page 5 of 22

Introduction and overview.

All cultured strains of M. bovis are subject to routine surveillance using spoligotyping. The results are stored in a database along with all relevant epidemiological data including GIS reference, herd of origin, owner, visible lesions or not, eartag number etc. In a similar manner the results of VNTR typing are stored in the same database to facilitate comparisons between the two techniques. This data is regularly 'mined' to provide summaries of spoligotype prevalence and geographical localisation. Epidemiological analysis has been used to provide informal briefings to DEFRA, ISG, the SVS and the scientific community. The data has also been used in preliminary analyses to provide 'proof-of-principle' to guide future research. A number of examples of such analyses are described below. Furthermore, summaries, spoligotyping data and epidemiological analyses are made available to the SVS on an ad hoc basis in response to specific request. The entire database was made available to lead Veterinary Officers of the SVS in March 2004 for their own contact tracing and epidemiological analyses and is regularly updated.

03/01. Epidemiological analysis by cluster identification of spoligotype x VNTR combinations. In February 2002 the database was analysed to determine the most prevalent spoligotypes in GB and the results used to identify the ten most common spoligotype patterns in GB (see 04/03). These results were fundamental for deciding the strains in the '10X10' strain collection for SNP analysis of strains of the most common spoligotypes (Objective 04) as well as for phenotypic analysis of the ten most common spoligotypes in GB (Objective 06). Furthermore, the database was used to select 10 representative isolates of each of the most common spoligotypes that were geographically diverse with variable VNTR types. The database was also analysed for the distribution of VNTR types of the two most common spoligotypes found in GB (type 9 and 17) and for the geographical localisation of other spoligotypes. These data were used as the basis for the analysis of the population structure of M. bovis in GB (Objective 05).

03/02 By close liaison with ISG, isolates from areas within the field trial and past badger removal operations will be identified for further analysis by the most discriminating typing methods.Molecular typing data was supplied to ISG for all cattle and badger isolates from the Randomised Badger Culling Trial areas. In all 10 trial areas, one or more spoligotypes of M. bovis were found to infect both cattle and badgers. In eight trial areas, at least one spoligotype was found in cattle but not in the badgers sampled on these initial culls. In three trial areas, spoligotypes were found in badgers that were not detected in cattle during the period under consideration. The local spatial associations between spoligotypes detected in cattle and badgers were investigated by calculating, for each infected badger, the distance to the nearest bovine with the same spoligotype, and to the nearest bovine infected with a different spoligotype, and comparing these distances using a paired sign test. The ratio between these two distances was consistently less than one (medians: adults 0.40, cubs 0.51) indicating that badgers were closer to cattle infected with the same spoligotype as themselves (sign tests adults p<0.001, cubs p=0.031). Similar associations were found with infected cattle tested in the 12 months following the culls, and the pattern was consistent across trial areas. M. bovis infections in badgers and cattle were spatially associated on a scale of 1-2km. Badgers and cattle infected with the same strain type of M. bovis were particularly closely correlated. These observational data support the hypothesis that transmission occurs between the two host species; however they cannot be used to evaluate the relative importance of badger-cattle and cattle-badger transmission. These results along with the underlying analysis have been published recently (Woodroffe et al., 2005). Isolates from the culling trial areas are now being targeted for VNTR analysis.

03/03 By outlier identification: Extending the approach of 03/01, types will be mapped and spatial outliers identified.Molecular epidemiological data has demonstrated that distinct genotypes of M. bovis are clustered in different geographical regions and these clusters have increased in size locally over time (so called “clonal expansion” - see Objective 05). New breakdowns occurring in geographically separate and previously TB-free regions of the country could be linked by combining genotyping and cattle movement data to the movement of cattle from TB hotspot areas as a result of restocking herds following the FMD outbreak in 2002. This resulted in the introduction of M. bovis isolates with spoligotypes that had not been seen in these areas of the country before (ISG 4th Report; Gopal et al submitted). For example, for 17 herd breakdowns in North-eastern England, reactor animals were traced to source herds from which M. bovis was isolated with the same spoligotype-VNTR combination (Gopal et al, submitted). These studies demonstrated that the combined use of the State Veterinary Service’s TB99 epidemiological questionnaire and Vet Net (disease management database) TB recording data, the British Cattle Movement Service’s Cattle Tracing System (CTS) data, and the VLA’s molecular typing database provides a powerful method for the investigating the source of confirmed bovine tuberculosis breakdowns in low prevalence areas.

Spoligotyping data has also been provided to University of Warwick to support their analysis of the consequences of herd restocking following FMD (SE3026).

SID 5 (2/05) Page 6 of 22

03/04 Data from the population genetics – especially phylogenetic trees – will be mapped, and possible type diffusion identified. On the basis of this exploratory analysis, population genetic anomalies will be identified and further genetic and/or sequencing will be undertaken to explain these.To further understand the geographical localisation of clones (spoligotypes) in GB and as a preliminary exercise to apply advanced molecular epidemiological techniques in a local 'on farm' situation an epidemiological case study of bovine TB in East Sussex was carried out in 2004. Data from the spoligotype/VNTR database, the cattle tracing system (CTS) database, geographical information system (GIS) data and a badger sett density map were used to investigate and analyse the frequency and distribution of strains in a single county. East Sussex has been traditionally regarded as a 'hot-spot' for bovine tuberculosis in the south-eastern part of England, although adjacent counties are free of the disease (Wilesmith, 1983). It was shown that the region of isolation of M. bovis in E.Sussex is highly geographically localised in 11 parishes in the south of the county and that the majority of strains, isolated from both cattle and badgers, represent members of a single clone with a distinct spoligotype pattern (spoligotype 13), VNTR pattern and chromosomal deletion. This molecular type was easily differentiated from similar spoligotypes isolated in the Republic of Ireland and is a good example of the application of extra molecular epidemiological techniques (VNTR and deletion analysis) used to differentiate otherwise identical spoligotypes. It was also shown that strains sharing the E. Sussex molecular pattern (with similar spoligotype, VNTR, and unique deletion) but recovered outside East Sussex were isolated from infected animals exported from specific E. Sussex farms. These data provide basic parameters for the rate of transmission of bovine TB between cattle. It was also found that badger sett density correlated more closely with outbreaks of bovine tuberculosis in cattle in E. Sussex than cattle density and it was found that surface geology was an important determinant of badger sett density.

03/05 Bovine tuberculosis pathology will be quantified by combining lesion distribution within the body with degree of dissemination, and a rank scoring applied. Most attention will focus on multi-reactor herd breakdowns, with the intention of identify either low virulence – which might be potential vaccine candidates – and high virulence – against which vaccine will have to be tested. Rather than restricting analysis to the limited pathology data from TB50 forms, analysis designed to look at associations between spoligotypes and the epidemiological features of an outbreak was performed. Of particular interest was the potential relationship between spoligotype and epidemiological features of the disease. These features included:

1) IRs as a proportion of IRs & reactors 2) The proportion of unconfirmed incidents in the same 10 x 10 km square3) Proportion of incidents detected first in the slaughterhouse4) Duration of restrictions 5) Percentage of reactors with visible lesions6) Percentage of samples from which M. bovis isolated.

Preliminary analysis revealed some intriguing differences across different molecular types of M. bovis (Goodchild et al., 2003). This analysis showed differences in numbers of inconclusive reactors across spoligotypes, and revealed that particular spoligotypes are more frequently detected on repeat testing. This suggests that clonal groups of M. bovis may have distinct phenotypes that are relevant to the control strategy. If further analysis, incorporating VNTR data, supports these initial findings, this may have important implications for disease control since spoligotype may impact on how well the skin test performs and how severe a breakdown is likely to be. This would open up the possibility of tailoring control measures to fit the pathogenicity and immunogenicity of the molecular type of the M. bovis isolate involved. For example, for particular spoligotypes early use of the Bovigam assay could prove highly beneficial.

The phylogenetic tree emanating from the population genetic analysis of M. bovis in GB (see Objective 05) was used to identify isolates suitable for performing comparative virulence studies in experimentally infected cattle as part of project SE3030. In these studies no differences in virulence (as measured by pathology scores) were detected between a spoligotype 9 and a spoligotype 35 isolate.

04Develop Single Nucleotide Polymorphism (SNP) analysis for maximum differentiation of M. bovis stains at the genetic level.

(04/01) The M. bovis AF2122/97 genome annotation was carried out using the M. tuberculosis H37Rv genome annotation as a template All differences between the genomes were noted and tabulated. For preliminary studies ten genes were selected that had a frameshift, a truncation, or any other difference that could cause a potential loss in function of the gene in the AF2122/97 genome. These genes were selected because they should maximise the number of single nucleotide polymorphisms (SNP) or, more correctly, single nucleotide mutations (SNMs) available: functionless sequences are considered neutral to natural selection and purifying selection would not be capable of removing any SNMs. The local alignment program Blastn was run on the DNA sequences of the selected genes to identify SNMs. The SNMs included transversions, transitions, 1bp indels.

SID 5 (2/05) Page 7 of 22

At the time, the M. bovis BCG Pasteur genome was in the process of being sequenced as part of SE3206 and was 99% finished. However, the sequence was available in contig form with unclosed gaps and inaccuracies due to vector contamination, etc. Nevertheless, a comparison of the M. bovis AF2122/97 and incomplete M. bovis BCG Pasteur genomes was carried out. To maximise the chances of observing SNMs associated with phenotypic differences we initially concentrated on genes associated with housekeeping functions such as energy metabolism, amino acid biosynthesis and macromolecule metabolism. The DNA sequence of each housekeeping gene in the M. bovis AF2122/97 genome and its ortholog in the M. bovis BCG Pasteur genome were selected using ACT. The nucleotide sequence of each gene was obtained and an alignment program (clustalX) was used to identify SNMs. The differences were then manually checked and noted. Of the 248 housekeeping genes selected and aligned, only 48 genes showed any polymorphism between the M. bovis AF2122/97 and M. bovis BCG Pasteur genomes (all others were identical in nucleotide sequence).

(04/02) The SNapshot system was applied to the detection of 10 SNMs in 10 strains of M. bovis. However, this method only generated information on the nucleotide sequence at one specific position. We have found that sequencing across the SNM (about 250bp upstream and downstream) gives the benefit of checking that the correct position within the genome is analysed and has the added benefit that we gain more information on nucleotide polymorphism within the target region of the genome. This is therefore our method of choice for SNM analysis.

(04/03, 04/04) A test panel of M. bovis isolates was made up of 10 isolates of each of the 10 most frequently encountered spoligotypes of M. bovis (see 03/01 and 03/02 and Table 2). Using sequencing rather than SNapshot (see 04/02) we analysed 91 potentially polymorphic sites in 89 isolates representing the diversity of spoligotypes most frequently recovered in GB. For strains with spoligotypes 9, 11 10, 17, 25 and 35 ten isolates of each spoligotype were analysed, for spoligotypes 12 and 13 nine isolates were analysed and for strains of spoligotype 20 eight isolates were analysed. Analysis of strains with spoligotype 22 has been problematic and we currently have data on 30 regions in 10 isolates. 44 of the potentially polymorphic sites were found to be phylogenetically informative and are shown in Table 3.

From this analysis we have identified a series of polymorphisms that are consistent with the previously suggested phylogenetic history of the M. tuberculosis complex including M. bovis (Brosch et al. 2002). Furthermore, the single nucleotide mutations clearly show that, within strains of bovine adapted M. bovis recovered in GB, strains of spoligotype 20, 25 and 35 are distinct from all others analysed (strains of spoligotype 9 and its children – see Objective 05). These two groups of strains differ in at least 9 SNMs (Figure 1). This results is consistent with the phenotypic analysis (see 06/17) which also clustered strains of these spoligotypes away from strains of spoligotype 9 and its children. It also suggests that spoligotype 9 may have emerged in GB as a result of an evolutionary bottleneck. There was no evidence for recombination in this data set.

One site was polymorphic within spoligotype 9 and its descendants: ilvG. This polymorphism is interesting because it links all spoligotype 17 strains (the second most common spoligotype in GB) to a specific VNTR type of a spoligotype 9 strain. These data can be used to suggest that we have identified a possible ancestral type for all spoligotype 17 strains.

Figure 1. Combined spoligotype and single nucleotide mutation phylogeny for the 10 most common spoligotypes (and vaccine strain BCG) recovered in GB.

SID 5 (2/05) Page 8 of 22

05Undertake population genetic analyses of the spoligotypes to define temporal and spatial rates of change of M. bovis in cattle and wildlife populations in GB.

To determine the population structure of M. bovis in Great Britain the patterns of migration and selection in strains collected from cattle and badgers and characterized by two molecular epidemiological methods, spoligotype and polymorphism at VNTR loci, were analysed. In total 11,500 isolates of M. bovis from GB that have been characterised by spoligotype were used in the analysis. Genetic exchange between cells is rare or absent in the M. tuberculosis-complex (see also results from Objective 04 above) so that, using spoligotypes, it is possible to recognise 'clones', with a recent common ancestor. Using a series of simulations based on increasingly complex models, we showed that the distribution of VNTR types in the commonest clone in the data set, spoligotype 09, is incompatible with random mutation and drift. The most plausible explanation is that M. bovis in Great Britain has undergone a series of 'clonal expansions' either caused by the spread of a favourable mutation, together with all other genes present in the ancestral cell in which the mutation occurred or as a result of a 'founder effect' as a clone invades a previously uninfected area. This work was published in Proceedings of the National Academy of Sciences, USA (Smith et al., 2003 - attached). (05/02, 05/03). This interpretation is supported by the geographical localisation of different genotypes. The relationship between clonal expansion and geographical localisation was explored in more depth in Hewinson et al., (2005) and is described below.

At the international level it is becoming clear that many nation states tend to have a dominant, and apparently locally evolved, clone or clonal complex. The clonal group of strains from the Cameroon, for example, are all marked by the absence of spacer 30 in their spoligotype pattern (Njanpop-Lafourcade et al., 2001). Although this spoligotype pattern is not unique to the Cameroon these data suggest that the strains evolved in that country from a single common ancestor. Furthermore, even neighbouring countries with extensive trading links show marked differences in the population structure of the disease. Strains from The Republic of Ireland, the British mainland and France differ, both qualitatively and quantitatively, in the spoligotypes recovered (Costello et al., 1999; Haddad et al., 2001). The second most common spoligotype in GB, spoligotype 17 responsible for over 20% of isolates over a twenty-year period, was not recovered in a population genetic analysis of strains from the Republic of Ireland or from France. The most common spoligotype in France, identical to the spoligotype pattern of the vaccine strain BCG, is rarely if ever recovered in cattle in either mainland GB or the Republic of Ireland. These results suggest that molecular typing will be a useful tool to detect import of disease on an international scale and highlight the usefulness of the international database www.Mbovis.org. that is run by VLA and was established as part of this project. The aim of the database is to facilitate the international exchange of spoligotype pattern data and to provide an authoritative central store of spoligotype data. This database is accessed through www.Mbovis.org and is designed to record and give authoritative names to spoligotype patterns derived from the M. bovis lineage of strains. The website has been functional since July 2003, is fully searchable, and currently incorporates over 500 unique spoligotype patterns with ancillary data. The database is constantly being updated as new patterns are added and extra utilities developed to facilitate data recovery and analysis.

Within Great Britain the theme of geographical localisation of molecular type is even more extreme than between countries. Eleven of the most commonly recovered spoligotypes in GB are responsible for virtually all of the bovine TB (Table 4). For four of them, over 79% of the isolates are found in one region or county (Table 5). Another six, have between 66% and 93% of isolates located in two, adjacent, regions or counties (Table 5). No other area contains more than 10% of these ten spoligotypes. The exception to geographical localisation of spoligotype is the most common spoligotype recovered, spoligotype 9, which tends to be dispersed at reasonable frequency throughout the South West of GB. However, even for strains of spoligotype 9 there is evidence of extreme geographical localisation of genotype. The application of a second molecular epidemiological technique, VNTR typing, has revealed marked polymorphism within strains of spoligotype 9. More than 900 strains from individual outbreaks of bovine TB in cattle caused by spoligotype 9 have been analysed and over 20 different VNTR types recorded. The three most common VNTR types of spoligotype 9 in GB, representing over 60% of the isolates, are geographically localised (over 86% for each VNTR type) in one or two, adjacent, counties.

As described above, we have previously shown that the frequency of recovery of M. bovis genotypes, spoligotypes and especially VNTR types of spoligotype 09 strains, is not compatible with a simple evolutionary process of random mutation and drift when applied to the whole population (Smith et al., 2003). We concluded that the data could be explained by a series of clonal expansions in which individual genotypes increase disproportionately in frequency with a related loss of rare genotypes. Geographical localisation of genotype is not necessarily a consequence of clonal expansion. A fully mixed population could show clonal expansion at the population level but have very little geographical localisation. We have suggested that clonal expansion can be explained either by selection or by a population sampling effect. If, as is not unreasonable, the population in GB consists of a series of geographically isolated populations with little communication between them, then a selected genotype will fix only in a local population and the overall population structure will appear as a series of clonal expansions. On the other hand, if M. bovis in GB has been subjected to a severe bottleneck in population

SID 5 (2/05) Page 9 of 22

size, such as that caused by the cattle testing and eradication program in the 1950's and 60's, then clonal expansion may be more readily explained by population sampling effects. We have suggested that ecological opportunity (more properly ‘founder effect’), the exploitation of a new region or host by one or a few genotypes, as a principal cause of clonal expansion by population sampling of M. bovis in GB. However, drift in a series of small isolated populations may also play a part.

In reality probably both selection and population sampling effects are involved in generating clonal expansion and geographical localisation of M. bovis in GB. It is both possible and worthwhile to tease apart the relative contributions of these population level forces to the geographical localization of M. bovis in GB in order to target disease control strategies more effectively. This remains a challenge for future research.

As part of this project we also developed an algorithm for generating phylogenies from spoligotype patterns of the direct repeat (DR) locus of strains of M. bovis and have used the phylogenies generated by this algorithm to analyse the evolution of the DR locus. The algorithm assumes that repeat units within spoligotype patterns are lost and never gained and that there is no recombination between strains. Under these assumptions deletions of spoligotype spacer units are used as markers of clonal lineages and, remarkably, show the direction of evolution. The algorithm was applied to 38 spoligotype patterns recovered from over 11,000 strains of M. bovis isolated in Great Britain over a 30-year period. The phylogeny generated by the algorithm revealed two clades of M. bovis with distinct differences in host preference: a clade representing spoligotype patterns from strains primarily isolated from cows and badgers (main bovine lineage) and a second clade representing strains primarily isolated from llamas (llama/microti lineage). We assayed deletions of chromosomal DNA to independently confirm the separation of these two clades. Further analysis of the phylogeny shows that the DR locus evolves primarily by the loss of single spacers and that homoplasies are rare both between and within clades. The main bovine lineage was clustered around spoligotype 09, the spoligotype most frequently recovered from cattle strains in the UK (Figure 2). Spoligotype 17 (the second most commonly recovered spoligotype) was directly descended from type 09. As a generalisation the population structure of this group can be described as one major, common spoligotype - type 09 - with many single step descendants clustered around it. Included in these direct descendants is type 17 - which has risen to high frequency in the population and is now generating descendant spoligotypes. This phylogenetic analysis suggests that M. bovis encompasses a number of clades showing distinct host-preferences and that differences in host adaptation can be recognised by spoligotype patterns (05/01). This concept was expanded to include all members of the M. tuberculosis complex and to describe each of these clades as ecotypes and will be published in Smith et al (2005 -attached). The spoligotype phylogeny was presented as a hypothesis for testing and has been combined with the results of the SNP analysis (see 04 above) to generate an integrated phylogeny of M. bovis in GB (Figs 1 and 2). (05/04).

SID 5 (2/05) Page 10 of 22

Figure 2. Phylogeny for the most common spoligotypes of M. bovis recovered in GB (and AN5 the strain of M. bovis used for the production of bovine PPD). Spoligotype 9 is the progenitor strain of approximately 94% of all strains of M. bovis in GB. Daughter strains undergo clonal expansion in specific geographical areas. CG38 represents a hypothetical ancestral strain.

Population Structure of M. bovis in GB in relation to use of AN5 for Bovine Tuberculin production.

Prior to 1975 the M. tuberculosis strains DT, C and PN were used at the VLA for the production of PPD for bovine diagnosis (Pritchard, 1988). These strains are the same as those employed for human tuberculin production and they were used since the yield of bacilli from glycerol-containing media was greater than that obtained from M. bovis strains. However, since 1975 PPD production has switched to use M. bovis AN5, a strain that was originally isolated in England circa 1948 and which is used worldwide for bovine PPD production (Paterson, 1948). Its acceptance as a standard for tuberculin production was principally based on its high yield of cell mass on glycerinated-media, a phenotype that was selected by repeated subculture of the bacillus on laboratory media Keating et al., 2005). This selection for a desirable phenotype by passage through artificial media has parallels with the method used by Calmette and Guérin to attenuate a strain of M. bovis to generate BCG (Calmette, 1927). Using genomic technologies it has been shown that during this in vitro culture the genome of BCG suffered a number of gene deletions and chromosomal rearrangements, one deletion removing the potent antigens ESAT-6 and CFP-10 (Behr et al., 1999; Mahairas et al., 1996). Hence it is possible that the genome of M. bovis AN5 underwent similar events during in vitro passage that could have removed or altered the expression of genes encoding potent antigens.

Despite its widespread use as a diagnostic reagent, M. bovis AN5 is poorly defined. Recently we characterised the genome of M. bovis AN5 by using DNA microarrays and molecular typing technology (Inwald et al., 2003). Spoligotyping of M. bovis AN5 was performed and the resulting patterns compared to the VLA spoligotype database that currently holds typing information on ~32,000 M. bovis strains isolated from 1975 to May 2005 (with ~95% of data for strains isolated since 1993). From the results obtained with M. bovis AN5 it was clear that its profile was not shared by any strains in the database. Moreover, the spoligotype pattern for AN5 appears above 94% of all M. bovis strains extant in GB (Figure 2) raising a possibility that tuberculin made from AN5 has been successful at eradicating strains above AN5 in the M. bovis phylogeny, but less successful at removing strains below AN5 in this phylogeny. As with spoligotyping, the VNTR pattern for M. bovis AN5 shows no similarity to profiles from ~8350 typed M. bovis isolates from 1997 to date. This again raises the possibility that the AN5 strain may not be optimal for the detection of infection by M. bovis strains currently prevalent in GB. This hypothesis is supported by the results in Objective 03/05 of epidemiological analyses that showed differences in numbers of inconclusive reactors across spoligotypes, and revealed that particular spoligotypes are more frequently detected

SID 5 (2/05) Page 11 of 22

on repeat testing (Goodchild et al., 2003). This suggests that clonal groups of M. bovis have distinct phenotypes that are relevant to the control strategy. Further investigations into the implications of this work are being taken forward in ROAME SE3220.

06 Develop and validate phenotypic analysis for the differentiation of M. bovis isolates to improve strain typing for local epidemiological studies.

Introduction and Summary of Results.

The application of genetic fingerprinting techniques to Mycobacterium bovis has allowed us to define clonal groupings that are geographically distinct (see Objective 05, Figure 2 and Table 5). An obvious question that arises when examining such phylogenies is whether these clonal groups have associated phenotypic characteristics that may explain their relative success as pathogens. Some evidence to indicate that this might be the case is discussed briefly under objective 03/05 where preliminary epidemiological analysis showed differences in numbers of inconclusive reactors across spoligotypes, or revealed that particular spoligotypes are more frequently detected on repeat testing. This suggests that clonal groups of M. bovis have distinct phenotypes that are relevant to the control strategy for bovine TB in GB. To explore this question further we used Fourier transform infrared (FT-IR) spectroscopy to examine 100 blinded strains of M. bovis of diverse spoligotype. FT-IR is a rapid whole-organism fingerprinting method, which generates a biochemical signature of the bacteria. Cluster analyses of the resulting spectra generated strain-groupings that closely mirrored a phylogeny generated from a combination of spoligotype and single nucleotide mutations. Hence, fingerprinting methods based on phenotype or genotype grouped the strains into similar clusters. These results indicate that clonal groupings of M. bovis share distinct phenotypic characteristics, possibly cell wall differences that may result in differences in virulence or transmission.

Results

(06/01) To select and cultivate 25 different M. bovis strains of known genotypes with representatives of the M. tuberculosis complex and other mycobacteria from VLA. A subset to be grown in replicate to test the reproducibility of the methods.Seven rapid-growing mycobacteria (supplied by VLA) were used to optimise the three chosen analytical techniques (FT-IR, MALDI-MS and ESI-MS) in order to study the phenotype of the M. bovis strains. These were chosen rather than M. bovis because they could be cultured outside of VLA, thereby allowing in depth assessment. And providing proof of principle that the approach was valid. Three biological replicates of 100 different M. bovis strains (10 samples of 10 spoligotypes chosen to maximise genotypic and geographical diversity; see 03/01 and Table 1) were selected for further study and were cultured on 7H10 (supplemented with OADC and antifungals) and pasteurised.

(06/02) To pasteurise the above and check that no mycobacteria are viable samples prior to transportation to UWA.Following pasteurisation of the samples 10% of the total sample number were selected randomly and re-culturing was attempted to check viability. The cultures were incubated for 4 weeks and then checked for visible growth. Samples were only sent to UWA when no visible growth was observed in any of the samples.

Milestone 06/03 To acquire FT-IR spectra of all the bacterial samples (from material prepared in milestone 06/02), and to assess reproducibility of culturing.FT-IR spectra of the seven rapid-growing mycobacteria were acquired to check the reproducibility of the technique. The following experimental conditions were investigated;

The cells were pasteurised (heated at 80C for 1 h), the pasteurised cells were re-cultured to check for viability. Spectra were acquired for the pasteurised and unpasteurised cells,

The effect of different growth temperatures (30C and 37C), growth media (7H9 or 7H10 both supplemented with OADC) and growth rates (cells were cultured until late exponential phase), and

Three sample preparation methods were investigated (viz. slurried into physiological saline and analysed directly, vortexed with glass beads).

Spectra of the seven rapid mycobacteria were acquired with (1) our high throughput scanner accessory, (2) via microspectroscopy and (3) using a diamond HATR (horizontal attenuated total reflectance) accessory available for FT-IR spectrometers.FT-IR spectra were also acquired of the 100 samples of M. bovis from milestone 06/01.

SID 5 (2/05) Page 12 of 22

(06/04) To acquire ESI-MS spectra of all the bacterial samples (from material prepared in milestone 06/02), and to assess reproducibility of culturing.

Initial investigations concentrated on the seven rapid-growing mycobacteria. However, we were unable to acquire wholly reproducible spectra. This may be due to the thick mycolic acid layer on the surface of the mycobacteria suggesting that, at least at this stage, ESI-MS is not suitable for whole organism fingerprinting.

(06/05) To acquire MALDI-MS spectra of all the bacterial samples (from material prepared in milestone 06/02), and to assess reproducibility of culturing.MALDI-MS spectra of the seven rapid-growing mycobacteria were acquired to check the reproducibility of the technique. The following were performed;1. sample preparation techniques were evaluated for protein extraction, sonication conditions were optimised

and two different solvent combinations were tested (chloroform and methanol and trifluoroacetic acid (TFA) and acetonitrile (ACN)),

2. cells were pasteurised to compare with protein profiles of un-pasteurised cells, the growth conditions tested above were employed here, and

3. MALDI-MS spectra of the 100 samples of M. bovis were acquired using the most suitable method determined above. This was ACN with TFA and sonication of cells.

(06/06) To perform cluster analysis of the three spectral types (data from milestones 06/03-06/05) and evaluate discriminatory levels of each technique based on the known genotypes of the 25 M. bovis selected and those from the representatives of the M. tuberculosis complex and other mycobacteria.Cluster analysis of the spectra acquired from milestones 06/03 and 06/05 were performed on the 100 samples of M. bovis. This involved principal component (PC), discriminant function (DF) and hierarchical cluster analysis (HCA). PCA is an unsupervised analysis which is used to reduce the dimensionality of the multivariate data whilst preserving the variance, this is performed prior to DFA. DFA is a supervised learning technique, which discriminates between groups on the basis of the retained principal components with a priori knowledge of which spectra were replicates. This was based on the replicates of the individual samples. The DFA minimises ‘within group’ variance and maximises ‘between group’ variance. HCA was used to construct dendrograms from the a priori group centres in the DFA space using scaled Euclidean distances and the dendrograms are produced using average linked clustering algorithms. These calculations were conducted in Matlab using in-house code.The reproducibility of the cluster analyses was validated to optimise the number of principal components and discriminant functions selected in the analysis. In this process the PCA and DFA was performed on a subset of the data, termed the ‘training data’. The remaining data was used to test the analysis, whereby the ‘test data’ was projected into the PC and DF space, to ensure the replicates from the ‘test data’ clustered with those from the ‘training data’. Members of the M. tuberculosis complex and other mycobacteria were not included in this analysis.

(06/07) To perform various supervised machine learning methods to assess the ability of (a) FT-IR, (b) ESI-MS, and (c) MALDI-MS for the identification of 25 M. bovis strains.It became evident during the course of our investigations that since FT-IR produced clusters that related to the known genotype of the M. bovis that supervised learning methods (viz. neural networks or partial least squares) were not necessary.

(06/08) To enhance the ESI-MS process for measuring metabolome. To make and assess the efficiency of cell lysis by sonication for the mycobacteria. To perform tandem MS to identify some of the dominant / discriminatory metabolites. Due to the problems identified in Milestone 06/04 this task was not possible.

(06/09) To enhance the MALDI-MS process for measuring the proteome. To make and assess reproducibility of protein extracts by 2-D gel electrophoresis. This will allow discovery of which proteins are being detected from the direct analysis on the whole bacteria and which are discriminatory. This task to be continued throughout the project.Protein extracts analysed on the MALDI-MS were separated using 2-D gel electrophoresis (different pH ranges were employed). Results indicated the most suitable protein extraction method for the MALDI-MS technique was acetonitrile / trifluoroacetic acid (0.1%), used in a 1:1 ratio.

(06/10) To test reproducibility of the methods. To cultivate the same M. bovis strains from 06/01. Then to check viability.The reproducibility of the techniques was assessed using the seven rapid growing mycobacteria. Spectra of these bacteria were collected and used to calibrate a PC-DFA model. The bacteria were cultured again over a time period of one month and three months for projection into the PC-DFA model.The 100 samples of M. bovis were initially cultivated for the analysis, these were subsequently re-cultured.

SID 5 (2/05) Page 13 of 22

(06/11) To collect (a) FT-IR, (b) ESI-MS and (c) MALDI-MS data from 06/10.FT-IR spectra of the samples collected in milestone 06/10 were collected.

(06/12) To assess reproducibility of the cluster analysis and supervised machine learning methods from data collected 06/11 and from 06/03-06/05.The spectra collected from the rapidly growing mycobacteria (milestone 06/10) were split into three data sets depending on the collection time. As described in Milestone 06/10, PCA and DFA was performed on the initial data set, the data from the subsequent two sample points (1 and 4 months later) was then projected into the PC-DFA space to determine if the samples clustered with the original samples. Whilst not all of the seven bacteria clustered together, the results were promising and demonstrated that there was a reasonable level of reproducibility.

(06/13) To select and cultivate 100 different M. bovis strains. Then to check viability.A second batch of M. bovis samples were cultured (see Table 2). Three biological replicates of each sample were cultured and pasteurised (as described above). The viability of 10% of the total sample number was assessed before transportation from University of Aberystwyth (UWA) to the University of Manchester due to the move of the subcontractor to Manchester.

(06/14) To collect (a) FT-IR, (b) ESI-MS and (c) MALDI-MS from 06/13.The samples were analysed for FT-IR on a new FT-IR instrument (Bruker, Equinox 55) at the University of Manchester. The samples were analysed in transmission mode, this was not possible on the instrument at UWA.The samples were not re-analysed on the MALDI-MS because the 100 M. bovis samples had been analysed on the MALDI-TOF at UWA.

(06/15) To analyse the data produced from 06/14 by cluster analysis and supervised machine learning methods.Analysis was performed on the new data collected from Milestone 06/14 as detailed in Milestone 06/06 to generate the strain groupings shown in Figure 3.

(06/16) To analyse SNPs developed by VLA using MALDI-MS genotyping on DNA.To date the analysis of the SNPs (SMNs in Objective 04) has not been performed. This is because the samples (600bp) generated in objective 04 for SNP analysis were not suitable for analysis using MALDI-MS (the mass of these oligos is beyond the mass range of this or any MALDI-MS instrument).

(06/17) To compile finalised reports on findings and suggest which of the phenotypic typing methods, and which metabolites and proteins may be discriminatory for the different spoligotypes of M. bovis.The FT-IR analyses using two different collection modes proved to be the most reproducible and suitable method for whole organism fingerprinting. The results have been written up and a manuscript produced. The manuscript is attached to this report to serve as a technical annexe (SE3020obj06.pdf), it also gives detailed results for this section. For brevity the main findings are listed below:1. Spoligotypes 25 and 35 were the most different as judged by this phenotypic typing method (Figure 3).2. This phenotypic analysis also indicated homogeneity within the isolates of spoligotypes 10, 17, 20, and 22

(indicating limited genetic variation within these groups (Figure 3).3. There was a great deal of heterogeneity within isolates of spoligotype 9.4. There was limited homogeneity within isolates of spoligotypes 11, 12 and 13.5. Analysis of two dominant spoligotypes in GB (9 and 17), indicated tight clustering of type 17, whereas

spoloigotype 9 was dispersed through the PC-DFA space. These results were consistent with the findings of Smith et al. (2003; Objective 05) that the phylogeny of spoligotype 9 is complex consisting of 22 VNTR types and is the ancestral type of 85% of M. bovis isolates in GB.

6. Cluster analyses of the resulting spectra generated strain-groupings that closely mirrored a phylogeny generated from a combination of spoligotype and single nucleotide mutations (compare Figure 2 with Figure 3). Hence, fingerprinting methods based on phenotype or genotype grouped the strains into similar clusters. These results indicate that clonal groupings of M. bovis share distinct phenotypic characteristics, possibly cell wall differences that may result in differences in virulence or transmission.

7. Transmission generated data is potentially the most appropriate for the analysis due to the lack of baseline shift and simple optical interrogation of the sample by transmission rather than reflectance would indicate greater spectral reproducibility.

SID 5 (2/05) Page 14 of 22

07Integrate all information to provide recommendations to Defra and ISG on the use of existing and novel typing methods for molecular epidemiological studies.

A number of formal and ad hoc meetings were held with SVS, ISG and Defra to update and advise on the results emerging from the use and analysis of molecular epidemiology data. In addition the SVS have been supplied with all spoligotyping data which is updated regularly. This has recently been extended to VNTR data.

The key recommendations that have emerged as a result of our findings, meetings and discussions are as follows:

1. A combination of spoligotyping and VNTR typing should be applied to all isolates and that VNTR should be optimised for use in GB (validation of the optimum set of VNTR loci for GB isolates identified in SE3017 is required). Optimised VNTR targets should be evaluated on 1,000 isolates that are submitted for culture and typing.

2. A combination of cattle movement data (CTS) and molecular typing data should be used to determine the contribution of cattle movement to the spread of bovine tuberculosis in GB.

3. Local epidemiological studies utilising molecular typing, CTS, TB99 data and local knowledge, should be performed to further elucidate the mechanism of spread and persistence of M. bovis.

4. The reasons underlying clonal expansion of isolates in GB should be determined as a matter of urgency as this will serve to identify how to control the spread of M. bovis within GB.

5. Association between strains and their phenotypic characteristics should be investigated further especially with respect to the impact of these characteristics on diagnosis.

6. A combination of molecular typing and CTS provides a useful decision making tool for the SVS and a decision support tool should be developed to maximise use of these data (see below).

Concept notes have been submitted to Defra take the work listed above forward.

Provision of decision support systems for the SVS.Through discussions with the SVS a clear picture has emerged of the type of decision support system that should be developed. This system would comprise a “point and click” Intranet application showing maps and tabular data of reactor animals that have had M. bovis isolates spoligotyped and VNTR profiled. For each spoligotyped isolates, information would be available on the location of the farm on which the reactor animal was identified, the past movement history of the animal, and probability estimates of this spoligotype being detected – on the basis of past data – on this farm. The development of statistical analysis for the complex spoligotyping dataset as part of this project along with other VLA initiatives including introduction of a web-based database system for project decision support and the use of XML-based technology to successfully display maps over the DEFRA intranet (“Spida v1”), will facilitate the development of such a decision support system. Following the development of a prototype system that was demonstrated to the SVS, a concept note for the development of this system was submitted to Defra in 2004. The aims of this next phase of development were: (1) to develop linkages between the VLA spoligotyping database and the web-based mapping system (“Spida v1”) so as to be able to automatically map the occurrence of the predominant spoligotype and VNTR pattern (where available) on breakdown farms; (2) to develop linkages with the BCMS cattle traceability system (CTS) to thus provide data on the past movement of typed animals and, (3) to integrate spatial statistical routines which determine the conditional probabilities of the occurrence of the commonest spoligotypes within geographical areas. The concept note is attached (SB4013 05-06.doc).

SID 5 (2/05) Page 15 of 22

TablesTable 1 Sample reference and Spoligotype designation of the 10 by 10 collection (First culture batch).

Sample Spoligo Sample Spoligo Sample Spoligo Sample Spoligo Sample Spoligo38/01 9209/01 111935/89 133781/92 20157/01 25293/01 9800/01 111932/89 13522/90 201258/01 25501/01 93295/99 11351/99 13519/90 20681/01 25894/01 91249/01 113979/00 135768/99 20888/92 251198/01 94925/90 11223/00 131637/99 201495/01 255523/90 9180/01 111685/96 136855/98 202145/01 256524/89 96512/89 116875/98 132646/96 202451/01 251210/90 9468/01 11954/92 131670/98 202880/99 25393/01 92043/99 112265/89 1321-140/01 20384/01 253558/00 95415/00 111659/92 132084/01 202396/01 25                   1304/01 1054/01 121121/01 173048/98 221207/01 35430/89 102368/01 12214/01 17207/01 221286/99 356460/89 106427/89 12128/01 17477/01 22186/01 35658/89 10 74046994 12319/01 17927/01 22507/01 357432/88 101975/90 12359/01 171535/01 222255/90 351855/89 105696/99 12484/01 17928/91 224119/96 35272/01 105344/98 12810/01 17276/92 223274/97 3524/90 101867/96 121020/01 1721-643/01 2221-

9176/0035

2463/01 102208/01 123459/90 174438/98 221307/01 355448/00 102433/01 121943/89 17288/01 2221-284/01 35

SID 5 (2/05) Page 16 of 22

Table 2 Sample reference, Spoligotype and VNTR designation of the10 by 10 collection (Second culture batch).

SID 5 (2/05) Page 17 of 22

Sample Spoligo VNTR Sample Spoligo VNTR Sample Spoligo VNTR61/0038/01 9 6554*33.1 61/5415/00 11 7554*33.1 61/3781/92 20 7554*33.161/0276/92 9 6554*33.1 61/894/01 11 7554*33.1 61/5768/99 20 7554*33.161/0293/01 9 7555*32.1 74046994 12 7454*33.1 61/0207/01 22 7524*33.161/0393/01 9 7554*33.1 61/0054/01 12 7454*33.1 61/0288/01 22 7524*33.161/0501/01 9 7554*33.1 61/1975/90 12 7454*33.1 61/0477/01 22 7524*33.161/1198/01 9 6554*33.1 61/2208/01 12 7454*33.1 61/0927/01 22 7524*33.161/1286/99 9 7555*32.1 61/2368/01 12 7434*33.1 61/0928/91 22 7524*33.161/2043/99 9 7554*33.1 61/2433/01 12 7434*33.1 61/1535/01 22 7524*33.161/3558/00 9 7524*33.1 61/5696/99 12 7454*33.1 61/1867/96 22 7524*33.161/5344/98 9 7555*32.1 61/0223/00 13 7353*33.1 61/3048/98 22 7524*33.161/5523/90 9 7554*33.1 61/0954/00 13 7353*33.1 61/4438/98 22 7524*33.161/6855/98 9 7555*32.1 61/1659/92 13 7353*33.1 61/0157/01 25 6554*23.161/0024/90 10 7554*33.1 61/1932/89 13 7356*33.1 61/0384/01 25 6554*23.161/0272/01 10 7554*33.1 61/1935/89 13 7356*33.1 61/0681/01 25 6554*23.161/0430/89 10 7554*33.1 61/2265/89 13 7353*33.1 61/0888/92 25 6554*23.161/0658/89 10 7554*33.1 61/3979/00 13 7353*33.1 61/1258/01 25 6554*23.161/1210/90 10 7554*33.1 61/0128/01 17 7555*33.1 61/1495/01 25 6454*23.161/1304/01 10 6554*33.1 61/0214/01 17 7555*33.1 61/2145/01 25 6554*23.161/1855/89 10 7554*33.1 61/0319/01 17 7555*33.1 61/2396/01 25 6554*23.161/2463/01 10 7554*33.1 61/0351/99 17 7555*33.1 61/2451/01 25 6554*23.161/5448/00 10 7554*33.1 61/0359/01 17 7555*33.1 61/2880/99 25 6554*23.161/6427/89 10 7554*33.1 61/0484/01 17 7555*33.1 61/6875/98 25 6554*23.161/6460/89 10 7554*33.1 61/0519/90 17 7555*33.1 21/0284/01 35 3354*33.161/6524/89 10 7554*33.1 61/0522/90 17 7555*33.1 21/0643/01 35 3354*33.161/7432/88 10 7554*33.1 61/0810/01 17 7555*33.1 21/9176/00 35 3354*33.161/0180/01 11 7554*33.1 61/1020/01 17 7555*33.1 61/0186/01 35 3354*33.161/0209/01 11 7554*33.1 61/1121/01 17 7555*33.1 61/0507/01 35 3354*33.161/0468/01 11 7554*33.1 61/1943/89 17 7555*33.1 61/1207/01 35 3354*33.161/0800/01 11 7554*33.1 61/3459/90 17 7555*33.1 61/1307/01 35 3354*33.161/0894/01 11 7554*33.1 21/0140/01 20 7554*33.1 61/2255/90 35 3534*33.161/1249/01 11 7554*33.1 21/140/01 20 7554*33.1 61/3274/97 35 3354*33.161/1685/96 11 7554*33.1 61/1637/99 20 7354*33.1 61/4119/96 35 3354*33.161/3295/99 11 7554*33.1 61/2084/01 20 7554*33.1 61/6512/89 (11?) xxxx2xx61/4925/90 11 9554*33.1 61/2645/96 20 7554*33.1 61/1670/98 (20?) 7555*33.1

Table 3. Phylogenetically informative single nucleotide mutations in strains of the 10 by 10 collection.

Please note that due to intellectual property issues this table was removed at the request of the contractor prior to Defra publication of this report. Reference to it is made in the body of the report.

As a substitute the contractor has provided the figure below:

SID 5 (2/05) Page 18 of 22

Table 4. Frequency of spoligotypes in GB.

Spoligotype patterns Frequency of isolation a

           

International Name b

VLA name

Number of breakdowns % of total

       

SB0140 9 3532 35.90SB0263 17 2322 23.60SB0274 11 1278 12.99SB0673 22 612 6.22SB0129 25 608 6.18SB0275 15 325 3.30SB0272 10 304 3.09SB0271 12 200 2.03SB0134 35 191 1.94SB0145 20 186 1.89SB0130 21 162 1.65SB0273 13 50 0.51SB0957 81 10 0.10SB0142 51 9 0.09SB0139 75 7 0.07SB0813 73 5 0.05SB1072 89 5 0.05SB1073 90 5 0.05SB0663 43 4 0.04SB0668 47 3 0.03SB0054 65 3 0.03SB0959 74 3 0.03SB0666 59 2 0.02SB0486 91 2 0.02SB0672 36 1 0.01SB0345 37 1 0.01SB0675 40 1 0.01SB0678 45 1 0.01SB0662 49 1 0.01SB0141 63 1 0.01SB0146 79 1 0.01SB1074 84 1 0.01SB1075 87 1 0.01SB1076 88 1 0.01

Total 9839 100.00

             

a The frequency of each spoligotype recovered from cattle was analysed in March 2005 from a database recording all spoligotypes of cultured strains of M. bovis since 1987 (Spoligotype Database, VLA Weybridge UK). The majority of the data was collected since 1997. Only one isolate was included from each herd breakdown of bovine tuberculosis.

b Assigned by www.Mbovis.org

SID 5 (2/05) Page 19 of 22

Table 5. Geographical localisation of the eleven most frequently recovered spoligotypes in GB.

                   

Spoligotype Areas that contain more than 10% of the total isolates of each spoligotype

               

% area % area % area

9 23 Cornwall 21 Devon 20 Dyfed

17 33Here. and Worc. a 33

Gloucestershire

11 71 Devon 22 Somerset22 51 Gwent 33 Here. and Worc.25 59 Staffordshire 20 Derbyshire15 89 Cornwall

10 79Gloucestershire

12 94 Cornwall

35 48Here. and Worc. 29 Shropshire

20 95 Cornwall21 41 Somerset 33 Avon                   

a Herefordshire and Worcestershire.

SID 5 (2/05) Page 20 of 22

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.

1. Brosch, R., S. V. Gordon, et al. 2002. A new evolutionary scenario for the Mycobacterium tuberculosis complex. Proc Natl Acad Sci U S A 996: 3684-9.

2. Behr M.A., Wilson, M.A., Gill, W.P., Salamon, H., Schoolnik, G.K., Rane, S., Small, P.M., 1999, Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science 284, 1520-1523.

3. Calmette, A., 1927, La Vaccination Preventive Contre la Tuberculose. Masson et Cie, Paris.4. Costello, E., O'Grady, D., Flynn, O., O'Brien, R., Rogers, M., Quigley, F., Egan, J., Griffin, J., 1999, Study of

restriction fragment length polymorphism analysis and spoligotyping for epidemiological investigation of Mycobacterium bovis infection. J Clin Microbiol 37, 3217-3222.

5. Diggle PJ, Zheng P, Durr P. 2005. Non-parametric estimation of spatial segregation in a multivariate point process. Applied Statistics. 54, 645-58.

6. Frothingham, R., Meeker-O'Connell, W.A., 1998, Genetic diversity in the Mycobacterium tuberculosis complex based on variable numbers of tandem DNA repeats. Microbiology 144 (Pt 5), 1189-1196.

7. Goodchild, A.V., De La Rua Domenech, R., Palmer, S., Dale, J., Gordon S.V., Hewinson, R.G., Clifton-Hadley, R.S., 2003. Association between molecular type and the epidemiological features of Mycobacterium bovis in cattle. In: Proceeding of the Society for Veterinary Epidemiology and Preventive Medicine, Warwick 31 March - 2 April 2003.

8. Haddad, N., Ostyn, A., Karoui, C., Masselot, M., Thorel, M.F., Hughes, S.L., Inwald, J., Hewinson, R.G., Durand, B., 2001, Spoligotype diversity of Mycobacterium bovis strains isolated in France from 1979 to 2000. J Clin Microbiol 39, 3623-3632.

9. Hewinson, RG, Smith NH, Vordermeier, HM, Gordon, SV. 2005. Recent advances in our knowledge of Mycobacterium bovis: a feeling for the organism. Veterinary Microbiology. In Press.

10. Inwald, J., Hinds, J., Palmer, S., Dale, J., Butcher, P.D., Hewinson, R.G., Gordon, S.V., 2003, Genomic Analysis of Mycobacterium tuberculosis Complex Strains Used for Production of Purified Protein Derivative. J Clin Microbiol 41, 3929-3932.

11. Kamerbeek, J., Schouls, L., Kolk, A., van Agterveld, M., van Soolingen, D., Kuijper, S., Bunschoten, A., Molhuizen, H., Shaw, R., Goyal, M., van Embden, J., 1997, Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 35, 907-914.

12. Keating, LA, Wheeler, PR, Mansoor, H, Inwald JK, Dale, J, Hewinson, RG, Gordon SV. The pyruvate requirement of some members of the Mycobacterium tuberculosis complex is due to an inactive pyruvate kinase: implications for in vivo growth . Molecular Microbiology 56 (1), 163-174.

13. Mahairas, G.G., Sabo, P.J., Hickey, M.J., Singh, D.C., Stover, C.K., 1996, Molecular analysis of genetic differences between Mycobacterium bovis BCG and virulent M. bovis. J Bacteriol 178, 1274-1282.

14. Paterson, A.B., 1948, The production of bovine tuberculoprotein. J. Comp. Path. 58, 302-313.15. Pritchard, D.G., 1988, A century of bovine tuberculosis 1888-1988: conquest and controversy. J Comp Pathol 99,

357-399.16. Njanpop-Lafourcade, B.M., Inwald, J., Ostyn, A., Durand, B., Hughes, S., Thorel, M.F., Hewinson, G., Haddad,

N., 2001, Molecular typing of Mycobacterium bovis isolates from Cameroon. J Clin Microbiol 39, 222-227.17. Smith N. H., J. Dale, J. Inwald, S. Palmer, S. V. Gordon , R. G. Hewinson, and J. Maynard Smith. The

population structure of Mycobacterium bovis in Great Britain: clonal expansion. Proc Natl Acad Sci U S A. 2003. 100:15271-5.

18. Smith N. H., K. Kremer, J. Inwald, J. Dale, J. R. Driscoll, S. V. Gordon, D. van Soolingen, R. G. Hewinson, and J. Maynard Smith. 2005 Ecotypes of the Mycobacterium tuberculosis complex. J. Theo. Biol (In Press).

19. Smith N. H., Kremer, K., J. Inwald, J., Dale, J., Driscoll, J.R., Gordon, S.V., van Soolingen, D., Hewinson, R.G., Maynard Smith, J., 2005, Ecotypes of the Mycobacterium tuberculosis complex. J. Theo. Biol (In Press).

20. Vaidyanathan, S. Winder, C. L., Wade, S. C., Kell, D. B., Goodacre, R. Sample preparation in matrix-assisted laser desorption ionization mass spectrometry of whole bacterial cells and the detection of high mass (> 20 kDa) proteins. 2002. Rapid Communications in Mass Spectrometry 16 (13), 1276-1286

21. Wilesmith JW (1983) Epidemiological features of bovine tuberculosis in cattle herds in Great Britain. J. Hyg., Camb. 90, 159-176.

22. Winder C.L., Carr, E., Goodacre R., Seviour R. The rapid identification of Acinetobacter species using Fourier transform infrared spectroscopy. 2004. Journal of Applied Microbiology 96, 328-330.

23. Winder CL and Goodacre R (2004). Comparison of diffuse-reflectance absorbance and attenuated total reflectance FT-IR for the discrimination of bacteria. The Analyst 129, 1118-1122.

24. Woodroffe R, Donnelly, CA, Johnston, WT, Bourne, FJ, Cheeseman, CL, Clifton-Hadley, RS, Cox DR, Gettingby, G, Hewinson, RG, LeFevre, AM, McInerney, JP, Morrison, IW. 2005. Spatial association of Mycobacterium bovis infection in cattle and badgers (Meles meles). Journal of Applied Ecology 42, 852-862.

25. www.Mbovis.org

References in bold have arisen out of this project. In addition a Review article has been requested by Nature Reviews in Microbiology. This manuscript is in preparation.

SID 5 (2/05) Page 21 of 22

SID 5 (2/05) Page 22 of 22