General enquiries on this form should be made to:Defra, Science Directorate, Management Support and Finance Team,Telephone No. 020 7238 1612E-mail: [email protected]
SID 5 Research Project Final Report
SID 5 (2/05) Page 1 of 19
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 OZ0405
2. Project title
Genotypic and phenotypic comparison of Y. enterocolitica from humans and animals
3. Contractororganisation(s)
Department of Food and Environmental Safety, Veterinary Laboratories Agency Department of Infectious and Tropical Diseases, London School of Hygiene and Tropical Medicine
54. Total Defra project costs £ 260,367
5. Project: start date................ 01 May 2002
end date................. 30 April 2005
SID 5 (2/05) Page 2 of 19
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.
Yersinia enterocolitica is a Gram negative bacterium which can cause diarrhoea and related, more serious, diseases in humans. This infection is believed to be foodborne. The 1999-2000 Defra national abattoir survey indicate that 6% of cattle, 13% of sheep and 26% of pigs were colonised with this potential pathogen. However, the role of veterinary strains in human disease is currently unknown.
Strains of Y. enterocolitica are defined on the basis of biotype (BT 1-5) and serotype. Biotypes 2-5 are generally considered pathogenic, on the basis of the presence of the major, well-characterised virulence determinants, including the plasmid pYV, and correlation with pathogenicity in a mouse infection model. Strains of BT 1a conversely do not possess these virulence determinants and have no pathogenicity in mice and are consequently considered non-pathogenic. Strains of BT1b are lethal in the mouse model and contain all the virulence determinants known but also have an additional high pathogenicity island. With the aim of determining the relationship between human and animal strains, phenotypic comparison using the standard bio-serotype approach was In collaboration with Dr T Chesty (HPA), initial characterisation of the strains by biotype and serotype showed that 58% of the animal isolates and 53% of the human isolates were Biotype (BT) 1a. The main recognised pathogenic Y. enterocolitica biotype isolated from livestock was BT3 (O:5,27) (35% of sheep, 22% of pigs and 4% of cattle) but this biotype was not detected in any of the human isolates investigated. The major pathogenic biotypes of strains isolated from humans were BT3 (O:9) (24%) and BT4 (O:3) (19%) whereas of the veterinary isolates investigated, only pigs (11%) carried BT3 (O:9) strains.
To further investigate the relationship between human and animal isolates, 88 strains were selected and further characterised by genotyping. In collaboration with Dr S. On, DFVF, Denmark, the technique of fluorescent Amplified Fragment Length Polymorphism (AFLP), was successfully established and used to genotype other enteric pathogens. AFLP primarily distinguished Y. enterocolitica strains according to their biotype, with
SID 5 (2/05) Page 3 of 19
strains separated into two distinct clusters; cluster A comprising largely of the putatively pathogenic biotypes (BT 2-4) and cluster B comprising of the putatively non-pathogenic biotype 1A strains and a single BT 1B isolate. Within these two genotypic clusters, sub-clusters were observed and these largely were serotype-associated. Similarity of AFLP profiles in BT3 (O:9) strains from pigs and humans appeared to confirm that pigs are a major source of human infection with this putatively pathogenic biotype. Evidence was also obtained that sheep may be a reservoir for human infections with BT4 (O:3). However, the results also suggested that some strains causing human disease do not come from veterinary sources identifiable at this time.
As part of a collaborative venture, a PhD student, under the supervison of Prof Brendan Wren, LSHTM, genotyped the 88 selected strains using a pan-Yersinia micro-array. The Pan-Yersinia species array included all sequences that were present in the sequenced strains Y. pestis CO92, Y. pestis KIM and Y. enterocolitica 8081 1b. In addition, unique sequences from Y. enterocolitica and Y. pseudotuberculosis strains, which were available from collaborators and had been deposited in Genbank, were also included (e.g. the O-antigen locus of Y. enterocolitica serotypes O: 3 and O: 9). From this data, the phylogeny revealed two distinct clades; a “pathogenic” clade primarily comprising pathogenic biotypes from humans, cattle, pigs and sheep and a “non-pathogenic” clade primarily comprising the non pathogenic biotype 1a human, cattle, sheep and pig isolates. Within the “pathogenic” clade the data split into two distinct groups, caused by the presence or absence of the virulence plasmid pYV, which was confirmed using GeneSpring. Loss of the virulence plasmid alone did not account for the clear divide of the pathogenic and non-pathogenic strains from each other and this strongly indicates that there are large chromosomal differences between the two groups. Interestingly, the basis for formation of subgroups via array analysis was different from the AFLP analysis. AFLP subgrouped isolates according to serotype, though there was a difference between human and porcine BT4 (O:3) isolates. Microarray analysis subgrouped the strains on an, as yet, unidentified genetic basis, with the human and porcine BT4 (O:3) isolates forming a statistically significant link. The ability of array analysis to compare the presence and absence of individual genes will allow the future identification of genetic markers responsible for the groupings.
In order to compare the isolates on the basis of phenotypic properties related to virulence, the 88 selected strains were assayed for their ability to adhere to, and invade, cultured epithelial cells in vitro. All of the isolates were capable of high levels of adhesion to epithelial cells, regardless of biotype or source. All isolates tested, excepting 3 non-motile BT1a strains, were also capable of invading HEp-2 cells. In general, BT1a strains were approximately 100-fold less invasive than the pathogenic biotype strains, but there was no correlation between source and invasive ability. To further the phenotypic comparison, the ability was assessed of the strains to survive within U937 human macrophages cells in vitro, and the levels of cytokines secreted by infected cells. All of the strains tested were capable of surviving within macrophages to the same degree. There were differences in cytokine secretion induced by challenge with the strains. BT3 (O:5,27) strains displayed increased secretion of the pro-inflammatory cytokines IL-6 and IL-8. The BT1a isolates were capable of modulating cytokine secretion in that the levels of IL-6 and IL-8 expressed were similar to those observed for the pathogenic biotypes. However, challenge with BT1a strains resulted in a large increase in secretion of the pro-inflammatory cytokine TNF
Investigation of the molecular basis for virulence requires the use of defined mutants but the production of such mutants in Y. enterocolitica has been problematic. Both the VLA and LSTHM laboratories developed successful methodologies to solve this. At the LSTHM, two genes were identified from the microarray analysis with potential roles in
SID 5 (2/05) Page 4 of 19
the pathogenicity of BT 1b strains. Using a Red recombinase system of mutagenesis these two genes ( a two component regulatory system gene and an Intimin like gene) were mutated and shown to play a role in motility and invasion in vitro. These mutants will now be tested in a mouse model. At the VLA the molecular basis of the non-invasiveness of the non-motile BT1a isolates was further investigated. Using the pKNG suicide vector method, the regulator of flagellin expression in strain 53/03 was successfully deleted. This is the first report of a mutant being constructed in a BT1a strain. The data obtained suggested that flagella are involved in invasion, but not adherence, of BT1a Y. enterocolitica, as well as survival within macrophages. The flagella were also immunomodulatory, and appeared to play a role in down-regulating TNF production and up-regulating the anti-inflammatory cytokine IL-10.
Although not an initial objective of this project the role of serum antibodies in livestock infections was investigated using serum from cattle and pigs experimentally infected as part of a previous Brucella project. These results showed that antibodies directed against a number of antigens of Y. enterocolitica, including flagellin and Yop virulence factors, protected against invasion of epithelial cells in vitro.
The findings of OZ0405 highlight a wide range of options for future work. The microarray data demonstrated that the presence of the pYV plasmid alone is not responsible for virulence. The data requires further analysis to determine whether there are any CDSs that are specific to human isolates. In addition, many new genetic markers have been identified which can distinguish Y. enterocolitica subgroups. Some of these may relate to phenotypic traits. Interestingly the results suggest that host response (immunopathology) is an important factor in the outcome of infection with the various biotypes. Clearly this requires further investigation in order to accurately determine the public health risk of Y. enterocolitica strains in the human food chain.
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).
The overall aimf of OZ0405 was to phenotypically and genotypically compare Y. enterocolitica isolates from humans and animals, in order to investigate the relationship between strains present in livestock and strains causing disease in humans.
Objective 1: To subtype cattle sheep and pig strains of Y. enterocolitica for comparison with human strains. Strains of Y. enterocolitica are defined on the basis of biotype (BT 1-5) and serotype. Biotypes 2-5 are generally considered pathogenic, on the basis of the presence of the
SID 5 (2/05) Page 5 of 19
major, well-characterised virulence determinants, including the plasmid pYV, and correlation with pathogenicity in a mouse infection model (2). Strains of BT 1a conversely do not possess these virulence determinants and have no pathogenicity in mice and are consequently considered non-pathogenic (2). Strains of BT1b are lethal in the mouse model and contain all the virulence determinants known but also have an additional high pathogenicity island (2). However, the correlation between the mouse model and human disease is unclear so the relevance of biotype to human virulence is remains unknown.
To investigate the relationship between livestock carriage of Y. enterocolitica and human disease, the biotypes/serotypes of strains recovered from the faeces of pigs, cattle and sheep at slaughter during a national abattoir survey in Great Britain in 1999-2000, were compared to those strains isolated from human cases of diarrhoea during the same period. This work was undertaken in collaboration with the Health Protection Agency (HPA) and overall 339 strains were investigated in this study. The detailed methods and results are published and the publication is given in Appendix 1.
The faecal carriage of Y. enterocolitica by cattle, sheep and pigs at slaughter was 6.3%, 10.7% and 26.1% respectively. Y. enterocolitica biotype (BT) 1a was the most frequently isolated biotype from livestock (58% of all strains isolated). Despite being putatively non-pathogenic this was the predominant biotype isolated from human cases (53%) over the same period. The predominant putatively pathogenic Y. enterocolitica biotype isolated from livestock was BT3 (O:5,27) (35% of sheep, 22% of pig and 4% of cattle isolates) but as this biotype was not detected in any of the human isolates investigated, this did not appear to be a reservoir for human infection. The major pathogenic biotypes of strains isolated from humans were BT3 (O:9) (24% of isolates) and BT4 (O:3) (19% of isolates). In contrast BT4 (O:3) strains were isolated from pigs (5% of isolates) and sheep (1% of isolates), but only pigs carried BT3 (O:9) strains (11% of isolates). Although this data suggests that pigs may be the primary reservoir for human pathogenic Y. enterocolitica infection, because of significant overlaps in phenotypes of the veterinary and human strains it was not possible to draw accurate conclusions about the role of these veterinary hosts as reservoirs of human infection. However, further investigations using methods with greater discriminatory power were considered likely to provide this information.
Amplified Fragment Length Polymorphism (AFLP) is a generic and highly discriminatory PCR-based molecular typing method which randomly samples the whole of the genome (10). This technique has been applied to many bacterial pathogens and has proved a highly valuable tool for enteric pathogens such as campylobacter (10). Initial work on the development of the AFLP technique was undertaken in collaboration with Dr S On (DFVF, Denmark). The detailed results of this work are in press and the manuscript is given in Appendix 2.
For AFLP investigations, a panel of 88 strains were selected to provide a range of biotypes, serotypes and host sources. This panel comprised 61 animal (35 pig, 9 cattle and 17 sheep) and 27 human isolates The strains illustrated a wide diversity in AFLP patterns but clustered into two broad groups; Clusters A and B. Cluster A primarily comprised strains of biotypes 2, 3 and 4, while cluster B only contained biotype 1 strains (Fig 1, appendix 2). The percentage similarity between these two clusters was only 41.15%. However, isolates within Cluster A ranged from between 66.8% to 95.2% similarity, while isolates within cluster B ranged from 68.6% to 83.4% similarity. Within Clusters A and B, sub-clusters were also observed. These sub-clusters were associated with phenotypic properties. Cluster A divided into 6 sub-clusters (A1 - A6). Sub-cluster A1 comprised eleven O:5,27 strains (nine isolates were BT 3 and two were BT4) and
SID 5 (2/05) Page 6 of 19
four strains that were non-serotypable but were all BT3. Isolates representing sub-cluster A2 were all serotype O:3 (BT4). Sub-cluster A3 was slightly more phenotypically heterogeneous with fourteen serotype O:9 isolates (12 of which were BT3 and two BT2), one serotype O:3 (BT4) isolate and two BT1A isolates of serotypes O:6,30 and O:5. Sub-cluster A4 comprised isolates of all the same phenotype (O:3 (BT4)) except one that was serotype O:6,30 (BT1a). Sub-clusters A5 and A6 contained only one strain each of serotype O:3 (BT4) and serotype O:19,8 (BT1a) respectively. Cluster B contained twenty one biotype 1a strains of varying serotypes. The remaining isolate was BT1b (serotype O:19). Cluster B also subdivided into 6 sub-clusters (B1- B6). Subcluster B1 contained five isolates of serotype O:5, one isolate of serotype O:6,30 and one non-serotypable isolate. Sub-cluster B2 contained four strains of serotype O:19,8 and one of isolate O:4,32. All but one isolate of sub-cluster B3 were serotype O:6,30. Sub-clusters B4 and B5 contained only 1 isolate each, of serotypes O:6,30 and O:5 respectively. Finally sub-cluster B6 contained the only isolate of biotype 1b (O:19) in the collection.
There were no clear associations of host source with AFLP patterns. Interestingly 6 of the 7 human BT4 O:3 strains were clustered together in sub-cluster A4 and had distinctly different patterns from the cluster of 4 pig BT4 O:3 strains in sub-cluster A2 suggesting that these strains were unrelated. In contrast the human BT3 O:9 strains in the collection clustered closely with the pig strains of the same phenotype confirming that pigs were the most likely reservoir of this infection in humans.
In conclusion investigation of the relationship between phenotype, using biotype/serotype, and genotype, using AFLP, and host source confirmed that pigs are potential sources of human infection at least with the putatively pathogenic strains of BT3 O:9. However, the results of comparisons of other biotypes is not so clear, though preliminary information suggests that sheep may be a source of human infection with BT4 O:3 strains. The sources of human infection with some putatively pathogenic biotypes/serotypes remains unknown but do not appear come from veterinary sources identifiable at this time. The role of BT 1a strains in human disease remains a conundrum. The prevalence of such isolates within livestock is clearly high and these studies suggest that such strains are associated with disease in humans despite the absence of known virulence determinants. Nevertheless, the identification of some BT1a strains within cluster A raises questions about the diversity of strains in this phenotype and the reflection of this on virulence potential.
Objective 2: To use DNA microarrays to investigate the genomic diversity of Y.enterocolitica from humans and animals and to identify and characterise up to 4 putative novel virulence determinants.
Over recent years the availability of whole genome sequences has enabled the direct comparison of the genomic content of bacterial strains using microarray technology.In 2001 the genome sequence of Yersinia pestis strain CO92 became available and was used to generate a DNA microarray for post-genomic studies. Subsequently in 2002, the sequence of Y. enterocolitica 8081 (a BT 1b serotype O:8 strain) was completed by The Sanger Centre. The selection of this strain for sequencing was important because traditionally Y. enterocolitica have been grouped into non-pathogenic BT 1a strains, pathogenic European strains and highly pathogenic American strains (generally BT 1b) of which strain 8081 is an example. It was hypothesised that this BT 1b would contain the genetic elements responsible for its virulence and would therefore be a highly useful comparator for microarray analysis. Microarray technology is an extremely effective method for whole genome analysis of bacterial isolates and was therefore used to compare the Y. enterocolitica isolates from livestock and humans.
SID 5 (2/05) Page 7 of 19
In this study the existing Yersinia pestis CO92 microarray was updated to generate a Pan-Yersinia species array to include all sequences that are present in the sequenced strains Y. pestis CO92, Y. pestis KIM and Y. enterocolitica 8081 1b. In addition, unique sequences from Y. enterocolitica and Y. pseudotuberculosis strains, which were available from collaborators and had been deposited in Genbank, were also included (e.g. the O-antigen locus of Y. enterocolitica serotypes O: 3 and O: 9 ) (Pan Yersinia v3.0 array).
The 88 Y. enterocolitica strains selected for previous AFLP studies in Objective 1 were typed using the pan-Yersinia microarray (genomotyping). These microarray studies were undertaken, by Sarah Howard at LSTHM, as a part of a PhD studentship funded in part from this project. These studies identified 1442 predicted coding sequences (CDSs) that were common to all of the 88 UK strains (core genes). Many of the CDSs within this list are necessary for a bacterial cell to perform all essential housekeeping functions including biosynthetic, cellular, metabolic and regulatory process.
A total of 227 chromosomal predicted CDSs were absent or highly divergent in all test strains but present in the highly virulent genome sequence strain Y. enterocolitica 8081 1b (Fig 2). These coding sequences may contribute to virulence and could potentially be important in the pathogenicity of 1B biotypes and are prime candidates for further investigation.
Recently Prof Wren’s group has developed new approaches to optimise the analysis of data from whole genome DNA microarray studies (comparative phylogenomics). Whole genome comparisons were carried out combined with Bayesian-based algorithms to model the phylogeny of the 88 strains. From this data, the Bayesian-based phylogeny (11) revealed two distinct clades supported unequivocally by Bayesian probabilities; a “pathogenic” clade comprising primarily of pathogenic biotypes (BT2-4) and a “non-pathogenic” clade comprising primarily of the non-pathogenic BT 1a isolates (Figure 1) Of the 88 strains there were two “non-conformer” BT 1a strains in the “pathogenic” clade and one “non-conformer” BT 4 strain in the “non-pathogenic” clade. These strains will be re-checked by microarray analysis and their biotypes will be confirmed. Within the “pathogenic” clade the data splits into two distinct groups. This is believed to be caused by the presence or absence of the virulence plasmid pYV and was confirmed using GeneSpring (shown on figure 1). Loss of the virulence plasmid alone does not account for the clear divide of the putatively pathogenic and non-pathogenic strains from each other and it strongly indicates that there are large chromosomal differences between these to groups. This is an important finding and demonstrated that the presence of the pYV plasmid alone is not responsible for virulence.
As previously mentioned American and European “pathogenic” strains have significant phenotypic differences. Eight American pathogenic strains were recently made available for investigation with the microarrays. These were highly similar to the control sequenced 8081 1b strain with only one or two genes differences (data not shown).
The analysis shows that, as with AFLP analysis, no distinct groups formed which had unique host specificity in either “pathogenic” or “non-pathogenic” clades. These results suggest that the genomic differences between isolates of the same biotype and serotype, irrespective of whether they were isolated from humans, pigs, cattle or sheep, are very small. By microarray the isolates, within the “pathogenic” clade, of common bio-serotypes only generally grouped together. This is in contrast to the AFLP data for the same strains where clustering was closely associated with bio-serotype. Thus it appears that the basis for the clustering of strains in the array analysis is different from
SID 5 (2/05) Page 8 of 19
that in the AFLP. These differences were appeared to have some impact when looking for host associations; for example the distinct clusters within the pig and human BT4 (O:3) strains seen by AFLP were not clear by microarray, while the close association between human and pig BT3 (O:9) strains was diffused over two subclades differentiated by the presence or absence of pVY. It is hoped that further studies using MacClade software might reveal the specific genes, which cause the isolates to fall into these distinct groups. The data will also be analysed to determine whether there are any CDSs that are specific to human isolates.
In addition to carrying out comparative phylogenomics the data has also been clustered using condition trees within GeneSpring 7. This allows all of the predicted CDSs to be visualised and colours them according to whether the CDS is present (yellow) or absent (blue) (fig 2). This highlighted many CDSs that were absent in both the non-pathogenic and pathogenic European strains compared to the highly pathogenic American strains (about 227 genes represented as blue lines or blue blocks across all strains) (fig 2). Surprisingly, the non-pathogenic strains (grouped on the right of fig 2 by vertical red lines) had less gene deletions than the pathogenic European strains compared to the highly pathogenic American strains. It may seem counterintuitive that the pathogenic strains should have more deletions, but can be explained through distinct lineages of parallel evolution of the three Y. enterocolitica subgroups (European non-pathogenic; European pathogenic and American pathogenic) over a considerable evolutionary time. Similar instances of so called “black holes” , ie deletions relating to higher virulence, have been reported for E. coli, Shigella species and Yersinia pestis (19). From a practical perspective, we now have dozens of new genetic markers that can distinguish
SID 5 (2/05) Page 9 of 19
Figure1. The phylogenetic relationship of all strains tested in a collection comprising of human strains from HPA and cattle, sheep and pig strains isolated during the national abattoir study (1999-2000). 100 = 100% of all phylogenies showing a given topology.
the three Y. enterocolitica subgroups, and have the possibility of relating these genetic differences to phenotypic traits.
Part of this objective was also to characterise up to 4 putative virulence factors of Y. enterocolitica. The well-recognised approach to such questions is to construct a mutant by defined gene deletion and then investigate the effect of this mutation on phenotypic properties by comparison of the mutant and the wild-type. However, in the past, the construction of defined mutants of Y. enterocolitica has proven particularly problematic. For the first time in both collaborative laboratories, mutant construction for Y. enterocolitica has been established.
Different approaches to the selection of genes for mutation were adopted in each laboratory. In the first approach the microarray data from this study, in conjunction with data from the Yersinia pestis genome project, was mined to identify genetic sequences with putative virulence significance. The second approach adopted the opposite strategy in that a well-recognised phenotypic property with presumed virulence significance was investigated at the genomic level.
SID 5 (2/05) Page 10 of 19
Figure 2. Condition tree. Shows data collected from all test strains. The strains were clustered using the absent gene list generated by Trinary GACK analysis with spearman correlation in GeneSpring 7. The branches representing each strain are coloured according to biotype.
Investigation of genetic differences identified by microarray:The target genes were selected from microarray analysis of the data and from the Yersinia pestis genome project. These results revealed that the coding sequence YE1398 was present in all strains tested and YE0694 was absent from all strains excepting Y. enterocolitica 8081. YE1398 shares 77% similarity to the YojN protein, which is involved in the RcsC-RcsB two component regulatory system. YE0694 shares 28% similarity with the E. coli Intimin adhesion protein. Both of these CDSs are pseudogenes in Y. pestis. Y. pestis is thought to have evolved from the enteric pathogen Y. pseudotuberculosis between 1,500-20,000 years ago mainly through gene loss through the acquisition of pseudogenes (1, 17). Using the rationale that YE1398 may play an important role in the pathogenicity of all Y. enterocolitica strains but that YW0694 now only important in the highly pathogenic strains and that both genes are no longer required and therefore functionally redundant in Y. pestis, they were selected for mutagenesis.
Defined Ye0694 and Ye1398 mutations were constructed utilising a red recombinase system (Appendix 3) in BT1b strains JB850V and WA-C, and confirmed by PCR and Southern blot. Characterisation of the YE0694 and YE1398 defined mutants demonstrated reduced (WA-C YE0694-), or no swarming motility (WA-C YE1398-) respectively. This suggests that both genes YE0694 have significant roles in swarming Further studies are under way to complement these mutants and to test their pathogenic potential in a murine model of Yersiniosis infection, in collaboration with Virginia Miller,
Investigation of the role of flagella:Preliminary investigations with naturally-occurring variants of Y. enterocolitica suggested that the flagella of BT 1a Y. enterocolitica strains play a role in invasion of epithelial cells in vitro, persistence in macrophages, and modulation of cytokine secretion in vitro. (This data is detailed in Appendix 4). To confirm this link, a clinical faecal BT1a isolate (53/03) was selected, and the regulator of the fleABC flagellin operon, fliA, was insertionally inactivated using the previously described pKNG101 suicide vector (9). Attempts to amplify the fleABC operon from BT1a isolates using primers designed against the sequence previously published proved unsuccessful. Therefore the fliA gene was chosen for mutagenesis. The fliA gene encodes a regulator responsible for inducing expression of the flagellin encoding genes during flagella synthesis, and has been successfully mutated to produce an aflagellate mutant in a BT1b strain (17). A Kanamycin cassette was successfully inserted into fliA of strain YE53/03. This is the first report of mutant construction in a clinical Y. enterocolitica isolate or in a BT1a isolate.
The data indicated that abolition of flagella expression resulted in the complete loss of the ability of the clinical strain YE53/03 to invade epithelial cells in vitro. Complementation of the mutant resulted in ~ 70% restoration of the invasive phenotype. Incomplete restoration is expected when complementation is performed using a plasmid encoded copy of the mutated gene (12). Interestingly the mutation had no affect on bacterial association with the tissue culture cells, suggesting that an inability to actively swim towards the cell layer, or to adhere to the epithelial cell surface were unimportant in this assay.
Further studies demonstrated that the aflagellate mutant was completely cleared by U937 human macrophages in vitro after six hours, whilst the wild type and complemented strain persisted. A time course experiment showed that the mutant was ingested by the macrophages to the same extent as the wild type, but was completely killed between 3 and 4 hours post-infection. To investigate this effect further the response of macrophages to infection with the wild type and mutant was studied. There was no difference in levels of the cytokine IL-6 (required for clearance of Y.
SID 5 (2/05) Page 11 of 19
enterocolitica infections) secreted by cells infected with either the mutant or wild type. However, U937 cells infected with the mutant, were unable to secrete high levels of the anti-inflammatory cytokine IL-10, whilst the wild type strain induced high levels of this cytokine. Complementation of the mutant resulted in ~50% restoration of the IL-10 secretion phenotype. Similarly, U937cells infected with the mutant strain increased the secretion ( ~2-fold) of the pro-inflammatory cytokine TNF- compared to the wild type and complemented organisms.
In order to provide further evidence to confirm the virulence properties of BT1a flagella, crude flagellin preparations, from strains 53/03 (a clinical isolate) and SJM2 ( fliA mutant) were prepared by size exclusion sandwich centrifugation of secreted flagellin. Bacteria, were grown in low salt, low temperature conditions to induce secretion of flagellin (17), and passed first through a 40kDa filter and then through a 20kDa filter. The resulting sample was analysed by SDS-PAGE and Western blot to confirm the major protein present was flagellin. Approximately 1g/ml of the purified flagellin was added to cultured U937 cells. The added flagellin had no obvious cytotoxic effect on the cells as determined microscopically, and there was no induced secretion of IL-10, in contrast to that observed with wild type whole cells. However, the flagellin-treated cells secreted lower (~2-fold) TNFlevels when exposed to flagellin from 53/03 compared to the equivalent protein sample prepared from SJM2. This data adds further proof that the flagella of BT1a Y. enterocolitica are immunomodulatory.
Objective 3: To investigate the virulence potential of human and animal strains using tissue culture models.
The results from objectives 1 and 2 confirm that there is a molecular basis for the division of strains in terms of biotype and serotype. However, the relationship between biotype and basis of pathogencity remain unanswered in two main areas; firstly the putative virulence differences in BT 1a and BT2-5 strains and secondly the fact that BT 1a strains which, although recovered from patients with enteric disease, are also commonly isolated from community based healthy controls (6).
In an attempt to extend the phenotypic and genotypic characterisation undertaken in Objectives 1 and 2 to include virulence-associated properties (virulotyping), the 88 selected isolates described previously were phenotypically characterised by use of a variety of in vitro assays. These included the ability of strains to invade cultured human epithelial cells, to escape from epithelial cells once internalised, and to survive encounter with macrophages. These assays mimic the host infection cycle of Y. enterocolitica, which invades intestinal epithelial cells, escapes via the baso-lateral membrane, and colonises the lamina propria, where it must avoid or survive encounters with host macrophages (2).
InvasivesnessThe first assay compared invasiveness, using HEp-2 human epithelial cells. The detailed methods and results are included in a manuscript in preparation (see appendix 5).
The results clearly demonstrated that all Y. enterocolitica biotypes are capable of adhering to epithelial cells in high numbers, with approximately 10% of the organisms added associating with the HEp-2 cells (~ 106 cfu/ml from 107cfu/ml added for infection with Multiplicity Of Infection (MOI) of 1:100). Isolates of the putatively pathogenic BTs, e.g. BT3 (O:5,27), BT3 (O:9), and BT4 (O:3) strains, all have the same levels of
SID 5 (2/05) Page 12 of 19
association to, and invasion of, epithelial cells. However, isolates of BT 2-4 are about 10 times more invasive than BT 1a strains, with invasion levels of 104-105 cfu/ml compared to 102-103 cfu/ml respectively. Nevertheless, the invasiveness of BT 1a isolates was significant when compared to other enteric pathogens, for example Campylobacter jejuni is substantially less invasive than this but is still considered an invasive pathogen. There was no relationship between host source and invasive potential for this organism.
Escape capacityThe ability of bacteria to divide in, and escape from, invaded cells, is considered a primary virulence property of Y. enterocolitica. This property enables the bacteria to spread laterally down the gut and traverse the intestinal epithelial basal layer. An escape assay was established to mimic this bacterial property in vitro, which involved performing a gentamicin protection assay as with the invasion assay, and then incubating the cells with internalised bacteria overnight in antibiotic free medium. The bacteria in the medium were then enumerated to ascertain the ability of the strain to escape from epithelial cells (7). The 88 selected strains were tested in this assay. A variation in the ability of strains to escape from cells was observed. Moreover, this variation to some extent reflected the observed invasive potential of these strains. Biotype 1a isolates originating from sheep displayed a weak escape phenotype, with none of the strains assayed displaying levels of escape above that for an E. coli DH5 control (102 cfu/ml.) All other strains tested displayed levels of escape of between 9.2 x 104 and 1 x 107 cfu/ml. BT 4 (O:3) isolates displayed lower levels of escape then the other pathotypes tested, with 2.4 x 10 4 cfu/ml recovered. None of the isolates tested displayed significant increases in the number of intracellular bacteria in the replication assay performed within HEp-2 cells.
Macrophage survivalThe ability of Y. enterocolitica isolates to survive macrophage engulfment is also considered a virulence property of this organism. Upon escape from intestinal epithelial cells, Y. enterocolitica colonises the lamina propria, which results in mass infiltration of macrophages and other phagocytic cells in an attempt to clear the infection. Y. enterocolitica avoids macrophages killing by either injecting the host macrophages with cytotoxins upon attachment (3), or by inhibiting the phagolysosomal process, allowing the organism to survive within the macrophages (16). An in vitro assay, using the U937 differentiated human macrophages cell line was established to mimic this bacterial property and test the survival capabilities of the 88 selected strains. The results for the average number of organisms recovered did not vary over 1-5 hours post challenge and there was no significant difference in the capacity of any of the BTs to survive within macrophages.
Modulation of cytokine secretionWhen macrophages encounter a pathogen, they produce a large number of cytokine signals, which are involved in regulating and inducing a relevant immune response to clearing the infection. In order to assess the ability of isolates to modulate cytokine secretion, supernatants from Yersina-infected U937 cells were harvested and assayed using commercial ELISA kits to detect IL-6 and IL-8 (key pro-inflammatory cytokines involved in clearing Y. enterocolitica infections (4, 13)); IL-10 ( an anti-inflammatory cytokine which down-regulates the immune response in an uninfected situation (15)); and TNF (the major pro-inflammatory cytokine released in response to intracellular pathogens (8)). Infections were carried out with 4 randomly selected isolates from each biotype/serotype group. Secretion of IL-6 was highest in cells infected with BT 3 (O:5,27) isolates compared to the other biotypes/serotypes tested. However, levels of IL-6 secretion were also slightly elevated by all biotype 1a strains tested. In contrast the BT 4 (O:3) and BT 3 (O:9) strains, regardless of source, down-regulation IL-6 secretion when
SID 5 (2/05) Page 13 of 19
compared to the other strain types.
Similarly secretion of IL-8 by infected U937 cells was also greatest for BT 3 (O:5,27) isolates. Moreover, the secretion of IL-8 induced by BT 1a strains, was also elevated compared to BT3 (O:9) and BT4 (O:3) strains, although secretion levels were slightly higher in the environmental BT1a isolates than in the human strains.
Levels of secreted IL-10 were similar for all of the isolates tested, and did not differ significantly from the levels secreted by uninfected cells. In contrast, U937 cells secreted large amounts of TNF- upon infection with all BT 1a strains. However, secretion induced by human and animal BT1a strains was consistently higher (~8 fold and ~4 fold respectively) than that observed upon infection with BT2-4 strains.
These results suggest that biotypes differ in their ability to modulate the host immune response. In particular, the results for BT3 (O:5,27) indicate that this BT is incapable of down-regulating levels of IL-6 and IL-8, resulting in an increased rate of clearance of infection. This may help to explain the absence of this BT in human isolates. The data for the BT1a isolates also suggests that infection with this BT leads to a different immune response to that observed for BT4 and BT3. This could result in a different type of infection and disease pathology to that classically reported in Y. enterocolitica infection with BT1b or BT2-5.
Objective 4: To evaluate two simple laboratory tests of putative virulence of Y. enterocolitica in human and animal strains.
To date there has been no validated, simple biochemical, serological, or genotypic test to identify pathogenicity of Y. enterocolitica isolates. The aim of this objective was to assess the ability of two previously reported simple tests in to distinguish putatively “pathogenic” from “non-pathogenic” isolates as defined by biotype.
Arabitol Utilisation TestThe arabitol test separates isolates on the basis of BT1a (able to utilise D-arabitol) and BT2-5 (unable ot utilise D-arabitol) (5, 14). The arabitol utilisation test was performed on the 88 selected isolates. The majority (92%) of the BT1a isolates, while only 4% of the BT 2-5 isolates, utilised arabitol. Thus, in light of the findings in Objective 3, and taking into account that CMOX medium can differentiate strains in the same way on the basis of presence or absence of the pYV plasmid, the ability of the arabitol utilisation test to define the pathogenicity of strains is questionable.
The Myf TestMyf is a fimbrillar antigen produced uniquely by Y. enterocolitica isolates of BT1b and BT 2-5. A diagnostic latex agglutination kit was developed as part of a previous Defra-funded project.
The Myf latex agglutination test was developed and manufactured at the VLA. However, after the start of this project, this test kit became unavailable, therefore, a Fluorescent Antibody Test (FAT) using anti-Myf antiserum serum raised in rabbits, produced and labelled by Biotechnology Dept, VLA Weybridge, was established and performed on all 88 selected isolates. The results indicated that only 81% of BT2-5 isolates were positive for Myf expression, however this was significantly higher than the 25% of BT1a isolates which were positive. In addition, there was no correlation between expression of Myf in
SID 5 (2/05) Page 14 of 19
BT1a strains and invasion or genotyping groups. These results suggest that Myf expression is a poor indicator of pathogenic potential.
Additional work undertaken:
The role of serum antibodies in cattle and pigs in protective immunity against Y. enterocolitica
Little is known about the bacterial factors involved in the colonisation of animals with Y. enterocolitica or the role of host immunity in this process. Interestingly as part of previous Defra-funded studies undertaken for the purposes of investigating the effect of cross-reacting immune responses to Y. enterocolitica on the diagnosis of Brucella spp infections, sera from experimentally infected cattle and pigs were available. Both the cattle and pigs were infected with a BT3 (O:9) strain, with serum harvested post mortem at day 7 post inoculation. In addition serum was collected 3 days prior to infection. These sera were utilised to investigate the specificity of antibodies produced directed against Y. enterocolitica. Whole cell protein preparations from 7 Y. enterocolitica strains, selected to represent each bio/serotype present in cattle, pigs, sheep and humans, and grown overnight in LB broth at 28oC, were separated by SDS-PAGE and analysed by western blotting. Both infected cattle and pigs elicted serum antibodies directed against cross-reactive antigens expressed by all isolates. However, the serum antibody response was strongest in both cattle and pigs, against the BT 3 (O:9) isolates, the same serotype used in the experimental challenge (Fig 1, appendix 6). In contrast, pre-infection sera gave only very weak reactivity.
To investigate the specificity of the antibody response against secreted proteins of Y. enterocolitica, Fop and Yop secreted proteins were produced under optimal growth and recovered from the bacterial culture supernatant of 7 strains. These preparations were then separated by SDS-PAGE, as before, and western blotted with the pig and cattle antisera. The antisera reacted with Fops from various biotypes, indicating the presence of cross-reacting antibodies (Fig 2, appendix 6). Using a polyclonal antiserum raised against the flagella of Y. enterocolitica BT 1b O:8 isolate (kindly supplied by G. Cornelis, University of Basle), at least one of these polypeptides (37Kda) was flagellin. Due to the absence of antibodies specific to Yop effectors we could not positively confirm the identity of the antigenic polypeptides in the Yop preparations, however the SDS-PAGE profile matched that previously reported for Yop secretion (17). As previously reported (2) BT 1a isolates did not express detectable Yop proteins.
As previously demonstrated invasiveness is a key virulence property for Y. enterocolitica. Pig and cattle post challenge antisera was incubated with HEp-2 cells for 10 mins prior to infection, resulting in a reduction of invasion by all biotype strains tested by up to ten-fold. Serum from pre-infected animals had no effect. This figure was significant (p < 0.001) for all the biotypes excepting one strain ( pig BT4 (O:3) strain (p <0.194)). These results suggest that the antibodies present in serum of colonised cattle and pigs offer some degree of protection against the invasiveness of Y. enterocolitica. However, interestingly the ability of the strains to survive in U937 cells was only slightly reduced, (up to 2-fold) by the addition of serum, suggesting that anti-Y. enterocolitica antibodies provides only a limited degree of protection against the ability of Y. enterocolitica to subvert host macrophages.
The detailed methods and results of this work have been submitted for publication (see appendix 6).
Conclusions:
SID 5 (2/05) Page 15 of 19
The results of this investigation suggest that, on the basis of phenotypic and genotypic studies, all Y. enterocolitica strains present in livestock have the potential to pose a threat to human public health, with the possible exception of BT3 O:5,27 strains. In particular, although previously assumed to be non-pathogenic, BT1a strains can both adhere to and invade epithelial cells, survive inside macrophages, and down-regulate the secretion by infected macrophages of IL-6 and IL-8, two key cytokines in clearance of Y. enterocolitica infections. This is particularly important as BT1a strains colonise a significant proportion of livestock.
For the first time, in a collaboration with Professor Wren LSTHM, DNA microarrays have been used to investigate the comparative genomics of Y. enterocolitica. These studies have indicated the molecular basis of the pathotype differences and provided insight into the evolution of Y. enterocolitica. Further work is clearly now required to further mine this data.
A prerequisite to future indentification of virulence associated genes in Y. enterocolitica, is the development of methods for defined deletion mutations. One strategy was adapted to delete two genes identified as being potentially associated with virulence on the basis of the DNA microarrays. This work is still ongoing with collaborators and could provide evidence for novel virulence factors. A second strategy adopted, for the first time, enabled the deletion of a locus in BT1a wild type isolates. This work clearly demonstrated the key role of flagellin in adherence, invasion and macrophage survival. This work will provide a significant impetus to future research on the virulence of Y. enterocolitica.
The role of host immune response in the outcome of infection is well recognised. In studies extra to the objectives, the role of anti-Y. enterocolitica antibodies, directed in part against secreted proteins, in the control of colonisation and possibly virulence was demonstrated using in vitro models of adherence and invasion. Most recently the importance of cellular immune responses in bacterial enteric infections has been recognised. For Y. enterocolitica such responses were assayed by the detection of the expression of the cytokines IL-8, IL-6, IL-10 and TNF by infected macrophage cells. Interestingly, these preliminary observations suggest that such responses are biotype-associated. Further studies are required particularly using systems which more closely reflect the whole animal, e.g.in vitro organ cultures, or a mouse chronic infection model.
References
1. Achtman, M., K. Zurth, et al. 1999 Yersinia pestis, the cause of plague, is a recently emerged clone of Yersinia pseudotuberculosis. PNAS 96: 14043-14048 2. Bottone, E. J. 1999. Yersinia enterocolitica: overview and epidemiologic correlates. Microbes and infection 1: 323-333.3. Cornelis, G. R., and H. Wolf-Watz. 1997. The Yersinia Yop virulon: a bacterial system for subverting eukaryotic cells. Molecular Microbiology 23:861-867.4. Denecker, G., S. Totemeyer, L. J. Mota, P. Troisfontaines, I. Lambermont, C. Youta, I. Stainier, M. Ackermann, and G. R. Cornelis. 2002. Effect of low- and high-virulence Yersinia enterocolitica strains on the inflammatory response of human umbilical vein endothelial cells. Infection and Immunity 70:3510-3520.5. Fantasia Mazzotti, M., M. G. Mingrone, and A. Martini. 1987. Fermenting ability of Yersinia spp. strains on esculin, L-fucose and D-arabitol. Microbiologica 10:413-6.6. Food Standards Agency. 2000. A report of the study of infectious intestinal Disease in England. The Stationary Office.
SID 5 (2/05) Page 16 of 19
7. Grant, T., V. Bennett-Wood, and R. M. Robins-Browne. 1999. Characterization of the interaction between Yersinia enterocolitica biotype 1A and phagocytes and epithelial cells in vitro. Infection and Immunity 67:4367-4375.8. Kampik, D., R. Schulte, and I. B. Autenrieth. 2000. Yersinia enterocolitica invasin protein triggers differential production of interleukin-1, interleukin-8, monocyte chemoattractant protein 1, granulocyte-macrophage colony-stimulating factor, and tumor necrosis factor alpha in epithelial cells: implications for understanding the early cytokine network in Yersinia infections. Infect Immun 68:2484-92.9. Kaniga, K., I. Delor, and G. R. Cornelis. 1991. A wide-host-range suicide vector for improving reverse genetics in gram-negative bacteria: inactivation of the blaA gene of Yersinia enterocolitica. Gene 109:137-141.10. Newell D.G, Frost J.A., Duim B., Wagenaar J.A., Madden R.H., van der Plas J., and On S.L.W. New developments in the subtyping of Campylobacter species. In Campylobacter 2nd Edition. Ed Nachamkin I. and Blaser M.J, ASM Washington DC PP 27-4411. Ronquist, F. & Huelsenbeck, J.P. Mr Bayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574 (2003)12. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: A laboratory manual, 2nd ed. Cold Spring Harbour Laboratory Press, New York.13. Schulte, R., and I. B. Autenrieth. 1998. Yersinia enterocolitica-induced interleukin-8 secretion by human intestinal epithelial cells depends on cell differentiation. Infect Immun 66:1216-24.14. Shehee, M. W., and M. D. Sobsey. 2004. Development of a L-rhamnose and D-arabitol supplemented MacConkey agar to identify pathogenic Yersinia enterocolitica among environmental Yersinias in swine production wastes. J Microbiol Methods 57:289-92.15. Sing, A., A. Roggenkamp, A. M. Geiger, and J. Heesemann. 2002. Yersinia enterocolitica evasion of the host innate immune response by V antigen-induced IL-10 production of macrophages is abrogated in IL-10-deficient mice. J Immunol 168:1315-21.16. Yamamoto, T., T. Hanawa, S. Ogata, and S. Kamiya. 1997. The Yersinia enterocolitica GsrA stress protein, involved in intracellular survival, is induced by macrophage phagocytosis. Infect Immun 65:2190-6.17. Young, B. M., and G. M. Young. 2002. YplA is exported by the Ysc, Ysa, and flagellar type III secretion systems of Yersinia enterocolitica. Journal of Bacteriology 184:1324-1334.18. Young, G. M., J. L. Badger, and V. L. Miller. 2000. Motility is required to initiate host cell invasion by Yersinia enterocolitica. Infection and Immunity 68:4323-4326.19. Wren, B. W. 2003. The Yersiniae - a model genus to study the rapid evolution of bacterial pathogens. Nature Reviews Microbiology. 1:55-64.
SID 5 (2/05) Page 17 of 19
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 18 of 19
McNally, A., Cheasty, T., Fearnley, C., Dalziel, R. W., Paiba, G., Manning, G., and Newell, D.G. (2004). Comparison of the biotypes of Yersinia enterocolitica isolated from pigs, cattle and sheep at slaughter and from humans with yersiniosis in Great Britain during 1999-2000. Letters in Applied Microbiology. 39(1): 103-108
McNally, A., Fearnley, C., On, S. L., Manning, G., Cheasty, T., and Newell, D. G. (In Press). Application of fluorescent Amplified Fragment Length Polymorphism to the comparison of human and animal Yersinia enterocolitica. Applied and Environmental Microbiology.
McNally, A., Brew, S.D., Manning, G., and Newell, D. G. (In Submission). Serum responses in pigs and cattle against Yersinia enterocolitica virulence factors. Veterinary Microbiology.
McNally, A, Schuller, S., Dalton, T., La Ragione, R. M., Manning, G., Phillips, A. D., and Newell, D. G. Yersinia enterocolitica isolates from humans and animals are adherent, invasive, and persist in macrophages in vitro regardless of biotype, but differ in modulation of cytokine secretion. Manuscript in Preparation.
McNally, A. and Newell, D. G. Flagella of biotype 1A Y. enterocolitica play a role in invasion of epithelial cells, persistence in macrophages, and modulation of cytokine secretion in vitro. Manuscript in Preparation.
McNally, A. Comparison of Y. enterocolitica from humans and animals. VLA Research directorate scientific group meeting. Oral presentation Jan 2005.
McNally, A. Comparison of Y. enterocolitica from humans and animals. Defra non-foodborne zoonoses review. Oral presentation Aug 2004.
Howard, S. Use of pan-Yersinia microarray for comparison of human and animal Y. enterocolitica. 3rd annual BGS microarray conference, Hinxton hall, Cambridge. Oral presentation May 2004.
Fearnley, C. Comparison of Y. enterocolitica isolates from humans and animals. AVTRW annual conference, Scarborough. Oral presentation April 2003.
McNally, A., Dalton, T., Manning, G., Cheasty, T., and Newell, D.G. Genotypic and phenotypic comparison of Y. enterocolitica from humans and livestock to assess zoonotic risk. Annual SGM conference, Trinity College, Dublin. Poster presentation September 2004.
Howard, S., McNally, A., Hinchcliffe, S., Prentice, M.B., Newell, D.G., and Wren, B.W. Use of pan-Yersinia microarray for comparison of human and animal Y. enterocolitica. 3rd annual BGS microarray conference, Hinxton hall, Cambridge. Poster presentation May 2004.
SID 5 (2/05) Page 19 of 19
