Genotypic Analysis of the Earliest Known Prehistoric Case of Tuberculosis in Britain
http://www.100md.com
微生物临床杂志 2005年第5期
Centre for Molecular Microbiology and Infectious Disease, Imperial College of Science, Technology, and Medicine, London
Ancient Monuments Laboratory, English Heritage Centre for Archaeology, Fort Cumberland, Eastney, Portsmouth, United Kingdom
ABSTRACT
The earliest known case of human tuberculosis in Britain dates to the middle period of the Iron Age, approximately 2,200 years before present. Bone lesions on the spine of a male skeleton excavated at Tarrant Hinton in Dorset, United Kingdom, show evidence of Pott's disease and are supported by molecular evidence of Mycobacterium tuberculosis complex DNA amplified by IS6110 PCR (19). In the present study, we used a further series of sensitive PCR methods to confirm the diagnosis of tuberculosis and to determine the genotype of the infecting strain. These tests demonstrated that this individual was infected with a strain of M. tuberculosis rather than Mycobacterium bovis. The strain had undergone the tuberculosis D1 deletion affecting the mmpS6 and mmpL6 genes and can therefore be identified as a member of the family of "modern" M. tuberculosis isolates. All evidence obtained was consistent with surviving mycobacterial DNA being highly fragmented in this case.
INTRODUCTION
Tuberculosis is the result of infection by members of the Mycobacterium tuberculosis complex. Disease in humans is caused predominantly by strains classified as M. tuberculosis, which can be subdivided into "ancient" and "modern" families on the basis of the tuberculosis D1 deletion event. Less frequently, human tuberculosis is caused by strains that are classified as Mycobacterium canetti and are thought to have diverged from the last common ancestor in the distant past or by strains that have undergone the region of difference 9 deletion event and are classified as Mycobacterium africanum or as Mycobacterium bovis (4). While tuberculosis has traditionally been thought of as a zoonotic disease transferred to humans in neolithic times, it is now clear that M. bovis—the pathogen responsible for disease in most other mammals—in fact corresponds to a more recent group of strains that have undergone a series of deletion events perhaps associated with adaptation to different host species. In light of the current diversity of strains responsible for human tuberculosis and of questions related to the evolution of M. tuberculosis, it was of interest to examine the phenotype of strains involved in cases of ancient disease.
We previously reported on a prehistoric case of tuberculosis in an Iron Age interment, burial 7, excavated from Tarrant Hinton in Dorset, United Kingdom (19). The remains comprised the skeleton of an adult male about 30 to 40 years old at the time of death. The lumbar vertebrae of this skeleton showed lytic lesions which were strongly suggestive of an infectious process. Damage to the vertebrae had resulted in 60° angular kyphosis of the spine. Various possibilities, including brucellosis, were considered, but tuberculosis, resulting in Pott's disease, was the preferred diagnosis. This conclusion was supported by biomolecular analysis of bone powder sampled from the burial. Extracts prepared from two lumbar vertebrae and a rib were PCR positive for the multicopy insertion element IS6110, an established marker of the M. tuberculosis complex (15). However, this test does not distinguish between the two most likely causative agents in humans, Mycobacterium tuberculosis and Mycobacterium bovis.
The aims of the present study were to determine whether the Tarrant Hinton individual had suffered from disease due to M. tuberculosis or M. bovis and to carry out further genotypic analysis of the Iron Age strain. One of the criteria used for authenticating ancient DNA (aDNA) studies is that they should provide consistent molecular data (6). We therefore reasoned that, given the reproducible IS6110 findings in this case, it should be possible to demonstrate the presence of other regions of the Mycobacterium tuberculosis genome and that these should make phylogenetic sense. We therefore modified several tuberculosis genotyping PCRs to amplify fragments of <120 bp and applied these to new samples of bone taken from the skeleton of burial 7.
Data recovered from archaeological human remains with osteological evidence of tuberculosis are often minimal. Nevertheless, such data are one means of testing models for the evolution of the M. tuberculosis complex (4, 24) and of establishing an absolute chronology for genetic variations and host adaptations seen in extant strains. In this context, the Tarrant Hinton case is significant, as it is the earliest known example of skeletal tuberculosis in Britain and was recently radiocarbon dated (19).
MATERIALS AND METHODS
Bone samples. Bone powder was collected by one of us (S.A.M.) from the body of one of each of the diseased thoracic vertabrae (T12) and lumbar vertebrae (LV2). Samples were also taken from trabecular bone within two ribs. One of these ribs had previously tested positive (19). Samples were taken by using sterile disposable scalpel blades and were placed in sterile tubes for transportation to the laboratory. Gloves were worn and changed when samples were obtained from different sites. A total of 420 mg of bone powder was obtained from LV2, 180 mg was obtained from T12, 240 mg was obtained from rib 1, and 400 mg was obtained from rib 2. Three extracts were prepared from LV2 and T12, and two extracts were prepared from each of the rib samples.
DNA extraction. DNA was extracted by using NucliSens kits from bioMerieux UK Ltd., Hampshire, United Kingdom. Bone powder was transferred to 9-ml guanidinium lysis buffer tubes, mixed on a rotating wheel (Stuart Rotator SB2) for 30 min, and heated to 95°C for 5 min. The samples then underwent three freeze-thaw cycles in liquid nitrogen. Thereafter, the manufacturer's instructions were followed for isolation and washing of the DNA on silica before elution in 60 μl Tris-EDTA buffer. Extraction negative controls, consisting of reagents but no bone powder, were processed at the same time. Whenever practical, the extracts were assayed before freezing. When this was not convenient, they were stored at –20°C before being assayed.
PCR. The oligonucleotide primers designed specifically for this study are shown in Table 1, which also shows the main cycling parameters for each assay and the amplicon sizes. Forty-five cycles of amplification were used for first-round stages. One microliter of product was used as a template in the nested and heminested second-round stages, when an additional 35 to 41 cycles were performed.
An Excite core kit (BioGene, Cambridge, United Kingdom) was used for all PCR amplifications. This kit is a uracil-N-glycosylase-ready kit suitable for real-time and routine hot-start PCR applications. SYBR green was included in the PCR master mixtures at a final dilution of 1/55,000, and reactions were performed and monitored with a Corbett Rotorgene 3000 real-time PCR platform and a final volume of 25 μl. Melting analysis was performed with Rotorgene software, and all products also were run on 3% agarose checker gels. Positive controls for the IS1081 and D1 PCR methods were run separately from the prehistoric extracts to minimize the chances of cross-contamination; the PCR machine used was a Hybaid Express programmed to match the Rotorgene 3000. Products for sequencing were separated on 3% low-melting-temperature agarose gels, and bands were excised and purified by using a Q-Bio gene Geneclean kit (Fisher Scientific, Leicestershire, United Kingdom).
Cycle sequencing. Cycle sequencing was performed on an Applied Biosystems 9700 PCR block with a dRhodamine dye terminator ready reaction kit (Applied Biosystems, Cheshire, United Kingdom). Products were sequenced with both forward and reverse primers, and the reactions were run on an ABI 310 genetic analyzer (Applied Biosystems).
Contamination control measures. Measures to prevent contamination of extracts with previously generated amplicons were applied throughout the study. These measures have been described in detail elsewhere (18). The focus was on physical barriers, with separate areas for extraction, PCR setup, and analysis. Surfaces and equipment (centrifuges, pipettes, and so forth) were cleaned before each assay. Filter tips were used routinely. Control extractions were used to monitor the efficacy of these procedures. Template (water) blanks were alternated with samples in the PCR machines to screen for random contamination. Positive controls were either omitted from amplifications or were run in separate laboratories to minimize the opportunities for introducing contamination.
RESULTS
A summary of the main molecular findings from the Tarrant Hinton case is shown in Table 2. Nested PCR for IS6110 was used initially to confirm that extracts prepared for the present study were also positive for M. tuberculosis complex DNA. Lumbar vertebra LV2 and one of the rib extracts were again shown to be positive. Further experiments concentrated on these extracts.
Initial attempts to amplify IS1081, a second multicopy locus characteristic of the M. tuberculosis complex, with outer primers F2 and R2 were unsuccessful. However, when the inner primers F2 and R3 were used, products of the correct size (113 bp) were obtained from LV2 and both rib extracts. This result probably reflects the highly degraded nature of the DNA fragments in this case. Sequencing of two of the IS1081 products confirmed their identities (Table 2).
Next we analyzed two loci commonly used to distinguish M. tuberculosis from M. bovis. A band of the expected size was obtained for the oxyR locus from both samples. Sequencing of these bands showed that the nucleotide at position 285 was G, consistent with this being a strain of M. tuberculosis rather than M. bovis (23). The oxyR reverse primer always produced clearer data around the region of the nucleotide at position 285, with amplicons generated from modern control DNA as well as the present case. We also applied a second method to confirm the species. The nucleotide at position 169 in the pyrazinamidase gene (pncA) was shown to distinguish between M. tuberculosis and M. bovis strains (22). Sequencing of the pncA band from the Tarrant Hinton case indicated that this nucleotide was C, confirming the strain as M. tuberculosis.
We next tested the lumbar vertebra and both rib extracts with both flanking and internal primers to ascertain the status of the tuberculosis D1 deletion. This M. tuberculosis-specific deletion has been used to divide evolutionarily "old" and "modern" strains (4). A product was obtained with DNA prepared from LV2 by using the flanking primers. The identity of this band was confirmed by sequencing. This finding identifies the strain as a "modern" M. tuberculosis isolate. We also observed that the internal primers generated an amplicon from the LV2 extract. This amplicon was sequenced and shown to be a primer-dimer artifact which happened to coincide with the size of the expected product.
IS6110 was amplified on one occasion from samples LV2 and rib 1, consistent with earlier PCR findings and sequencing (19). oxyR285 was amplified on two separate occasions from samples LV2 and rib 1. IS1081 was amplified twice from the LV2 extract and on one occasion each from samples rib 1 and rib 2. The D1 PCR was repeated twice each with two separate extracts of sample LV2, with the same result. IS6110, oxyR285, and pncA were tested without positive controls to minimize the risk of contamination with modern strains. Positive controls for IS1081 and D1 were included but were run on separate PCR machines for the same reason. With the exception of oxyR285 from sample rib 2, all PCR products were successfully sequenced.
The use of the Corbett real-time PCR platform was found to be particularly suitable for the initial optimization of PCR methods, for monitoring blanks during amplifications, and for allowing the termination of second-round amplifications after the optimum number of cycles.
DISCUSSION
We reexamined bone samples from burial 7 at Tarrant Hinton, containing the remains of an adult male about 40 years old at the time of death. The skeleton (Fig. 1) showed evidence of tubercle causing Pott's disease of the spine, which in this case would have resulted in an angular kyphosis of about 60° (Fig. 2). We previously reported on the main osteological findings of the skeleton (19). This is the earliest known case of tuberculosis in Britain.
Our earlier study coupled the paleopathological evidence with aDNA analysis; attempts were made using PCR to amplify several loci on the tuberculosis genome. Only one of these, for IS6110, was successful. As this insertion element is multicopy in many extant M. tuberculosis strains, the inference was that the causative species was probably M. tuberculosis, but definitive proof was lacking, as IS6110 is also present in M. bovis in single copy and in other members of the M. tuberculosis complex (M. canetti, M. africanum, Mycobacterium microti, M. bovis, and M. bovis BCG) (7). There are no known skeletal manifestations of tuberculosis which enable a distinction to be made between these mycobacteria on purely pathological grounds.
The aDNA PCR methods for tuberculosis tried previously on burial 7 relied on the preservation of DNA fragments of between 131 and 211 bp, lengths which might reasonably be expected to survive in archaeological material, given suitable environmental conditions (21). These loci and the amplicon sizes were as follows: katg463 (163 bp), katG203 (142 bp), gyrA95 (131 bp), oxyR285 (150 bp), RD7 (211 bp), and mtp40 (152 bp). All of the above were unsuccessful. The exception to this was IS6110, with a first-round amplicon size of 123 bp and a size of 92 bp after nesting. This method is exquisitely sensitive for detecting tuberculosis DNA and is usually the first to be applied for identifying cases in archaeological studies (1, 12, 27). However, given the difference in sensitivities between IS6110 and other methods, there is clearly still a need for confirmatory tests for this pathogen in archaeological material and also for those which will provide useful genotyping data.
Multicopy genes and repeated sequences make good targets for aDNA studies (16). We have therefore applied a confirmatory PCR for tuberculosis which targets another multicopy element, IS1081. This is typically present in six copies in members of the complex (5). We have also modified the oxyR and pncA methods, which distinguish between M. tuberculosis and M. bovis (22, 23), for use with smaller DNA fragments. New primers have been designed to determine the status of the TbD1 deletion. It has been suggested that loss of this region was an early event in the evolutionary scenario of the M. tuberculosis complex which can be used to discriminate truly "ancestral" from "modern" strains (4).
Data from the revised biomolecular methods indicate that Pott's disease present in burial 7 from Tarrant Hinton was certainly due to a strain of M. tuberculosis. A PCR product was obtained with the TbD1 flanking primers, which indicated the deletion had occurred. As the skeleton had been radiocarbon dated to 2,286 ± 25 (mean ± standard deviation) years before present, which gives a calibrated date range of between 400 and 230 BC, our findings show that the TbD1 deletion had occurred by this period in the British Iron Age. Loss of this region is common to strains from all three major groups of M. tuberculosis on the Musser system of typing (24). Retention of the region is found in only a proportion of the early group 1 lineage and also in all strains of M. bovis (4).
The extent of M. bovis infection in Iron Age herds—and hence the risk of human exposure—is unknown. M. bovis has the greatest host range of any other member of the Mycobacterium tuberculosis complex, and in this middle period of the British Iron Age (600 to 100 BC), cattle, sheep, pigs, goats, and wild species, such as deer, would have been typical elements of the mixed farming economy. In fact, faunal remains from sites across Wessex indicate that sheep were predominant by this time, having taken over from cattle. This was probably due to their ability to survive on higher pastures with minimal requirement for water and for their wool, an important raw material for Iron Age spindle looms. However, cattle would have still been important as symbols of status and kept as beasts of traction as well as for their milk and other dairy products which might have been stored for consumption outside periods of lactation (8).
The search for residual genetic signatures of some pathogens in prehistoric material has proved problematical. An example is the search for treponemal DNA in skeletons with evidence of venereal or endemic syphilis undertaken to answer the enigma of its origins and to link skeletal lesions with species (20). Early indications that DNA from this pathogen might persist (17) have not been borne out by later studies, leaving many workers doubtful whether DNA from this organism will ever feature prominently in bioarchaeological research (3). Similarly, the survival of DNA from Yersinia species remains controversial, with some groups finding a relatively high occurrence in remains of victims of the Black Death (10), an observation which others have been unable to replicate (13).
Tuberculosis appears to be singular in this respect, in that there is a growing body of papers which report recovery of tuberculosis DNA from human and animal remains (2, 11). This can in part be explained by the readily identified nature of skeletal lesions and by the extremely robust character of the mycobacterial cell wall (25). A degree of similarity of the bacterial envelope with Mycobacterium leprae probably explains why some groups have successfully demonstrated this pathogen in cases of lepromatous leprosy (9, 14, 26).
Our experience with the various PCR methods indicated that DNA in this case was particularly degraded compared to others we have examined (18). The first study on this skeleton indicated that fragments of tuberculosis DNA of larger than 130 bp did not survive, and this finding was borne out by the observation in the present study that the IS1081 method failed to generate an amplicon when outer primers were used (135 bp) but was successful when inner primers were used (113 bp) (Table 2). In addition, the main findings were consistent between bone samples taken from different anatomical sites (Table 2), supporting the hematogenous spread of the disease throughout the body. The genotyping data that we obtained were in agreement with the proposed evolutionary scenario for the M. tuberculosis complex. Both the D1 internal and pncA PCR methods occasionally produced misleading bands at high cycle numbers in the presence of aDNA extracts. These were difficult to distinguish from the genuine products using either agarose gel chromatography or melt analysis on the real-time platform. Therefore, it is worth stressing the importance of sequencing all PCR products, particularly those in the size range of 90 to 115 bp, in order to distinguish them from primer-dimer and other nonspecific artifacts of PCR, such as those that might result from template switching.
A number of guidelines have been suggested for validating data from prehistoric material (6). Generally, these are helping to raise the degree of scientific rigour applied to aDNA projects. However, not all criteria have been applied in all studies, often for reasons of cost, lack of expertise, or limited sample availability. The last was certainly a factor here, restricting the number of additional PCR tests which could be used to genotype the strain further. However, reproducibility is central in the aDNA field, and the opportunity to reexamine the Tarrant Hinton remains has allowed us to identify the species as M. tuberculosis. A major obstacle to further scrutiny of the case would be the large sample size needed to recover the pathogen DNA, making additional destructive sampling difficult to justify.
In summary, sensitive and careful PCR analysis has allowed us to derive genotypic data from an Iron Age case of tuberculosis. The ability to obtain genotypic data by exploring the limited archaeological resource is one means of establishing an absolute chronology for polymorphisms which probably reflect changes in pathogen virulence and host susceptibility over time through the selective expansion of successful clones in the various lineages of the M. tuberculosis complex.
ACKNOWLEDGMENTS
We are very grateful to the staff at the Priest's House Museum, Wimborne, Dorset, United Kingdom, for allowing further sampling from the skeleton of burial 7. Thanks are due also to Richard Hopkins, who developed and optimized the oxyR PCR for aDNA applications, and to Alan Graham, who provided background information on the Tarrant Hinton site.
REFERENCES
Baron, H., S. Hummel, and B. Herrmann. 1996. Mycobacterium tuberculosis complex DNA in ancient human bones. J. Archaeol. Sci. 23:667-671.
Bathurst, R. R., and J. L. Barta. 2004. Molecular evidence of tuberculosis induced hypertrophic osteopathy in a 16th-century Iroquoian dog. J. Archaeol. Sci. 31:917-925.
Bouwman, A. S. 2004. Treponemal diseases and the prospects for biomolecular archaeology. PhD thesis. University of Manchester Institute of Science and Technology, Manchester, United Kingdom.
Brosch, R., S. V. Gordon, M. Marmiesse, P. Brodin, C. Buchrieser, K. Eiglmeier, T. Garnier, C. Gutierrez, G. Hewinson, K. Kremer, L. M. Parsons, A. S. Pym, S. Samper, D. van Soolingen, and S. T. Cole. 2002. A new evolutionary scenario for the Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA 99:3684-3689.
Collins, D. M., and D. M. Stephens. 1991. Identifcation of an insertion sequence, IS1081, in Mycobacterium bovis. FEMS Microbiol. Lett. 67:11-15.
Cooper, A., and H. Poinar. 2000. Ancient DNA: do it right or not at all. Science 289:1139.
Cousins, D., S. Williams, E. Liebana, A. Aranaz, A. Bunschoten, J. Van Embden, and T. Ellis. 1998. Evaluation of four DNA typing techniques in epidemiological investigations of bovine tuberculosis. J. Clin. Microbiol. 36:168-178.
Cunliffe, B. 1993. Wessex to AD 1000. Longman Group, United Kingdom Ltd., London, United Kingdom.
Donoghue, H. D., J. Holton, and M. Spigelman. 2001. PCR primers that can detect low levels of Mycobacterium leprae DNA. J. Med. Microbiol. 50:177-182.
Drancourt, M., G. Aboudharam, M. Signoli, O. Dutour, and D. Raoult. 1998. Detection of 400-year-old Yersinia pestis DNA in human dental pulp: an approach to the diagnosis of ancient septicemia. Proc. Natl. Acad. Sci. USA 95:12637-12640.
Fletcher, H. A., H. D. Donoghue, G. M. Taylor, A. G. M. van der Zanden, and M. Spigelman. 2003. Molecular analysis of Mycobacterium tuberculosis DNA from a family of 18th century Hungarians. Microbiology 149:143-151.
Gernaey, A. M., D. E. Minnikin, M. S. Copley, R. A. Dixon, J. C. Middleton, and C. A. Roberts. 2001. Mycolic acids and ancient DNA confirm an osteological diagnosis of tuberculosis. Tuberculosis 81:259-265.
Gilbert, M. T. P., J. Cuccui, W. White, N. Lynnerup, R. W. Titball, A. Cooper, and M. B. Prentice. 2004. Absence of Yersinia pestis-specific DNA in human teeth from five European excavations of putative plague victims. Microbiology 150:341-354.
Haas, C. J., A. Zink, G. Palfi, U. Szeimies, and A. G. Nerlich. 2000. Detection of leprosy in ancient human skeletal remains by molecular identification of Mycobacterium leprae. Am. J. Clin. Pathol. 114:428-436.
Hellyer, T. J., L. E. Desjardin, M. K. Assaf, J. H. Bates, D. M. Cave, and K. D. Eisenach. 1996. Specificity of IS6110-based amplification assays for Mycobacterium tuberculosis complex. J. Clin. Microbiol. 34:2843-2846.
Hofreiter, M., D. Serre, H. N. Poinar, M. Kuch, and S. Pbo. 2001. Ancient DNA. Nat. Rev. Genet. 2:353-359.
Kolman, C. J., A. Centurion-Lara, S. A. Lukehart, D. W. Owsley, and N. Tuross. 1999. Identification of Treponema pallidum subspecies pallidum in a 200-year-old skeletal specimen. J. Infect. Dis. 180:2060-2063.
Mays, S., G. M. Taylor, A. J. Legge, D. B. Young, and G. Turner-Walker. 2001. Aleopathological and biomolecular study of tuberculosis in a medieval skeletal collection from England. Am. J. Phys. Anthropol. 114:298-311.
Mays, S., and G. M. Taylor. 2003. A first prehistoric case of tuberculosis from Britain. Int. J. Osteoarchaeol. 13:189-196.
Morton, R. S., and S. Rashid. 2001. The syphilis enigma: the riddle resolved Sex. Transm. Infect. 77:322-324.
Pbo, S. 1990. Amplifying ancient DNA, p. 159-166. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White (ed.), PCR protocols: a guide to methods and applications. Academic Press, Inc., New York, N.Y.
Scorpio, A., and Y. Zhang. 1996. Mutations in pncA, a gene encoding pyrazinamidase/nicotinamidase, cause resistance to the anti-tuberculous drug pyrazinamide in tubercle bacillus. Nat. Med. 2:662-667.
Sreevatsan, S., P. Escalante, X. Pan, D. A. Gillies, S. Siddiqui, C. N. Khalaf, B. N. Kreiswirth, L. G. Adams, T. Ficht, V. S. Perumaalla, M. D. Cave, J. D. van Emden, and J. M. Musser. 1996. Identification of a polymorphic nucleotide in oxyR specific for Mycobacterium bovis. J. Clin. Microbiol. 34:2007-2010.
Sreevatsan, S., X. Pan, K. E. Stockbauer, N. D. Connell, B. N. Kreiswirth, T. S. Whitman, and J. M. Musser. 1997. Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent dissemination. Proc. Natl. Acad. Sci. USA 94:9869-9874.
Takade, A., A. Umeda, M. Matsuoka, S.-I. Yoshida, M. Nakamura, and K. Amako. 2003. Comparative studies on the cell structures of Mycobacterium leprae and M. tuberculosis using the electron microscopy freeze-substitution technique. Microbiol. Immunol. 47:265-270.
Taylor, G. M., S. Widdison, I. N. Brown, D. B. Young, and T. Molleson. 2000. A case of lepromatous leprosy from 13th century Orkney. J. Archaeol. Sci. 27:1133-1138.
Zink, A. R., C. Sola, U. Reischl, W. Grabner, N. Rastogi, H. Wolf, and A. G. Nerlich. 2003. Characterization of Mycobacterium tuberculosis complex DNAs from Egyptian mummies by spoligotyping. J. Clin. Microbiol. 41:359-367.(G. Michael Taylor, Dougla)
Ancient Monuments Laboratory, English Heritage Centre for Archaeology, Fort Cumberland, Eastney, Portsmouth, United Kingdom
ABSTRACT
The earliest known case of human tuberculosis in Britain dates to the middle period of the Iron Age, approximately 2,200 years before present. Bone lesions on the spine of a male skeleton excavated at Tarrant Hinton in Dorset, United Kingdom, show evidence of Pott's disease and are supported by molecular evidence of Mycobacterium tuberculosis complex DNA amplified by IS6110 PCR (19). In the present study, we used a further series of sensitive PCR methods to confirm the diagnosis of tuberculosis and to determine the genotype of the infecting strain. These tests demonstrated that this individual was infected with a strain of M. tuberculosis rather than Mycobacterium bovis. The strain had undergone the tuberculosis D1 deletion affecting the mmpS6 and mmpL6 genes and can therefore be identified as a member of the family of "modern" M. tuberculosis isolates. All evidence obtained was consistent with surviving mycobacterial DNA being highly fragmented in this case.
INTRODUCTION
Tuberculosis is the result of infection by members of the Mycobacterium tuberculosis complex. Disease in humans is caused predominantly by strains classified as M. tuberculosis, which can be subdivided into "ancient" and "modern" families on the basis of the tuberculosis D1 deletion event. Less frequently, human tuberculosis is caused by strains that are classified as Mycobacterium canetti and are thought to have diverged from the last common ancestor in the distant past or by strains that have undergone the region of difference 9 deletion event and are classified as Mycobacterium africanum or as Mycobacterium bovis (4). While tuberculosis has traditionally been thought of as a zoonotic disease transferred to humans in neolithic times, it is now clear that M. bovis—the pathogen responsible for disease in most other mammals—in fact corresponds to a more recent group of strains that have undergone a series of deletion events perhaps associated with adaptation to different host species. In light of the current diversity of strains responsible for human tuberculosis and of questions related to the evolution of M. tuberculosis, it was of interest to examine the phenotype of strains involved in cases of ancient disease.
We previously reported on a prehistoric case of tuberculosis in an Iron Age interment, burial 7, excavated from Tarrant Hinton in Dorset, United Kingdom (19). The remains comprised the skeleton of an adult male about 30 to 40 years old at the time of death. The lumbar vertebrae of this skeleton showed lytic lesions which were strongly suggestive of an infectious process. Damage to the vertebrae had resulted in 60° angular kyphosis of the spine. Various possibilities, including brucellosis, were considered, but tuberculosis, resulting in Pott's disease, was the preferred diagnosis. This conclusion was supported by biomolecular analysis of bone powder sampled from the burial. Extracts prepared from two lumbar vertebrae and a rib were PCR positive for the multicopy insertion element IS6110, an established marker of the M. tuberculosis complex (15). However, this test does not distinguish between the two most likely causative agents in humans, Mycobacterium tuberculosis and Mycobacterium bovis.
The aims of the present study were to determine whether the Tarrant Hinton individual had suffered from disease due to M. tuberculosis or M. bovis and to carry out further genotypic analysis of the Iron Age strain. One of the criteria used for authenticating ancient DNA (aDNA) studies is that they should provide consistent molecular data (6). We therefore reasoned that, given the reproducible IS6110 findings in this case, it should be possible to demonstrate the presence of other regions of the Mycobacterium tuberculosis genome and that these should make phylogenetic sense. We therefore modified several tuberculosis genotyping PCRs to amplify fragments of <120 bp and applied these to new samples of bone taken from the skeleton of burial 7.
Data recovered from archaeological human remains with osteological evidence of tuberculosis are often minimal. Nevertheless, such data are one means of testing models for the evolution of the M. tuberculosis complex (4, 24) and of establishing an absolute chronology for genetic variations and host adaptations seen in extant strains. In this context, the Tarrant Hinton case is significant, as it is the earliest known example of skeletal tuberculosis in Britain and was recently radiocarbon dated (19).
MATERIALS AND METHODS
Bone samples. Bone powder was collected by one of us (S.A.M.) from the body of one of each of the diseased thoracic vertabrae (T12) and lumbar vertebrae (LV2). Samples were also taken from trabecular bone within two ribs. One of these ribs had previously tested positive (19). Samples were taken by using sterile disposable scalpel blades and were placed in sterile tubes for transportation to the laboratory. Gloves were worn and changed when samples were obtained from different sites. A total of 420 mg of bone powder was obtained from LV2, 180 mg was obtained from T12, 240 mg was obtained from rib 1, and 400 mg was obtained from rib 2. Three extracts were prepared from LV2 and T12, and two extracts were prepared from each of the rib samples.
DNA extraction. DNA was extracted by using NucliSens kits from bioMerieux UK Ltd., Hampshire, United Kingdom. Bone powder was transferred to 9-ml guanidinium lysis buffer tubes, mixed on a rotating wheel (Stuart Rotator SB2) for 30 min, and heated to 95°C for 5 min. The samples then underwent three freeze-thaw cycles in liquid nitrogen. Thereafter, the manufacturer's instructions were followed for isolation and washing of the DNA on silica before elution in 60 μl Tris-EDTA buffer. Extraction negative controls, consisting of reagents but no bone powder, were processed at the same time. Whenever practical, the extracts were assayed before freezing. When this was not convenient, they were stored at –20°C before being assayed.
PCR. The oligonucleotide primers designed specifically for this study are shown in Table 1, which also shows the main cycling parameters for each assay and the amplicon sizes. Forty-five cycles of amplification were used for first-round stages. One microliter of product was used as a template in the nested and heminested second-round stages, when an additional 35 to 41 cycles were performed.
An Excite core kit (BioGene, Cambridge, United Kingdom) was used for all PCR amplifications. This kit is a uracil-N-glycosylase-ready kit suitable for real-time and routine hot-start PCR applications. SYBR green was included in the PCR master mixtures at a final dilution of 1/55,000, and reactions were performed and monitored with a Corbett Rotorgene 3000 real-time PCR platform and a final volume of 25 μl. Melting analysis was performed with Rotorgene software, and all products also were run on 3% agarose checker gels. Positive controls for the IS1081 and D1 PCR methods were run separately from the prehistoric extracts to minimize the chances of cross-contamination; the PCR machine used was a Hybaid Express programmed to match the Rotorgene 3000. Products for sequencing were separated on 3% low-melting-temperature agarose gels, and bands were excised and purified by using a Q-Bio gene Geneclean kit (Fisher Scientific, Leicestershire, United Kingdom).
Cycle sequencing. Cycle sequencing was performed on an Applied Biosystems 9700 PCR block with a dRhodamine dye terminator ready reaction kit (Applied Biosystems, Cheshire, United Kingdom). Products were sequenced with both forward and reverse primers, and the reactions were run on an ABI 310 genetic analyzer (Applied Biosystems).
Contamination control measures. Measures to prevent contamination of extracts with previously generated amplicons were applied throughout the study. These measures have been described in detail elsewhere (18). The focus was on physical barriers, with separate areas for extraction, PCR setup, and analysis. Surfaces and equipment (centrifuges, pipettes, and so forth) were cleaned before each assay. Filter tips were used routinely. Control extractions were used to monitor the efficacy of these procedures. Template (water) blanks were alternated with samples in the PCR machines to screen for random contamination. Positive controls were either omitted from amplifications or were run in separate laboratories to minimize the opportunities for introducing contamination.
RESULTS
A summary of the main molecular findings from the Tarrant Hinton case is shown in Table 2. Nested PCR for IS6110 was used initially to confirm that extracts prepared for the present study were also positive for M. tuberculosis complex DNA. Lumbar vertebra LV2 and one of the rib extracts were again shown to be positive. Further experiments concentrated on these extracts.
Initial attempts to amplify IS1081, a second multicopy locus characteristic of the M. tuberculosis complex, with outer primers F2 and R2 were unsuccessful. However, when the inner primers F2 and R3 were used, products of the correct size (113 bp) were obtained from LV2 and both rib extracts. This result probably reflects the highly degraded nature of the DNA fragments in this case. Sequencing of two of the IS1081 products confirmed their identities (Table 2).
Next we analyzed two loci commonly used to distinguish M. tuberculosis from M. bovis. A band of the expected size was obtained for the oxyR locus from both samples. Sequencing of these bands showed that the nucleotide at position 285 was G, consistent with this being a strain of M. tuberculosis rather than M. bovis (23). The oxyR reverse primer always produced clearer data around the region of the nucleotide at position 285, with amplicons generated from modern control DNA as well as the present case. We also applied a second method to confirm the species. The nucleotide at position 169 in the pyrazinamidase gene (pncA) was shown to distinguish between M. tuberculosis and M. bovis strains (22). Sequencing of the pncA band from the Tarrant Hinton case indicated that this nucleotide was C, confirming the strain as M. tuberculosis.
We next tested the lumbar vertebra and both rib extracts with both flanking and internal primers to ascertain the status of the tuberculosis D1 deletion. This M. tuberculosis-specific deletion has been used to divide evolutionarily "old" and "modern" strains (4). A product was obtained with DNA prepared from LV2 by using the flanking primers. The identity of this band was confirmed by sequencing. This finding identifies the strain as a "modern" M. tuberculosis isolate. We also observed that the internal primers generated an amplicon from the LV2 extract. This amplicon was sequenced and shown to be a primer-dimer artifact which happened to coincide with the size of the expected product.
IS6110 was amplified on one occasion from samples LV2 and rib 1, consistent with earlier PCR findings and sequencing (19). oxyR285 was amplified on two separate occasions from samples LV2 and rib 1. IS1081 was amplified twice from the LV2 extract and on one occasion each from samples rib 1 and rib 2. The D1 PCR was repeated twice each with two separate extracts of sample LV2, with the same result. IS6110, oxyR285, and pncA were tested without positive controls to minimize the risk of contamination with modern strains. Positive controls for IS1081 and D1 were included but were run on separate PCR machines for the same reason. With the exception of oxyR285 from sample rib 2, all PCR products were successfully sequenced.
The use of the Corbett real-time PCR platform was found to be particularly suitable for the initial optimization of PCR methods, for monitoring blanks during amplifications, and for allowing the termination of second-round amplifications after the optimum number of cycles.
DISCUSSION
We reexamined bone samples from burial 7 at Tarrant Hinton, containing the remains of an adult male about 40 years old at the time of death. The skeleton (Fig. 1) showed evidence of tubercle causing Pott's disease of the spine, which in this case would have resulted in an angular kyphosis of about 60° (Fig. 2). We previously reported on the main osteological findings of the skeleton (19). This is the earliest known case of tuberculosis in Britain.
Our earlier study coupled the paleopathological evidence with aDNA analysis; attempts were made using PCR to amplify several loci on the tuberculosis genome. Only one of these, for IS6110, was successful. As this insertion element is multicopy in many extant M. tuberculosis strains, the inference was that the causative species was probably M. tuberculosis, but definitive proof was lacking, as IS6110 is also present in M. bovis in single copy and in other members of the M. tuberculosis complex (M. canetti, M. africanum, Mycobacterium microti, M. bovis, and M. bovis BCG) (7). There are no known skeletal manifestations of tuberculosis which enable a distinction to be made between these mycobacteria on purely pathological grounds.
The aDNA PCR methods for tuberculosis tried previously on burial 7 relied on the preservation of DNA fragments of between 131 and 211 bp, lengths which might reasonably be expected to survive in archaeological material, given suitable environmental conditions (21). These loci and the amplicon sizes were as follows: katg463 (163 bp), katG203 (142 bp), gyrA95 (131 bp), oxyR285 (150 bp), RD7 (211 bp), and mtp40 (152 bp). All of the above were unsuccessful. The exception to this was IS6110, with a first-round amplicon size of 123 bp and a size of 92 bp after nesting. This method is exquisitely sensitive for detecting tuberculosis DNA and is usually the first to be applied for identifying cases in archaeological studies (1, 12, 27). However, given the difference in sensitivities between IS6110 and other methods, there is clearly still a need for confirmatory tests for this pathogen in archaeological material and also for those which will provide useful genotyping data.
Multicopy genes and repeated sequences make good targets for aDNA studies (16). We have therefore applied a confirmatory PCR for tuberculosis which targets another multicopy element, IS1081. This is typically present in six copies in members of the complex (5). We have also modified the oxyR and pncA methods, which distinguish between M. tuberculosis and M. bovis (22, 23), for use with smaller DNA fragments. New primers have been designed to determine the status of the TbD1 deletion. It has been suggested that loss of this region was an early event in the evolutionary scenario of the M. tuberculosis complex which can be used to discriminate truly "ancestral" from "modern" strains (4).
Data from the revised biomolecular methods indicate that Pott's disease present in burial 7 from Tarrant Hinton was certainly due to a strain of M. tuberculosis. A PCR product was obtained with the TbD1 flanking primers, which indicated the deletion had occurred. As the skeleton had been radiocarbon dated to 2,286 ± 25 (mean ± standard deviation) years before present, which gives a calibrated date range of between 400 and 230 BC, our findings show that the TbD1 deletion had occurred by this period in the British Iron Age. Loss of this region is common to strains from all three major groups of M. tuberculosis on the Musser system of typing (24). Retention of the region is found in only a proportion of the early group 1 lineage and also in all strains of M. bovis (4).
The extent of M. bovis infection in Iron Age herds—and hence the risk of human exposure—is unknown. M. bovis has the greatest host range of any other member of the Mycobacterium tuberculosis complex, and in this middle period of the British Iron Age (600 to 100 BC), cattle, sheep, pigs, goats, and wild species, such as deer, would have been typical elements of the mixed farming economy. In fact, faunal remains from sites across Wessex indicate that sheep were predominant by this time, having taken over from cattle. This was probably due to their ability to survive on higher pastures with minimal requirement for water and for their wool, an important raw material for Iron Age spindle looms. However, cattle would have still been important as symbols of status and kept as beasts of traction as well as for their milk and other dairy products which might have been stored for consumption outside periods of lactation (8).
The search for residual genetic signatures of some pathogens in prehistoric material has proved problematical. An example is the search for treponemal DNA in skeletons with evidence of venereal or endemic syphilis undertaken to answer the enigma of its origins and to link skeletal lesions with species (20). Early indications that DNA from this pathogen might persist (17) have not been borne out by later studies, leaving many workers doubtful whether DNA from this organism will ever feature prominently in bioarchaeological research (3). Similarly, the survival of DNA from Yersinia species remains controversial, with some groups finding a relatively high occurrence in remains of victims of the Black Death (10), an observation which others have been unable to replicate (13).
Tuberculosis appears to be singular in this respect, in that there is a growing body of papers which report recovery of tuberculosis DNA from human and animal remains (2, 11). This can in part be explained by the readily identified nature of skeletal lesions and by the extremely robust character of the mycobacterial cell wall (25). A degree of similarity of the bacterial envelope with Mycobacterium leprae probably explains why some groups have successfully demonstrated this pathogen in cases of lepromatous leprosy (9, 14, 26).
Our experience with the various PCR methods indicated that DNA in this case was particularly degraded compared to others we have examined (18). The first study on this skeleton indicated that fragments of tuberculosis DNA of larger than 130 bp did not survive, and this finding was borne out by the observation in the present study that the IS1081 method failed to generate an amplicon when outer primers were used (135 bp) but was successful when inner primers were used (113 bp) (Table 2). In addition, the main findings were consistent between bone samples taken from different anatomical sites (Table 2), supporting the hematogenous spread of the disease throughout the body. The genotyping data that we obtained were in agreement with the proposed evolutionary scenario for the M. tuberculosis complex. Both the D1 internal and pncA PCR methods occasionally produced misleading bands at high cycle numbers in the presence of aDNA extracts. These were difficult to distinguish from the genuine products using either agarose gel chromatography or melt analysis on the real-time platform. Therefore, it is worth stressing the importance of sequencing all PCR products, particularly those in the size range of 90 to 115 bp, in order to distinguish them from primer-dimer and other nonspecific artifacts of PCR, such as those that might result from template switching.
A number of guidelines have been suggested for validating data from prehistoric material (6). Generally, these are helping to raise the degree of scientific rigour applied to aDNA projects. However, not all criteria have been applied in all studies, often for reasons of cost, lack of expertise, or limited sample availability. The last was certainly a factor here, restricting the number of additional PCR tests which could be used to genotype the strain further. However, reproducibility is central in the aDNA field, and the opportunity to reexamine the Tarrant Hinton remains has allowed us to identify the species as M. tuberculosis. A major obstacle to further scrutiny of the case would be the large sample size needed to recover the pathogen DNA, making additional destructive sampling difficult to justify.
In summary, sensitive and careful PCR analysis has allowed us to derive genotypic data from an Iron Age case of tuberculosis. The ability to obtain genotypic data by exploring the limited archaeological resource is one means of establishing an absolute chronology for polymorphisms which probably reflect changes in pathogen virulence and host susceptibility over time through the selective expansion of successful clones in the various lineages of the M. tuberculosis complex.
ACKNOWLEDGMENTS
We are very grateful to the staff at the Priest's House Museum, Wimborne, Dorset, United Kingdom, for allowing further sampling from the skeleton of burial 7. Thanks are due also to Richard Hopkins, who developed and optimized the oxyR PCR for aDNA applications, and to Alan Graham, who provided background information on the Tarrant Hinton site.
REFERENCES
Baron, H., S. Hummel, and B. Herrmann. 1996. Mycobacterium tuberculosis complex DNA in ancient human bones. J. Archaeol. Sci. 23:667-671.
Bathurst, R. R., and J. L. Barta. 2004. Molecular evidence of tuberculosis induced hypertrophic osteopathy in a 16th-century Iroquoian dog. J. Archaeol. Sci. 31:917-925.
Bouwman, A. S. 2004. Treponemal diseases and the prospects for biomolecular archaeology. PhD thesis. University of Manchester Institute of Science and Technology, Manchester, United Kingdom.
Brosch, R., S. V. Gordon, M. Marmiesse, P. Brodin, C. Buchrieser, K. Eiglmeier, T. Garnier, C. Gutierrez, G. Hewinson, K. Kremer, L. M. Parsons, A. S. Pym, S. Samper, D. van Soolingen, and S. T. Cole. 2002. A new evolutionary scenario for the Mycobacterium tuberculosis complex. Proc. Natl. Acad. Sci. USA 99:3684-3689.
Collins, D. M., and D. M. Stephens. 1991. Identifcation of an insertion sequence, IS1081, in Mycobacterium bovis. FEMS Microbiol. Lett. 67:11-15.
Cooper, A., and H. Poinar. 2000. Ancient DNA: do it right or not at all. Science 289:1139.
Cousins, D., S. Williams, E. Liebana, A. Aranaz, A. Bunschoten, J. Van Embden, and T. Ellis. 1998. Evaluation of four DNA typing techniques in epidemiological investigations of bovine tuberculosis. J. Clin. Microbiol. 36:168-178.
Cunliffe, B. 1993. Wessex to AD 1000. Longman Group, United Kingdom Ltd., London, United Kingdom.
Donoghue, H. D., J. Holton, and M. Spigelman. 2001. PCR primers that can detect low levels of Mycobacterium leprae DNA. J. Med. Microbiol. 50:177-182.
Drancourt, M., G. Aboudharam, M. Signoli, O. Dutour, and D. Raoult. 1998. Detection of 400-year-old Yersinia pestis DNA in human dental pulp: an approach to the diagnosis of ancient septicemia. Proc. Natl. Acad. Sci. USA 95:12637-12640.
Fletcher, H. A., H. D. Donoghue, G. M. Taylor, A. G. M. van der Zanden, and M. Spigelman. 2003. Molecular analysis of Mycobacterium tuberculosis DNA from a family of 18th century Hungarians. Microbiology 149:143-151.
Gernaey, A. M., D. E. Minnikin, M. S. Copley, R. A. Dixon, J. C. Middleton, and C. A. Roberts. 2001. Mycolic acids and ancient DNA confirm an osteological diagnosis of tuberculosis. Tuberculosis 81:259-265.
Gilbert, M. T. P., J. Cuccui, W. White, N. Lynnerup, R. W. Titball, A. Cooper, and M. B. Prentice. 2004. Absence of Yersinia pestis-specific DNA in human teeth from five European excavations of putative plague victims. Microbiology 150:341-354.
Haas, C. J., A. Zink, G. Palfi, U. Szeimies, and A. G. Nerlich. 2000. Detection of leprosy in ancient human skeletal remains by molecular identification of Mycobacterium leprae. Am. J. Clin. Pathol. 114:428-436.
Hellyer, T. J., L. E. Desjardin, M. K. Assaf, J. H. Bates, D. M. Cave, and K. D. Eisenach. 1996. Specificity of IS6110-based amplification assays for Mycobacterium tuberculosis complex. J. Clin. Microbiol. 34:2843-2846.
Hofreiter, M., D. Serre, H. N. Poinar, M. Kuch, and S. Pbo. 2001. Ancient DNA. Nat. Rev. Genet. 2:353-359.
Kolman, C. J., A. Centurion-Lara, S. A. Lukehart, D. W. Owsley, and N. Tuross. 1999. Identification of Treponema pallidum subspecies pallidum in a 200-year-old skeletal specimen. J. Infect. Dis. 180:2060-2063.
Mays, S., G. M. Taylor, A. J. Legge, D. B. Young, and G. Turner-Walker. 2001. Aleopathological and biomolecular study of tuberculosis in a medieval skeletal collection from England. Am. J. Phys. Anthropol. 114:298-311.
Mays, S., and G. M. Taylor. 2003. A first prehistoric case of tuberculosis from Britain. Int. J. Osteoarchaeol. 13:189-196.
Morton, R. S., and S. Rashid. 2001. The syphilis enigma: the riddle resolved Sex. Transm. Infect. 77:322-324.
Pbo, S. 1990. Amplifying ancient DNA, p. 159-166. In M. A. Innis, D. H. Gelfand, J. J. Sninsky, and T. J. White (ed.), PCR protocols: a guide to methods and applications. Academic Press, Inc., New York, N.Y.
Scorpio, A., and Y. Zhang. 1996. Mutations in pncA, a gene encoding pyrazinamidase/nicotinamidase, cause resistance to the anti-tuberculous drug pyrazinamide in tubercle bacillus. Nat. Med. 2:662-667.
Sreevatsan, S., P. Escalante, X. Pan, D. A. Gillies, S. Siddiqui, C. N. Khalaf, B. N. Kreiswirth, L. G. Adams, T. Ficht, V. S. Perumaalla, M. D. Cave, J. D. van Emden, and J. M. Musser. 1996. Identification of a polymorphic nucleotide in oxyR specific for Mycobacterium bovis. J. Clin. Microbiol. 34:2007-2010.
Sreevatsan, S., X. Pan, K. E. Stockbauer, N. D. Connell, B. N. Kreiswirth, T. S. Whitman, and J. M. Musser. 1997. Restricted structural gene polymorphism in the Mycobacterium tuberculosis complex indicates evolutionarily recent dissemination. Proc. Natl. Acad. Sci. USA 94:9869-9874.
Takade, A., A. Umeda, M. Matsuoka, S.-I. Yoshida, M. Nakamura, and K. Amako. 2003. Comparative studies on the cell structures of Mycobacterium leprae and M. tuberculosis using the electron microscopy freeze-substitution technique. Microbiol. Immunol. 47:265-270.
Taylor, G. M., S. Widdison, I. N. Brown, D. B. Young, and T. Molleson. 2000. A case of lepromatous leprosy from 13th century Orkney. J. Archaeol. Sci. 27:1133-1138.
Zink, A. R., C. Sola, U. Reischl, W. Grabner, N. Rastogi, H. Wolf, and A. G. Nerlich. 2003. Characterization of Mycobacterium tuberculosis complex DNAs from Egyptian mummies by spoligotyping. J. Clin. Microbiol. 41:359-367.(G. Michael Taylor, Dougla)