AMERICAN JOURNAL OF PHYSICAL ANTHROPOLOGY 114:298-311 (2001)
Paleopathological and Biomolecular Study of Tuberculosis in a Medieval Skeletal Collection From England
S. Mays,1* G.M. Taylor,2 A.J. Legge,3 D.B. Young,2 and G. Turner-Walker4
ancient Monuments Laboratory, English Heritage, Fort Cumberland, Eastney, Portsmouth P04 9LD, UK ^Department of Infectious Diseases and Microbiology, Imperial College School of Medicine at St. Mary's, London W2 IPG, UK
^Faculty of Continuing Education, Birkbeck College, London WC1B 5DQ, UK
^Institute of Archaeology and Cultural History, Vitenskapsmuseet, Norwegian University of Science and Technology, N-7034 Trondheim, Norway
KEY WORDS      ancient DNA; Wharram Percy; paleopathology; Mycobacterium tuberculosis; Mycobacterium bovis
ABSTRACT Nine human skeletons of medieval date from a rural English burial site show signs of skeletal tuberculosis. They were subject to polymerase chain reaction (PCR) assays aimed at detecting traces of DNA from infecting mycobacteria, with the purpose both of confirming the paleopathological diagnosis of tuberculosis and determining in individual cases whether disease was due
to M. tuberculosis or M. bovis. In all nine cases, evidence for M. tuberculosis complex DNA was found, and in all instances it appeared that disease was due to M. tuberculosis rather than M. bovis. The significance of the findings for understanding tuberculous infection in rural agrarian communities in medieval England is discussed. Am J Phys Anthropol 114:298-311, 2001.   © 2001 Wiley-Liss, inc.
Tuberculosis is an infectious disease which may be acute or chronic, and which may affect any organ in the body, including the skeleton. The principal causative agents of tuberculous disease in humans are two bacteria of the Mycobacterium tuberculosis complex, M. tuberculosis and M. bovis. Mycobacterium tuberculosis is an acid-fast bacillus which has microbial and biochemical characteristics almost identical with Mycobacterium bovis, which is responsible for causing disease in diverse animal species, particularly cattle, as well as in humans (O'Reilly and Daborn, 1995).
Transmission of M. tuberculosis is generally by inhaled droplet infection from person to person, so disease caused by this microorganism generally begins as a respiratory infection. M. bovis may be acquired via inhalation of contaminated aerosol from live infected animals or carcasses, producing pulmonary disease. It may also be transmitted by consumption of contaminated milk or meat, in which case the primary lesion is located in the cervical lymph nodes (due to oropharangeal penetration by ingested bacteria) or the alimentary tract (Aufder-heide and Rodriguez-Martin, 1998).
Skeletal tuberculosis is, almost without exception, a result of hematogenous spread of infection from soft-tissue foci. Skeletal disease only occurs in a minority of instances; data from the preantibiotic era indicate that about 5-7% of cases showed bone changes (Steinbock, 1976). The bony lesions permit
the identification of tuberculosis in skeletal remains, and indeed abundant palaeopathological evidence for the disease has accumulated from around the world (e.g. Morse, 1967; Steinbock, 1976; Buikstra, 1981; Roberts and Manchester, 1995; Ortner, 1999).
Today, most human cases of tuberculosis are caused by M. tuberculosis rather than M. bovis, but the extent to which this applies to earlier populations is unclear. Because M. tuberculosis is spread person to person by droplet infection, it is a population density-dependent disease (Manchester, 1984). On the other hand, infection with M. bovis is favored by close human-animal contact. Before effective steps were taken in the mid-20th century to eliminate tuberculosis from cattle, the disease was rife in British herds (Collins and Grange, 1983), and the proportion of tuberculous disease in the human population due to M. bovis was much greater than it is today (Rich, 1951).
It has been suggested (Manchester, 1991) that in antiquity there was a baseline of M. bovis infection, upon which was superimposed M. tuberculosis disease. Although disease caused both by M. tubercu-
*Correspondence to: S. Mays, Ancient Monuments Laboratory, English Heritage, Fort Cumberland, Fort Cumberland Road, Eastney, Portsmouth P04 9LD, UK. E-mail: simon.mays@english-heritage.org.uk
Received 16 February 2000; accepted 19 November 2000.
8 2001 WILEY-LISS, INC.
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
299
losis and M. bovis is likely to have affected rural and urban populations in the past, the balance between the two may have differed in different communities, M. tuberculosis being favored in crowded urban centres, while small, scattered, agrarian communities may have harbored a greater proportion of cases due to M. bovis. M. tuberculosis disease may have risen in importance through time as a consequence of increasing urbanization and aggregation of population (Manchester, 1991).
From the point of view of tracing the history of tuberculosis, it is clearly of interest to investigate the balance between disease caused by the two bacilli in earlier human groups. Although disease due to M. tuberculosis and due to M. bovis may each affect the skeleton (Davies et al., 1984), it is impossible to distinguish the two forms of tuberculous disease using osteological criteria, as bony lesions caused by the two pathogens are anatomically similar (Ortner, 1999). However, this distinction is possible using ancient DNA. Analysis of human remains has demonstrated that fragments of M. tuberculosis complex DNA may survive over several millennia both in soft-tissue remnants (Salo et al., 1994; Nerlich et al., 1997; Pap et al., 1999), and in bone (Spiegelman and Lemma, 1993; Arriaza et al., 1995; Baron et al., 1996; Taylor et al., 1996; Baxar-ias et al., 1998; Braun et al., 1998; Faerman et al., 1999; Horáčková et al., 1999; Haas et al., 1999; Nuorala, 1999; Dutour et al., 1999) in remains of infected individuals. A variety of modern genomic techniques discriminate between members of the M. tuberculosis complex (Espinoza de los Monteros et al., 1998; Kamerbeek et al., 1997; Aranaz et al., 1996), and recent work (Taylor et al., 1999) has demonstrated the potential value of applying such techniques in paleopathological cases.
The purpose of the present work was to investigate tuberculous disease in an early cemetery sample. Likely cases of tuberculosis were identified using conventional paleopathological examination. Analyses of DNA extracted from these skeletons were then used, firstly to try to confirm the paleopathological diagnosis of tuberculosis, and secondly to attempt to determine, in individual cases, whether M. tuberculosis or M. bovis was the causative agent. In an attempt to investigate the presence of tuberculosis in the farmyard animals of this community, some animal bone fragments from the settlement associated with the cemetery, bearing bony signs of infection, were also analyzed.
MATERIALS
The skeletal material for this study comes from the rural agrarian site at Wharram Percy, England. Wharram Percy is a deserted medieval village, situated in the Yorkshire Wolds, an area of chalk uplands in northern England (Fig. 1). The site was subject to a long-running research excavation (Be-resford and Hurst, 1990). An important focus of the fieldwork at Wharram Percy was St Martin's church
Fig. 1.   Location of Wharram Percy.
and churchyard; excavations here resulted in the recovery of 687 articulated human skeletons. The burials, which date primarily from the 10th-16th century AD, are mainly of ordinary peasants who lived at the village of Wharram Percy or at other settlements and farmsteads in this rural parish. Several aspects of the Wharram Percy collection render it particularly suitable for present purposes. The chalk geology at the site has resulted in excellent macroscopic bone preservation, facilitating paleopathological study. Only those resident in the parish were accorded burial in the churchyard, so the skeletal material represents a geographically denned population. As well as the churchyard, large areas of the settlement have been subject to archaeological investigation. This provides additional background information concerning living conditions, aiding interpretation of the paleopathological data.
OSTEOLOGICAL METHODS
All human burials from Wharram Percy have been examined osteologically by one of us (S.M.). Nine show signs of tuberculosis, and are the subject of the present work. Pathological changes were recorded using gross and radiographic examination. Tuberculosis was diagnosed according to the morphology of the lesions and their skeletal distribution.
300
S. MAYS ET AL.
Account was taken of the diagnostic criteria presented by Ortner and Putschar (1985), Steinbock (1976), Aufderheide and Rodriguez-Martin (1998), and Kelley and El-Najjar (1980). In brief, the more important of these include: predilection of lesions for the lumbar and thoracic vertebral bodies and the large joints of the appendicular skeleton; lesions that are mainly destructive, producing cavitation into the cancellous bone of the long-bone ends or vertebral bodies with little new bone formation; sequestra are characteristically absent; frequent vertebral body collapse with spinal kyphosis; and spinal disease is generally monofocal.
All nine burials showing evidence of tuberculosis are adult. Sex was established using dimorphic aspects of the pelvis and skull (Brothwell, 1981). Age at death was estimated using dental wear, calibrated using the juvenile part of the assemblage (Brothwell, 1981; Mays et al., 1995; Miles, 1963).
Study of faunal remains from Wharram Percy (Stevens, 1992; Pinter-Bellows, 1992) has indicated the presence of a number of cattle bone fragments showing signs of infection in the form of periostitis. Although tuberculosis in cattle may on occasion affect the bones (Lignereux and Peters, 1999), adequate paleopathological criteria for the diagnosis of the disease in faunal remains from archaeological sites have yet to be formulated. In cattle, as in man, chronic lesions in the lungs are a prominent feature of tuberculous disease (Hardie and Watson, 1992). In humans, it appears that chronic pulmonary tuberculosis may be associated with periosteal reactions on the ribs (Kelley and Micozzi, 1984; Roberts et al., 1994), so it seems reasonable to speculate that this may also be the case for cattle, although there is currently no evidence available to confirm this (Lignereux and Peters, 1999). Although rib periostitis cannot be considered diagnostic for pulmonary tuberculosis in cattle, periostitic bovine rib fragments probably form the most promising faunal remains available for biomolecular analysis as far as the present purposes are concerned. Three cattle rib fragments with periostitis were located in the Wharram Percy faunal collections; these were subjected to DNA analyses.
BIOMOLECULAR METHODS
DNA extracts were analyzed for the presence of M. tuberculosis complex. A number of further analyses were performed to discriminate between M. tuberculosis and M. bovis. Two human skeletons with no bony infectious lesions, and for which there were clear indications that death was unrelated to infectious disease (both individuals showed fatal cranial trauma), were used as controls. The DNA study was performed by G.M.T. in the Department of Medical Microbiology, Imperial College, London.
The skeletal lesions of brucellosis may resemble those of tuberculosis. Like tuberculosis, skeletal brucellosis shows a predilection for the vertebral bodies and large joints, and forms lytic lesions with
little reactive bone except in the healing phase (Ortner and Putschar, 1985). The spread of brucellosis to human subjects depends upon contact with animal vectors, so those in agrarian communities like Wharram Percy might be considered to have been at risk. In consideration of the above, analysis for brucella DNA was also performed on the nine diseased human skeletons.
The samples for biomolecular analysis were taken from cancellous bone from the interior of a diseased element from each of the nine human suspected cases of tuberculosis. Cancellous bone tissue was also taken from the two human skeletons serving as controls. Bone samples, in the form of surface scrapings, were taken from the three bovine rib fragments showing periostitis. DNA was extracted from bone powder (100-200 mg) using a Nuclisens® guani-dinium/silica kit (Organon Teknica, Boxtel, The Netherlands). A total of three extracts was prepared from all but one of the suspected tuberculosis cases; the exception was burial NA046, where only a single extract was prepared. A single extract was prepared from the bovine rib samples.
Polymerase chain reaction (PCR)
General PCR conditions, reagents, and thermal cyclers were common to all methods and have been described in detail elsewhere (Taylor et al., 1999). In heminested methods, 1 julI of first-round products was used to "spike" the second amplification step.
Measures were taken to avoid contamination, as described in detail elsewhere (Taylor et al., 1996, 1997). Briefly, bone extracts were prepared in a laboratory previously unused for tuberculosis research. Thereafter, a two-laboratory, three-workstation strategy was adopted so that PCR setup, amplification, and analysis of products were all performed in different areas. Autoclaved disposable plasticware, dedicated reagents, and pipettes with aerosol-resistant tips were used. Extraction and template blanks were included in every PCR assay, and were always free from product, indicating that contamination was well-controlled.
The procedures detailed below were carried out on the human bone samples. Two of these procedures, PCR for IS6110 and RD7, were also carried out on the three animal bone samples.
Mycobacterium tuberculosis complex.   In an
attempt to confirm the palaeopathological diagnosis of tuberculosis, PCR analyses for the RNA polymerase beta gene (rpoB) and the repetitive insertion element IS6110 were performed to detect M. tuberculosis complex.
rpoB. A pair of biotinylated primers was used to amplify a 157-base pair (bp) fragment of the RNA polymerase beta (rpoB) gene, accession number L27989. The sequence of the primers, and analysis of the product using the INNO-LiPA line probe assay to confirm M. tuberculosis complex, have been
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
301
described recently for other samples (Taylor et al., 1999).
IS6110. A nested PCR for the IS6110 insertion element was performed as previously described (Taylor et al., 1996). This is a modification of the method originally reported by Thierry et al. (1990).
Species-defining polymorphisms. The following procedures were carried out in an attempt to distinguish M. tuberculosis from M. bovis.
mtp40. PCR for mtp40 was performed as previously described (Taylor et al., 1999). The mtp40 element can be used to identify differences within the M. tuberculosis complex; it is present in the vast majority of M. tuberculosis isolates and absent from the great majority of M. bovis isolates (Liebana et al., 1996; Weil et al., 1996). Absence of mtp40 from M. bovis strains is associated with a deletion event, RD5, which is 8,964 bp in size and located between genome positions 2,626,067-2,635,031 on the H37Rv reference strain (Gordon et al., 1999).
oxyR. Primers were designed to amplify a fragment from the oxyR pseudogene of the M. tuberculosis H37Rv strain (accession number U16243). A heminested procedure was used. The sequence of the forward primer (Fl) was 5' -CGCGCTGTCAGAGCT-GACTTT-3', and of the reverse primer (Rl) 5'-TCT-GCGGAATCAGTGTCACC-3'. Forty-three cycles of amplification were performed. Cycling conditions were 94°C for 20 sec, 66°C for 30 sec, and 72°C for 30 sec, generating a 150-bp product. In the heminested second round (127-bp product), primer Rl was used with an inner forward primer F2 of sequence 5'-TTGTGACTGCATGAGGGGC-3'. A further 35 cycles of amplification were performed. The stages were 94°C for 10 sec, 62°C for 30 sec, and 72°C for 15 sec. A final elongation cycle was performed at 72°C for 2 min. In some experiments, inner primers F2 and Rl were used for 43 cycles with 5 julI of bone extract to assess the requirement for heminested PCR.
pncA The pyrazinamidase (pncA) gene is involved in susceptibility or resistance to the antitu-berculous drug pyrazinamide. A heminested PCR method was evaluated. In first-round amplification, a forward primer (Fl) 5'-ATC-AGC-GAC-TAC-CTG-GCC-GA-3' was used with reverse primer (Rl) 5'-GAT-TGC-CGA-CGT-GTC-CAG-AC-3' followed, in the heminested procedure, with Fl and a second reverse primer (R2) 5'-GGA-GTA-CCG-CTG-ACG-CAA-TG-3'. These result in a final product of 140 bp. Cycling conditions were 43 cycles at 66°C for 30 sec, 72°C for 20 sec, and 94°C for 20 sec, followed by one cycle at 72°C for 3 min. In the heminested stage, 35 cycles of amplification were performed as for first round, except that the annealing temperature was 65°C.
Spoligotyping. Spoligotyping (spacer oligotyping) provides simultaneous detection and typing of M. tuberculosis complex members. Spoligotyping relies on PCR amplification of variable sequences specifically found within the direct repeat (DR) region of the genome of M. tuberculosis complex members, in combination with their hybridization to a checkerboard of known variable elements or spacers (Kamerbeek et al., 1997). The DR consists of repeated sequences 36 bp long, interspersed by nonrepetitive DNA spacers (35-41 bp long). Hybridization patterns vary between isolates. M. bovis strains can be distinguished from M. tuberculosis by the absence of the terminal five spacers (nos. 39-43) at the 3' end of the DR region. Spoligotyping was performed as described by Kamerbeek et al. (1997), with the modification of an increase in cycle number to 45 cycles.
M. bovis-speciRc fragments. PCR for an M. bovis-specific fragment was performed using primers JB21 and JB22, as described by Rodriguez et al. (1995). Cycling parameters were modified to 45 cycles at 94°C for 25 sec, 62°C for 30 sec, and 72°C for 30 sec.
A second M. bovis-speci&c PCR was developed to amplify shorter fragments of DNA. The target was a length of DNA flanking deletion region 2 (RD2). The primers span a 12.7-kb fragment of the M. tuberculosis genome which is not present in M. bovis (Zu-märraga et al., 1999a). This deletion is now known as RD7 and extends from bp 2,208,003-2,220,721 on the M. tuberculosis genome (Gordon et al., 1999). Under the PCR conditions used, no product is obtained from M. tuberculosis; in M. bovis the primers generate a band of 211 bp. The sequence of these was: (forward) 5'-ACTTCAGTGCTGGTTCGTGG-3' and (reverse) 5'-ATCTTGCGGCCCAATGAATC-3'. Cycling conditions were 43 cycles at 66°C for 30 sec, 72°C for 20 sec, and 94°C for 20 sec, followed by one cycle at 72°C for 2 min.
Brucellosis. The target for amplification was insertion sequence IS6501, first reported in Brucella ovis (Ouahrani et al., 1993). This occurs in all strains of Brucella bacteria so far tested, and is present in variable copy number (5-35) depending on the species. A heminested strategy was adopted. The primer sequences were (Fl) 5'-TAACCGAT-TATTTGTCGACGC-3' and (Rl) 5'-GCG-TGGAC-TTTCGATATGGT-3'. In the second round, primer Fl was used with the second reverse primer (R2) 5' -CATGGAAAAGCCTTTCCCAT-3'. Primers were synthesised on an Applied Biosystems model 381A DNA synthesiser. Amplification consisted of 35 cycles at 94°C for 40 sec, 60°C for 30 sec, and 72°C for 15 sec, followed by one cycle of 72°C for 2 min. The final product of 147 bp is situated from nucleotides 165-311 inclusive in accession no. X71024.
PCR minimum detection limits (Table 1)
Tuberculosis. The minimum detection limits for all the tuberculosis methods were established using
302
S. MAYS ET AL.
TABLE 1. Minimum detection limits of PCR methods applied to Wharram Percy cases1
Product
PCR method size (bp)    Cycles    Minimum detection
rpoB single round	157	43			150 fg
IS6110 nested	93	35	X	35	15 fg
mtp40 single round	152	43			1.5 pg
oxyR heminested	127	43	X	35	1.5 pg
oxyR inner primers	127	43			1.5 pg
M. bovis RD7	211	43			7.4 pg
pncA heminested	140	43	X	35	34 pg
Brucella heminested	147	35	X	35	34 pg
1 bp, base pairs; fg, femtogram (i.e.,			10	-15	g); pg, picogram (i.e.,
io-12 g).
serial dilutions of a stock extract containing 300 |jLg/ml of partially purified genomic DNA from the H37Rv strain of M. tuberculosis.
Sensitivity of the "in-house" M. bovis PCR for RD7 was similarly determined using serial dilutions of partially purified genomic DNA purified from heat-killed M. bovis strain AF 2122/97. A sensitivity of 10 femtograms has been reported for the M. bovis-spe-cific PCR method of Rodriguez et al. (1995).
Brucella. B. suis DNA clone (clone 1330), containing 675 |jLg/ml DNA, was used. Serial dilutions of the clone in high-pressure liquid chromatography (HPLC)-grade water were prepared to establish the minimum detection limits of the PCR. A 10~2 dilution served as a positive control in all PCR reactions.
The most sensitive tests, the IS6110 and rpoB PCRs, were applied to the samples from the two human skeletons used as controls.
Automated DNA sequencing
Cycle sequencing of PCR products was performed on a Hybaid Touchdown with the ABI Dye Terminator Ready Reaction Kit (Perkin-Elmer Applied Bio-systems), according to the manufacturer's protocol, with analysis on an ABI 310 Genetic Analyser.
Gel electrophoresis
Routine gel electrophoretic analysis of products was performed on either 2% or 3% w/v agarose gels. Products for sequencing were subsequently run on 0.8% w/v LMP agarose (Gibco BRL, Life Technologies). Bands were excised from the gel with a sterile scalpel blade and were purified using either the NucleiClean™ DNA isolation kit (Sigma-Aldrich) or the Geneclean II kit (Bio 101, CA).
RESULTS Osteological analyses
The age at death and sex of the humans showing signs of tuberculosis, together with a summary of the locations of bone lesions, are shown in Table 2. Some of the pathological changes are illustrated in Figure 2a-e. The most common locations for lesions were the spine and the hip, although other bones are also affected, most frequently the ribs (Table 2). In all cases where ribs were involved, lesions were located near the proximal ends and represent extensions of vertebral foci of infection. No evidence for periostitis on the visceral surfaces of ribs, of the form described by Kelley and Micozzi (1984) and Roberts et al. (1994) as reflecting pulmonary infection, was found among the suspected tuberculosis cases.
Each of the three bovine rib fragments showed periostitis. One specimen (S95 Context 2432) showed woven bone on its visceral surface; the other two (S85 Context 22 and S95B Context 2053) showed a mixture of woven and partially remodelled bone deposits on their lateral surfaces.
Biomolecular analyses
Bone samples from suspected human and animal cases and controls were screened for the presence of DNA from infecting bacteria, using PCR assays directed towards a series of genetic loci with denned specificity patterns. The results are summarized in Table 3.
TABLE 2. Summary of osteological results1
Lesions
Burial	Sex	Age	Date	Spine	Hip	Other
EE056	M?	50 +	900-1400 AD*	Yes (t, 1)	No	Yes (ribs)
G438	F	25-35	1060-1170 AD*	Yes (t, 1)	No	Yes (ulna, ribs, ilium)
G482	M	50 +	1060-1170 AD*	Yes (1)	No	No
NA026	F	35-45	900-1400 AD*	Yes (t)	-	Yes (ribs, scapula)
NA046	M	50 +	890-1170 AD	Yes (1)	No	No
NA112	M	50 +	890-1160 AD	No	Yes	No
NA197	M	50 +	900-1400 AD*	No	Yes	No
SAO 13	F	35-45	1270-1410 AD	Yes (t)	No	Yes (ribs)
WC0142	M	Adult	900-1700 AD*	Yes (1, s)	Yes	Yes (ilium, femora, tibiae
fibula, metatarsal, calcanei, foot phalanx)
1 Age, estimated age at death in years; Date, calibrated radiocarbon dates for individual burials (95% confidence limits), except * 95% confidence limits for dates of use of the part of the churchyard in which the burial was found (given where there is no date available on individual burial); Lesions, lower-case letters give location of lesions in spine: t, thoracic; 1, lumbar; s, sacral; —, part of skeleton missing.
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
303
rpoB. A region of the mycobacterial rpoB gene was amplified by PCR, and the product was analyzed for the presence of specific oligonucleotide sequences using the INNO-LiPA line probe assay. A PCR product was obtained in all eight samples tested from suspected cases. Only five of these produced the M. tuberculosis complex pattern during subsequent hybridization (Fig. 4). PCR product from the other three samples hybridized to a single band on the INNO-LiPA test strips, a result similar to that obtained in the case of nontuber-culous mycobacteria such as might be found in the environment. The two human control samples were negative for rpoB.
IS6110. All but two of the human suspected tuberculosis cases were positive for IS6110, a specific marker for the M. tuberculosis complex (Fig. 3). The three animal bone specimens were negative for IS6110, as were the two human control samples. The IS6110 element is present in multiple copies in the genome of most M. tuberculosis isolates and in one or several copies in M. bovis. While detection of IS6110 identifies the presence of a member of the M. tuberculosis complex, it does not therefore distinguish M. tuberculosis from M. bovis.
mtp40. Seven cases were positive for mtp40, suggesting the presence of M. tuberculosis rather than M. bovis.
oxyR The sequence of the oxyR pseudogene has been found to differ at nucleotide 285. In M. tuberculosis isolates there is a guanine at this position, and in M. bovis an adenine (Sreevatsan et al., 1996). Amplification of the oxyR pseudogene was positive in six cases; in each there was a guanine at position 285, consistent with identification of M. tuberculosis.
pncA The sequence of the pncA gene, encoding a pyrazinamidase enzyme, differs between M. tuberculosis (which is pyrazinamide-sensitive) and M. bovis (pyrazinamide-resistant) by a mutation at nucleotide 169. In M. tuberculosis strains base 169 is cytosine, whereas in M. bovis it is guanine (Scorpio and Zhang, 1996). Although a PCR reaction for pncA often generated bands at or near the expected size, sequencing was difficult due to double or multiple base-calling. This may have been due to amplification of related genes from contaminating environmental bacteria, and precluded the use of the pncA assay for discriminatory typing.
Spoligotyping. Three cases were successfully spoligotyped, generating patterns suggestive of M. tuberculosis in each case (Fig. 5). Burial NA112 was spoligotyped twice from the same extract, and good reproducibility of the pattern was observed. However, typing from two separate extracts from burial NA026 yielded patterns with several discrepancies.
This could arise if different fragments of the DR region happened to be preserved in different parts of the bone sample, and raises a note of caution in interpretation of spoligotype patterns from ancient DNA.
M. bovis-specific fragments. Rodriguez et al. (1995) described a PCR reaction for M. bovis, using primers JB21 and JB22. This assay produced no bands of the expected size when applied to the human bone samples. However, this method is based on amplification of fragments of about 500 bp, and a negative result could reflect the extent of DNA degradation in the Wharram Percy material. Nevertheless, PCR for the shorter 211-bp fragment RD7 of M. bovis was also negative in all the Wharram Percy samples we examined, allowing some confidence that M. bovis was truly absent. Preliminary experiments confirmed that this method amplified M. bovis strain AF 2122/97, M. bovis BCG, and several clinical isolates causing disease, but failed to amplify the H37Rv and CDC1551 reference strains of M. tuberculosis. PCR for the RD7 fragment was also applied to the three animal bone samples, and in each case proved negative.
IS6501. All human samples were negative by heminested PCR for the insertion sequence IS6501, present in multicopy in all species of Brucella which produce disease in humans (Ouahrani et al., 1993). Preliminary experiments (not shown) confirmed that the assay detected various biovars of B. suis, B. ovis, B. melitensis, and B. abortus, as well as B. neotomae, B. canis, and B. maris.
In summary, for all nine of the suspected tuberculosis cases in the human skeletons, evidence from at least two distinct genetic loci allowed identification of the presence of DNA from organisms belonging to the M. tuberculosis complex. DNA from other related organisms was detected in some assays (e.g., the rpoB amplification) and is most probably a result of contamination of the samples with environmental bacteria. Each of the nine cases gave one or more positive results in assays designed to distinguish M. tuberculosis from M. bovis. In every case, the observed genotype resembled present-day M. tuberculosis rather than M. bovis. There was no evidence for Brucella infection.
The three animal bone specimens proved negative for the RD7 PCR, and for the analysis for IS6110, the most sensitive of the methods used in this work for detection of M. tuberculosis complex.
DISCUSSION
The biomolecular analyses confirmed the presence of M. tuberculosis complex in each of the nine humans in whom tuberculous disease was suspected on osteological grounds. This adds to a growing corpus of paleopathological cases in which osteological diagnosis of tuberculosis from dry bones has been supported by DNA evidence (e.g., Baron et al., 1996;
304
S. MAYS ET AL.
Fig. 2.
Taylor et aL, 1996; Baxarias et aL, 1998; Braun et aL, 1998; Haas et al., 1999). The concordance between osteological and biomolecular results strengthens the notion that tuberculosis can be reliably diagnosed osteologically in ancient human skeletons, despite the fact that many conditions produce skeletal lesions which may, to varying degrees, resemble it (Buikstra and Cook, 1981, their Table 2; Aufderheide and Rodriguez-Martin, 1998).
In most instances, the osteological lesions clearly fulfilled the criteria for diagnosis of tuberculosis pre-
sented in the published literature. The exception to this was burial NA046, where the sole pathological lesions consisted of superficial cavitation on the body of the fourth lumbar vertebra (Fig. 2c) and a similar, smaller erosion on the fifth. Our favored diagnosis on osteological grounds was that these represented early paradiskal lesions of tuberculosis, on the basis of their morphological resemblance to lesions of this type identified in skeletal remains from documented cases of tuberculosis (cf. Kelley and El-Najjar, 1980, their Fig. 1). However, other
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
305
Fig. 2. Osseous lesions suggestive of tuberculosis in some of the Wharram Percy skeletons, a: Burial G438. Ninth thoracic through second lumbar vertebrae, showing destruction and collapse of vertebral bodies with spinal kyphosis, b: Burial SA013. Third through seventh thoracic vertebrae. T4-6 show erosions on the lateral/anterior aspects of their bodies, c: Burial NA046. Fourth lumbar vertebra, showing a small lytic lesion on the right side of the upper surface of its body, d: Burial SA013. Left sixth rib, showing lytic lesions at its head, e: Burial NA197. Right innominate, showing destruction of the acetabulum.
TABLE 3. Summary of biomolecular results1
TB       INNO LiPA TB TB DR region Brucella
rpoB    (rpoB product)    IS6110    mtp40     TB oxyR spoligotype M. bovis IS6501
Burial number (2) (1) (3) (3)        285 (1) (1) RD7 (1) (1)
Human suspected tuberculosis cases
EE056	+(2)	M. tb complex	+(3)	+(2)	—	—	nd	-(1
G438	+(2)	M. tb complex	+(3)	+(1)	+(G)	nd	-(1)	-(1
G482	+(2)	Atypical pattern	+(3)	+(2)	+(G)	—	-(1)	~(1
NA026	+(2)	M. tb complex	+(3)	-(2)	+(G)	M. tb	-(1)	-(1
NA046	nd	nd	+(1)	-(1)	+(G)	M. tb	-(1)	-(1
NA112	+(2)	Atypical pattern	-(3)	+(2)	+(G)	M. tb	nd	~(1
NA197	+(2)	M. tb complex	+(3)	+(1)	—	—	-(1)	-(1
SAO 13	+(2)	M. tb complex	-(3)	+(2)	+(G)	—	-(1)	-(1
WC0142	+(2)	Atypical pattern	+(1)	+(1)	—	—	-(1)	~(1
Control samples2								
EE003	-(1)	nd	-(1)	nd	nd	nd	nd	nd
G304	-(1)	nd	-(1)	nd	nd	nd	nd	nd
Animal bone samples2								
S85, Context 22	nd	nd	-(1)	nd	nd	nd	-(1)	nd
S95, Context 2432	nd	nd	-(1)	nd	nd	nd	-(1)	nd
S95B, Context 2053	nd	nd	-(1)	nd	nd	nd	-(1)	nd
1 Numbers in parentheses at heads of columns are number of extracts tested by that PCR. Numbers in parentheses after results are numbers positive ( + ) or negative ( —). tb, tuberculosis; (G), guanine; nd, not determined.
2 One extract prepared.
diagnoses could not be confidently excluded on lesion morphology alone. For example, these lesions might also be consistent with other infections (e.g., brucellosis: Ortner and Putschar, 1985), malignant neoplastic disease (e.g., metastatic carcinoma: Kelley and El-Najjar, 1980), or trauma (Maat and Mastwijk, 2000).
Although the osteological indicators were ambiguous, NA046 was positive for three markers of the M. tuberculosis complex: IS6110, spoligotyping, and oxyR. The results from the latter two analyses indicated the presence of M. tuberculosis rather than M. bovis. The lack of an mtp40 PCR product in this instance may simply reflect the lower sensitivity of
306
S. MAYS ET AL.
WT
WT
Alypio&l
Alyph ■ M
Fig. 3. IS6110 PCR. Gel electrophoretic analysis of nPCR products of 2% agarose gel. Lanes 1 and 12, DNA 100-bp ladders. Lane 2, water blank. Lane 3, H37Rv positive control. Lane 4, burial G438. Lane 5, burial NA026. Lane 6, burial WC0142. Lane 7, burial NA112. Lane 8, burial G482. Lane 9, burial NA197. Lane 10, burial EE056. Lane 11, burial SA013. Arrow denotes IS6110 PCR product.
this method compared to IS6110 (Table 1). Although the biomolecular results mean that we can be confident that this case was positive for M. tuberculosis, it is possible, given that the great majority of tuberculosis cases show no bony signs (Steinbock, 1976), that NA046 had tuberculous infection which had not affected the skeleton and that the spinal lesions simply reflect some unrelated, coincidental condition. However, given their morphological resemblance to lesions in documented cases of tuberculosis, a more parsimonious explanation is that the spinal changes are indeed early tuberculous lesions, and that the DNA analyses in this instance enabled a firm diagnosis which would otherwise have remained elusive.
Our experience with several different PCR methods on the Wharram Percy human bone material highlights three general points relating to the analysis of ancient DNA. Firstly, it confirms that it is advantageous to use methods that are targeted to amplify shorter fragments, which are likely to survive in ancient material. Secondly, the observation that not all PCRs were successful in each specimen indicates the value of applying several PCR techniques and of repeat sampling. Finally, it is important to be able to detect pathogen DNA in mixed samples containing contaminating soil microorganisms; this proved problematic in the case of rpoB and pncA assays in the present study.
1
SA13
Atypical
EE 56
WT
NA197
WT
G482
WT control
NA 112
Khi-I,
WCO 142
NA 26
GA3&
H37Hv
Waler
Fig. 4. rpoB PCR. Analysis oirpoB PCR products from eight of the suspected tuberculosis cases, using in INNO LiPA rif.TB kit. Five cases show typical wild-type (WT) pattern, indicative of rifampicin-sensitive M. tuberculosis complex strains (hybridization to probes S1-S5), and three show an atypical pattern, with hybridization only to probe S2.
Severe diagenesis, of the type thought to be caused by soil-dwelling microorganisms, has consistently been found by one of us (G.T.-W.) in histological sections of medieval human bone from Wharram Percy (Fig. 6). Severe alteration of bone by microbiological attack is typical of sites like Wharram Percy, which lie on calcareous geology (Millard and Hedges, 1995). Although it has been suggested (Hagelberg et al., 1991; Colson et al., 1997) that there may be a correlation between DNA survival and the quality of the bone microstructure, the present results demonstrate the potential for analysis of ancient pathogen DNA to produce useful pa-leopathological data, even in bone where preservation at the histological level is poor. This may be consistent with the idea that the tough cell wall characteristic of mycobacteria aids survival of their DNA during the decay process (Spiegelman and Do-noghue, 1999).
All nine tuberculosis cases from the Wharram Percy churchyard appear to represent infection with M. tuberculosis rather than M. bovis. The only other study of which we are aware which has successfully distinguished M. tuberculosis and M. bovis disease in ancient human remains is an earlier report from our group (Taylor et al., 1999), in which two British medieval cases of tuberculosis were studied and, as in the present work, M. tuberculosis rather than M. bovis was found to be the causative agent. However, in contrast to the present study, those burials came from a cemetery from a large, crowded urban center,
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
307
1
10
20
30
40
Fig. 5. Spoligotyping. Hybridization analysis of the DR region of M. tuberculosis complex organisms amplified from three Wharram Percy cases and controls. Rows 1 and 2, burial NA026 analyzed from two different extracts. Rows 3 and 4, burial NA112 analyzed twice from the same extract. Row 5, burial NA046. Row 6, water blank. Row 7, M. bovis BCG control. Row 8, M. tuberculosis H37Rv strain. Row 9, soil extract. The numbers along the bottom are spacer numbers.
f-. i.■■.   k«» ass
.....-1öia|~
3*;
1
Fig. 6. Scanning electron micrographs from bone from a Wharram Percy burial. Although the outer 100 u,m of bone preserve some histological detail, across most of the section advanced diagenesis is apparent. The great majority of histological structures have been destroyed or obscured by postmortem alteration of bone tissue, so that in general only gross features such as Haversian canals remain visible. The severe diagenetic changes seen in this specimen are typical of those in medieval human bone from Wharram Percy.
where it was no surprise to confirm the presence of M. tuberculosis.
Analysis of historical samples of M. bovis DNA will be of particular interest in the context of the rapidly expanding knowledge generated by charac-
terization of contemporary mycobacterial genomes. While it is common speculation that M. tuberculosis adapted for human infection following transfer from domesticated livestock, recent characterization of genomic deletion events casts doubt on this scenario
308
S. MAYS ET AL.
and suggests that it is M. bovis rather than M. tuberculosis which has undergone more extensive adaptation (Behr et al., 1999; Gordon et aL, 1999). The time-scale for these deletion events is unknown; the possibility should at least be considered that medieval M. bovis might have more closely resembled contemporary M. tuberculosis or a common progenitor of the M. tuberculosis complex. It is known from studies of contemporary M. bovis isolates that some do share molecular markers more commonly associated with M. tuberculosis strains. Examples include M. bovis isolates retaining the mtp40 element (Weil et al., 1996) and isolates from sealions and fur seals with pncA and oxyR polymorphisms usually associated with M. tuberculosis (Zumarraga et al., 1999b). Due to the variety of methods which have been used in the present study to differentiate M. bovis, it seems unlikely that this possibility could explain our findings at Wharram Percy. Further, all isolates of M. bovis we have previously examined, including those from fur seal, have tested PCR-pos-itive for RD7 (Taylor, unpublished observations). This implies erasure of the 12.7-kb fragment downstream from M. bovis deletion region 2 (RD2) (Zumarraga et al., 1999a). Deletion of this 12.7-kb fragment from the M. bovis genome may have been an early event in the evolution of the M. tuberculosis complex.
At Wharram Percy it is clear that cattle formed an important part of the site's economy. Analyses of animal bones recovered from the settlement (Ryder, 1974; Stevens, 1992; Pinter-Bellows, 1992) indicate that sheep formed about 40-60% of the faunal remains, and cattle about 30-40%, with pigs 11% or less. Cattle were likely the most important source of meat. Finds of the remains of pottery vessels which would have been used for dairy products attest to the inclusion of these foods in diets (Beresford and Hurst, 1990). It was usual in the medieval period for humans and livestock to share the same living space (Lamond, 1890), and at Wharram Percy, excavations have uncovered the remains of peasant long-houses (Beresford and Hurst, 1990) which would have housed both human and bovine inhabitants, cattle being stalled at one end while the family lived at the other (Fig. 7).
In a community like Wharram Percy, where people had close contact with cattle, and where beef and dairy products formed important parts of the diet, it might be expected that there would have been significant risk of human infection from M. bovis. It is therefore perhaps surprising that all cases of tuberculosis identified here appear to have been caused not by M. bovis but by M. tuberculosis.
One possible explanation for the absence of M. bovis disease is simply that the Wharram Percy cattle may have been free of tuberculosis. Despite the evidence that prior to the steps taken to eliminate tuberculosis in cattle in the 20th century it was rife in British herds (Collins and Grange, 1983), it cannot be stated with certainty that it was present
Fig. 7. Reconstruction of a peasant long-house at Wharram Percy. These dwellings housed both human and bovine inhabitants. ®English Heritage Photo Library.
in cattle in medieval times. The three bovine rib samples were analyzed in an attempt to determine whether there was any evidence for tuberculosis among Wharram Percy cattle. That results proved negative may reflect poor DNA survival or that the periostitis shown by these specimens was due to some other disease. The results obtained do not therefore allow us to comment on whether or not tuberculosis was present in domestic stock at Wharram Percy.
If it is accepted that domestic stock were not the main source of tuberculosis among the human inhabitants at Wharram Percy, alternative models for infection from human sources need to be considered. As was stated above, spread of pulmonary tuberculosis is favored by high population densities. The population density of the parish served by Wharram Percy church is difficult to determine precisely, and varied greatly during the medieval period. Nevertheless, estimates of approximately 1-6 persons per square kilometer (calculated from figures presented in Bell, 1987), with a maximum reached during the late 13th-early 14th century, are probably not too wide of the mark. This is in line with population density figures estimated for Yorkshire as a whole in the medieval period (Smith, 1988). These figures are indicative of a sparsely populated area by medieval standards (Smith, 1998).
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
309
Rural Yorkshire would not appear to have provided an ideal environment for a population density-dependent disease such as pulmonary tuberculosis. Nevertheless, it does appear that tuberculosis can be maintained even in scattered populations (Black, 1975), so it might be suggested that the human form of the disease was endemic among rural communities in this area in the medieval period. However, the degree to which tuberculosis can be maintained in human groups of small size in the absence of an extrinsic reservoir of relatively large size is unclear (Buikstra and Cook, 1981). Population aggregates in urban centers may act as reservoirs of tuberculous infection, which may then "pulse out" to small rural communities in their hinterlands (Buikstra and Cook, 1981). Wharram Percy lies about 20 miles from York, which, with a population of about 12,000 in the medieval period (Goldberg, 1986), was easily the largest regional center, and was indeed one of the largest cities in England. Excavated human remains from York showing tuberculous lesions (Stroud and Kemp, 1993) confirm the presence of the disease there in medieval times.
In the medieval period, populations in urban settlements were maintained by immigration from surrounding rural areas. Russell (1948) offers documentary evidence that Wharram Percy would have lain well within the hinterland for migration of individuals to become resident in York, and analysis of skeletal demographic data (Mays, 1997a) suggests emigration from Wharram Percy to urban centers such as York. Ceramic and other artifactual evidence from Wharram Percy indicates trade with local towns such as Malton (Hurst, 1984). At Wharram Percy, a few remains of marine fauna (Ryder, 1974) have been recovered, and consistent with this, carbon stable isotope determinations (Mays, 1997b) have shown that marine foods were a significant source of protein in human diets here. Marine foods may have been obtained direct from fishing ports, such as Scarborough or Hull, or indirectly from trade with inland market towns.
The above serves to remind us that, despite its rural location, Wharram Percy was not an isolated community, but rather had a network of social and economic links with a range of other settlements. Perhaps contact with reservoirs of the disease in larger population centers may have helped maintain tuberculosis in the countryside. If the Wharram Percy results are anything to go by, it may be that models such as this, which invoke person-to-person transmission of disease, are more appropriate for understanding rural tuberculosis in the medieval period than reference to local animal vectors.
ACKNOWLEDGMENTS
The Brucella PCR was developed and validated by Mary Crossey (Department of Medicine, St. Mary's NHS Trust, London), in collaboration with Dr. Steve Spencer (Brucella Research Group, Veterinary Laboratory Agency, Weybridge Surrey, UK), who kindly
supplied the B. suis clone. Thanks are due to Stephanie Pinter-Bellows for access to some of her unpublished data and for help in locating the pathological animal bones.
LITERATURE CITED
Aranaz A, Liebana E, Mateos A, Dominguez L, Vidal D, Domingo M, Gonzolez O, Rodriguez-Ferri EF, Bunschoten AE, van Em-bden JD, Cousins D. 1996. Spacer oligonucleotide typing of Mycobacterium bovis strains from cattle and other animals: a tool for studying epidemiology of tuberculosis. J Clin Microbiol 34:2734-2740.
Arriaza BT, Salo W, Aufderheide AC, Holcomb TA. 1995. Pre-Columbian tuberculosis in northern Chile: molecular and skeletal evidence. Am J Phys Anthropol 98:37-45.
Aufderheide AC, Rodriguez-Martin C. 1998. The Cambridge encyclopaedia of human paleopathology. Cambridge: Cambridge University Press.
Baron H, Hummel S, Herrmann B. 1996. Mycobacterium tuberculosis complex DNA in ancient human bones. J Archaeol Sci 23:667-671.
Baxarias J, Garcia A, Gonzalez J, Perez-Perez A, Tudo BG, Gar-cia-Bour CJ, Campillo DD, Turbon ED. 1998. A rare case of tuberculous gonarthropathy from the middle ages in Spain: an ancient DNA confirmation study. J Palaeopathol 10:63-72.
Behr MA, Wilson MA, Gill WP, Salamon H, Schoolnik GK, Rane S, Small SM. 1999. Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science 284:1520-1523.
Bell R. 1987. Wharram. A study of settlement on the Yorkshire Wolds. Volume III Wharram Percy: the church of St. Martin. Society for Mediaeval Archaeology monograph series no. 11. London: Society for Mediaeval Archaeology.
Beresford M, Hurst J. 1990. Wharram Percy: deserted mediaeval village. London: Batsford/English Heritage.
Black FL. 1975. Infectious disease in primitive societies. Science 187:515-518.
Braun M, Cook DC, Pfeiffer S. 1998. DNA from Mycobacterium tuberculosis complex identified in North American, pre-Columbian human skeletal remains. J Archaeol Sci 25:271-277.
Brothwell DR. 1981. Digging up bones, 3rd ed. Oxford: Oxford University Press/British Museum (Natural History).
Buikstra JE, editor. 1981. Prehistoric tuberculosis in the Americas. Northwestern University Archaeological Program, scientific papers no. 5. Evanston: Northwestern University.
Buikstra JE, Cook DC. 1981. Pre-Columbian tuberculosis in west-central Illinois: prehistoric disease in biocultural perspective. In: Buikstra JE, editor. Prehistoric tuberculosis in the Americas. Northwestern University Archaeological Program, scientific papers no. 5. Evanston: Northwestern University, p 115-139.
Collins CH, Grange JM. 1983. The bovine tubercle bacillus.
J Appl Microbacteriol 55:13-29. Colson I, Bailey JF, Vercauteren M, Sykes B, Hedges REM. 1997.
The preservation of ancient DNA and bone diagenesis. Ancient
Biomol 1:109-117. Davies PD, Humphries MJ, Byfield SP, Nunn AJ, Darbyshire JH,
Citron KM, Fox W. 1984. Bone and joint tuberculosis. A survey
of notifications in England and Wales. J Bone Joint Surg [Br]
66:326-330.
Dutour O, Palfi G, Brun J-P, Berato J, Panuel M, Haas CJ, Zink A, Nerlich AG. 1999. Morphological, palaeoradiological and palaeomicrobiological study of a French mediaeval case of tuberculous spondylitis with cold abscess. In: Palfi, Deak J, Hutas I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 395-400.
Espinoza de los Monteros LE, Galan JC, Gutierrez M, Samper S, Garcia Marin JF, Martin C, Dominguez L, de Rafael L, Ban-quero F, Gomez-Mampaso E, Blazquez J. 1998. Allele-specific PCR method based on pncA and oxyR sequences for distinguishing Mycobacterium bovis from Mycobacterium tuberculosis: intra-specific M. bovis pncA polymorphism. J Clin Microbiol 36:239-242.
310
S. MAYS ET AL.
Faerman M, Jankauskas R, Gorski A, Bercovier H, Greenblatt CL. 1999. Detecting Mycobacterium tuberculosis DNA in mediaeval skeletal remains from Lithuania. In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 371-376.
Goldberg PJP. 1986. Female labour, service and marriage in the late mediaeval urban north. North Hist 22:18-38.
Gordon SV, Brosch R, Billaut A, Gamier T, Eiglemeier K, Cole ST. 1999. Identification of variable regions in the genomes of tubercle bacilli using bacterial artificial chromosome arrays. Mol Microbiol 32:643-655.
Haas CJ, Zink A, Molnár E, Marcsik A, Dutour O, Nerlich AG, Pálfi G. 1999. Molecular Evidence for Tuberculosis in Hungarian Skeletal Samples. In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 385-391.
Hagelberg E, Bell LS, Allen T, Boyde A, Jones SJ, Clegg JB. 1991. Analysis of ancient bone DNA: techniques and applications. Philos Trans R Soc Lond [Biol] 333:399-407.
Hardie RM, Watson JM. 1992. Mycobacterium bovis in England and Wales: past, present and future. Epidemiol Infect 109:23-33.
Horáčková L, Vargová L, Horváth R, Bartoš M. 1999. Morphological, roentgenological and molecular analyses in bone specimens attibuted to tuberculosis, Moravia (Czech Republic). In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 413-417.
Hurst JG. 1984. The Wharram research project: results to 1983. Mediaeval Archaeol 28:77-111.
Kamerbeek J, Schouls L, Kolk A, van Agterveld M, van Soolingen D, Kuijper S, Bunshoten A, Molhuizen H, Shaw R, Goyal M, van Embden J. 1997. Simultaneous detection and strain differentiation of Mycobacterium tuberculosis for diagnosis and epidemiology. J Clin Microbiol 35:907-914.
Kelley MA, El-Najjar MY. 1980. Natural variation and differential diagnosis of skeletal changes in tuberculosis. Am J Phys Anthropol 52:153-167.
Kelley MA, Micozzi MS. 1984. Rib lesions in chronic pulmonary tuberculosis. Am J Phys Anthropol 65:381-386.
Lamond D. 1890. Walter of Henley's husbandry. London: Longman Green.
Liebana E, Aranaz A, Francis B, Cousins D. 1996. Assessment of genetic markers for species differentiation within the Mycobacteria tuberculosis complex. J Clin Microbiol 34:933-938.
Lignereux Y, Peters J. 1999. Elements for the retrospective diagnosis of tuberculosis on animal bones from archaeological sites. In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 339-348.
Maat GJR, Mastwijk RW. 2000. Avulsion injuries of the vertebral endplates. Int J Osteoarchaeol 10:142-152.
Manchester K. 1984. Tuberculosis and leprosy in antiquity: an interpretation. Med Hist 28:162-173.
Manchester K. 1991. Tuberculosis and leprosy: evidence for interaction of disease. In: Ortner DJ, Aufderheide AC, editors. Human paleopathology. Current syntheses and future options. Washington, DC: Smithsonian Institution Press, p 23-35.
Mays S. 1997a. Life and death in a mediaeval village. In: de Boe G, Verhaege F, editors. Death and burial in mediaeval Europe. Zellik: Instituut voor het Archeologisch Patrimonium, p 121-125.
Mays S. 1997b. Carbon stable isotope ratios in mediaeval and later human skeletons from Northern England. J Archaeol Sei 24:561-567.
Mays S, de la Rua C, Molleson T. 1995. Molar crown height as a means of evaluating existing wear scales for determining age at death in human skeletal remains. J Archaeol Sei 22:659-670.
Miles AEW. 1963. The dentition in the assessment of individual age in skeletal material. In: Brothwell DR, editor. Dental anthropology. Oxford: Pergamon. p 191-209.
Millard AR, Hedges REM. 1995. The role of environment in uranium uptake by buried bone. J Archaeol Sei 22:239-250.
Morse D. 1967. Tuberculosis. In: Brothwell DR, Sandison AT, editors. Diseases in antiquity. Springfield, IL: Charles C. Thomas, p 249-271.
Nerlich AG, Haas CJ, Zink A, Szeimies U, Hagedorn HG. 1997. Molecular evidence for tuberculosis in an ancient Egyptian mummy. Lancet 350:1404.
Nuorala E. 1999. Tuberculosis on the 17th century man-of-war Kronau. Int J Osteoarchaeol 9:344-348.
O'Reilly LM, Daborn CJ. 1995. The epidemiology of Mycobacterium bovis infections in animals and man: a review. Tubercle Lung Dis [Suppl] 76:1-46.
Ortner DJ. 1999. Paleopathology: implications for the history and evolution of tuberculosis. In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 255-261.
Ortner DJ, Putschar WGJ. 1985. The identification of pathological conditions in human skeletal remains. Washington, DC: Smithsonian Institution Press.
Ouahrani S, Michaux S, Sri Widada J, Bourg G, Tournebize R, Ramuz M, Liautard J-P. 1993. Identification and sequence analysis of IS6501, an insertion sequence in Brucella spp.: relationships between genomic structure and the number of IS6501 copies. J Genet Microbiol 139:3265-3273.
Pap I, Józsa L, Repa I, Bajzik G, Lakhani SR, Donoghue H, Spigelman M. 1999. 18-19th century tuberculosis in naturally mummified individuals (Vác, Hungary). In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 421-428.
Pinter-Bellows S. 1992. The vertebrate remains from sites 94 and 95. In: Milne G, Richards JD, editors. Wharram. A study of settlement on the Yorkshire Wolds, volume VII. Wharram Percy: two Anglo-Saxon buildings and associated finds. York University archaeological publications 9. York: University of York, p 69-79.
Rich R. 1951. The pathogenesis of tuberculosis, 2nd ed. Springfield, IL: Charles C. Thomas.
Roberts C, Manchester K. 1995. The archaeology of disease, 2nd ed. Stroud: Sutton.
Roberts CA, Lucy D, Manchester K. 1994. Inflammatory lesions of ribs: an analysis of the Terry Collection. Am J Phys Anthropol 95:169-182.
Rodriguez JG, Mejia GA, Del Portillo P, Patarroyo ME, Murillo LA. 1995. Species-specific identification of Mycobacterium bovis by PCR. Microbiology 141:2131-2138.
Russell JC. 1948. British mediaeval population. Albuquerque: University of New Mexico Press.
Ryder ML. 1974. Animal remains from Wharram Percy. Yorkshire Archaeol J 46:42-52.
Salo WL, Aufderheide AC, Buikstra J, Holcomb TA. 1994. Identification of Mycobacterium tuberculosis DNA in a pre-Columbian Peruvian mummy. Proc Natl Acad Sei USA 91:2091-2094.
Scorpio A, Zhang Y. 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.
Smith R. 1988. Human resources. In: Astill G, Grant A, editors. The countryside of mediaeval England. Oxford: Blackwell, p 188-212.
Spiegelman M, Donoghue H. 1999. Mycobacterium tuberculosis DNA in archaeological specimens. In: Pálfi G, Dutour O, Deák J, Hutáš I, editors. Tuberculosis past and present. Budapest: Golden Books/Tuberculosis Foundation, p 353-360.
Spiegelman M, Lemma E. 1993. The use of polymerase chain reaction (PCR) to detect Mycobacterium tuberculosis in ancient skeletons. Int J Osteoarchaeol 3:137-143.
Sreevatsan S, Escalante P, Pan X, Gillies DA, Siddiqui S, Khalaf CN, Kreiswirth BN, Adams LG, Ficht T, Perumaalla VS, Cave MD, van Emden JD, Musser JM. 1996. Identification of a polymorphic nucleotide in oxyR specific for Mycobacterium bo-vis. J Clin Microbiol 34:2007-2010.
Steinbock RT. 1976. Palaeopathological diagnosis and interpretation. Springfield, IL: Charles C. Thomas.
Stevens P. 1992. The vertebrate remains from site 39. In: Milne G, Richards JD, editors. Wharram. A study of settlement on the
BIOMOLECULAR STUDY OF MEDIEVAL TUBERCULOSIS
311
Yorkshire Wolds, volume VII. Wharram Percy: two Anglo-Saxon buildings and associated finds. York University archaeological publications 9. York: University of York, p 67-69.
Stroud G, Kemp RL. 1993. Cemeteries of St. Andrew, Fishergate. The archaeology of York 12/2. York: York Archaeological Trust.
Taylor GM, Crossey M, Saldhana J, Waldron T. 1996. DNA from Mycobacterium tuberculosis identified in mediaeval human skeletal remains using polymerase chain reaction. J Archaeol Sei 23:789-798.
Taylor GM, Rutland P, Molleson TI. 1997. A sensitive PCR method for the diagnosis of plasdmodium species DNA in ancient human remains. Ancient Biomol 1:193-203.
Taylor GM, Goyal M, Legge AJ, Shaw RJ, Young D. 1999. Geno-typic analysis of Mycobacterium tuberculosis from mediaeval human remains. Microbiology 145:899-904.
Thierry D, Cave MD, Eisenach KD, Crawford JT, Bates J, Gicquel B, Guesdon JL. 1990. IS6110, an IS-like element of Mycobacterium tuberculosis complex. Nucleic Acids Res 18:188.
Weil A, Pliyatis BB, Butler WR, Woodley CL, Shinnick TM. 1996. The mtp40 gene is not present in all strains of Mycobacterium tuberculosis. J Clin Microbiol 34:2309-2311.
Zumärraga M, Bigi F, Alito A, Romano MI, Cataldi A. 1999a. A 12.7 kb fragment of the Mycobacterium tuberculosis genome is not present in Mycobacterium bovis. Microbiology 145:893-897.
Zumärraga MJ, Bernardelli A, Bastida R, Ouse V, Loureiro J, Cataldi A, Bigi F, Alito A, Castro Ramos M, Samper S, Otal I, Martin C, Romano M. 1999b. Molecular characterisation of mycobacteria isolated from seals. Microbiology 145:2519-2526.