:5. . 15...... a: I.
. “ 3.5... .2

.1

a 9...

n. a
n. .
runawwfi.

.33 . . .2.
hair...

:9
. . r...
I ‘a..:..L1..S.
A . {Ix}.

: 3.. .
3.31.1.2... ((23.1. ‘ r,
23:... s33. .33. x .3. .‘ua’.m.s,
.1553... 3...}..13ua19534
.1337... \R7,O«¢... ..
.H. £391.... v.2... \
2:... .Ia Luv 1
E... ii!
. :1; ~
3.9;}...3315
3.1.6 .pId.....:.I.:. . It;
I ‘33. . .. mu
V xiv: r .-
wavots. I...“ a
a... 3.1.Hanlns.
a
, 5.3.). I.
inst Lu: HR“... aLV‘lCafi...
,1 ..
12.....tL‘vanoa...) .I.

:3;qu

in...

i .1 A q
255155.... . 3. :2 .557“
.\ 1%... $3.. .5! .. 1.; 03...}.
.t. c :. 7.9. . 301:1.) sl‘Uar
1.... \suva‘asfiic9s

31.. 32v
4! In): 1.,
A. V

..
93a... :11:
I .? .
"Alina. ,. n.
3.351....
x. at
.
A“. azw}
5: .17... .i
fiflflflufi.
. c r. .1. x: .5
._\L.4.ot..uv!..\,lh.

$1. a 41.“...wa \.
. x .
tr: . Q
“.1 “Jr“

I 0.1-... .
. .5
. :2 14.3w. N.
(its... .4 .‘t

. h ‘1‘ .
Luahvszfiikin.

5h.
23.. .52.:
. .4 a... \n 59,:
2.31.}. 9.x... . ‘
:5 . 1!. t
\ await: I:

 

2007

This is to certify that the
thesis entitled

MOLECULAR IDENTIFICATION OF PATHOGENS IN ANCIENT
SKELETAL REMAINS FROM BUTRINT AND DIAPORIT ALBANIA

presented by

Michael Joseph Mutolo

has been accepted towards fulfillment
of the requirements for the

MS. degree in Forensic Science

 

 

.‘\
a .
I

97%”? a ,4 7/7...»

W Professor’s Signature

/2.// «ll/o!

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
MICI Hyatt State
University

 

 

 

I-t-o-o-o-u-o---o-o-o-c-u-o-----a----------o----‘--o-a--—--a.

..-._.-o-o-n-n-I-a----s-o-n-s-o-o-n--o-n----u-.--~nO

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

APP 2 2 20m
tuna

'EE)‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd—p.1

MOLECULAR IDENTIFICATION OF PATHOGENS IN
ANCIENT SKELETAL REMAINS FROM BUTRINT AND DIAPORIT ALBANIA

By

Michael Joseph Mutolo

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTERS OF SCIENCE
School of Criminal Justice

2006

ABSTRACT

MOLECULAR IDENTIFICATION OF PATHOGEN S IN ANCIENT SKELETAL
REMAINS FROM BUTRINT AND DIAPORT ALBANIA

By
Michael Joseph Mutolo
Paleopathologists and anthropologists examine skeletal lesions to better
understand diseases that afflicted ancient societies. Traditional analyses of skeletal
remains have limited success in identifying the particular organism or pathogen
responsible. These limitations can be overcome by the addition of molecular techniques
to screen for pathogen DNA present in the remains. Analysis of skeletal material from a
World Heritage Site in Butrint and Diaporit, Albania demonstrated that pathogen DNA
analysis could be a useful tool to the study of ancient cultures. Bone samples were
collected from five sets of remains that showed pathologies, three late Roman (5th — 7th
century AD) and two from the late Medieval period (11th — 13th century AD). Skeletal
lesions seen in these remains had commonality indicating that the same pathogen may
have been involved. In particular, lesions present on the vertebrae and ribs of the late
Medieval samples were almost identical. Bone DNA extracts were screened for the
presence of DNA from the causative agents of tuberculosis and brucellosis. Real-time
PCR results indicated that the vertebrae and ribs of the late Medieval samples contained
Brucella spp. DNA. This provides novel information on the ancient inhabitants of a
culturally rich site that would have otherwise been unobtainable. This also represents the
first time brucellosis has been molecularly identified in an ancient society and proved that
molecular techniques can be used to investigate diseases processes in other skeletal

populations and cultures.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... v
LIST OF FIGURES ........................................................................................................... vi
INTRODUCTION .............................................................................................................. 1
Ancient Butrint, Albania ................................................................................................. 2
Ancient Diaporit, Albania ............................................................................................... 6
Anthropological Analysis of Skeletal Remains from Butrint and Diaporit .................... 7
Skeletal Pathogens and their Identification through Traditional Methods ..................... 8
Tuberculosis Infection .................................................................................................. 11
Skeletal Tuberculosis Infection ..................................................................................... 12
Previous Osteological Studies of Skeletal Tuberculosis ............................................... l3
Skeletal Brucellosis ....................................................................................................... 15
Molecular Analysis of Ancient DNA ........................................................................... 16
Ancient Pathogen DNA Analysis ................................................................................. 18
Molecular Identification of Tuberculosis in Ancient Remains ..................................... 19
Goal of this Project ....................................................................................................... 21
MATERIALS AND METHODS ...................................................................................... 22
Description of Burials: .................................................................................................. 22
Bones Collected for Analysis ........................................................................................ 29
Bone Sample Preparation .............................................................................................. 32
Extraction of DNA ........................................................................................................ 33
DNA Amplification using Polymerase Chain Reactions .............................................. 34
Prevention of Contamination during Ancient DNA Analysis ...................................... 37
Sensitivity Assays of the MTB Complex PCR Primers ................................................ 38
Differentiation of M. bovis and M. tuberculosis ........................................................... 40
M. tuberculosis Mtp40 Primer Optimization and Primer Specificity Assays ............... 41
Sequencing of Mitochondrial Hypervariable Region 1 ................................................ 41
Real-Time PCR Primer Design for Tuberculosis, Brucella and Mitochondrial DNA .43
Optimization of Real-Time Primers .............................................................................. 44
Real-Time PCR on Ancient Bone Samples .................................................................. 46
RESULTS ......................................................................................................................... 48
Bone Sample Preparation and DNA Extraction ............................................................ 48
MTB Complex Primer Optimization and Sensitivity Assays ....................................... 49
Optimization of Methods used to Differentiate M. tuberculosis and M. bovis ............ 51
Mtp40 Primer Specificity Assays ................................................................................. 53
Screening of Bone Extracts for Tuberculosis DNA ...................................................... 54

iii

Spiking Human Bone DNA Extracts with M. tuberculosis Genomic DNA ................. 58

Mitochondrial DNA Amplification ............................................................................... 59
Hypervariable Region I Sequencing ............................................................................. 61
Real-Time PCR Optimization ....................................................................................... 64
Real-Time PCR Results of Voegtly Cemetery Samples ............................................... 65
Real-Time PCR Results for Butrint and Diaporit Pathological Samples ..................... 70
DISCUSSION ................................................................................................................... 83
Bone Preparation for DNA Extraction .......................................................................... 84
DNA Extraction and PCR ............................................................................................. 86
Contamination and Ancient DNA Analysis .................................................................. 87
Human mtDN A Analysis to Assess the Quality of DNA in Skeletal Remains ............ 89
Screening for Tuberculosis in Voegtly Cemetery Material using IS6110 .................... 92
Screening for Tuberculosis in the Voegtly Cemetery Material using OxyR and Mtp40
....................................................................................................................................... 93
Screening for Tuberculosis in Butrint and Diaporit Skeletal Material ......................... 95
Assessing Bones for Pathogen DNA using Real-Time PCR ........................................ 97
Tuberculosis, Brucellosis, and Albania ...................................................................... 100
APPENDIX A ................................................................................................................. 107
BIBLIOGRAPHY ........................................................................................................... 109

iv

LIST OF TABLES

Table 1: Description of Individual Burials ...................................................................... 23
Table 2: Butrint and Diaporit Control and Livestock Samples ........................................ 30
Table 3: Voegtly, Butrint and Diaporit Pathological Bone Samples ............................... 31
Table 4: Primer Sequences and Parameters ..................................................................... 36
Table 5: Calculation of M. tuberculosis and M. bovis Genome Weight ......................... 39
Table 6: Real-Time PCR Primers .................................................................................... 44
Table 7: Combination of Concentrations Used for Primer Optimization. ........................ 45
Table 8: Summary of Analysis Conducted on Pathology and Questioned Bones ........... 47
Table 9: Minimum Number of Genomic Copies Detectable ........................................... 51
Table 10: Control/Livestock Extracts and IS6110 Experiments ...................................... 55
Table 11: Questioned Pathological Samples and 186110 Experiments ........................... 56
Table 12: Ranges of HVl Sequence Obtained for each Bone ......................................... 62
Table 13: Consensus Sequences Derived from Bone Samples for Pathogen Burials ...... 63
Table 14: Optimum Primer Concentration and Average Ct Values for Primer Pairs ...... 64
Table 15: Summary of Real-Time Results for Voegtly Burial 32 Rib mtDN A .............. 69
Table 16: Summary of Real-Time Results for Voegtly Burial 32 Rib IS6110 ................ 69
Table 17: Summary of Real-Time Results ....................................................................... 77
Table 18 (a — f): Ct Values and Tm for MtDNA Positive Reactions ............................... 78
Table 19 (a - e ): Ct Values and Tm for 186501 Positive Reactions ............................... 80
Table 20 (a — b): Ct Values and Tm for Bcsp31 Positive Reactions ............................... 82
Table 21: HVl Sequences from Individual Bone Samples. .......................................... 107

LIST OF FIGURES

Figure 1: Maps Depicting the Location of Butrint ............................................................ 3
Figure 2: Archaeological Sites of Butrint ........................................................................... 5
Figure 3: Archaeological Sites of Diaporit ........................................................................ 6
Figure 4: Butrint and Diaporit Pathological Descriptions and Photographs ................... 24
Figure 5: Location of Albanian Burials ........................................................................... 28
Figure 6: Variability of Aqueous Layers ......................................................................... 48
Figure 7: Primer Sensitivity of 186110 on M. tuberculosis DNA Dilutions .................... 50
Figure 8: Differentiation of M. bovis from other Members of the MTB Complex ......... 52
Figure 9: Mtp40 Specificity ............................................................................................. 53
Figure 10: PCR Screening for the Presence of IS6110 in Voegtly Samples ................... 54
Figure 11: Screening of Vertebrae Samples for IS6110 using Nested PCR .................... 57
Figure 12: 136110 PCR on Bone DNA Extracts Spiked with M. tuberculosis DNA ...... 58
Figure 13: Semi-Nested HVl Results using DNA Extracted from Long Bones ............. 59
Figure 14: HVl Semi-Nested PCR of Vertebrae Extracts ................................................ 60
Figure 15: MtDNA Ampflication Plot for Voegtly Burial 32 Rib .................................. 66
Figure 16: MtDNA Dissociation Curve for Voegtly Burial 32 Rib ................................ 67
Figure 17: 186110 Dissociation Curve for Voegtly Burial 32 Rib .................................. 68
Figure 18: MtDNA Amplification Plot for 4015 Vertebra .............................................. 72
Figure 19: MtDNA Dissociation Curve for 4015 Vertebra ............................................. 73
Figure 20: Real-Time IS6501 Amplification Plot for 4015 Vertebra .............................. 74
Figurep21: Real-Time IS6501 Dissociation Curve for 4015 Vertebra ............................. 75

vi

 

 

206)

Figure 22: Agarose Gel of Real-Time PCR Pathogen Results for 4015 Vertbra ............ 76

vii

I) 00’)

 

 

 

 

 

INTRODUCTION

Throughout history a number of diseases and pathogens have plagued mankind.
Epidemics of cholera, small pox, bubonic plague, malaria and tuberculosis have spread
rapidly through various ancient and modern civilizations. In the 14th century over
twenty-five million Europeans died from the bubonic plague (Zivanovic 1982). In the
19’h century the number one cause of death for working class citizens in the United States
was tuberculosis (Zivanovic 1982). Past civilizations’ interactions with these pathogens
have created a desire in modern scientists to understand how epidemics directly impact
today’s societies.

Paleopathologists and physical anthropologists attempt to understand not only the
manifestation of diseases in an individual, but also how diseases affect an entire society.
Paleopathology involves the study of ancient human and animal remains for evidence of
disease. Zimmerman and Kelly (1982) described multiple goals of paleopathology. The
first is to understand how a pathogen has spread throughout a particular society and how
the society’s reaction to the epidemic has evolved over time. By studying an ancient
society’s response to disease, paleopathologists hope to understand how modern
pathogens would behave in future epidemics. Another goal is to identify which diseases
were present in a society by examining the characteristic lesions and markings left on
mummified and skeletal tissue. This final goal is also important to physical and forensic
anthropologists who attempt to identify abnormalities and pathologies in their analysis of

human skeletal remains.

Ancient Butrint, Albania

Zickel and Iwaskiw (1994) discussed the complex history of Albania and the
struggles to preserve its culture. In 1944 the communist government took control of
Albania and banned other nations, including the United States and Great Britain, from
entering the country. As a result, the United States and nations of Western Europe had
only limited knowledge of the culture and history of Albania. In the 19705 the United
Nations Education, Science, and Cultural Organization (UNESCO) designated several
locations in Albania as World Heritage sites. However, due to entrance restrictions
imposed by the communist regime, archaeological and historical investigations were
limited to scientists employed by the Albanian government. In the 1990s, the communist
regime ended and scientists from other countries were once again allowed to enter the
region. These scientists brought with them new research ideas and began the excavation
of historical sites in conjunction with the Butrint Foundation, Packard Humanities, and
the Albanian Institute of Archeology (Butrint Foundation Website). A joint report issued
by UNESCO and the International Council on Monuments and Sites in 1997 stated that
many of the historical sites had been looted and artifacts were destroyed. This report
called upon scientists worldwide to aide in the preservation of ancient sites throughout
Albania.

One historical site that was of particular interest was the ancient city of Butrint.
Butrint is seated in the ‘cradle of Western civilization’ between Greece, Italy, former
Yugoslavia, and the Adriatic Sea (Figure 1(a)). The city is located on a small peninsula
of marshland in the southwest coast of Albania and is surrounded by three bodies of

water including Lake Butrint, an inland lagoon, and a series of straits (Figure 1(b)).

Figure 1: Maps Depicting the Location of Butrint

(a). Location of Albania in relation to other southeastern European countries and the
Adriatic Sea. (b). The city of Butrint (red box) in relation to surrounding bodies of water.

a).

 

 

A .

(mum “ .3:

  

YUGOSLAVIA

iHE PORTER
Y'i’sfli! M"

Adriatic “NANA titling": ‘0}.
mm. a r AL 1 Dink/4
Sea
Elman'
‘7!
, .,V|0fé
‘ ~ .
i‘IIALY , S'wrot
. z ; cum
\\ W GREECE
o 13' com t".
o )0 com. ‘33 5*”

  
    
     
   
 

 

 

(b)-

 

(Butrint Foundation Website)

The Butrint Foundation has published several reports on the history of the ancient

city (UNESCO 1999). Over the ages the site of Butrint has been inhabited by a number

of different cultures including Greeks, Romans, Byzantines, and Venetians. The Greek

Chaonian Tribe first founded the city in the 8th century BC. By the 4th century BC the

site was dedicated to Aesclepius, the Grecian God of Medicine, and had taken on the

appearance of a Greek polis with temples, theaters, and public buildings. In 44 BC the

city was seized by the Roman Empire and became part of the Illyrian province. Butrint

expanded considerably under Roman control, and eventually was made a diocese of a

bishOp during the height of early Christianity. This period also marked the construction
of Roman infrastructure, such as roads and aqueducts. The city played a vital role as a
port for the Roman Empire and a diverse assemblage of peoples (European, Middle
Eastern and African) are believed to have used the site as a trade route.

Butrint began to decay in the 5’h and 6h centuries when the collapse of the Roman
Empire caused infrastructure failure (UNESCO 1999). A Slavic invasion in the 7‘h
century AD led to abandonment of the city, but the site regained its prosperity under the
Byzantine administration during the 9th century. Little is known about the city following
this period until it came under Venetian control in 14th century AD. Power struggles
between the Venetians and Ottoman Turks continued for several centuries. In the late
Middle Ages waters from the marsh and surrounding lakes flooded the city leading to its
ultimate abandonment.

Butrint has been well characterized archaeologically (UNESCO 1999). Four sites
have been excavated and are currently being studied and preserved by the Albanian
Institute of Archeology (Figure 2). The majority of the archaeological work has been
conducted on the T riconch palace that sits near the shore of the Adriatic Sea. This palace
was believed to have reached its peak use during the 4th and 5th centuries AD. It was
abandoned during the 6th and 7th centuries but later was re-occupied by the Venetians.

Other sites of interest include a Greek amphitheatre and an early Christian baptistery.

Figure 2: Archaeological Sites of Butrint

Photograph (a) depicts an aerial view of the archaeological sites and ancient city of
Butrint (Butrint Foundation Website). Photograph (b) is an image of the early Christian
Baptistery (Murray 2005). Photograph (c) depicts an image of the Triconch palace while
(d) is the Greek amphitheater (Murray 2005 and Butrint Foundation Website).

 

Ancient Diaporit, Albania

On the eastern shores of Lake Butrint sits an additional archaeological site known
as Diaporit (Figure 3). Archeological evidence suggests that this site was first occupied
during the late 3rd century BC when a small villa was established (Butrint Foundation
Annual Report 2004). The Romans first entered the area in the early 1" century AD and
from 50 to 80 AD established a large villa. Excavations have revealed several features of
this Roman villa including fountains, courtyards, elaborate mosaics, and a roman
bathhouse. Archeological evidence of mosaics and a church indicates that Christianity

arrived to the area between the 5“ and 7’” century AD.

Figure 3: Archaeological Sites of Diaporit

Map depicted in (a) shows the location of Diaporit in relation to other excavation sites
currently being studied in Albania (Butrint Foundation Annual Report 2004). Photograph
(b) is of the excavated Roman Villa (Fenton 2006).

(a).

  

 

 

 

Anthropological Analysis of Skeletal Remains from Butrint and Diaporit

Excavations of the Butrint and Diaporit sites led to the discovery of several
burials, which remained uninvestigated until very recently. Dr. Todd Fenton and
graduate students from Michigan State University conducted anthropological analysis to
characterize the inhabitants of the sites. Skeletal remains were analyzed to access the
individuals’ age and sex. According to the Butrint Foundation’s 2004 annual report
Fenton and his students had thus far completed skeletal analysis on 80 individuals from
the Diaporit site and 105 from Butrint. Bone samples from a number of skeletons were
collected and brought to Michigan State University for molecular analysis.

Some of these individuals displayed skeletal lesions that may be associated with
pathologies such as tuberculosis or brucellosis. Both diseases have been identified in
ancient and modern day Albanian societies (World Health Organization 2000). It is
believed that the Roman Empire’s collapse, combined with failure of infrastructures and
conflicts over the region enhanced societal decay (Butrint Foundation Annual Report
2004). Poor living conditions, lack of resources, and the introduction of foreign cultures

could have led to the spread of disease and epidemics, which expedited this process

Skeletal Pathogens and their Identification through Traditional Methods

According to Zimmerman and Kelley (1982), the characterization of pathogens in
human skeletal remains can be a challenging process. Many pathogens that are able to
infect bone do not lead to the manifestation of skeletal abnormalities or lesions. Lack of
these skeletal lesions is believed to be a result of rapid recovery or death immediately
following infection. F urtherrnore, if abnormalities or skeletal lesions are produced they
are rarely distinctive enough to identify the particular organism causing the disease.
Long term or chronic infection is usually required for the manifestation of skeletal
lesions.

Zimmerman and Kelley (1982) described two broad categories of bone infection:
pyogenic and granulmatous. Pyogenic organisms such as staphylococci and streptococci
cause infections that result in pus formation. These organisms can infect any bone in the
body, producing inflammation in the marrow, cortex, and periosteum. Pyogenic
infections of bone typically arise from bacteria in the blood stream or as an outgrth of
soft tissue. Once localized in bone, the infection can become chronic without medical
treatment. Infected bone may display an irregular thickened appearance and numerous
cloacae (channels or holes through the cortex). Pyogenic lesions are characterized by
bone erosion with sequestra (small areas of new bone regeneration).

Granulomatous organisms produce pockets of infection that contain masses of
bacterial cells and host leukocytes. Examples include the pathogens that cause leprosy,

tuberculosis, syphilis and some fungal infections. These organisms can infect any bone

in the body, resulting in areas of erosion. In contrast to pyogenic infection,
granulomatous organisms erode bone while inhibiting sequestra or bone regeneration.

Ortner and Putschar (1981) described some of the traditional methods used by
paleopathologists and anthropologists to study skeletal lesions, including gross analysis,
X-ray examination, chemical analysis, and microscopic examination. Gross and X-ray
analyses involve the examination of skeletal material to provide a description and
understanding of the disease process that causes skeletal abnormalities. Anthropologists
use these methods to identify the number of lesions present and describe the overall
appearance and quality of the bone tissue. Identification of abnormalities and their
distribution throughout the skeleton can provide further information on the disease and its
manifestation.

Microscopic and chemical based methods tend to be more limited than gross or
X-ray analyses. Chemical tests detect the levels of certain amino acids and minerals,
such as fluorides and iron, in bone tissue. Von Endt and Ortner (1982) used amino acid
analysis to identify potential anemia caused by malnutrition in the skeletal remains of
American Southwest Pueblo Indians. Bone proteins in children that showed signs of
skeletal anemia had 25% less amino acids that contained hydroxyl groups (serine,
threonine, and tyrosine) and acidic-side chains (aspartatic and glutamic acid) compared to
children who had died from trauma. Enzymes that require an iron co-factor synthesize
these amino acids. Based on their test results, Von Endt and Ortner (1982) hypothesized
that the children were experiencing anemia due to iron deficiencies caused by a limited

diet.

Microscopy, or histological analysis, relies on examination of osteoblasts and
osteoclasts. Osteoblasts are cells that secrete bone, while osteoclasts remove portions
causing the production of the bone matrix. Paget’s disease is an example of a metabolic
bone abnormality that can be identified using microscopy. This disease simultaneously
causes an increase in osteoclast and osteoblast activity, resulting in a bone matrix that
appears fragmented when viewed microscopically.

These traditional techniques are limited and directly influenced by the condition
of the skeletal remains and bone samples being studied (Ortner and Putschar 1981).
Highly fragmented material may be difficult to examine, and understanding how
pathologies affect the skeleton as a whole is nearly impossible. Microorganisms found in
soil can also damage bone tissue, creating abnormalities that are similar to disease
processes. One way to surpass these limitations is to combine skeletal examination with
molecular genetic techniques.

Several published articles have described the identification of pathogen DNA in
ancient human remains. It is believed the presence of pathogen DNA in bone could
indicate either acute or chronic infection during an individual's lifetime. Past genetic
analysis of ancient DNA has proved successful in identifying the presence of a wide
range of pathogens, including malaria, leprosy, plague, tuberculosis, and syphilis in a
number of ancient societies (Rafi et al. 1994, Drancourt and Raoult. 2004, Sallares and

Gomzi. 2001, Hunnis 2004, Mays et al. 2002).

10

Tuberculosis Infection

Despite progress in modern healthcare, medicine, and improved living standards,
tuberculosis is still one of the most common diseases, killing an average of 1.7 million
people a year worldwide (WHO 2006). The disease is caused by the infection of lung or
soft tissue by small, acid fast, gram—positive bacteria. These bacteria are members of the
Mycobacterium Tuberculosis Complex (MTB complex), which includes a number of
pathogenic Mycobacterium species (M. bovis, M. tuberculosis, M. microtti, M. afiicanus,
etc.) and infect a wide range of hosts. The disease in humans is primarily caused by M.
tuberculosis or M. bovis.

Mays et a1. (2002) and other authors have summarized the history of tuberculosis.
M. bovis is thought to be the primary infectious organism that inflicted ancient societies.
It is transmitted by the inhalation of contaminated aerosol from livestock or through the
ingestion of contaminated meat or milk. Ingestion of the bacterium results in foci or
small nodules of infection in the intestinal walls and lymph nodes. Since the 19th century
the majority of tuberculosis cases have been caused by the spread of M. tuberculosis from
person to person through the inhalation of contaminated aerosol. Industrialized and
Western societies are particularly hard hit by the disease since its spread seems to be
dependent on population-density. Once the organism is inhaled, pulmonary infection
ensues by the formation of foci in the lung. The foci, if not healed by the individuals’
immune system, will progress, spreading first to surrounding lymph nodes and then to the
blood stream. The spread of bacteria to the blood stream can result in secondary

infection of soft, skeletal, and organ tissue. The degree of secondary infection is highly

ll

dependent on the virulence of the organism, the size of the inocculum, and the resistance

level of the individual’s immune system.

Skeletal Tuberculosis Infection

Ortner and Putschar (2001) characterized the infection of the skeleton and bone
tissue by tuberculosis bacilli. Skeletal manifestations of the disease are more common in
unhealthy adults and young children. It is estimated that 50 to 90 percent of cases of
tuberculosis occur in individuals between 3 and 15 years of age. Skeletal lesions may
begin to develop at this time and usually persist well into adulthood. It is estimated that
only 5 to 7 percent of skeletal tuberculosis infections result in lesion production
(Zimmerman and Kelley 1982). However, this number may be an underestimate due to
the variability and expression of lesions. One of the difficulties in diagnosing
tuberculosis in skeletal populations is that there are no absolute criteria for osteological
identification.

The most common site of skeletal tuberculosis infection is the spinal column and
vertebrae (Ortner and Putschar 2001). The part of the spinal column involved is almost
exclusively the vertebral bodies; while involvement of the vertebral arches is rare. It is
believed that the bacilli travel through the blood stream, up the paravertebral plexus and
attack the center, the anterior surface, or the paradiscal region of the vertebral body.
Uehlinger (1970) conducted a study on 62 individuals that had been diagnosed with
tuberculosis at autopsy. He found that the most common vertebrae involved were

thoracic (particularly T6 -— T12) and lumber (L1 — L5). Vertebral infections are lytic and

12

produce lesions that Mays et al. (2002) described as circular resorptions present on the
anterior surface of the vertebrae. The bacteria will continue to cause these lesions and
destruction until the vertebrae collapse. Vertebral collapse caused by tuberculosis, also
known as Pott’s disease, is characterized by “kyphosis”, or angling of the vertebral
column followed by fusion of the spine.

Zimmerman and Kelley (1982) reported that weight-bearing joints could also
become infected with tuberculosis. The bacilli enter the joint from adjacent bone or
tissue membranes that line the surrounding areas. Once situated in the joint, the bacteria
erode bone tissue creating lesions on articulating surfaces. Extra-vertebral tuberculosis

has also been found to affect ribs as well as bones of the limbs, skull and pelvis.

Previous Osteological Studies of Skeletal Tuberculosis

Kelley and El-Naj jar (1980) examined the natural variation and patterns of
skeletal tuberculosis in the Hamann-Todd osteological collection. This collection
contains approximately 500 individuals who demonstrate signs of tuberculosis infection.
Twenty-six individuals were selected, using medical records or death certificates that
documented them as suffering from skeletal tuberculosis. Gross and X-ray analyses were
conducted to characterize the location and severity of the lesions. Their results indicated
that the most common site of lesions was the lower spine (57.7%) from T11 to L5. The
authors commented that 40 - 50% of the skeletons examined showed signs of
tuberculosis in other bones. Rib involvement was the second most common site with
34.6% of individuals having lesions on the side of the rib facing the lungs or at the site of

vertebral articulation. Skeletal tuberculosis was also noted in the pelvis, knee, sternum,

13

shoulder, clavicle, and femur. Thirty-eight percent of the individuals showed signs of
tuberculosis infection in multiple bones.

Ortner and Putschar (1981) stated that rib lesions associated with tuberculosis are
generally found in two locations: near the rib head (the portion of the rib that articulates
with the vertebral body) or on the visceral surface (the posterior surface of the ribs that
faces the lungs) due to the spread of overlying pulmonary foci. Kelley and Micozzi
(1984) examined the association between rib lesions and pulmonary tuberculosis in the
Hamann-Todd collection. Seventy individuals were selected that showed signs of
skeletal tuberculosis. The authors reported that 55.7% of the skeletons examined
contained evidence of lesions on the rib visceral surface. Based on these results they
hypothesized that rib lesions could have been produced by adjacent pulmonary lesions.
They concluded that there is a strong correlation between rib lesions and pulmonary
tuberculosis in the Hamman-Todd Collection.

Roberts et al. (1994) conducted another study of tuberculosis rib infection by
examining the Terry anatomical collection. Out of 255 documented tuberculosis cases,
62% showed visceral surface rib lesions. However, 22% of individuals that died from
non-tubercular pulmonary diseases (i.e. pneumonia, bronchitis, emphysema) and 15% of
individuals with non-pulmonary causes of death showed similar lesions. The authors
concluded that visceral rib lesions are not a definitive indicator of pulmonary tuberculosis

and could be associated with other pathogenic infections.

14

Skeletal Brucellosis

There are several types of pathologies that can resemble skeletal tuberculosis,
including secondary malignant bone tumors, fractures of the vertebrae, pyogenic
osteomyelitis, Paget’s disease, rheumatoid arthritis, brucellosis, and fungal infections
(Kelley and El-Najar 1980). Ortner and Putschar (1981) described the infection of
skeletal material by Brucella spp; a gram-negative coccobacilla that causes the disease
brucellosis. Three members of Brucella spp. are pathogenic to domestic livestock
including B. abortis, B. suis, and B. melitensis. The latter two have also been associated
with human infection and are primarily transmitted through ingestion of the bacteria. B.
melitensis is common in goats and sheep fiom the Mediterranean and is transmitted
through the ingestion of infected meat or milk products. Similarly, B. suis infects
domestic swine and is transmitted through meat. B. abortis, a strain that can cause
miscarriages in cattle and horses, may also infect humans in extremely rare cases.

Brucellosis in humans usually manifests as a chronic lung infection followed by
recurring fevers. Skeletal tissue will become involved with chronic infection, especially
with B. melitensis, and can vary from 2 — 70% of cases (Jaffe, 1972). F oci of the bacteria
spread to cancellous bone from the bloodstream and produce skeletal lesions. Once in
cancellous bone the foci will produce lytic cavities and the overlying cortex may become
perforated with little evidence of healing. The vertebral column and pelvis are the most
common sites of infection with each containing multiple foci. As a result the lesions
produced appear very similar to those of skeletal tuberculosis. A major difference

between tuberculosis and brucellosis infection is that Brucella spp. tend to be less

15

destructive than members of the MTB complex. As a result brucellosis does not cause

severe degradation of joints or kyphosis of the vertebral column.

Molecular Analysis of A ncient DNA

The analysis of ancient DNA can provide a vast amount of information about an
individual and a society. In our laboratory, ancient DNA analysis has been used to study
family groups, genetic diversity, human migration patterns, and ancestry (Clemmer 2005,
Rennick 2005, Murray 2006). Data obtained from such studies can compliment
anthropological analysis by providing novel information that would not otherwise be
obtainable.

DNA extracted from ancient remains is often highly degraded and of low copy
number. The amount of DNA template must be increased before genetic analysis can be
conducted. Polymerase chain reaction (PCR) is a method used by molecular biologist to
amplify small amounts of DNA. PCR involves the use of thermal cycling in which a
sample of DNA is heated and cooled in a precise pattern in the presence of Taq DNA
polymerase. This results in the amplification of a specific region of DNA sequence. The
boundaries of the amplified sequence are defined by short pieces of DNA, known as
oligonucleotide primers. Primers are designed so that they are complimentary to DNA
sequences that flank the desired product (often termed the amplicon). This allows
molecular biologist to target a particular DNA region of interest. Amplification of DNA
using PCR is exponential: after 30 cycles, billions of copies of the target region can be

produced from a very small amount of starting material. The sensitivity of PCR has been

16

greatly increased by the development of nested PCR. In this process, additional cycles
are conducted on the amplified template using primers internal to those used in the first
round. This allows for the amplification of extremely small amounts of starting template,
such as those found in skeletal remains.

In ancient material, nuclear (chromosomal) DNA is often too degraded to analyze.
Instead analysis of mitochondrial DNA is conducted. Mitochondria are organelles
located in the cell that are responsible for producing energy. In humans, mitochondria
contain a circular genome that is 16,569 bases in size. This genome has both protein
coding and a non-coding segment; the latter is also known as the control region and is
further divided into two hypervariable (HV) regions: HVl and HV2. The non-coding
nature of the hypervariable regions has allowed mutations to accumulate due to a lack of
selection against them. As a result, unrelated individuals tend to contain on average 3%
base differences (i.e. polymorphisms) in the DNA sequences of their HV regions
(Hummel 2003). Polymorphisms have accumulated over time and are passed from
generation to generation through maternal lineages. Maternal ancestors will share the
same mtDNA sequence, termed a haplotype, which generally differs from sequences
found in unrelated individuals. In forensic science, the sequences of mtDNA HV regions
are compared to each other and to a reference sequence (Anderson et al. 1981) to identify
remains as belonging to a family group. Previous studies in our laboratory have utilized
PCR and mtDNA analysis on Albanian ancient skeletal material to look at genetic
diversity and maternal relatedness of ancient individuals (Clemmer 2005, Murray 2006,

Rennick 2005).

17

Ancient Pathogen DNA Analysis

DNA analysis can aide in the identification of the specific pathogen that produced
skeletal lesions. Pathogen DNAs may be preferentially amplified by designing primers
that target sections of the pathogens genome, without binding to any host sequences.
Shawar (1993) described the use of PCR to identify members of the MTB complex in
lung tissue and sputum from pathological samples. PCR was conducted to screen for the
presence of a mobile genetic element, IS6110, which is found in the genome of MTB
complex members. The copy number of 186110 varies by species, with M. tuberculosis
having 8 — 20 copies and other members of the complex having as few as one. Because it
is unique to Mycobacterium spp., amplification of 186110 is a quick and reliable method
to screen samples for a member of the MTB complex.

Other PCR based methods have been developed to differentiate members of the
MTB complex. Most efforts have focused on differentiating M. tuberculosis and M.
bovis. A method developed by Sreevatsan et al. (1996) involves the amplification of the
OxyR gene to determine which base is present at position 285. All M. bovis strains
contain an adenine (A) at this site, while all other MTB complex members contain a
guanine (G). Screening for this conserved polymorphism has become a diagnostic tool
for M. bovis infection.

Differentiating M. tuberculosis from other members of the MTB complex has
proven more difficult because they share 95 — 99% genetic homology. Methods to
identify M. tuberculosis as the pathogenic agent usually involve screening for species-

specific genes, such as Mtp40. Parra et al. (1991) identified the Mtp40 gene as a 402 bp

18

segment of DNA that encodes a protein specific to M. tuberculosis. The absence of
Mtp40 from other members of the MTB complex has allowed it to be utilized as a

diagnostic tool for M tuberculosis infection.
Molecular Identification of Tuberculosis in Ancient Remains

While most of the assays described above were developed for modern samples,
the methods have been modified to identify tuberculosis in ancient remains. In 2001,
Mays et al. published a paper describing the use of PCR to screen for MTB complex
DNA in nine human skeletons from the Wharton Perry Collection. Gross and X-ray
analyses were used to characterize lesions on the vertebral bodies and joints of nine
individuals. According to the authors skeletal lesions included “cavitations of the
cancellous bone. . .with little new bone formation...” Spinal kyphosis and vertebral body
collapse were also seen in several of the remains. Material was collected from the inside
of lesions from each individual and from two control individuals. DNA extracted from
this material was first screened for IS6110 using nested PCR. Bones from seven of the
nine individuals resulted in a positive PCR reaction, indicating that a member of the
MTB complex’s DNA was present. Further analysis of the OxyR and Mtp40 gene
determined the infectious agent was M. tuberculosis.

Mays et al. (2002) continued this work by investigating the relationship between
visceral rib lesions and tuberculosis. Ribs from the seven positive burials identified in
the 2001 studies were screened along with ribs from seven burials that lacked rib lesions.

IS6110 PCR products were obtained from one of the positive burials and two ribs from a

19

control burial. The authors concluded that rib lesions do not appear to be directly
associated with tuberculosis infection in the Wharton Perry population. They also stated
that molecular techniques could be used to detect pathological infection even when
skeletal lesions are absent.

Mays et al.’s studies (2001 and 2002) were limited to the analysis of bones from a
single collection. Zink et al. (2005) expanded this research by attempting to isolate MTB
complex DNA from different skeletal populations. The first group included vertebrae
from recent cases of systemic tuberculosis that were collected at autopsies conducted in
Munich from 1990 — 2000. An additional 12 vertebral samples were collected as controls
from individuals that died from other causes. Historical remains from a necropolis in
Egypt (1550 — 500 BC) and a 14th century German cemetery were also tested. Egyptian
bone samples consisted of vertebrae from 36 skeletons: 5 showing signs of tuberculosis,
12 showing pathologies of unknown origin, and 19 that appeared normal. The Germen
cemetery sample set consisted of 51 long bones from skeletons that showed signs of
thickening, porosity and periosteal reaction (morphological changes in bone surface). .

In the modern group, 8 out of 12 cases tested positive for the MTB complex.
Results obtained from the historical Egyptian collection varied in the degree of visible
skeletal lesions. In total, 7 out of 36 vertebrae resulted in positive amplification of
186110. This included 3 out of 5 that showed tuberculosis characteristic lesions, 2 out of
12 that showed unknown pathologies, and 2 out of 19 that appeared normal. Ten out of
51 of the German long bone samples tested positive. Zinc et al. (2005) concluded that the
use of IS61 10 as a molecular tool for the identification of MTB complex DNA is feasible

with both modern and ancient materials. They also concluded that molecular screening

20

techniques for pathogens can be informative in instances were skeletal lesions are

minimal.

Goal of this Project

The identification of pathogens in ancient and modern skeletal remains is of great
interest to both paleopathologist and anthropologists. Molecular techniques have proven
their ability to aide in identifying the specific organism responsible for skeletal lesions.
This is accomplished by screening bone samples for the presence of pathogen DNA. The
goal of this project was to utilize molecular techniques, such as PCR, SNP analysis, and
real-time PCR, to analyze skeletal remains from the ancient cities of Butrint and Diaporit
in present day Albania. An anthropological study of remains from these sites had
identified several individuals that showed pathologies consistent with tuberculosis or
brucellosis. Bone samples were screened for the DNA of pathogens that cause these
diseases, the first time that such methods were applied to ancient skeletal remains from
Albania. This research not only showed that molecular techniques could complement
osteological analysis in the identification of pathogens, but also provides information on

the lives and deaths of individuals that inhabited a culturally rich site.

21

MATERIALS AND METHODS

Description of Burials:

Bone samples from eleven human and three livestock skeletons were used in this
study. Table 1 provides a summary of the location, time period, and biological profiles of
the burials. Bones from two individuals, burials 32 and 629, originated from a cemetery
in Pittsburgh, PA, and had previously tested positive for M. tuberculosis DNA in a study
conducted by Ubelaker et al. (2003). The remaining nine individuals were from Butrint
and Diaporit, Albania. Dr. Todd Fenton and graduate students divided the individuals
into three categories: tuberculosis/brucellosis consistent, questioned pathologies, and
controls. Individuals, 2722 and 4015 showed skeletal lesion whose appearance were
consistent with previous reported cases of skeletal tuberculosis or brucellosis (Mays et al.
2002). The questioned pathology group consisted of skeletal remains from three burials,
213, 319 and 5010, which showed minimal damage or only contained a subset of the
characteristic lesions described by Mays et al. (2002). Bones from four skeletons that
contained no signs of pathologies were selected to serve as controls. More detailed
information on the pathological burials for the Butrint and Diaporit sites are outlined in

Figure 4.

22

Table 1: Description of Individual Burials

A description of remains sampled in this study. The information is arranged by
categories assigned during osteological examination and the site where the burials were
found. Time Period, Sex, and Age at Death information for the Voegtly cemetery burial
was obtained from Uberlaker et a1. (2003). Dr. Todd Fenton and graduate students
provided information on the Albania burials. Date following ‘Location’ indicates the
year in which the burial was excavated. Information was undetermined for some

individuals (Undet).

Voegtly Samples

32 ' PA 1987
629 ' PA 1987

Control les
10 ' ' 2000
584 ' ' 2002
3023 Butrint 2001
3060 Butrint 2001

les
't 2001
Butrint 2002
TB/Brucella Pathol ca]
2272 Butrint 1995

4015 Butrint 2002

Livestock les

M5207
M5267

M5509

Butrint 2003
Butrint 2003

Butrint 2003

Late 18
Late 18

Roman
Roman 5 — Cen

Medieval

Medieval

Roman 5 — Cen
Roman 5 — Cen
Roman 5 — Cen

les

Medieval (1 1 — 13
Cen
Medieval

Late 5
Late 5 Century

Late 5

23

Male
Female

Female
Undet
Undet
Undet

Male
Female
Female

Adult
Adult

Undet

Undet

Undet
Adolescent

Adolescent

Adolescent

 

Figure 4: Butrint and Diaporit Pathological Descriptions and Photographs

Descriptions of niberculosislbrucellosis consistent and questioned pathology burials are
provided. Images show examples of pathological lesions found on the skeletal remains.
(Osteological data provided by Dr. Todd Fenton and his graduate students).

(a) Individual 213:

Skeletal remains of an adult male from Diaporit dated to the Roman era in the 5"I - 7"I
Century AD. Pathologies included osteophytic lipping on lumbar vertebrae (left image)
while cervical and thoracic vertebrae showed evidence of bony growths. The ulna
contained poritic lesions and sclerotic bone at the proximal and distal ends (right image).
The left femur and tibia (not shown) appeared expanded with sclerotic bone at both ends.

   

(b) Individual 319:

Skeletal remains of an adult female from Diaporit dated to the Roman era in the 5" — 7"
Century AD. Pathologies included osteophytic lipping and porosity on cervical
vertebrae, porosity of the ribs (left image), and lesions on the parietals of the skull (right
image). Both tibiae and one fibula (not shown) displayed expansion and reactive woven
bone.

 

24

(c) Individual 5010

Skeletal remains of an adult female dated to the Roman era in the 5"I — 7‘” Century AD
that were excavated from the Triconch Palace in Butrint. Poratic hyperostosis was seen
throughout the skeleton. Oval and circular lesions were present on the frontal and
parietal bones of the skull (area outlined by the white box). Cervical vertebrae showed
osteophytic lipping and extensive periosteal reactions were noted on the humeri, femora,
and tibia (not shown).

 

25

(d) Individual 2272

Medieval (11" - 13" Century AD) skeletal remains of an adolescent male excavated from
the Baptistery in Butrint. Multiple circular lesions and areas of erosion were present on
the thoracic vertebrae. Ribs showed signs of cortical expansion and the fibula showed
some expansion of the metaphyses (not shown). The femoral shafts had myostosis
ossifications and contained porosity and expansion of the head (not shown).

 

 

26

(e) Individual sons

Medieval (11" — 13" Century AD) skeletal remains of an adolescent male excavated from
the Merchant House in Butrint. Multiple circular lesions and areas of erosion were
present on several thoracic vertebrae. Ribs showed thickening, orbicular expansion and
lesions on the anterior surface with porosity on the visceral surface (not shown).

 

 

 

 

 

i
luv-(-

v I
_ — —._. — — .M

 

 

/

Li? M to“? .. fleets;

 

 

27

Figure 5: Location of Albanian Burials

(a). Distribution of burials by excavation site. (b). Site plan of Butrint with general
position of burials marked in red. Burial 2722 was found near the baptistery. 5010 was
located inside the Triconch palace. 4015, 3023, 3060 were found inside the merchant
house that is located near the palace.

A

 

(a).

 

 

Paints om , .\ /
/ Triancul Castle 3%

(b)-

t.
'9’

r" Triconch Paraoe'x _
I

- Bathhouse

.A

 

 

28

Bones Collected for Analysis

The Voegtly cemetery samples from burials 32 and 629 consisted of fragmented
ribs and femur cross-sections (Table 3). A forensic biology graduate student from
Michigan State University collected whole vertebrae and fragments of ribs, pelvises,
skulls, metacarpals, metatarsals, and long bones from the Albanian remains (Murray
2005). Larger bones that showed signs of pathology were sectioned using a Dremel tool
fitted with a cutting wheel. Butrint and Diaporit bone samples available for DNA
analysis from the control group and livestock remains are outlined in Table 2. The

tuberculosis/brucellosis consistent and questioned bones are outlined in Table 3.

29

Table 2: Butrint and Diaporit Control and Livestock Samples

Bones samples collected for DNA analysis are listed and arranged by burial number.

Thoracic Vertebrae
Left Rib
Fibula
Tibia

M

Thoracic Vertebrae
Pelvis
Left Rib

Femur

Radius
Metatarsal

Ulna
Calcanium

Vertebrae

Ulna
Rib
Phalanx
Radius

 

30

Table 3: Voegtly, Butrint and Diaporit Pathological Bone Samples

Bones collected for DNA analysis and brief description of pathologies noted by Ubelaker
et al. (2003) (indicated by *) or Fenton (2006) arranged by burial number.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Burial Bonel . .. Pathologics Described During Osteological
Number Sample ' .4 Analysis
32* Rib Lesions on Cervical Vertebrae
Femur
629* Rib Ossified nodules and lesions on Vertebrae
213 Thoracic Vertebrae Adjacent to Three Vertebrae that Showed
Pathology
Ischium Small Amount of Pathology
Femur -
Ulna Lesion on
Fibula Pathology on Distal End
Metacarpal -
319 Thoracic Vertebrae -
Rib Ribs show signs of Perosity
F emur -
Fibula Shows signs of Pathology on Distal Half
Skull Shows signs of Pathology
Ulna -
Metacarpal -
5010 Thoracic Vertebrae
Rib Skeletal lesions on Multiple Ribs
F emur -
Fibula -
Humerous Skeletal lesions present
Metacarpal -
4015 Thoracic Vertebrae Multiple circular lesions on the anterior
Rib Rib thickening and trabecular expansion; minor
porosity visceral surface
F emur -
Pelvis -
Metacarpal -
2272 Thoracic Vertebrae Multiple circular lesions on the anterior
Rib Most ribs exhibit cortical expansion
Pelvis
Radius -
Fibula Expansion of metaphysis
Metacarpal -
F emur Porosity on femoral neck; expansion of distal
metaphyses

 

 

 

 

31

 

Bone Sample Preparation

Digestion buffer (50 mM EDTA, 0.5% SDS, 20 mM Tris pH 8.0), bone wash
solution (10 mM EDTA, 0.5% SDS, 20 mM Tris pH 8) and water were sterilized by
passing them through a 0.22-micron filter. A Dremel tool, drill parts, rotary tool, sanding
wheel, 1/ 16 inch drill bits, weigh paper, weigh boats, sterile 1.5 mL microcentrifuge
tubes, and filtered liquids were UV irradiated at 6 joules per cm2 for 5 minutes.

Bones were cleaned and prepared for sampling in a UV sterilized Cleanspot
Workstation. For whole bone and large bone fragments (e. g., a whole vertebra or large
section of femur) a small portion of the surface was prepared for sampling. The area was
rinsed with 1 — 2 mL of bone wash and then cleaned using sterile swabs. The bone was
transferred to a weigh boat and allowed to air-dry. Swabbed areas were gently sanded
using a Dremel rotary tool, fitted with a 1/4 inch sanding band, until the outer brown
surface material was removed and clean cortical bone revealed. A second wash of the
sanded area was conducted to remove any residual dust generated during sanding.
Finally, the area was rinsed using sterile water and the fragment allowed to air dry. Bone
powder was generated for the DNA extraction process by drilling lightly into the sanded
area using a Dremel rotary tool fitted with a 1/16 inch drill bit. Twenty-five to 50 mg of
bone powder was collected on a piece of weigh paper, transferred to a 1.5 mL sterile
microcentrifuge tube, and the final weight recorded.

An alternative method was used for smaller bones and bone fragments (e. g.,
metacarpals and ribs). Following the initial rinse, the Dremel tool was fitted with a 1/16

inch drill bit and small holes or nicks, approximately a millimeter deep, were made to

32

remove the brown surface material. The area surrounding the holes was scrubbed using
sterile swabs soaked in buffered wash solution. A new sterile drill bit was attached to the
Dremel tool and the initial surface holes were deepened to 1 to 2 millimeters. Bone
powder was collected and weighed as described above. Five hundred uL of digestion
buffer and 5 uL of proteinase K (20 mg/mL) were added to each tube containing bone
powder. A 1.5 mL sterile microcentrifuge tube containing 500 pL of digestion buffer and
5 111.. of proteinase K (20 mg/mL) was prepared alongside the samples to serve as a
reagent blank. Samples and associated reagent blanks were vortexed and placed in a

55°C incubator for 24 hours.

Extraction of DNA

DNA was extracted from the bone powder using an organic method. An equal
volume of saturated phenol (500 uL) was added to each sample, followed by vortexing
for approximately 15 seconds, and centrifugation at 14,000 rpm for 5 minutes. The
aqueous layer was transferred to new sterile 1.5 mL microcentrifuge tubes. An additional
phenol extraction step was conducted if the aqueous layer was colored (ranging from
light pink to brown). This was required for most vertebrae and rib samples.

An equal volume of chloroform was added to the aqueous layer. Samples were
vortexed for approximately 15 seconds and centrifuged at 14,000 rpm for 5 minutes. The
resulting aqueous layer was loaded onto Microcon filter devices and pushed through the
column by centrifuging the samples at 10,000 X g for 10 minutes (the Microcon protocol

calls for 14,000 X g; however, this was lowered to prevent soil/residual powder from

33

tearing or clogging the column). The filtrate was discarded and 300 uL of sterile TE (10
mM Tris, lmM EDTA pH 8.0) was pushed through the column by centrifugation for 10
minutes at 12,000 X g. An additional two TE washes were conducted using this method.
DNA was eluted off the column using 20 uL of sterile TE buffer. A one in ten dilution of
the extracted DNA was made using sterile TE and used in PCR reactions. DNA samples

and dilutions were stored at -20°C.

DNA Amplification using Polymerase Chain Reactions

Nested PCR was carried out for both MTB complex DNA and human
mitochondrial DNA. Primer sequences and PCR parameters are outlined in Table 4. The
first round of PCR consisted of 20 uL reactions using 19 uL of master mix and 1 uL of
bone DNA extract (1:10 dilution of stock extract). A master mix was prepared using 2
uL of Eppendorf Hot Master Taq Buffer, 2 uL of 30 ug/uL bovine serum albumin, 2 uL
of 2 ug/uL deoxynucleoside S’-triphosphates, 0.4 uL of 20 M forward, 0.4 uL of 20 M
reverse primer and IU of Hot Master Taq. Reagent blanks, negative and positive controls
were conducted with each set of PCR reactions. Four uL of the PCR products, from
initial amplifications, were electrophoresed on a 3% agarose gel. Ethidium bromide
staining of the gel was conducted followed by UV visualization.

Semi-nested or nested PCR was conducted on those samples that contained no
visible or a faint product. Semi-nested PCR involved the preparation of a master mix, as
described above, replacing the bovine serum albumin with sterile water, and using a

primer that was internal to either the initial forward or reverse primer. Nested PCR was

34

conducted for the MTB complex reactions using forward and reverse primers internal to
the initial primer pair. One pL of the amplified template DNA from the first round PCR
was added to 19 uL of the second round PCR master mix. PCR parameters used for the
second round are also outlined in Table 4. If no band was visualized in the initial PCR
agarose gel then 20 to 25 additional amplification cycles were conducted. If a faint
product was visualized then 10 to 15 additional cycles were conducted. Following the
second round of amplification 4 [L of the nested PCR products were electrophoresed on
a 3% agarose gel and visualized with UV light.

Controls were examined for the presence of PCR product to assess the validity of
the results. If a product was seen in the reagent blank the number of cycles used in the
second round of PCR was lowered. DNA was re-extracted from the bone sample if a

product was visible in the reagent blank after using this lower cycle number.

35

Table 4: Primer Sequences and Parameters

PCR conditions and primer sequences used in the amplification of mitochondrial and
MTB Complex DNA are outlined. An initial denaturation step of 94 °C for 5 minutes
was conducted prior to cycling. This was followed by a final extension step of 72 °C for
7 minutes.

 

 

 

 

Primers (5’ -) 3’) Product Parameters Cycle
Size N 0.
(hp)
IS6110* P1 ctcgtccagcgccgcttcgg 94°C, 30 sec
P2 cctgcgagcgtaggcgtcgg 123 68°C 1 min 35
72°C 30 sec
183 ttcggaccaccagcacctaa 94°C, 30 sec 20 — 25
[$4 tcggtgacaaaggccacgta 92 64°C 1 min
72°C 30 sec
OxyR“ F cgcgctgtcagagctgacttt 94°C, 20 sec
R tctgcggaatcagtgtcacc 150 66°C 30 sec 35
72°C 30 sec
F2 ttgtgactgcatgaggggc 127 94°C, 30 sec 20 — 25
R tctgcggaatcagtgtcacc 62°C 30 sec
72°C 30 sec
SNP(R) acgcactgcacgacggtggccagc
Mtp40i F ctggtcgaattcggtggagt 94°C, 30 sec
R atggtctccgacacgttcgac 152 64°C 1 min 35
72°C 30 sec
F2 gcaaagttgaacgctgaggt 131 94°C, 30 sec 20 — 25
R atggtctccgacacgttcgac 63°C 1 min
72°C 30 sec
HVl 15989 ccatgcttacaagcaagt 94°C, 30 sec
16207 acttgcttgtaagcatgggg 218 56°C 1 min 35
72°C 1 min
16057 aagtattgactcacccatca
16207 acttgcttgtaagcatgggg 150 94°C, 30 sec 10 — 20
56°C 1 min
72°C 1 min

 

 

 

 

 

 

 

as = Mays et al. (2001) i = Forward and Reverse from Fletcher et a1. (2003);

Nested primers for Mtp40 and OxyR SNP designed in house.

36

Prevention of Contamination during Ancient DNA Analysis

The Armed Forces DNA Identification Laboratory (AF DIL) has developed a
series of protocols to prevent contamination of ancient DNA sources (Edson et al. 2004).
These include the use of personal protective equipment, running appropriate controls,
laboratory set-up, and tracking of contamination. All of these measures were utilized in
this study. During the extraction process and PCR preparation, facemasks, disposable
sleeves, and two pairs of latex gloves were always worn. Laboratory equipment was
rinsed with 10% bleach, washed with detergent and subjected to UV light. Reagent
blanks, negative, and positive controls were run with every experiment. DNA extracted
from a control bone that was known to not contain the pathogen (e. g., bones from
individuals who died from trauma, fresh skeletal material, etc) was used as an additional
control.

Contamination of the bone samples, DNA extracts, and PCR reactions by
pathogen control and exogenous human DNA was of great concern. Dilutions of DNA
extracted from M tuberculosis, M. bovis and B. abortus were used as positive controls
when screening for the presence of pathogen DNA. All stock bacterial DNA and
dilutions were stored in a -20°C freezer located in a separate laboratory from the main
extraction laboratory. PCR experiments were set up by aliquoting the master mix and
adding bone DNA extract in a CleanSpot PCR hood. The negative control and bone
samples were then placed in a thermocycler and held at 4°C. Bacterial DNA was added

to the positive PCR reaction in a separate laboratory, placed in the thermocycler away

37

from other samples, and the thermocycler started. Care was taken to never return to the

extraction laboratory or handle bone samples after using bacterial DNA.

Sensitivity Assays of the MT B Complex PCR Primers

Methods and primers designed for the detection of MTB complex DNA were
assayed for their sensitivity using bacterial genomic DNA. Three 1:1000 dilutions of M.
tuberculosis and M. bovis DNA were prepared. The optical density, at 260 nm, was
determined using UV spectroscopy. Absorbencies were used to estimate that the M.
tuberculosis genomic DNA stock had a concentration of 700 ng/uL and M. bovis stocks
were at 935 ng/uL. The relative molecular weights of a single genomic copy of M

tuberculosis and M bovis were calculated using the method outlined in Table 5.

38

Table 5: Calculation of M. tuberculosis and M. bovis Genome Weight

The total number of base pairs for the bacterial genomes were obtained from the Sanger
Institute website (www.sanger.ac.uk/). Molecular weight of the genome in Daltons was
estimated by multiplying the total number of bases by the average molecular mass of an
NT or G/C base pair (~660 Daltons). This was used to determine the approximate mass
of the genome in femtograms (1 femtogram =10“ grams).

Length 01' Single . Dalton 9 . .
(.enmne ()ne (.enmnre

(Up) (5;)

Species (.‘ennme Base I’nir . Uram
.\l\\ 1|):rlt n _ ‘7 .
(lip) \I“ (1):!) u U \ (nmersmn

 

M. tuberculosis 44] 1532 660 291 161 l 120 1.67E-24 4.86E-15

 

M. bovis 4345492 660 2868024720 1.67E-24 4.79E~15

 

 

 

 

 

 

 

 

It was estimated that a single copy of M tuberculosis or M. bovis 's genome
equaled 4.86 femtograms (fg) and 4.79 fg respectively. Using these estimates the M
tuberculosis stock contained approximately 1.44 x 108 genomic copies/uL and the M
bovis stock approximately 1.95 x 10 8 genomic copies/ pL. Serial dilutions of stocks
were prepared so that PCR reactions varied the amount of template from 0.1 to 10,000
genomic copies. PCR inhibition was also assayed by spiking bone sample DNA extracts
with low copy numbers of M tuberculosis DNA (1 to 100 copies). Nested PCR was then
conducted on the samples to see if the bacterial DNA could be preferentially amplified, in

the presence of human DNA, and visualized on an agarose gel.

39

Differentiation of M bovis and M tuberculosis

Semi-nested PCR was conducted for the OxyR gene using the master mix
protocol outlined above. Primers and PCR parameters used to target this gene are
outlined in Table 4. PCR products were electrophoresed on a 4% agarose gel to
determine if amplification was successful and provide an estimate of the amount of DNA
present.

A SNP assay was designed with the intent of analyzing any samples that produced
positive OxyR products. Methods were first optimized using M tuberculosis and M
bovis DNA. SNP reactions were prepared by adding 2 11L of shrimp alkaline phosphatase
(SAP) and 1 uL of exonuclease I (Exol) to the Oxy R PCR reactions. Reactions were
incubated at 37°C for 60 minutes followed by denaturing of the enzymes at 75°C for 15
minutes. Single base primer extension assays were conducted using a CEQ SNP-Primer
Extension Kit (Beckman). SNP premix was prepared that contained DNA polymerase,
reaction buffer, ddATP, and ddGTP CEQ dye terrninators. Ten uL SNP reactions were
set up using 5.5 uL of SNP premix, 1 uL of either 0.2 M or 2 pM OxyR SNP Primer
(Table 4), 0.5 uL of DNA template and 3 uL of sterile water. Reactions were subjected
to 25 cycles of 96°C denaturation for 10 seconds, 55°C annealing for 5 seconds, and
72°C elongation for 30 seconds. One pL of SAP was added to each reaction followed by
a 60-minute incubation at 37°C, to remove unincorporated dye terrninators, and a 75°C
hold to deactivate the enzyme.

SNP analyses were conducted by adding 0.5 -— 1 uL of the SNP product, 0.5 uL of

Beckman-80 size standard, and 38.5 — 39 uL of Sample Loading Solution. SNP-21

40

parameters (capillary temperature 50°C, denature at 90°C for 60 seconds, inject at 2.0 kV
for 30 seconds, and separate at 6.0 kV for 21 minutes) were used for data collection.
Raw data were analyzed using the Fragment Analysis Program associated with the CEQ

8000.

M tuberculosis Mtp40 Primer Optimization and Primer Specificity Assays

Primers to target the Mtp40 gene were designed and optimized by Kristy Bachus.
The primers were used in a semi-nested PCR method designed to target M tuberculosis.
Experiments were also conducted to determine the specificity of the Mtp40 primers to M
tuberculosis. Primer sets were tested on DNA extracted from a variety of sources
including: E. coli, soil from a Michigan agricultural site (Meyers 2006), soil from
Voegtly cemetery, human blood, human saliva, ancient human bone, fresh human bone,
M tuberculosis, and M bovis. Only reactions that contained M tuberculosis genomic
DNA produced a PCR product of the appropriate size. Specificity assays, such as the one

described above, were also conducted on all other MTB complex primer pairs.

Sequencing of Mitochondrial Hypervariable Region 1

DNA sequencing was conducted on PCR amplicons using the HVl primers

outlined in Table 4. PCR reactions were purified by diluting the reaction in a total

volume of 300 uL sterile water. Diluted reactions were loaded onto a Montage PCR

41

Centrifugal Device and centrifuged at 1000 X g for 15 minutes. PCR products were
eluted from the column by resuspending them in their initial volume using sterile water.

Ten uL sequencing reactions were prepared using 4 uL of CEQ DTCS Quick
Start Kit, 1 uL of 20 mM forward or reverse primer, variable volume of DNA template
(~6 ng) and sterile water. Reactions were subjected to the following sequencing
parameters: denaturation at 96°C for 20 seconds, annealing at 50°C for 20 seconds, and
elongation at 60°C for 4 minutes for 30 cycles. Reverse 16207 often produced poor
sequencing results. Increasing the annealing temperature for 16207R to 60°C and the
extension to 65°C provided more accurate sequencing results (Murray 2005).
Sequencing reactions were halted by adding 2.5 uL of a stop solution (1 uL of 3 M
NaOAc pH 5.3, 0.5 uL glycogen, 0.2 uL of 500 mM EDTA, and 0.8 uL of sterile water)
to each sample.

DNA was precipitated by transferring the sequencing reactions to 1.5 mL sterile
micocentrifuge tubes and adding 30 uL of 95 % cold ethanol. The samples were then
vortexed and centrifuged at 14,000 rpm for 15 minutes to produce a DNA pellet. The
supernatant was carefully removed and 200 [L of cold 70% ethanol was added. Samples
were centrifuged for 3 minutes at 14,000 rpm and the supernatant was removed. The
70% ethanol wash was repeated followed by vacuum drying for 15 — 20 minutes.

DNA pellets were resuspended in 40 uL of Sample Loading Solution, vortexed
gently, loaded into a 96 well CEQ sample plate, and covered with a drop of mineral oil.
Samples were run using the LFR-l program associated with the CEQ 8000 Genetic
Analysis System (capillary temperature 50°C, denature at 90°C for 120 seconds, inject at

2.0 kV for 15 seconds and separate at 4.2 kV for 60 minutes).

42

Raw data were analyzed using the Beckman CEQ Sequencing Analysis software.
Sequences were aligned using BioEdit (Hall 1999) and compared to the reference
sequence (Anderson et al.1981). Forward and reverse HVl sequences were obtained for
a long bone (femur, ulna, or radius), a rib or pelvis, and a thoracic vertebra from each
burial. Those bone samples that showed polymorphisms from the Anderson et al. (1981)
reference sequence at a particular base were noted. DNA sequences obtained from each
bone were compared to one another to produce a consensus sequence for that individual.
Results were also compared to the HVl mtDNA sequences of the analyst to rule out
contamination. PCR reactions were repeated if any contamination was noted, followed

by re-extraction of DNA from the bone if contamination was seen a second time.

Real- Time PCR Primer Design for Tuberculosis, Brucella and Mitochondrial DNA

DNA sequences of genes specific to the MTB Complex (186110, OxyR’ and
Mtp40) and Brucella spp. (IS6501 and Bcsp31) were obtained from the NCBI Website
(Sequence Ascension No: 186110 (X17348), OxyR’(DQ056361), Mtp40 (M57952),
IS6501 (M94960) and Bcsp31(M20404)). These were uploaded into Applied Biosystems
Primer Express TM software, and primer pairs were selected that would amplify a 58 — 63
bp region of the target gene. Real-time PCR primers for human mitochondrial HVl were
also used. These had been designed previously by Michael Gehring (2004) and amplified
a 118 base pair region of HVl. An additional set of HVl primers (Fl6l90: R16251) that

amplified a 61 bp region were also used (Table 6).

43

Table 6: Real-Time PCR Primers

Primers are organized based on their target species and shown with 5’-) 3’ directionality.

 

 

 

 

 

 

 

'l'arget Species Forward (5‘ -)3’) Reverse (5' 93’)
IS6110 MTB Complex gcttagcggcgggacaa gccgacgcggtctttaaaa 62
Oxy R MTB Complex gcgacgaatcggtttggt gcaagacgctggtaggacttct 63
Mtp40 M Tuberculosis cggcgaaatgacaatgca ggtccggtggcattcgt 65
IS6501 Brucella spp. cgcgcggtggattgac agcggtaggccgatagca 58
Bcsp3l Brucella spp. gcgttgggagcgagctt ccagtgccgatacggaaaaa 59
HVl Human mtDNA (F 16190) (R16251) 61

ccatgcttacaagcaagt ggagttgcagttgatgtt
HVl Human (F 1 6400) (D-Loop) 1 18

MtDNA accatcctccgtgaaatcaa accctgaagtaggaaccaga

 

 

 

 

 

 

 

Optimization of Real-T ime Primers

Real-time PCR primers were optimized according to the SYBR® Green PCR
Masterrnix Protocol (ABI 2005). Twenty-five uL reactions were prepared that contained
12.5 [L of the SYBR Green PCR Masterrnix (2X), 2.5 pL of 500 — 9000 nM forward and
reverse primer, 0.5 ng/uL of DNA template, and the final volume adjusted with sterile
water. Dilutions of 200 pM primer stocks were made so that the final forward or reverse
primer concentrations could be adjusted to 50 nM, 300 nM, and 900 nM respectively.
Primer optimization was conducted by setting up reactions that contained the
combinations of forward and reverse primer concentrations outlined in Table 6. A small

set of reactions that contained the improper template for that primer pair were set up to

44

test primer specificity (e. g., 0.5 ng of human DNA was added to reactions that contained

TB specific primers). All reactions were run in duplicate in a 96 well optical plate.

Table 7: Combination of Concentrations Used for Primer Optimization.

Primer concentrations are in nanomolar (nM) and are shown as a ratio of forward to
reverse primer.

 

 

 

F (900nM) F (300nM) F (5011M)
R (900nM) 900900 300900 50900
R (300nM) 900300 300300 50300
R (50nM) 90050 30050 50:50

 

 

 

 

 

 

Real-time PCR reactions were loaded onto an ABI Prism 7000 thermocycler. An
initial 50°C hold for 2 minutes was conducted to activate the enzyme Uracil N-
Glycosylase. This was followed by a lO-minute 95°C enzyme deactivation step that
simultaneously activates the AmpliTaq Gold DNA polymerase. Reactions were
subjected to forty cycles of 95°C denaturation for 15 seconds and one-minute
annealing/extension steps at 60°C. Thermal cycling was concluded with the generation
of a dissociation curve by ramping of the sample from 64°C to 95°C. Results were
analyzed using the ABI 7000 Series software. According to the SYBR Green PCR
Mastermix protocol the primer combination that produced the lowest Ct value with the
smallest standard deviation is optimal, so these concentrations were used in all future

real-time reactions.

45

Real-Time PCR on Ancient Bone Samples

Twenty-five 11L real—time PCR reactions were prepared using the optimal primer
concentrations. Two and a half 11L of template from bone DNA extracts were added to
each 25 uL reaction. Additional reactions were set up for each bone that contained 2.5
uL of a 1:10 dilution of the initial extract. All reactions were run in duplicate and
screened using primers for the six target genes outlined in Table 6. Negative controls,
reagent blanks and appropriate positive controls were run alongside each set of reactions
to check for contamination and proper amplification.

Real-time analysis was conducted on the ABI 7000 Prism using the same
conditions described above with the exception of 50 cycles rather than 40 as this has been
found to produce better results with degraded DNA template (Andreasson et al. (2003).
Amplification plots and dissociation curves of the bone samples were compared to those
of the positive control to assess whether or not the DNA extract contained the appropriate
product. Bone samples that showed positive amplification were run on a 4% agarose gel

and stained with ethidium bromide.

46

Table 8: Summary of Analysis Conducted on Pathology and Questioned Bones
Summary of tests conducted on each bone sample. TB PCR= 186110 and OxyR

screening. MtDNA PCR = (X = 15989:16207 amplification) and (X* = 15989:16207
amplification and sequencing). RT = Real-Time PCR; followed by the region targeted.

Rib

Thoracic Vert

Ischium
F emur
Ulna

><><><><

M

Thoracic Vert
Rib
F emur
Fibula
Ulna

><><><><><><

M

Thoracic Vert

Thoracic
Vertebrae
Rib
F emur
Fibula

M

><><><>< ><

Thoracic
Vertebrae
Rib
F emur
Pelvis

M

><><><>< ><

Thoracic
Vertebrae
Rib
Pelvis
Radius
Fibula

><><><><>< ><

 

47

RESULTS

Bone Sample Preparation and DNA Extraction

The bones most often used for analysis were thoracic vertebrae, femora, and ribs.
Bones from the pelvis were tested if a burial did not contain a rib, while other long bones
(ulna, tibia or fibula) were examined if femora were absent. Drilling of long bones
resulted in clean white powder obtained from the cortical bone. In contrast, samples with
a high amount of spongy or cancellous material, such as vertebrae and ribs, contained a
large amount of dirt and debris. These bones usually resulted in a yellow to brown
colored powder.

Following digestion of the bone powder the resulting aqueous layer was clear for
long bone samples and red to brown for others (Figure 6). While some of the red or
brown color could be removed during the organic extraction it was impossible to remove
all of it. This resulted in DNA extracts from vertebrae and rib samples having a light

brown or black coloring.

Figure 6: Variability of Aqueous Layers

Color variation of bone preparations from
burial 4015 (image taken following first
phenol extraction). On the left is

white/1i ght yellow colored extract
resulting from the digestion of femur bone
powder. The tube in the middle is the
typical coloration of a rib extract while the
tube on the right is contains digested
powder from the vertebral body.

 

48

MT B Complex Primer Optimization and Sensitivity Assays

Primer concentrations and annealing temperatures were optimized on M
tuberculosis and M bovis DNA. All MTB complex primer pairs gave a clean product
using 20 M of forward and reverse primer except IS6110. The primer pair designed to
target 186110 was optimized at a concentration of 10 11M forward primer to 10 11M
reverse primer. Optimum annealing temperatures were determined using a temperature
gradient ranging from 55°C to 70°C. All primer pairs had optimum annealing
temperatures close to those previously published and are reported in Table 4.

Primer sensitivity assays were conducted using 100,000 to 0.1 copies of the M
tuberculosis or M bovis genome as template in PCR reactions. Minimum detection
levels for the external primer sets ranged from 10 — 150 genomic copies after 35 cycles of
amplification (Table 9). Semi-nested PCR was required for reactions that contained
fewer than 100 copies of template to produce a visible product. Following an additional
25 cycles of amplification sensitivity levels ranged from 1 — 15 copies depending on the

primer pair. An example of a primer sensitivity assay for 186110 can be seen in Figure 7.

49

Figure 7: Primer Sensitivity of [$6110 on M tuberculosis DNA Dilutions

Ten pL PCR reactions were prepared that contained serial dilutions of M tuberculosis
template. Following amplification 4 11L of each reaction were loaded onto a 4% agarose
gel. (a). PCR using IS61 10 forward and reverse primers on M. tuberculosis DNA
template ranging from 45000 picoMoles (pM) to 0.04 pM (~ 100,000 to 0.1 genomic
copies). IS61 10 PCR products are indicated on the gel by the white arrow. (b). Nested
PCR was conducted on the 4.5, 0.45 and 0.045 pM (~10, 1 and 0.1 copies) samples

depicted in Figure 7(a). PCR products obtained are outlined in the white box. A 50 base
pair ladder is indicated by ‘50 bp.’

(a).

Negative
45000 pM

n.
.0
0
Ln
.
-
-
-
-
.
.
.

N

(b)- ' l .
. 150 bp

. 100 bp

. 50bp

 

50

Table 9: Minimum Number of Genomic Copies Detectable

Primer sets were tested on 100,000 to 0.1 copies of M tuberculosis (M. TB) and M bovis
(M. Bovis) genomic DNA. The lowest number of genomic copies detectable is

represented by the value in parentheses. Number under the gene name indicates the
number of cycles used. N_25 = Number of cycles used in nested or semi-nested PCR.

 

 

Species 1861 10 1861 10 OxyR Oxy R Mtp40 .\11p4()
(35) (N_25) (35) (N25) (35) (N 25)
M. TB 45 fg 4.5 fg 450 fg 45 fg 700 fg 70 fg
(~10) (~ 1) (~95) (~ 10) (~145) (~ 15)
M. Bovis 450 fg 45 fg 450 fg 45 fg
N/A N/A
(~95) (~9) (~ 95) (~10)

 

 

 

 

 

 

 

 

Optimization of Methods used to Differentiate M tuberculosis and M bovis

Figure 8 depicts results from an OxyR SNP assay optimization experiments. Two
sets of SNP extension reactions were prepared; the first used 0.2 1.1M of the OxyR SNP
primer and the second used 0.02 M of SNP primer. Reactions were initially conducted
with 1 1.1L of PCR product as template. This resulted in a large amount of pull-up and
non-target artifact peaks at both primer concentrations. A 1:10 dilution of the initial PCR
product was used in future reactions. M tuberculosis produced a black peak at 19.37
corresponding to a cytosine. The M bovis electropherogram shows a blue peak, thymine,
at 21.04. DNA base complementarity indicates that M tuberculosis has a guanine at
position 285 and M bovis contains an adenine. The negative and human DNA controls

did not produce any peaks.

51

 

Figure 8: Differentiation of M. bovis from other Members of the MTB Complex

Results from SNP analysis of base 285 in the OxyR gene. The top two
electropherograms correspond to SNP reactions using human DNA and negative control
respectively. The third represents results from reactions containing M tuberculosis DNA
(black peak at 19.37), while the fourth is from M bovis DNA (blue peak at 21.04). The
red peak at 13 is a size standard.

200000
2 150000
.9»
m 100000
0
,5
a 50000 14.66

0 I: 16.98

 

 

 

 

 

 

 

 

 

 

 

 

 

200000
2 150000
.9
co 1 00000
g. 50000
a 13
o g 12. .L
200000
2 150000 19.37
.9 _
to 100000
0
5" 50000 13 16.01
' - 19.21
a A .
200000
.1; 21.04
: 150000
.9
g 100000
)5
0 5°°°° 10.48 13 17.58
0 r r r r n - 1/\r ;' r ,
1 a 1
o 10 20

52

Mtp40 Primer Specificity Assays

Only reactions containing M tuberculosis genomic DNA produced a product of
the appropriate size for both the external and semi-nested primer pairs (Figure 9). The
[$6110 and OxyR primer sets were specific to M tuberculosis and M bovis and did not
cross-react with DNA from any other sources. However, when the forward OxyR primer
was used with the OxyR SNP primer cross-reactivity with DNA from human saliva was
noted.

Figure 9: Mtp40 Specificity

A one hundred and fifty bp product was only visible in the reaction containing M
tuberculosis DNA (indicated by arrow). Bacterial DNA extracted from E. Coli,
agricultural soil (Agr. Soil), and Voegtly cemetery soil (Voegtly) did not produce any
visible product. Human DNA from blood, saliva, fresh bone (F. bone) and ancient bone

(A. bone) were also negative. 50 base pair ladder (50 bp), negative control (Neg) and
empty lanes (empty) are indicated.

 

53

Screening of Bone Extracts for Tuberculosis DNA

DNA was extracted from each Voegtly bone sample three times and PCR
screening for the IS61 10 element was conducted twice on each extraction. Initial
extractions of a femur and rib sample from burial 32 yielded a 123 bp product for IS6110
(Figure 10). DNA was re-extracted from the rib and femur of this burial to determine the
reproducibility of results. Only DNA extracted from the rib yielded a reproducible
amplicon. Burial 32 extracts did not produce any visible product when used with the
OxyR or Mtp40 primer sets. Burial 629 did not produce positive results for any MTB

complex assays.

Figure 10: PCR Screening for the Presence of 186110 in Voegtly Samples

IS6110 nested PCR on Burials 32 and 629 run alongside a 25 bp ladder (25 bp). Letters
indicate the type of bone (F = femur; R =rib) and 1:10 is a tenfold dilution of the neat
reaction. Products approximately 123 bp in size can be seen in the 32 rib sample and its
dilution (as indicated by arrow). These products correspond to the expected amplicon
size for IS61 10 (123 bp).

629R1:10-

b
‘.".
1-
LL
N
t")

 

54

 

Bone samples from Butrint and Diaporit that were screened for IS6110 are
outlined in Table 10 and Table 11 along with the number of times DNA was extracted.
PCR reactions to screen for [$6110 were conducted a minimum of three times on
vertebra, rib/pelvis and a long bone from each individual. Additional bones, such as
metatarsals and other long bones, were assayed only in burials that showed signs of
pathology. No visible products for IS6110 were obtained from any of the bone extracts.

An example of experimental results from vertebrae samples is provided in Figure 11.

Table 10: Control/Livestock Extracts and 186110 Experiments
Number of times each bone sample from the control and livestock groups were drilled

(indicated by the ‘Number of Extracts’). The number of experiments in which that bone
extract was used to screen for the MTB complex is also provided.

Thoracic
Left Rib

Tibia

Thoracic Vertebrae
Radius

Metatarsal

Ulna
Calcanium

Vertebrae

Ulna
Radius

 

55

Table 11: Questioned Pathological Samples and 186110 Experiments

The number of time each bone sample from the tuberculosis/brucellosis consistent and
questioned pathology groups were drilled (indicated by the ‘Number of Extracts’). The
number of experiments in which that bone extract was used to screen for the MTB
complex is also provided.

Thoracic Vertebrae

Z
a
00

Ischium
F emur
Ulna

Thoracic Vertebrae
Rib
Femur

Thoracic Vertebrae
Rib
Femur

Thoracic Vertebrae
Thoracic Vertebrae

N
N
N
N
N
N
N
N
N
N
N

2

Thoracic Vertebrae (Neural
Arch
Rib
Femur
Pelvis

Thoracic Vertebrae
Thoracic Vertebrae

22 2222

Thoracic Vertebrae
Rib
Pelvis
Radius
Fibula

22222

 

56

Figure 11: Screening of Vertebrae Samples for 186110 using Nested PCR

DNA extracted from vertebrae was screened for the 156110 element. Reactions were
conducted as a neat reaction and at 1:10 dilutions. (a) Reactions following 35 cycles
using the 186110 external primers. (50 bp = 50 bp ladder, Empty=no sample loaded, Neg
= negative control, RB = reagent blank, and Pos = positive control). Only the positive
control produced an amplicon of the appropriate size (arrow). (b) Nested PCR for
IS6110 conducted for an additional 30 cycles. Only the positive control produced an
amplicon (93 bases) of the appropriate size.

(a).

1,213? ..
.2722 - .2
2722110
40151210
5010110

e:
.‘i':
..
.n'
f.
N

‘ ‘2711 1:18?

10“" -_

a
a

4015 1:101;
5010 1

 

57

Spiking Human Bone DNA Extracts with M tuberculosis Genomic DNA

The sensitivity limits of the IS6110 PCR assay, previously determined using
bacterial DNA, were tested in the presence of vertebrae DNA extracted from burials 213,
319, 2722, 4015, and 5010. The first experiment consisted of spiking each reaction with
~100 copies of the M tuberculosis genome prior to amplification. Gel electrophoresis of
the PCR product indicated that it was possible to detect 100 copies of the M tuberculosis
genome, in the presence of human DNA, without using nested PCR (Figure 12a). A
second series of experiments was conducted assaying the ability to detect one copy of the
M tuberculosis genome using semi-nested PCR. A single copy of the genome was
detectable in three out of five samples (Figure 12b).

Figure 12: 186110 PCR on Bone DNA Extracts Spiked with M. tuberculosis DNA
Figure (a) depicts the detection of 100 copies of M tuberculosis genomic DNA using the
PCR method designed to target IS61 10 (PCR products are indicated by the white arrow).
Figure (b) depicts the IS6110 nested PCR protocols ability to detect concentrations as
low as one copy of the M tuberculosis genome (PCR products are outlined by white box)
in the presence of human DNA extracted from vertebrae (V). Negative control (Neg),
positive control (TB (* = over amplified», human control DNA spiked with one copy of
M tuberculosis (Human +TB) and human control DNA (Human)) were also conducted in

(a).
(a .

E1111»;

101111.111

3
on
U)
—
-_.
. _.
.-
.-
-
-
-
-

 

58

 

 

Mitochondrial DNA Amplification

PCR reactions for mtDNA HV1 were conducted on DNA isolated from vertebrae,
ribs/pelvises, and long bones. Initially a neat reaction (1 1.1L of DNA extracted from bone
into 19 11L master mix) and a 1:20 serial dilution were used. However, after repeated
experiments it was found that neat reactions were usually inhibited or did not amplify.
Given this, semi-nested PCR needed to be conducted on all extracts in order to obtain a
product for sequencing. Long bone samples required fewer cycles during the second
round of amplification than vertebra and rib samples. PCR reactions of long bone
samples also had a higher tendency to over amplify. Over amplification was corrected by
dropping the number of cycles during semi-nested PCR from 20 — 25 to 10 - 15 (Figure
13).

Figure 13: Semi-Nested HV1 Results using DNA Extracted from Long Bones

Amplification of 1:10 dilutions of mtDNA extracted from long bones. Numbers
correspond to burial number and letter to bone type (U=ulna, R=radius, and F=femur).
(a). Results following 35 cycles of amplification. The double band in the positive control
is an artifact witnessed with the primer set 15989F216207R. (b & 0). Products obtained
after 10 and 20 cycles of semi-nested PCR. Products in (b) are examples of those used
for cycle sequencing, while (c) are over amplified. The white arrows indicate PCR
products of the appropriate sizes.

 

59

DNA extracted from ribs and vertebrae required a higher number of cycles than
long bones. Following the initial 35 cycles an additional 20 to 25 cycles of PCR were
conducted. Rib and vertebra DNA extracts, especially samples from burials 213 and
5010, had a high incidence of inhibition. The bones from individual 213 had a red color,
which was carried over to the DNA extracts. Similarly, the bones from individual 5010
contained a greater amount of soil than the other burials. Despite additional washes
during the sample preparation and selecting areas for drilling that did not contain a large
amount of debris, the extracts were black in color. Inhibition of the PCR reactions was
overcome by the addition of BSA in the first round and also by diluting the initial DNA
extracts at 1:20 rather than 1:10. An example of inhibition can be seen in the lane
containing reaction 5010V in Figure 12 (b). This is indicated by the lack of amplification

of the product and the absence of primer dimer.

Figure 14: HV1 Semi-Nested PCR of Vertebrae Extracts

Example of semi-nested PCR results on a region of HV 1 in mitochondrial DNA. Figure
(a) shows reactions following 35 cycles of amplification. Figure (b) shows the same
samples after nesting the reactions for 25 cycles using internal primer set 16057F:16207.
The white arrows indicate PCR products of the appropriate sizes.

 

60

Hypervariable Region I Sequencing

Table 12 lists the bone samples used for sequencing and the ranges of HV1
sequences obtained. Forward and reverse sequences from vertebrae, ribs and a long bone
were obtained for burials 213, 319, 2722 and 5010. Sequence was not obtainable from
burial 4015’s vertebra despite successful amplification using semi-nested PCR. Failure
appeared to be due to a lack of DNA template in the sequencing reactions or the results

could not be analyzed using the CEQ 8000 software.

61

Table 12: Ranges of HV1 Sequence Obtained for each Bone

The range of sequence obtained for HV1 organized by burial number and bone type.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Burial Bone I, Range of HV1 Sequence Total Number of Bases
Sample :. Obtained

213 Ulna 16059 —- 16208 149
Rib 16082 — 16204 122
Vertebra 16060 — 16205 145
319 Femur 16069- 16213 144
Rib 16081 — 16203 122
Vertebra 16082 — 16204 122
2722 Femur 16088 — 16216 128
Rib 16058 — 16203 145
Vertebra 16058 — 16206 148
4015 F emur 16060 — 16206 146
Rib 16058 — 16207 149
5010 Femur 16098 - 16204 106
Rib 16058 — 16203 145
Vertebra 16058 — 16205 147

 

 

 

 

 

62

 

 

The sequencing results from each bone are outlined in Appendix A: HV1
Sequences from Individual Bone Samples. These sequences were compared to one
another and the Anderson et al. (1981) reference to determine a consensus sequence for
each individual. A degenerate base was called if two bones from the same individual had
a different base at a single site (for example burial 5010 rib and vertebra showed a C at
position 16103; while the femur showed a T; therefore the position was called a Y). A
comparison of consensus sequences is provided in Table 13. The sequence of the analyst
throughout this region is also provided to demonstrate that sequences obtained were not a

result of contamination.

Table 13: Consensus Sequences Derived from Bone Samples for Pathogen Burials

Consensus sequence for each burial compared to the Anderson et al. (1981) reference
sequence. A lack of a letter in a box indicates that the individual contained the same base
as the reference at that particular position. Grey boxes denote areas where sequence was
not obtained at that site for that particular individual. R = A or G; M= A or C; Y = C or
T.

Anderson A

Burial #
213
319
2722
4015
5010

 

63

Real- Time PCR Optimization

Different combinations of forward and reverse primer concentrations were used to
optimize real-time PCR conditions. The concentrations that produced the lowest Ct value
and smallest standard deviation between duplicates were selected and used for all further
studies. Concentrations, average Ct and standard deviations for the all genes are outlined

in Table 14.

Table 14: Optimum Primer Concentration and Average Ct Values for Primer Pairs

Primer concentration ratios that gave the lowest Ct and smallest standard deviations are
reported for each gene as the concentration of forward to reverse in nM.

Primer Target Optimum Primer Average (‘1

Concentration
Forward/Reverse (11.\’l)

 

 

 

 

 

 

 

IS6110 300:300 16.80
OxyR 900:300 22.40
Mtp40 300:900 19.31
IS6501 900:900 19.02
Bcsp3l 900:900 21.1 1
HV1 (61 bp) 300:300 20.70
HV1 (118 bp) 300:300 19.86

 

 

 

 

 

64

 

Real- Time PCR Results of Voegtly Cemetery Samples

Real-time PCR reactions were conducted on DNA extracts from Voegtly
cemetery’s burial 32 that had previously tested positive for IS6110. Amplification of
mtDNA (Figure 15) and IS6110 were successful. Dissociation curves of PCR products
obtained from bone DNA extracts were compared to those of positive controls. If the
curves contained similar overall patterns (i.e. the plots of the positive and questioned
samples had a similar shape) and average melting temperatures then the extracted
samples and the positive controls were assumed to have amplified the same product.
Figure 16 shows the dissociation curve for the mtDNA reactions. Human control DNA
and reactions containing burial 32 rib DNA had similar average Tm (754°C and 762°C
respectively). Figure 17 shows the dissociation curve for the IS6110 PCR reactions. The
average Tm for the M tuberculosis IS6110 product was the same as the product obtained
from 32 rib (81 .4°C). Amplification seen in the negative controls or reagent blank
usually had a lower Tm and most likely resulted from primer dimer or small artifact
products. This indicates that both mtDNA and the IS6110 element were successfully
amplified from the burial 32 rib reactions. Results verified that MTB complex DNA was

present in the bone extract and indicates the individual had tuberculosis.

65

 

-1 _ 9 _ o M . _ a H o _ o _ v _ o r a H .1 1L:

89.52 loud 11 . .

Savavncovnvvvnvnv wvovunonhnnnunvnnnnnsagasnggcnnunuunoNOwopspopowvwner wan-pa o s o o v m up

80.00. p

Manda. w

 

 

 

. m
«8.09 p I
m.

98.00%

0009.06

£9580 “Eamon <75 58:2
-. r0950;
saga

12313983 8...... 5333...... .1.

 

 

#2303. £8585
a sl_g.lwfil.u.l;fl',.l|.].

 

.eouombxo ofi 80¢ :53 “common 05 mm on: :0on 05 was .9980 gunmen a m_
on: 3% 98:96 05 mm at ham 23 .5382 “we 05 m_ in man 68033 8m £8388 ENG at mm 395 can £838 9563 <75
58:: .238: Eoomoaozc 03:20: am SEQ .m> “3:52 296 me 38:23 mm BE .83 eozmoEEEa mug 08228.2 48.29.: 58::

am an seam 2.9.5 .5 8: 558515. <29: "2 9...»...—

66

9...?

>1. ..

86

 

 

 

 

R
\\<.

 

new

111 1 11 1 l -l.ll 2...

// \ E £9280 “Eamon <20 5:53 _

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0°va 839:8 <20 a: mm wag—:55”. £8582 v5

UoNéh me E... amounts ca was $29 75:8 SEE—c flea—=8 0238 2:. iguanamoa mun: woe use 50% 05 8e 8:83 “common use
£838 03832 .3238“ Mom Mmm 3:5 0839: 8:: 298m 93 0:3 22. 6:: 88:33 98 30:9» 05 .3 3332: 2a omen:
£5 E 295:8 032mg .350 nouflooamv 05 mo 0332.59 .m> 8383th no cove—Sm a ma $260.5 Mom 5298 we seaflocmma

an an 3.5m 2325 .8 2:5 3382.: <29: a: ea»...—

67

 

 

 

 

 

 

 

 

 

 

 

 

 

£9980 0230a 05 55 832 EH a 5:5 8:88 oEoomméoe a 8328.8 Ame MN 0 8:82 mum Baum—mac 22.

.001; .8 EH owflozw as 3: 08388 <29 at mm wfififleoo 82888 95 .8 28 can A<2Q 8.88 5:55 £9980 3E8.“ 50m
£53 Howey. use 8.88 gunmen 05 808an nouaowzmfim .8 mews o: 523 8:: 29:8 we“ 52w 22. @395 8: En someone.“
28:86 05 6:: up. 05 E @8885 £ 8288c Mom AMNMV <29 at mm 3.85 05 .8 0:0 6:: 29:8 can 0:3 05 .3 @88va
93 £888 328m .023 82880me 03 ,8 gaming .m> 58883. .8 coco—Ha a mm 8898 mom 2 SE .8 8:308me

am an 3:8 282$ he 3:6 8:28an ESE "S 2.55

68

Table 15: Summary of Real-Time Results for Voegtly Burial 32 Rib mtDNA

Ct-values and melting temperature (Tm) data obtained during the amplification of
mtDNA extracted from burial 32 rib. HV1 (61 bp) indicates that the smaller 61 bp
primer set was used in these experiments. The Ct values refer to the cycle number in
which the product crossed a set threshold. Reactions that did have any detectable
amplification are indicated by Undetermined. *=duplicate reaction

   

mCR'RéfiétiM—wwalum H-zl‘mtt’cl—

 

 

 

 

 

 

Negative Undetermined 71.2
Reagent Blank Undetermined 69.2
HV1 (61 bp) 33.73 75.9
HV1 (61 bp)* 37.89 76.5
HV1 Positive 20.71 75.2
HV1 Positive* 20.66 75.6

 

 

 

 

Table 16: Summary of Real-Time Results for Voegtly Burial 32 Rib 186110

Ct—values and Tm data obtained during the amplification of 186110 extracted from burial
32 rib. One of the reactions that contained 32 rib DNA produced a product with a similar
Tm as the M. tuberculosis positive control. Reactions that did have any detectable
amplification are indicated by Undetermined. * = duplicate reaction

   

 

 

 

 

 

 

 

- .. 'Favaluw‘
Negative Undetermined 71.5
Reagent Blank Undetermined 87.8
186110 43.32 81.4
IS61 10* Undetermined 82.0
Positive Control 27.52 81.4
Positive Control* 27.05 81.4

 

 

 

 

69

Real- Time PCR Results for Butrint and Diaporit Pathological Samples

Real-time PCR experimental results for burial 4015 vertebra are provided in
Figures 18 — 21. Mitochondrial DNA and Brucella spp. genes (IS6501 and Bcsp31) were
successfully amplified. The amplification plot for mtDNA is shown in Figure 18 and the
dissociation curve in Figure 19. PCR reactions containing human control DNA and bone
DNA extract produced similar average melting temperatures (754°C and 762°C). The
amplification curve and dissociation curve for IS6501 are depicted in Figure 20 and 21.
DNA extracts from 4015 vertebra appeared to have amplified the same product as the B.
abortus control DNA (average Tm 812°C). The negative control of this sample had a
lower Tm of 77 .8°C. The IS6501 negative control, positive control, Bcsp31, and IS6501
real-time PCR reactions were electrophoresed on an agarose gel (Figure 22). Results
indicated that the product produced in the negative control was different from those in the
samples and positive control. Furthermore, the gel confirmed that the PCR products
amplified from the bone were the same size as the positive control.

A summary of real-time PCR result can be found in Table 17. Using real-time
methods, mtDNA was successfully amplified from burial 319 vertebra, 2722 ribs and
4015 vertebra, ribs and femur. More importantly, the ribs of individual 2722 produced an
amplicon for both Bcsp31 and IS6501, while 4015 produced amplicons for both genes
from vertebra and rib samples. Results were verified by conducting real-time PCR for
Brucella spp. on DNA extracts that had been prepared four months earlier, prior to

receiving B. abortus control DNA in the laboratory. The Ct values for all positive

70

reactions are outlined in Tables 18 — 20. Results indicate that Brucella spp. DNA was

present in the remains of individuals 2722 and 4015.

71

 

_ I _ a _ u h n _ n H u _ o 1th1in11L. . 1 u. 1 L. w 1. H.

.‘33'833:‘3'.‘II‘R§""§IIK8«~&'=2nun—=39:=3.o u an. n a .
:::::::j,i,.:,:;ff
, , ,.. ,,fl , , ,W, ,7 ,
WW ,,,i :: W 3 ,,
,W, , ,

 

 
            

, ,, ,

, ,g, ,3, i

,7 ",7, ,,H,,”

,,,,,,,W _V, , ,
,

i

W

+ .1, a z .. .1 “1. .1. .,+ M, rm 3, L 1 .1. 1.

, , W , 7 , ,
. , , ,, 3 ,, 803832 T

-,.r. T, 8.82

"am

L k E‘—

 

 

.g.

 

 

 

 

 

p SBA-IIEOEIQ :
.88 2: 8 882?: 88 m_ 8.2: 80mm»: 2: dozwoczmfia Rhona—$8:
box: :85 m_ £5 85 $8085 A3 oEwEV 03:0 839.8me 06 £96.50: .8288 a 88:88 88 8.980 guano: 2F finer—2,
m :8 .3 «.8885 m_ can 85—95 28:82 0823:: 05 .8 28 >80 88383 93 88:82 Snot? m Se :5 <29 8.83
gm A38: 80898:: 0388.5 :M 38D .m> 838:2 2&0 .3 @0888 m_ 83 .88 8:363:28 MDm 28252 <ZQE 55::

.325» 23. .5. 8: 5:85.35. <29: a. £5...

72

 

 

 

 

 

 

via—magmas 1 1 1 1Q 1 1 1 1i: 1 1 \fing 3H9

3.980 02332

 

 

 

.".‘
°
BMW/WOO

 

1 11 12°

 

 

    

 

22:50 BEE <2: 5:5:

 

wad $to> m

 

and

 

 

 

 

 

 

 

 

 

 

£53 Bows.— ofi mflmomvi on: :2 05 233 $535: .0: Eu £838 03892
53:00am“: not? 23 00380 a: a Rs :5» 23V 83%.»: <2: 58.? m § 333:8 28:88 Be 22: .858
«88:5 28:80 3325 .933 nouumoOmmmu ofimo 353:5 .m> chafing“: mo corona a ma 8::ng Mom 4295 we 5:308me

nun—33> mgv you 9:39 notfiocuafi <29: 5— unaut—

73

 

 

3 J I Im a I— . '— h I.— c I.— u J v .— n I n ' p .-
86 ,JF‘WiiiLJ “““““ [_
.8'3'339Ueugfllfiflnaaafiflgflifin‘flfifi...2:.—3333:2u . _ a o . n «p
iv

 

1, 4,,.4: , ,fi,,,<, H

;,f,d, f,f:f,j,f:;;

, ,_ if ,,i, , _,m,,
,W , , W

 

 

 

 

 

 

 

 

 

 

 

A on: son Ema 8388 :53 «common 2: E :08 ms? :o:8m:a8u oz 6:: :83 06 3 @8865 3 5:82 35:8 ”Eamon
05 E :03 cm? 33 5:32;; .on: 36:93 :5 Exam 5:3: 05 3 con—865 8m 3383.. AN: 23 Sfio> m 3:; Snot?

m 89. 3:5 233 8:: >>o=9h :5 2:: Em: 2: a: 850% a 8:85.95 <29 35:8 3536 .m ABE: E0823: 0330.5

é £5 .9 h3:52 2&0 & 8.2%.. a 83 Be got? 23 3:5 5 83 852%; «B 05.32 882 an” 532%

«.5023» m2? .3.— BE =c3aa§35< Snow— oEwH-tcm 6N Piaf.—

74

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Swfloafilgfl r I. J,— Iu‘. JJ . 1|? n U ‘ . J
I a 3 ' an as ' 3
\‘1‘w‘1‘1‘ .2
f 2:3 55> 28 f
‘ / ‘ m
»
Rd
59380 .2
3336 .m

 

 

 

 

 

 

 

 

 

 

 

J53 “downp— 2: E :08 33 8:12: oz 0%.: mo EH a 5:5 5: 852; a “63:8: 8?
35:8 0339: 2F damage EH 5 3: Am :3 :36 Ban—83 <75 Snot? m :3 3.5: mam—38:8 30:82 was 275 3:3:
.5 £838 03:8: .955 533083: 2: mo 0333.5: .m> 0558388 no 8:253 a mm 835i #5.: 832 .«c 5:32.35

«sot; 28 3. 25.0 5:882: 389 os_e-_8m "a 2.qu

75

 

.0N30 0:388:90 0:: :0 0:02:88 :0: 0:00:05 030:0 02:3

0:.r in“. 33050.: 88.: <ZQ “00:30:80 8053 :08 380:: 8:03? m 8:. 0:: 0:00:05 0:30: .8388 2:63 0 8038:: 00:00:96
0:: .8 0:0 3:8 8:: 08 Z ._m%0m 8&0 0:00Em :8 «00:80:00 8:: 08:000.: Mom 08: :00”: A0: .Smom: in: 0:00.95 8‘: :0:00:0m
8:: 08:80: Mon: 0E:-_00m 3: 008500 08: 0:: E :00m 003 8:03:98 02 d: Gm: win: 0:03:80 mks: 0:: :0 08:50::

8.: @0588 8:: 08:08: :0: 08:80: A3 .9060: :0::0::0> .8 :08 380: 0:: :8: 80:03: <75 H <2: .22 m 8:: :8 <2

m 5:. z: “00:00:05 mm Em 0 .8 E: 0:03 08:80: 0:00:95 .996 20:80 03:00: :80 .302: 08:80 03:36: .55 00:83 8030:
02080—0 ._0m :08 E S: 003 8200— :: mm < 08:000.: MD: 08:80: :8: “005050 0:03:08 <ZQ 58:8: .8 £0w 088w<

a..::._0> m2... 8.: 0:303— :0u0::am mom 0E_H-_00~— .8 .00 08.83: ”Nu 0.53:—

Table 17: Summary of Real-Time Results

Bone extracts used in real-time assays. HV1-RT indicates real-time PCR assays for

mtDNA HVland IS6501-RT/Bcsp31-RT identifies real-time assays for Brucella spp. A
positive result (Pos) indicates that the sample amplified and the product had a similar Tm
and dissociation curve as the control DNA. All bones were negative in MTB complex

real-time assays. * = Older Extracts Drilled 5/7/06 — 5/20/06 ; all others are from
8/10/06. :1: = Extract was very dark in color; additional 1:10 dilution conducted

Burial

 

219

Bone/Extract
Vertebra

llVl-R'I‘

[8650 l -R'l‘

Bcsp3l-R'l

 

319

Vertebra

 

5010

Vertebra

 

2722

Vertebra“

 

Vertebra Neural Arch

 

Vertebra Body

 

Vertebra Lesions

 

Rib*

Pos

 

Rib

Pos

Pos

 

Rib (1:10p:

Pos

Pos

 

Radius*

 

4015

Vertebra“

Pos

 

Vertebra Neural Arch

Pos

 

Vertebra Body

 

Vertebra Lesions

 

Rib*

Pos

 

Rib

 

 

 

Femur“

 

Pos

 

 

 

77

 

Table 18 (a — 1): Ct Values and Tm for MtDNA Positive Reactions

Ct-values and Tm data obtained during the amplification of mtDNA fi'om burials 319
vertebra, 2722 rib, 4015 vertebra, 4015 rib and 4015 femur. Lower and higher than
average Tm values seen in negative controls and reagent blanks are the result of non-
specific amplicons/primer dimers. All DNA extracts were prepared in August of 2006
except those indicated which were prepared four months earlier in May of 2006 (* =
duplicate reactions). Reactions that did have any detectable amplification are indicated

by Undetermined.

(a) Burial 319 Vertebra

 

 

 

 

 

 

P('R Reaction (1 Value 'l'm (of)
Negative Undetermined 70.3
Reagent Blank Undetermined 69.2
HV1 (61 bp) 33.73 76.5
HV1 (61 bp)* 33.86 76.8
Positive Control 20.71 75.2
Positive Control“ 20.66 75.6

 

 

 

 

(b) Burial 2722 Rib and 2722 Rib (1:10) Dilution

Ne 've
Blank
HV1 61
HV1 61 *

HV1 61 1:10
HV1 61 1:10*

Positive Control
Positive Control“

Undetermined
Undetermined
27.66
40.5
3 1.02
3 l .77
18. 1 1
19

(c) Burial 4015 Vertebra* (May 2006 Extract)

 

 

 

 

 

 

 

PCR Reaction (‘t Value 'I‘m (0(7)
Negative Undetermined 80.4
Reagent Blank 34.00 78.1
HV1 (61 bp) 36.89 76.5
HV1 (61 bp)* Undetermined 67.6
Positive Control 22.16 74.4
Positive Control* 21.73 76.5

 

 

 

 

78

 

 

(d) Burial 4015 Vertebra Neural Arch

 

 

 

 

 

 

 

 

Negative 34.12 69.0
Reagent Blank Undetermined 70.5
HV1 (61 bp) 30.27 76.0
HV1 (61 bp) 32.15 76.4
Positive Control 20.53 75.2
Positive Control” 20.32 75.5

 

(e) Burial 4015 Rib* (May 2006 Extract)

 

 

 

 

 

 

PCR Reaction ('t Value 'I‘m (00)
Negative Undetermined 70.3
Reagent Blank Undetermined 69.2
HV1 (61 bp) 32.72 76.8
HV1 (61 bp) 34.16 76.8
Positive Control 20.71 75.2
Positive Control“ 20.66 75.6

 

 

 

 

(t) Burial 4015 Femur

Negative Undetermined 80.4

 

 

 

 

 

 

 

 

Reagent Blank 34.00 78.1
HV1 (61 bp) 35.46 76.5
HV1 (61 bp) 33.95 76.8

Positive Control 22.16 74.4
Positive Control“ 21.73 76.5

 

79

 

 

 

Table 19 (a - e): Ct Values and Tm for IS6501 Positive Reactions

Ct-values and melting temperature (Tm) data obtained during the amplification of IS6501
from burials 2722 rib, 4015 rib, and 4015 vertebra. Lower Tm values were seen in some
negative controls and reagent blanks as a result of non-specific amplicons/primer dimers.
All DNA extracts were prepared in August of 2006 except those indicated which were
prepared in May of 2006 (* = duplicate reactions). Reactions that did have any
detectable amplification are indicated by Undetermined.

(a) Burial 2722 Rib* (DNA extracted in May of 2006)

Negative 37.76 69.2

 

 

 

 

 

 

 

Reagent Blank 36.34 69.2
IS6501 34.74 78.7
IS6501* 36.32 79.0
Positive Control 20.19 81.7
Positive Control“ 2024 31.7

 

 

 

(b) Burial 2722 Rib and Rib (1:10)

N 've
Blank
IS6501
IS6501*

IS6501 1:10
IS6501* 1:10
Positive Control
Positive Control*

Undetermined
Undetermined

31.65
32.57
30.93
21.17
18.77
18.41

(c) Burial 4015 Rib* (DNA extracted in May of 2006)

  

l’(’R Reaction

   

(‘t Value

     

 

   

'l‘m (0C)

 

 

 

 

 

 

 

Negative 37.76 69.2
Reagent Blank 36.34 69.2
IS6501 34.8 79.9
IS6501* 37.03 78.7
Positive Control 20.19 81.7
Positive Control* 20.24 81.7

 

 

 

80

(d) Burial 4015 Neural Arch

 

 

 

 

 

 

|’('R Reaction (‘t Value 'l‘m ("(‘)
Negative 35.62 77.2
Reagent Blank 38.2 78.4
IS6501 34.54 81.2
IS6501* 31.31 81.2
Positive Control 22.00 81.2
Positive Control” 22.12 81.2

 

 

 

 

(e) Burial 4015 Rib* (DNA extracted in May of 2006)

 

 

 

 

 

 

l’(‘R Reaction (‘t Value
Negative 37.76 69.2
Reagent Blank 36.34 69.2
IS6501 34.74 78.7
IS6501* 36.32 79.0
Positive Control 20.19 81.7
Positive Control“ 20.24 81.7

 

 

 

 

81

 

 

Table 20 (a — b): Ct Values and Tm for Bcsp3] Positive Reactions

Ct-values and melting temperature (Tm) data obtained during the amplification of
Bcsp31 from burial 4015 rib and vertebra. Lower Tm values were seen in some negative
controls and reagent blanks as a result of non-specific amplicons/primer dimers. All
DNA extracts were prepared in August of 2006 except those indicated which were
prepared in May of 2006 (* = duplicate reactions). Reactions that did have any
detectable amplification are indicated by Undetermined.

(a) Burial 4015 Vertebra Neural Arch

N tive 35.62
Blank 38.20
31 34.52

1 * Undetermined
Positive Control 19.90
Positive Control* 18.84

 

(b) Burial 4015 Rib (DNA extracted in May of 2006)

Ne 've 32.57
Blank 33.08
31 31.45

1 * 3 1 .84
Positive Control 24.06
Positive Control“ 22.39

 

82

 

DISCUSSION

Previous research conducted in our laboratory has shown how the analysis of
mtDNA from ancient skeletal remains can provide information that would otherwise be
unobtainable. Clemmer (2005), Rennick (2005) and Murray’s (2006) molecular studies
on remains from a tumulus located at Kamenica, in present day Albania, allowed for a
better understanding of relatedness, burial patterns, migration, and overall genetic
diversity of the region. The research presented here demonstrates an additional way in
which DNA analysis can provide information about an individual and their society.

Diseases have been an intricate part of human history since ancient times.
Understanding the interaction between ancient people and pathogens is not only of
cultural interest, but may also shed light on how human response to epidemics has
evolved over time. The analysis of diseases in a culture is traditionally accomplished
through the visual examination of remains. Anthropologists include descriptions of
pathologies in biological profiles when examining both ancient and modern skeletons.
Skeletal pathologies provide insight into the lives of an individual (such as diet, health,
medical practices, and possible cause of death) and in a modern forensic science context
can aid in human identification based on comparative medical history. Paleopathologists
examine the distribution of these lesions in skeletal populations to better understand how
epidemics may have affected the society as a whole. The knowledge gained from
osteological analysis is limited because skeletal lesions produced by two different disease
processes can share striking similarities. As a result, educated guesses on the organisms

that could be responsible are made without ever identifying the exact causative agent.

83

Relying on diagnostic skeletal lesions alone may lead to the misidentification of a
disease process, an inaccurate description of the skeletal remains, or a gross
underestimate of an endemic pathogen. Ancient pathogen DNA analysis can be used to
verify skeletal lesions or abnormalities as originating from a specific organism. This can
aid in the development of criteria for the identification of pathogens in skeletal material
and provide a better understanding of bone tissues response to infection. Furthermore,
the combination of traditional osteological methods and pathogen DNA analysis provides
a powerful tool to identify the specific organism infecting an individual and provide a

better understanding of the disease’s distribution throughout a culture.

Bone Preparation for DNA Extraction

In this study, the preparation of bone samples varied greatly based on bone type,
size, and condition. Bones with a high percentage of cortical material, such as long
bones, were the easiest to process as they contained a small amount of soil that could be
removed by rigorous cleaning. In contrast, bones with a higher percentage of cancellous
material, such as vertebrae, were the most difficult to process. The woven nature of the
bone provided a porous structure that trapped dirt and debris. Several different methods
were attempted to effectively clean cancellous bone. Swabbing the surface proved
difficult because small pieces of cotton became caught on the material. Scrapping of the
bone with a sterile spatula only resulted in breakage. An alternative method was tried in
which the dirt was flushed out of the crevices. One milliliter of digestion buffer was

initially used for this process, but led to the direct disintegration of a large portion of the

84

bone. Using a wash with a lower concentration of EDTA (10 mM instead of 50 mM)
prevented digestion during this step. EDTA is a chelating agent that binds divalent
cations, such as the calcium found in bone, and previous studies have shown this
chemical’s ability to decalcify and break down a bone sample (Sarsfield 2000). The poor
quality of the cancellous bone matrix may have allowed it to be completely degraded by
the digestion buffer. Flushing of the bone with the alternate wash, followed by gentle
swabbing, proved to be the most effective method for cleaning this material prior to
drilling.

Drilling cancellous bone also proved to be more difficult than cortical bones.
Many more holes were required to generate the 25 — 50 mg of powder necessary for DNA
extraction. Sections of cancellous bone also had a higher tendency to break when drilled.
Reducing the speed of the Dremel tool and holding the drill bit steady, while stabilizing
the bone with forceps, prevented excess fragmentation. Small and brittle bones also had
to be handled differently. For example, rib material fragmented or broke in half when
drilled at a ninety-degree angle. In this case it was found that holding the drill almost
parallel to the bone, while dragging the bit along the surface of the bone, prevented
excess destruction of the material. Other samples, such as those from the Voegtly
cemetery, were impossible to drill. In these instances the bone surface was washed
extensively, small pieces of bone were broken off with forceps and crushed in digestion

buffer.

85

DNA Extraction and PCR

Extraction of DNA from digested bone powder often required the removal of
water-soluble components that produced a reddish brown or black color. Performing
additional phenol extractions removed a portion of the coloration, yet also led to the loss
of some of the DNA template. Pushing the sample through a Microcon also reduced
pigmentation of the extract. However, the aqueous layer sometimes left a small amount
of residual debris on the column membrane. This debris would clog or tear the
membrane if the samples were centrifuged at the maximum speed allowed. Lowering the
speed of centrifugation, from 14,000 X g to 10,000 X g prevented damage to the
Microcon as well as loss of the DNA samples in the filtrate. A maximum of two
additional phenol extractions and three TE washes on the Microcon column removed the
highest amount of coloration from the sample without sacrificing DNA template yield.

PCR inhibition was seen with many of the DNA extracts that retained coloration.
O’Rourke et al. (2000) found that a large amount of soil-derived PCR inhibitors co-
extract with DNA, including tannins, humic acids, and fulvic acids. These inhibitors
usually bind directly to Taq polymerase, lowering the success of, or preventing, the
amplification of the template DNA. Wilson (1997) reported that diluting or purifying
DNA extracts and the addition of a compound with a greater affinity for the inhibitor than
the polymerase are the most effective methods to overcome inhibition. The utility of
these three methods was demonstrated in this and previous studies (Halvorson 2005,
Rennick 2005, Murray 2006). All burials in this study required at least a 1:10 dilution of

the DNA extracts in order to amplify successfully. Some extracts (2722 vertebra, 5010

86

 

unfathaqr Pi L .. ‘1.”

 

vertebra, and 213 ischium) retained a dark brown to black color and needed to be diluted
1:20. Diluting DNA extracts, while lowering the amount of starting template,
simultaneously decreases the amount of inhibitors present. If inhibition can be overcome
while retaining enough starting DNA, then PCR can still be successful. BSA, which is
believed to compete with Taq DNA polymerase for the binding of inhibitors, was also
added to each reaction. Finally, DNA was concentrated and purified by the use of
Microcon spin columns and TE washes. Successful amplification of ancient DNA using
these methods demonstrated their ability to reduce PCR inhibition associated with

skeletal material.

Contamination and Ancient DNA Analysis

Ancient DNA analysis requires a high level of sterility to prevent contamination
by modern DNA. Contamination can lead to erroneous results, complicate future
molecular work, and potentially impair DNA analysis from valuable skeletal material.
Wearing personal protective equipment, sterilizing supplies, working under a UV
sterilized hood, using different labs for extraction and amplification, and separating
positive control DNAs from samples, were all used to prevent contamination in this
study. Despite these stringent methods, positive reagent blanks and negative controls for
mtDNA were obtained in some experiments.

Contamination usually originated from the analyst or one of the bone samples. If
the analyst’s sequence was obtained then new PCR reagents were used and the

experiment was repeated. PCR reactions were also repeated if the reagent blank or

87

negative control contained the same sequence as one of the bone samples. In this case
DNA was re-extracted from the bone to access the validity of the sequences and ensure
that they did not arise from exogenous DNA. Determining the source of contamination
not only allowed for the development of methods to prevent it in future experiments but
also served as a means of quality control.

Pathogen DNA screening experiments required a few extra precautions. Taylor et
al. (1999) recommended that stringent measures be taken including using separate
laboratories for DNA extraction and PCR that had not previously been used for
microbiology or pathogen DNA work. In this study, 186110, OxyR, and Mtp40 primer
optimization and sensitivity assays were conducted in a separate laboratory from bone
experiments. Initially these experiments were conducted using a relatively high
concentration of bacterial DNA directly from the stock solution. This led to
contamination of all reactions following nested PCR. Contamination probably resulted
from the analyst handling the reactions, pipetting, or aerosol (particles of liquid, which
may contain minute concentrations of DNA, that are passed through the air from
pipetting, opening tubes, etc.). Using a 1:1000 dilution of the bacterial stock as the
highest concentration reduced incidences of contamination. Additional safeguards were
also implemented such as changing gloves when switching labs, storing stock genomic
bacterial DNA in a separate box from dilutions, and ensuring that reactions with the
highest concentration of DNA were handled last.

Contamination seen in the optimization experiments also set forth policies in the
handling of bone samples. The majority of optimization experiments were completed

before the Albanian bone samples were received in the laboratory. In addition, bone

88

L
3 0“,

 

 

screening and bacterial DNA experiments were never conducted on the same day.
Taking these precautions ensured that bacterial DNA samples and questioned bone
samples were kept isolated from one another during the entire screening process. Taylor
et al. (1999) suggested that positive controls should be used sparingly (i.e. not utilized in
every PCR experiment) when conducting ancient DNA analysis. A problem arises that
without using positive controls the validity of the results become questionable. An
alternative method utilized by Zink et al. (2005) was the use of a positive bone sample
rather then bacterial DNA. This was attempted here with the Voegtly material but the
irreproducibility of amplification was problematic. Therefore, a better method for future
ancient pathogen DNA studies would be to use very low concentrations of bacterial DNA
as positive controls, while practicing methods that prevent and monitor levels of

contamination.

Human mtDNA Analysis to Assess the Quality of DNA in Skeletal Remains

Previous studies have shown that human mtDNA can be extracted and amplified
from degraded biological sources including skeletal remains, fingernails, and hairs
(O’Rourke et al. 2000). F oran (2006) compared the relative degradation of nuclear DNA
and mtDNA over time. His results indicated that DNA preservation could be affected by
multiple factors, including copy number, chromatin structure, cellular location, and
transcriptional activity. MtDNA is retrievable from degraded biological sources because
it has a high copy number and may be protected by both the cellular and mitochondrial

membrane. Successful extraction and amplification of mtDNA provides evidence that

89

pathogen DNA, if present, has the potential to be recovered and amplified from the
skeletal material.

MtDNA was isolated from vertebra, rib, and long bone samples from each
skeleton. Edson et al. (2004) reported that bone type directly influences the success rate
of mtDNA extraction, amplification, and sequencing. The authors reviewed casework
conducted by AF DIL. PCR of DNA template extracted from long bones had a success
rate (measured by the ability to amplify and sequence the control region) of 94.79%,
while rib DNA had a rate of 96.15% and vertebrae DNA a rate of 85.71%. Similar
results were found in this study with template from long bones requiring fewer PCR
cycles to amplify and producing cleaner sequence than DNA from ribs and vertebrae.
This indicates that either the femur samples were less physically degraded, had a higher
copy number of preserved DNA template, or had less PCR inhibition.

The rib DNA amplification success rate was much lower in this study than those
reported by the AF DlL. There are several possible reasons for this discrepancy. First,
the Albanian samples tested were much older (the 13m century is the latest time period)
than those typically studied by the AF DIL (World War II to present). Also, ribs tend to
be very fragile and degrade more rapidly than long bones. Second, the AFDIL examines
remains from all over the world while the Albanian samples were located in a relatively
closed environment. Weather, temperature, soil conditions, bacteria, and moisture could
all have affected DNA quality and quantity. Finally, the sample size of the AF DIL study
is immensely larger than that of the Albanian bones (1021 vs. 17). The analysis of more
long bones, ribs, or vertebrae fiom Albania could reveal that these bones have similar

success rates as those reported by the AF DIL. However, previous research conducted in

90

[\l

30'.

 

 

 

our laboratory on Albanian skeletal material has shown success rates that were lower than
those produced by AFDIL (Rennick 2005, Murray 2006). Differences in rates are
therefore most likely caused by variation in the condition and age of the material.

PCR reactions for mtDNA HV1 were conducted until an amplicon that could be
sequenced was attained from a long bone, vertebra, and rib. Sequences were obtained
from these bones for all burials with the exception of the vertebra from burial 4015. This
bone contained a large amount of irremovable debris and soil. MtDNA PCR products
obtained from this sample occasionally appeared smeared or light on an agarose gel. The
DNA in the vertebra may have been degraded or extracts may have contained a high
concentration of inhibitors. The presence of inhibitors is supported by the fact that
extracts from this sample were black in color. Furthermore, a dilution of the DNA
extract did allow for amplification but sequencing reactions still failed. In general, bone
DNA extracts that were light brown, red, or clear amplified successfully using a 1:10
dilution. In instances where the extracts were dark brown or black the sample had to be
diluted further (in this study a 1:20 or higher dilution was required). Darker reactions
seemed to contain a higher concentration of inhibitors, which needed to be lowered for
the Taq polymerase to work efficiently.

Following successful amplification, HV1 sequences obtained from the vertebra,
rib, and long bone were aligned and their polymorphisms compared. The analysis of
human DNA when conducting pathogen-screening studies is essential. Successful
amplification of mtDNA is a good indicator that, if present, the pathogen DNA should be
detectable, as both the human and pathogen DNA were subjected to the same

environmental conditions over a similar period of time. In addition, human mtDNA is

91

expected to be in the remains, unlike pathogen DNA whose presence is uncertain.
Human mtDNA amplification serves as a means to assess whether or not the samples are
inhibited or if the DNA has completely degraded. Sequencing of mtDNA products
ensures that the amplicon obtained originated from the skeletal material and helps rule

out contamination by exogenous DNA.

Screening for Tuberculosis in Voegtly Cemetery Material using IS61 10

The screening of DNA extracts for the IS6110 insertion element provided an
effective way to detect members of the MTB complex. Methods developed in this study
were first tested on Voegtly cemetery burials 32 and 629. Ubelaker et al. (2003) reported
success at identifying MTB complex DNA in bone samples from both of these
individuals. Only DNA extracts from burial 32 gave a positive product while those of
629 did not produce an amplicon after several attempts. Initially, DNA extracted from
the Voegtly material was to be used as a positive control for the screening of ancient
material. Irreproducibility of results made the use of these bones as positive controls
impossible.

Ubelaker et al. (2003) had similar difficulties and only obtained consistent
amplification following whole genome amplification (primer extension pre-amplification
(PEP)) of the skeletal DNA extract. In this method PCR reactions are conducted with
random oligonucleotide primers to amplify all DNA present in a sample, prior to
conducting reactions that target a specific region. PEP was not utilized in this study

because previous work has shown low success rates of whole genome amplification on

92

o 0'.

F131
1.
c. E
‘ I
I'.
'0
b

 

LL:

 

 

DNA from Albanian skeletal material (Barber and F oran 2006). PEP is not species-
specific and amplifies all DNA present in the reaction. This method could be beneficial
when the amount of pathogen DNA template is at a low concentration (i.e. acute
infections or highly degraded material), however PEP will also increase even minute
amounts of exogenous DNA or contamination in the extract. The risk of contamination
leading to erroneous results is therefore magnified when PEP is utilized. Care should be
taken when employing this method in molecular pathogen screening experiments because

the validity of results is highly questionable.

Screening for Tuberculosis in the Voegtly Cemetery Material using OxyR and Mtp40

Burial 32 rib DNA extracts were also screened for the presence of the OxyR and
Mtp40 genes. Despite repeated attempts, all PCR reactions containing 32 rib template
were negative. Mays et al. (2001 and 2002) had similar results in their molecular assays
of medieval remains from the Wharram Percy collection. In the initial study, DNA was
extracted from the inside of lesions located on bones (mostly vertebrae) of nine
individuals. Seven of these tested positive for IS6110 and Mtp40, while only 6 tested
positive for OxyR. Rib DNA from each of these individuals was also screened, resulting
in one testing positive for IS6110 and none for the OxyR gene.

Amplification of IS61 10 may have been more successful because the genomes of
the MTB complex members contain 5 — 20 copies of this insertion element and only
single copies of the OxyR and Mtp40 genes. PCR assays that utilized primer sets for the

IS6110 element were able to detect a lower copy number of the bacterial genome than

93

those designed for Mtp40 and OxyR (detection level of 1 vs. 100 copies). Mays et al.
(2002) also found that assays for IS6110 were more sensitive than those designed to
target OxyR or Mtp40 sets (3 vs. 300 copies). The higher copy number of the 186110
insertion element appears to directly affect the sensitivity and ability to detect the
pathogen DNA. Future molecular work on analyzing pathogen DNA from ancient
remains would benefit from designing PCR assays that target regions of the genome that
are multi-copy.

The inability to amplify the OxyR and Mtp40 genes from the IS6110 positive
Voegtly material prevented the identification of which species of the MTB complex
material was present. However, the methods developed here may aide in the
differentiation of M. tuberculosis and M. bovis in future studies of ancient remains. The
SNP present in the OxyR gene is used to classify a member of the MTB complex as M.
bovis (Baker 2004). Traditionally, sections of the gene are amplified and sequenced in
order to determine what base is present at the species-specific polymorphic site (Baker
2004). Sequencing of this gene has also been used in studies of ancient skeletal material
(Donoghue 1998, Mays et al. 2001, Mays et al. 2002). DNA sequencing requires that a
rather large (150 — 200 bp minimum) piece of DNA be extracted and amplified from the
samples. Obtaining and successfully sequencing this size fragment of DNA could be
difficult in highly degraded remains. Single nucleotide extension assays are more
effective because the size of the amplicon is smaller (often 40 — 60 bp), the assay is very
rapid, and base determination of a single site can be conducted without scrutinizing a

large amount of sequence. SNP assays can also be multiplexed so a number of species-

94

specific polymorphic sites are assayed and a profile for the identification of a particular
organism produced.

A semi-nested Mtp40 PCR method was also developed in this study. Previous
researchers (Mays et al. 2001, Fletcher et al. 2003) used primer sets that produced a
relatively small amplicon (152 bases) and attempted to increase detection limits by using
43 cycles of PCR rather than 35. Mays et al. (2001) reported a minimum detection level
of 300 copies using this approach. Primers developed by Mays el al. (2001) were used as
the external set for a semi-nested assay designed here. Semi-nested PCR assays
improved the sensitivity allowing for the detection of 15 copies. Based on these results
semi-nested PCR of the Mtp40 gene may allow for the identification of M. tuberculosis
in ancient material with a higher efficiency. Further testing would need to be conducted

since the Voegtly bone still failed to yield a product for Mtp40 using this method.

Screening for Tuberculosis in Butrint and Diaporit Skeletal Material

Several sets of skeletal remains from Butrint and Diaporit contained lesions.
Similar vertebral and rib lesions were identified on two sets of remains located at the
Butrint sites 2722 and 4015. Burial 2722 was from the 11th — 13th century AD and
contained the skeletal remains of an adolescent male. This burial was excavated from the
Baptistery and the bones showed multiple lesions on the vertebrae and ribs. The second
set of skeletal remains, burial 4015, was excavated from the Merchant’s house, and
contained an adolescent male from the 11‘11 — 13th century. One other burial from the

Butrint site, 5010, and two others from Diaporit, 213 and 319, showed pathologies that

95

 

 

were not entirely consistent with 2722 and 4015. While these burials had porosity of the
ribs, long bones, and vertebrae they also contained distinctive skeletal lesions. Examples
of distinctive pathologies on these remains include lesions on the parietals of the skull
(319 and 5010) and bony projections/ growths on the vertebrae (213). These pathologies
were not consistent among the five pathological skeletons (213, 319, 2722, 4015 and
5010) utilized in this study. However, Fenton (2006) hypothesized that tuberculosis or
brucellosis could still produce these types of lesions and abnormalities based on gross
analyses of the remains and comparison to previous studies.

The vertebral and rib lesions seen on the Albanian remains were consistent with
the lesions described by Mays et al. (2001 and 2002) in their study of the Wharrum Percy
collection The causative agent behind these lesions was determined to be M.
tuberculosis. This led to the assumption that tuberculosis was also the causative agent
behind the lesions noted on the Butrint and Diaporit bones. However, none of the
Albanian bones screened gave a positive result for the MTB complex DNA afier a total
of 330 assays.

Negative PCR results for pathogen screening assays are difficult to interpret
because they can indicate either that the person was never infected or that the pathogen
DNA was present but is no longer detectable. Additional experiments were used to
assess whether or not the bones were negative for the pathogen. MtDNA analysis
confirmed that DNA was preserved in the Albanian bone samples. Primer sensitivity and
spiking assays showed that one copy of the tuberculosis genome could be identified, even
with a large background of human DNA. Based on these experiments, even if only one

copy of the M. tuberculosis or M. bovis genome remained in the Albanian bones then it

96

should have been detected. Experiments such as these allow the analyst to claim that the

individuals tested were negative for infection with more confidence.

Assessing Bones for Pathogen DNA using Real-Time PCR

Following the negative results obtained using traditional PCR; the decision was
made to try real-time PCR methodologies. In traditional PCR, amplified product is
electrophoresed on an agarose gel and visualized by staining. If the concentration of the
resulting amplicon is too low, no band will be seen and the results will be interpreted as
negative. The real-time system allows the amount of PCR product to be monitored after
each cycle. This is beneficial because even a minute amount of amplification can be
detected. In addition, highly degraded DNA template can be amplified because the
products produced are generally very small (~60 bp). Real-time PCR seems to be a better
platform for screening for pathogen DNA because it can address the issues of low copy
number and degradation.

Real-time PCR of human mtDNA was conducted on bone samples to once again
assess preservation, inhibition, and to determine PCR conditions for reactions containing
bone DNA extract (i.e. how much template should be added, the ideal final volume of
reactions, etc). Real-time PCR amplification of human mtDNA was first conducted using
primers designed by Gehring (2004). Amplification was not seen using this primer pair,
perhaps because the targeted region was too large (118 bp) for the degraded template.
The decision was made to use a set of primers that target a smaller region of HV1 (6lbp),

and successful amplification was obtained for 8 out of 19 bone extracts (Voegtly 32 rib;

97

L

30‘

 

 

 

 

Albanian 319 vertebrae, 2722 rib, 4015 vertebrae, rib, and femur). These numbers are
lower than expected when compared to nested PCR results.

One explanation for this is that the mtDNA primers were not designed using
Primer Express software. This software is manufactured by Applied Biosystems to
design optimal primer sets for use on the ABI Real-time PCR Systems. As a result, the
melting temperatures (Tm) of the amplicons were lower than what is traditionally ideal
for real-time assays (~70°C vs.~80°C). This made it difficult to differentiate between the
actual mtDNA product and non-specific products/primer dimers on the dissociation
curve. An additional problem could have been that the designed primers targeted regions
within HV1. Primer sets for HV1 are typically designed to target conserved stretches of
DNA, because people tend to have a high level of variation outside of these regions. It is
possible that the mtDNA primers used in the real-time assay targeted a variable rather
than conserved region. If an individual’s DNA contained polymorphisms at the primer-
binding site then amplification would not occur. A better method would be to target a
gene whose sequence is conserved among all humans’ mtDNA genome (e.g., cytochrome
B). These would allow the assessment of mtDNA preservation and amplification with
greater efficiency.

To target the causative agents of brucellosis and tuberculosis additional real-time
PCR reactions were performed. Real-time assays that screened for the MTB complex
were designed to target IS6110, OxyR and Mtp40 genes. Only Voegtly burial 32 rib
gave positive amplification of IS6110; all Albanian samples tested negative. The
Voegtly material produced a similar dissociation curve and Tm as the tuberculosis

positive control DNA. These results confirmed the findings that individual 32 was

98

 

positive for 186110, previously published by Ubelaker et al. (2003) and presented in this
study. They also demonstrate real-time PCR’s capability of detecting pathogen DNA and
identifying tuberculosis infection. More importantly, they provide further evidence that
the lesions found on the Butrint and Diaporit skeletal material did not originate from a
member of the MTB complex.

Consequently, the decision was then made to screen bones for the causative
agents of brucellosis, since this disease produces similar lesions as skeletal tuberculosis.
Two individuals from the Butrint site, burials 2722 and 4015, tested positive for
brucellosis. Both of these individuals were adolescent males whose remains dated to the
11th —— 13‘h century AD. Pathologies present on their vertebrae and ribs were very similar
to one another (Figure 5). DNA extractions from bone samples produced real-time PCR
products that had similar dissociation curves and melting temperatures as control DNA
from B. abortus. Two genes were targeted in the real-time assays, 186501 and Bscp3 l ,
with the former having a higher success rate (i.e. successful amplification).

Three rib extracts (prepared May of 2006; August of 2006, and a 1:10 dilution of
an extract prepared in August of 2006) from burial 2722 were positive for [$6507. The
products produced from the samples had an average Tm of 806°C compared to the
positive controls Tm of 81.5. Amplification of IS6501 was also seen for two rib extracts
(May of 2006 and August of 2006) and one vertebrae extract from burial 4015. A single
rib and vertebra extract from burial 4015 produced positive results for Bscp3 l. The
greater success rate of IS6501 amplification compared to Bscp3l could be because
IS6501 is an insertion element and present in the genome 5 — 35 times (Ouahrani 1993).

Results in this study show that both the MTB complex 186110 and the Brucella spp.

99

 

IS6501 had greater success than assays for single copy genes. This provides further
evidence that future assays designed to screen for pathogens would benefit from
designing primers that target multi-copy elements.

Similarity of the melting temperatures between the bone extracts and B. abortus
positive controls indicates that the same amplicon was most likely produced. These
results were supported using an agarose gel, which showed that the amplicons were the
same size. Both IS6501 and Bcsp3] are specific to the three members of Brucella spp.
that are pathogenic. The amplification of these products from bone indicates that these
two individuals likely had brucellosis during their lifetime. As a result Brucella spp is
the probable causative agent behind the skeletal lesions. Osteological examination
showed that the lesions of the two individuals were extremely similar and potentially
caused by the same disease process. The molecular confirmation of the same pathogen in

bone samples from these individuals solidifies these claims.

Tuberculosis, Brucellosis, and Albania

Both tuberculosis and brucellosis are common diseases found in modern day
Albania. A study published by the WHO (2004) reported that 53 out of every 100,000
Albanians were positive for tuberculosis. A similar number of brucellosis cases were
reported in 2002 with 40 out of every 100,000 Albanians having active infection. This
places Brucella spp. infection as the number one food born illness in modern day Albania
(WHO 2002). Brucellosis is such a health risk to both humans and livestock that

countries worldwide, including Albania, have developed surveillance systems for the

100

disease. In Albania, the Minisz of Health and National Institute of Public Health
require the reporting of any human positive cases, while animal infections are monitored
by the Institute of Veterinary Medicine.

lncidences of these diseases in modern Albania, and the rest of Europe, have led
many scientists to believe that both tuberculosis and brucellosis were present throughout
the ancient Mediterranean basin. An osteological study conducted by Capasso (1999)
found skeletal lesions that appeared to be consistent with brucellosis in 17.4% of adult
skeletons from the town of Herculaneum. These individuals died during the eruption of
Mount Vesuvius in 44 AD. Brucella spp. was proposed to be the causative agent behind
the skeletal lesions based on gross and X-ray examination. This hypothesis was made
due to similarity of the lesions to modern cases of skeletal brucellosis. Capasso (2002)
also examined remnants of bacteria in preserved dairy products located at the Helcaneum
site using scanning electron microscopy. The author identified several cocci-bacteria that
had the appearance, size, and shape of members of Brucella spp. While these findings
were consistent with brucellosis, other bacteria or pathogens might have been responsible
for the skeletal damage. Capasso studies (1999 and 2002) confirmed that the ancient
people in the Mediterranean experienced disease processes that could be consistent with
tuberculosis or brucellosis. However, no molecular analysis has been performed on these
remains to identify the causative agent.

The molecular data presented in this study provides concrete evidence that
brucellosis was present in the ancient Mediterranean basin through the analysis of a
subset of human skeletons from Butrint, Albania. Brucellosis was identified in the

remains of two male adolescents from the 11th — 13th century AD. Both sets of remains

101

(Burials 2722 and 4015) displayed almost identical skeletal pathologies on their
vertebrae. Similarities between these two individuals’ age, the time period of the
remains, and manifestation of the disease allow for several hypotheses regarding the
history of the site as well as the health and life of its inhabitants.

The most common method of Brucellosis transmission is the ingestion of infected
milk or meat from livestock (CDC 2005). A lack of pasteurization or sanitization of
animal products would transmit the disease to humans. Younger individuals are more
susceptible to the disease than those who acquire it later in their lifetime. Geyik et al.
(2002) conducted a study in which 283 human brucellosis cases from a hospital in
southeast Turkey were reviewed. The authors identified 195 individuals with
musculoskeletal involvement and compared the relative ages of the patients. Out of these
195 individuals 20% were under the age of 15 years old while 63% were between the
ages of 15 and 45 years old. The two individuals that tested positive in this study are
close to the two age ranges were skeletal infection would be frequent. The possibility
exists that the two males were fed infected meat or milk as young children and acquired
an infection that progressed until their death in their early teens. F enton (2006) stated
that several other individuals were found at the site that exhibited skeletal pathologies
similar to those seen in burials 2722 and 4015, during collections conducted in the
summer of 2006. These individuals showed signs of minor vertebral lesions and possible
healing indicating either acute infection or the ability to sustain infection more readily.
Bone growth supports Brucella spp. infection since tuberculosis generally inhibits the
process to a greater extent (Ortner and Putschar 2001). This provides support for the

hypothesis that the inhabitants of the ancient site were acquiring Brucella spp. from

102

consuming livestock products. This mode of transmission continues to be common in
present day Albania with the majority of brucellosis cases originating from the ingestion
of infected meat or dairy in rural communities (WHO 2002).

Another possibility is that the disease was obtained from being in close contact
with infected livestock. According to the Center for Disease Control and Prevention
(CDC 2005) brucellosis can also be transmitted in rare cases by the handling of infected
livestock or the bacterium. The CDC reports that there is a higher susceptibility to
brucellosis among individuals who work with animals in general. In 1998 Wallach et al.
examined an outbreak of human Brucellosis among farm workers in Argentina.
According to the authors the workers had been employed on a farm where roughly 500
out of 2200 goats showed severe signs of Brucella spp. infection The causative agent
was later identified as B. meltinsis and 60 workers were tested for the pathogen. Thirty-
three out of sixty workers tested positive with twenty-one of these individuals having
close contact with the infected herd, 11 having occasional contact and one having no
contact. Studies such as these suggest that individuals employed in animal care such as
veterinarians, farmers, animal handlers and laboratory workers have a higher rise of
contracting the disease via close contact.

Livestock rearing and milk production are traditional economic activities in
Albanian societies and even today accounts for 47% of agricultural production (Shundi
2004). Individuals involved in this field, including young males, are often required to
live in close proximity of the herd and care for sick livestock. The fact that both infected
individuals were adolescents could indicate that younger members of the culture worked

more closely with animals in the raising, herding or care of livestock. The two young

103

males may have been actively involved in livestock care and as a result acquired infection
due to their greater exposure to sick livestock compared to other members of the society.

A number of claims can also be made regarding the history of the site based on
this molecular data. The human skeletons from Butrint and Diaporit that display lesions
consistent with brucellosis come from two main time periods: the 5th — 7’h Century AD
and the 11th — 13th Century AD. Both of these periods can be characterized as having
harsh living conditions due to the collapse of the Roman infiastructure in the late 6th
century. As a result, from the 7th — 11‘h Century AD the region experienced a downfall in
public health. In addition, the area was exposed to outside populations by the invasions
of Slavic, Byzantine and Venetian cultures. A decrease in overall health and the
introduction of new people may have resulted in the spread of diseases. This could
include brucellosis which was molecularly confirmed in this study to be present during
the 11th — 13th Century AD

The age of these remains (11th - 13th Century AD) also corresponds to the period
directly before the city flooded and was ultimately abandoned. During that time the city
was in ruins, the site had been subjected to centuries of war, and waters from the
surrounding lakes and marshes were beginning to rise. The inhabitants of the city were
mostly ‘squatters’ living in and around the ruins of the Triconch palace. It is believed
that these individuals were in very poor health. Anthropological analyses of skeletal
remains from the period indicate a high child mortality rate, incidences of disease, and
population wide anemia and malnutrition caused by harsh living conditions (Butrint
Foundation Annual Report 2004). The skeletal remains of burials 2722 and 4015 are a

representation of such conditions due to the vertebral lesions and their young age at

104

death. The molecular identification of the causative agent centuries later would also
suggest a high pathogen load. In addition, the confirmation of brucellosis suggests that
the population was experiencing diseases and possibly epidemics at this time period.
Malnutrition and anemia would have only accelerated the spread of pathogens and
produced an increase in child mortality. Diseases such as brucellosis may have played a
direct role in the downfall of the society as the waters rose and the city was finally
abandoned.

The research presented here has provided novel and interesting information on
remains from a World Heritage site. The identification of brucellosis in ancient Butrint
contributes directly to understanding the inhabitants’ health and lifestyles. Plans to
conduct further anthropological and molecular research are being developed. In the
summer of 2006, Dr. Todd Fenton and graduate students identified additional individuals
that contained signs of skeletal pathologies. Molecular examination of these skeletal
remains may lead to the identification of more individuals with brucellosis or possibly the
identification of different pathogens. Additional research could also be conducted to
identify which pathogenic Brucella spp. is present in the remains. According to Omar
and Putschar (2001) B. melitensis causes a greater level of bone erosion and damage than
other Brucella spp. It is also believed that this species is common among goat
populations in the Mediterranean and as a result is responsible for human brucellosis
throughout the area. The molecular confirmation of B. melitensis in the Albanian
samples would solidify these claims and provide novel information regarding both the

inhabitants of Butrint and the disease process as a whole.

105

To the best of the researcher’s knowledge this is the first time Brucella spp. has
been identified in ancient skeletal remains. The results show how beneficial molecular
identification of pathogens is to anthropological and paleopathological studies. Scientists
now have the ability to identify the causative agents behind skeletal lesions. As a result,
the research provides a vast amount of knowledge about the inhabitants and history of a
World Heritage site. This research proved that brucellosis, as speculated, was present in
the Mediterranean basin and afflicted ancient societies. It also provided knowledge on
the manifestation of Brucella spp. infection in bone. More importantly, it indicates that
brucellosis arrived to Butrint as the city was beginning to decay and may have been
directly involved in its abandonment and the poor health of the inhabitants. This proves
that the molecular identification of pathogens can be beneficial in understanding the lives
and histories of ancient societies. Anthropologists and paleopathologists now have a new
tool and ability to better understand the diseases and epidemics that have plagued

mankind and ancient cultures worldwide.

106

 

Agigwé 05838:. 0:30:80 .0:_:0:u-<v 0:0. 8:: :0 800.08 000: 0:: 8:00:05 2052 a: 00:0: 0:0 08:20::5 60:58 :0: 0.03 00:00:00
"0:23: .85 00:00:08 00:0:0:0: 8:3: :0 :0 830:5. 0:: 0: “00:09:00 @0008 0:0: 0:80:00 03: 8:00:00." 0.: 085% 0:0: :000 :8: 08:00:00 _ >E

00—95% 28: 12.2.5.5 E0...: 80:0:50m SE "3 030,—.

< XHQZm—mnjx

107

 

3:

.A0:_:0:w-0 05:02:? 05028-0 05:00.35 0:0 :0:: :0 80008 000: 0:: m::00_::_ 20:2 x: :08: 0:0 80:08::5 005050 :0: 003 00:00:08
"803.0: EEO 00:03:00 00:08.:0: ammo: :0 :0 880050. 0:: 0: 8:00:80 8:00:08 0:0: 0:08:00 03: 00:00:05" L 08800 0:0: :000 80.: 80:00:08 : >:

00—95% 23— 10:33::— E0._.: 000:0:00m Fr: Aug—:88: mm 030,—.

< Van—293:0.

108

BIBLIOGRAPHY

Anderson, 5., AT. Bankier, B.G. Barrel], M.H.L. deBruijn, A.R. Coulson, J. Drouin, I.C.
Eperon, D.P. Nierlich, B.A. Rose, F. Sanger, R.H. Schreier, A.J.H. Smith, R.
Staden, and LG. Young. 1981. Sequence and organization of the human
mitochondrial genome. Nature. 290:457 — 465.

Barber, A and D. Foran. 2006. The Utility of Whole Genome Amplification for Typing
Compromised Forensic Samples. Journal of Forensic Sciences. 51: 1344 — 1349.

Butler, J .M. 2005. Forensic DNA Typing: biology, technology, and genetics of STR
markers. London: Academic Press.

Butrint Foundation. Butrint Foundation Annual Report 2004.
http://www.butrintfound.dial.pipexcom/publications/ar04.pdf. Accessed Aug
2006.

Capasso, L. 1999. Brucellosis at Herculaneum (79 AD). International Journal of
Osteoarchaeology. 9: 277 — 288.

Capasso, L. 2002. Bacteria in two-millennia-old cheese, and related epizoonoses in
Roman populations (abstract). Journal of Infectious Disease. 45: 122—127

Center for Disease Control and Prevention (CDC). Division of Bacterial and Mycotic

Diseases: Brucellosis. 2005.
http://www.cdc.gov/NCIDOD/DBMD/DISEASEINFO/brucellosis_g.htm -

howcommon Accesed Aug 2006.

Clemmer, V. 2005. Maternal Relatedness within Double Burials of an Ancient Albanian
Tumulus. Thesis for Degree of MS. School of Criminal Justice. Michigan State

University.

Donughue, H.D., M. Spigelman, J. Zias, A.M. Gemaey-Child, and DE. Minnikin. 1998.
Mycobacterium tuberculosis complex DNA in calcified pleura from remains 1400
years old. Applied Microbiology. 27: 265 — 269.

Drancourt, M. and D. Raoult. 2004. Molecular detection of Yersinia pestis in dental pulp.
Microbiology. 150: 263 — 264.

Edson, S.D., J.P. Ross, M.D. Coble, T.J. Parsons and SM. Barritt. 2004. Naming the
Dead: Confronting the Realities of Rapid Identification of Degraded Skeletal
Remains. Forensic Science Review. 16: 63 — 90.

Fenton, T. E-mail and personal communication. 7/1/2005.

109

Fenton, T. E-mail and personal communication. 8/1/2006.

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.

Foran, D. R. 2006. Relative degradation of nuclear and mitochondrial DNA: an
experimental approach. Journal of Forensic Sciences. 51: 766 — 770.

Gehring, M. E. 2004. The Recovery and Analysis of Mitochondrial DNA from Exploded
Pipe Bombs. Thesis for the Degree of MS. School of Criminal Justice. Michigan
State University.

Geyik, M.F., A. Our, K Nas, R. Cevik, J Sarac, and B. Dikici. 2002. Musculoskeletal
involvement in brucellosis in different age groups: a study of 195 cases. Swiss
Medical Weekly. 132: 98 — 105.

Hall TA. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis
program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41:95-98.

Hummel, S. 2003. Ancient DNA typing: Methods, Strategies, and Applications. New
York: Springer-Verlag Berlin Heidelberg.

J affe, H. L. 1972. Metabolic, Degenerative and Inflammatory Diseases of Bones and
Joints. Philadelphia: Lea and Febiger.

Kelley, M., and MA. Micozzi. 1984. Rib Lesions in Chronic Pulmonary Tuberculosis.
American Journal of Physical Anthropology. 652381 — 386.

Kelley, M., and M.Y. El-Najjar. 1980. Natural Variation and Differential Diagnosis of
Skeletal Changes in Tuberculosis. American Journal of Physical Anthropology.
52:153 — 167.

Mays, S., G.M. Taylor, A]. Legge, D.B. Young, and G. Tumer-Walker. 2001.
Paleopathological and Biomolecular Study of Tuberculosis in a Medieval Skeletal
Collection from England American Journal of Physical Anthropology. 114:298 —
31 l.

Mays, S., E. Fysh, and GM. Taylor. 2002. Investigation of the Link Between Visceral
Surface Rib Lesions and Tuberculosis in a Medieval Skeletal Series from England
Using Ancient DNA. American Journal of Physical Anthropology. 19:27 — 36.

Meyers, Melissa. 2006. The Utility of microbial dna and terminal restriction fragment

length polymorphism analysis in the forensic examination of soil. Thesis for the
Degree of MS. School of Criminal Justice. Michigan State University.

110

Misner, L. 2004. Predicting mtDNA Quality based on Bone Weathering and Type. Thesis
for Degree of MS. School of Criminal Justice. Michigan State University.

Ortner, DJ, and WC. J. Putschar. 1981. Identification of Pathological Conditions in
Human Skeletal Remains. Washington. 29 — 52.

O’Rourke, D.H., G. Hayes, and SW. Carlyle. 2000. Ancient DNA Studies in Physical
Anthropology. Annual Review of Anthropology. 29:217 — 242.

Parra, C.A., L.P. Londono, P. Del Portillom and ME. patarroyo. 1991. Isolation,
characterization and molecular cloning of a specific Mycobaterium tuberculosis
antigen gene: identification of a species-specific sequence. Infection and

Immunity. 59: 3411 — 3417.

Rafi, A., M. Spiqelman, J. Stanford, E. Lemma, H. Donoqhue, and J. Zias. 1994.
Mycobacterium leprae DNA from ancient bone detected by PCR. Lancet. 343:
1360 — 1361.

Rennick, S. 2005. Genetic Analysis of Monumental Structure within the Kamenica,
Albania Tumulus. Thesis for the Degree of MS. School of Criminal Justice.
Michigan State University.

Roberts, C., D. Lucy, and K. Manchester. 1994. Inflammatory Lesions of Ribs: An
Analysis of the Terry Collection. American Journal of Physical Anthropology.
952169 - 182.

Sallares, R. and S. Gomzi. 2001. Biomolecular archaeology of malaria. Ancient
Biomolecules. 3: 195 — 213.

Shawar, R.M., F.A. el-Zaatari, A. Nataraj and J.E. Clarridge. 1993. Detection of
Mycobacterium tuberculosis in clinical samples by two-step polymerase chain
reaction and nonisotopic hybridization methods. Journal of Clinical Microbiology.
31: 61 — 65.

Shundi, A. Country Pasture/Forage Resource Profiles: Albania. 2004.
http://www.fao.org/AG/agp/agpc/doc/Counprof/Albania/albania.htm. Accessed
December 2006.

Sreevatsan, S., P. Escalante, X. Pan, D.A. Gilles, S. Siddiqui, C.N. Khalaf, B.N.
Kreiswirth, P. Bilfani, L.G. Adams, T. Ficht, V.S. Perumaalla, M.D. Cave , J.D.
can Embden and J .M. Musser. 1996. Identification of a polymorphic nucleotide in
OxyR specific for Mycobacterium bovis. Journal of Clinical Microbiology. 34:
2007 — 2010.

111

Taylor, G.M., M. Goyal, A.J. Legge, R.J. Shaw, and D. Young. 1999. Genotypic analysis
of Mycobacterium tuberculosis from medieval human remains. Microbiology.
145: 899 — 904.

Ubelaker DH, and EB. Jones, ed. 2003. Human remains from Voegtly Cemetery,
Pittsburgh, Pennsylvania. Smithsonian Contributions to Anthropology Series,
Number 46, Smithsonian Institution Press: Washington DC.

Uehlinger, E. 1970. Epidemiology of tuberculosis: Pathological anatomy. Pneumology.
143: 221 - 234.

UNESCO. Butrint Albania: Advisory Board Evaluation. 1999.
httLJ/whc.unesco.oryarchive/advisory body evaluation/570bis.pdf. Accessed
Aug 2006.

UNESCO. World Heritage Centre. 2006. http://whc.unesco.org/pg.cfm. Accessed Aug
2006.

Von Endt, D.W., and D. J. Ortner. 1977. Amino acid analysis of bone in a suspected case
of iron deficiency anemia. American Journal of Physical Anthropology. 47: 165.

Von Endt, D.W., and D. J. Ortner. 1982. Amino acid analysis of bone from a possible
case of prehistoric iron deficiency from the American southwest. American
Journal of Physical Anthropology. 59: 377 — 385.

Wallach, J ., L.E. Samartine, A. Efron, and P. C. Baldi. 1998 "Human infection by
brucella melitensis: an outbreak attributed to contact with infected goats. FEMS
Immunology and Medical Microbiology Microbiology. 19: 315 — 321.

Wilson I. 1997. Inhibition and Facilitation of Nucleic Acid Amplification. Applied and
Environmental Microbiology. 63: 3741 — 3751.

World Health Organization (WHO). WHO Surveillance Programme for Control of
Foodbome Infections and Intoxications in Europe: 8t Report on Albania. 2002.
http://www.bfr.bund.de/intemet/8threport/CRs/alb.pdf -
search=%22WHO%2OBrucellosis%20Albania%22. Accessed Aug 2006.

World Health Organization (WHO). Albania Country Profile. 2006.
http://www.who.int/countries/alb/en/, . Accessed Aug 2006.

World Health Organization (WHO) Albania: Tuberculosis Prevalence/Incidence. 2004.
http://www.who.int/GlobalAtlas/predefinedReports/TB/PDF Files/AL 2004 Bri

efpdf. Accessed Aug 2006.

Zickel, R. and W.R. Iwaskiw. 1994. Albania: A country study, 2nd Ed. Library of
Congress: Federal Research Division.

112

Zimmerman, M.R., and MA. Kelley. 1982. Atlas of Human Paleopathology. New York:
Praeger Publishers.

Zink, A.R., W. Grabner, and A.G. Nerlich. 2005. Molecular Identification of Human
Tuberculosis in Recent and Historic Bone Tissue Samples: The Role of
Molecular Techniques for the Study of Historic Tuberculosis. American Journal
of Physical Anthropology. 126: 32 — 47.

Zivanovic, S. Infectious and Contagious Diseases. 1982. In: Ancient Diseases. London:
Methuen & CO: 217 - 243.

113