Ancient DNA from Mycenae Grave circle

This is an article form ournal of Anrchaeological Science that I Purchased. I do not know much if I have the right to put it in this forum so if te thread could be locked for any visitor outsider to the forum ... or remove it if moderators consider it necessary
================================================== =====
a r t i c l e i n f oArticle history:
Received 28 January 2008
Received in revised form 7 April 2008
Accepted 21 April 2008
Keywords:
Ancient DNA
Kinship analysis
Mitochondrial DNA
Mycenae
Polymerase chain reaction
*********************************************
a b s t r a c t

The richness of the burials in Grave Circle B at Mycenae, Greece indicates that the 35 people interred
there held elite status during their lifetimes 3500 years ago. It has been speculated that the groups of
burials represent different dynasties or branches of the same family. To test this hypothesis, we carried
out an exhaustive ancient DNA (aDNA) study of 22 of the skeletons. We were unable to identify nuclear
aDNA in any specimen, but we obtained authentic mitochondrial aDNA sequences for four individuals.
The results were compared with facial reconstructions and interpreted within the archaeological context
represented by the organisation of the graves and the positions of the burials within the graves. We
conclude that the contemporaneous male G55 and female G58 skeletons, which both possess the UK
mitochondrial haplogroup, were brother and sister. The implication is that G58 was buried in Grave
Circle B not because of a marital connection but because she held a position of authority by right of birth.
The results illustrate the difficulty in using aDNA to study kinship relationships between archaeological
specimens, but also show that aDNA can advance understanding of kinship when used to test hypotheses
constructed from other evidence.
 2008 Elsevier Ltd. All rights reserved.

1. Introduction

The Bronze Age citadel at Mycenae in Greece is one of the most
evocative prehistoric sites in all of Europe. The legendary home of
Agamemnon and Clytemnestra, Mycenae held a natural attraction
for early antiquarians in the years before its first systematic study
by Heinrich Schliemann in the 1870s. Schliemann’s famous telegram,
sent during his excavation of Grave Circle A in 1876, stating
that he had ‘‘gazed upon the face of Agamemnon’’, turned out to be
erroneous for the burials that he had uncovered predated the
Trojan War by some four centuries, but his excavations were
nonetheless significant as they established Mycenae as one of the
richest and, by implication, most powerful of the Aegean states
during the 17th to 12th centuries BC.
The burials in Grave Circles A and B span some five generations
during the period 1650–1500 BC, Grave Circle B predating A with
possibly 50 years’ overlap between the two. The Grave Circles
therefore date to the very beginning of the Mycenaean age, the
cusp of the Middle to Late Helladic periods when Mycenae was
establishing itself as a dominant trading and political power in
the eastern Mediterranean. Within Grave Circle B there is a
development from simple cist burials to larger, deeper and richer
Shaft Graves, while Grave Circle A comprises six Shaft Graves. Many
of the burials in Grave Circle B were accompanied by weapons,
pottery and/or gold ornaments; one, in Grave G, was buried with
a face-mask made of electrum (a naturally occurring gold–silver
amalgam). Generally they were less well endowed than the remarkable
gold-laden burials in Circle A, but the richness of both
Grave Circles leaves little doubt that their occupants were elite
members of early Mycenaean society. These considerations prompt
questions about the relationships within this elite group, in particular
whether the individuals were members of a single family or
small number of families who had established themselves as the
ruling dynasty in earlyMycenae. Schliemann believed that the gold
face-masks of Grave Circle A were portraits of the owners and noted
their dissimilarities (Schliemann, 1876), but this observation did
not hinder speculations that the 19 individuals buried in Circle A
were members of a single family (Mylonas, 1957). When Grave
Circle B was discovered in 1951 this assumption gained weight as
the graves in this cemetery appear to be laid out in four groups,
each group spanning the different phases of the site, reminiscent of
family plots. Dr J.L. Angel, who examined the Circle B skeletons in
1954, felt he could visualise the faces that lay over the skulls, and
suggested family relationships (Angel, 1973). Forty years later,
modern techniques of facial reconstruction were applied to the
seven best preserved skulls, the results suggesting that these seven
individuals fall into three groups, the ‘heart-shapes’ comprising
G55, G58 and A62, the ‘long faces’ of G51 and Z59, and the ‘beaky
face’ of B52, with S131, the earliest of these seven burials, having
features in common with both of the first two types (Musgrave
et al., 1995; Prag and Neave, 1997) (Fig. 1).
Ancient DNA (aDNA) analysis has the potential to identify kinship
patterns between groups of skeletons, maternal relationships
being revealed by mitochondrial DNA (mtDNA) typing, paternal
ones by studying markers such as short tandem repeats (STRs) on
the Y chromosome, and general family relationships by typing
autosomal STRs. Realisation of this potential has, however, been
hampered by the problems inherent in aDNA research and
although success has been achieved with historic material (Gill
et al., 1994; Gerstenberger et al., 1999; Dudar et al., 2003; Gilbert
et al., 2007) there have been few authenticated kinship studies with
archaeological specimens. Here we describe an exhaustive aDNA
study of the Grave Circle B skeletons which has resulted in
assignment of mtDNA haplogroups to four skeletons, providing
insights into the family relationships between these individuals.
2. Material and methods

Samples were taken, mostly from mandibles or clavicles to
ensure that the same individual was not inadvertently sampled
twice, from 22 of the 35 skeletons present in Grave Circle B. The
outer 1–2 mm of each bone sample was removed with a sterile
scalpel (Bouwman et al., 2006) and the bone UV irradiated (254 nm,
120,000 mJ cm2, 25 min with the bone rotated 180 between
each exposure) to minimise the risk of contaminating DNA on the
surface being carried over to the extraction process. Each bone was
sealed in a DNA-free plastic bag and crushed into powder. Between
0.49 and 0.51 g of bone powder was measured into 1.5 ml microfuge
tubes. DNA extractions were carried out as previously
described (Bouwman and Brown, 2005). All reagents and buffers
were made up with ultrapure water. Samples were processed in
batches of five, each batch was accompanied by three extraction
blanks (full extraction procedure followed but with no bone
material) and subsequently with two PCR blanks (PCR set up with
ultrapure water rather than bone extract). A standard PCR of 50 ml
contained 2.5 ml of bone extract, 1  buffer (150 mM Tris–HCl pH
8.0, 500 mM KCl), the appropriate amount of MgCl2, 200 mM each
dNTP, 100 ng each primer, 1% bovine serum albumin and 1.25 units
AmpliTaq Gold DNA polymerase (Applied Biosystems). Cycling
conditions were: 4 min at 94 C; followed by 44 cycles of 1 min at
the appropriate annealing temperature, 1 min at 72 C and 1 min at
94 C; followed by 1 min at the annealing temperature and 10 min
at 72 C. Primer sequences, MgCl2 concentrations and annealing
temperatures are given in Supplementary Table 1. Those PCRs
marked in Supplementary Table 1 with an asterisk were combined
in two multiplex (triplex and hexiplex) systems as previously
published (Butler et al., 2003; Coble and Butler, 2005). The Mg2þ
concentrations and annealing temperatures for all the other PCRs
were optimised by us and these PCRs were combined in multiplex
systems that will be described elsewhere (A.S.B. & T.A.B., manuscript
in preparation). PCRs were examined by electrophoresis in 3%
agarose gels. Bands were purified using Qiaquick columns (Qiagen),
and DNA was ligated into pCR2.1 TOPO (Amersham) and cloned
in Escherichia coli TOP10F0 cells. Recombinant plasmid DNA was
purified using Qiaquick columns and sequenced with the ABI Big
Dye sequencing kit (Amersham).

3. Results

We tested samples from 22 skeletons with PCRs directed at 7
mtDNA loci (spanning most of hypervariable region 1 and part of
region 2), 17 autosomal STRs, 5 Y chromosome STRs, 2 Y chromosome
single nucleotide polymorphisms (SNPs), and the sex-specific
region of the amelogenin locus. The results of all PCRs are given in
Table 1. Nuclear amplification products were occasionally obtained,
but too sporadically for the results to be authenticated. With mitochondrial
PCRs, 18 of the 22 samples never gave an amplification
product of the correct size, or if they did then that product was
considered to be non-endogenous to the sample because it was
accompanied by contaminated negative controls, was entirely
made up of sequences containing an unusual mutation at position
16172 possessed by A.S.B., who performed all the extractions and
PCRs, or was not human mtDNA. The other four samples (G55, G58,
Z59 and A62) gave sequences which were considered to derive, at
least in part, from endogenous DNA. The sequences obtained from
G55 and G58 corresponded to mitochondrial haplogroup UK, those
from Z59 corresponded to haplogroup U5a1 or U5a1a, and those
from A62 matched the Cambridge Reference Sequence, compatible
in the region sequenced with various haplogroups including H,
HV1, J, U, U3 and U4 (but not UK, U5a1 or U5a1a). Details of all
sequences and the deduction of haplogroups are given in Supplementary
Material.

4. Discussion
4.1. Ancient human DNA

Authentication criteria for aDNA research have been discussed
extensively in the literature, in particular the difficulty in applying
a full set of validation tests to material that is available only in small
quantities and without associated animal remains. Understandably,
we were allowed to take only minimal sized samples from the
Mycenaean skeletons, in some cases sufficient for just a single DNA
extraction. It is recommended that in situations such as this, researchers
take a cognitive and self-critical approach to authentication
of results (Gilbert et al., 2005), and this iswhatwe attempt to do here.
First, we carried out the work in accordance with the standard
criteria of authenticity for aDNA research (Cooper and Poinar, 2000)
as far as was possible. Ancient DNA extractions and PCRs were
set up in independent, physically isolated labs, each with an
ultrafiltered air supply maintaining positive displacement pressure
and a managed access system. Within these rooms, specimens were

handled and extractions were prepared within a Class II biological
safety cabinet, and PCRs were set up within a laminar flow hood. All
extractions were accompanied by negative controls in which the
entire extraction procedure was performed without bone material,
and all PCRs were accompanied by negative controls containing
water instead of DNA extract: if any negative control gave a band
after PCR then the accompanying test PCRs were disregarded. The
lengths of the template molecules were assessed by carrying out
the MtZ PCR (Supplementary Table 1) with all bone extracts, this
PCR giving a product of 425 bp, generally considered too long to
result from genuine aDNA. An MtZ product therefore indicates an
extract that has probably become contaminated with modern DNA.
In this study, the MtZ PCR never gave a product. Because of the
small amounts of material that were available, itwas not possible to
perform replicate extractions for all skeletons (Table 1). However,
replicate extracts were prepared for each of the four skeletons to
which haplogroups were assigned, and in most cases the critical
nucleotide positions from which haplogroups were deduced were
sequenced in more than one PCR product. All sequences were
obtained after cloning. Because of the limited amounts of material
it was not possible to divide the bone samples so that portions
could be sent to a second lab for independent testing, and, similarly,
there was insufficient material to carry out tests aimed at determining
the overall level of biomolecular preservation in the
Mycenaean bones. Corroboration of the human results could not be
sought through study of associated animal remains, as there were
none. We did not quantify the number of starting templates in the
extracts by real-time PCR, because we do not believe that knowing
the number of starting templates gives any assurance that these
are aDNA rather than contaminants. We did, however, introduce
one additional precautionary step in that we removed the outer
1–2 mm of each bone prior to preparation of extracts. We have
shown that even after extensive handlingmost of the contaminating
DNA in a bone resides in the outer 1–2 mm (Bouwman et al., 2006),
and that very little redistribution occurs if the bone is washed (M.M.
Mundee, A.S.B. and T.A.B., manuscript in preparation).
Second, we considered if aDNA is likely to be present in the
material being studied. The most important consideration in this
regard is not the chronological age of the specimens but their
thermal history, as DNA degradation occurs more rapidly at higher
temperatures. A useful measure is ‘thermal age’ (Smith et al., 2003),
which is calculated from the temperature history of a site and its
geographical location. From experimental studies of DNA breakdown,
plus a consideration of the oldest authenticated detections
of aDNA, it has been estimated that the limit for DNA preservation
is approximately 19,000 years at 10 C: hence specimens from any
site that has a thermal age normalised to 10 C at >19,000 years are
unlikely to contain aDNA. When these calculations are applied
to Greece, a thermal age of 19,000 years at 10 C corresponds to
approximately 3600 chronological years (Chilvers, 2004), suggesting
that Grave Circle B (3600–3500 years ago) is at the very limits
for aDNA preservation and that local factors such as exposure to
water and time since excavation (Pruvost et al., 2007) will be key
to determining if aDNA can be recovered. This conclusion is in
agreement with other work that we have carried out with material
from Greek sites dating to the period 2000–1700 BC (Chilvers et al.,
in press). For example, we have been unable to detect aDNA in
bones from Lerna (Argolis), where the skeletons were buried in
relatively shallow graves in a marshy region, but at Kouphovouno
(Laconia), where the burials were deeper and DNA analysis took
place very soon after excavation, we have detected both mitochondrial
and nuclear aDNA in 8 of the 20 skeletons that we have
examined. Similar results were obtained in a previous study of
Greek bones from a broader time range (Evison, 2001) and also in
a small pilot project that we previously carried out with the Mycenaean
remains (Brown et al., 2000).
Third, we considered the likely pattern of contamination of the
Grave Circle B bones with modern DNA. Grave Circle B was excavated
in 1952–1954, using the procedures current at that time and
hence without precautions to prevent DNA contamination. We
established, however, that since excavation the bones have not
been extensively handled, except by Dr J.L. Angel who carried out
the osteological examination in 1954. The fact that contaminating
DNA might date back to the 1950s suggests that these molecules
would be partially degraded and hence low molecular weight,
reducing the value of theMtZ PCR as a contamination screen, though
not reducing the effectiveness of surface removal as a means of
decreasing carryover of contaminants into extracts. We therefore
surmise that although contamination of these skeletons is possible,
extensive contamination by multiple individuals is unlikely.
Eighteen of the specimens that we studied did not contain
detectable endogenous DNA. The other four skeletons each contained
endogenous DNA from a single source. For G55 and G58 this
endogenous DNA is of the UK haplogroup, for Z59 it is U5a1 or
U5a1a, and for A62 it is any of several haplogroups including H,
HV1, J, U, U3 and U4, but not UK, U5a1 or U5a1a. It is possible that
these endogenous DNAs derive from modern contamination, if: (1)
G55 and G58 were handled by one or more individuals of the UK
haplogroup, who did not handle, or at least did not contaminate,
any of the other 20 specimens that we studied; (2) Z59was handled
by a different individual who did not contaminate any of the other
21 bones; and (3) A62 was similarly handled by another different
individual who did not contaminate any of the other 21 bones. This
scenario is possible but we consider the alternative explanation,
that these endogenous DNAs are ancient in origin, to be more likely.

4.2. Comparisons between mtDNA haplogroups and
facial reconstructions


Facial reconstructions had previously been made for all four of
the skeletons to which we could assign mtDNA haplogroups (Musgrave
et al., 1995; Prag and Neave, 1997). G55 and G58 are an adult
male and female, respectively, both with ‘heart-shaped’ faces, and
both buried together in the same grave (see Fig. 1). This pair are
noteworthy as the male is the only person in Grave Circle B to be
buried with a face-mask, andG58 is one of only four skeletons – of 35
in all – that is definitely female, an under-representation that suggests
that those females that are buried in Grave Circle B each had
some special reason for being there. It has previously been suggested
that G55 and G58 have a family relationship (Prag and Neave,1997),
a hypothesis that is supported by their shared mtDNA haplogroup.
UK is present in 4.9–6.6% of modern Europeans (Richards et al.,
2000), making it relatively unlikely that G55 and G58 share this
haplogroup by chance rather than by shared maternal descent. They
were probably contemporaneous as the remains of G58 were still
articulated at the time theyweremoved aside to allowthe interment
of G55, suggesting that they were buried within a few months of
each other. Their maternal relationship could therefore indicate that
they were brother and sister.
Z59 is an adult male from a different grave, a member of the
‘long face’ group and hence possibly not a member of the same
family as G55 and G58. Again, this hypothesis is supported by the
mtDNA data, as Z59 has a different haplogroup to G55 and G58 and
hence belongs to a different maternal lineage. Z59 is looked on as
one of the ‘founder fathers’ of the Grave Circle elite, as his is a relatively
early burial. In the absence of consistent Y chromosome or
autosomal DNA data we are unable to add anything to the suggestion
that he might be the genetic father or grandfather of some
of the younger burials. A62 is another adult male, the third member
of the group with ‘heart-shaped’ faces, and so possibly a member of
the same family as G55 and G58. Although A62 gave the least
complete DNA results, the partial sequence that was obtained ruled
out the possibility that he has the UK haplogroup, and hence he
does not share maternal descent with G55 and G58. A non-maternal
relationship (e.g. the same father but different mothers) cannot,
however, be discounted.

5. Conclusions

Whether or not a group of skeletons buried in proximity to one
another represent the members of a single family is one of the
central questions arising when human remains are excavated at
archaeological sites of any age. Although there has been success in
using aDNA to study such relationships in a historic context (Gill
et al., 1994; Gerstenberger et al., 1999; Dudar et al., 2003; Gilbert
et al., 2007), our results illustrate the difficulty in applying this type
of analysis to archaeological remains, in which aDNA is generally
poorly preserved and the problems caused by contamination with
modern DNA are more acute. Nonetheless, we show that when
hypotheses about kinship can be constructed from existing evidence
then the limited aDNA data obtainable from archaeological
remains can be used to test those hypotheses and advance
understanding. At Mycenae, the facial reconstructions previously
carried out with the Grave Circle B skeletons, together with the
organisation of the graves and the positions of the burials within
the graves, led to the hypothesis that the contemporaneous male
G55 and female G58 might share a family relationship. The aDNA
results confirm this hypothesis and, taking all the evidence into
account, support the conclusion that G55 and G58 were brother and
sister. The implication is that G58 was buried in this high status and
male-dominated Grave Circle not because of a marital connection
but because she held a position of authority by right of birth. DNA
analysis has therefore enabled us to glimpse the factors contributing
to the organisation of the higher echelons of society at the
beginning of the Mycenaean age.
Acknowledgements
We thank E. Palaiologou and colleagues (Ephorate of Prehistoric
and Classical Antiquities, Nafplion) for material from Grave Circle B.
We also thank H. Clark and E. French (British School at Athens) and
R. Arnott (University of Birmingham) for assistance with permits
and for advice and encouragement, and an anonymous referee for
valuable comments that enabled us to improve the manuscript. The
work was funded by the Leverhulme Trust and the Institute for
Aegean Prehistory.