Staying out in the cold: glacial refugia and mitochondrial DNA phylogeography in ancient European brown bears.

Abstract

Models for the development of species distribution in Europe typically invoke restriction in three temperate Mediterranean refugia during glaciations, from where recolonization of central and northern Europe occurred. The brown bear, Ursus arctos, is one of the taxa from which this model is derived. Sequence data generated from brown bear fossils show a complex phylogeographical history for western European populations. Long-term isolation in separate refugia is not required to explain our data when considering the palaeontological distribution of brown bears. We propose continuous gene flow across southern Europe, from which brown bear populations expanded after the last glaciation.

Full text

Page 1

Molecular Ecology (2007) 16, 5140–5148doi: 10.1111/j.1365-294X.2007.03590.x
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
Blackwell Publishing Ltd
FAST-TRACK ARTICLE
Staying out in the cold: glacial refugia and mitochondrial
DNA phylogeography in ancient European brown bears
CRISTINA E. VALDIOSERA,* NURIA GARCÍA,*† CECILIA ANDERUNG,‡§ LOVE DALÉN,*¶
EVELYNE CRÉGUT-BONNOURE,** RALF-DIETRICH KAHLKE,†† MATHIAS STILLER,‡‡
MIKAEL BRANDSTRÖM,‡ MARK G. THOMAS,§ JUAN LUIS ARSUAGA,*†
ANDERS GÖTHERSTRÖM*‡ and IAN BARNES¶
*Centro Mixto UCM-ISCIII de Evolución y Comportamiento Humanos, c/Sinesio Delgado 4 Pabellon 14, 28029 Madrid, Spain,
†Departamento de Paleontología, Universidad Complutense de Madrid, Facultad de Ciencias Geologicas, Ciudad Universitaria 28040
Madrid, Spain, ‡Evolutionary Biology Center, Uppsala University, Norbyvägen 18, SE-752 36 Uppsala, Sweden, §Department of
Biology, University College London, Wolfson House, 4 Stephenson Way, London NW1 2HE, England, UK, ¶School of Biological
Sciences, Royal Holloway, University of London, Egham, Surrey TW20 0EX, England, UK, **Muséum Requien, 67, rue Joseph Vernet,
84000 Avignon, France, ††Forschungsinstitut Senckenberg, Forschungsstation für Quartärpaläontologie, Am Jarkobskirchhof 4, 99423
Weimar, Germany, ‡‡Max Planck Institute for Evolutionary Anthropology, Deutscher Platz 6, D-04103 Leipzig, Germany
Abstract
Models for the development of species distribution in Europe typically invoke restriction
in three temperate Mediterranean refugia during glaciations, from where recolonization of
central and northern Europe occurred. The brown bear, Ursus arctos, is one of the taxa from
which this model is derived. Sequence data generated from brown bear fossils show a
complex phylogeographical history for western European populations. Long-term isolation
in separate refugia is not required to explain our data when considering the palaeontological
distribution of brown bears. We propose continuous gene flow across southern Europe,
from which brown bear populations expanded after the last glaciation.
Keywords: ancient DNA, expansion contraction model, gene flow, last glacial maximum, phylo-
geography, recolonization
Received 24 June 2007; revision accepted 18 September 2007
Introduction
Extensive climatic fluctuations during the late Quaternary
have influenced the demographic history, genetic diversity,
and present-day distribution of many species (Webb &
Bartlein 1992). Current models propose that the onset of
maximal glacial conditions restricted many temperate plant
and animal species into unglaciated refugia. Three main
European refugia have been proposed for the Pleistocene:
the Iberian, Italian and Balkan peninsulas (Bennett et al. 1991;
Taberlet & Bouvet 1994; Hewitt 1996). According to the
expansion/contraction model (E/C), populations survived
in these southern refugia during Pleistocene glaciations
and expanded into mainland Europe as the glaciers retreated
(Hewitt 1996). Both the confinement of populations to
southern refugia and the subsequent range expansion
following glacial retreat should result in population
bottlenecks that reduce genetic variation in the recolonized
areas (Hewitt 1996; Taberlet et al. 1998; Hewitt 1999; Randi
2003; Rowe et al. 2004). Over multiple glaciations, including
the last glacial maximum (LGM) 23 000–18 000 years ago
(Kukla et al. 2002), refugial populations would have been
isolated from one another for extended periods of time and
consequently would be expected to exhibit significant
genetic divergence from each other (Hewitt 1996; Hewitt
2000). However, the general application of this model to
different taxa has been questioned because of palaeontolo-
gical evidence (Stewart & Lister 2001), DNA sequence data
(Rowe et al. 2004; Kotlik et al. 2006) and pollen records
(Willis et al. 2000).
As the modern phylogeographical distribution of brown
bear haplogroups is consistent with origins in the three
major European refugia, the brown bear has served as one
of the model species supporting a scenario of glacial refugia
Correspondence: Cristina E. Valdiosera, c/Sinesio Delgado No. 4
Pabellon 14. 28029 Madrid, Spain. Fax: +34 91-8222855; E-mail:
cvaldiosera@isciii.es

Page 2

GLACIAL REFUGIA AND BROWN BEAR PHYLOGEOGRAPHY 5141
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
and postglacial recolonization of central and northern
Europe (Taberlet et al. 1994, 1998; Hewitt 1999, 2000, 2001,
2004). There is a clear division into two main mitochondrial
lineages in modern European brown bear populations.
These populations are divided into those carrying an eastern
lineage (clade IIIa, Leonard et al. 2000), which is composed
of Russian, northern Scandinavian and eastern European
populations, and those carrying a western lineage (clade I,
Leonard et al. 2000), which is composed of two subgroups,
one believed to originate from the Iberian Peninsula,
including southern Scandinavian bears and the Pyreneean
populations; and the other from the Italian–Balkan penin-
sulas (Taberlet et al. 1994; see however Kohn et al. 1995). In
addition, based on the subfossil record in northwestern
Moldova and mitochondrial DNA data from modern
populations, a Carpathian refuge has also been proposed
(Sommer & Benecke 2005; Saarma et al. 2007). Two contact
zones have been identified where the two main lineages
meet each other, one in Sweden (Taberlet et al. 1995) and the
other one in Romania (Kohn et al. 1995). However, because
of historical human activity (Taberlet et al. 1994; Waits et al.
1999), the current geographical distribution of brown bears
in Europe is fragmented and reduced to just a few small
populations in the West and some larger ones in the East.
Furthermore, brown bears are now extinct in Central Europe,
thus interpretation of the present-day phylogeographical
pattern may reveal little about the glacial demographic
history. Therefore, analyses of bear fossil remains throughout
their historical range can provide direct information of the
past population structure.
Previous ancient DNA work on European brown bears
(two individuals from Austria) has been used to suggest a
different phylogeographical structure in brown bears before
the LGM, to that in modern-day populations (Hofreiter
et al. 2004). In this study, we focus on glacial and postglacial
populations, particularly those from Western Europe.
Assuming (i) that the E/C model is a major factor in the
formation of the phylogeographical structure seen in
modern brown bear populations (Taberlet et al. 1994; Hewitt
1996), and (ii) the classical view of Mediterranean refugia
for temperate species, from which recolonization of central
and northern Europe took place, two predictions can be
made for postglacial genetic patterns of brown bears: (i)
haplotypes not associated with the traditional refugia
should not occur in mainland Europe after the LGM, and
(ii) populations in the refugia should be reciprocally mono-
phyletic due to long-term isolation. Other temperate spe-
cies of small mammals, currently inhabiting Central and
northern Europe, such as the pygmy (Sorex minutus) and
common shrews (Sorex araneus), and the bank vole (Clethri-
onomys glareolus) have proven not to have a Mediterranean
peninsular (i.e. Iberian, Italian or Balkan) origin (Bilton
et al. 1998), and thus do not fit the traditional E/C model.
Significant differences between major clades, and clear
spatial patterning of these clades in modern (Taberlet et al.
1994) and past (Barnes et al. 2002) populations, attributed
to maternal philopatry, make brown bears a tractable and
extensively studied taxon for ancient DNA studies (Leonard
et al. 2000; Barnes et al. 2002; Hofreiter et al. 2004). In order
to investigate the role played by ancient glacial refugia in
shaping the current haplotype distributions, we analysed
mitochondrial DNA haplotype distribution and nucleo-
tide diversity in ancient brown bear samples from Spain,
Southern France, Germany, Italy and Romania, with dates
ranging from the Late Pleistocene to Late Holocene (Fig. 1).
Materials and methods
DNA extraction
A total of 66 bones and teeth of brown bear specimens from
different sites (Fig. 1) were collected from several museum
collections (see Supplementary material) and extracted
together with 43 water negative controls.
The following procedure was employed to all samples:
Pieces of bone and tooth of about 1 cm3 were cleaned on
the surface with 1 m HCL, and then ground to powder under
liquid nitrogen using a Spex 6700 Freezer Mill according to
the manufacturer’s instructions. Samples were extracted
twice, using both solvent and silica-binding approaches
Fig. 1 Map of geographical distribution of sample localities. Sites
of sampled fossil/subfossil Ursus arctos: 1, Navacepeda; 2,
Cantabria; 3, Cuevas del Somo; 4, Atapuerca/Cueva Major; 5,
Mont Ventoux; 6, Mühlberg; 7, Wysburg near Weisbach; 8, Bad
Frankenhausen; 9, Dienstedt; 10, Hohle Fels Cave; 11, Grotta
Beatrice; 12, Pestera Baltagul.

Page 3

5142 C. E. VALDIOSERA ET AL.
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
(Yang et al. 1998; Leonard et al. 2000). Finally, an independent
replication was carried out using the method of Anderung
et al. (2005) in five samples (see below). In every extraction,
150–250 mg of bone powder was used, except for the
Atapuerca sample, where 500 mg were used.
DNA amplification
Amplification was carried out using two sets of primers
designed to amplify 111-bp and 135-bp fragments (primers
not included) from the control region of the mitochondrial
genome. Primer sequences are as follows: for the short
fragment (111 bp) designed in this study: URSUSF1—136–156
CAGCACCCAAAGCTAATGTTC and URSUSR1—273–290
GCACGAKMTACATAGGGG; for the long fragment
(135 bp), we used primers L16164 and H16299 from (Hänni
et al. 1994). Polymerase chain reaction (PCR) amplifica-
tions were carried out with 1 μL of DNA extract and 1 U
of Platinum Taq HiFi Polymerase (Invitrogen) when using
phenol–chloroform DNA extraction protocol and 5 μL of
DNA extract and 2 U of HotStar Taq (QIAGEN) when using
the silica-based protocol. Reaction conditions were perfor-
med as in Leonard et al. (2000). PCR products were purified
using a MiniElute PCR Purification Kit according to the
manufacturer’s instructions before cycle sequencing. The
sequencing reaction was purified with isopropanol (once
at 80% and then 70%) and formamide before PCR products
were screened with an ABI PRISM 3100 genetic analyser.
Extractions were carried out in an ancient DNA laboratory
at the University College London and repeated in a different
ancient DNA laboratory at Centro Mixto UCM-ISCIII de
Evolución y Comportamiento Humanos in Madrid. Four
out of five samples were successfully replicated in an inde-
pendent laboratory in Uppsala, Sweden (MV4 L6 714, MV4
L5 1184, MV4 L6 851, Cueva de las motas 33-1). A fifth sample
(Vallecampo) replicated in Uppsala University showed
sequence variation; in this case, the sample was re-extracted
twice. Every sample, from which there was enough material
and gave both short and long DNA sequences, was sent for
radiocarbon dating except the two bears from Romania,
which were indirectly dated, that is associated with multiple
other remains from the same site (Table 1); all dates are
represented in radiocarbon years. Every sample was extra-
cted and amplified at least twice, in addition to the inde-
pendent laboratory replication. In four cases, we observed
sequence variation in the same sample. In those cases, we
applied the majority rule consensus (Krause et al. 2006)
considering two out of three sequences (from three differ-
ent amplifications) the correct one for samples MV4 L6714
and MV4162 and three out of four sequences for samples
Remanie and Vallecampo.
Phylogenetic analyses
Phylogenetic relationships were estimated using maximum
likelihood and Bayesian inference. Maximum-likelihood
Table 1 Samples analysed. Ancient brown bear samples that yielded the complete sequence of 193 bp, geographical site, assigned date/
cultural context and haplotype. Samples from Romania marked with a * were not AMS dated directly but could be associated with other
bear remains dated from the same site to 3rd century BC
SampleSite C14 date bp
HaplotypeAccession no.
Vallecampo
C.Motas 33-2
C.Motas 33-1
Mv4 K3 99
Atapuerca
Gbcm2
Asturias
GEE
Mv4 L6 714
Mv4 L5 1184
Mv4 Mr-204–48
Mv4 M5 162
Mv4 L6 851
Mv4 Remanie
Hem
A3
A5
A9
A12
Romania1
Romania2
Cuevas del Somo (Burgos, Spain)
Cuevas del Somo (Burgos, Spain)
Cuevas del Somo (Burgos, Spain)
Mont Ventoux (Vaucluse, France)
Atapuerca-Cueva Mayor (Burgos, Spain)
Grotta Beatrice Cenci (Abruzzo, Italy)
Cantabria Cave (Cantabria, Spain)
Cuevas del Somo (Burgos, Spain)
Mont Ventoux (Vaucluse, France)
Mont Ventoux (Vaucluse, France)
Mont Ventoux (Vaucluse, France)
Mont Ventoux (Vaucluse, France)
Mont Ventoux (Vaucluse, France)
Mont Ventoux (Vaucluse, France)
Navacepeda (Avila, Spain)
Dienstedt (Thuringia, Germany)
Mühlberg (Thuringia, Germany)
Bad Frankenhausen (Thuringia, Germany)
Wysburg (Thuringia, Germany)
Pestera Baltagul (Romania)
Pestera Baltagul (Romania)
7500 ± 55
—
4624 ± 45
1570 ± 35
17 440 ± 425
16440 ± 65
—
5380 ± 45
4645 ± 40
3845 ± 40
1750 ± 30
1790 ± 55
6525 ± 50
3445 ± 40
350 ± 40
1665 ± 35
1770 ± 35
5210 ± 35
XII–XIV century
III century*
III century*
Ua8
Ua8
Ua9
Ua7
Ua20
Ua4
Ua10
Ua8
Ua13
Ua12
Ua17
Ua7
Ua11
Ua15
Ua6
Ua18
Ua22
Ua15
Ua15
Ua23
Ua23
EF488487
EF488502
EF488503
EF488495
EF488504
EF488488
EF488489
EF488490
EF488496
EF488492
EF488493
EF488491
EF488494
EF488505
EF488497
EF488501
EF488498
EF488499
EF488500
EF488506
EF488507

Page 4

GLACIAL REFUGIA AND BROWN BEAR PHYLOGEOGRAPHY 5143
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
trees and all model parameters (Ti/Tv = 100.0, I = 0.344,
alpha = 0.452) were estimated with phyml (Guindon et al.
2005). Substitution models were compared with likelihood-
ratio tests when nested, and with the Aikake Information
Criterion. The Hasegawa–Kishino–Yano + invariant sites +
gamma model of sequence evolution was used to generate
Bayesian posterior probabilities. Markov chain Monte Carlo
sampling was performed as implemented in the phylo-
genetic analysis software mrbayes (Huelsenbeck et al. 2001;
Ronquist & Huelsenbeck 2003), using 5 000 000 iterations,
sampling every 1000, with the first 1250 samples (25%)
discarded as burn-in. Node support was calculated accord-
ing to 500 bootstrap replicates (maximum likelihood) and
posterior probabilities (Bayesian inference) (Fig. 2a).
Fig. 2 (a) Maximum-likelihood tree of
ancient and modern brown bear sequences.
Numbers above nodes represent bootstrap
values, below nodes are posterior proba-
bilities. (b) Minimum-spanning network for
haplotypes in the western (W) clade. Missing
haplotypes are shown with a dot and
colours correspond to geographical origin.

Page 5

5144 C. E. VALDIOSERA ET AL.
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
A minimum-spanning network based on pairwise dif-
ferences among haplotypes in 30 sequences was constructed
using arlequin (version 2.000) (Fig.2b) (Schneider etal. 2000).
To test for a possible founder effect due to recolonization
events (Edmonds et al. 2004) in the central European bears,
we used estimated pairwise difference nucleotide diversity
(π, Nei 1987; Nei & Miller 1990). The sequence data were
grouped into two data sets (peninsular n = 16 and main-
land n = 17) (Table 2). A value for π, with confidence inter-
vals, was estimated through nonparametric bootstrapping
with 10 000 bootstrap replicates.
Results
From 66 samples (see Supplementary material) of bone or
tooth, 21 yielded reproducible sequences for both targeted
fragments (Accession GenBank nos EF488487–EF488507)
(Table 1). Contamination was not detected in any of the 43
water negative controls. We obtained 16 radiocarbon dates
for samples from The Ångström Laboratory, Uppsala Univer-
sity, Sweden, using the accelerator mass spectrometry (AMS)
method (insufficient sample was available for specimens
A12, C. Motas 33-2 and Asturias) (Table 1). One more date
(sample gbcm2) was generated at Centro di Datazione e
diagnostica at the Universita degli Studi di Lecce and
provided by the Soprintendenza per i Beni Archeologici
dell’Abruzzo, Chieti, Italy. The two Romanian samples
were dated through association with other samples from
the same site that yielded largely overlapping AMS dates.
Phylogenetic analysis
The 21 ancient sequences obtained in this study (Fig. 1),
together with 19 previously published sequences obtained
from the GenBank database (see Supplementary material)
(Taberlet et al. 1994; Barnes et al. 2002; Hofreiter et al. 2004),
were used to infer a phylogenetic tree. We used the cave bear
(Ursus spelaeus), the evolutionary closest distinct species as an
outgroup (Loreille et al. 2001; Hofreiter et al. 2002; Fig. 2a).
Using Bayesian, maximum-likelihood and neighbour-
joining approaches, we obtained a tree topology similar to
that previously inferred (Leonard et al. 2000; Barnes et al.
2002); eastern and polar bear clades (IIIa and IIb) were well
differentiated from the western clade. However, the western
European populations showed a complex glacial and post-
glacial phylogeographical structure, making it difficult to
separate haplogroups according to their geographical
origin and showed very low levels of population differen-
tiation as opposed to the strong phylogeographical struc-
ture expressed in present-day populations (Taberlet et al.
1994). For example, one sample originating in Iberia (dated
to 17 440 ± 425 bp) yielded a haplotype (Ua20) that falls into
the clade believed to originate in Italy. Similarly, another
sample originating in Italy (dated to 16 440 ± 65 bp) yielded
a haplotype (Ua4) grouping with the clade hypothesized to
originate from Iberia (Taberlet et al. 1994). Moreover, western
(Ua15, Ua18) and eastern (Ua22) haplotypes dating to
5210 ± 35, 1665 ± 35, and 1770 ± 35 bp, respectively, were
identified in Germany (Fig. 2b).
Minimum-spanning network
Because of the complexity shown during the LGM and
post-LGM in the phylogeographical structure of western
European populations and the difficulty to associate haplo-
groups to a specific peninsular origin, in order to develop
a better understanding for the haplotype distrtibution of
this clade, we constructed a minimum-spanning network
Table 2 Samples used for π test. Geogra-
phical site and age of the samples assigned
to ‘mainland’ Europe and ‘Peninsular
Groups’ used to compare nucleotide
diversity, respectively
‘Mainland’ Europe group AgePeninsular group Age
Dalarna (Sweden)
Norway
Pyrenees
Mv4 M5162 (France)
Mv4 L51184 (France)
Mv4 MR 204–48 (France)
Mv4 L6714 (France)
Mv4 L6851 (France)
Mv4 Remanie (France)
Mv4 K3 99 (France)
A5 (Germany)
A9 (Germany)
A3 (Germany)
A12 (Germany)
Scotland
Ramesch (Austria)
Winden (Austria)
Present
Present
Present
1790 ± 55 bp
3845 ± 40 bp
1750 ± 30 bp
4645 ± 40 bp
6525 ± 50 bp
3445 ± 40 bp
1570 ± 35 bp
1770 ± 35 bp
5210 ± 35 bp
1665 ± 35 bp
XII-XIV century
Late Holocene
47420 bp
39940 bp
Abruzzo (Italy)
Bulgaria
Croatia
Greece
Cantabria (Iberia)
Slovenia
Asturias (Iberia)
GEE (Iberia)
HEM (Iberia)
C.Motas 33–1 (Iberia)
C.Motas 33–2 (Iberia)
Atapuerca (Iberia)
Gbcm2 (Italy)
Vallecampo (Iberia)
Romania1
Romania2
—
Present
Present
Present
Present
Present
Present
—
5380 ± 45 bp
350 ± 40 bp
4624 ± 45 bp
—
17 440 ± 425 bp
16 440 ± 65 bp
7500 ± 55 bp
III century
III century
—

Page 6

GLACIAL REFUGIA AND BROWN BEAR PHYLOGEOGRAPHY 5145
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
of modern (n = 10) (Taberlet et al. 1994; Kohn et al. 1995)
and ancient (n = 20) sequences originating from different
geographical regions across Europe: Scotland (Barnes et al.
2002), Austria (Hofreiter et al. 2004) Iberia, Italy, France,
Germany and Romania (Fig. 2b). A total of 21 haplotypes
were obtained, spanning a time range from 17 440 ± 425 to
the present.
In the southern French site of Mont Ventoux 4 (Crégut-
Bonnoure et al. 2005), there are six haplotypes (Ua7, Ua11–13,
Ua15 and Ua17) distributed throughout the network, span-
ning a time range of 6525 ± 50–1570 ± 35 bp (Fig. 2a, b).
Interestingly, the three oldest samples yielded haplotypes
that form a monophyletic group (Ua11–13) positioned
between the hypothesized Iberian and Italian/Balkan clades
in the minimum-spanning network. These three novel
haplotypes dated to 6525 ± 50, 3845 ± 40, and 4645 ± 40 bp,
respectively, form a unique and previously undescribed,
post-LGM European lineage.
Nucleotide diversity
We could not find any significant difference in estimated π
(P = 0.28; determined through bootstrapping 10 000 repli-
cates) between the three combined peninsulas (Iberia, Italy
and Balkans, π = 0.0388) and the mainland (P = 0.0455).
Discussion
We did not observe the patterns of phylogeographical
structuring that are expected for a peninsular origin for
populations inhabiting mainland Europe after the LGM.
Instead, we found high diversity in the samples from Mont
Ventoux 4 in southern France as well as in four German
sites (Bad Frankenhausen, Dienstedt, Wysburg, Mühlberg).
We found no evidence for reciprocal monophyly in the
peninsulas during or after the LGM, as would be expected
following the long-term isolation of these populations.
Thus, we fail to provide support for the two predictions
made based on an E/C model; our data indicated that the
classic glacial refugium model is insufficient to explain the
postglacial genetic history of European brown bears. There
have been objections to the general application of the
refugium model to other taxa. Palaeontological evidence
(Stewart & Lister 2001), as well as DNA sequence data (Rowe
et al. 2004; Kotlik et al. 2006) and pollen records (Willis et al.
2000) are not always in agreement with an E/C scenario.
In a previous study, two pre-glacial brown bear sequences,
among other Late Pleistocene species, were used to argue for
a lack of phylogeographical structure shortly before the
last glacial maximum (Hofreiter et al. 2004). The sequences
were retrieved from two cave sites in Austria, Winden (a
western haplotype) and Ramesch (an eastern haplotype).
Interestingly, the Winden cave is located at the East of the
Ramesch cave, this latter one located at the west of the
Romanian contact zone. The lack of phylogeographical
structure has been explained as the consequence of exten-
sive mixing between populations following a postglacial
expansion, whereby the phylogeographical structure
formed during long-term isolation in southern glacial ref-
ugia was eroded during postglacial expansion by extensive
gene flow. Our data show a similar mixing of mitochon-
drial types, however, this time during the LGM within gla-
cial refugia populations in the Iberian and Italian peninsulas.
These two findings together suggest that any lack of phyl-
ogeographical structure before the last glaciation continued
during the last glacial maximum. Moreover, during the
Holocene, Mont Ventoux 4 (in southern France) presents
three different haplogroups at different time periods, each
comprising mitochondrial haplotypes of hypothesized
Iberian (Ua7 dated to 1790 ± 55 bp and 1570 ± 35 bp) and
Italian (Ua15 and Ua17 dated to 3445 ± 40 bp, 1750 ± 30 bp,
respectively) populations, and three more haplotypes (Ua11–
Ua13 dated to 6525 ± 50, 3845 ± 40, and 4645 ± 40 bp,
respectively) that are not associated with any of the three
glacial refugia from where the mainland European popu-
lations are suggested to have originated. Similarly, high
genetic diversity is found in Germany, with haplotypes
belonging to the eastern clade and the Italian/Balkan clade
present in German populations during the Holocene.
Finding two different lineages (eastern and western) in
the ancient German samples could be explained by the
existence of a contact zone in this area. This does not, how-
ever, apply to the site of Mont Ventoux 4, where not only
the two known western (Iberian and Italy/Balkans) line-
ages are present, but also one that has no counterpart in
studied ancestral refugia. As the three lineages do not have
a temporal overlap, one can speculate that the presence of
this novel lineage could be due to population replacement,
with a refugium in southern France and an extinction event
during the early Holocene. However, the number of obser-
vations is insufficient to test the serial distribution at this site.
An alternative scenario, more compatible with these data
and the anomalies associated with each of the predictions
under the E/C model, is that brown bears were not restricted
to Mediterranean peninsulas during the LGM, but also sur-
vived in mainland southern Europe. Although our data do
not indicate a clear phylogeographical structure, there is a
tendency for haplotypes to group according to a specific
geographical area in an east–west cline. Such a pattern
would be expected under an isolation by distance model
over a continuous range, as has been suggested previously
for European brown bears (Randi et al. 1994). The lack of
a clear phylogeographical structure before (Hofreiter et al.
2004), during and after the LGM is compatible with
restricted but continuous long-term gene flow among geo-
graphical regions within southern mainland Europe and
the Mediterranean peninsulas (Avise et al. 1987). This
scenario would also explain the lack of monophyly in the

Page 7

5146 C. E. VALDIOSERA ET AL.
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
proposed refugial populations, the existence of highly diver-
gent haplotypes (Ua4 and Ua20) in the peninsulas and the
presence of apparently unique lineages outside the penin-
sulas. Finally, this is more in concordance with not finding
significantly lower diversity in mainland Europe would be
expected following a recolonization event, creating a
founder effect.
There are other forms of evidence that are also in con-
cordance with this scenario. Floral evidence from France
(Renault-Miskovsky & Leroi-Gourhan 1981) as well as
macroscopic charcoal fossils from central Europe (Willis
et al. 2000) have lead to the suggestion of pockets of ther-
mophilus trees in Europe during the LGM, an environment
that could well sustain bear populations. Although there is
little evidence of brown bears in the fossil record during the
LGM, some fossil remains have been reported from a
number of caves: in the Western Carpathians, Lisková, Važec
and Vyvieranie caves, attributed to late glacial period
(Sabol 2001) while Moravany Lopata II (Slovakia) (Musil
2003), Cosauti I (David et al. 2003) and Ciuntu (Borziac et al.
1997) (Moldova) yielded Ursus arctos remains between
24 100 ± 800 bp and 17 030 ± 180 bp. In France, six localities
have yielded remains of U. arctos: Faurous (Gironde; with
Magdalenian industry; Gilbert 1984), Bois Ragot (Goueix,
Vienne: level b dated 11 030 ± 140 bp), Duruthy (Sorde-
l’Abbaye, Landes, level 3 with Magdalenian industry;
Delpech 1983), Harzabaletako Karbia (Aussurucq, Pyrénées-
Atlantiques; 29 200 ± 100 bp; Clot & Duranthon 1990),
Oilascoa (Saint-Michel, Pyrenées-Atlantiques; 18 720 ±
350 bp; Clot et al. 1990), Le Rond du Barry (Solignac, Haute-
Loire) and Solutre, an open cave system in southern France
(layer 6; 21 600 ± 700, 22 650 ± 500, and 23 200 ± 700 bp; Evin
et al. 1994). In Austria, stratigraphic layers with brown bear
presence have been dated to 22 180 ± 190 bp at Willendorf II
(Vogel & Zagwijn 1967; Haesaerts et al. 1996) and between
18 890 ± 140 bp and 19 380 ± 90 bp at Grubgraben (Damblon
et al. 1996; Terberger & Street 2002; Musil 2003). Brown bear
remains have also been recovered in Paviland Cave (Goat’s
Hole), UK, dated to 17 670 ± 140 cal. bp (Aldhouse-Green &
Pettitt 1998), these latter suggesting a cold tundra-steppe
environment. Thus, the fossil record does indicate a wide
geographical range of brown bears across mainland Europe
before and during the LGM. Moreover, the geographical
range of 47 LGM sites containing temperate mammal ele-
ments, such as roe deer, red deer and red fox during 23 000–
16 000 bp, clearly shows a distribution which differs
from the classical view of populations restricted to glacial
Mediterranean refugia (Sommer & Nadachowski 2006).
Brown bears, which have served as a model species for
the refugium model, were probably never dependent on a
refugium system. During the LGM, permafrost and cold
tundra-steppe dominated the region between the ice sheet
covering northern Europe and the mountains separating
the Mediterranean peninsulas from the rest of the continent
(Hewitt 1999). This environment is similar to the habitats
occupied by brown bears today in Alaska, Canada and
Siberia (McLoughlin et al. 2000). Furthermore, there is
evidence in the fossil record of brown bears inhabiting
Beringia during the LGM (Leonard et al. 2000; Barnes et al.
2002). If this was indeed a suitable habitat during the LGM,
then the cold tundra-steppe of southern mainland Europe
is likely to have been as well. Considering that brown bears
represent one of the most adaptable species in Holarctic
ecosystems and the flexibility of their diet (Pasitschniak
1993), which can vary from nearly strict vegetarian to full
carnivore, it seems probable that bears could have occu-
pied a similar ecological niche in France, northern Italy and
Eastern Europe during the LGM.
Theoretical studies have shown that phylogeographical
breaks can evolve without barriers to gene flow, especially
when dispersal distances are low and population size is
decreasing (Irwin 2002). Rather than being the result of
repeated isolation in peninsular refugia, the appearance of
phylogeographical structure in modern-day brown bear
populations may thus be the result of the high degree of
female philopatry in brown bears (Randi et al. 1994) and
the severe reduction in population size during the Late
Holocene (Servheen 1990). Following this, the different
brown bear lineages in Europe would not represent evolu-
tionary significant units, and would instead be the result of
a recent fragmentation caused by human activities.
This study highlights the problem with using mitochon-
drial DNA data to identify units for conservation, since the
current distribution of haplotypes may not accurately
relate to species’ history. Indeed, there is increasing evidence
that phylogeographical inference from modern data con-
flicts with results obtained from ancient DNA (Leonard
et al. 2000; Barnes et al. 2002; Hofreiter et al. 2004; Haak et al.
2005; Dalén et al. 2007; Leonard et al. 2007). Complex demo-
graphic histories, including bottlenecks and replacements,
are difficult to infer using modern genetic data alone. The
pre-glacial (Hofreiter et al. 2004) and postglacial history of
bear populations in Europe reflects this problem, and is
best studied with a temporal data set, including samples
dating to periods before major anthropogenic manipula-
tion took place, and also connected to important climatic
changes. The extent to which this is applicable to other taxa
should be explored.
Acknowledgements
We would like to acknowledge Dr Colin Smith for helpful com-
ments and discussion to improve the manuscript, Universidad de
Burgos (Ana I. Ortega), Museo Nacional de Ciencias Naturales,
Madrid (Begoña Sanchez), Museo de Ciencias Naturales de Alava
(Carmelo Corral), Institutul de Speologie ‘Emil Racovita’, Bucarest
(Dr Silviu Constantin), the speleologists from ‘Ursii’ Caving Club
(Razvan Arghir), Benedetto Sala and Paolo Mazza, Thüringer
Landesamt für Denkmalpflege und Archäologie, Weimar (Dr

Page 8

GLACIAL REFUGIA AND BROWN BEAR PHYLOGEOGRAPHY 5147
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
Habil. S. Ostritz and Dr D. Walter), Soprintendenza per i Beni
Archeologici dell’Abruzzo, Chieti (Maria Adelaide Rossi), for
kindly providing access to fossil material. Enrique Sacristán for
providing one unpublished datation. Cniio for providing labora-
tory facilities.
Financial support was provided by the Consejería de Medio
Ambiente, Ordenación del Territorio e Infraestructuras del Principado
de Asturias (DGVI 1253/03), Ministerio de Ciencia y Tecnología
of the government of Spain (Project No. CGL2006-13532-C03-02)
and Consejo Superior de Investigaciones Cient’ficas (UAC2003-
0034).
References
Aldhouse-Green S, Pettitt P (1998) Paviland Cave: contextualizing
the ‘Red Lady.’ Antiquity, 72, 756–772.
Anderung C, Bouwman A, Persson P et al. (2005) Prehistoric
contacts over the Straits of Gibraltar indicated by genetic analysis
of Iberian Bronze Age cattle. Proceedings of the National Academy
of Sciences, USA, 102, 8431–8435.
Avise JC, Arnold J, Ball RM et al. (1987) Intraspecific phylogeo-
graphy: the mitochondrial DNA bridge between population
genetics and systematics. Annual Review of Ecology and Systematics,
18, 489–522.
Barnes I, Matheus P, Shapiro B, Jensen D, Cooper A (2002) Dynamics
of Pleistocene population extinctions in Beringian brown bears.
Science, 295, 2267–2270.
Bennett KD, Tzedakis PC, Willis KJ (1991) Quaternary refugia of
north European trees. Journal of Biogeography, 18, 103–115.
Bilton DT, Mirol PM, Mascheretti S et al. (1998) Mediterranean
Europe as an area of endemism for small mammals rather than
a source for northwards postglacial colonization. Proceedings of
the Royal Society B: Biological Sciences, 265, 1219–1226.
Borziac IA, Allworth-Jones P, French C et al. (1997) The upper
Palaeolithic site of Ciuntu on MIddle Pruth, Moldova: a
multidiscplinary study and reinterpretation. Proceedings of the
Prehistoric Society, 63, 285–301.
Clot A, Duranthon F (1990) Les mammifères fossiles du Quater-
naire dans les Pyrénées. Muséum d’Histoire naturelle de Toulouse,
159.
Crégut-Bonnoure E, Argant A, Argant J et al. (2005) Les cavités
karstiques du Mont Ventoux (Vaucluse, France): des ours brun
(Ursus arctos L.) et des milieux d’altitude à l’Holocène. Bulletin
du Musée d’Histoire Naturelle de Toulouse, 141, 5–18.
Dalén L, Nyström V, Valdiosera C et al. (2007) Ancient DNA
reveals lack of postglacial habitat tracking in the arctic fox.
Proceedings of the National Academy of Sciences, USA, 104, 6726–6729.
Damblon F, Haeserts P, van der Plicht J (1996) New datings and
considerations on the chronology of Upper Palaeolithic sites in
the Great Eurasiatic Plain. Préhistoire Européene, 9, 177–132.
David AI, Nadachowski A, Pascaru V, Wotjal P, Borziac I (2003)
Late Pleistocene mammal fauna from the Late Palaeolithic
butchering site Cosauti 1, Moldova. Acta Zoologica Cracoviensa,
46, 85–96.
Delpech F (1983) Les faunes du Paléolithique supérieur dans le
Sud-Ouest de la France. Cahiers du Quaternaire, 6, 1–453.
Edmonds C, Lillie A, Cavalli-Sforza L (2004) Mutations arising in
the wave front of an expanding population. Proceedings of the
National Academy of Sciences, USA, 101, 975–979.
Evin J, Bintz P, Monjuvent G (1994) Human settlements and the
last deglaciation in the French Alps. Radiocarbon, 33, 345–357.
Gilbert A (1984) Contribution à l’étude des faunes de la fin des
temps glaciaires et du début des temps post-glaciaires. Thèse de
Doctorat, Université de Bordeaux I, 321.
Guindon S, Lethiec F, Duroux P, Gascuel O (2005) phyml online — a
web server for fast maximum likelihood-based phylogenetic
inference. Nucleic Acids Research, 33, W557–W559.
Haak W, Forster P, Bramanti B et al. (2005) Ancient DNA from the
first European farmers in 7500-year-old Neolithic sites. Science,
310, 1016–1018.
Haesaerts P, Damblon F, Bancher F, Trnka G (1996) Revised
stratigraphy and chronology of the Willendorf II sequence,
Lower Austria. Archaeologia Austriaca, 80, 25–42.
Hänni C, Laudet V, Stehelin D, Taberlet P (1994) Tracking the
origins of the cave bear (Ursus spelaeus) by mitochondrial DNA
sequencing. Proceedings of the National Academy of Sciences, USA,
91, 12336–12340.
Hewitt GM (1996) Some genetic consequences of ice ages, and
their role in divergence and speciation. Biological Journal of the
Linnean Society, 58, 247–276.
Hewitt GM (1999) Post-glacial re-colonization of European biota.
Biological Journal of the Linnean Society, 68, 87–112.
Hewitt GM (2000) The genetic legacy of the Quaternary ice ages.
Nature, 405, 907–913.
Hewitt GM (2001) Speciation, hybrid zones and phylogeography
— or seeing genes in space and time. Molecular Ecology, 10, 537–549.
Hewitt GM (2004) Genetic consequences of climatic oscillations in
the Quaternary. Philosophical Transactions of the Royal Society of
London. Series B: Biological Sciences, 359, 183–195.
Hofreiter M, Capelli C, Krings M et al. (2002) Ancient DNA
analyses reveal high mitochondrial DNA sequence diversity
and parallel morphological evolution of late Pleistocene cave
bears. Molecular Biology and Evolution, 19, 1244–1250.
Hofreiter M, Serre D, Rohland N et al. (2004) Lack of phylogeography
in European mammals before the last glaciation. Proceedings of
the National Academy of Sciences, USA, 101, 12963–12968.
Huelsenbeck JP, Ronquist F, Nielsen R, Bollback JP (2001)
Bayesian inference of phylogeny and its impact on evolutionary
biology. Science, 294, 2310–2314.
Irwin D (2002) Phylogeographic breaks without geographic barriers
to gene flow. Evolution, 56, 2383–2394.
Kohn M, Knauer F, Stoffella A, Schröder W, Pääbo S (1995)
Conservation genetics of the European brown bear — a study
using excremental PCR of nuclear and mitochondrial sequences.
Molecular Ecology, 4, 95–103.
Kotlik P, Deffontaine V, Mascheretti S et al. (2006) A northern
glacial refugium for bank voles (Clethrionomys glareolus).
Proceedings of the National Academy of Sciences, USA, 103, 14860–
14864.
Krause J, Dear PH, Pollack JL et al. (2006) Multiplex amplification
of the mammoth mitochondrial genome and the evolution of
Elephantidae. Nature, 439, 724–727.
Kukla GJ, Clement AC, Cane MA, Gavin JE, Zebiak SE (2002) Last
interglacial and early glacial ENSO. Quaternary Research, 58, 27–
31.
Leonard J, Vilà C, Fox-Dobbs K et al. (2007) Megafaunal extinctions
and the disappearance of a specialized wolf ecomorph. Current
Biology, 17, 1146–1150.
Leonard JA, Wayne RK, Cooper A (2000) Population genetics of Ice
Age brown bears. Proceedings of the National Academy of Sciences,
USA, 97, 1651–1654.
Loreille O, Orlando L, Patou-Mathis M et al. (2001) Ancient DNA
analysis reveals divergence of the cave bear, Ursus spelaeus, and
brown bear, Ursus arctos, lineages. Current Biology, 11, 200–203.

Page 9

5148 C. E. VALDIOSERA ET AL.
© 2007 The Authors
Journal compilation © 2007 Blackwell Publishing Ltd
Mcloughlin P, Case R, Gau R et al. (2000) Population delineation of
barren-ground grizzly bears in the central Canadian Arctic.
Wildlife Society Bulletin, 30, 728–737.
Musil R (2003) The middle and upper palaeolithic game suite in
central and Southeastern Europe. In: Neanderthals and Modern
Humans in the European Landscape during the Last Glaciation (eds
Andel TV, Davis W), pp. 167–190. McDonald Institute for
Archaeological Research monographs, Cambridge, UK.
Nei M (1987) Molecular Evolutionary Genetics. Columbia University
press, New York.
Nei M, Miller JC (1990) A simple method for estimating average
number of nucleotide substitutions within and between
populations from restriction data. Genetics, 125, 873–879.
Pasitschniak M (1993) Mammalian species; Ursus arctos. American
Society of Mammalogists, 439, 1–10.
Randi E (2003) Conservation genetics of carnivores in Italy. C. R.
Biol., 326 (Suppl. 1), S54–S60.
Randi E, Gentile L, Boscagli G, Huber D, Roth HU (1994) Mito-
chondrial DNA sequence divergence among some west
European brown bear (Ursus arctos L.) populations. Lessons for
conservation. Heredity, 73 (5), 480–489.
Renault-Miskovsky J, Leroi-Gourhan A (1981) Palynologie et
archéologie: Nouvellex resultants due Paléolithique supérior au
Mésolithique. Bulletin de l’Association Française Pour Lètude de
Quaternaire, 7–8, 121–128.
Ronquist F, Huelsenbeck JP (2003) mrbayes 3: Bayesian phylo-
genetic inference under mixed models. Bioinformatics, 19, 1572–
1574.
Rowe K, Heske E, Brown P, Paige K (2004) Surviving the ice:
northern refugia and postglacial colonization. Proceedings of the
National Academy of Sciences, USA, 101, 10355–10359.
Saarma U, Ho SY, Pybus OG et al. (2007) Mitogenetic structure of
brown bears (Ursus arctos L.) in northeastern Europe and a new
time frame for the formation of European brown bear lineages.
Molecular Ecology, 16, 401–413.
Sabol M (2001) Fossil and subfossil findings of brown bears from
selected localities in Slovakia. Slovak Geological Magazine, 7, 3–
17.
Schneider S, Roessli D, Excoffier L (2000) ARLEQUIN: A Software for
Population Genetics Data Analysis. Genetics and Biometry
Laboratory, Department of Anthropology, University of Geneva,
Geneva, Switzerland.
Servheen C (1990) The Status and Conservation of the Bears of the
World. International Conference Bear Research Management,
British Columbia, Canada.
Sommer R, Benecke N (2005) The recolonization of Europe by
brown bears Ursus arctos Linnaeus, 1758 after the Last Glacial
Maximum. Mammal Review, 35, 156–164.
Sommer RS, Nadachowski A (2006) Glacial refugia of mammals in
Europe: evidence from fossil records. Mammal Review, 36, 251–265.
Stewart J, Lister A (2001) Cryptic Northern refugia and the origins
of the modern biota. Trends in Ecology & Evolution, 16, 608–613.
Taberlet P, Bouvet J (1994) Mitochondrial DNA polymorphism,
phylogeography, and conservation genetics of the brown bear
Ursus arctos in Europe. Proceedings of the Royal Society B: Biological
Sciences, 255, 195–200.
Taberlet P, Swenson JE, Sandegren F, Bjarvall A (1995) Localization
of a contact zone between two highly divergent mitochondrial
DNA lineages of the brown bear Ursus arctos in Scandinavia.
Conservation Biology, 9, 1255–1261.
Taberlet P, Fumagalli L, Wust-Saucy AG, Cosson JF (1998)
Comparative phylogeography and postglacial colonization
routes in Europe. Molecular Ecology, 7, 453–464.
Terberger T, Street M (2002) Hiatus or continuity? New results for
the question of pleniglacial settlement in Central Europe. Antiquity,
76, 691–698.
Vogel JC, Zagwijn WH (1967) Groningen radiocarbon dates VI.
Radiocarbon, 9, 63–106.
Waits LP, Sullivan J, O’Brien SJ, Ward RH (1999) Rapid radiation
events in the family Ursidae indicated by likelihood phylogenetic
estimation from multiple fragments of mtDNA. Molecular
Phylogenetics and Evolution, 13, 82–92.
Webb T, Bartlein P (1992) Global changes during the last 3 million
years: climatic controls and biotic responses. Annual Review of
Ecology and Systematics, 23, 141–173.
Willis KJ, Rudner E, Sumegi P (2000) The full-glacial forests of
Central and southeastern Europe. Quaternary Research, 53, 203–
213.
Yang D, Eng B, Waye J, Dudar J, Saunders S (1998) Technical note:
improved DNA extraction from ancient bones using silica-
based spin columns. American Journal of Physical Anthropology,
105, 539–543.
Cristina Valdiosera: Ancient DNA, phylogeography and genetic
diversity of brown bears. Dr. Nuria García: Palaeobiology, evolution
and ecology of carnivores. Dr. Cecilia Anderung: Ancient DNA
techniques and cattle domestication. Dr. Love Dalèn: Ancient
DNA, evolution and ecology of mammoths. Dr. Evelyne Crégut-
Bonnoure: Palaeobiology and morphology of Pleistocene Fauna.
Dr. Ralf-D. Kahlke: Palaeobiology and Quaternary large mammals.
Mathias Stiller: Ancient DNA, evolution, ecology and phylo-
geography. Mikael Brandström: Evolutionary Genomics. Dr.
Mark G. Thomas: Genetic anthropology. Prof. Juan Luis Arsuaga:
Evolutionary history of hominids. Dr. Anders Götherström:
Ancient DNA techniques development and cattle domestication.
Dr. Ian Barnes: Genetic changes of mega fauna during the Late
Pleistocene.
Supplementary material
The following supplementary material is available for this article:
Table S1 Previously published sequences. Ancient and modern
brown bear samples taken from GenBank data base with their cor-
responding accession number, geographical location, assigned age
and haplotype
Table S2 Sample information
This material is available as part of the online article from:
http:/ /www.blackwell-synergy.com/doi/abs/
10.1111/j.1365-294X.2007.03590.x
(This link will take you to the article abstract).
Please note: Blackwell Publishing are not responsible for the con-
tent or functionality of any supplementary materials supplied by
the authors. Any queries (other than missing material) should be
directed to the corresponding author for the article.