Content uploaded by Francisco Petrucci-Fonseca
Author content
All content in this area was uploaded by Francisco Petrucci-Fonseca on Dec 21, 2015
Content may be subject to copyright.
Mitochondrial DNA Sequence Variation
in Portuguese Native Dog Breeds:
Diversity and Phylogenetic Affinities
ANA ELISABETE PIRES,LAHOUSSINE OURAGH,MOHSEN KALBOUSSI,JOSE
´
MATOS,
F
RANCISCO PETRUCCI-FONSECA, AND MICHAEL W. BRUFORD
From Centro de Biologia Ambiental, Faculdade de Cie
ˆ
ncias da Universidade de Lisboa, Rua Ernesto Vasconcelos, Edifı
´
cio
C2-3 Piso, 1749-016 Campo Grande, Portugal (Pires and Petrucci-Fonseca); Instituto Nacional de Engenharia, Tecnologia
e Inovacxa
˜
o, Estrada do Pacxo ao Lumiar, 22, Edifı
´
cio E, Molecular Biology Group, 1649-038 Lisboa, Portugal (Pir es and Matos);
Laboratoire d’Analyses Ge
´
ne
´
tiques Ve
´
te
´
rinaires, Institut Agronomique et Ve
´
te
´
rinaire Hassan II, BP 6202-Instituts,
10101-Rabat, Morocco (Ouragh); Department of Biology, Faculty of Sciences of Tunis, University El Manar, Tunisia
(Kalboussi); and Cardiff University, School of Biosciences, P.O. Box 915, CF10 3TL, Cardiff, United Kingdom (Pires and Bruford).
Address correspondence to A. E. Pires at the address above, or e-mail: pireseg@cf.ac.uk.
Abstract
In an extensive survey of the genetic diversity in Portuguese dogs, we have examined an 887-bp fragment of the mitochondrial
DNA (mtDNA) from 8 Portuguese, 1 Spanish, and 2 North African native dog breeds, including village dogs from Portugal
and Tunisia. Forty-nine haplotypes were found in the 164 individuals analyzed, with private haplotypes being found in several
breeds. For example, the Castro Laboreiro Watchdog, a rare breed from a small and isolated region in Portugal, was mono-
morphic for mtDNA and possessed a new haplotype, which may be provisionally considered a breed-specific marker. Phy-
logenetic analyses recapitulated 4 major clades identified in other studies, but new haplotypes, grouping within a clade that was
previously thought as geographically restricted, were detected in Portugal and Morocco. Portuguese village dogs showed no
genetic differentiation from nonnative dogs or from local breeds of the areas in which the village dogs were sampled. Although
Iberian and North African dog breeds possessed breed-specific mtDNA haplotypes, no significant geographic structure could
be detected among them. There is no evidence for introgression of North African haplotypes in Iberian dogs, contrary to
previous results for other domestic animals.
It is now well established that the ancestor of the dog, Canis
familiaris, is the gray wolf Canis lupus. Gray wolves and dogs
differ by only 4.6% (12 substitutions in a 261-bp fragment of
the D-loop) in their mitochondrial DNA (mtDNA) (Vila` and
others 1997), and this comparison is the lowest value be-
tween any pair of canid species. Other lines of evidence also
place the wolf as the dog ancestor, such as morphology
(Wayne 1986), behavior (Zimen 1981), and vocalization
(Lorenz 1975; Zimen 1981).
Currently, the domestic dog is the most abundant canid,
with a global population of around 400 million animals, and
has the widest geographic distribution of any domestic spe-
cies (Coppinger R and Coppinger L 2002). Among mammals,
domestic dogs are remarkable by the range of natural varia-
tion they exhibit in terms of morphological and behavioral
traits, even if the maintenance of dog breeds is guided by
established standards, such that within each breed animals
often show an uniform and distinctive morphology and be-
havior (Ostrander and Comstock 2004). There are more than
400 dog breeds worldwide, and their economic importance is
substantial (Clutton-Brock 1984).
Although dogs were present in what is now Portugal by
the Mesolithic (Cardoso 2002), the first breeds (probably
used for hunting) are known from mosaic depictions dating
from the Roman occupation (Braga 2000), and it is probable
that the Phoenicians had previously brought hound-type
dogs into the country (Veiga 2001). Currently, there are 10
working native dog breeds, recognized by the Portuguese
Kennel Club, that as a group demonstrate remarkable phe-
notypic diversity. Seven of these dog breeds are also recog-
nized by the Fe´de´ration Cynologique Internationale. These
breeds are thought to have a relatively recent origin although
their foundation dates are unknown. Breed standards date
mostly from the first half of the 20th century, and intensive
selection for standards started only 20–30 years ago.
Recent demographic changes are predicted to have im-
pacted the genetic diversity of Portuguese breeds. For exam-
ple, in the past, transhumance (twice-yearly migration of
Journal of Heredity
doi:10.1093/jhered/esl006
ª The American Genetic Association. 2006. All rights reserved.
For permissions, please email: journals.permissions@oxfordjournals.org.
1
Journal of Heredity Advance Access published July 3, 2006
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
livestock, shepherds, and their various livestock guarding dog
breeds [Martı
´
n and others 1994]), which has not occurred
in Portugal since the 1990s (Silva 2000) and in Spain since
the 1950s (Coppinger R and Coppinger L 2002), is likely
to have facilitated gene flow among livestock guarding
dog populations and breeds when they were brought into
contact during these seasonal migrations. Further, some Por-
tuguese dog breeds have gone through periods of critically
low population size (Vasconcelos 1995; Gomes 2003) due
to rural emigration, the abandonment of agricultural and
hunting practices, and major reductions of wolf populations
in some regions ICN (1997). With fewer wolves in the wild,
shepherds began using dogs with no specific skills for live-
stock guarding, and the subsequent decrease in the use of
livestock guarding dog caused a reduction in their number.
Finally, the popularity of dog shows created subsets of de-
mographically isolated dogs within some breeds (i.e., show vs.
working dogs). Portuguese livestock guarding dogs are a
good example of this, where both subsets have become dis-
tinct, not only morphologically (C. Cruz, personal commu-
nication) but also genetically (Petrucci-Fonseca and others
2000). Most of the demographic data collected from the Por-
tuguese Kennel Club refers only to show dogs, and therefore,
the number of registrations per year does not necessarily re-
flect the total population at any given time. Almost all breeds
show a common pattern, concerning the evolution of their
numbers, of an increase in the number of registries in the
Portuguese Kennel Club by the end of the 1970s (Gomes
2003).
Spain and the Maghreb region (Morocco, Tunisia, and
Algeria) are geographically proximate to Portugal and are
all within the same zoogeographic region: the Mediterranean
subregion of the Palaeartic (Roger and others 1998). Histor-
ically, the Iberian Peninsula has had a close connection with
North Africa mainly due to the Moorish occupation, which
lasted for 8 centuries (Brito and others 1992; Ribeiro and
Saraiva 2004). The Moors introduced crops and livestock
to Iberia (Cymbron and others 1999), and it is possible that
livestock guarding dogs were traded as well. In order to assess
genetic variation between domestic dog populations from
North Africa and Iberia, the 2 internationally recognized
North African dog breeds, the Aidi and Sloughi, and mongrel
dogs from Tunisia were included in our study.
MtDNA is a powerful tool for estimating levels of genetic
diversity, phylogenetic structure, and recent demographic
history in domestic animals, but its use beyond these appli-
cations is more limited (Bruford and others 2003), especially
because it only provides information about the female lineage
(Avise 2004). This limitation is a drawback when studying
domestic animals because male-mediated gene flow is usually
more pronounced among them. For example, the application
of mtDNA markers in domestic dogs (e.g., Okumura and
others 1996; Tsuda and others 1997; Vila` and others 1997;
Savolainen and others 2002) has showed little correspon-
dence between mitochondrial lineages, geographic structure,
or traditional breed classification, and many breeds contain
haplotypes shared by other breeds scattered over different
phylogenetic clades.
The most recent phylogenetic analysis of mtDNA in do-
mestic dogs, including samples from Europe, Asia, Africa,
and Arctic America, assigned mtDNA sequences to 6 clades:
A, B, C, D, E, and F (Savolainen and others 2002). At least
5 wolf matrilines are considered to be present at the origin of
the domestic dog population because all clades but F are
intermingled with wolf sequences. Clades A, B, and C contain
95.9% of the sampled dog haplotypes and are represented
in all geographic regions (clade A) or in all except America
(clades B and C). Geographically more restricted are the
clades D, E, and F, which, respectively, are found in Turkey,
Spain, and Scandinavia; Japan and Korea; and Japan and Siberia.
This is the first comprehensive study of mtDNA varia-
tion of Portuguese dogs and some other geographically ad-
jacent populations—the Spanish mastiff, Aidi, Sloughi, and
Tunisian mongrel. Table 1 summarizes the function, country
of origin, and conservation status of these breeds. The major
aim of this investigation was to determine the mtDNA ge-
netic variability of Portuguese dogs and to evaluate the geo-
graphic context of genetic variation in these and neighboring
dog populations.
Methods
Sampling and DNA Extraction
Blood, hair, and tissue samples from 164 animals belonging
to 13 different breeds or populations of dogs were collected
at dog shows, from breeding kennels, and from distinct loca-
tions in their historical breed regions (Table 1). Animals were
selected based on morphological standards and on informa-
tion about their ancestry in order to exclude related animals
back to the third generation whenever possible, although
complete information on ancestry was not always available.
In the Portuguese Warren hound, 3 body size types occur:
small, medium, and large. Hence, individuals from this breed
were sampled more extensively. Since the 1960s, one Portu-
guese Water dog breeder has exported dogs to outside of
Portugal, mainly to the United States (Molinari 1993), and
one animal from the United States is included, possibly rep-
resenting an extinct lineage in Portugal. Samples were also
collected from dogs kept in municipal kennels or animal shel-
ters, whose phenotypes could not be assigned to any recog-
nizable breed (hereafter Portuguese village dogs) as defined
by McDonald and Carr (1995). We selected those individuals
from diverse regions such as the Azores, where the Azores
cattle dog originated, the Estrela Mountain region, where the
Estrela Mountain dog originated, and Alentejo, where the
Alentejo Shepherd dog originated. In Tunisia, there are no
formally established dog breeds, and consequently, samples
were collected randomly within the local population. To min-
imize the risk of both the effects of genetic relatedness
among individuals and introgression from other breeds
and/or village dogs, we sampled for each breed, working ani-
mals with unknown ancestry only when from remote regions
within the historical range of each breed.
The number of registered females for each breed was col-
lected directly from Portuguese Kennel Club archives. The
Journal of Heredity
2
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Table 1. Information on the dog breeds used in this study. LGD, livestock guarding dogs; LHD, livestock herding dogs; HD, hunting dogs; WD, water dog. Harmonic means of
breeding females (2–8 years of age) were calculated taking into account the entire time span of records at the Portuguese Kennel Club for each breed. Demographic data only available for
Portuguese autochthonous dog breeds
Breed/population of dogs
Number of
samples Locality Function
Harmonic mean
female population
size (time span)
a
Current female
population size
Current
conservation
status
Total number
of haplotypes
b
Nonshared
haplotypes
b
New haplotypes
(global context)
c
Castro Laboreiro Watchdog
(CLWD)
14 Portugal LGD 6 (1932–2001) 435 Endangered 1 1 1
Estrela Mountain dog (EMD) 15 Portugal LGD 7 (1932–2001) 3692 Vulnerable 8 3 2
Alentejo Shepherd dog (ASD) 15 Portugal LGD 12 (1933–2001) 1303 Vulnerable 8 3 0
Azores cattle dog (ACD) 14 Portugal LHD 81 (1985–2001) 1468 Vulnerable 4 0 0
Portuguese Sheepdog (PSD) 8 Portugal LHD 57 (1954–2001) 414 Endangered 2 1 0
Portuguese Pointer (PP) 9 Portugal HD 43 (1932–2001) 2167 Vulnerable 2 0 0
Portuguese Warren hound
(PWH)
27 Portugal HD 44 (1932–2001) 1483 Vulnerable 13 6 4
Portuguese Water dog (PWD) 11 Portugal WD 34 (1946–2001) 932 Endangered 5 3 0
Spanish mastiff (SM) 10 Spain LGD ———621
Aidi 10 Morocco LGD ———841
Sloughi (SL) 10 Morocco HD ———652
Tunisia dogs (Tun) 9 Tunisia —— — — 620
Portuguese village dogs
(PortVillageDogs)
12 Portugal —— — — 10 7 2
Total 164 37 13
a
Time span refers to the time period since the first Portuguese Kennel Club records for each breed until 2001.
b
Total number of haplotypes and number of nonshared haplotypes detected in the Iberia and North Africa regarding the 887-bp mtDNA fragment.
c
New haplotypes based on the trimmed 554-bp mtDNA fragment described above.
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
3
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
number of potentially breeding females for each breed over
time was calculated as the harmonic mean of breeding
females since the first registries until the year 2001 (the ma-
jority of females only breed from the age of 2 up to 8 years
old) because this is the least biased stationary estimator when
comparing fluctuating, rapidly expanding or contracting,
populations (Frankham and others 2002). These calculations
are based on the Portuguese Kennel Club records only and
mainly involve show dogs. Currently, these are the most ac-
curate demographic data available due to a lack of a reliable
census of working dogs for each breed.
Blood samples (2 ml) were taken into vacutainers contain-
ing ethylenediaminetetraacetic acid (EDTA) (10% w/v) and
kept frozen until processed. Hairs, including the bulb, were
plucked (up to 40–50 per individual) and kept dry. Tissue
samples (ear biopsies) were preserved in dimethyl sulfoxide
(DMSO)–salt buffer (20% DMSO, 0.25 M Na–EDTA, and
NaCl to saturation, pH 8.0) at 20C. DNA was extracted
from whole blood and tissue according to standard protein-
ase K/Phenol–Chloroform protocols (Sambrook and Russel
2001), followed by an extra ethanol precipitation step, or us-
ing the Nucleospin Blood QuickPure kit (Macherey-Nagel,
Du
¨
ren, Germany) following the supplier’s recommendations.
DNA was extracted from hair bulbs in a Chelex 20% solution
(protocol adapted from Walsh and others 1991).
DNA Amplification, Purification, and Sequencing
An mtDNA fragment of 887 bp, comprising a segment of the
cytochrome b, the tRNA-Thr, the tRNA-Pro, and a segment
of the control region, was amplified using a single pair of
primers (all primers are numbered in relation to the complete
mtDNA sequence determined by Kim and others 1998):
L15210 5#-ACA TGA ATT GGA GGA CAA CCA GT-3#
(a shortened version of Shields and Kocher’s [1991]
L15774 primer) and H16097 5#-TAT GTC CTG TGA
CCA TTG ACT GA-3# (S. Funk, Institute of Zoology, Lon-
don). L15210 plus the following 5 internal primers were used
for sequencing: H15377 5#-TTT GAG TCT TAG GGA
GGG CG-3# (designed by A. E. Pires), L15360 (Hoelzel
and others 1991), L15515 5#-GTG TCA GTA TYT CCA
GGT-3# (designed by A. E. Pires), L15805 (Southern and
others 1988), and H16075 5#-GCA CCT TGA TYT TAT
GCG T-3# (designed by A. E. Pires).
Polymerase chain reaction (PCR) amplifications were
performed in a 25-ll reaction volume with 100–200 ng of
DNA from blood and tissue samples or 8–10 ll of extract
from hair bulbs, as template, 1 U of Taq polymerase (Gibco
[now Invitrogen], Paisley, UK), 1 reaction buffer (Gibco
10 stock contains 200 mM Tris–HCl pH 8.4 and 500 mM
KCl), 1.5–2 mM MgCl
2
(Invitrogen, Paisley, UK), 20 lMof
each deoxynucleoside triphosphate (ABgene, Epsom, UK),
0.5 lM of each primer, and Sigma water. Nonacetylated bovine
serum albumin (MBIFermentas, Vilnius, Lithuania) was in-
cluded at a concentration of 0.024 lg/ll for amplification
of blood-extracted DNA and 0.48 lg/ll for hair bulbs-
extracted samples. Amplification reactions were performed
in a GeneAmp PCR System 9700 (Perkin Elmer) thermal cy-
cler, with an initial denaturation step of 94 C for 3 min;
followed by 40 cycles (60 cycles for hair samples) of dena-
turation at 94 C (40 s), annealing at 55 C (30 s), and exten-
sion at 72 C (1 min); followed by a final extension step at
72 C for 7min. For hair samples, we used a ramp between
the annealing and extension steps at 50% of the maximum
speed (programmable in the thermocycler) to make the tem-
perature transition slower and increase the number of stable
primer–template complexes undergoing extension in each cy-
cle. A negative control was included in each set of amplifi-
cations. PCR products were separated on 1.5% agarose gels,
cleaned using the GeneClean Turbo kit (Qbiogene, Carlsbad,
CA), and eluted in 35 ll of elution solution. Purified products
were sequenced in both directions using the ABIPrism
BigDye Terminator Cycle Sequencing Ready Reaction Kit
(version 2.0) and following the manufacturer’s instructions.
Products were separated on a semiautomated DNA analyzer
(ABI 377) and sequences were edited, assembled, and aligned
using the program SEQUENCHER 3.1.2 (Gene Codes
Corporation) and submitted to GenBank using SEQUIN
(http://www.ncbi.nlm.nih.gov/Sequin/).
Population and Phylogenetic Analyses
DNA Polymorphism
Arlequin version 2.0 (Schneider and others 2000) was used to
compute indices of genetic diversity. Haplotype diversity H
and nucleotide diversity p were determined for each popula-
tion. Pairwise genetic distances between haplotypes were cal-
culated under the Tamura–Nei model (Tamura and Nei 1993),
after excluding positions with insertion/deletions. Heteroge-
neity of substitution rates per site across the D-loop region
was taken into account and gamma distribution parameters
(Yang 1994) estimated with the best-fitting model of sequence
evolution, as determined by MODELTEST version 3.06
(Posada and Crandall 1998). Pairwise F
ST
values were estimated
and significance at the 5% level determined with 10 000 per-
mutations. Arlequin was also used to carry out an analysis of
the molecular variance (AMOVA; Excoffier and others 1992).
Three groups were defined taking into account the geographic
origin of the samples: Portugal, Spain, and North Africa.
Sequences with no missing data were collapsed into hap-
lotypes using the program COLLAPSE (Version 1.0), avail-
able at http://darwin.uvigo.es. All haplotypes were compared
with the DNA sequence information stored in the GenBank
database. This was conducted using the ‘‘basic local alignment
search tool’’ (BLAST) program (Altschul and others 1990),
available through the National Center for Biotechnology
Information (Bethesda, MD). Chiperm version 1.2 (Posada
2000) was used to test for significant associations between
haplotypes and breeds or dog populations using the algorithm
developed by Hudson and others (1992). The significance of
the v
2
statistics was approximated by Monte Carlo simulation
(Roff and Bentzen 1989) by permuting the contingency
tables; 100 000 permutations were performed.
Phylogenetic relationships among haplotypes (887 bp)
were estimated using the neighbor-joining (NJ) method in
PAUP version 4.0b10 (Swofford 2002). In order to analyze
4
Journal of Heredity
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
the sequences reported in this study in a wider context, 54
haplotypes of 554 bp (accession numbers: AF531664,
AF531668, D83609, AF531670, AF531671, AF531672,
AF531674, AF531675, D83606, AF531679, AF531682,
AF531685, AF531687, AF531656, AF531693, AF531696,
AF531700, AF531702, AF531710, AF531658, AF531731,
AF531732, AF531733, AF531734, D83607, D83625,
D83634, AB007402, AF531722, AF531723, AF531724,
AF531725, AB007380, D83601, AF531728, AF531729,
AF531730, AF531715, D83636, AF531717, AF531718,
AF531719, AF531720, AF531721, AF531735, AF531736,
AF531737, AF531738, AF531740, AF531739, AF531741,
D83611, D83637, AB007381) were imported from GenBank
to represent the 6 major dog clades described by Savolainen
and others (2002). These haplotypes represent 20 of the 75
haplotypes described by Savolainen and others (2002) for
clade A and all haplotypes described for clades B, C, D,
E, and F. Haplotypes from wolves, golden jackal, and coyote
were also included (accession numbers: AF008140,
AF008137, AF008142, AF008135, AF008138, AF008139,
AF008141, AF184048, AF008158). From the 3 Iberian wolf
haplotypes described by Vila` and others 1997 (W1, W2, and
W3), we selected W1 as a representative of wolves from Ibe-
ria. Among W1, W2, and W3, there are only 5 variable posi-
tions out of 261. In order to construct an NJ tree integrating
sequences from the studies mentioned above, the 887-bp
fragments generated in this study were trimmed to 554 bp.
The best-fitting model of sequence evolution was determined
by MODELTEST version 3.06 (Posada and Crandall 1998).
The reliability of the nodes was assessed with 1 000 bootstrap
iterations (Felsenstein 1985).
An intraspecific gene genealogy was inferred using the
median-joining network algorithm (Bandelt and others 1999)
in NETWORK, version 4.1, available at http://www.
fluxus-engineering.com. Gaps were treated as evolutionary
events, and data were analyzed with all characters weighing
equally. The tolerance parameter e was set to 0. Haplotype
frequencies were converted into proportional areas in the
graphical output.
Results
Genetic Diversity and Differentiation
The primers L15210/H16097 amplified a sequence compris-
ing part of the cytochrome b gene (112 bp), the tRNA-Thr
and tRNA-Pro genes (135 bp), and the hypervariable portion
at the 5# end of the control region (640 bp). Forty-nine dif-
ferent haplotypes were obtained from 164 individuals (acces-
sion numbers: AY706476–524) (Figure 1). Forty-eight
phylogenetically informative sites were found (5% of the total
fragment), and, of these, only positions 4, 56, 145, 162, 206,
and 225 are outside the control region. Changes corre-
sponded to synonymous transitions. A strong transitional
bias of 39.2:1 was observed, a common finding in the
mtDNA of mammals (Graur and Li 2000). Due to indels,
3 sites contained gaps (positions 255, 722, and 729). Nucleotide
frequencies for the entire fragment were A 5 0.28420,
C 5 0.26180, G 5 0.15250, T 5 0.30150. The low guanine
(G) content is a common result in vertebrate mtDNA (e.g.,
Tamura and Nei 1993).
Of the 49 haplotypes found in this study, a BLAST search
conducted against the sequences in GenBank revealed that 40
are new or unique haplotypes. Differences between the 40
new haplotypes and the dog sequences deposited in Gen-
Bank ranged from 1 to 15 nucleotides in 887 bp (0.1–
1.7%). Twelve of the 49 haplotypes found in this study
are shared among the 13 dog populations sampled, and
the remaining 37 haplotypes are breed or dog population spe-
cific (Table 1). The total number of haplotypes found per
breed or dog population (shared and nonshared) varies be-
tween 1 and 13 (Table 1). The number of nonshared haplo-
types, that is, haplotypes that are only found in 1 of the 13
dog groups of this study varies between 0 and 7 (Table 1).
Within the sampled individuals, 11 haplotypes, 4 from Aidi,
5 from Sloughi, and 2 from Tunisian dogs, were only found
in dogs from North Africa (Table 1). With the exception of
the Portuguese Warren hound, Aidi and Sloughi were the
breeds showing the highest number of nonshared haplotypes
(4 and 5, respectively, Table 1). Portuguese village dogs
showed the highest value of haplotypic diversity (0.97 ±
0.04), followed by Aidi (Table 2). Among the remaining
breeds, only the Portuguese Warren hound and Alentejo
Shepherd dog showed haplotype diversities higher than
0.90. Haplotype diversity was not correlated to sample size
(R
2
5 0.0324, graph not shown). Among the endangered
Portuguese native dogs, Portuguese Water dog had a relatively
high haplotypic diversity (0.82 ± 0.08).
Nucleotide diversity per site (p ± SD) was on average
high, ranging from 0.014 ± 0.007 (Estrela Mountain dog)
to 0.002 ± 0.001 (Portuguese Pointer), with the exception
of the Castro Laboreiro Watchdog that showed no diversity
(Table 2). This statistic is not highly sensitive to sample size,
as already noted by Vila` and others (1999). Castro Laboreiro
Watchdog showed a single and private or breed-specific hap-
lotype (H25) in the 14 analyzed animals, of which 11 were
working dogs and 3 were show dogs. Estrela Mountain
dog, Alentejo Shepherd dog, and Spanish mastiff showed
similar values of genetic diversity. It is noteworthy that high
levels of nucleotide diversity were also found in the Northern
African dogs (Table 2).
Genetic differentiation among breeds and dog popula-
tions was statistically significant (v
2
5 1177.675, P ,
0.001). Pairwise F
ST
ranged from nearly 0 (between Aidi
and Portuguese Village dogs) to 0.801 (between Castro
Laboreiro Watchdog and Portuguese Pointer) (Table 2). Pair-
wise F
ST
values between Portuguese village dogs and dogs
from outside of Portugal (Spanish mastiff, Aidi, Sloughi,
and Tunisia dogs) were near 0. Castro Laboreiro Watchdog
and Portuguese Sheepdog were the only breeds with all pair-
wise F
ST
values statistically significant. Castro Laboreiro
Watchdog, Portuguese Sheepdog, Portuguese Pointer, and
Portuguese Water dog, the least diverse breeds, showed
the greatest genetic differentiation values from the Portu-
guese village dogs (significant at the 5% level), which repre-
sent the ‘‘background’’ mtDNA variation of dogs in Portugal.
5
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Within the livestock guarding dogs group (Castro Laboreiro
Watchdog, Estrela Mountain dog, Alentejo Shepherd dog,
Spanish mastiff, and Aidi), Castro Laboreiro Watchdog
was the most isolated, with no mtDNA variation and signif-
icant F
ST
values for all breed comparisons (a 5 0.05) (all F
ST
values 0.396). Estrela Mountain dog and Alentejo Shep-
herd dog showed little genetic differentiation from each
other. Spanish mastiff showed substantial differentiation
from all Portuguese livestock guarding dogs, although to
a much lesser extent from Estrela Mountain dog and Alentejo
Shepherd dog. F
ST
between Spanish mastiff and Aidi (the only
African livestock guarding dog in this study) was not signif-
icant (0.040). The insular Azores cattle dog shows the least
differentiation from the Alentejo Shepherd dog (F
ST
5 0.027
and not significant). Portuguese Warren hound showed sig-
nificant F
ST
values with the breeds Portuguese Sheepdog,
Castro Laboreiro Watchdog, Portuguese Pointer, and Azores
cattle dog. All F
ST
values among Portuguese Warren hound
subpopulations (varieties) were near 0 (data not shown).
F
ST
values between each breed and the dogs that compose
the background genetic population (Portuguese village dogs)
of their region ranged from 0.066 to 0.073, and none were
significant (data not shown). AMOVA showed subdivision
between breeds (U
ST
5 0.171, P , 0.0001), with around
80% of the variation found within populations (breeds). Var-
iation among breeds accounted for practically 20% of the to-
tal genetic variability found. No variation could be attributed
to geographic structure for the entire data set.
Phylogenetics
The best-fitting model of sequence evolution determined by
MODELTEST (hierarchical likelihood ratio test statistic) was
the HKY85 þ I þ G model (Hasegawa and others 1985) with
gamma distribution correction for rate heterogeneity among
sites. The shape parameter of the distribution (a) and propor-
tion of invariable sites (I) were 0.6306 and 0.8646, respec-
tively. Only clades with bootstrap values equal or higher
than 70% are considered here (see Hillis and Bull 1993).
Figure 2 shows an NJ unrooted phylogram of all haplo-
types found in the samples of this study (887 bp). Mitochon-
drial lineages are not clustered by geographic origin or
traditional breed classification: the same haplotypes occur
in different populations and are scattered throughout differ-
ent clades. Forty-nine percent of the haplotypes found in this
study were in clade A and with all breeds represented, al-
though with weak bootstrap support. Clades B, C, and D
are supported by a high bootstrap value (83%) and include
all breeds except Castro Laboreiro Watchdog (H25). Clade
D, previously described as being regionally restricted to
Turkey, Spain, and Scandinavia (Savolainen and others
2002), includes Estrela Mountain dog and Alentejo Shepherd
dog haplotypes from Portugal (H38 and H40) and Aidi hap-
lotypes from Morocco (H2 and H6) (AY706513, AY706515,
AY706477, and AY706481 respectively) (Figures 2 and 3).
Few nodes showed significant bootstrap values, but most
major partitions are well supported. When 54 domestic
dog haplotypes from Savolainen and others (2002) and wild
canid sequences from Vila` and others (1997) (fragments of
554 bp) were included, the structure of the NJ tree is very
similar to that obtained by Savolainen and others (2002) (data
not shown). We recovered all clades except E and F. Regard-
ing the 554-bp sub–data set of the sequences generated in this
study, we identified one new haplotype for clade D
(AY706481) obtained from Morocco. For the other clades,
12 new haplotypes were found: in clade A, 5 (Portugal and
Morocco, H19, H25, H41, H42, H44); clade B, 3 (Portugal
and Morocco, H30, H34, H45); and clade C, 4 (Portugal and
Spain, H16, H27, H29, H32) (AY706494, AY706500,
AY706516, AY706517, AY706519, AY706505, AY706509,
AY706520, AY706491, AY706502, AY706504, AY706507,
respectively). The following haplotypes were breed or
population-specific within this study: Castro Laboreiro
Watchdog, 1 (H25); Estrela Mountain dog, 2 (H41, H42);
Portuguese Warren hound, 4 (H29, H30, H32, H34); Spanish
mastiff, 1 (H27), Aidi, 1 (H6); Sloughi, 2 (H44, H45) and Por-
tuguese village dogs, 2 (H16, H19) (Table 1). The W1 Iberian
wolf haplotype grouped with dog haplotypes from clade B,
along with haplotypes from all breeds except Castro
Laboreiro Watchdog, Portuguese Sheepdog, Aidi, and Por-
tuguese Pointer. The C. aureus and C. latrans sequences were
both highly distinct, as expected.
The haplotype network (Figure 3) shows the same rela-
tionships as the NJ tree (Figure 2), but more detail is evident.
Eight missing haplotypes (median vectors, mv) were detected.
In clade A, all breeds are represented, but 4 median vectors
are also included. The most frequent haplotypes, H7 and H10
(AY706482 and AY706485), respectively, are present in 16
(7 dog populations) and in 19 (6 dog populations) animals.
These frequent haplotypes are connected to other haplo-
types, from which they mostly differ by singletons. Haplo-
type H25 (AY706500) in clade A is only one mutational
step (transition in position 422) different from haplotype
H28 (AY706503). Haplotype H28 is shared by the Portu-
guese Pointer, Alentejo Shepherd dog, and Portuguese War-
ren hound breeds. Clade A is connected to Clade C by 11
mutational steps, including 2 median vectors. Clade C is con-
nected to Clade B by the same number of steps. In the net-
work, Clade D comprises 4 divergent lineages and is
separated from the rest of the haplotypes by 8 steps. This
clade comprises only 3 breeds in this study, whereas clades
B and C represent 9 and 8 dog populations, respectively.
Figure 1. Base substitutions and gaps found among 164 dog mitochondrial sequences. Forty-nine haplotypes are listed. Dashes (-)
indicate gaps, and dots (.) indicate matches regarding the reference nucleotide sequence (Haplotype 10, the most frequent).
Position 1 corresponds to the base 15211 on the cytochrome b gene in the mitochondrial genome sequence of Kim and others
(1998). Nucleotide sequence data from this study have been assigned the GenBank accession numbers AY706476–524.
7
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Haplotypes H7, H10, H21, and H5 were shared between Por-
tuguese and African breeds.
Discussion
Genetic Diversity and Differentiation
Because most haplotype diversities were high (Table 2), we
would expect the v
2
test implemented in Chiperm to be pow-
erful in detecting genetic differentiation between breeds or
populations (Hudson and others 1992). Genetic differentia-
tion was especially apparent due to the number of private
haplotypes found in several breeds. The lack of differentia-
tion between Portuguese village dogs and dogs from outside
Portugal (Spanish mastiff, Aidi, Sloughi, and Tunisia dogs) is
probably a consequence of the high diversity found in all
these breeds and/or populations.
Portuguese Warren hound subtypes (varieties) are not dif-
ferentiated using mtDNA. Portuguese Water dog was listed
in the Guinness Book of World Records in 1970 as the
world’s rarest dog breed (Molinari 1989) and has come close
to extinction twice in the last century, with a total of only 50
animals in 1974 (Molinari 1993). However, the genetic diver-
sity of Portuguese Water dog is surprisingly high. In Portu-
guese Kennel Club books, registrations for this breed have
been less than 50 animals per year for 40 years and have never
exceeded 450 animals per year. Although this breed is very
popular in the United States, only one individual could be
sampled from there (a dog imported from the United States
and living in Portugal), and other Portuguese Water dogs liv-
ing in Portugal share its haplotype (H11). Further efforts
need to be undertaken to test more individuals from lineages
exported to the United States from Portugal in the 1960s
(Molinari 1993). Concerning Castro Laboreiro Watchdog,
even though it is a livestock guarding dog like Estrela Moun-
tain dog, Alentejo Shepherd dog, and Spanish mastiff, it was
never involved in the large transhumance movements in
which all other livestock guarding dog breeds participated
(Geraldes 1996) and has been isolated for long periods. It
was, instead, involved in small, geographically restricted
migrations within a remote area, and contact with other
dog populations is likely to have been scant. This is a possible
explanation for its mtDNA uniqueness. Although working
animals lacking the breed standard phenotype can be found
in the area, these animals were not sampled for this study.
However, they might be a possible addition to future analy-
ses. Expected heterozygosity values inferred using microsat-
ellites loci indicate that Castro Laboreiro Watchdog is one of
the least diverse breeds in Portugal (Amaro 2001; AE Pires,
F Simo
˜
es, F Petrucci-Fonseca, and MW Bruford, in prepara-
tion). Within the dogs sampled for this study, position 422
shows a transition (A to G) that separates Castro Laboreiro
Watchdog from all other dogs. Therefore, this nucleotide can
provisionally be considered a breed-specific marker in the
study area. As can be seen in the network (Figure 3), this hap-
lotype differs by one base from a more common haplotype
present in several dog breeds (Portuguese Pointer, Portu-
guese Warren hound, and Estrela Mountain dog). Probably
Table 2. Haplotype diversity, nucleotide diversity, and average pairwise differences between haplotypes within populations (underlined values) and F
ST
values (below the diagonal).
Significant values at the 5% level are in bold. Abbreviations of breed names as in Table 1
CLWD EMD ASD ACD PSD PP PWH PWD SM Aidi SL Tun
PortVillage
Dogs
Haplotype
diversity
(H ±SD)
0.00 ± 0.00 0.90 ± 0.05 0.92 ± 0.04 0.66 ± 0.09 0.25 ± 0.18 0.50 ± 0.13 0.91 ± 0.04 0.82 ± 0.08 0.84 ± 0.10 0.96 ± 0.06 0.89 ± 0.08 0.83 ± 0.13 0.97 ± 0.04
Nucleotide
diversity
(p ±SD)
0.000 ± 0.000 0.014 ± 0.007 0.010 ± 0.006 0.009 ± 0.005 0.004 ± 0.002 0.002 ± 0.001 0.012 ± 0.006 0.010 ± 0.006 0.008 ± 0.005 0.011 ± 0.006 0.011 ± 0.006 0.007 ± 0.004 0.011 ± 0.006
CLWD
0.00 ± 0.00
EMD 0.472
12.26 ± 5.87
ASD 0.513 0.055
9.25 ± 4.51
ACD 0.550 0.136 0.027
8.33 ± 4.11
PSD 0.799 0.269 0.227 0.296
3.38 ± 1.93
PP 0.801 0.317 0.359 0.401 0.415
1.53 ± 1.01
PWH 0.403 0.039 0.055 0.093 0.135 0.241
10.420 ± 4.90
PWD 0.674 0.058 0.199 0.278 0.410 0.504 0.038
8.71 ± 4.36
SM 0.473 0.148 0.152 0.118 0.223 0.197 0.103 0.320
6.87 ± 3.53
Aidi 0.473 0.014 0.053 0.102 0.218 0.191 0.070 0.223 0.040
9.54 ± 4.79
SL 0.460 0.048 0.138 0.173 0.215 0.187 0.016 0.098 0.058 0.020
9.90 ± 4.95
Tun 0.518 0.163 0.228 0.230 0.237 0.103 0.109 0.305 0.014 0.035 0.00
5.92 ± 3.12
PortVillageDogs 0.396 0.071 0.081 0.062 0.125 0.144 0.013 0.175 0.034 0.004 0.018 0.016
9.79 ± 4.82
8
Journal of Heredity
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Castro Laboreiro Watchdog as it is today, an mtDNA mono-
morphic dog breed, is the result of a severe bottleneck from
a more diverse ancestral population of Castro Laboreiro
Watchdogs. Considering all haplotypes from Savolainen
and others (2002) and Vila` and others (1997), the guanine
in position 422 is found in one dog from East Asia (isolate:
A56, accession number: AF531707). However, the Castro
Laboreiro Watchdog haplotype remains unique when com-
pared with this East Asian dog due to a difference in another
nucleotide position (position 430). Although less likely,
this haplotype may represent a lineage tracing back to the
livestock guarding dogs dispersal from Southwest Asia (as-
sociated with human migrations), where livestock was first
domesticated (Mannion 1999).
The low levels of genetic differentiation between Estrela
Mountain dog and Alentejo Shepherd dog are expected based
on historical information: these breeds have a probable com-
mon origin, and until recently both have been in close contact
due to the transhumance practice (Alpoim 1999). In contrast,
genetic differentiation between Spanish mastiff and both
Estrela Mountain dog and Alentejo Shepherd dog is more
pronounced as genetic exchange that might have occurred
during the long period of extensive transhumance is likely
to have been more restricted. At the origin of the Azores
cattle dog, the Portuguese native breeds Castro Laboreiro
Watchdog, Estrela Mountain dog, and Alentejo Shepherd
dog are considered founders (Amaral and Veiga 2004).
Our results only support a major contribution by the Alentejo
Shepherd dog because the F
ST
value between the Azores
cattle dog and the Alentejo Shepherd dog is the lowest value
and is not significant.
It is interesting to note that livestock guarding and herd-
ing dog breeds, namely Estrela Mountain dog, Alentejo Shep-
herd dog, and Azores cattle dog, do not have a statistically
significant divergent mtDNA composition from that of
the ‘‘background’’ gene pool in the region where they are
from, probably due to the high haplotypic diversity revealed
by the Portuguese village dog population. Finally AMOVA
results suggest that, although breeds share haplotypes, there
is breed differentiation. Further sampling in Spain and North
H18
H36
H41
H13
H4
H42
H28
H7
H20
H10
H47
H14
H35
H48
H8
H43
H44
H1
H19
H46
H24
H15
H21
H40
H6
H38
H2
H34
H9
H49
H17
H45
H31
H30
H11
H12
H33
H27
H5
H39
H16
H29
H32
H23
H3
H37
0.0005 substitutions/site
H25
H26
H22
83
82
84
99
92
Figure 2. Unrooted NJ phylogram of dog haplotypes based on 887 bp, comprising partial sequences of cytochrome b, tRNA-
Thr, tRNA-Pro, and control region. Letters A–D indicate clades consistently supported with different tree-building methods (data
not shown). Support values are indicated at nodes when found in at least 70% of 1000 bootstrap replicates.
9
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Africa may shed more light on the geographic structure of
Iberian and North African dog breeds.
The high levels of diversity in African dogs could be
explained by their antiquity and/or reported introgression.
Both Aidi and Sloughi are considered rare (Miguil 1986; Ouar
1994). At the beginning of the 20th century, most of the dog
population in the Atlas mountainous area comprised Aidi
dogs. In contrast, by 1994, only 49% of the dog population
corresponded to Aidi breed individuals, and in parallel with
a reduction in numbers, there was documented introgression
with hunting dogs (Ouar 1994). This breed may also be less
intensively managed than many European breeds. The breed
standard was described in 1969, but the majority of animals
are free ranging for guarding purposes (Ouar 1994), making
possible introgression with feral dogs and/or other dog
breeds. The date of origin of the Sloughi remains speculative,
but it is considered an ancient breed. Representatives of Af-
rican Sighthound-like dogs date back to more than 7000 years
(Giudicelli 1975). At present, it is mainly used to hunt desert
hares and gazelles and indirectly to protect goat and sheep
herds by hunting their predator, the golden jackal (Miguil
1986). A number of factors have contributed to its present
low population size, including a decrease of hunting activities
using dogs, the popularity of firearms, and a reduction in its
geographic distribution (Miguil 1986). Finally, the high ge-
netic diversity observed in Tunisian dogs is expected because
the dogs sampled belong to the local village dog population
from no specific breed. Tunisian dogs show considerably less
differentiation from the other African dogs and from the
Spanish mastiff than from the Portuguese native dog breeds.
African domestic dogs were found to have distinct haplo-
types and haplotype frequencies from Portuguese dogs,
and although some haplotypes are shared, no south–north
cline in frequency could be detected in our data as found
in cattle, for which admixture was detected by Cymbron
and others (1999).
Figure 3. Median-joining network (e 5 0) depicting phylogenetic relationships among, and breed assignment of, all dog breed
mtDNA haplotypes from this study. Circled areas are proportional to the corresponding haplotype frequency. If individuals
of different breeds shared one haplotype, a pie diagram illustrates the respective proportions. Reticulations are observed.
Median vectors (mv) produced by the network software, representing missing or not sampled haplotypes, are illustrated by
black dots.
10
Journal of Heredity
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Phylogenetic Analysis
The lack of an obvious underlying geographic or breed-
related structure in the NJ tree could indicate that most of
the breeds are derived from the same ancestral stock and re-
tain many of the ancestral maternal haplotypes and/or that
extensive introgression has occurred in the process of breed
formation and development. Similar results were obtained by
Tanabe and others (1991) based on blood proteins; Okumura
and others (1996), Tsuda and others (1997), Vila` and others
(1997), and Savolainen and others (2002) based on mtDNA
control region; and Jordana and others (1999) using morpho-
logical and behavioral characteristics. Another explanation
may be that most breeds have originated too recently, and
mutations that account for breed differentiation have yet
to occur and become fixed. Other researchers have found
dog breeds to be highly differentiated. For example, Parker
and others (2004) sampling purebred dogs from closed gene
pools found that breeds could be genetically differentiated
using biparentally inherited microsatellites markers. How-
ever, by sampling only 5 unrelated pedigreed animals from
each breed, the authors did not carry out an extensive and
representative sampling across each breed, thereby maximiz-
ing the probability of detecting a signal of breed structure.
Most of the haplotypes in this study clustered within
the 3 major phylogenetic clades A, B, and C described in
Savolainen and others (2002), implying that Iberian dogs
are unlikely to be the result of an additional domestication
event. All clades found by Savolainen and others (2002) were
extensively represented in the phylogenetic analysis, although
the proportion in clade A (26.6%) was reduced relative to
their study. Therefore, the 8 breed-specific haplotypes found
in Iberian and North African dogs for clades B, C, and D
represent additional diversity in these clades. This unique
diversity is potentially important information to the conser-
vation of domestic dog genetic variation in this region.
The application of single-genetic locus surveys is not
without problems (Avise 2004), and this work should be aug-
mented by neutral nuclear DNA markers and genes associ-
ated with the phenotypes under selection. Currently, studies
based on nuclear DNA markers such as amplified fragment
length polymorphisms and autosomal microsatellites are on-
going (AE Pires, F Simo
˜
es, F Petrucci-Fonseca, and MW
Bruford, in preparation). MtDNA does not account for
male-mediated gene flow, a dominant force in domestic an-
imal evolution (Petit and others 2002). Male dog microsatel-
lites loci from the nonrecombining portion of the canine Y
chromosome are already available (Olivier and others 1999;
Bannash and others 2005), and their analysis among breeds
revealed a high breed and geographic structure (Bannash and
others 2005). Similarly, it is likely that the study of the Y chro-
mosome can bring additional detail to the phylogenetic anal-
ysis of the domestic dog breeds analyzed here. Other possible
avenues of research involve the analysis of functional genes.
For example, Fondon and Garner (2004) studying the associ-
ation of DNA mutations, in particular intragenic tandem
repeats in developmental genes, with phenotypic variability,
allowed identification of a possible molecular mechanism
responsible for morphological and functional variation. Neff
and others (2004) by using a pharmacogenetic mutation as
a molecular marker were able to unravel the emergence of
formally recognized breeds. Another example is the work
of Vila` and others (2005) who analyzed variation in genes
from the major histocompatibility complex (MHC) poten-
tially under balancing selection to infer past relationships be-
tween domestic mammals and their wild ancestors. The
authors found high levels of genetic diversity suggesting that
domestic mammals might have acquired much of their cur-
rent genetic diversity by backcrossing several times with their
wild ancestors. Typing nuclear genes from the MHC in each
of the breeds found to contain unique mtDNA haplotypes in
our study would probably reveal additional MHC alleles.
Implications for Conservation
Dog breeds should be viewed as dynamic populations, in
which historic phenomena such as admixture, introgression,
genetic isolation, and drift have occurred (Neff and others
2004). Portuguese dog breeds have experienced major demo-
graphic fluctuations over the past 20 generations, with a con-
sistent decline through disuse of breeds in a working context,
followed by a rapid expansion in many cases due to the fash-
ion of showing dogs and for the pet trade. Concerning its
mtDNA, Iberian and North African breeds show, in general,
considerable diversity, including the possession of some
newly described sequences. In this study that focused on
marginal working dog breeds, several haplotypes emerge
as novel, and conservation planning for these breeds should
therefore take into account the unique genetic characters
highlighted here. Following Weitzman’s concept for live-
stock that involves the use of a ‘‘diversity function’’ (Bruford
and others 2003), Castro Laboreiro Watchdog and Portu-
guese Sheepdog would become priorities for conservation
on the basis of mtDNA alone. The considerable mtDNA di-
versity in the dogs of Iberia and North Africa prevents, with
one exception, its use as a marker for breed designation,
and these dog breeds should be studied with several other
kinds of nuclear markers available to accurately assess breed
distinctiveness.
Acknowledgments
The research was supported by Fundacxa
˜
o para a Cie
ˆ
ncia e a Tecnologia (grant
PRAXIS XXI/BD/21677/99) and AGRO (project 31106). We thank Carla
Cruz, Silvia Ribeiro, the dog owners, and breed clubs who contributed to this
study; Margarida Gomes for helping collecting data from the Portuguese
Kennel Club; Carlos Fernandes for advice in the laboratory and data analyses;
Isabel Amorim, Carlos Fernandes, 2 anonymous reviewers, and the corre-
sponding editor for helpful discussions and comments on the manuscript.
We also thank the Portuguese Kennel Club for facilitating access to the da-
tabase on dog registries.
References
Alpoim JA. 1999. O Rafeiro do Alentejo. Monografia da Racxa [Alentejo
Shepherd Dog. Breed monography]. Monforte, Portugal: Ca
ˆ
mara Municipal
de Monforte.
11
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local
alignment search tool. J Mol Biol 215:403–10.
Amaral AJ, Veiga V. 2004. O Ca
˜
o de Fila de S. Miguel [The Azores cattle
dog]. Vila Franca do Campo, Portugal: Ca
ˆ
mara Municipal de Vila Franca do
Campo.
Amaro A. 2001. Contribuicxa
˜
o para a Caracterizacxa
˜
o da Estrutura Gene´tica de
Racxas Caninas Auto´ctones Portuguesas. Estudo do Grupo de Ligacxa
˜
oda
Transferrina em Racxas Caninas Tipo Mastim [Contribution to the genetic
characterization of the Portuguese autochthonous dog breeds. Study of
transferrin ligation group in mastiff type breeds] [MSc dissertation]. Porto,
Portugal: Universidade do Porto.
Avise JC. 2004. Molecular markers, natural history and evolution. Sunder-
land, MA: Sinauer Associates, Inc.
Bandelt H-J, Forster P, Ro
¨
hl A. 1999. Median-joining networks for inferring
intraspecific phylogenies. Mol Biol Evol 16:37–48.
Bannash DL, Bannash MJ, Ryun JR, Famula TR, Peterson NC. 2005.
Y chromosome haplotype analysis in purebred dogs. Mamm Genome 16:
273–80.
Braga PD. 2000. Histo´ria dos Ca
˜
es em Portugal. Das Origens a 1800 [Dogs’
history in Portugal from origin to 1800]. Lisboa, Portugal: Hugin Editores,
Lda.
Brito RS, Fabia
˜
o C, Macı
´
as S, Mattoso J, Torres C. 1992. Histo´ria de Portu-
gal. Antes de Portugal [History of Portugal. Before Portugal]. Volume 1.
Lisboa, Portugal: Cı
´
rculo de Leitores, Lda and Authors.
Bruford MW, Bradley DG, Luikart G. 2003. DNA markers reveal the com-
plexity of livestock domestication. Nat Rev Genet 4:900–10.
Cardoso JL. 2002. Pre´-Histo´ria de Portugal [Pre-history of Portugal]. Lisboa,
Portugal: Editorial Verbo.
Clutton-Brock J. 1984. Dog. In: Mason IL, editor. Evolution of domesticated
animals. London, UK: Longman.
Coppinger R, Coppinger L. 2002. Dogs: a new understanding of canine or-
igin, behavior, and evolution. Chicago, IL: University of Chicago Press (Trd).
Cymbron T, Loftus RT, Malheiro MI, Bradley DG. 1999. Mitochondrial
sequence variation suggests an African influence in Portuguese cattle. Proc
R Soc Lond B Biol Sci 266:597–603.
Excoffier L, Smouse P, Quattro JM. 1992. Analysis of molecular variance
inferred from metric distances among DNA haplotypes: application to hu-
man mitochondrial DNA restriction data. Genetics 131:479–91.
Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the
bootstrap. Evolution 39:783–91.
Fondon JW III, Garner HR. 2004. Molecular origins of rapid and continuous
morphological evolution. Proc Natl Acad Sci USA 101:18058–63.
Frankham R, Ballou JD, Briscoe DA. 2002. Introduction to conservation
genetics. Cambridge, UK: Cambridge University Press.
Geraldes AD. 1996. Brandas e inverneiras: particularidades do sistema
agro-pastoril crastejo [Particularities of the Castro Laboreiro’s agricultural
and pastoral system]. Lisbon, Portugal: Instituto da Conservacxa
˜
oda
Natureza/Parque Nacional da Peneda-Gere
ˆ
s.
Giudicelli B. 1975. Situation Actuelle du Sloughi en Algerie [PhD disserta-
tion]. Lyon, France: Ecole Nationale Ve´te´rinaire.
Gomes MS. 2003. Racxas Caninas Auto´ctones Portuguesas. Contributo para o
seu Estudo Gene´tico e Demogra´fico [Portuguese autochthonous dog breeds.
A contribution to their genetic and demographic study] [Licenciatura disser-
tation]. Santare´m, Santare´m: Escola Superior Agra´ria de Santare´m.
Graur D, Li W-H. 2000. Fundamentals of molecular evolution. 2nd ed.
Sunderland, MA: Sinauer Associates.
Hasegawa M, Kishino H, Yano T. 1985. Dating of the human-ape splitting by
a molecular clock of mitochondrial DNA. J Mol Evol 22:160–74.
Hillis DM, Bull JJ. 1993. An empirical test of bootstrapping as a method for
assessing confidence in phylogenetic analysis. Syst Biol 42:182–92.
Hoelzel AR, Hancock JM, Dover GA. 1991. Evolution of the cetacean
mitochondrial D-Loop region. Mol Biol Evol 8:475–93.
Hudson RR, Boos DD, Kaplan N. 1992. A statistical test for detecting geo-
graphic subdivision. Mol Biol Evol 9:138–51.
ICN, 1997. Conservacxa
˜
o do lobo em Portugal [Wolf conservation in Portu-
gal]. Lisboa, Portugal: Instituto da Conservacxa
˜
o da Natureza. 231 p.
Jordana J, Manteca X, Ribo O. 1999. Comparative analysis of morphological
and behavioral characters in the domestic dog and their importance in the
reconstruction of phylogenetic relationships in canids. Genet Mol Biol
22:49–57.
Kim KS, Seong EL, Jeong HW, Ha JH. 1998. The complete nucleotide se-
quence of the domestic dog (Canis familiaris) mitochondrial genome. Mol
Phylogenet Evol 10:210–20.
Lorenz K. 1975. Foreword. In: Fox MW, editor. The wild canids: their
systematics, behavioural ecology and evolution. New York: Van Nostrand
Reinhold.
Mannion AM. 1999. Domestication and the origins of agriculture: an ap-
praisal. Prog Phys Geogr 23:37–56.
Martı
´
n PG, Angioni G, Raverdy Ch, Ojanen U, Este´banez NL, Martı
´
nez BG,
Chamorro GV, Lenberg JG, Ilera FA, Alonso AC, Albaladejo PF, Bernal
AM, Toledo AA, Dı
´
ez LM, Canada AM, Garcı
´
a MM, Pascual MR, Cardoso
JJ. 1994. Por los Caminos de la Trashumancia [On the Tanshumance paths].
Le´on, Spain: Junta de Castilla y Le´on.
McDonald DW, Carr GM. 1995. Variation in dog society: between resource
dispersion and social flux. In: James Serpell, editor. The domestic dog: its
evolution, behaviour and interactions with people. Cambridge: Cambridge
University Press. p 199–216.
Miguil A. 1986. Contribution a l’e´tude du Sloughi au Maroc [PhD disserta-
tion]. Rabat, Morocco: Institut Agronomique et Veterinaire Hassan II.
Molinari C. 1989. El Cao de Aguas Portugues. Um perro pescador [The
Portuguese Water dog. A fisher dog]. La Revista del Perro No. 29. Madrid,
Spain: Ediciones Guau, SA.
Molinari C. 1993. The Portuguese Water dog. 1st ed. Mafra, Portugal: Author.
Neff MW, Robertson KR, Wong AK, Safra N, Broman KW, Slatkin M,
Mealey KL, Pedersen NC. 2004. Breed distribution and history of
canine mdr1-1D, a pharmacogenetic mutation that marks the emergence
of breeds from the collie lineage. Proc Natl Acad Sci USA 101:
11725–30.
Okumura N, Ishiguro N, Nakano M, Matsui A, Sahara M. 1996. Intra- and
interbreed genetic variations of mitochondrial DNA major non-coding
regions in Japanese native dog breeds (Canis familiaris). Anim Genet
27:397–405.
Olivier M, Breen M, Binns MM, Lust G. 1999. Localization and character-
ization of nucleotide sequences from the canine Y chromosome. Chromo-
some Res 7:223–33.
Ostrander EA, Comstock KE. 2004. The domestic dog genome. Curr Biol
14:R98–9.
Ouar F. 1994. Contribution a l’e´tude de la situation du chien de l’Atlas:
re´partition ge´ographique, e´cologie et caracte´ristiques morphome´triques
[PhD dissertation]. Rabat, Morocco: Institut Agronomique et Veterinaire
Hassan II.
Parker HG, Kim LV, Sutter NB, Carlson S, Lorentzen TD, Malek TB,
Johnson GS, DeFrance HB, Ostrander EA, Kruglyak L. 2004. Genetic
structure of the purebred domestic dog. Science 304:1161–4.
Petit E, Ballou F, Excoffier L. 2002. Mammalian population genetics: why
not Y? Trends Ecol Evol 17:28–33.
Petrucci-Fonseca F, Pires AE, Ribeiro S, Almendra L, Clemente A, Collacxo
MT, Matos J, Simo
˜
es F. 2000. Ca
˜
es de gado e a conservacxa
˜
o do lobo em
Portugal [Livestock guarding dogs and wolf conservation in Portugal].
Galemys 12:135–48.
12
Journal of Heredity
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
Posada D. 2000. Testing geographic association—Chiperm. Version 1.2.
Provo, UT: Department of Biology, Brigham Young University.
Posada D, Crandall KA. 1998. MODELTEST: testing the model of DNA
substitution. Bioinformatics 14:817–8.
Ribeiro A, Saraiva JH. 2004. Histo´ria de Portugal. A Formacx a
˜
odo
Territo´rio—da Lusita
ˆ
nia ao Alargamento do Paı
´
s [History of Portugal.
Territory formation—from Lusitania to country expansion]. Matosinhos,
Portugal: Quidnovi
.
Roff DA, Bentzen P. 1989. The statistical analysis of mitochondrial DNA
polymorphisms: v
2
and the problem of small samples. Mol Biol Evol 6:
539–45.
Roger L, Boxshall G, Clark P. 1998. A dictionary of ecology, evolution and
systematics. 2nd ed. Cambridge, UK: Cambridge University Press.
Sambrook J, Russel DW. 2001. Molecular cloning—a laboratory manual. 3rd
ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Savolainen P, Zhang YP, Luo J, Lundeberg J, Leitner T. 2002. Genetic
evidence for an East Asian origin of domestic dogs. Science 298:1610–3.
Schneider S, Roessli D, Excoffier L. 2000. Arlequin version 2.000: a software
for population genetics data analysis. Geneva, Switzerland: Genetics and
Biometry Laboratory, University of Geneva.
Shields GF, Kocher TD. 1991. Phylogenetic relationships of North
American ursids based on analysis of mitochondrial DNA. Evolution 45:
218–21.
Silva RF. 2000. Transuma
ˆ
ncia no Portugal central: diversidade e organizacxa
˜
o
do territo´rio [Transhumance in central Portugal: diversity and land organi-
zation]. In: Associacxa
˜
o da defesa do patrimo´nio Arouquense, editor.
Montemuro—au
´
ltima rota da tansuma
ˆ
ncia. Arouca: Associacxa
˜
o da defesa
do patrimo´ nio Arouquense, Escola Superior Agra´ria de Viseu. p 55–69.
Southern SO, Southern PJ, Dizon AE. 1988. Molecular characterization of
a cloned dolphin mitochondrial genome. J Mol Evol 28:32–42.
Swofford DL. 2002. PAUP*. Phylogenetic analysis using parsimony (*and
other methods). Version 4. Sunderland, MA: Sinauer Associates.
Tamura K, Nei M. 1993. Estimation of the number of nucleotide substitu-
tions in the control region of mitochondrial DNA in humans and chimpan-
zees. Mol Biol Evol 10:512–26.
Tanabe Y, O
ˆ
ta K, Ito S, Hashimoto Y, Sung YY, Ryu JK, Faruque MO. 1991.
Biochemical-genetic relationships among Asian and European dogs and the
ancestry of the Japanese native dog. J Anim Breed Genet 108:455–78.
Tsuda K, Kikkawa Y, Yonekawa H, Tanabe Y. 1997. Extensive interbreeding
occurred among multiple matriarchal ancestors during the domestication of
dogs: evidence from inter- and intraspecies polymorphisms in the D-loop
region of mitochondrial DNA between dogs and wolves. Genes Genet Syst
72:229–38.
Vasconcelos RC. 1995. Racxas de Ca
˜
es Portugueses [Portuguese dog breeds].
2nd ed. Lisboa, Portugal: Editorial Presencxa.
Veiga V. 2001. O Podengo Portugue
ˆ
s [Portuguese Warren hound breed].
Lisboa, Portugal: Author.
Vila` C, Amorim IR, Leonard JA, Posada D, Castroviejo J, Petrucci-Fonseca
F, Crandall KA, Ellegren H, Wayne RK, 1999. Mitochondrial DNA phylo-
geography and population history of the grey wolf Canis lupus. Mol Ecol
8:2089–103.
Vila` C, Savolainen P, Maldonado JE, Amorim IR, Rice JE, Honeycutt RL,
Crandall KA, Lundeberg J, Wayne RK. 1997. Multiple and ancient origins of
the domestic dog. Science 276:1687–9.
Vila` C, Seddon J, Ellegren H. 2005. Genes of domestic mammals augmented
by backcrossing with wild ancestors. Trends Genet 21:214–8.
Walsh PS, Metzger DA, Higuchi R. 1991. Chelex 100 as a medium for single
extraction of DNA for PCR-based typing from forensic material. Biotech-
niques 10:506–13.
Wayne RK. 1986. Cranial morphology of domestic and wild canids: the in-
fluence of development on morphological change. Evolution 40:243–61.
Yang Z. 1994. Maximum likelihood phylogenetic estimation from DNA
sequences with variable rates over sites: approximate methods. J Mol Evol
39:306–14.
Zimen E. 1981. The wolf: his place in the natural world. London: Souvenir
Press.
Received August 26, 2005
Accepted May 4, 2006
Corresponding Editor: Robert Wayne
13
Pires et al.
Mitochondrial DNA of Portuguese Domestic Dogs
by guest on September 14, 2015http://jhered.oxfordjournals.org/Downloaded from
