Amphibia-Reptilia 29 (2008): 523-533
β-fibrinogen intron 7 variation in Discoglossus (Anura:
Discoglossidae): implications for the taxonomic assessment of
morphologically cryptic species
Guillermo Velo-Antón
1,*
, Iñigo Martínez-Solano
2
, Mario García-París
3
Abstract. The generalized use of nuclear introns in combination with mitochondrial DNA data in molecular systematic and
intraspecific phylogeographical studies is providing new insights into the complex evolutionary histories of taxa surviving
the Quaternary glaciations. Previous studies have highlighted the suitability of the beta-fibrinogen intron 7 (β-fibint7)for
phylogenetic and phylogeographic studies in a wide variety of taxa, including amphibians. Here we use sequences of this
marker to assess inter- and intraspecific variation in Discoglossus (Discoglossidae), with special emphasis on geographic
patterns of genetic structure in the Iberian Peninsula, where recent studies have questioned the taxonomic status of D.
jeanneae. We obtained 81 sequences of β-fibint7 from samples including all currently recognized species except D.
montalentii and 37 populations in the Iberian Peninsula and compared levels of genetic variation with those observed in
a fragment of similar length of the mtDNA gene cytochrome b. The sequence of β-fibint7 in Discoglossus is the shortest
described for amphibians so far, 378 base pairs. In general, we found low levels of variability (only 26 parsimony-informative
sites in the dataset), with no alternatively fixed haplotypes in samples attributed to D. galganoi or D. jeanneae basedontheir
mtDNA. Values of pairwise sequence divergence between non-Iberian species ranged from 1.1% to 4.5% (13.3% to 20.9% in
mtDNA). The patterns observed in samples from the Iberian Peninsula are consistent with either incomplete lineage sorting
or ongoing gene flow between D. galganoi and D. jeanneae. We conclude by reviewing the genetic evidence available to
address the taxonomic status of Iberian species of Discoglossus.
Keywords: beta-fibrinogen intron 7, Discoglossus, Iberian Peninsula, mtDNA, nuclear intron, systematics.
Introduction
The use of sequence data from nuclear in-
trons in molecular systematics and in phylogeo-
graphic studies has become increasingly preva-
lent, mostly in combination with mitochondrial
DNA (mtDNA) data (Dolman and Moritz, 2006;
Heckman et al., 2007; Leavitt et al., 2007).
In those cases, the analyses of several poten-
tially unlinked molecular markers, with very
different inheritance profiles and mutation rates
across a comprehensive geographic set of sam-
ples is crucial to bridging the gap between gene
1 - Grupo de Ecoloxía Evolutiva, Departamento de
Ecoloxía e Bioloxía Animal, Universidade de Vigo,
E.U.E.T. Forestal, Campus Universitario, 36005 Pon-
tevedra, Spain
2 - Department of Ecology and Evolutionary Biology,
University of Connecticut, 75 N Eagleville Road, Unit
3043, Storrs, Connecticut 06269, USA
3 - Museo Nacional de Ciencias Naturales, C.S.I.C., c/ José
Gutiérrez Abascal 2, 28006 Madrid, Spain
∗
Corresponding author; e-mail: guillermov@uvigo.es
and species trees (Maddison, 1997; Templeton,
2004). Combining information from nuclear
and mtDNA markers thus constitutes a power-
ful tool to infer complex evolutionary histories,
like those exhibited by species that survived
the glaciations of the Quaternary in allopatric
refugia and underwent several successive cy-
cles of isolation (retreat) and admixture (expan-
sion) following climatic changes (Knowles and
Carstens, 2007).
Amongst the increasing list of introns used
in molecular systematics, one of the most com-
mon is the beta-fibrinogen intron 7 (β-fibint7).
The availability of nearly-universal primers and
its relatively fast evolutionary rate have proven
useful in recovering phylogenetic relationships
in a variety of vertebrate taxa, from amphibians
to birds and mammals (Prychitko and Moore,
1997; Weibel and Moore, 2002; Yu and Zhang,
2005; Sequeira et al., 2006; Gonçalves et al.,
2007). In some cases, high levels of intraspecific
variation have allowed characterization of phy-
©
Koninklijke Brill NV, Leiden, 2008. Also available online - www.brill.nl/amre
524 G. Velo-Antón, I. Martínez-Solano, M. García-París
logeographic patterns, sometimes correspond-
ing closely to those observed at the mtDNA
level (Godinho et al., 2006).
Discoglossus and its close fossil relatives
conform one of the most ancient clades of
frogs, with an independent evolutionary history
since the Lower Cretaceous (Sanchiz, 1998; San
Mauro et al., 2004), and constitute an interest-
ing group to further evaluate the potential util-
ity of this marker. The genus includes six ex-
tant species endemic to the western Mediter-
ranean Basin that have been widely studied
from morphological, molecular and ecological
perspectives (Capula et al., 1985; Glaw and
Vences, 1991; Capula and Corti, 1993; Vences
and Glaw, 1996; Fromhage et al., 2004; Zan-
gari et al., 2006). Discoglossus galganoi and D.
jeanneae are endemic to the Iberian Peninsula;
D. scovazzi is present in north Africa west of
the Moulouya River; D. pictus auritus is also
present in north Africa east of the Moulouya
River, with introduced populations in NE Iberia
and south-eastern France; D. pictus pictus in-
habits Sicily, Malta and Gozo; D. montalentii
is found in Corsica and finally, D. sardus is
found in Corsica, Sardinia and smaller islands in
the Tyrrhenian Sea. Whereas phylogenetic rela-
tionships between taxa have been well resolved
(Fromhage et al., 2004; Zangari et al., 2006),
important questions remain unresolved or con-
troversial, most notably the taxonomic status of
the Iberian endemic D. jeanneae Busack, 1986.
The most recent and comprehensive molecular
dataset published so far (Zangari et al., 2006)
found little genetic differentiation in allozymes
between populations assigned to this species
and its sister taxon, D. galganoi, also endemic
to the Iberian Peninsula, and suggested recogni-
tion as subspecies (D. galganoi galganoi and D.
g. jeanneae) as the most appropriate taxonomic
arrangement for the two Iberian taxa. However,
despite the low genetic distances, both Iberian
species were recovered as monophyletic, imply-
ing some level of nuclear differentiation.
Our aim in this study is to evaluate the phylo-
genetic and phylogeographic utility of β-fibint7
in Discoglossus, with special attention to ge-
ographic patterns of variation in the Iberian
Peninsula. We compare our results with pub-
lished and new mtDNA sequences from the
same individuals used in this study and discuss
the genetic evidence available to address the
taxonomic status of D. jeanneae.
Materials and methods
Sampling
We obtained tissue samples from all extant species of
Discoglossus except D. montalentii. The final dataset in-
cluded a total of 68 samples from Iberian Discoglossus (D.
galganoi and D. jeanneae) collected at 37 localities across
the Iberian Peninsula, three samples of D. scovazzi from one
locality in Morocco, two samples of D. pictus from two lo-
calities in Tunisia, and eight samples of D. sardus from Cor-
sica and Sardinia (see table 1, fig. 1).
DNA amplification and alignment
Total genomic DNA was extracted from ethanol-preserved
tissues using a standard proteinase K/phenol chloroform ex-
traction protocol. A total of ∼510 base pairs (bp) corre-
sponding to beta-fibrinogen intron 7 and part of the exon
flanking regions were amplified via polymerase chain re-
action (PCR) using the primers FIB-B7U and FIB-B7L
(Prychitko and Moore, 1997) in all 81 samples used in this
study. PCR reactions consisted of 38 cycles with a denatur-
ing temperature of 94
◦
C (1 min), annealing at 50
◦
C(1min)
and extension at 72
◦
C (1 min). PCR reactions were per-
formed in a total volume of 13 μl with 30-60 ng of tem-
plate DNA, including 0.15 μl of Taq polymerase (Biotools,
5 U/ml), 0.5 μl of each primer (10 mmol/l), 0.5 μlofdNTPs
(10 mmol/l), 0.7 μlofMgCl
2
(25 mmol/l) and 1.25 μlofre-
action buffer (Biotools, Tris-HCl, pH = 8.3). Double-strand
templates were cleaned using sodium acetate and ethanol to
precipitate the PCR products and then re-suspended in 20-
25 μl of ddH
2
O. Sequencing reactions were performed for
both strands and sequenced on an ABI PRISM 3700 DNA
sequencer following the manufacturer’s instructions.
For comparisons, we compiled a mtDNA dataset (70 in-
dividuals) including most of the samples sequenced for β-
fibint7. This dataset is composed of previously published
partial sequences of the cytochrome b (cob) gene (Martínez-
Solano, 2004), complemented with 24 new sequences. De-
tails of amplification and sequencing of cob are provided
in Martínez-Solano (2004). Sequences of D. montalen-
tii were downloaded from GenBank (Accession Numbers:
AY347430 and AY347431) and included in the final cob
alignment for reference.
Sequences were compiled, edited and aligned manu-
ally with ProSeq v.2.91 (Filatov, 2002). Heterozygous nu-
cleotide positions were identified as double peaks in the
β-fibint7 variation in Discoglossus 525
Table 1. Sampling localities, including latitude and longitude coordinates, voucher numbers, sampling sizes and haplotypes observed in β-fibint7 and cob datasets. Region numbers refer
to population groups in fig. 1.
Species Country Region Locality Vouchers b-fibint7 cob Latitude Longitude b-fibint7 haplotype cob haplotype
D. galganoi Portugal 1 Cernache de Bonjardim IMS 338, 339 2 2 39
◦
48
36
N8
◦
10
33
WI,V,VII VI
2 Vagos IMS 324, 325 2 2 40
◦
32
24
N8
◦
32
00
W I I, IV
3 Póvoa do Varzim IMS 373, 374 2 2 41
◦
22
60
N8
◦
46
00
W III, VIII I, V
Spain 4 Pontevedra: Isla de Ons ONS 1, 2 2 2 42
◦
21
55
N8
◦
56
17
W I, III I
5 Coruña: Torre de Hercules Coruña 1, 2 2 2 43
◦
23
14
N8
◦
24
9
W I, III II
6 Lugo: Gomeán IMS 215, 228, 229 3 2 42
◦
56
06
N7
◦
24
18
W I, III II
7 Asturias: Tineo IMS 218, 219, 220, 222 4 3 43
◦
20
18
N6
◦
24
49
W I, III I
8 León: Reliegos IMS 4, 6 2 2 42
◦
28
35
N5
◦
21
17
W I, II, III I, III
9 Zamora: Codesal IMS 230 1 2 41
◦
58
15
N6
◦
22
54
WI,II I
10 Burgos: Quintanilla-Escalada IMS 51, 52 2 2 42
◦
47
52
N3
◦
46
18
WI,II I
13 Madrid: Soto del Real IMS 34 1 1 40
◦
45
57
N3
◦
46
10
WI,X VI
13 Madrid: Cenicientos IMS 17 1 1 40
◦
15
14
N4
◦
29
57
WII VI
13 Madrid: Navalagamella IMS 60 1 1 40
◦
28
00
N4
◦
07
00
WI,II VI
13 Madrid: Villamantilla IMS 47, 48 2 2 40
◦
22
21
N4
◦
08
11
WI,II VI
13 Madrid: Cerceda IMS 540 1 1 40
◦
42
00
N3
◦
55
60
W II, X VI
14 Ávila: Monbeltrán IMS 42 1 1 40
◦
15
30
N5
◦
01
12
WI VI
15 Toledo: El Real de San Vicente IMS 234, 235 2 2 40
◦
08
20
N4
◦
41
31
WI,II VI
17 Badajoz: Mérida IMS 217, 232, 233 3 3 38
◦
57
37
N6
◦
21
36
W I, II, VI, IX VI
18 Ciudad Real: Poblete IMS 44, 45, 46 3 2 38
◦
56
19
N3
◦
58
55
W I, II VIII
22 Sevilla: El Ronquillo IMS 32 1 1 37
◦
42
05
N6
◦
06
57
W VI, X VII
23 Huelva: Rosal de la Frontera IMS 533, 534 2 1 37
◦
58
00
N7
◦
13
00
W III, IV, XIII VI
526 G. Velo-Antón, I. Martínez-Solano, M. García-París
Table 1. (Continued).
Species Country Region Locality Vouchers b-fibint7 cob Latitude Longitude b-fibint7 haplotype cob haplotype
D. jeanneae Spain 10 Burgos: Entrambosríos IMS 49, 50 2 2 43
◦
02
53
N3
◦
42
02
WI,II IX
11 Zaragoza: Romanos IMS 53 1 1 41
◦
07
40
N1
◦
16
36
WII IX
12 Guadalajara: Mondéjar IMS 26 1 1 40
◦
19
53
N3
◦
06
28
WI,II IX
12 Guadalajara: Driebes IMS 27 1 1 40
◦
14
55
N3
◦
02
41
WI,II IX
12 Guadalajara: Yebes IMS 62 1 1 40
◦
50
55
N3
◦
28
01
WI,II IX
12 Guadalajara: Uceda IMS 63 1 1 40
◦
07
38
N3
◦
12
12
WII IX
13 Madrid: Aoslos IMS 1, 2, 10, 11, 12 5 4 41
◦
03
15
N3
◦
36
08
W I,II,X IX, X
13 Madrid: Valdilecha IMS 22 1 1 40
◦
16
60
N3
◦
17
60
WII IX
13 Madrid: Belmonte de Tajo IMS 58 1 1 40
◦
08
48
N3
◦
25
37
WII IX
13 Madrid: Chinchón IMS 59 1 1 40
◦
07
60
N3
◦
25
00
WI,II IX
16 Cuenca: Belinchón IMS 73 1 2 40
◦
03
05
N3
◦
02
51
WI IX
19 Albacete: Ojos de Villaverde GVA135 1 1 38
◦
48
51
N2
◦
22
06
WVI XI
20 Jaén: Iznatoraf IMS 250, 251, 252 3 2 38
◦
08
60
N3
◦
01
60
W I, VI, XI IX
20 Jaén: Santuario Nuestra Señora IMS 253, 254, 255, 256 4 4 38
◦
05
59
N4
◦
01
04
W II, XI IX
Cabeza
21 Málaga: Fuengirola IMS 257 1 1 36
◦
31
60
N4
◦
37
00
WI IX
24 Cádiz: San José del Valle IMS 33, 36, 37 3 1 36
◦
36
23
N5
◦
24
20
W I, II, IX, XII IX
D. scovazzi Morocco Ifrane IMS 18, 19, 28 3 2 33
◦
31
60
N5
◦
06
00
W XIV, XV, XVI XII, XIII
D. pictus Tunisia Gafsa Oasis Sud GVA 301 1 1 34
◦
23
18
N8
◦
47
50
E XVII, XVIII XIV
Monastir GVA 302 1 1 35
◦
45
33
N10
◦
48
49
EXIX,XX XV
D. sardus France Corsica GVA 73, 74, 75, 76, 79, 81 6 4 41
◦
39
32
N9
◦
11
23
E XXI, XXII XVI, XVII, XVIII
Italy Sardinia GVA 85, 86 2 2 40
◦
59
21
N9
◦
12
37
E XXI, XXII XVI
D. montalentii France Corsica 0 2 XIX, XX
Total 81 74
β-fibint7 variation in Discoglossus 527
Figure 1. Sampling locations of D. galganoi and D. jeanneae in the Iberian Peninsula. Pie charts on maps represent sampling
localities (pooled based on geographic proximity, see “Regions” in table 1) and haplotype frequencies for cob (1A) and
β-fibint7 sequences (1B). On the right, statistical parsimony networks depicting relationships between haplotypes, with sizes
proportional to their frequencies. Numbers in fig. 1A represent population groups as listed in table 1.
electropherograms (Brumfield et al., 2003) and coded ac-
cording to standard IUPAC ambiguity codes. All sequences
obtained for this study were deposited in GenBank (Acces-
sion numbers EU744884-EU744911).
Genetic analyses
We used the Bayesian approach implemented in PHASE
version 2.1 (Stephens et al., 2001) to reconstruct haplotypes
for heterozygous individuals. Only haplotypes with poste-
rior probabilities (pp) over 0.90 were included in the analy-
ses.
Haplotype networks for both nuclear and mitochondrial
datasets were constructed and compared using the statisti-
cal parsimony algorithm implemented in TCS version 1.21
(Clement et al., 2000) treating indels (regardless of their
size) as single mutational events (fig. 1). Intra- and inter-
specific corrected (Kimura-2 parameter) genetic distances
between samples were estimated with the software package
MEGA 3.1 (Kumar et al., 2004).
Two approaches were used to determine the presence
of recombination in the nuclear gene. First, the minimum
number of recombination events (Rm) (Hudson and Kaplan,
1985) was determined using DNASP version 4.10.9 (Rozas
et al., 2003). Additionally, the Phi test implemented in
SplitsTree4 (Huson and Bryant, 2006) was used to calculate
the pairwise homoplasy index (
w
) of Bruen et al. (2006)
to test for recombination. The ability of this statistic to
distinguish recurrent mutation from recombination has been
demonstrated through simulations (Bruen et al., 2006).
In order to test the phylogenetic potential of the mark-
ers used we performed standard phylogenetic analyses on
both nuclear and mtDNA datasets. These included Max-
imum Parsimony (MP), Maximum Likelihood (ML) and
Bayesian analyses. MP and ML analyses were performed
with PAUP* 4.0b10 (Swofford, 2003). First, we analyzed
the data to determine the best fitting substitution model for
the maximum likelihood and Bayesian analyses, which was
selected with the Akaike information Criterion (AIC) as im-
plemented in ModelTest 3.7 (Posada and Crandall, 1998).
For MP and ML analyses the heuristic search algorithm
was performed with 10 random additions of sequences and
tree-bisection-reconnection (TBR) branch swapping. Statis-
tical support for the resulting topologies was assessed us-
ing nonparametric bootstrap (Felsenstein, 1985), with 1000
pseudorreplicates for MP and 100 for ML. Bayesian analy-
ses were carried out with Mr. Bayes v3.1 (Ronquist and
Huelsenbeck, 2003), defining three partitions correspond-
ingto1st+ 2nd + 3rd codon positions in the cob dataset.
The β-fibint7 alignment was analyzed as an unpartitioned
dataset. Analyses were run for 10
6
generations with a sam-
pling frequency of 10
4
generations. Of the resulting 1000
