Page 6 - BMC_1471

BMC Evolutionary Biology 2008, 8:56 http://www.biomedcentral.com/1471-2148/8/56

Table 1: Pairwise F ST between groups of West Mediterranean green toads based on sequences of the mitochondrial d-loop. All F ST -
values are significant (p < 0.0000).
B. balearicus B. balearicus B. balearicus B. boulengeri Offshore B. boulengeri North
Corsica+Sardinia Balearic Islands Apennine Peninsula Islands, N. Africa African Mainland
B. balearicus Balearic Islands 0.60662
B. balearicus Apennine Peninsula 0.75859 0.49501
B. boulengeri Offshore Islands, N. Africa 0.91530 0.91812 0.95058
B. boulengeri North African Mainland 0.77888 0.74269 0.82886 0.16073
B. siculus 0.97543 0.98188 0.98815 0.96950 0.87081

population growth for the green toads on the African ported (posterior support = 100%) clade that differs from
mainland, as well as for all boulengeri considered together; all sampled African and Eurasian green toads (Figure 2).
log likelihood tests rejected no growth scenarios for both For all mitochondrial and nuclear markers, Bufo n.sp.
groupings (Table 2). Tajima's D calculated for the main- shows a much greater genetic distance from Italian main-
land group indicated some support for population expan- land B. balearicus than from B. boulengeri, which inhabits
sion (-1.63869, 0.10 > p > 0.05); however, the mismatch all of North Africa. Taken together, as for the two preced-
distribution analysis did not support a growth model. ing taxa, phylogenetic divergence of Bufo n. sp. is evident
and we acknowledge this by describing it as new species
Tropomyosine intron sequences in African samples from (Bufo siculus, see below).
Libya (loc. 36), Tunisia (loc. 29) and Lampedusa (loc. 26)
exhibit the same tree topology as the mitochondrial mark- Bufo viridis (Laurenti, 1768) and old isolates
ers and constitute a highly supported clade (posterior The haplotype group representing this taxon, as previ-
probability = 96%) that differs substantially from tropo- ously shown by Stöck et al. [28], was found exclusively in
myosine sequences from two Sicilian green toads (loc. the very northeastern part of Italy, northeast of the Po
21). The very conserved RAG-1 marker shows two African River (loc. 41, 42). So far, two very well supported mono-
samples in a polytomy with Eurasian green toads, from phyletic and geographically widespread mtDNA groups of
which both B. balearicus and two samples each from Sicily green toads are known from the Balkan region. The hap-
and Lampedusa differ. lotype group dominating Asia Minor (B. variabilis) and
that of B. viridis (Figure 1: blue and light green; [28])
Bufo n. sp apparently co-occur in Greece. However, neither of the
This taxon was identified on most of Sicily (loc. 21-23a) two was found off the Adriatic coast on the Croatian
and two close islands (Favignana, loc. 25; Ustica, loc. 24). islands of Krk and Cres (loc. 43, 44). Instead, a very well
Demographic analyses performed in Fluctuate estimated θ supported separate clade (posterior support = 100%)
to be an order of magnitude lower than that estimated for formed by two sequences of toads from those islands
all other green toad groups examined except the mainland revealed an apparent relict group, which is more closely
balearicus group (Table 2). Although the log likelihood related to B. viridis than to B. variabilis (Figure 2), but fur-
test did not permit rejection of a no growth scenario, the ther data are required to confirm this relationship.
very low initial θ estimated by the analysis indicates that
few individuals may have founded the lineage. Similarly, Age of African-Sicilian vicariant separation
the observed mismatch distribution is not statistically Divergence time estimates for the principal mitochondrial
identical to the expected distribution under a sudden (control region and 16S rRNA) clades recovered are
growth model (p = 0.077). However, the unimodal peak shown in Table 3. We provide the 95% highest posterior
is shifted to the left of the distribution and very tightly density intervals (95% HPD, that is, the shortest intervals
matches the expected distribution for a recently expanded that include 95% of the posterior sampled values) as well
population (Figure 3f). Finally, Tajima's D did not indi- as the mean of the sampled values. Time estimates are
cate deviation from neutrality, and thus a scenario of pop- based on the reconstruction of the most common recent
ulation expansion cannot be invoked with this measure. ancestor for the mtDNA control regions under the coales-
cent (see Material & Methods section for details). The 95%
Bufo n.sp. exhibits reciprocal monophyly for both mito- HPD, although overlapping, show consistent values
chondrial markers (and using the tropomyosine intron within and between clades. Divergence between African
marker on a subsample of individuals) with all other and Sicilian haplotypes falls within the range of 0.6 and
groups/taxa (Figure 2). The RAG-1 phylogram shows 3.5 My [mean 1.8 My, around the Pliocene/Pleistocene
toads from Sicily and Lampedusa forming a well-sup- boundary (~1.8 My)] for the control region. Divergence

Page 6 of 19
(page number not for citation purposes)