Page 12 - Colliard_ali_2010

Colliard et al. BMC Evolutionary Biology 2010, 10:232 Page 12 of 16
http://www.biomedcentral.com/1471-2148/10/232

In addition to drift and dispersal, selection often con- Important caveats obviously apply to our crossing
tributes crucially to the maintenance of narrow hybrid experiments, mainly due to the practical difficulties in
zones [39]. We think that the one under study is no obtaining reproducing individuals from the field.
exception, and that selection against nuclear introgres- Absence of replicates limits the power of our inferences,
sion is very likely to also play a role. First, F 1 -hybrids and we lack the exact controls for intraspecific matings
from our experimental crosses showed reduced fitness (although the latter is compensated by our long experi-
(Table 3). Outcomes from reciprocal crosses were asym- ence of breeding green toads in the lab, which allowed
metric, which often occurs in interspecific crosses (as us to provideintraspecificcontrolsfromother Western
documented e.g. in hybrids between green toads and Palearctic lineages; Table 3). Results from these preli-
Bufo calamita [40] or B. bufo [41,42]. If the small size minary crosses must be clearly considered as provi-
and altered coloration observed in males from the B. sional, but are worth reporting here as a support for our
balearicus × B. siculus cross translate into lower survival main conclusions gathered from field population-genet-
or fertility in the field, then nuclear markers are indeed ics data.
expected to show less introgression than mtDNA. Sec- Though many hybrid zones have been documented in
ond, survival was drastically affected both in backcrosses amphibians (see Background), few provide the data
and in F 2 -hybrids (F 1 ×F 1 ), which should strongly required to calibrate the speciation process, in terms of
reduce introgression. Third, the currently known ranges reliable time of divergence and patterns of introgression
of both taxa in Sicily roughlycoincide withthe borders in the field. A notable exception is provided by fire- and
of semiarid (B. balearicus) and arid (B. siculus)climates yellow-bellied toads, which constitute one of the best-
[43]. Adaptation of these two genomes to different cli- studied anuran hybrid zones. Bombina bombina and B.
mates may select against hybrids, and potentially stabi- variegata, thought to have diverged during Upper Mio-
lize the contact zone. cene or Lower Pliocene [3.5 Mya; [49-51]] hybridize in
The sex-specific phenotypic effects in the F 1 as well as narrow, stable zones maintained by selection and disper-
the dramatic hybrid breakdown observed both in back- sal [51]. Strong selection against hybrids is generated
crosses and in the F 2 generation (F 1 × F1) were expected both by hybrid breakdown [21,52], and by environment-
from the classical Dobzhansky-Muller model of specia- dependent selection against toads in mismatched habi-
tion [44], arising from the confrontation of incompatible tats [53].
genes and the disruption of co-adapted sets of genes. However, mitochondrial introgression in Bombina
Except for sex chromosomes in the heterogametic sex, seems more limited than in Sicilian green toads, with
F 1 -hybrids inherit complete sets of genes from both par- mtDNA clines similar to or even steeper than those of
ental species and should thus suffer little from co-adap- nuclear loci [allozymes; [54]]. The divergence between
tation losses. We do not know, however, whether sex- B. siculus and B. balearicus is slightly more recent (Plio-
specific differences in F 1 phenotypes (dwarfed males) Pleistocene, ca 2.7 Mya), but, despite higher mitochon-
conform to Haldane’s rule [45], because sex-determina- drial DNA introgression, we found almost no admixture
tion mechanisms are unknown for Sicilian green toads at supposedly neutrally evolving nuclear microsatellite
(as for most amphibians). Results from related species loci, suggesting stronger selection to keep gene pools
suggest a XX/XY-system for B. variabilis (as “B. viridis“) apart.
from Asia Minor [46,47], but a ZZ/ZW-system for taxo-
nomically undefined green toads from Moldavia [B. viri- Conclusions
dis/B. variabilis contact zone?, [48]]. Contrasting with While anuran species that diverged > 8 Mya may exhibit
F 1 , backcrosses and F 2 (F 1 ×F 1 ) inherit imbalanced partial or complete hybrid inviability in the laboratory,
numbers of genes from both parental species due to as recently shown for instance by a combination of
recombination in F 1 , and may thus lack crucial alleles at experimental crosses and molecular divergence time
complementary loci. Inbreeding presumably also played estimates for the Fejevarya limnocharis group [55], there
aroleinour F 2 crosses, which may partially explain the is accumulating evidence that anurans with Plio-(Pleisto)
additional mortality relative to the breakdown observed cene divergence tend to be reproductively isolated under
in backcrosses (Table 3). However, inbreeding is also natural conditions. Secondary contacts between Austra-
expected to affect F 1 ×F 1 crosses in the field, given the lian hylids of Plio-Pleistocenedivergencehaveprovided
scarcity of hybridization events. It is also worth noting evidence for allopatric speciation driven by reinforce-
that a few backcrossed tadpoles survived metamorpho- ment mechanisms, characterized by highly asymmetric
sis. Among such rare individuals, females are most likely F 1 -viability in experimental crosses [8]. Neither F 1 and
responsible for the mitochondrial introgression and F 2 -hybrids nor backcrosses could be identified at a con-
signs of nuclear allele leakage observed in the field. tact zone between two Hyla species with an estimated