Page 3 - Delicado_Machordom_2015

D. Delicado et al. Evolutionary patterns of Pseudamnicola

risk for extinction (Strong et al. 2008). Conversely, only a and the large subunit rDNA (16S rRNA), as well as the
few species of P. (Pseudamnicola) occur in mountainous nuclear large subunit rDNA (28S rRNA) were amplified
springs and most of them dwell streams and lakes, in which with the primers: LCO1490 and HCO2198 (Folmer et al.
the environmental conditions are more variable and waters 1994) for COI fragment, 16Sar-L and 16Sbr-H (Palumbi
remain connected. In an attempt to compare genetic pat- et al. 1991) for 16S, and F63.2 and LSU3 primers for 28S
terns among populations occurring in upland and lowland (Park & Foighil 2000) modified by Benke et al. (2009). The
streams, Hughes (2007) recovered information of several PCR cycling conditions were as described in Delicado et al.
freshwater animal groups, including mollusks, and con- (2013), including the annealing temperatures: 48, 50 and
cluded that genetic differences of lowland populations are 51 °C for COI, 16S and 28S, respectively. Products were
generally lower than those inhabiting in headwaters loca- sequenced in an ABI 3730 XL sequencer (Life Technolo-
tions. Despite this study did not compare upstreams and gies, Carlsbad, CA, USA) using a Big Dye Terminator Kit
downstreams populations belonging to the same taxonomic (Life Technologies).
group, its results suggest that habitat type is a very influen-
tial factor on dispersal capabilities and, therefore, on the Phylogenetic study design
genetic structure of the populations. Once the new sequences for each individual were obtained,
Thus, based on this background, in this study we gath- these were edited in Bioedit v. 7.0.5.3 (Hall 1999) and
ered for the first time mitochondrial and nuclear DNA compiled together with the other 164, gleaned from Gen-
sequences data of species from both subgenera to: (i) build Bank, in three individual data matrices corresponding to
a molecular phylogeny and thus compare the time and each gene fragment. Sequences of the three partitions
mode of diversification of both, the strict springsnails and could be unambiguously aligned using Se-Al version 2.0a11
lowland stream, subgenera, (ii) conduct an independent (Rambaut 2002).
analysis to assess the potential effect of habitat transition First, data matrices for each gene were analysed separately
on the divergence of the lineages, iii) estimate their ances- and subsequently combined to reconstruct the phylogenetic
tral biogeographic areas of diversification and (iv) finally trees. A partition homogeneity test (ILD) (Mickevich &
clarify the systematic status of the genus Pseudamnicola and Farris 1981; Farris et al. 1994), implemented in PAUP* 4.0
its subgenera. On the whole, through this multidisciplinary b10 (Swofford 2002), was used on the different partitions
study, we aim at providing a holistic overview of the evolu- (i.e. each gene) to check congruence among data for the dif-
tionary framework of an organismal group linked to differ- ferent genes. Prior to the phylogenetic analyses, we
ent inland aquatic ecosystems. employed jModelTest v. 0.1.1 (Posada 2008) under Akaike’s
information criterion (AIC; Akaike 1974) to identify the best
Materials and methods molecular evolutionary model of nucleotide substitution that
Samples and sequences fits for each data set. For the COI partition, the model
To assess Pseudamnicola evolutionary relationships, we selected was HKY+I (Invariable sites) +G (rate variation
examined a total of 202 individuals from 91 localities among sites) model (Hasegawa et al. 1985) and for 16S and
(Fig. 1) belonging to the genus and two out-group species, 28S fragments GTR+I+G (Generalized time-reversible
Peringia ulvae (Pennant, 1777) and Mercuria emiliana (Pala- model; Tavar e 1986). Phylogenetic inference was obtained
dilhe, 1869). The sequences of these out-groups were by conducting maximum likelihood (ML), maximum parsi-
acquired from GenBank under accession numbers mony (MP) and Bayesian inference (BI) methods.
JX081779–80, JX081888–89 and JX081990–91 (Delicado Maximum parsimony analyses were performed in
et al. 2013). In-group localities included: 19 for P. (Pseu- PAUP* through a heuristic search with a tree bisection
damnicola) species from the Ibero-Balearic region (previ- and reconnection TBR swapping algorithm, including ten
ously sequenced in Delicado et al. 2014), 51 for random stepwise additions. ML analyses were conducted in
P. (Corrosella) species (generated by Delicado & Ramos PHYML v. 2.4.4 (Guindon & Gascuel 2003) using the
2012; Delicado et al. 2012, 2013) and 21 for additional evolutionary models selected by jModelTest. Clade support
P. (Pseudamnicola) populations from other areas of the for the MP and ML phylogenies was assessed by nonpara-
Mediterranean basin (Fig. 1, Table S1). metric bootstrapping (Felsenstein 1985) using 1000 pseu-
doreplicates in each case. BI was run using the software
DNA isolation, amplification and sequencing MrBayes version 3.1.2 (Huelsenbeck 2000; Huelsenbeck &
DNA of a total of 40 specimens from those 21 supplemen- Ronquist 2001), performing two independent and parallel
tary Mediterranean populations were isolated in this work runs of four Metropolis-coupled chains with 5 million gen-
following the CTAB protocol of Wilke et al. (2006). Two erations each, and sampling one tree per 1000 replicates.
mitochondrial genes, cytochrome c oxidase subunit I (COI) After assessing convergence between runs by checking that

ª 2015 Royal Swedish Academy of Sciences, 44, 4, July 2015, pp 403–417 405