Molecular studies and phylogenetic analyses
DNA was extracted from small pieces of foot muscle from representative specimens by use of a QIAGEN DNA extraction kit for animal tissue (Qiagen, Hilden) following the standard procedure of the manual. An approximately 900 base pair long fragment of the 16S gene was amplified by PCR using the primers 16S3F and 16S4Ra (Hyman et al., 2007). Whenever we failed to amplify the whole fragment due to DNA fragmentation as typically encountered in extracts from older museum specimens, we amplified two overlapping shorter fragments or even performed nested PCRs by using the internal primers 16S3R and 16S4F (Hyman et al., 2007). In addition, an 823 base pair long fragment of the COI gene was amplified by using the primers LCOH1940 (Folmer et al., 1994) and COI-H865 (5’- TACYATTGTRGCAGCTGTAAA-3’; designed herein). For samples with highly fragmented DNA we performed a nested PCR using the primers LCOH1490 and HCOI2198 (Folmer et al., 1994) to amplify a 655 base pair long fragment. Reactions were performed using standard protocols with annealing temperatures / elongation times of 55 °C / 90 s for 16S and 60 s 50 °C / 60 s for COI, respectively. Both strands of PCR fragments were purified and cycle sequenced by use of the PCR primers. Electropherograms were corrected for misreads and forward and reverse strands were merged into one sequence file using CodonCode Aligner v. 3.6.1 (CodonCode Corp., Dedham, MA). Sequences of the previous helicarionid study Hyman et al. (2007) were retrieved from GenBank and included in our dataset while all newly produced sequences have been deposited in GenBank under the accession numbers KY662298-662378, KY662388-KY662468.
The 16S sequences were aligned using the online version of MAFFT (version 7) available at http://www.mafft.cbrc.jp/alignment/server/ by employing the iterative refinement method E-INS-i suitable for sequences with multiple conserved domains and long gaps (Katoh et al., 2002). Uncorrected p-distances between sequences were calculated by using the phylogenetic software MEGA7 (Kumar et al., 2016) under the option ‘pair-wise deletion of gaps’. Prior to the phylogenetic analysis, we used the online version of Gblocks (Castresana, 2000) available at http://www.molevol.cmima.csic.es/castresana/Gblocks_server.html to remove ambiguously aligned positions from the 16S alignment by enabling all options allowing for a less stringent selection. Each mtDNA fragment was checked for saturation using the test implemented in DAMBE (Xia and Lemey, 2009). The best-fit model of nucleotide substitution was identified for each gene partition separately using the model proposal function of MEGA7.
The aligned 16S and COI sequences were then concatenated into one partitioned data set and a maximum likelihood-based method of tree reconstruction was employed to estimate phylogenetic relationships. We analysed the concatenated and partitioned sequence dataset using the program raxMLgui (version 1.5) (Silvestro and Michalak, 2012). Nodal support of the best ML tree was estimated by performing 10 independent runs each with 200 thorough bootstrap replicates.