Phylogenetic position of Loricifera inferred from nearly complete 18S and 28S rRNA gene sequences

Background Loricifera is an enigmatic metazoan phylum; its morphology appeared to place it with Priapulida and Kinorhyncha in the group Scalidophora which, along with Nematoida (Nematoda and Nematomorpha), comprised the group Cycloneuralia. Scarce molecular data have suggested an alternative phylogenetic hypothesis, that the phylum Loricifera is a sister taxon to Nematomorpha, although the actual phylogenetic position of the phylum remains unclear. Methods Ecdysozoan phylogeny was reconstructed through maximum-likelihood (ML) and Bayesian inference (BI) analyses of nuclear 18S and 28S rRNA gene sequences from 60 species representing all eight ecdysozoan phyla, and including a newly collected loriciferan species. Results Ecdysozoa comprised two clades with high support values in both the ML and BI trees. One consisted of Priapulida and Kinorhyncha, and the other of Loricifera, Nematoida, and Panarthropoda (Tardigrada, Onychophora, and Arthropoda). The relationships between Loricifera, Nematoida, and Panarthropoda were not well resolved. Conclusions Loricifera appears to be closely related to Nematoida and Panarthropoda, rather than grouping with Priapulida and Kinorhyncha, as had been suggested by previous studies. Thus, both Scalidophora and Cycloneuralia are a polyphyletic or paraphyletic groups. In addition, Loricifera and Nematomorpha did not emerge as sister groups. Electronic supplementary material The online version of this article (doi:10.1186/s40851-015-0017-0) contains supplementary material, which is available to authorized users.


Introduction
Since its first description as a new phylum [1], Loricifera has been one of the most enigmatic metazoan phyla. Although only 35 loriciferan species have been described worldwide, the actual species diversity is higher, as many new species await description [2][3][4][5][6]. All known loriciferan species are microscopic (80-800 μm) and occur in marine sediments, such as mud, sand, and shell gravel. The most extreme habitat for Loricifera is the hypersaline anoxic deep basin in the Mediterranean Sea, where members of this phylum are metabolically active [6,7]. Our knowledge of loriciferan life cycles is also only fragmentary, given the recent findings of new life cycles and larval types [3][4][5]8].
The present study investigated the phylogenetic position of phylum Loricifera within Ecdysozoa using nearly complete 18S and 28S rRNA sequences. Also of interest was the phylogenetic status of the taxa Scalidophora and Cycloneuralia.

Sampling and DNA sequencing
The loriciferan specimen used in this study was collected from Ise Bay, Japan, northwestern Pacific (34°9.77′N, 136°51.40′E, 161-174 m depth) during a cruise of the TR/V Seisui-maru (Mie University) on 21 November 2013. A sediment sample was collected with a biological dredge, subsequently frozen to prevent DNA degradation, and sent to the laboratory. In the laboratory, meiofaunal specimens were extracted by floatation [30] with Ludox® HS 40. The extracted sample was sorted under a stereomicroscope, and a single adult loriciferan specimen (Fig. 1a) was obtained and preserved in 99 % EtOH for DNA extraction.
Total genomic DNA was extracted [31] from the specimen with a DNeasy Tissue Kit (Qiagen, Tokyo). After DNA extraction, the exoskeleton was mounted in Fluoromount G® as a hologenophore (Fig. 1b). The loriciferan specimen was identified as Rugiloricus sp. based on the morphology of the hologenophore.
Nearly complete 18S rRNA (18S) and 28S rRNA (28S) genes sequences were amplified by PCR using previously published primer sets and conditions [31]. All nucleotide sequences were determined by direct sequencing with a BigDye Terminator Kit ver. 3.1 (Life Technologies, Co., USA) and a 3730 DNA Analyzer (Life Technologies, Co., USA). Sequence fragments were assembled by using MEGA 5 [32]. After assembly, 18S (1872 bp) and 28S (3450 bp) sequences were deposited in GenBank under accession numbers LC032019 and LC032020.
The chi-square test in Kakusan4 [35] indicated that the base composition of each dataset was significantly homogeneous.
Before the analyses, the optimal substitution model was determined with Kakusan4 to be the general timereversible model with the gamma distribution (GTR + Γ). Phylogenetic trees were constructed by maximum likelihood (ML) implemented in raxmlGUI 1.2 [36,37], and Bayesian inference (BI) implemented in MrBayes 3.2.1 [38,39]. Nodal support for the ML tree was assessed through analyses of 1000 bootstrap pseudoreplicates. For BI, Markov-chain Monte-Carlo searches were performed with four chains, each of which was run for 1,000,000 generations, with trees sampled every 100 generations. Stationarity was evaluated by monitoring likelihood values graphically. The initial 20 % of trees from each run were discarded as burn-in, and the remaining trees were used to construct majority-rule consensus trees Theristus agilis ○ and determine the Bayesian posterior probability for each clade [39].
Phylogenetic evaluation of loricifera, scalidophora, and cycloneuralia The clade we detected consisting of Loricifera, Nematoida, and Panarthropoda received high nodal support (96/1.00), but the phylogenetic position of Loricifera within this clade remains unclear, as support for the node grouping Loricifera with Panarthropoda was quite low (63/0.66). However, the scalidophoran phyla Priapulida and Kinorhyncha together comprised a clade with high nodal support (100/1.00) to the exclusion of Loricifera, which instead grouped in a highly supported (96/1.00) clade with Nematoida and Panarthropoda. Our results thus do not support both the 'Scalidophora hypothesis, ' in which Loricifera comprises a clade with Kinorhyncha and Priapulida, and the 'Loricifera + Nematomorpha  hypothesis'. Our trees also indicated non-monophyly for Cycloneuralia, as Loricifera and Nematoida showed closer relationships to Panarthropoda than to other cycloneuralian phyla (Priapulida and Kinorhyncha).

Evaluation of synapomorphies for scalidophora and cycloneuralia
Morphological synapomorphies have previously been proposed that uniting the scalidophoran phyla (Loricifera, Priapulida and Kinorhyncha) and the cycloneuralian phyla (Scalidophora plus Nematoda and Nematomorpha). Putative synapomorphies [11] among Loricifera, Priapulida, and Kinorhyncha include (1) an introvert that has short, spinose scalids that are staggered in arrangement and triradiate in cross-section, and that has (2) inner and outer retractor muscles; (3) a compound filter of protonephridia consisting of two or more terminal cells; (4) basally thickened cuspidate spines; and (5) sensory organs (flosculi) with external cuticular micropapillae and a central pore.
The most important synapomorphy proposed for cycloneuralians is the collar-shaped circumoral brain consisting of a ring neuropil [20,21]. Our results failed to support the monophyly of either Scalidophora or Cycloneuralia, and the putative synapomorphies supporting these groups thus need to be reevaluated. With regard to the monophyly of Loricifera + Nematoida + Panarthropoda that we detected, three possible topologies among these groups (Fig. 3)   evolutionary scenarios for the three scalidophoran phyla (Priapulida, Kinorhyncha, Loricifera). If Loricifera is the sister taxon of Panarthropoda (Fig. 3a) or of Nematoida (Fig. 3b), the most parsimonious scenario is that 'scalidophoran' characters arose independently in Loricifera and in the common ancestor of Priapulida + Kinorhyncha and represent convergent characters. Alternatively, if Loricifera is basal in the Loricifera + Nematoida + Panarthropoda clade (Fig. 3c), the most parsimonious interpretation is that the common ancestor of Ecdysozoa possessed 'scalidophoran' characters, which the common ancestor of Nematoida and Panarthropoda subsequently lost. In all three topologies (Fig. 3), the most parsimonious evolutionary scenario for 'cycloneuralian' characters is that they originated once in the common ancestor of Ecdysozoa and were lost once in the common ancestor of Panarthropoda. In other words, the 'cycloneuralian' characters are plesiomorphic in ecdysozoans.

Conclusions
We reconstructed the phylogeny of ecdysozoan phyla using nearly complete 18S and 28S rRNA gene sequences, and our results suggested a new hypothesis for the phylogenetic position of Loricifera. These results did not support the previously proposed 'Scalidophora' or the 'Loricifera + Nematomorpha' clades, but detected a 'Loricifera + Nematoida + Panarthropoda' clade with rather high nodal support. Cycloneuralia emerged as paraphyletic, with high nodal support. Relationships among phyla in the 'Loricifera + Nematoida + Panarthropoda' clade were not well resolved, and phylogenetic analysis using transcriptomic or genomic data will be necessary to reconstruct the relationships within this clade, and to elucidate evolutionary transitions within Ecdysozoa.

Availability of supporting data
The data sets supporting the results of this article are included within the article and its additional files.