Phylogenetic Position of the East Asian Endemic Genus Hattoria (Anastrophyllaceae) Based on Chloroplast DNA Sequences
Total Page:16
File Type:pdf, Size:1020Kb
Hattoria 9: 41–52. 2018 Phylogenetic position of the East Asian endemic genus Hattoria (Anastrophyllaceae) based on chloroplast DNA sequences Tomoyuki KATAGIRI & Yuya INOUE Hattori Botanical Laboratory, Obi 6–1–26, Nichinan City, Miyazaki 889–2535, Japan Author for correspondence: Tomoyuki KATAGIRI, [email protected] Abstract The phylogenetic position of the East Asian monospecific genus Hattoria has been controversial since its description in 1961. Over the years it has been considered as either a member of subfam. Jamesonielloideae in Lophoziaceae/Jungermanniaceae, subfam. Jungermannioideae in Jungermanniaceae, or Anastrophyllaceae, due to its morphological characters related to the genera with unlobed leaves, such as Jamesoniella, Scaphophyllum or Gottschelia. In the present study we provide molecular evidence that supports its inclusion in the family Anastrophyllaceae and Hattoria as a distinct genus resolved and nested in a Tetralophozia–Plicanthus clade. Our phylogenetic analyses suggest that unique morphological characters, such as unlobed and entire leaves and bracts shown in Hattoria, be considered an autapomorphy for the genus/species and also support recognition of the genus Hattoria as distinct from the other genera of the family. Some sporophyte characters including morphology of the seta are newly given for Hattoria yakushimensis (Horik.) R.M.Schust. Introduction Hattoria R.M.Schust. is a monospecific leafy liverwort genus endemic to East Asia. The generitype, Hattoria yakushimensis (Horik.) R.M.Schust., was originally described from Yakushima Island as Anastrophyllum yakushimense by Horikawa (1934). Based on its unique morphological characters the new genus, Hattoria, was proposed in Lophoziaceae s.lat. by Schuster (1961). These characters include: (1) unlobed and entire leaves and bracts, (2) strongly concave leaves mostly hyaline-margined, (3) lack of underleaves and bracteoles, (4) lack of gemmae, (5) exclusively intercalary branching, (6) distally pluriplicate perianths with the mouth contracted. However, the systematic placement of the genus has fluctuated over the last five decades. Kitagawa (1965, 1966) and Inoue (1966, 1974) supported Schuster's treatment of Hattoria in Lophoziaceae, and Hattoria was considered to be closely related to Jamesoniella or Gottschelia and placed in Lophoziaceae subfam. Jamesonielloideae. Schuster (1970) suggested its close relationships to the monospecific genus Scaphophyllum of Jungermanniaceae subfam. Scaphophylloideae, but later he included Hattoria in Jungermanniaceae subfam. Jamesonielloideae (Schuster 1979). Grolle (1971) and Váňa (1973) removed Hattoria from Jungermanniaceae subfam. Jamesonielloideae and assigned it to the 41 Jungermanniaceae subfam. Jungermannioideae, while Crandall-Stotler et al. (2009) included Hattoria in Scapaniaceae s.lat. with Lophozia and Anastrophyllum. Recent molecular evidence has given a narrower circumscription of the families Lophoziaceae and Scapaniaceae, and the new family Anastrophyllaceae was proposed by Söderström et al. (2010). The three genera, Jamesoniella, Scaphophyllum, and Gottschelia, which have been considered to be related to Hattoria, were all segregated from Lophoziaceae based on molecular evidence. Jamesoniella was revealed to be paraphyletic and synonymized in Syzygiella and placed in Adelanthaceae (De Roo et al. 2007; Feldberg et al. 2010). Scaphophyllum was synonymized in Solenostoma (Feldberg et al. 2009), and the generitype Gottschelia schizopleura was considered to belong Scapaniaceae or Cephaloziellaceae (Feldberg et al. 2009, 2013; Váňa et al. 2013; Söderström et al. 2016). In their recent study of the systematics of liverworts, Söderström et al. (2016) provisionally placed Hattoria in the family Anastrophyllaceae due to lack of molecular evidence. In the present study, we use DNA sequence data to investigate the phylogenetic position of Hattoria based on fresh samples collected in the type locality, Yakushima Island, Japan. We investigate (1) whether the genus should be placed in Adelanthaceae, Anastrophyllaceae, or Lophoziaceae, (2) whether molecular evidence supports the recognition of Hattoria as a distinct genus, and (3) which genus or species should be resolved sister to Hattoria. Materials and Methods Plant material and morphological study The molecular and morphological analyses of Hattoria yakushimensis were conducted based on fresh material with mature sporophytes collected in the type locality of the species, Yakushima Island, Japan in 2017 (Katagiri 4281 in NICH). Structural analyses were made by manual dissection of the plant material using a dissecting microscope. Molecular phylogenetic analyses Three phylogenetic markers were selected for the present analyses: chloroplast ribulose 1,5-bisphosphate carboxylase/oxygenase large subunit (rbcL), ribosomal protein S4 (rps4) genes, and the tRNA trnLUAA-trnFGAA region including the trnL intron and the trnL-trnF intergenic spacer (trnL-F). The sequences of Hattoria yakushimensis were newly obtained. The protocol for DNA extraction, PCR amplification and sequencing, and primer sequences followed earlier publications (Taberlet et al. 1991; Souza-Chies et al. 1997; Tsubota et al. 1999, 2015; Masuzaki et al. 2010; Inoue et al. 2011, 2012; Suzuki et al. 2013; Inoue & Tsubota 2014). Sequences obtained in the present study have been submitted to the DNA Data Bank of Japan (DDBJ), a member of International Nucleotide Sequence Database Collaboration (INSDC). The newly obtained accession numbers are LC376047 for rbcL, LC376048 for rps4, and LC376049 for trnL-F. The supposed ingroup and outgroup species were selected based on De Roo et al. (2007), Yatsentyuk et al. (2004), Vilnet et al. (2010), and our preliminary analyses using Jungermanniales accessions. A total of 60 OTUs were examined in the present analysis (see Appendix). 42 Sequences of the three regions were aligned separately by using the program MAFFT ver. 7.380 (Katoh & Standley 2013) with some manual adjustment on the sequence editor of MEGA ver. 7.0.21 (Kumar et al. 2016). Start and stop codons were removed, and gaps were treated as missing data. The resulting total length was 2,634 bp. and the final data matrix consisted of 60 OTUs. Phylogenetic analysis using the concatenated sequences of rbcL, rps4 and trnL-F was performed based on a maximum likelihood (ML) method (Felsenstein 1981), and the approximate unbiased (AU) test (Shimodaira 2002, 2004) in the final stage of the analysis scheme. Prior to the phylogenetic reconstruction, Kakusan4 (ver. 4.0.2016.11.07, Tanabe 2011) was used to determine the appropriate substitution model and partitioning scheme for our data based on corrected Akaike information criterion (AICc: Sugiura 1978). Phylogenetic trees were constructed using the following three program packages to obtain the candidate topologies: (1) RAxML ver. 8.2.10 (Stamatakis 2014) with ML method using the equal mean rate model among codon positions (GTR+Γ for all regions) with 1,000 heuristic searches; (2) PAUPRat (Sikes & Lewis 2001) over PAUP* ver. 4.0b10 (Swofford 2002) with the maximum parsimony (MP) method (Fitch 1971) to implement Parsimony Ratchet searches (Nixon 1999) using the Parsimony Ratchet search strategy with random weighting of each character in fifty 200 iteration runs; (3) MrBayes ver. 3.2.6 (Ronquist et al. 2012) with Bayesian inference (BI) method using the proportional model among codon positions (GTR+ Γ for all codon positions of rbcL and third codon position of rps4; K80+Γ for first and second codon positions of rps4; HKY85+Γ for trnL-F) with 10,000,000 generations, sampling trees every 1,000 generations. A 50% majority-rule consensus tree was calculated after the convergence of the chains and discarding 25% of the sampled trees as burn-in. Based on the ML criteria, re-calculation of likelihood values for each tree topology was performed by PAUP with the GTR+Γ+ Invariant model which is the best fitted model for our data. The set of candidate topologies was evaluated by the AU test and Bayesian posterior probability (PP) calculated by BIC approximation (Schwarz 1978; Hasegawa & Kishino 1989) using CONSEL ver. 0.20 (Shimodaira & Hasegawa 2001). A 50% majority-rule condensed tree for the topologies with high ranking log-likelihood values that passed both AU and PP tests was also computed by MEGA. Supporting values more than 50% obtained by CONSEL were overlaid to assess the robustness of each branch of the condensed topology: AU test (AU), bootstrap probabilities (BP), and Bayesian posterior probabilities (PP) are shown on or near each branch (AU/BP/PP). Results Phylogenetic analyses The concatenated data matrix had a total length of 2,634 bp, of which 1,058 (40.1%) were variable, and 733 (69.2% of the variable sites) were parsimony-informative. A total of 82 topologies were obtained from the three analyses: 13 ML topologies by RAxML; 68 MP by PAUPRat over PAUP*; and one BI by MrBayes. More detailed topologies were searched through the obtained trees using a log-likelihood measure. One 50% majority-rule condensed tree was obtained for the three topologies with high-ranking log-likelihood values that passed the both AU and PP tests as shown in Fig. 1. The resultant tree confirmed six clades 43 Figure 1. A majority-rule condensed tree based on chloroplast sequences (rbcL, rps4, trnL-F) with supporting values: AU test (AU), bootstrap probabilities (BP), and Bayesian posterior probabilities (PP). Asterisks indicate that all three supporting values are 100%. corresponding