Selaginella guihaia (Selaginellaceae): A new spikemoss species from southern China and northern Vietnam around the Gulf of Tonkin

Abstract Selaginella guihaia sp. nov. (Selaginellaceae), a new species of spikemoss from southern China and northern Vietnam around the Gulf of Tonkin (Beibu Gulf), is described and illustrated. Morphological and molecular comparisons of the new species with other similar species (S. doederleinii, S. ornata and S. trachyphylla) are provided. The morphological and molecular evidence clearly indicates S. guihaia is a distinct species. Morphologically S. guihaia differs from other species by its obviously white–margined leaves, the ventral leaves scabrous on upper surfaces throughout the basiscopic or also rarely present on upper halves, and the ovate axillary leaves.


Introduction
Selaginella P. Beauv. (Selaginellaceae) is the largest lycophyte genus with about 700-800 species and distributed in all the continents except Antarctica (Jermy 1990;Tryon and Lugardon 1991;Zhang 2004;Zhang et al. 2013;Zhou and Zhang 2015;Zhou et al. 2016;Weststrand and Korall 2016a, b;PPG I 2016). However, the highest species diversity occurs in the tropics and subtropics. The genus is characterized by the presence of rhizophores, single veined leaves with ligule, sporangia borne axillary on the upper surface of sporophylls and bearing two types of spores (heterospory) (Webster 1992). Among several herbaria collections of Selaginella doederleinii Hieron., we found that the leaves of some specimens are obvious white-margined with ventral leaves that are often scabrous on the upper surface. We also observed and collected similar plants in the field. These turned out to be a rather common and widely distributed undescribed species in the mountainous areas of southern China (Guangxi and Hainan) and North Vietnam around the Gulf of Tonkin (Beibu Gulf ). With evidences from morphological characters and molecular analysis, we described these plants as a new species herein.

Materials and methods
Morphology characters were examined from the dried herbarium specimens studied from PE (herbaria acronyms according to Thiers 2016). All the characters were examined under stereomicroscope using NIS-Elements D 3.10 imaging software from Nikon Instruments (http://www.nikoninstruments.com). Voucher specimens (see Appendix 1) are deposited at PE.
We downloaded 60 sequences ITS and rbcL from Genbank representing 32 species in Selaginella and those species involves the major clades of the phylogenetic analysis of Selaginella (Zhou et al. 2016). In this study, we newly sequenced four species, including two samples of the possible new taxon and four samples of its putative relatives, S. ornata (Hook. & Grev.) Spring, S. doederleinii Hieron. and S. trachyphylla A. Braun. Total genomic DNA was isolated from silica-dried material using the Plant Genomic DNA Kit (Tiangen Biotech, Beijing, China) following the manufacturer's protocols. One plastid region rbcL and one nuclear region ITS were amplified for the possible new taxon and its putative closely related taxa. The rbcL region was amplified with newly designed primers rbcL 192F (5' CACGTGGACTACCGTTTGGA3') and 1324R (TACCCTCAAGAGCGGGATCA3'). The primers were designed in Primer 3.0 (Untergasser et al. 2012) using the published chloroplast genomes of Selaginella moellendorffii Hieron. (Smith 2009) and S. uncinata (Desv.) Spring (Tsuji et al. 2007). The PCR protocol of rbcL region followed Zhou et al. (2016). The ITS region was amplified using the primers and PCR protocol described in Arrigo et al. (2013). All PCR products were directly sequenced using ABI 3730XL analyzer (Applied Biosystems, Foster City, California, USA). Newly obtained sequences were assembled with Con-tigExpress and then aligned with the downloaded sequences using Clustal X v.1.83 (Thompson et al. 1997) followed by manual adjustment in BioEdit v.7.1.11 (Hall 1999). The full length of the ITS region were sequenced but only 5.8S and part of ITS2 region were used because of a large number of insertions and deletions in ITS1 and ITS2 (Zhou et al. 2016); the ambiguous regions were excluded prior to analysis as previously done in similar studies (Arrigo et al. 2013;Zhou et al. 2016). ILD (Incongruence Length Difference) test (Farris et al. 1995) was performed on PAUP* v.4.0b10 (Swofford 2002) to test if there is conflict between nuclear and chloroplast genes. The combined dataset (rbcL and ITS) were analyzed with the maximum parsimony (MP), maximum likelihood (ML) and Bayesian inference (BI) methods. MP analyses were carried out using PAUP* v.4.0b10 (Swofford 2002). All characters were weighted equally and gaps were treated as missing data. The most parsimonious trees were obtained with heuristic searches of 1000 replicates random stepwise sequence addition(RAS), with tree bisection-reconnection (TBR) branch swapping, and 10 trees from each random sequence addition were saved were used to obtain the most parsimonious trees. MP bootstrap values (MP BS ) were calculated with 1000 replicates. jModelTest 0.1.1 (Posada 2008) was used to select the appropriate substitution model for ML and BI analyses. The ML analyses were conducted using the web server RAxML-HPC2 on XSEDE (Stamatakis 2014), and ML bootstrap values (ML BS ) were calculated applying 1000 bootstrap replicates with the GTRCAT substitution model. Bayesian analyses and posterior probability (BI PP ) calculation were conducted in Mr-Bayes 3.2.6 (Ronquist et al. 2012) implemented on the CIPRES Science Gateway Portal (Miller et al. 2010). Four Markov chain Monte Carlo chains were run, each beginning with a random tree and sampling one tree every 1000 generations of 10 000 000 generations. After checking all the ESS>200 in Tracer v1.5 (Rambaut et al. 2009), the first 25% of samples were discarded as burn-in, and the remaining trees were used to calculate a 50% majority-rule consensus topology and posterior probability values.

Results
The ILD test results showed no obvious conflict existing between the two datasets, rbcL and ITS (P =0.02). Thus, the datasets were combined. The combined data matrix included up to 1460 nucleotides for each of the 36 taxa with 845 constant characters, 512 parsimony informative characters, consistency index (CI) = 0.56, retention index (RI) = 0.79. The three phylogenetic analyses (MP, ML, BI) inferred congruent topologies. The Best ML tree is presented in Figure 1.
The molecular evidence showed that two samples of Selaginella guihaia were grouped together with strong support (BS =99, PP =0.99), and then formed a moderately supported clade with S. doederleinii and S. commutata Alderw.
Our phylogenetic analyses and morphological evidence reveal that the possible new taxon Selaginella guihaia is different from the morphologically similar species S. doederleinii, S. ornata and S. trachyphylla that co-occur in the same region. The overall morphology and the growth habit of S. guihaia resemble those of S. ornata, however, the former has monomorphic (vs. dimorphic) sporophylls. Furthermore, consistent with the sporophylls variations, S. guihaia and S. ornata were separately placed into two large clades in the molecular phylogenetic tree. Although S. doederleinii and S. trachyphylla were placed closely with S. guihaia in the molecular phylogenetic analysis, the distinct whitemargined leaves of S. guihaia is different from both these species. The ventral leaves of S. guihaia are scabrous near the lower part of leaf epidermis but rarely on the upper part, whereas hte ventral leaves of S. trachyphylla are scabrous throughout the leaf epidermis.
Conservation status. We evaluated the conservation status of Selaginella guihaia according to the IUCN (2012) criteria for risk assessment: S. guihaia falls into the Least Concern (LC) category. S. guihaia is in fact known from many localities from southern China and northern Vietnam around the Gulf of Tonkin.