Two new hybrids of the genus Diplazium (Athyriaceae) from Japan

Abstract In this study, we describe the ferns Diplazium × kanayamaense hyb. nov. and D. × tsukushiense hyb. nov. and further compare them to parental species D. chinense, D. deciduum and D. fauriei in terms of morphological characteristics, plastids and nuclear DNA markers. These new hybrids have been determined to be endemic to western Japan. The International Union for Conservation of Nature and Natural Resources status was evaluated for D. × kanayamaense as endangered (EN) and D. × tsukushiense as critically endangered (CR).


Introduction
One of the important characters of Japanese fern flora is the richness of hybrids. Japanese pteridologists have been recognising morphological variations of Japanese ferns and they recognised many hybrids confuse identification of Japanese ferns (Tagawa 1959;Nakaike 1992;Ebihara 2017). However, some Japanese hybrid ferns still are not described in Latin or English languages and they do not have scientific names. Especially, Athyriaceae has still many undescribed hybrids (Ebihara 2017) and there is still not enough evidence to support a combination of their hypothesised parents.
Diplazium has been identified as the largest genus of Athyriaceae (PPG I 2016). It reportedly has 300-400 species (Wei et al. 2013;Liu et al. 2018), of which 26 species, eight varieties and 25 hybrids are recorded in Japan. Of these, 15 hybrids still do not have scientific names (Ebihara 2017). Hori and Murakami (2019) reported reticulate evolution of apogamous and sexual species in the D. hachijoense complex and found four undescribed apogamous species, based on plastid and nuclear markers. However, describing these ferns has been difficult as DNA phylogenies suggest that several undetected species are present in this complex. Otherwise, providing a description for these hybrids is easy when parents are endemic to Japan. In this study, we focused on D. ×toriianum Sa. Kurata, D. mettenianum complex (Ohta and Takamiya 1999) and D. chinense (Baker) C.Chr. Kurata (1960) has described D. ×toriianum [D. mettenianum (Miq.) C.Chr.× D. squamigerum (Mett.) Matsum.] in terms of morphological characteristics. Subsequently, Ohta and Takamiya (1999) (Ebihara 2017).
Meanwhile, Tsutsui (1988) found one undescribed fern from Kyushu, Japan. He considered the fern to be a hybrid of D. chinense and D. fauriei, based on its morphological characteristics; this claim was supported by Ebihara (2017). Further, Kanemitsu (2019) has also found one undescribed fern nearby that might be a hybrid of D. chinense and D. deciduum. However, both Tsutsui (1988) and Kanemitsu (2019) did not provide scientific names with appropriate descriptions. Thus, this study identifies two new hybrids, D. ×kanayamaense (D. chinense × D. deciduum) and D. ×tsukushiense (D. chinense × D. fauriei), with descriptions based on morphological characteristics, plastids and nuclear DNA markers.

Plant materials and DNA extraction
Total DNA for molecular analyses was extracted from silica-dried leaves using cetyltrimethylammonium bromide, as previously described (Doyle and Doyle 1990).
For conservation assessment, area of occupancy (AOO) and extent of occurrence (EOO) were estimated using GeoCAT (Bachman et al. 2011); default settings for grid size were also applied.

Plastid and nuclear DNA sequencing
We sequenced plastid trnL-F and nuclear AK1 gene following methods from Hori and Murakami (2019), but with modified conditions for polymerase chain reaction-singlestrand conformation polymorphism (PCR-SSCP) analysis of the nuclear AK1 gene. Electrophoresis of AK1 PCR products used 50% MDE gels (Lonza, Basel, Switzerland) containing 2% glycerol at 15 °C for 16 h at 300 V or 5% glycerol at 15 °C for 20 h at 300 V, followed by silver staining. To sequence the bands separated on the SSCP gels, the polyacrylamide gel was dried after silver staining by sandwiching the gel between Kent paper and a cellophane sheet on an acrylic backplate at 55 °C for 4 h. To extract the DNA, a piece of the DNA band was peeled from the dried gel by using a cutter knife and was incubated in 50 μl of TE buffer (10 mM Tris-HCl and 1 mM EDTA, pH 8.0) at 4 °C overnight. The supernatant solution was used as a template for further PCR amplification with the same primer set employed for the original PCR amplification. Sequence information obtained from voucher materials is provided in Table 1 and the Appendix 1.

Molecular analysis
The data set of plastid trnL-F phylogeny reflects what we directly sequenced from all the materials. In the dataset of nuclear AK1 phylogeny, we used all the alleles which we separately picked up from PCR-SSCP gels. Sequences were aligned using MUSCLE (Edgar 2004) and assessed with Bayesian Inference analysis using MrBayes 3.2.6 (Ronquist et al. 2012) and Maximum Parsimony (MP) analysis with MEGA X software (Kumar et al. 2018). Indels were treated as missing characters in the analysis. The best-fit model (trnL-F: GTR+G model; AK1: HKY+G model) of sequence evolution for DNA regions was selected using jModelTest 2.1. 10 (Darriba et al. 2012). Four Markov Chain Monte Carlo loops were run simultaneously and sampled every 100 of 1 million simulations. Tracer 1.7.1 (Rambaut et al. 2018) was used in examining posterior distributions of all parameters and associated statistics, including estimated sample sizes. The first 2500 sample trees from each run were discarded as burn-in periods. An MP tree was obtained using a subtree pruning-regrafting algorithm (Swafford et al. 1996) at search level 1, where initial trees were obtained by random addition of sequences (10 replicates). Confidence levels for monophyletic groups were estimated with 1,000 MP bootstrap pseudo-replicates.

Plastid and nuclear DNA phylogenetic trees
We sequenced 709-736 bp fragments of the trnL-F intergenic spacer from different specimens. The aligned trnL-F matrix was 753 bp, of which 140 bp (18%) were parsimony-informative. We then sequenced 280-520 bp of the AK1 intron for each specimen, yielding a 574 bp aligned matrix, of which 78 bp (13%) were parsimonyinformative. The accessions of DNA sequences were listed in Appendix 2. The 50% majority consensus trees resulting from Bayesian Markov Chain Monte Carlo Bayesian (B/MCMC) analysis of plastid trnL-F and nuclear AK1 gene are shown in Figures 1, 2, respectively.
Diplazium chinense, D. ×kanayamaense and D. ×tsukushiense in trnL-F phylogeny displayed haplotype 1. Diplazium ×toriianum exhibited the same haplotype 7 as  If the hybrids partly (or incompletely) shared the nuclear DNA allele of parents (in such a case, the hybrid had only nuclear allele A1A2C, A1A2K, A1A2H, A1A2I etc.), we need to assume the relationships between unknown species and present hybrids. In D. chinense, there was only one allelic type. There were different allelic combinations in Diplazium fauriei (DJ or DI) and D. deciduum (CK or CHK). However, the allele of hybrids (D.  Figure 3 Diagnosis. Diplazium ×kanayamaense has been determined to be similar to D. ×toriianum in having 1-pinnate pinnatifid pinnae curved to an apex. However, lobes of D. ×toriianum are obtuse at the apex and scales are more entire on the margin. In contrast, lobes of D. ×kanayamaense are acute at the apex and scales show small projections on their margins. Description. Terrestrial summergreen fern. Rhizomes: creeping, non-branched, black, 10-15 × 0.5-0.8 cm in diam., closely set with roots and persistent, densely clothed with old stipe bases, glabrous; fronds: 2-5 per rhizome; stipes: purplish-green, 8-11 × 0.2-0.3 cm in diam., glabrous in middle to upper sections, sparsely clothed with dark brown scales (2.0-4.0 × 0.5-1.0 mm, with small projection on margin) of basal sections, lanceolate; blades: fresh green on adaxial surface, 1-pinnate pinnatifid, 1-pinnate at the apex, 15-26.5 × 8-23 cm, ovate; rachises: purplish-green, glabrous; pinnae: 9-10 pairs, ascending, lanceolate, alternate or opposite, petiolated (2-4 mm long), serrate to lobed, curved from base to apex, acute at the apex, sessile near the apex of blades, widely spaced, lowest pair slightly reduced or the same as second, second lowest pair usually largest, 15-17 cm × 1.5-3 cm; pinnules: alternate, 9-10 pairs on the basal sections of the blade, reduced distally, ovate to lanceolate, entirely to shallowly serrated, acute at apex in basal part of blade, obtuse at the apex in the middle section of blades, vein-free, single or double, close to or reaching to the margin, 5-7 pairs in the middle lobe; the most basiscopic pinnules on the lowest pinnae: occasionally absent, clearly short, independent from the costa, 2-10 mm × 1.5-4.0 mm; sori: long linear-or J-shaped, 1.0-3.0 mm long, on the middle of veinlets, 4-10 pairs per ultimate segment, persistent; indusia: cloudy white or brown, same shape as sori, entire, persistent; spores: absent or irregular-shaped, abortive.
Specimens examined. Japan. Kyushu: Fukuoka Prefecture, Fukuoka City, Sawaraku, Mt. Kanayama,33°28'35.89"N,130°19'23.57"E,alt Distribution and ecology. Diplazium ×kanayamaense has been identified to be from Kyushu, western Japan (Figures 3, 5). The species is observed on soil under semievergreen forest near streams. This hybrid is endemic to Japan and exists in a population of approximately 124 individuals with juveniles, although parents, D. chinense and D. deciduum, were both absent near its side.
Conservation status. IUCN Red List Category. Based on estimates from Geo-CAT, the EOO of D. ×kanayamaense was 0.002 km 2 . AOO of D. ×kanayamaense was 4.0 km 2 . There were only approximately 124 individuals in the type locality and the population size is not decreasing. According to IUCN (2012) criteria, this hybrid is endangered (EN). A formal evaluation of endangerment can be summarised by the following IUCN hierarchical alphanumeric coding system of criteria: EN D.  Figure 4 Diagnosis. D. ×tsukushiense is likened to D. fauriei with fronds 1-pinnate at the apex. However, lower pinnae of D. fauriei are not lobed, finely serrated on the margin. In contrast, lower pinnae of D. ×tsukushiense are lobed deeply and 1-pinnate pinnatifid.
Conservation status. IUCN Red List Category. Based on estimates from GeoCAT, the EOO of D. ×tsukushiense is 0.001 km 2 . The known AOO of D. ×tsukushiense is 4.0 km 2 . Only 10 individuals are found in the type locality and population size is decreasing because of illegal waste dumping in forests. Therefore, this hybrid should be considered critically endangered (CR), as per the IUCN (2012) criteria. A formal evaluation of endangerment is summarised by the following IUCN hierarchical alphanumeric coding: CR B1ab (i, iv, v)+B2ab (i, iv,v)+C1+C2 a (i, ii) b+D.

Discussion
The parents of D. ×kanayamaense and D. ×tsukushiense have been determined to be D. chinense, D. deciduum and D. fauriei. These three species are rather common in western and southern Japan. Therefore, hybridisation amongst these three species is natural to occur more frequently. However, the distribution of D. ×kanayamaense and D. ×tsukushiense was very narrow in the northern part of Kyushu. We suppose mixed large populations of parents and environmental conditions supported the establishment of hybridisation in the northern part of Kyushu.
We also found that there were differences between the distribution of hybrids and parents. In the type locality of D. ×tsukushiense, the allelic composition of D. fauriei did not match D. ×tsukushiense. We surveyed a wide area around type localities, but eventually, we found the parental individual (allelic type, DI) of D. fauriei in a location 30 km away from the type locality of D. ×tsukushiense. The difference in the distribution of parents and hybrids suggested hybridisation can decrease or cause the extinction of populations of parents.
This study could not estimate the ploidy level of these hybrids because of the difficulty of cultivation. However, for parents of these hybrids, previous cytological studies were well studied by using enough individuals, including type locality and around areas of hybrids (Ohta and Takamiya 1999;Takamiya et al. 2000). Previous cytological studies reported ploidy levels and reproductive modes of parents as follows: D. chinense, diploid sexual (Mitui 1968) or tetraploid sexual (Takamiya et al. 2000); D. deciduum, hexaploid sexual (Ohta and Takamiya 1999;Takamiya 2006); and D. fauriei, tetraploid sexual or hexaploid sexual (Ohta and Takamiya 1999;Takamiya 2006). In addition, Takamiya (2006) reported D. ×tsukushiense (D. chinense × D. fauriei) as a tetraploid sterile. Therefore, hexaploid D. fariei had no relationship with D. ×tsukushiense. Our materials can be also tetraploid sterile because we collected samples from the same place as Takamiya (2006). We assumed that the ploidy level of D. ×kanayamaense can be pentaploid sterile, based on ploidy levels of D. chinense (tetraploid) and D. deciduum (hexaploid). We do not expect the existence of diploid D. chinense because Takamiya et al. (2000) showed enough cytological data of tetraploid D. chinense, which were derived from the populations that were sampled across the distribution range of D. chinense in Japan. We show the relationships of D. ×kanayamaense, D. ×tsukushiense and its relatives in Figure 6.
The respective plant size of D. ×kanayamaense and D. ×tsukushiense shows different characteristics. Diplazium ×kanayamaense is smaller than its parents D. chinense and D. deciduum, but D. ×tsukushiense is intermediate between D. chinense and D. fauriei (Table 2). In D. ×kanayamaense and D. ×tsukushiense, roots and rhizome both seem to be too weak to survive and difficult to cultivate, especially as most individuals of D. ×kanayamaense are juvenile, which are 10 cm tall or less. Therefore, environmental stability is important to maintain individual fern hybrids. The locality of D. ×kanayamaense has remained unchanged for years, whereas the locality of D. ×tsukushiense seemed to be altered due to illegal dumping activities. Thus, we expect that the discovery of these two new hybrids can assist the conservation efforts for Japanese fern flora.