Reinstatement of the corticioid genus Leifia (Hymenochaetales, Basidiomycota) with a new species L.brevispora from Hubei, Central China

Abstract The monotypic genus Leifia was previously considered to be a later synonym of Odonticium. With the morphological and phylogenetic evidence provided by an additional four East Asian specimens, we propose to reinstate Leifia as an independent genus in Hymenochaetales. Leifia morphologically differs from Odonticium by its grandinioid hymenophore with hyphal strands, numerous thick-walled cystidia with an invaginated apical end and narrowly and thick-walled basidia. The phylogeny generated from the current data set of ITS and 28S regions indicates that Leifia forms a sister clade to Odonticium. Besides the generic type Leifiaflabelliradiata in the Leifia clade, two specimens, collected from Hubei, Central China, are newly introduced as Leifiabrevispora. This new species is the second species of Leifia and differs from the generic type by its shorter basidiospores and distribution in warm-temperate to subtropical areas in East Asia. The additional two specimens, collected from Da Lat, Viet Nam, differ morphologically, both from each other and from known species of Leifia, but more samples need to be examined before further taxonomic decisions can be made.

Till now, Larsson et al. (2006) is the single paper which includes the species Odonticium flabelliradiatum in a phylogenetic analysis. Although Odonticium flabelliradiatum grouped with O. romellii (S. Lundell) Parmasto, the generic type of Odonticium and two species of Repetobasidium J. Erikss. with a full Bayesian posterior probability (BPP) support in the Rickenella Raithelh. clade of Hymenochaetales, Larsson et al. (2006) considered that this clade might not be reliable due to the lack of morphological similarities and still used the name Leifia flabelliradiata rather than O. flabelliradiatum. However, no further taxonomic opinion relating to Leifia was provided in Larsson et al. (2006).
In 2017, four specimens close to Odonticium flabelliradiatum were collected from Central China and Vietnam, which draw our attention to the taxonomic status and diversity of Leifia. Based on morphological and molecular evidence, we propose the reinstatement of Leifia and reveal a higher diversity of this genus.

Materials and methods
Specimens studied are deposited in the herbarium of Institute of Applied Ecology, Chinese Academy of Sciences (IFP). Morphological photos were taken with a digital camera Canon E12 (Tokyo, Japan) in the field. Morphological observations were made with Nikon SMZ 645 and SMZ 1000 stereomicroscopes and a Nikon Eclipse 80i light microscope (Tokyo, Japan) at magnifications up to 1000×. Microscopic procedures followed Hjortstam et al. (1987). Basidiocarp sections were prepared in Melzer's reagent, lactic acid Cotton Blue (CB) and 3% potassium hydroxide (KOH). All microscopic measurements were made in CB. When presenting the variation of basidiospore sizes, 5% of the measurements were excluded from each end of the range and are given in parentheses. The following abbreviations are used in the text: L = mean basidiospore length (arithmetic average of all measured basidiospores), W = mean basidiospore width (arithmetic average of all measured basidiospores), Q = variation in the L/W ratios between the specimens studied, n (a/b) = number of spores (a) measured from given number (b) of specimens.
The four specimens newly collected were subjected to polymerase chain reaction (PCR) directly with the Phire Plant Direct PCR kit (Finnzymes Oy, Espoo, Finland), following the manufacturer's instructions. The nuc rDNA ITS1-5.8S-ITS2 (ITS barcode) and 28S regions were amplified using the primer pairs ITS1-F (Gardes and Bruns 1993) or ITS5 and ITS4 (White et al. 1990) and LR0R and LR7 (Vilgalys and Hester 1990), respectively. The PCR procedure was as follows: initial denaturation at 98°C for 5 min, followed by 39 cycles at 98 °C for 5 s, 59 °C for 5 s (ITS region)/48 °C for 5 s (28S region) and 72 °C for 5 s, with a final extension at 72 °C for 10 min. The PCR products were sequenced at the Beijing Genomics Institute, China, with the same primers used for PCR. All newly generated sequences were deposited in GenBank (Table 1).
The current dataset for phylogenetic analysis was mainly adopted from Larsson et al. (2006), where, to avoid redundance, taxa in the Rickenella clade including Leifia flabelliradiata were mostly referred to, while taxa in other clades were representatively selected (Table 1). Sistotrema brinkmannii (Bres.) J. Erikss. was selected as an outgroup taxon. Besides taxa in Hymenochaetales, Protodontia piceicola (Kühner ex Bourdot) G.W. Martin and Exidiopsis calcea (Pers.) K. Wells from Auriculariales were also included as additional ingroup taxa. The ITS and 28S datasets were separately aligned with MAFFT 7.110 (Katoh and Standley 2013) with the G-INI-I option (Katoh et al. 2005) and then the two resulting alignments were concatenated as a single alignment deposited in TreeBASE (study no. 23768). The best-fit evolutionary model for this concatenated alignment was estimated as GTR+I+G with jModel Test (Guindon and Gascuel 2003;Posada 2008). Maximum likelihood (ML) and Bayesian Inference (BI) methods were conducted to perform phylogenetic analysis, respectively, using raxmlGUI 1.2 (Silvestro and Michalak 2012;Stamatakis 2006) and MrBayes 3.2 (Ronquist et al. 2012). In the ML analysis, bootstrap (BS) values were tested under the auto FC option (Pattengale et al. 2010). In the BI analysis, two independent runs were employed. Each run had four chains of 10 000 000 generations and started from random trees. Chain convergence was determined with Tracer 1.5 (http://tree.bio. ed.ac.uk/software/tracer/). After sampling every 1000th generation, the first 25% of sampled trees was removed, whereas the other 75% was subjected to construction of a 50% majority consensus tree and calculation of BPPs. The ML and BI methods generated congruent topologies in main lineages. Therefore, the topology generated in the ML analysis is presented and the BS values and BPPs, simultaneously above 50% and 0.7, respectively, are shown at the nodes.
To further differentiate the taxa of Leifia, the distance matrix of the alignment of their ITS sequences (5.8S and ITS2 region) were estimated using MEGA5 (Tamura et al. 2011) under the parameters of maximum composite likelihood model, uniform rates amongst sites and pairwise deletion of gaps/missing data treatment.

Results
From four studied specimens, four ITS and four 28S sequences were newly generated (Table 1). These sequences were incorporated in the dataset of Larsson et al. (2006) with an emphasis of taxa in the Rickenella clade. The current dataset included 62 taxa, each with an ITS and a 28S sequence. The concatenated alignment had 2426 characters. The BS search in the ML analysis stopped after 350 replicates. In the BI analysis, all chains were converged as suggested by the effective sample sizes of all parameters above 3300 and by the potential scale reduction factors close to 1000.
The current phylogeny ( Figure 1) recovered Hymenochaetales as a strongly supported clade (94%, 1.00). Amongst Hymenochaetales, the Oxyporus (Bourdot & Galzin) Donk clade, the Kneiffiella P. Karst. clade, the Hyphodontia J. Erikss. clade and the Hymenochaetaceae clade were recovered like those in Larsson et al. (2006), although the latter two clades received no statistical support ( Figure 1). The so-called Coltricia Gray clade in Larsson et al. (2006) here consisted entirely of corticioid species currently referred to Lyomyces P. Karst., Palifer Stalpers & P.K. Buchanan and Xylodon (Pers.) Gray, while Coltricia perennis (L.) Murrill nested within the Hymenochaetaceae clade ( Figure  1). The Rickenella clade of Larsson et al. (2006), the focus group for this study, did not group together well, but Odonticium romellii and Leifia flabelliradiata formed a strongly supported clade (91%, 1.00; Figure 1) like that in Larsson et al. (2006). The four newly sequenced specimens, also in this clade, had a closer relationship with L. flabelliradiata (100%, 1.00; Figure 1) than with Odonticium. Besides the lack of morphological similarities between Odonticium and Leifia, the branch length separating Odonticium from Leifia and related taxa also indicated that the two genera should be treated as independent.
Notes. The grandinioid hymenophore, simple-septate hyphae, distinctly thickwalled cystidia with an invaginated apical end and ellipsoid to subovate basidiospores with a straight or concave side, indicate that the new species is the second member of Leifia. Moreover, the phylogeny inferred from the ITS and 28S dataset also confirm the taxonomic position of L. brevispora. The generic type of Leifia, L. flabelliradiata, differs from L. brevispora by having longer basidiospores (4.5-5.5 × 2-2.5 μm) and a distribution in Europe (Eriksson et al. 1981).

Discussion
In this study, the newly generated ITS and 28S sequences were incorporated into the dataset of Larsson et al. (2006) and, in the resulting phylogeny (Figure 1), clades are labelled A-F as in Larsson et al. (2006). The differences of phylogeny observed between the current study and Larsson et al. (2006) might reflect that the ITS and 28S dataset itself is not enough to reliably resolve the relationships within Hymenochaetales. Similar to Larsson et al. (2006), Leifia formed a sister lineage to Odonticium with strong support in the current phylogeny ( Figure 1). The five taxa of Leifia and the two of Odonticium were clearly separated and recovered as independent, fully supported clades. Morphologically, Leifia is well distinguished from Odonticium by its grandinioid hymenophore with hyphal strands, numerous thick-walled cystidia with an invaginated apical end and narrowly and thick-walled basidia (Eriksson et al. 1981). Therefore, we propose to resurrect Leifia as an independent genus in Hymenochaetales.
Amongst the four newly sequenced taxa in Leifia clade, LWZ 20170820-46 and LWZ 20170820-48 represent the new species L. brevispora, while LWZ 20171015-36 and LWZ 20171015-38, both collected from Bidoup Nui Ba National Park, Da Lat, Viet Nam, seem to represent two undescribed taxa. LWZ 20171015-36 differs from L. brevispora and L. flabelliradiata by fairly thick basidiocarps and LWZ 20171015-38 differs by having basidia and basidioles that swell in KOH. Moreover, LWZ 20171015-38 grows on fallen branches of Pinus, while the other three specimens were all collected from angiosperm substrates. Although the morphological characters of LWZ 20171015-36 and LWZ 20171015-38 are unique in Leifia, we feel more samples need to be examined before describing them as new species.