A new species, Fordiophyton tereticaule, from China, is described and illustrated here based on morphological and molecular evidence. It is morphologically similar to F. faberi in having erect stems, slightly oblique and membranous leaf blades, broadly ovate to suborbicular bracts, and oblong petals, but differs by the terete stems, densely puberulous petioles, and elliptic leaf blades. Our phylogenetic analyses based on plastid genome and nrITS data indicate that this new species is clustered with four Fordiophyton species of Yunnan but placed far apart from F. faberi. An updated key to the genus is also provided.

phylogeny, Sonerileae, taxonomy, Yunnan

Fordiophyton Stapf, a genus belonging to the tribe Sonerileae of the family Melastomataceae, is endemic to China and Vietnam with 15 species currently known (Chen 1984; Chen and Renner 2007; Ning and Liu 2010; Zeng et al. 2016a, b; Dai et al. 2019; Dai et al. 2020). All species occur in South China except F. phamhoangii (V.T.Pham, C.T.Vu & Ranil) T.V.Do & Ying Liu which is endemic to Vietnam. Fordiophyton faberi Stapf was designated as the type of Fordiophyton (Deng and Wu 2004). This genus is characterized as usually having raphides, 4-merous flowers and eight distinctly dimorphic and unequal stamens without ventral tubercles or a dorsal spur at the connective base (Deng and Wu 2004; Chen and Renner 2007; Zeng et al. 2016a). Recent phylogenetic studies have shown that Fordiophyton is close to Blastus Lour. and Bredia Blume (Zeng et al. 2016a; Zhou et al. 2019a, b).

During a recent survey in Malipo County, Yunnan Province, China in November 2019, we encountered an interesting Fordiophyton species with terete stems. At first glance, it looked like another two species also occurring in Yunnan, viz. F. faberi Stapf or F. strictum Diels because of its erect stems and membranous leaf blades, but those two species have very different quadrangular stems.

After careful examination of specimens of Fordiophyton species from China and Vietnam, and referring to the relevant references (Chen 1984; Pham 2000; Chen and Renner 2007; Ning and Liu 2010; Zeng et al. 2016a, b; Dai et al. 2019; Dai et al. 2020), we were unable to match this unknown species with any previously recorded species. We thus describe it here as a new species. To evaluate its phylogenetic position and relationships with congeneric species, phylogenetic analysis based on plastid genome and nrITS data was performed.

A total of 43 species from 13 genera (including 14 species of Fordiophyton) in the Sonerileae were sampled (Suppl. material 1: Table S1). Sarcopyramis napalensis Wall. and Sonerila cantonensis Stapf were selected as outgroup taxa according to Zhou et al. (2019a). All the plastid genome and nrITS sequences were downloaded from Genbank except that of this unknown species.

Silica-gel dried leaves of this unknown species were sent to Novogene (Tianjin, China) to extract total genomic DNA for library (350 bp) preparation for genome skimming sequencing. Paired-end (150 bp) sequencing was conducted on Illumina NovaSeq 6000 (San Diego, CA, USA), generating ca. 20 Gb raw data. After quality control of the raw data by fastp v.0.23 (Chen et al. 2018), ca. 6 Gb paired reads were extracted for the plastid and rDNA assembly by GetOrganelle v.1.7 (Jin et al. 2020), and the plastid genome (MK994846) of F. faberi and the rDNA (KM117261) of Wisteria floribunda (Willd.) DC. were used as the reference. Plastid Genome Annotator (Qu et al. 2019) and Geneious Prime 2019 (<www.geneious.com>) were used for the annotation of the plastid genome. The nrITS sequence of this sample was extracted from the rDNA by Geneious.

All plastid genome sequences were aligned by MAFFT v.7.4 (FFT-NS-i × 1000 strategy) after removing one inverted repeat region of each sample (Katoh and Standley 2013). The nrITS data were aligned in MEGA v.7 using MUSCLE (Edgar 2004; Kumar et al. 2016). Maximum likelihood analyses were conducted by IQTREE v.1.6 using SH-aLRT test and ultrafast bootstrap (UFBoot) feature (–alrt 1000 –bb 1000) on CentOS v.7.6 (Nguyen et al. 2015; Hoang et al. 2018).

The aligned plastid genome matrix contained 140 425 bp, of which 13 003 bp (9.26%) are variable and 3 790 bp (2.70%) are parsimony informative. The aligned nrITS matrix contained 724 bp, of which 324 bp (44.75%) are variable and 216 bp (29.83%) are parsimony informative. The best-fit model (TVM+F+R3) for plastid genome matrix and best-fit model (GTR+F+I+G4) for nrITS matrix were automatically chosen by IQTREE according to Bayesian Information Criterion. The phylogenetic analysis based on the plastid genome (Fig. 1) indicated that Fordiophyton is diphyletic and clustered with Blastus, Bredia and two species of Phyllagathis with weak support (SH-aLRT 76%, UFBoot 49%). Most species of Fordiophyton formed a well-resolved clade (Clade A), which includes the type of the genus, while only one species, F. breviscapum (Clade B), was placed sister to Blastus. Clade A can be further divided into two subclades (A1 and A2) with strong support (SH-aLRT 100%, UFBoot 100%). The unknown species is sister to F. strictum and three other Fordiophyton species in Subclade A2 with strong support (SH-aLRT 100%, UFBoot 100%). The phylogenetic analysis based on nrITS included all Fordiophyton species except F. damingshanense S.Y.Liu & X.Q.Ning and F. degeneratum (C.Chen) Y.F.Deng & T.L.Wu which supported that Fordiophyton is not monophyletic and the unknown species is clustered with four species of Yunnan and one species of Vietnam (Suppl. material 2: Fig. S1).

Figure 1. 

Phylogenetic relationships of Sonerileae based on plastid genome. The numbers near the nodes are the values of the SH-aLRT test (left) and ultrafast bootstrap (right). The unknown species is indicated by an asterisk.

The phylogenetic analyses confirmed that Fordiophyton is polyphyletic and consists of two clades (Zhou et al. 2019a, b; Dai et al. 2020). In Clade A, the unknown species was clustered with four other sympatric Fordiophyton species from Yunnan in Subclade A2. The forked and curved anther base of the longer stamens seems to be a synapomorphy of this subclade. In contrast, the species in Subclade A1 usually have longer stamens with unforked anther bases and are distributed in Guangdong, with the exception of the widespread F. faberi (occurring in almost every province of southern China) which, unusually, has longer stamens with forked anther bases like species in Subclade A2.

Morphologically, the unknown species is more similar to F. faberi than any other Fordiophyton species in both vegetative and reproductive characters, such as the erect stems, long petioles, slightly oblique and membranous leaf blades, broadly ovate to suborbicular bracts, and non-auriculate base of the calyx lobes. But the phylogenetic analyses revealed the unexpected placement of the unknown species in a different subclade from F. faberi (as shown in Fig. 1). A detailed comparison of this unknown species, F. faberi, and F. strictum (as a representative of species in Subclade A2) is provided in Table 1.

In Clade B, F. breviscapum is sister to Blastus with strong support (SH-aLRT 100%, UFBoot 100%) but far apart from two species of Kerriothyrsus C.Hansen Clade (Fig. 1). However, the phylogenetic analysis based on nrITS indicated that it is sister to the Kerriothyrsus Clade with weak support (Dai et al. 2020; Suppl. material 2: Fig. S1). Further study is needed to resolve the generic assignment of this species.

We thank Mr Xi-Bing Guo for his kind assistance with our field survey and Mr Ding-Han Cui for the beautiful line-drawing. This study was funded by Guangzhou Science and Technology Project no. 202102021219.