Morphological and Molecular Characterizations of Three Species of the Genus Synura (Synurales, Chrysophyceae) from China

Abstract

Three species of the genus Synura from China are described here. Morphological observations and molecular phylogenetic analyses were conducted for three specimens collected at different locations in China. The described morphological features included cell size, scale size, spines, keels, and struts. Molecular analyses based on multiple genetic markers (SSU and LSU rDNA and internal transcribed spacer rDNA) were used to determine the phylogenetic positions of the three Synura species. Morphologically, specimen GZ201017 collected in Guizhou Province was characterized by a well-developed keel and lanceolate scales; specimen SX210304 collected in Shanxi Province was characterized by a less-developed keel and poor silicification; and specimen GD201126 collected in Guangdong Province was characterized mainly by spines with blunt ends or two small teeth on the tips. The morphotypes GZ201017, SX210304, and GD201126 corresponded to the original descriptions of Synura petersenii, S. glabra, and S. longitubularis, respectively. This discovery laid a foundation for the molecular phylogeny of the genus Synura and an enhanced understanding of Synura diversity and distribution in China.

1. Introduction

The genus Synura was established by Ehrenberg in 1834 and contains colonial flagellates whose cells are covered with imbricate silica scales, the ultrastructure of which is the primary attribute used to distinguish species [1,2]. To date, the genus Synura has been recognized as a distinct genus which is exclusively distributed in freshwater. There are currently 57 species accepted in the AlgaeBase taxonomy, and molecular evidence is accessible for 34 species [3,4]. The cells of Synura often have two to four types of scales, consisting of caudal scales, spineless body scales, spine-bearing body scales, and tubular apical scales [5]. Species identification is conducted according to the following characteristics: scale size, keel size, number and distance of struts, spine size, spine tip, and hexagonal meshwork. Generally, body scales are mainly used to distinguish amongst species [6,7,8]. Early classifications of Synura species were made mainly by observing features such as cell size and shape using light microscopy [1]. Subsequently, the application of electron microscopy has revolutionized the classification of Synura species [9,10,11,12,13,14,15,16,17]. However, the accuracy of the systematics and biodiversity of Synura cannot be guaranteed by relying simply on traditional morphological observations. Indeed, with the rapid advances in molecular technology, the section divisions of the genus Synura have been improved. The first proposed classification scheme divided the Synura genus into two sections on the basis of scale ultrastructure: Petersenianae and Uvellae [9]. Molecular reconstruction data then showed that section Synura and section Peterseniae represented two distinct evolutionary subclades on the phylogenetic tree, which was consistent with their morphological groupings [18]. Additional subgenera have been identified in subsequent studies [10,13,15,17]. According to the classification in 1974, the genus contained three sections: Synura, Petersenianae Petersen et Hansen ex Balonov et Kuzmin (1974: 1682), and Lapponicae Balonov et Kuzmin (1974: 1685) wherein section Lapponicae replaced section Uvellae proposed in 1956 [7,9,10]. Péterfi and Momeu accepted the classification scheme proposed in 1974, and the series Splendidae was recognized [10,15]. In addition, in 2013, Škaloud et al. proposed a five-section classification comprising Echinulatae, Peterseniae, Spinosae, Splendidae, and Synura, but no formal descriptions were given [11]. In 2016, new sectional ranks were proposed wherein Jo et al. suggested that the genus Synura should be further divided into the following three sections: Synura, Peterseniae, and Curtispinae [7]. As currently recognized, these three sections are clearly distinguishable, and the molecular classification scheme is consistent with that based on morphology [7,8]. More recently, Škaloud et al. and Jo et al. studied the diversity within the genus Synura, but this genus needs to be further resolved [7,8,16,19,20]. Furthermore, genetic markers, including SSU, LSU, and ITS, have been widely used in the phylogenetic analysis of Chrysophyta [7,8,18,19,20,21,22,23]. Therefore, three genes were selected to investigate species within the Synura genus in this study.

The genus Synura is distributed widely around the world, but rarely has been reported in Asia [6,24]. Some species are widespread, while others are extremely rare and distributed in specific habitats, even inhabiting cold and acidic conditions [24,25,26]. The first report of the genus Synura in China was by Skvortsov (1961), but that report was not based on scale ultrastructure. Since then, several additional species of the genus Synura have been reported in China based on their ultrastructure [9,27,28,29,30]. In addition, Pang and Wang described the stomatocyst of Synura petersenii in 2012 and described a new species S. morusimila in 2013 [31,32]. Wei and Yuan reported S. glabra and S. petersenii in 2014 and S. bjoerkii, for the first time, in 2015 [33,34]. To date, the Synura studies in China have lacked the molecular evidence that is necessary to enhance and validate the research on Synura diversity and distribution in this region.

Three Synura specimens were collected in Guizhou, Guangdong, and Shanxi provinces, respectively. Molecular phylogenetic analysis was conducted on the basis of concatenated SSU rDNA, LSU rDNA, and ITS rDNA sequences, and the identification results were verified using morphological characteristics. The aims of this study were to (1) describe three species of the genus Synura using a combination of morphological and molecular techniques; (2) infer the phylogenetic relationships among Synura species in this study; (3) compare morphological characterizations among the three species and closely related species; and (4) comprehend the species diversity and geographical distribution of Synura in China. This study has significantly improved our knowledge of the diversity of Synura species and their regional distribution in China in addition to serving as a regional resource for the biodiversity of freshwater Chrysophyta.

2. Materials and Methods

2.1. Collection and Culture

Three specimens (GD201126, GZ201017, and SX210304) were collected from Guangdong Province, Guizhou Province, and Shanxi Province in China (Figure 1,

Table 1. Collection information and voucher numbers for taxa analyzed in this study.

Isolate	Locality with Longitude and Latitude	Habitat Type	Collection Date	Collector	Voucher Number
GD201126	the Xianxia waterfall, Huizhou, Guangdong Province, China (23.6494° N, 113.8851° E)	a pond near a waterfall	26 November 2020	Nini Cui	SXU-GD201126
GZ201017	the Xinpu Wetland Park, Zunyi, Guizhou Province, China (27.7024° N, 107.0180° E)	a lake in the park	17 October 2020	Qi Liu	SXU-GZ201017
SX210304	the Long Korean Road, Changzhi, Shanxi Province, China (36.0619° N, 113.0049° E)	a lake near a factory	4 May 2021	Chen Su	SXU-SX210304

). A plankton net with a mesh size of 20 μm was used to collect samples, and the samples were transferred to the laboratory as soon as possible. Single strains were isolated by Pasteur capillary pipette under an inverted microscope and placed into a uni-algal culture in DY-IV medium with MES. The pH of the medium was 6.8. The cultures were maintained at 14–16 °C under a 12 h: 12 h of light: dark cycle with 1000 lux of illumination. The incubation period was 8 days, after which the culture was expanded. Voucher specimens were preserved in 4% formaldehyde. Voucher specimens were deposited in the Herbarium of Shanxi University (SXU), Shanxi University, Taiyuan, Shanxi Province, China.

2.2. DNA Extraction, Amplification, and Sequencing

Algal sediments pellets were obtained after centrifugation of 1 mL culture in the exponential growth phase for 5 min at room temperature. A plant DNA extraction kit (Sangon Biotech, Shanghai, China) was used to extract the total DNA from the pellets. SSU, LSU, and ITS rDNA were amplified using polymerase chain reaction (PCR) in a total volume of 50 μL, containing 37.75 μL ddH₂O, 5.0 μL 10 × buffer, 4.0 μL 2.5 mM dNTPs, 0.25 μL Taq DNA polymerase (Sangon Biotech, Shanghai, China), 1.0 μL of each primer (10 mM), and 1.0 μL of genomic DNA. The SSU gene was PCR amplified using the 18S_1F, 18S_9R, 18S_4F, and 18S_12R primers, and the LSU gene used the 28S_25F, 28S_861R, 28S_736F, 28S_1440R, 28S_1228F, 28S_2160R, 28S_2038F, and 28S_2812R primers [35,36]. According to Wee et al. and White et al., the amplification of the ITS marker used the primers KN1.1 and ITS4 [22,37]. The SSU, LSU, and ITS genes all used the following cycle: 94 °C for 5 min, 35 cycles of 94 °C for 30 s, annealing temperature for 30 s and 72 °C for 2 min, and final 72 °C for 10 min. The reactions were undertaken in a MyCycler thermal cycler (Bio-Rad, Hercules, CA, USA). The annealing temperature changed depending on the primer. The temperature for 18S_1F, 18S_9R, and 28S_736F, and 28S_1440R was set to 49 °C and that for 28S_1228F, 28S_2160R, 28S_2038F, and 28S_2812R was set to 51 °C. The temperature for 18S_4F and 18S_12R was set to 54 °C and that for 28S_25F and 28S_861R was set to 47 °C. The annealing temperature of the primers KN1.1 and ITS4 was set to 48 °C. The PCR products, along with their amplification primers, were sent to BGI Tech Corporation (Beijing, China) where they were sequenced on an ABI 3730XL sequencer. The sequence data of SSU rDNA (OM267653, OM285147, and OM267663), LSU rDNA (OM267664, OM285146, and OM285148), and ITS rDNA (OP811172, OP811173, and OP811174) were submitted to GenBank.

2.3. Phylogenetic Analysis

Using MAFFT version 7, the sequence data obtained by sequencing in BGI Tech Corporation were aligned with those of other Synura species and outgroup taxa, downloaded from GenBank, by nucleotide blasting [38]. The SSU rDNA, LSU rDNA, and ITS rDNA molecular data of the genus Synura were collected (49 SSU rDNA sequences from 26 species, 44 LSU rDNA sequences from 24 species, and 56 ITS rDNA sequences from 30 species). The sequences of SSU, LSU, and ITS were concatenated on the basis of the methods of Zhang et al. (2020) [39]. The concatenated SSU, LSU, and ITS sequence set was 5418 base pairs, of which 1400 (25.84%) were variable and 1192 (22.00%) were parsimony informative. T, C, A, and G exhibited average compositions of 26.3%, 20.0%, 27.5%, and 26.2%, respectively. BioEdit v7.2.1 was used to cut the untrimmed ends to produce the same length alignments [40]. MEGA 5.0 was employed to calculate the pairwise genetic P-distances between individual samples [41]. The outgroup taxa Neotessella volvocina and N. lapponica were chosen on the basis of previous studies [7,8]. The appropriate model was built using the software PartitionFinder 2 with all algorithm and BIC criterion (for BI: Subset (1)(2)(3) = GTR + I + G; for ML: Subset (1)(2)(3) = GTR + I + G) [42]. Maximum likelihood (ML) phylogenies were inferred using the IQ-TREE under edge-linked partition model for 5000 ultrafast bootstraps, as well as the Shimodaira–Hasegawa-like approximate likelihood-ratio test [43,44,45]. Moreover, Bayesian inference (BI) phylogenies were inferred using MrBayes 3.2.6 under the partition model (2 parallel runs, 3,000,000 generations), in which the initial 25% of sampled data were discarded as burn-in [46]. The Figtree 1.4.2 software was applied to redact the resultant phylogenetic trees (http://tree.bio.ed.ac.uk/software/figtree/ (accessed on 5 December 2022)). Adobe Illustrator CS5 (Adobe System, San Jose, CA, USA) was used to optimize the graphics of all trees.

2.4. Morphological Observations

For morphological observations, fresh specimens were observed under a BX-51 Olympus (Olympus, Tokyo, Japan) fitted with a digital camera (DP72) for imaging. For electron microscopy, the cultured Synura strains were directly transferred and dried onto aluminum stubs in an oven. The aluminum stubs were sputter-coated with gold for 40 s and examined with a scanning electron microscope (Phenom-prox, Eindhoven, The Netherlands).

3. Results

3.1. Molecular Analysis

Pairwise distances based on concatenated SSU, LSU, and ITS sequences between the examined species and the outgroup taxa are listed in Table S1. Two species: Neotessella volvocina and N. lapponica were used as outgroups to root the phylogenetic tree. Similar topologies were produced by both methods, i.e., BI and ML. Therefore, only the BI trees containing all of the supporting values obtained on the nodes are displayed in Figure 2.

In the phylogeny based on multiple genetic markers (SSU, LSU, and ITS) (Figure 2), Synura was segmented into four primary clades: A, B, C, and D. The section Synura was segmented into two clades: C and D. The two clades were composed of the members Synura splendida and S. uvella, both with strong supporting values of 1.00/100. Clade A was divided further into A1, A2, and A3. S. macracantha diverged at the bottom of Clade A. Subclade A1 consisted of S. petersenii, S. americana, S. macropora, S. borealis, S. laticarina, S. conopea, S. soroconopea, S. sungminbooi, S. heteropora, S. lanceolata, S. hibernica, S. truttae, S. glabra, and S. kristiansenii. The species collected from Guizhou was clustered together with S. petersenii with a high supporting value (1.00/100), and the distance between them was smaller than the intraspecific distance (0.0067 vs. 0.0087). S. americana was not monophyletic, and S. macropora was revealed as closely related to S. americana (1.00/100). S. borealis was closely related to S. laticarina (1.00/99). The S. sungminbooi strains shared a tight relationship with S. soroconopea and S. conopea. A high supporting value (1.00/96) indicated that the S. heteropora and S. lanceolata strains were strongly connected to S. truttae and S. hibernica. Specimen SX210304 collected from Shanxi Province was in a highly supported clade (1.00/100) with S.glabra, and they formed a monophyletic lineage. In addition, the distance between specimen SX210304 and S. glabra was smaller than the intraspecific distance within Synura (0.0064 vs. 0.0087). The monophyletic S. kristiansenii strains had strong supporting data (1.00/100). Subclade A2 comprised S. asmundiae and S. bjoerkii, which were related with strong supporting value (1.00/95). Clade B was further subdivided into B1, B2, and B3. S. mammillosa was closely related to S. leptorrhabda and S. echinulata with strong support (1.00/100). S. multidentata diverged at the base of Clade B1. S. sphagnicola represented a monophyletic lineage with strong supporting data (1.00/100). S. mollispina and S. spinosa were very closely related (1.00/100), and S. curtispina was closely related to S. longitubularis (1.00/100). The species collected from Guangdong Province was closely related to S. longitubularis with high support (1.00/100). Additionally, the distance between the species collected from Guangdong Province and S. longitubularis was smaller than the intraspecific distances within the Synura genus (0.0022 vs. 0.0087). S. synuroidea branched from Clade B2 at its base. S. punctulosa made up Subclade B2 and the strain diverged at the bottom of Clade B.

3.2. Morphological Characterization

We observed the scale ultrastructures of the specimens (GZ201017, SX210304, and GD201126) by light microscope and scanning electron microscope. The results were consistent with the molecular results, which assisted in confirming the taxonomic status of these three new species records as Synura petersenii, S. glabra, and S. longitubularis. Of these, two species (S. petersenii and S. glabra) belonged to section Peterseniae and the other belonged to section Curtispinae. The scale ultrastructural characteristics of section Peterseniae and section Curtispinae of the genus Synura considered in this study are listed in Tables S2 and S3.

Specimen GD201126 of S. longitubularis was collected from the Xianxia waterfall (23.6494° N, 113.8851° E) in Guangdong Province. Its morphological features are shown in Figure 3 (1–3) and Figure 4. Its colonies were oval and 44–94 × 28–53 μm, and its cells were globular and 11–19 × 10–15 μm in diameter, with ovate or ellipsoidal scales covering the entire cell surface. Its body scales, 2.7–3.5 × 2.0–2.9 μm, were arranged in the shape of petals with a thickened posterior rim surrounding roughly one-half of the scale perimeter. A hexagonal meshwork was present on the distal areas of its scales. Its spines were 0.6–1.5 × 0.2–0.4 μm, with distal ends tapering to acute angles or terminating in two small teeth on the tip.

S. petersenii samples collected from the Xinpu Wetland Park (27.7024° N, 107.0180° E) in Guizhou were saved in the Herbarium of Shanxi University (SXU), Shanxi University, Taiyuan, Shanxi Province, China. The morphological characteristics of the sample are illustrated in Figure 3 (4–6) and Figure 5. Its colonies were petaloid and 29–38 μm in diameter, and cells were pyriform, 12–17 × 11–14 μm, and entirely covered by lanceolate scales. Its body scales, 2.7–4.2 × 1.2–2.2 μm, had well-developed keels and a great many struts on the base plate. A posterior rim covered one-half to two-thirds of the scale perimeters. The keels, 2.0–3.2 × 0.4–0.8 μm, frequently ended in acute tips and were adorned by medium-sized pores. The keels were wider in the anterior region. Numerous small pores decorated its basal plate. Numerous struts (19–26) extended regularly from the keel to the edge of the scale and interconnected the transverse folds irregularly. Struts were spaced 0.13–0.34 μm apart.

The morphological characteristics of the S. glabra samples collected from the Long Korean Road (36.0619° N, 113.0049° E) in Changzhi, Shanxi Province, are illustrated in Figure 3 (7–9) and Figure 6. Its colonies were globular, 47–58 μm in diameter, and its cells were oval, 7–11 × 6–9 μm. Its body scales were ovate, 2.2–3.1 × 1.3–2.2 μm, with narrower edges. The keels, 1.1–2.1 × 0.3–0.6 μm, were less-developed and narrow. Its keels were decorated with medium-sized pores and were concave. The transverse folds were not connected with any of the struts (16–20), which extended regularly from the keels to the edges of the scales. Struts were spaced 0.14–0.28 μm apart.

4. Discussion

The classification of Synura species has been based mainly on the ultrastructure of body scales, and the classification scheme changed four times from 1956 to 2016 [7,9,10,11]. Similarly, the classification of Tessellaria also changed continually in the last century. Playfair first established the genus Tessella in 1915 but changed the illegitimate genus name to Tessellaria in 1918 [47,48]. Subsequently, Playfair created a new family Tessellariaceae for the genus in 1921 [49]. Updating the classification scheme in 1974, Petersen and Hansen created section Lapponicae for Synura lapponica [10]. Kristiansen reclassified the genus Tessellaria as part of the family Synuraceae in 2007 [6]. However, this family classification was generally ignored [50,51,52]. In 2013, S. lapponica was transferred to Tessellaria [11]. Škaloud et al. revised the previous classification scheme and removed the genus Tessellaria from the family Synuraceae on the basis of ultrastructural differences in scales and molecular evidence [11]. In 2016, Jo et al. proposed replacing the illegitimate name Tessellaria with Neotessella and proposed new sectional ranks [7]. In this study, we selected two species of the genus Neotessella as outgroups in our molecular phylogenetic analysis.

Previous taxonomic studies of the genus Synura have been based primarily on traditional morphology focusing on the ultrastructure of body scales [2]. In recent years, molecular tools have made great contributions to assessing phylogenetic placement [7,21]. The first molecular analysis was performed by Wee et al. to investigate genetic variability in S. petersenii [22]. In 2010, Boo et al. confirmed the high degree of cryptic, species-level diversity in the S. petersenii complex, creating the first multigene phylogeny [53]. The first taxonomic assessment was performed by Škaloud et al.; six genetic lineages of the S. petersenii species complex were recognized as separate species on the basis of morphological data and multiple genetic markers [20]. Škaloud et al. proposed three new species: S. vinlandica, S. fluviatilis, and S. cornuta, and discovered significant cryptic diversity within the core lineages of Synura [8]. Four new species, including S. borealis, S. heteropora, S. hibernica, and S. laticarina, were described and characterized in 2014 [16]. Jo et al. performed phylogenetic analyses on the basis of multiple gene sequences and proposed new sectional divisions: Synura, Peterseniae, and Curtispinae [7]. These publications clearly show that phylogenetic analyses using gene sequences have become increasingly popular and important in the identification and classification of Synura species.

In this study, we focused on the lineage of the genus Synura. According to the phylogenetic trees based on concatenated SSU, LSU, and ITS sequences, specimen GD201126 and S. longitubularis, a new species described by Jo et al. in 2016, were closely related [7]. Moreover, phylogenetic studies showed that S. longitubularis and S. curtispina shared a close relationship, which was consistent with phylogenetic trees presented by Jo et al. and Škaloud et al. [7,8]. Additionally, the results of maximum likelihood (ML) and Bayesian inference (BI) phylogenies based on concatenated SSU, LSU, and ITS sequences positioned GZ201017 and S. petersenii and SX210304 and S. glabra, respectively, within the same clades. Furthermore, both S. petersenii and S. glabra were assigned to section Peterseniae. The taxa collected from Guizhou was identified as S. petersenii, specimen SX210304 was confirmed to be S. glabra, and specimen GD201126 was identified as the first report of S. longitubularis in China, mainly on the basis of the molecular evidence. Note, however, these species identifications require more specimens and molecular data to improve their reliability, a task to be completed in future studies.

Traditional taxonomy relies on morphological observations. A key classification feature of the genus Synura is the presence or absence of keels or spines on the base plates of the scales. Therefore, the taxonomic statuses of the three Synura species in this study were further verified by morphological observations. The typical characteristics of S. petersenii are a well-developed keel and many struts, whereas the body scales in S. glabra have a less-developed keel and low silicification in S. glabra. S. longitubularis is characterized mainly by a blunt spine tip or a spine tip with 2–3 teeth. In 1929, Korshikov measured the size of S. petersenii scales to be 4 µm × 2 µm, which was consistent with the observations of Petersen and Hansen (3.6–4.7 × 2.2–2.5 µm) and Škaloud et al. (3.8–4.6 × 1.8–2.3 µm) [9,20,54]. The dimensions of S. glabra were 2.5–3.3 × 1.7–2.1 µm and 2.7–3.0 × 2.0–2.2 µm in the description of Petersen and Hansen and Škaloud et al., respectively [9,20]. In the description by Jo et al. in 2016, the body scales of S. longitubularis were 2.2–3.9 × 1.6–2.4 μm, and the spines were 0.9–1.6 × 0.3–0.5 μm [7]. The morphologies of GZ201017, SX210304, and GD201126 collected in this study were consistent with the characteristics of S. petersenii, S. glabra, and S. longitubularis, respectively. The scales of specimen GZ201017 and S. petersenii were similar in size, but the specimen in this study had smaller scales, longer keels (2.0–3.2 × 0.4–0.8 µm), and fewer struts (19–26). Specimen SX210304 was similar to S. glabra, exhibiting a less-developed keel and several struts (16–20), and the struts were spaced 0.14–0.28 µm apart and were not interconnected by transverse ribs. In addition, the specimen representing the first report of S. longitubularis in China had body scales arranged in the shape of petals. Its spines were 0.6–1.5 × 0.2–0.4 μm in size, and the distal ends were blunt or had two small teeth on the tips. Morphological comparisons of silica scales revealed that all novel clades were broadly similar to the previously described taxa, with minor differences. Although morphology has historically been the central basis of Synura classification, similar morphological features and cryptic taxa may lead to incorrectly estimating the species diversity. Siver and Lott reported a high morphological diversity in scales of the genus Synura and recognized the necessity for molecular analyses [55]. Molecular approaches are highly sensitive to species, and morphological and molecular approaches can be used to complement each other. In some cases, molecular evidence must be referenced in the identification of the species’ taxonomic statuses. Indeed, it is necessary to conduct molecular analyses of the Synura species to fully comprehend the species diversity of the genus in China.

The SSU, LSU, and ITS genes are frequently used as DNA barcodes in molecular phylogenetic analyses of freshwater Chrysophyta, which have also been widely used in the molecular studies of the genus Synura [7,8,16,20,21]. However, our knowledge of Synura diversity remains meager and limited. S. elipidosa and S. falcata collected in China were validly described by Skvortzov in 1961, but without electron microscopy analysis [27]. Until 2012, S. elipidosa was assigned to section Peterseniae and S. falcata to section Synura, but these assignments were not based on molecular data [20]. Similarly, the studies of statocysts by Pang and Wang did not involve molecular evidence [31,32]. The history of phylogenetic revision based on morphological observations emphasizes the critical need for molecular phylogenetic analyses of Synura species worldwide, especially in China. This study, by supplying the molecular sequences of three Synura species from China, has contributed data and a theoretical foundation for future taxonomic studies and further revealed the Synura species and geographic diversity in China.

5. Conclusions

Morphological features and molecular phylogenetic studies both agree in their assignments of taxonomic statuses of three Synura species in China. GZ201017, a specimen collected in Guizhou Province, was morphologically distinguished by a well-developed keel and lanceolate scales, whereas poor silicification and less-developed keels were exhibited by specimen SX210304. In addition, specimen GD201126 was distinguished mainly by spines whose distal end was blunt or had two small teeth on the tips. Given that the morphologies of the three specimens in this study were slightly different from their respective type specimens, the identification of these taxa collected from China was based primarily on molecular evidence and further verified by morphological characteristics. The molecular information of three Synura species from China was supplied in this study, providing molecular evidence and a theoretical basis for molecular phylogenetic analyses of the freshwater Chrysophyta genus. This will help enrich our knowledge of the species diversity and geographical distribution of the genus Synura in China.