Taxonomic position of Eriocycla (Apiaceae): insights from molecular and morphological evidence

Abstract

Speciation arises from multifaceted factors, making phenotype-based classifications unreliable. Integrative taxonomy has advanced significant breakthroughs in taxonomically challenging groups like Apiaceae, which is characterized by highly convergent morphological traits across species. The genus Eriocycla (Apiaceae) has long presented persistent taxonomic uncertainties. While morphological similarities initially supported Eriocycla as Seseli sect. Eriocycla, phylogenetic studies consistently resolve Eriocycla within the tribe Echinophoreae, contrasting with Seseli (tribe Selineae). Integrated morphological and molecular analyses were conducted here to resolve this taxonomic conflict. Phylogenetic reconstructions based on nuclear ribosomal DNA and plastomes all supported that Seseli delavayi and Seseli nortonii formed a stable monophyletic group with Eriocycla nuda and Eriocycla pelliotii within Echinophoreae, separate from Seseli. Plastome comparisons across 14 taxa revealed structural conservation in E. nuda, E. pelliotii, S. delavayi, and S. nortonii, particularly in inverted-repeat and single-copy regions, distinct from that of other Seseli species. A unique inversion involving the trnY–GUA, trnD–GUC, and trnE–UUC genes was detected in E. nuda and E. pelliotii but absent in S. delavayi and S. nortonii. Shared morphological characteristics, including glabrous stem bases, basally free bracteoles, and prominent calyx teeth, further support their affinity with Eriocycla. We therefore propose to recognize Eriocycla as a separate genus rather than as Seseli sect. Eriocycla and reclassifying S. delavayi and S. nortonii into it. In conclusion, this study not only revealed the phylogenetic position of the tribe Echinophoreae but also resolved the long-standing taxonomic controversy surrounding Eriocycla and Seseli.

This study addresses the long-standing taxonomic uncertainties surrounding the genus Eriocycla (Apiaceae), which has often been classified within Seseli due to morphological similarities. Using both morphological and molecular data, including nuclear ribosomal DNA and plastomes, we resolve Eriocycla as a distinct genus within the tribe Echinophoreae. Phylogenetic analyses confirm that Eriocycla species, along with Seseli delavayi and Seseli nortonii, form a monophyletic group separate from Seseli. Plastome comparisons and shared morphological traits further support this reclassification. This work clarifies the evolutionary relationships within Echinophoreae and resolves the taxonomic issues surrounding Eriocycla.

Introduction

The plants of Apiaceae family, widely distributed over the world, are characterized by (compound) umbels and mericarps (Pimenov and Leonov 1993). Due to convergent evolution and hybridization (Dávalos et al. 2014, Zou and Zhang 2016), related genera often exhibit highly similar morphological features. Consequently, traditional classification systems based solely on phenotypic traits frequently fail to reflect the true phylogenetic relationships within the family (Drude 1898). With the advent of molecular biology, an increasing number of molecular markers, including the internal transcribed spacer (ITS), external transcribed spacer (ETS), and matK gene, have been employed to investigate phylogenetic relationships in Apiaceae (Spalik et al. 2004, Valiejo-Roman et al. 2006, Zhou et al. 2009, Downie et al. 2010, Doğan Güner and Duman 2013, Xu et al. 2021, Song et al. 2025). Numerous groups have been demonstrated to be polyphyletic, including the genera Peucedanum L., Ligusticum L., and Angelica L. (Zhou et al. 2020, Jiang et al. 2022, Liu et al. 2024a, 2024b, Ren et al. 2025). These genera share characteristics such as broad geographical distributions, high species diversity, and indistinct morphological boundaries with related groups. While phylogenetic methods can reveal relationships among species, they often fail to provide stable diagnostic characters. Therefore, the integrative taxonomy (synthesizing morphological, phylogenetic, genomic, and complementary methods) has emerged as a major trend in recent taxonomic studies (Dewar et al. 2025, He et al. 2025, Yuan et al. 2025). Integrative taxonomy not only reconstructs evolutionary relationships from multifaceted evidence but also establishes stable morphological boundaries, providing a foundation for field identification.

As one of the largest genera in Apiaceae, Seseli comprises 125–140 species and is distributed in the Old World from western Europe and northwestern Africa to China and Japan (Pimenov and Leonov 1993, 2004, Plunkett et al. 2018). The genus was characterized by: absence or abundance of linear, acuminate bracts exceeding pedicels in length, calyx teeth absent, and ovate striate fruits (Linné and Salvius 1754). Due to its wide distribution across diverse climate zones and diverse habitats, this genus exhibits remarkable morphological diversity, and its taxonomy has been debated since its establishment (Drude 1898, Schischkin 1950, Pimenov et al. 2012, Pimenov 2017, Liu et al. 2024a, 2024b). A central controversy involves whether the genus Eriocycla should be merged into Seseli.

Eriocycla Lindl., established with Eriocycla nuda Lindl. as the type species, was initially characterized by Lindley (1835) based on its densely pubescent fruits, prominent ribs, solitary vittae in each furrow, and two vittae on the commissure. However, overlapping fruit characteristics with Pituranthos Viv., particularly in mericarp morphology, led Clarke (1879) to synonymize Eriocycla under Pituranthos. Wolff (1927) reinstated and expanded Eriocycla, transferring six species into this genus. Comparative fruit anatomical studies of nine Eriocycla species revealed that Eriocycla and Seseli members do not differ in carpological features up to the genus level (Pimenov and Kljuykov 2000). Consequently, Eriocycla was reclassified as Seseli sect. Eriocycla (Lindl.) Pimenov & Kljuykov, with E. nuda designated as the section’s type species, and Seseli nortonii Fedde ex H.Wolff was treated as a synonym of E. nuda due to morphological similarities. According to extensive morphological examination of specimens, Pimenov (2017) synonymized all Chinese Eriocycla members with Seseli.

Early phylogenetic analyses using nrITS, rpl16, and rps16 intron regions resolved most Seseli species within the tribe Selineae (Zhou et al. 2009, Banasiak et al. 2013). However, Lyskov et al. (2022) employed ITS + ETS and psbD–trnT sequences to reconstruct the phylogeny of Seseli, placing E. nuda within the tribe Echinophoreae. Based on ITS sequences, Degtjareva et al. (2011) inferred the phylogenetic positions of 36 Seseli species, revealing that Seseli delavayi Franch., a narrow endemic to the Hengduan Mountains, was resolved within Echinophoreae.

According to field observations and the checking of collected materials in the type locality, we observed that S. nortonii and S. delavayi showed shared characteristics with E. nuda and Eriocycla pelliotii, including glabrous stem bases, basally free bracteoles, and prominent calyx teeth—features distinct from Seseli tortuosum and other Seseli taxa (Doğan Güner and Duman 2013). Recent studies have demonstrated that pollen characteristics serve as crucial evidence for resolving taxonomic problems in Apiaceae (Xiao et al. 2021, Song et al. 2024, 2025). However, comprehensive morphological investigations for Eriocycla are scarce, with only preliminary observations reported by Shu and Sheh (2001). Morphological and molecular similarities indicate the need for additional evidence to re-evaluate the taxonomic position of Eriocycla and related Seseli taxa.

Drawing on recent revisions in Apiaceae (Cai et al. 2022, Liu et al. 2024a, 2024b, Ren et al. 2025, Song et al. 2025), 3 concatenated datasets were obtained for phylogenetic reconstructions: 34 nuclear ribosomal DNA (nrDNA) (ITS + ETS), 35 plastid fragments (matK + rbcL + rps16 intron), and 42 plastomes [79 shared coding sequences (CDS)]. Comparative analyses of plastomes from 14 Eriocycla and Seseli species were conducted, supplemented by morphological and anatomical characteristics. With the integrated evidence, we aimed to (i) re-evaluate the taxonomic status of Eriocycla; (ii) reveal plastid characteristics of Eriocycla and related Seseli taxa; and (iii) provide new insights into the phylogeny of Eriocycla and Seseli, resolve the long-standing controversy between these genera.

Materials and methods

Taxon materials

Field surveys were conducted near the type localities of four species: E. nuda (Purang County, Xizang, China), E. pelliotii (Wushi County, Xinjiang, China), Seseli d. (Yuanmou County, Yunnan, China), and Seseli n. (Tingri County, Xizang, China), with over 10 individuals per species collected for morphological and molecular analyses.

Freshly collected leaves were dried in silica gel and stored. Voucher specimens referenced in this study are deposited at the Herbarium of Sichuan University (S.-D.Z.) (see Supplementary Table S1).

Morphological observation

Habitat photographs and external morphological features of E. nuda, S. nortonii, and S. delavayi were documented during field surveys. Microstructural observations of inflorescences (bracts, calyx teeth, petals) and mature mericarps (shape, ribs, vittae) were conducted using a stereomicroscope (SMZ25, Nikon Corp., Tokyo, Japan). For pollen morphology, 10 dry, mature, fully developed anthers per species were selected to examine gross morphology (equatorial view, polar view), germinal furrow structure, and exine ornamentation under a JSM-7500F scanning electron microscope. Pollen dimensions were measured using MATO v2.1 (Liu et al. 2023) with 10 replicates per species.

Morphological comparisons of S. nortonii and S. delavayi with other Seseli and Eriocycla taxa were based on specimen information, Flora of China (2005), and relevant literature (Linné and Salvius 1754, Pimenov and Kljuykov 2000, Doğan Güner et al. 2011, Doğan Güner and Duman 2013, Cai et al. 2022). Morphological terminology was followed according to Kljuykov et al. (2004), Ostroumova (2021), and Shu and Sheh (2001) .

DNA extraction, sequencing, assembly, and annotation

From our accumulated materials, 2 Eriocycla species (including the type species E. nuda and E. pelliotii) and 13 Seseli taxa (S. nortonii, S. delavayi, Seseli mairei, Seseli intramongolicum, Seseli aemulans, Seseli asperulum, Seseli coronatum, Seseli eriocephalum, Seseli glabratum, Seseli incisodentatum, Seseli squarrulosum, Seseli valentinae, and Seseli yunnanense) were selected, representing >80% coverage of both genera’s diversity in China. Total genomic DNA of these 15 species was extracted from silica gel-dried leaves using the modified CTAB method (Pahlich and Gerlitz 1980). Then, we used the primers ITS-4 (5′-TCCTCCGCTTATTGATATGC-3′), ITS-5 (5′-GGAAGTAAAAGTCGTAACAAGG-3′) (White et al. 1990), 18S-ETS (5′-ACTTACACATGCATGGCTTAATCT-3′) (Baldwin and Markos 1998), and Umb-ETS (5′-GCGCATGAGTGGTGAWTKGTA-3′) (Logacheva et al. 2010) to amplify the ITS and ETS regions. Polymerase chain reactions (PCRs) were conducted in 30 μl volume reactions with 2 μl plant total DNA, 1.5 μl forward primer, 1.5 μl reverse primer, 15 μl volume 2× Taq MasterMix (CWBIO, Beijing, China), and 10μl ddH₂O. The PCR amplification programme was set as follows: initial denaturation at 94°C for 4 min, followed by 36 cycles (denaturation at 94°C for 45 s, annealing at 52°C for 70 s, and extension at 72°C for 90 s), and a final extension at 72°C for 10 min. All PCR products were sent to Sangon (Shanghai, China) for sequencing. Then, the software DNASTAR SeqMan Pro v7.1.0 (Burland 2000) were used to assemble the ITS and ETS sequences.

For plastomes, total genomic DNA libraries of 15 species were generated via Illumina platform, and the 150 bp paired-end reads were created at Novogene (Beijing, China). We employed fastP v0.15.0 (-n 10 and -q 15) to filter the raw data (Chen et al. 2018), and high-quality reads were assembled for the whole plastomes by GetOrganelle v1.7.7.0 (Jin et al. 2020). Genome annotation was performed using Plastid Genome Annotator (Qu et al. 2019) with Seseli montanum (KM035851) as the reference. Then, we conducted manual refinements and extracted the matK gene, rbcL gene, and rps16 intron from annotated plastomes using Geneious Prime v9.0.2 (Kearse et al. 2012). Plastome circular maps for E. nuda, E. pelliotii, S. delavayi, and S. nortonii were generated via the online programme CHLOROPLOT (Zheng et al. 2020). Gene rearrangements across the four plastomes were analysed using Mauve Alignment v2.4.0 (Darling et al. 2004) implemented in Geneious Prime v9.0.2 (Kearse et al. 2012).

The 15 plastomes, ITS, and ETS sequences have been officially submitted in GenBank (see Supplementary Table S1).

Comparison of plastome structure

Comparative genomic analyses were conducted on 14 plastomes. The borders between inverted repeat (IR) and single-copy (SC) regions were visualized using CPJSdraw (Li et al. 2023). Sequence divergence was assessed using the mVISTA programme in Shuffle-LAGAN mode (Frazer et al. 2004), with E. nuda as the reference.

For codon usage analyses, 53 common SC CDS were extracted from these 14 plastomes after removing CDS <300 bp long (Wright 1990) and concatenated using PhyloSuite v1.2.2 (Zhang et al. 2020). Relative synonymous codon usage (RSCU) was calculated using the CodonW v1.4.2 programme (Peden 2000), and visualization was implemented in TBtools (Chen et al. 2020) through heatmap construction.

Phylogenetic analyses

To reconstruct the phylogeny of Eriocycla and Seseli species, three datasets were assembled and processed: (i) nrDNA: ITS + ETS from 34 Apiaceae taxa. Fifteen ETS sequences were newly generated; the remainder was sourced from NCBI. Sequences were aligned using MAFFT v7.308 (Katoh and Standley 2013) and concatenated with PhyloSuite v1.2.2 (Zhang et al. 2020). (ii) Plastid fragments: The matK gene, rbcL gene, and rps16 intron from 35 taxa. All sequences were obtained from NCBI. Alignment and concatenation followed the same MAFFT-PhyloSuite workflow. (iii) CDS: 79 shared SC CDS extracted from 42 complete plastomes. All plastomes were sourced from NCBI, including 15 generated by our team. CDS regions were aligned and concatenated using PhyloSuite v1.2.2 (Zhang et al. 2020; Supplementary Table S1).

Maximum likelihood (ML) analyses were conducted in RAxML v8.2.8 (Stamatakis 2014) based on the best-fit GTR + GAMMA model with 1000 bootstrap replicates. The Bayesian inference was performed using MrBayes v3.2.7 (Ronquist et al. 2012) after the program ModelTest v3.7 (Posada and Crandall 1998) calculated the best-fitting models of nucleotide substitutions under the Akaike information criterion (AIC). Four Markov chains were run for 10 million generations, sampling trees at 1000 generations intervals following a 25% burn-in. Consensus topologies from both analyses were visualized and annotated in FigTree v1.4.2 (Rambaut 2015).

Results

Morphological characteristics

Field investigations and inflorescence observations of E. nuda revealed significant morphological divergence from the previous description of S. tortuosum (Doğan Güner and Duman 2013). Key distinguishing traits included the absence of fibrous remnant sheaths at stem bases (Fig. 1c), contrasting with its presence in S. tortuosum. Additionally, E. nuda exhibited three to five linear bracts (Fig. 1g) compared with the absence of bracts in S. tortuosum, along with basally free bracteoles (Fig. 2) rather than connate. Floral morphology further differentiated the taxa, with E. nuda possessing pale yellow petals and prominent calyx teeth, whereas S. tortuosum displayed pale purple petals and obsolete calyx teeth. Notably, these traits displayed exceptional consistency among E. nuda, E. pelliotii, S. nortonii, and S. delavayi, while distinctly differentiating them from other examined Seseli species (see Supplementary Table S2).

Figure 1.

Eriocycla nuda. (a) Habit, (b) general morphology, (c) root, (d) basal leaves, (e) inflorescence, (f) infructescence, (g) bracts, and bracteoles.

Figure 2.

Morphology of bracteoles, petals and calyx teeth. (a) E. nuda, (b) S. nortonii, (c) S. delavayi.

Fruit morphology analyses revealed both shared and divergent characteristics. The mericarps of E. nuda, S. delavayi, and S. nortonii exhibited an oblong shape, while E. pelliotii displayed an ovoid form. Compression patterns further differentiated the taxa: E. nuda and E. pelliotii showed dorsal compression, whereas S. delavayi and S. nortonii were laterally compressed. All four species shared pubescent fruits, rounded ribs, and identical vittae configurations, with one vittae per furrow and two on commissure (see Supplementary Fig. S1).

The investigations of pollen morphology provided additional diagnostic evidence: while S. tortuosum exhibits short rod-like pseudo-cerebroid ornamentation in both equatorial and polar views (Degtjareva et al. 2011), all four studied species exhibit striato-cerebroid ornamentation in equatorial view and elongato-reticulate ornamentation in polar view (Fig. 3). However, polar perforations were observed in E. nuda and S. delavayi, contrasting with their absence in E. pelliotii and S. nortonii. In addition, the quantitative pollen metrics of these four species were divergent, with polar axis lengths of 24.70–33.63 μm, equatorial axis lengths of 12.07–17.46 μm, and P/E axis ratios of 1.64–2.05 (Table 1).

Figure 3.

The overview, equatorial view, germ furrow, polar view and exine ornamentation of pollen grains. (a) E. nuda, (b) E. pelliotii, (c) S. nortonii, (d) S. delavayi.

Table 1.

Pollen features of four species.

Taxa	E. nuda	E. pelliotii	S. nortonii	S. delavayi
Equatorial view	Superrectangle	Superrectangle	Superrectangle	Subrectangular
Polar view	Trilobate circular	Subtriangular	Trilobate circular	Trilobate circular
Exine ornamentation	Striate cerebriod	Striate cerebriod	Striate cerebriod	Striate cerebriod
Polar ornamentation	Elongate reticulate, perforate	Elongate reticulate	Elongate reticulate	Elongate reticulate, perforate
Aperture	Pleurotreme	Pleurotreme	Pleurotreme	Pleurotreme
P (μm)	30.61 (29.10–32.57)	24.70 (23.41–25.97)	33.63 (31.47–34.94)	26.62 (25.36–28.20)
E (μm)	17.46 (16.84–18.83)	12.07 (11.53–13.15)	16.68 (15.78–18.10)	16.21 (15.47–17.15)
P/E	1.76	2.05	2.02	1.64

Plastome features of Eriocycla nuda, Eriocycla pelliotii, Seseli nortonii, and Seseli delavayi

The plastomes of these four taxa ranged in length from 153 859 bp (S. delavayi) to 154 566 bp (E. pelliotii) (Fig. 4). All plastomes exhibited the typical quadripartite structure, comprising a pair of IRs, (25 104–25 386 bp) separating the large SC (LSC, 85 660–86 920 bp) and small SC (SSC, 17 427–17 465 bp) regions (see Supplementary Table S3). The total GC content was consistent at 37.6%, with IR regions demonstrating higher GC values (42.8–42.9%) compared with LSC (35.7–35.8%) and SSC (31.1–31.2%) regions. Each plastome contained 134 genes, including 86 CDS, 37 tRNA genes, and 8 rRNA genes. Mauve alignment revealed an inversion of the trnY–GUA, trnD–GUC, and trnE–UUC genes in E. nuda and E. pelliotii (see Supplementary Fig. S2).

Figure 4.

Plastome map of E. nuda, E. pelliotii, S. nortonii, and S. delavayi. Genes positioned outside the outer circle are transcribed in a clockwise direction, while those inside are transcribed counterclockwise. Functional groups of genes are distinguished by colour coding. Annotations in the upper left highlight a gene inversion event specific to the regions indicated for E. nuda and E. pelliotii.

Comparative analyses of plastomes

The LSC/IRb (JLB) and IRa/LSC (JLA) borders of E. nuda and E. pelliotii were located within the rpl12 gene and between the rpl23 and trnH–GUG genes, respectively, a pattern distinct from the other ten Seseli species but highly congruent with S. nortonii and S. delavayi (Fig. 5). Notably, S. delavayi exhibited the shortest rpl12 gene segment in the LSC region (785 bp) and the longest in the IRb region (685 bp) among the analysed plastomes. Two distinct types of IRb/SSC (JSB) borders were identified in these 14 plastomes: the first type occurred within the ndhF gene in E. pelliotii, S. delavayi, S. valentinae, and S. intramongolicum, while the second type was positioned between the ycf1 and ndhF genes in E. nuda, S. nortonii, S. eriocephalum, S. glabratum, S. coronatum, S. asperulum, S. mairei, S. aemulans, S. intramongolicum, and S. incisodentatum. The JLA borders of S. nortonii, S. delavayi, E. nuda, and E. pelliotii were located between the rpl23 and trnH–GUG genes, whereas this junction occurred between rpl22 and trnH–GUG genes in S. eriocephalum and between the trnL–CAA and trnH–GUG genes in other Seseli species.

Figure 5.

Comparison of the boundary positions of LSC, SSC, and IR regions for 14 plastomes. The coloured boxes represent functional genes and truncated fragments. The figure does not indicate the sequence length and only highlights the relative changes at or near the IR/SC boundaries.

The mVISTA results demonstrated higher sequence conservation in coding regions compared with non-coding regions, with inverted IR regions being more conserved than SC regions (see Supplementary Fig. S3). When using E. nuda as the reference, S. nortonii and S. delavayi exhibited substantial sequence divergence from other Seseli species, showing closer affinity to Eriocycla taxa. Polymorphisms were observed in highly divergent regions, including rps16–trnQ–UUG, trnE–UUC–trnT–GGU, trnT–UGU–trnL–UAA, rpl32–trnL–UAG, and petA–psbJ.

Analysis of 53 selected CDS across the 14 species identified 21 168–21 212 codons, with leucine (Leu) displaying the highest number of codons (2221–2244) and cysteine (Cys) showing the lowest (215–224) (see Supplementary Fig. S4). The RSCU values ranged from 0.33 to 2.02, where the codons AUG and UGG exhibited RSCU values of 1.00, indicating no codon usage bias. Among the 30 codons with RSCU > 1.00, all terminated with a purine (A/U) except UUG. The arginine codon CGA showed RSCU values of 1.35–1.36 in E. nuda, E. pelliotii, S. nortonii, and S. delavayi, compared with values ≥1.38 in the remaining 10 Seseli species.

Phylogenetic analyses

Phylogenetic reconstructions based on nrDNA and plastid fragments datasets robustly supported the monophyly of E. nuda, E. pelliotii, S. nortonii, and S. delavayi (nrDNA: BS = 100%, PP = 1.00, Fig. 6a; plastid fragments: BS = 100%, PP = 1.00, Fig. 6b), clustering these taxa within the tribe Echinophoreae. In contrast, the type species S. tortuosum and other Seseli members resolved within the tribe Selineae, forming a highly divergent paraphyletic group, demonstrating distant phylogenetic affinities (nrDNA: BS = 100%, PP = 1.00; plastid fragments: BS = 97%, PP = 1.00). Furthermore, the high-resolution CDS tree generated through concatenation methods exhibited identical topologies (BS = 100%, PP = 1.00, Fig. 7). Notably, across all three phylogenetic trees, S. nortonii formed a well-supported sister group with E. nuda (nrDNA: BS = 100%, PP = 1.00; plastid fragments: BS = 100%, PP = 1.00; CDS: BS = 100%, PP = 1.00) (Figs 6 and 7), the type species of Eriocycla.

Figure 6.

Phylogenetic trees inferred from Bayesian inference and Maximum Likelihood. Branch support was assessed using bootstrap percentage of ML and the followed posterior probability of BI. The solid circle represents maximum support in both two analyses (Bootstrap value = 100%, Posterior probability = 1.00); (−) represents the values <50%. (a) Phylogenetic tree of 34 Apioideae taxa based on nrDNA, (b) phylogenetic tree of 35 taxa based on plastome fragments.

Figure 7.

Phylogenetic tree inferred from 42 taxa based on 79 shared CDS, with carpological characteristics annotated at corresponding nodes. Scale bars: 0.5 mm (dorsal side views), 0.5 mm (transverse sections).

Discussion

In this study, we reconstructed the phylogenetic relationships of Eriocycla and Seseli based on 3 datasets (34 nrDNA sequences, 35 plastid fragments, and 42 plastomes) and obtained high-resolution trees for these 2 genera. Eriocycla nuda and E. pelliotii, two species previously classified within Seseli sect. Eriocycla, formed a well-supported monophyletic clade within the tribe Echinophoreae. This finding contradicts the treatment proposed by Pimenov and Kljuykov (2000). Based on fruit anatomical analyses of E. nuda and E. pelliotii, they concluded that traits such as fruit compression type, pubescence, and the number of vittae align these species most closely with Seseli, leading them to establish the Seseli sect. Eriocycla. However, the Apiaceae family is notoriously one of the most complicated families of flowering plants, with highly diverse fruit characteristics (Pimenov et al. 2019, Wojewódzka et al. 2019, Kljuykov et al. 2021), particularly within world-wide complex genera like Seseli, and we cannot exclude the influence of homoplasy and hybridization. Our results of the phylogeny also indicate we could not simply merge Eriocycla into Seseli. Furthermore, based on combined morphological and molecular evidence, Lyskov et al. (2020, 2022) transferred one species from Seseli sect. Eriocycla to Semenovia Regel & Herder and described two others as the new genus Shomalia Lyskov. The frequent taxonomic revisions, combined with the profound phylogenetic distance separating its core species (E. nuda, E. pelliotii) from Seseli, strongly indicate that the concept of Seseli sect. Eriocycla is no longer tenable. Our phylogenetic analyses also revealed that two Seseli taxa, S. nortonii and S. delavayi, resolve within the tribe Echinophoreae, while other Seseli species studied, including the type species S. tortuosum, fall into the tribe Selineae. The phylogeny of these taxa splits definitively into two major clades: one within Echinophoreae comprising E. nuda, E. pelliotii, S. nortonii, and S. delavayi; and another scattered within Selineae encompassing all remaining Seseli species examined. Comparative plastome analyses largely corroborated this division, the four Echinophoreae species share congruent plastome structures at the LSC/IRb (JLB: within the rpl12 gene) and IRa/LSC (JLA: between rpl23 gene and trnH–GUG genes) borders, a type distinct from other Seseli members (JLB: within ycf2 gene; JLA: between trnL–CAA and trnH–GUG genes). Regarding morphology, E. nuda, E. pelliotii, S. delavayi, and S. nortonii share key morphological characteristics absent in the other Seseli species studied, including glabrous stem bases, basally free bracteoles, and prominent calyx teeth (Supplementary Table S2). According to Doğan Güner et al. (2011) on the pollen features of Seseli, S. tortuosum exhibits short rod-like pseudo-cerebroid ornamentation in both equatorial and polar views, whereas in this study, the four Echinophoreae taxa displayed striato-cerebroid ornamentation in equatorial view and elongato-reticulate ornamentation in polar view.

Pimenov and Kljuykov (2000) treated S. nortonii as a synonym of E. nuda due to similarities in carpological features. These two species formed a solid sister group in our nrDNA-based and cpDNA-based trees, but a unique inversion of the trnY–GUA, trnD–GUC, and trnE–UUC genes was detected in E. nuda but absent in S. nortonii. Significant morphological divergences also exist between the two taxa: S. nortonii possesses densely hispid and thickly leathery basal leaves with laterally compressed fruits, contrasting with the pubescent and thickly membranous leaves and dorsally compressed fruits of E. nuda. These results demonstrate that adopting S. nortonii as a synonym of E. nuda is unsuitable.

Therefore, based on integrated evidence from phylogenetics, plastome structures, and morphological characteristics, we formally propose reinstating Eriocycla as an independent genus. S. nortonii and S. delavayi are hereby transferred to Eriocycla as Eriocycla nortonii comb. nov. and Eriocycla delavayi comb. nov., respectively. These findings provide new insights into the phylogeny of the tribe Echinophoreae. Regarding Seseli, one of the largest polyphyletic groups in the Apiaceae family, Cai et al. (2022) proposed defining this genus in a narrow sense (Seseli s.s.), comprising nine core members. Nevertheless, critical gaps in morphological and molecular data, particularly for the type species S. tortuosum, mean achieving a scientifically robust reclassification of Seseli in its entirety remains an arduous task requiring expanded sampling and multidisciplinary analyses.

Conclusion

This study reconstructs the phylogeny of Seseli and Eriocycla through three datasets: (i) 34 nrDNA(ITS + ETS); (ii) 35 plastid fragments (matK + rbcL + rps16 intron); (iii) 42 plastomes (79 shared CDS). These analyses robustly resolve E. nuda, E. pelliotii, S. nortonii, and S. delavayi as a well-supported monophyletic group within the tribe Echinophoreae, demonstrating distant phylogenetic relationships from S. tortuosum and other studied Seseli members of the tribe Selineae. Comparative plastome analyses congruently corroborated these distinctions: the four taxa exhibited structural conservation at JLB (within the rpl12 gene) and JLA (between the rpl23 and trnH–GUG genes) borders, alongside convergent codon usage bias for the CGA codon. The genus Eriocycla is characterized by glabrous stem bases, basally free bracteoles, and prominent calyx teeth, which clearly differentiate it from S. tortuosum and other Seseli species. Integrating morphological and molecular evidence confirms Seseli sect. Eriocycla does not constitute a natural group. Thus, we propose reinstating Eriocycla as an independent genus and transferring S. nortonii and S. delavayi to this genus, recognizing them as E. nortonii comb. nov. and E. delavayi comb. nov. In summary, our study resolved a long-standing controversy in Apiaceae family, providing new insights into the phylogenetic relationships of the tribe Echinophoreae and establishing a reference framework for future revisions of Seseli.

Appendix

Taxonomic treatment

Eriocycla Lindl.

Type species:  Eriocycla nuda Lindl. (Fig. 1)

Diagnosis

This genus can be easily distinguished from Seseli by glabrous stem bases (vs. fibrous remnant), numerous bracts (vs. absent or 1–2), basally free bracteoles (vs. connate), prominent calyx teeth (vs. obsolete), and rounded mericarp ribs (vs. keeled or filiform).

Eriocycla nortonii (Fedde ex H.Wolff) J.Cai & S.D.Zhou, comb. nov. (Fig. 8)

Figure 8.

Seseli nortonii. (a) Habit, (b) general morphology, (c) root, (d) basal leaves, (e) inflorescence, (f) infructescence, (g) bracts and bracteoles.

Basionym:  Seseli nortonii Fedde ex H.Wolff (1930, p. 329)

Type: China, Xizang: Dingri County, Kada Valley, crevice, 3600–4000 m a.s.l., 16 June 1922, Norton 108 (holotype: K!—barcode K000697458).

Diagnosis

Eriocycla nortonii closely resembles E. nuda but differs by the following characters: basal leaves densely covered with white-hispid hairs (vs. white pubescence), ultimate segments thickly coriaceous (vs. membranous to papery), involucral bracts 10 (vs. 3–5), fruits laterally compressed (vs. dorsally compressed), and pollen lacking polar perforations (vs. present).

Description

Plants 30–70 cm. Roots cylindrical, yellowish-brown. Stems solitary or clustered, purplish at base, hollow, striate, sparsely pubescent, branched from lower parts; stem base lacking fibrous remnants. Basal leaves numerous; petioles long, sheaths membranous-margined, hirsute; leaf blade broadly rhombic, 2–3-pinnate; ultimate segments ovate, margins serrate; rachis and abaxial surface densely white-hispid, adaxial surface sometimes glabrous; texture thickly coriaceous. Compound umbels with elongated peduncles; involucral bracts 10, linear-lanceolate, densely white-hispid, shorter than rays; rays 10–15, unequal; umbellules many-flowered; bracteoles 10, linear-lanceolate, densely hirsute, basally free. Petals pale yellow, suborbicular, apex emarginate with inflexed lobule, abaxially densely white-hispid. Calyx teeth subulate, prominent. Styles short, erect; stylopodium depressed-conical. Mericarps oblong, densely white-hispid, laterally compressed; ribs rounded; vittae one per furrow, two on commissure.

Distribution and habitat

Endemic to Dingri County, Xizang, China. Grows in arid rocky slopes or crevices at 3600–4000 m a.s.l.

Additional specimens examined

China, Xizang: Dingri County, Qudang Village, 28° 07′ N, 87° 31′ E, 3737 m a.s.l., 4 August 2022, J. Cai & J.Q. Lei CJ202208040101 (SZ); Qudang Village, 28° 25′ N, 87° 38′ E, 3772 m a.s.l., 4 August 2022, J. Cai & J.Q. Lei CJ202208040102 (S.-D.Z.); Zhaxizong Village, 28°34′N, 87°19′E, 3947 m a.s.l., 4 August 2022, J. Cai & J.Q. Lei CJ202208040201 (S.-D.Z.).

Eriocycla delavayi (Franch.) J. Cai & S. D. Zhou, comb. nov. (Fig. 9)

Figure 9.

Seseli delavayi. (a) Habit, (b) general morphology, (c) root, (d) basal leaves, (e) inflorescence, (f) infructescence, (g) bracts and bracteoles.

Basionym: Seseli delavayi Franch. (1894: 130).

Type: China, Yunnan: Heqing County, Lo-ho-chan Mountain, calcareous hills above Che-tong near Tapintze, 1500 m a.s.l., 4 September 1885, J.M. Delavay 2024 (lectotype: P!—barcode P04197843; isotype: K!—barcode K000697465)

Diagnosis

Eriocycla delavayi is distinguished from other Eriocycla species by its ternatisect basal leaves with lanceolate segments, subequal rays, ovoid fruits densely covered with white-hispid hairs, and elliptic to subrectangular pollen.

Description

Plants 50–90 cm, densely white-hispid throughout. Roots conical, yellowish-brown, woody. Stems solitary, branched above middle, cylindrical, densely white-hispid; stem base lacking fibrous remnants. Basal leaves ternatisect, segments sessile, linear-lanceolate, parallel-veined, both surfaces densely white-hispid. Compound umbels with elongated peduncles; involucral bracts 5–10, linear-lanceolate, acuminate, densely white-hispid, shorter than rays; rays 8–10, subequal; bracteoles 5–7, linear-lanceolate, densely hispid, basally free, longer than pedicels. Petals yellow, obovate, apex emarginate with inflexed lobule, abaxially densely white-hispid. Calyx teeth subulate, prominent. Styles very short, concealed within stylopodium; stylopodium conical. Mericarps oblong, densely white-hispid, laterally compressed; ribs rounded, obscured by indumentum; vittae one per furrow, two on commissure.

Distribution and habitat

Endemic to Yunnan, China (Heqing, Yuanmou). Grows on sunny grassy slopes at 1200–1800 m a.s.l.

Additional specimens examined

China, Yunnan: Heqing County, Tapintze, 1500 m a.s.l., 4 September 1885, J.M. Delavay 2024 (P); Tapintze, 1700 m a.s.l., 16 September 1929, R.C. Qin 24646 (PE); Tapintze, 1400 m a.s.l., 12 August 1963, 6463 (KUN); Yuanmou County, Wuguo Village, 1701 m a.s.l., 16 November 2021, J. Cai & H.H. Qin CJ0202111160201 (S.-D.Z.).

Key to species of Eriocycla

1a. Basal leaves ternatisect, ultimate leaf segments linear-lanceolate; mericarp laterally compressed, densely hispidE. delavayi

1b. Basal leaves (1-)2-pinnate, ultimate leaf segments ovate; mericarp laterally or dorsally compressed, densely pubescent2

2a. Mericarp oblong, laterally compressedE. nortonii

2b. Mericarp oblong or ovoid, dorsally compressed3

3a. Mericarp oblong; ultimate leaf segments deeply serrate on marginsE. nuda

3b. Mericarp ovoid; ultimate leaf segments sparsely serrate on marginsE. pelliotii

Supplementary data

Supplementary data is available at AoB Plants online.

Author contributions

X.-J.H. and S.-D.Z. designed the research. J.C. collected the samples; Y.-Z.L. analysed the data. Y.-Z.L. and Z.-X.L. prepared the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (grant no. 32470216 and 32170209).

Data availability

All newly generated DNA sequences have been submitted to NCBI (https://www.ncbi.nlm.nih.gov), and the GenBank accession numbers with other detailed information are provided in Supplementary Table S1.

Ethics approval

The collection of all samples fully adheres to national and local regulations. The plant samples used in the study are neither listed as protected species nor collected from national parks or nature reserves. Under relevant national and local laws, no specific permits were required for their collection.