We successfully procured eight distinct species of Lepidagathis (Fig. 1), with five demonstrating an exclusive endemic presence within India. To substantiate our research, voucher specimens were meticulously prepared for each collected plant sample and have been duly archived at the Department of Botany, Shivaji University, Kolhapur (Suppl. material 2: table S1). A geographical distribution of these sampled species within India can be observed in Fig. 2. The data of the remaining species were derived from our prior comprehensive research (Kadam et al. 2023b).

Figure 1. 

Sampled Indian Lepidagathis species a. Lepidagathis cristata, b. L. dalzelliana, c. L. fasciculata, d. L. incurva, e. L. mahakassapae, f. L. purpuricaulis, g. L. shrirangii, h. L. ushae.

Figure 2. 

Distribution map of all sampled Lepidagathis species.

The fresh leaf material of Lepidagathis cristata Willd, L. dalzelliana S.More, Mane, M.Sawant & H.S.Bhosale, L. fasciculata (Retz.) Nees, L. incurva Buch.-Ham. ex D.Don, L. mahakassapae S.More, M.Sawant, H.S.Bhosale & Kambale, L. purpuricaulis Nees, L. shrirangii Natekar, Kambale & Chandore, and L. ushae Borude, Gosavi & Chandore were used to extract the total genomic DNA using the DNeasy® Plant Mini Kit (QIAGEN, Germany). The quality and integrity of the extracted DNA were rigorously assessed through gel electrophoresis on a 2% agarose gel.

For PCR amplification we employed the set of genetic markers, ITS, trnL-F, and trnS-G, and PCR reactions setup and conditions were adapted from Kadam et al. (2023b). The purity and quality of the resulting PCR products were assessed via electrophoresis on a 2% agarose gel. Subsequently, the amplified genes were further purified and bidirectionally sequenced at Macrogen Corporation (Seoul, South Korea). For reference, we have documented the accession numbers for our generated Lepidagathis sequences alongside retrieved sequences from other Lepidagathis species and outgroup taxa, which are accessible in Table 1. To construct comprehensive DNA sequence data matrices, we assembled data from a total of 26 taxa, encompassing 23 Lepidagathis specimens, and three Barleria species as an outgroup. To account for any missing sequences, we supplemented the dataset with blank sequences.

After the sequencing run, the DNA sequences were analyzed, edited, and assembled using CodonCode Aligner version 9.0.2, developed by CodonCode Corporation. Subsequently, multiple sequence alignment was conducted using MEGA 10, as described by Kumar et al. (2016), utilizing the MUSCLE program developed by Edgar (2004). Further refinement of sequences in each aligned region was carried out using BMGE v 1.1 (Criscuolo and Gribaldo 2010). The data incongruence (ILD) test (Farris et al. 1994) was carried out using PAUP 4.0a 152 (Swofford 2002) to assess the phylogenetic congruence between the nuclear and chloroplast datasets. The test was performed with 1000 heuristic replicates, and the results showed no significant conflict between the two datasets (p-value = 0.001). Additionally, the topology of all individual phylogenies was largely consistent. As a result, all subsequent phylogenetic analyses were conducted on the combined dataset.

For a better understanding of the relationships among Lepidagathis taxa, we employed both Bayesian Inference (BI) and Maximum Likelihood (ML) methods to construct phylogenies based on nuclear, chloroplast, and combined (nuclear + chloroplast) datasets. The best-fit nucleotide substitution models for each sequence dataset were determined using the jModelTest 2 program (Darriba et al. 2012) based on the Akaike information criterion (AIC). The GTR+G model was found to be the best fit for combined datasets (ITS+trnS-G+trnL-F) and thus, it was employed for phylogeny construction. The ML analysis was executed with RaxML-HPC v.8.0 (Stamatakis 2014) via the XSEDE resource through the CIPRES science gateway (Miller et al. 2010) (https://www.phylo.org/). We employed the rapid bootstrap algorithm, conducting 1000 bootstrap replicates to obtain support values. However, BI phylogenetic analyses were carried out using MrBayes v.3.2.7a (Ronquist and Huelsenbeck 2003) on XSEDE. In the Bayesian analysis, we conducted Markov Chain Monte Carlo (MCMC) with four separate runs, each comprising 50,000,000 generations. These runs consisted of three heated chains and one cold chain, with tree sampling occurring every 1000 generations. The initial 10% of trees were discarded as burn-in, and the remaining trees were used to generate a 50% Majority-rule consensus and All-compatible group Bayesian tree with posterior probability values for each node.

Biogeographic regions were delineated by considering the distribution patterns of all Lepidagathis species. The distribution of Lepidagathis was categorised as follows: (A) America, (B) Africa and Arabia, (C) India and Sri Lanka, (D) Eurasia up to Wallace’s Line, and (E) the Pacific, (areas east of Wallace’s Line and Australia). To analyze the historical biogeography, the S-DIVA (Statistical Dispersal-Vicariance Analysis) was conducted using an All-compatible Bayesian tree in RASP v 4.2 (Yu et al. 2015). To ensure the reliability of the biogeographic analysis, 382 binary trees were employed for running the S-DIVA analysis.

This study incorporated sequence data from three distinct genomic regions, namely nrITS, trnL-F, and trnS-G, with respective sequence lengths of 546, 387, and 666 base pairs. These sequences were combined to form a composite matrix comprising both plastid and nuclear loci, which was used to construct molecular phylogenies of the Lepidagathis genus using ML and BI methods. This comprehensive dataset, encompassing nuclear and chloroplast sequences, encompassed 26 different species and comprised a total of 1599 characters, as detailed in Suppl. material 1. The resulting phylogenetic analysis based on this combined dataset effectively segregated Lepidagathis into two major clades: Clade A (PP = 1 and BS = 100) and Clade B (PP = 1 and BS = 100) (Fig. 3). Furthermore, the combined dataset provided robust support, as indicated by higher BI PP and ML BS, for delineating the evolutionary relationships between the Barleria and Lepidagathis lineages.

Figure 3. 

Bayesian phylogenetic tree based on the combined (ITS + trnS-G + trnL-F) dataset. Bayesian posterior probability values and Maximum Likelihood bootstrap values (BI PP / MLBS) are provided above branches. The species sampled from India are highlighted in red color and * represents bootstrap value less than 50.

The biogeographic analysis of Lepidagathis, using the S-DIVA method, was performed on a combined dataset specifically curated for this study. We categorized distribution areas into five regions, considering both the historical distribution of ancient supercontinents and the distribution patterns of Lepidagathis species. The resulting analysis revealed a maximum S-DIVA value of 1390.6670, which serves as robust evidence supporting our conclusions regarding ancestral range inference (Fig. 4). S-DIVA analysis unveiled a complex biogeographic history, marked by 20 dispersal events and 5 vicariance events that have significantly influenced the current distribution patterns observed in Lepidagathis. Notably, we identified two instances of global extinction events at Node 50 and Node 51, which occurred when a descendant lineage inherited a range different from that of its parent lineage. Furthermore, Node 48 indicated a dispersal event, with an 80% likelihood that Eurasia (D) may have served as the originating region for the common ancestor of Lepidagathis (Fig. 4). The distribution codes for Lepidagathis species used in the biogeographic analysis along with the output data from the RASP program for S-DIVA analysis, particularly focusing on the significant nodes are mentioned in Suppl. material 2.

Figure 4. 

Biogeographic analysis of Lepidagathis based on the Bayesian All-compatible groups tree. * (Black colour), ranges with probabilities < 5% are hidden and lumped together and reported.

The molecular phylogenetic analysis, based on the combined nrITS+cpDNA region, reveals that Lepidagathis forms a monophyletic group with robust BS PP and ML BS values. We concluded that for future phylogenetic investigations in this plant group, researchers can enhance their sampling size and reconstruct the phylogeny based on the combined (ITS + trnS-G + trnL-F) dataset.

As discussed in previously published work (Kadam et al. 2023b), the correct identification of Lepidagathis remains a challenging task due to historical misinterpretations, with several new genera mistakenly derived from it. This study, like the previous one, supports systematic revisions, including the incorporation of Lophostachys and Acanthura species into Lepidagathis. The lack of a universal classification system for Lepidagathis, likely due to its confusing morphology, persists, and resolving the taxonomic issues requires a proper classification system. The phylogenetic tree presented here confirms the existence of two major clades within Lepidagathis: Clade A and Clade B. All the Indian endemic species, including L. clavata, L. mahakassapae, L. dalzelliana, L. sabui, L. ushae, L. shrirangii, and L. cristata are clustered together in Clade A.

Recently, Bramhadande and Nandikar (2023) summarized the genus Lepidagathis in India, grouping species based on morphological similarities. This artificial classification aims to improve understanding and utilizes traits such as endemism, growth habits, leaves, flowers, and ovules to categorize 38 Indian taxa. Notably, the authors synonymized L. ushae with L. prostrata Dalzell. However, King et al. (2023) later recognized L. ushae as a distinct species following detailed morphological analyses of both fresh and herbarium specimens. Based on this, we have adopted the recognition of L. ushae as a separate species. According to Bramhadande and Nandikar’s (2023) classification, L. cristata, L. shrirangii, and L. lutea fall under Group A, characterized by decumbent herbs or subshrubs with glabrous to scabrid, linear to narrowly ovate leaves, sessile and acute; axillary spikes forming a congested, globose head near the roots, with mucronate-spinescent bracts. In our phylogeny, L. shrirangii, described by Natekar et al. (2019), and its closely related species L. lutea are positioned as sister taxa with robust support. Both species originate from the Konkan region of Maharashtra, India; however, L. cristata is not included in this clade (Fig. 3).

In Group C, Bramhadande and Nandikar (2023) placed L. sabui and L. mahakassapae based on their prostrate shrub habit, axillary or terminal erect spikes, and two ovules; all these species are endemic to peninsular India. Since L. ushae also shares these traits, we have included it in Group C as well. Our phylogeny supports this grouping as these species are placed in the same clade. However, L. dalzelliana, which was placed in Group H by Bramhadande and Nandikar (2023), appears in our phylogeny alongside species from Group C. Notably, L. mahakassapae shares a close relationship with L. sabui but can be distinguished by its oblanceolate acuminate leaves, pubescent terminal elongated spikes, large floral structures, small sterile bracts, and lanceolate broader segments of the bracteole (More et al. 2022). Similarly, L. dalzelliana bears a resemblance to L. clavata in appearance but differs in having long lanceolate, hairy bracts, and spatulate, oblanceolate to ovate-lanceolate glabrous leaves (More et al. 2023). In addition, L. mahakassapae and L. dalzelliana are also closely related species, sharing a sister relationship in their evolutionary lineage (Fig. 3). Both of these species inhabit high-altitude plateaus within the Satara district (Maharashtra) and coexist in the same geographical area. They are characterized as perennial, decumbent, prostrate, sub-shrubs, hairy bracts and seeds. The primary distinguishing features between them include flower color, hairiness of bracts, nerve number of bract, nature of bracteole, color of seed hairs and flowering and fruiting time (More et al. 2022; More et al. 2023). Similarly, L. sabui and L. ushae are found in lower-altitude lateritic plateaus and share a close phylogenetic relationship according to our research. These two species are also characterized as perennial, procumbent, prostrate, rigid herb, cylindrical stem, pinkish flower color (Borude et al. 2020; Chandore et al. 2020). In addition, L. clavata which was placed in Group D by Bramhadande and Nandikar (2023), appears in our phylogeny alongside species from Group H (L. rigida and L. cuspidata). Furthermore, L. cuspidata, L. rigida, L. clavata, L. dalzelliana, L. mahakassapae, L. sabui, and L. ushae co-occur in the same region, with many of them also described from the Konkan region of Maharashtra. Consequently, our phylogenetic analysis groups these species together, indicating their close relationship.

According to Bramhadande and Nandikar (2023), L. fasciculata (Retz.) Nees is the lone species from Group E, and it is placed near L. cristata in our phylogeny, though with weak support. L. incurva Buch.-Ham. ex D.Don and L. purpuricaulis Nees are classified as Group F species based on shared characteristics, such as their erect to decumbent herb or shrub habit, linear to elliptic-ovate leaves, and axillary or terminal inflorescences. However, our phylogeny separates these species into distinct clades. Additionally, both sampled and adapted species of L. incurva cluster together with strong ML support.

In summary, our phylogeny aligns with certain aspects of Bramhadande and Nandikar’s (2023) classification but diverges in others. A universal classification system for Lepidagathis is crucial, and a more comprehensive, robust molecular phylogeny is needed to resolve the remaining taxonomic issues.

The study of evolutionary history through molecular phylogeny is essential for gaining a precise understanding of biogeographical evolution (Ali et al. 2012). Surprisingly, no biogeographical investigation has been conducted on Lepidagathis thus far. Our S-DIVA analysis suggests that Eurasia (D) may serve as the probable center of origin for Lepidagathis. Within Lepidagathis, we observe a further division into Clade A, which is distributed across Eurasia and subsequently diverged in India and America (Node 42). In contrast, Clade B expanded initially in Eurasia (Node 47) and later dispersed into America (Nodes 43 and 44).

The shared ancestry of the former Lophostachys species (Lepidagathis chiapensis, L. sessilifolia, and L. uxpanapensis), the previous Acanthura mattogrossensis (now L. riedeliana), and L. alopecuroidea are traced back to a common origin during the migration from the Old World to the New World. (McDade et al. 2008). Our phylogenetic analysis positions these species collectively, corroborating the notion of their shared ancestry, observed at nodes 45 and 42, corresponding to the mentioned dispersal event.

It’s worth noting that species within Lepidagathis have received relatively little attention in terms of molecular phylogenetic research, and many remain unsequenced. However, it is important to emphasize that this study represents a preliminary exploration, shedding light on potential avenues for future molecular investigations into Lepidagathis.

The authors have declared that no competing interests exist.

No ethical statement was reported.

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (NRF-2016R1A6A1A05011910). APJ is thankful to Principal, The New College, Kolhapur and MAHAJYOTI for financial assistance [MAHAJYOTI/Nag./Fellowship/2021-22/1042 (275)]. RNM is thankful to the Head, Department of Botany and Principal, Balwant College, Vita for providing necessary facilities.

The study was designed by Suhas K. Kadam. Material was collected in the field by Rohit N. Mane and Akshay P. Jangam; data analysis was done by Suhas K. Kadam and Asif S. Tamboli; creation of the first draft of the manuscript was performed by Suhas K. Kadam, Yeon-Sik Choo and Jae-Hong Pak. All authors read and approved the final manuscript.

Suhas K. Kadam https://orcid.org/0000-0002-2396-4932

Rohit N. Mane https://orcid.org/0000-0002-3198-6816

Asif S. Tamboli https://orcid.org/0000-0003-2146-670X

Akshay P. Jangam https://orcid.org/0000-0002-4916-4220

Yeon-Sik Choo https://orcid.org/0000-0001-7307-325X

Jae Hong Pak https://orcid.org/0000-0001-9085-3741

The data that support the findings of this study are openly available in Science Data Bank, at https://doi.org/10.57760/sciencedb.15983.