Taxonomy of Acantholimon species in southern Iran: A molecular phylogenetic approach

Introduction

Acantholimon Boiss. (Plumbaginaceae) is a highly diverse genus comprising approximately 321 accepted species distributed from southeastern Europe to Central Asia (POWO, 2025). It predominantly thrives in the Irano-Turanian floristic regions, particularly in Iran, Turkey, and Afghanistan (Kubitzki, 1993; Hassler, 2023; POWO, 2025). Historically, the genus Acantholimon in Iran has been documented to comprise 84 species according to Flora Iranica (Rechinger & Schiman-Czeika, 1974) and 79 species in the Flora of Iran (Assadi, 2005a). However, recent taxonomic work has identified several new species (Assadi, 2004; Assadi, 2005b; Assadi & Mirtadzadini, 2006; Assadi & Zeraatkar, 2020; Mahmoodi & Assadi, 2021; Bordbar & Mirtadzadini, 2022; Bibak et al., 2024). Following these latest discoveries, the recognized total number of Acantholimon species in Iran has consequently reached 83. The genus is predominantly found in the alpine and subalpine steppe ecosystems of the Irano-Turanian region, a unique biogeographical area defined by its arid to semi-arid climates. It has developed remarkable adaptations to survive in these harsh environments, such as xerophytic features like cushion-shaped formations and spiny leaves, which allow it to withstand drought conditions and thrive in nutrient-deficient soils (Moharrek et al., 2019). The genus significantly contributes to soil erosion prevention due to its robust root systems and enhances regional biodiversity. Acantholimon, exhibiting an overall species endemism rate of 82% (the highest level recorded among all Iranian genera), ranks as the fourth most endemic-rich genus in Iran, with 54 endemic species (Assadi, 2006; Ghahremaninejad et al., 2025). Kerman, the central Alborz, Isfahan, and ChaharMahal-Bakhtiari provinces are among the regions in Iran with the highest levels of Acantholimon endemism (Khajoei Nasab & Khosravi, 2020). Approximately 22 species of Acantholimon have been recorded in Southern Iran, the majority of which exhibit regional endemism. Kerman Province contains the highest concentration of these endemic species within Southern Iran, accounting for 16 records. More recently, three novel species, A. saadii Assadi & Zeraatkar (2020), A. assadii Mirtadz. & Bordbar (2022), and A. brevispicatum Bibak, Kazempour, Assadi (2024), have been formally described from this region. Few molecular investigations have been published for the genus Acantholimon. Phylogenetic analyses utilizing both nuclear and plastid DNA markers consistently support close evolutionary relationships among Acantholimon, Cephalorhizum Popov & Korovin, Limonium Mill., Armeria Willd., and Psylliostachys (Jaub. & Spach) Nevski. Molecular data strongly suggest that Acantholimon is not a monophyletic genus, indicating that its current taxonomic classification is artificial (Moharrek et al., 2014; Moharrek et al., 2017). Furthermore, combined biogeographic and molecular dating analyses estimate the origin of Acantholimon s. l. to be in eastern Iran-Afghanistan, with subsequent diversification events spanning the Late Miocene through the Pliocene epochs (Moharrek et al., 2019). A taxonomic study of Acantholimon species in southern Iran was conducted, utilizing sequences from both nuclear and chloroplast genomes to assess their phylogenetic relationships. A diagnostic key and a distribution map were ultimately developed for these species.

Materials and Methods

Taxonomy study

Approximately 45 specimens of Acantholimon were collected during the spring and summer seasons between 2020 and 2023 from the southern region of Iran. Additionally, herbarium sheets housed at the Research Institute of Forests and Rangelands (TARI) were examined (Table 1). Specimens are also deposited in the TARI and Tarbiat Modares University Herbarium (TMUH). Morphological characters were studied both in the field and in the herbarium. The collected samples were identified using the taxonomic revision of the genus Acantholimon (Mobayen, 1964), Flora Iranica (Rechinger & Schiman-Czeika, 1974), and Flora of Iran (Assadi, 2005a). Data about both field observations and herbarium studies are summarized in Table 1. Distribution mapping was performed using ArcGIS version 10.6.1 (Figure 1).

 Figure 1. Distribution map of Acantholimon in southern Iran.

Table 1. Voucher information and GenBank accession number of examined species in molecular analysis.

(*) nrDNA ITS sequences for those taxa were obtained from GenBank

Molecular study

The research provided new nrDNA ITS sequences for 20 species (28 accessions) and rpl32-trnL _(UAG) sequences for 17 species (31 accessions). Additionally, fifteen nrDNA ITS sequences were obtained from GenBank. Nuclear and plastid markers were combined to create a dataset for 45 taxa. Details on species, accessions, voucher information, and GenBank accession numbers are summarized in Table 1. Dried leaf samples were used to extract total genomic DNA using a modified CTAB protocol (Doyle & Doyle, 1987). The nrDNA ITS region was amplified with primers AB101F and AB102R (Douzery et al., 1999). The rpl32-trnL _(UAG) spacer was amplified with rpl32-F and trnL _(UAG)-R primers following Shaw et al. (2007). PCR amplification was performed in a microtube containing 8 μl of deionized water, 10 μl of 2× Taq DNA polymerase master mix (Red, Amplicon), 0.5 μl of each primer (at 10 pmol/μl), and 1 μl of template DNA. The nrDNA ITS region was initially predenatured at 94°C for 4 minutes. Then, 35 cycles were run: 1 minute at 94°C for denaturation, 1 minute at 55°C for primer annealing, and 90 seconds at 72°C for extension. Final extension was at 72°C for 5 minutes. The PCR protocol for the rpl32-trnL _(UAG) region included an initial predenaturation of 5 minutes at 80°C, followed by 35 cycles of 1 minute at 94°C for denaturation, 1 minute and 60 seconds at 51.5°C for primer annealing, and 60 seconds at 65°C for extension. It concluded with a 7-minute extension at 72°C. PCR products were separated on 1% agarose gels in 1x TBE buffer (pH 8) and stained with ethidium bromide. The amplified products were sequenced with appropriate primers and sent to Pishgam Inc. for Sanger sequencing. Psylliostachys beludshistanicus Roshkova was considered an outgroup, based on Moharrek et al. (2014, 2017).

Phylogenetic analyses

A phylogenetic study of 20 Acantholimon taxa was conducted using maximum likelihood (ML) and Bayesian inference (BI). The maximum likelihood analysis was performed with the W-IQ-TREE web server (Trifinopoulos et al., 2016). MrBayes version 3.2.6 (Ronquist et al., 2012) was used for the Bayesian phylogenetic analysis, which was conducted on the CIPRES Science Gateway version 3.3 (Miller et al., 2010). The best nucleotide substitution model for each locus was identified using jModelTest (Posada, 2008) based on the Akaike Information Criterion (AIC) via the Phylemon 2.0 web server (Sánchez et al., 2011), while GTR models selected by ModelFinder were applied for nrDNA data, and GTR+G models were used for plastid and combined datasets in the ML analyses.

Results

The current study’s findings identify the following characters as valuable for taxonomy: inflorescence type and length; the number and length of flowers; bract shape and length; the number, size, and shape of bracteoles; calyx shape and color; calyx vein patterns; and the indumentum (or indument) of the inner calyx. Three datasets were analyzed: the nrDNA ITS sequence alignment, which included 45 accessions, spanned 622 nucleotide positions, yielding 53 parsimony-informative sites. The rpl32-trnL (UAG) matrix comprised 34 individuals, 1158 base pairs, and 67 parsimony-informative sites. Furthermore, the comprehensive combined dataset (nrDNA ITS + rpl32-trnL (UAG)) consisted of 1795 nucleotide sites, including 143 parsimony-informative sites. Detailed statistics for each dataset and their combination are presented in Table 2. For visualization, a Bayesian 50% majority-rule consensus tree was constructed, displaying both PP and BP values on the branches.

Table 2. Dataset and tree statistics from separate and combined analyses of the nuclear and chloroplast regions.

Results from the nrDNA ITS tree (Figure 2):

The tree inferred from the nrDNA ITS data resolved two major clades: Clade A and Clade B. Clade A is basal and comprises two subclades: A1 and A2. Subclade A1 included sixteen accessions: two of Acantholimon spinicalyx (Sect. Tragachantina), twelve of A. scorpius (Sect. Tragachantina), and two of A. austro-iranicum (Sect. Microstegia). Subclade A2 consisted of thirteen accessions, including A. modestum (2 accessions; Sect. Tragachantina), A. sirchense (1; Sect. Glumeria), A. chlorostegium (1; Sect. Tragachantina), A. albocalycinum (2; Sect. Glumeria), A. haesarense (1; Sect. Acantholimon), A. kermanense (2; Sect. Acantholimon), A. cupreo-olivascens (2; Sect. Glumeria), A. shirazianum (1; Sect. Microstegia), and A. mirtadzadinii (1; Sect. Acantholimon). Clade B was resolved into two subclades: B1 and B2. Subclade B1 contained two accessions of A. flexosum (Sect. Acantholimon). Subclade B2 comprised thirteen accessions from the Acantholimon section: A. festucaceum (2), A. zaeifii (2), A. flexosum (1), A. asphodelinum (2), A. serotinum (2), A. brevispicatum (2), A. hormozganense (1), and A. eschkerense (1).

Figure 2. The 50% majority-rule consensus trees were inferred from maximum likelihood analysis using nrDNA ITS sequences. The numbers above the branches represent the maximum likelihood analysis's bootstrap percentage (BP) and posterior probability (PP) of Bayesian inference, respectively.

Results from the rpl32-trnL (UAG) Dataset (Figure 3):

The rpl32-trnL (UAG) matrix also revealed two distinct clades, A and B. The first diverging subclade, A1, included fourteen individuals from the Tragachantina section: two accessions of A. spinicalyx and twelve of A. scorpius. Subclade A2 contained eleven accessions: A. sirchense (1; Sect. Glumeria), A. haesarense (1; Sect. Acantholimon), A. modestum (2; Sect. Tragachantina), A. kermanense (1; Sect. Acantholimon), A. chlorostegium (1; Sect. Tragachantina), A. albocalycinum (2; Sect. Glumeria), A. austro-iranicum (2; Sect. Microstegia), A. cupreo-olivascens (1; Sect. Glumeria), and A. mirtadzadinii (1; Sect. Acantholimon). Clade B contained subclades B1 and B2. B1 included one accession of A. festucaceum (Sect. Acantholimon), while B2 comprised seven Acantholimon section accessions: A. flexosum (1), A. serotinum (1), A. asphodelinum (1), A. zaeifii (2), and A. brevispicatum (2). Phylogenetic analysis employing the combined ITS and rpl32-trnL (UAG) datasets resulted in a tree topology that clearly segregated the sampled accessions into two primary clades, designated Clade A and Clade B (Fig. 4). Clade A further resolved into two well-supported subclades, designated A1 and A2. Subclade A1, representing the basal lineage within Clade A, was composed of sixteen accessions. This group was dominated by taxa from the Tragachantina section (A. spinicalyx and A. scorpius), alongside two accessions identified as A. austro-iranicum (Sect. Microstegia). Subclade A2 included thirteen accessions distributed across four distinct sections. These accessions comprised: one A. haesarense and two A. kermanense (Sect. Acantholimon); two A. albocalycinum, one A. sirchense, and two A. cupreo-olivascens (Sect. Glumeria); one A. chlorostegium and two A. modestum (Sect. Tragachantina); and one A. shirazianum (Sect. Microstegia). Furthermore, one accession of A. mirtadzadinii (Sect. Acantholimon) was nested within this subclade. Clade B contained subclades B1 and B2. B1 included two accessions of A. festucaceum (Sect. Acantholimon), while B2 comprised thirteen Acantholimon section accessions: A. flexosum (3), A. serotinum (2), A. asphodelinum (2), A. zaeifii (2), A. eschkerense (1), and A. brevispicatum (2).

Figure 3. The 50% majority-rule consensus trees were inferred from maximum likelihood analysis using rpl32-trnL _(UAG) sequences. The numbers above the branches represent the maximum likelihood analysis's bootstrap percentage (BP) and posterior probability (PP) of Bayesian inference, respectively.

Figure 4. The 50% majority-rule consensus trees were inferred from the combined dataset (nrDNA ITS and rpl32-trnL _(UAG)). The numbers above the branches represent the maximum likelihood analysis's bootstrap percentage (BP) and posterior probability (PP) of Bayesian inference, respectively.

Discussion

The genus Acantholimon, with 16 endemic species, exhibits a remarkably high degree of endemism in southern Iran, particularly within Kerman Province. This region is characterized by extensive high-altitude areas and is topographically and floristically connected to the southern Zagros Mountains. The abundance of endemic taxa suggests a complex evolutionary history shaped by geographic isolation, microhabitat specialization, and adaptive radiation (Hedge & Wendelbo, 1978). The topology of phylogenetic trees inferred from nrDNA ITS and rpl32‑trnL _(UAG) sequences reveals incongruent placement of certain taxa. These incongruences may result from natural processes such as hybridization, introgression, chloroplast capture, and incomplete lineage sorting (Wendel & Doyle, 1998). This study, which employs the chloroplast marker rpl32‑trnL _(UAG) in comparison with the nuclear nrDNA ITS region, may provide a clearer depiction of species relationships. This could be attributed to its higher sequence variability, as indicated by 67 parsimony‑informative sites versus 53 in the ITS loci. Moreover, the chloroplast genome is maternally inherited and exhibits a relatively conserved structure. The uniparental mode of chloroplast inheritance tends to minimize the complications associated with hybridization events, which are more frequent in biparentally inherited nuclear genomes, thereby potentially yielding clearer lineage patterns. Nevertheless, integrating data from both nrDNA and cpDNA markers generally provides a more robust and comprehensive phylogenetic resolution. The results of this study, together with those of Moharrek et al. (2017), indicate that the sections Acantholimon (= Staticopsis), Glumeria, Microstegia, and Tragacanthina do not form monophyletic groups (Figures 2–4). The section Acantholimon (= Staticopsis) represents the most species-rich group within this genus in Southern Iran, comprising 11 species. Molecular analyses suggest that the species belonging to this section are resolved into two distinct clades within the phylogenetic trees (Figures 2–4). Morphological evidence divides the Southern Iranian Acantholimon species into two corresponding groups. Group 1, characterized by a densely arranged terminal spike and multi-flowered spikelets possessing more than two bracteoles, encompasses A. mirtadzadinii, A. assadii, and A. kermanense. Group 2 exhibits either dense or lax spikes and features single-flowered, two-bracteolate spikelets, comprising A. hormozganense, A. serotinum, A. eschkerense, A. asphodelinum, A. flexosum, A. festucaceum, A. zaeifii, A. haesarense, and A. brevispicatum. Molecular analyses reveal that Group 1 species are nested within subclade A2, alongside species from other sections. Conversely, Group 2 resides entirely within clade B, with the notable exception of A. haesarense (which is also placed in subclade A2). This partitioning demonstrates a high degree of concordance between the morphological and molecular classifications within this section; however, it must be noted that no molecular data are available for A. assadii (Figures 2, 4). Three species of the Glumeria section (A. sirchense, A. albocalycinum, and A. cupreo-olivascens) are resolved into three distinct positions within subclade A2 (Figures 2–4). This section is morphologically characterized by lax spikes, single or multi-flowered spikelets possessing more than two bracteoles, and spikelets that are consistently shorter than the nodes. All species within subclade A2 are endemic to Kerman Province, with the exception of A. shirazianum. Furthermore, these species largely inhabit restricted mountainous areas and exhibit overlapping geographical distributions, suggesting the formation of a distinct phytogeographical cluster. Despite their molecular proximity, clear morphological distinctions are observed among them. The Microstegia section comprises two species: A. austroiranicum and A. shirazianum, which are allocated to subclades A1 and A2, respectively. The key morphological characteristics defining this section include heteromorphic leaves, dense spikelets containing one to four (1–4) flowers, and a conspicuously pilose, infundibular calyx. A. austroiranicum exhibits a close phylogenetic affinity with A. scorpius, as evidenced by their similar geographical distribution across southern Kerman, potentially indicating the formation of a distinct phytogeographical cluster. The species A. scorpius, A. spinicalyx, A. modestum, and A. chlorostegium belong to the section Tragacanthina, which is resolved into three distinct positions across subclades A1 and A2 in the phylogenetic analysis (Figures 2, 3, 4). Common morphological traits shared by these species include heteromorphic leaves, deciduous vernal leaves, lax or dense spikes, and spikelets that are single-flowered and two-bracteate. A. scorpius (Jaubert & Spach) Boissier is an endemic species widely distributed throughout Iran, extending into the transitional zone between the Irano-Turanian (IT) and Saharo-Sindian regions (Assadi, 2006). According to the nrDNA ITS tree (Figure 2), individuals of A. scorpius populations cluster into a single subclade, adjacent to two individuals of A. spinicalyx, despite observed morphological variation in characters such as calyx size, calyx limb structure, calyx vein patterns, and bracteole shape. The co-occurrence of one A. spinicalyx individual within the A. scorpius population (subclade A1) suggests a close phylogenetic relationship or potential hybridization, necessitating further molecular investigation. While A. spinicalyx is characterized by a colored calyx, in contrast to the white calyx of A. scorpius, the two species are nevertheless separated based on chloroplast and combined phylogenetic analyses (Figs. 2, 3s). Except for clade B, establishing consistent morphological characters that clearly correspond to the molecular topology of Acantholimon species distributed in southern Iran remains difficult. The morphological diversity observed among these species does not consistently reflect their phylogenetic relationships as inferred from molecular data. This incongruence indicates that classical taxonomy, which relies primarily on morphological traits such as leaf indumentum, calyx structure, and spike arrangement, has limited compatibility with molecular phylogenetic evidence. The lack of clear diagnostic morphological features across most clades suggests that convergent evolution or ecological adaptation may obscure phylogenetic signals within the genus. Therefore, integrating molecular markers with detailed morphometric and ecological analyses will be essential to achieve a more reliable systematic framework for Acantholimon in southern Iran. Although species of the genus Acantholimon are less susceptible to overgrazing because their leaves are rigid and woody, the primary conservation threats stem from habitat destruction due to the expansion of mining activities and the construction of access roads in mountainous areas. These disturbances are critically important for narrowly distributed endemic species in southern Iran, such as A. shirazianum, A. brevispicatum, A. hormozganense, A. haesarense, A. saadii, A. assadii, A. mirtadzadinii, A. sirchense, and A. cupreo-olivascens. The combination of small population sizes and restricted ranges renders these specific taxa particularly vulnerable to ongoing habitat fragmentation and anthropogenic disturbance.

Diagnostic Key to the Acantholimon species in Southern Iran

1. Inside the calyx pilose..................................................................................................................................................................... 2
2. Inside the calyx glabrous ............................................................................................................................................................... 3
3. Spikes terminal, exceeding the leaves; spikelets 2–5; bracteoles oblong............................................................ A. austroiranicum
4. Spikes terminal, not exceeding the leaves; spikelets 5–9; bracteoles oblanceolate ................................................ A. shirazianum
5. Spikelets 1-flowered, 2-bracteate .................................................................................................................................................. 4
6. Spikelets 1- or many-flowered, more than 2-bracteate ................................................................................................................ 16
7. Lower leaves persistent, monomorphic, or broader than upper leaves………………………………………………………….. 5

4- Lower leaves deciduous, heteromorphic (broader and shorter than upper leaves) ……………………………………..……... 12

5. Calyx white, veins colored ............................................................................................................................................................. 6
6. Calyx colored, veins colored ....................................................................................................................................... A. serotinum
7. Spikes unbranched ......................................................................................................................................................................... 7
8. Spikes branched ............................................................................................................................................................................. 8
9. Stems short (15–20 mm), not exceeding the leaves; spike short (5–10 mm); bracteoles ovate; petals oblong.............................. A. brevispicatum
10. Stems longer (30–45 mm), exceeding the leaves; spike longer (15–25 mm); bracteoles elliptic; petals spathulate............................ A. zaeifii
11. Spikes very dense, spikelets imbricate; bracts and bracteoles purple ............................................................................................ 9
12. Spikes lax, spikelets not imbricate; bracts and bracteoles brown ...................................................................... A. hormozganense
13. Spikes racemose ........................................................................................................................................................................... 10
14. Spikes terminal ......................................................................................................................................................... A. haesarensis
15. Spikelets predominantly terminal, occasionally on lateral branches................................................................... A. asphodelinum
16. Spikelets predominantly on lateral branches.............................................................................................................................. 11
17. Spikes with long lateral branches; spikelets lax, bracteoles with long apiculate …………….............................. A. festucaceum
18. Spikes with short lateral branches; spikelets dense, bracteoles with short apiculate or obtuse ................................ A. flexuosum
19. Spikes dense, axis not zigzag; calyx infundibular ..................................................................................................................... 13
20. Spikes lax, axis zigzag; calyx tubular .............................................................................................................................. A. saadii
21. Calyx white ................................................................................................................................................................................ 14
22. Calyx colored ............................................................................................................................................................ A. spinicalyx
23. Spikelets shorter than 9 mm; calyx shorter than 9 mm............................................................................................................... 15
24. Spikelets longer than 9 mm; calyx longer than 8 mm ................................................................................................. A. scorpius
25. Spikes lax; calyx milky white; calyx veins not reaching tips of lobes …………………………………........... A. chlorostegium
26. Spikes dense; calyx white to brown; calyx veins reaching tips of lobes .................................................................. A. modestum
27. Spikes terminal; spikelets longer than internodes ...................................................................................................................... 17
28. Spikes not terminal; spikelets shorter than internodes ............................................................................................................... 19
29. Spikes dense; calyx light to dark brown ........................................................................................................................ A. assadii
30. Spikes terminal; calyx purple .................................................................................................................................................... 18
31. Spikes exceeding or nearly the leaves; leaves 2 mm in width; bracteoles apiculate ........................................... A. mirtadzadinii
32. Spikes exceeding or very much longer than leaves; leaves 1 mm in width; bracteoles obtuse ............................. A. kermanense
33. Calyx white ........................................................................................................................................................ A. albocalycinum
34. Calyx purple or brown ............................................................................................................................................................... 20
35. Bracts lanceolate, reaching to the middle or two-thirds of bracteoles ….……………………………………….…. A. sirchense
36. Bracts ovate, reaching up to one-third of bracteoles .................................................................................... A. cupreo-olivascens

Acknowledgements

The authors thank the Herbarium for their support of the Research Institute of Forests and Rangelands (TARI) for their cooperation. This work was funded by the Tarbiat Modares University Research Council.