Cousinia Cass. (Asteraceae, Cardueae) is a large genus, being the 50th largest genera among flowering plants. It is represented by approximately 671 species (Frodin, 2004; Ulukuş and Tugay, 2020; Uysal et al., 2022; POWO, 2025). The species of Cousinia are typically found in arid regions, with a preference for upland vegetation. They are mainly distributed in the mountainous areas of Turkey, Iran, Afghanistan and Central Asia (Knapp, 1987). Central Asia, home to ca. 260 species, is considered the most significant region for these plants (Tscherneva, 1993).

Cousinia was first described by Cassini (1827). de Candolle, (1824) listed 34 species of Cousinia, organized into three categories according to the morphology of their capitula. In an early effort to comprehensively classify the genus Cousinia, von Bunge (1865) categorized 126 species of Cousinia into 23 sections, primarily based on growth habit and capitular features, such as the texture of receptacular bristles, shape of receptacular bracts, corolla color, color of the anther tube, and whether or not hairs were present on the anther tube. Bunge’s classification served as a foundation for later studies, including Boissier’s Flora Orientalis (1875) and its supplement (1888), which documented 141 species of Cousinia organized into 14 sections. The next significant classification was by Winkler in 1892 (Winker, 1892), who emphasized capitulum morphology for taxonomy like Bunge’s classification. In a notable divergence, Kuntze (1891) proposed merging Cousinia with the genus Arctium L. Winkler’s work (1892, 1897) expanded to include 267 species divided into three subgenera (Oligochaete, Dichacantha, and Eucousinia). Bunge’s approach laid the groundwork for a more natural classification system as later elaborated by Rechinger (1953, 1972) and Tscherneva (1962, 1988). Rechinger’s Flora Iranica (1972) included over 350 species across 58 sections, addressing the flora of the Iranian Plateau, Turkmenistan, Afghanistan, and the mountainous regions of Pakistan. Similarly, Tscherneva’s treatment (1962, 1988) for the Flora of the USSR cataloged around 260 species in 50 sections, covering regions of Central Asia and the Caucasus. According to Sennikov (2010), the sections Alpinae Bunge (Carduncellus (Juz.) Rech. f.) and Subappendiculatae Tschern. were discovered to be varied in their involucre structure, anther appendage shape, and basal leaves. Consequently, the classification now includes three sections instead of two, with a new section, Tianschanicae Sennikov, being distinguished from Alpinae. Additionally, Cousinia knorringiae has been reassigned from the Alpinae section to the Subappendiculatae section, while Cousinia omissa and Cousinia subappendiculata have been moved from Subappendiculatae to Tianschanicae.

Molecular studies started at the end of the 1990s. The phylogeny of the Cardueae tribe, studied by Häffner and Hellwig (1999) using ITS sequence data and later supported by Garcia-Jacas et al. (2002) through ITS and matK data, demonstrated a close relationship among the genera Cousinia, Arctium, Saussurea, and Jurinea. The molecular analysis of the Arctium-Cousinia complex by Susanna et al. (2003) identified two main lineages: the Arctioid clade, which includes Arctium s.str. and certain Cousinia subgenera (Hypacanthodes, Cynaroides, Hypacanthium, and Schmalhausenia), and the Cousinioid clade, consisting of Cousinia subg. Cousinia. This classification was supported by subsequent studies (Susanna et al., 2006; López-Vinyallonga et al., 2009). In 2011, López-Vinyallonga et al. transferred all Arctioid species to Arctium based on analyses of ITS, trnL-trnF, and matK sequences. Mehregan and Kadereit (2009) investigated potential hybridization among 214 species of Cousinia and related genera, concluding that there was no evidence of hybridization between Cousinia s.str. and other clades within the Arctium-Cousinia complex, nor between annual and perennial species of Cousinia s.str. Atazadeh et al. (2021) focused on species delineation within the sections Cynaroideae and Platyacanthae, but their ITS-based phylogenetic tree did not separate these sections. Kalouti et al. (2022) examined the monophyly of Cousinia sect. Stenocephalae, along with sect. Albidae and Cousinia, finding that none of these sections are monophyletic based on ITS data.

While various studies have used genetic sequence data from nuclear ITS regions and short chloroplast markers to explore relationships within the genus (Häffner and Hellwig, 1999; Garcia-Jacas et al., 2002; Susanna et al., 2003, 2006; López-Vinyallonga et al., 2009, 2011; Mehregan and Kadereit, 2009; Atazadeh et al., 2021; Kalouti et al., 2022), none have focused on the complete structure of the chloroplast (cp) genome, and only one cp genome of Cousinia is currently available in GenBank (Accession No. PP525141).

To address these gaps in understanding the evolutionary relationships within Cousinia, we sequenced the cp genomes of six Cousinia species. These include Cousinia rhodantha Kult. (Southwestern Pamir-Alay), C. pseudodshizakensis Tschern. & Vved. (Northwestern Pamir-Alay), C. proxima Juz. (Gissar Range), C. subcandicans Tschern. (Gissar Range), C. orthacantha Tschern. (Southwestern Pamir-Alay, Afghanistan), and C. rotundifolia C.Winkl. (Pamir-Alay), exhibit highly restricted distributions (Tscherneva, 1962), emphasizing their endemism and conservation importance. These species play crucial ecological roles in their native mountainous habitats, contributing to soil stabilization, supporting pollinators, and maintaining local biodiversity.

The primary objectives of our study were: 1) to analyze the cp genome structures of Cousinia species, 2) to identify the most divergent regions that can aid in species differentiation, 3) to make phylogenetic inferences to determine the position of Cousinia.

In this research, we performed sequencing of six species of Cousinia and carried out a comparative analysis of their cp genomes. This is the first report of complete cp. genomes from the genus Cousinia. They were found to be similar to those of other genera in Asteraceae with typical double-stranded circular tetramer structures and sizes ranging from 152,555 bp to 155,619 bp (Xing et al., 2019; Wu et al., 2019; Yun and Kim, 2022; He et al., 2023). The cp. genomes of six Cousinia species encoded 131 genes like most other species of Asteraceae, pseudogenes ycf1 and rps19 were also detected (Liu et al., 2024). A total of 16 intron-containing genes were identified in six plants of the genus Cousinia. These included 14 genes with one intron, and two genes with two introns. These intron-containing genes can play important roles in regulating gene expression (Wu et al., 2023). The total GC contents of the complete cp. genomes among the seven species of the Cousinia varied 37.67 to 37.75 ( Table 1 ). Among the LSC, SSC, and IR regions, the IR regions had the highest GC contents (43.09 to 43.12%), followed by the LSC (35.81 to 35.88) and SSC regions (31.35 to 31.55%). The IR regions had the highest GC contents among the four regions, which is likely due to the presence of two copies of rRNAs (rrna4.5, rrna5, rrna23, and rrna16) within this region (Shahzadi et al., 2020; Liu et al., 2024).

In this study, the repetitive sequences observed in the cp genomes of the seven Cousinia plant species displayed no significant heterogeneity in terms of their number, type, or length. Among the repetitive sequence types, reverse repeats were limited to only one in all species, while palindromic (P) repeats ranged from 18 to 21, and forward (F) repeats ranged from 20 to 23. Notably, complementary (C) repeats were absent in all seven Cousinia species. These long repetitive sequences play crucial roles in gene recombination and sequence structure variation (He et al., 2023). Furthermore, they may serve as valuable markers for differentiating Cousinia species, offering potential applications in species identification and phylogenetic studies (Gao et al., 2018).

SSRs, also called microsatellites, are widely utilized as valuable molecular markers in species identification and phylogenetic studies due to their high substitution rates (Upadhyay et al., 2010). In this study, a total of 273 SSRs were identified in the complete cp. genomes of seven Cousinia species, 158 were found to be mononucleotide repeats, constituting the majority (57.8%) of all SSRs ( Figure 4 ). Previous studies have suggested that this may be related to natural selection and genetic mutation (He et al., 2023; Zhang et al., 2023).

Vascular plant cp genomes usually exhibit the contraction or expansion of boundaries within different genera or even within the same genus, which is the main factor leading to variations in the length and number of genes in various species (Gu et al., 2024). The LSC/IRb boundary was generally located between the rps19 and rpl2 genes in the plastid genomes of six Cousinia species and Arctium lappa (NC042724). However, in Cousinia thomsonii (PP525141), the boundary was located within the rps19 gene. Upon further analysis of rps19 using Geneious v9.0.2, an illegal start codon was detected, suggesting potential annotation or sequencing issues that require further investigation. Moreover, the orientation of the SSC region in C. thomsonii was found to be forward, in contrast to the reverse orientation observed in other Cousinia species and A. lappa. This unique structural variation highlights the genomic divergence within the group. In C. thomsonii and A. lappa, the expansion of the IR region resulted in partially duplicated ycf1 genes being located in the IRb region, thereby generating a pseudogene ycf1 at the IRb/SSC boundary. While most land plants typically have pseudogenes in their cp genomes at the SSC/IRa boundary (Cai et al., 2023), the six Cousinia species analyzed did not exhibit annotated pseudogenes at this boundary. This absence is likely due to nucleotide insertions or deletions (indels) that disrupted the formation of a stop codon, preventing the annotation of pseudogenes in these regions.

Cp genomes provide essential resources for phylogenetic and taxonomic studies due to their conserved structure and variability in specific regions. In this study, we identified regions with nucleotide diversity (Pi) values exceeding 0.005, including trnE-UUC – rpoB, trnL-UAA – trnF-GAA – ndhJ, rbcL, and ycf1. These highly variable regions represent potential molecular markers for resolving phylogenetic relationships within the genus Cousinia. Previous studies have also demonstrated the effectiveness of cp markers such as trnL-trnF, trnL-trnT-rps4, and rpl32-trnL in differentiating Cousinia species (Susanna et al., 2006; López-Vinyallonga et al., 2009, 2011). The results highlight the utility of cp genome variability for species-level discrimination and evolutionary studies in Cousinia.

The phylogenetic analysis of the complete plastomes demonstrated congruent topologies, underscoring the reliability of plastome data for resolving phylogenetic relationships within the Cardueae tribe. One of the notable findings is the monophyletic clustering of most subtribes, with the exception of Saussureinae, indicating that this subtribe might be polyphyletic. This result aligns with previous studies suggesting complexities in Saussureinae’s taxonomy and phylogenetic placement (Herrando-Moraira et al., 2020). The basal position of the Carlininae subtribe is consistent with earlier phylogenetic reconstructions, supporting its potential as an ancestral lineage within Cardueae (Herrando-Moraira et al., 2019). Within the Arctiinae subtribe, the monophyletic grouping of Cousinia species with Arctium lappa supports previous findings by López-Vinyallonga et al. (2009, 2011), who identified a close phylogenetic relationship between these genera.

In this study, we investigated the molecular and morphological distinctions within Cousinia, particularly focusing on anther appendage shape as a taxonomic trait, alongside molecular phylogenetic data. Our results revealed two major molecular clades among the Cousinia species, which show some alignment with traditional sectional classifications, though with notable exceptions and inconsistencies. The first molecular clade, which includes C. pseudodshizakensis, C. rhodantha, and C. rotundifolia, corresponds to the Alpinae section. This molecular grouping is consistent with morphological observations, where species in the Alpinae section exhibit short, acute or obtuse anther appendages with an arched outline. However, the morphological data also suggest that C. princeps from the Alpinae section deviates in its anther appendage shape, which aligns more closely with species in the Tianschanicae and Subappendiculatae sections. This discrepancy indicates that while anther appendage morphology provides useful diagnostic features, it may not be entirely sufficient to classify species at a sectional level, especially when taxonomic boundaries overlap. The second molecular clade consists of species from the Dichotomae and Homalochaete sections, including C. orthacantha, C. proxima, C. subcandicans. Morphologically, these species share similar anther appendage traits narrow, elongated, acute, and toothed appendages in the Homalochaete section and short, acute, toothed appendages in the Dichotomae section. The congruence between molecular data and the anther appendage morphology for these species supports the close evolutionary relationship between the Dichotomae and Homalochaete sections. This finding is consistent with previous work by Rechinger (1972) and Sennikov (2010), who highlighted the taxonomic significance of anther appendage morphology in Cousinia. These traits serve as valuable markers in identifying evolutionary relationships within sections. To resolve the current uncertainties in the intersectional classification of Cousinia, particularly concerning the Alpinae, Tianschanicae, and Subappendiculatae sections, further molecular and morphological analyses are essential.

The comparative analysis of nrDNA ITS sequences, cp data, and morphological traits provided critical insights into the phylogenetic relationships within Cousinia. While plastome phylogenies demonstrated stable and congruent topologies, supporting well-defined monophyletic clades, ITS based phylogenies were largely consistent with plastome data, with only Carlina libanotica unexpectedly positioned within the Carduinae subtribe rather than Carlininae. This anomaly highlights the limitations of ITS as a sole marker for species delimitation. In particular, taxa belonging to the groups Stenocephalae, Albidae, and Cousinia had unclear phylogenetic positions, establishing paraphyletic or polyphyletic groupings. These results are consistent with earlier studies by Kalouti et al. (2022), which also reported evolutionary complexity in these areas.

Morphological data, particularly anther appendage shape, showed partial alignment with molecular clades, reinforcing its taxonomic relevance. For instance, the Alpinae section displayed relatively strong molecular-morphological congruence. Similarly, species from the Dichotomae and Homalochaete sections exhibited consistent molecular clustering, with anther appendage morphology further supporting their evolutionary affinity. These findings highlight the necessity of an integrative taxonomic approach, where plastome data offer robust phylogenetic resolution, ITS markers capture finer-scale divergence but require cautious interpretation, and morphological traits provide critical but sometimes homoplastic diagnostic features. A combined strategy incorporating plastome, nuclear, and morphological data is essential for refining the taxonomy and phylogenetic framework of Cousinia.

This study compares the entire cp genome sequences within Cousinia genus for the first time. There were 131 genes found in the cp genomes of six different Cousinia species, which varied in size from 152,553 to 152,619 bp. Higher nucleotide polymorphism levels were found in trnE-UUC – rpoB, trnL-UAA – trnF-GAA – ndhJ, rbcL, and ycf1 genomic regions, indicating possible markers for species differentiation. Within Cousinia, two different clades were identified by phylogenetic analysis using whole cp genomes. In addition, anther appendage morphology investigation revealed C. princeps from Alpinae section would need to be reclassified. Additionally, the phylogenetic analysis of the Cardueae tribe and its subtribes, along with the subtribal relationships within the genus, suggests that cp genome data holds significant potential for revising evolutionary and taxonomic frameworks within both Cousinia and broader Asteraceae studies. These results are preliminary because of the small number of species that were examined. Enhancing the intersectional classification within Cousinia requires additional study using a larger sample size.