Introduction

Accurate taxonomic identification, particularly at the intraspecific level (e.g., subspecies), plays a crucial role in understanding evolutionary processes, such as local adaptation and incipient speciation, and is essential for effective conservation planning. In taxa with wide distributions, morphological variability and hybridization potential, subspecies classifications may reflect underlying ecological divergence, cryptic genetic structure, or phenotypic plasticity. At the same time, anthropogenic changes such as habitat fragmentation, invasive species, and climate shifts exert increasing pressure on natural populations. Then assessing genetic variation within and among populations is essential for predicting their potential to adapt and persist under such changing conditions.

Senecioneae is the largest tribe in the family Asteraceae, represented by over 160 genera (more than 3,000 species), with new genera still being added [1,2]. The tribe is a diverse and taxonomically complex characterized by broad ecological amplitude [3], morphological variation, and frequent hybridization. Senecio L. is the largest genus in this tribe and one of the largest genera of flowering plants containing very different species, some are poisonous [4], while others are ornamental or have antimicrobial properties used in folk medicine [5,6]. According to various authors, genus contains 1,250–1,400 species [2,3,7–11], in Poland is represented by 11 species [12] or 27 species [13] depending on the classification in current taxonomy..

The large size of the genus, together with the extraordinary morphological variation both between and within species of Senecio, as well as widespread interspecific hybridization in this genus and other genera of Senecioneae [11,14,15], have greatly hampered attempts to create infrageneric classifications of Senecio. Furthermore, the considerable diversity of cytotypes within the tribe [2,16,17] contributes to its complex taxonomy and poorly understood evolutionary history. Recently, molecular and taxonomic studies [2,3,7,11,18,19] have shown the dynamic nature of classifications within the tribe Senecioneae with some groups being split off while others have been merged [20]. A notable example is the genus Jacobaea Mill., which now encompasses 45 species, including J. erucifolia, formerly classified within Senecio sect. Jacobaea (Mill.) Dumort [18].

Jacobaea erucifolia (L.) G.Gaertn., B.Mey. & Scherb. is a highly variable taxon with a range that extends across a large part of Europe, as well as Western and Central Asia. In Poland, it is mainly found in the southeastern part of the country and generally has a scattered distribution. Moreover, it was, until recently, a poorly recognized species, often even incorrectly identified due to its resemblance to the closely related Jacobaea vulgaris. Depending on the source, two subspecies are most commonly distinguished within this species in studies concerning European flora: subsp. erucifolia and subsp. tenuifolia (J. Presl & K. Presl) [21]. According to Iamonico and Managlia [8], even four subspecies are distinguished. In addition to the two mentioned above, two others: subsp. arenarius Soó and subsp. praealta are recognized.

Podsiedlik et al. [21,22], based on morphological analyses, demonstrated the presence of both subspecies (erucifolia and tenuifolia) in Poland. According to the authors, of the 31 inventoried preserved sites, four represent the subspecies erucifolia (an Atlantic taxon found in the western part of Poland, in Pomerania), while the remaining 27 represent the subspecies tenuifolia (a subcontinental taxon with a scattered distribution in the southern regions of Poland). Similar data, obtained from the morphological variability of the leaves of the studied herbarium materials, were published by Kruk and Sobisz [23]. All of these studies agreed that the vast majority of specimens examined were of the subspecies tenuifolia, while specimens designated as erucifolia constituted a minority. On the other hand, the presence of these intraspecific taxa in Poland raises some doubts, as it is not included in the available data on J. erucifolia localities. According to the only map of the distribution of J. erucifolia in Europe [24], only the nominative subspecies erucifolia is recorded in Poland. Furthermore, preliminary genetic analyses employing RAPD markers, conducted on material from 10 of the 23 previously morphologically analyzed populations (comprising 100 specimens), did not support the differentiation of the subspecies. RAPD markers showed that populations (considered to be subspecies. tenuifolia) clustered together with populations of subspecies. erucifolia in two main groups, which appeared to correlate more strongly with habitat characteristics than with taxonomic distinctions (see Supplementary material). While morphological studies identified these subspecies as distinct [21], the genetic evidence did not corroborate such differentiation. It is important to note that this research was limited to material currently available within Poland, without comparative analysis involving typical representatives of these subspecies from their broader geographic ranges. Due to a high variation between and within Jacobaea species, morphological discrimination between them is sometimes problematic and often prevents unequivocal identification. Additionally, inter-specific morphological differences may be blurred by the propensity for interspecific hybridization, a common phenomenon within this genus in natural habitats. Based on the results obtained, it was hypothesized that there are no two different subspecies in the Polish flora, and the morphological variability among populations likely reflects environmental adaptation, such as phenotypic plasticity, the development of ecotypes, or hybridization. This interpretation aligns with observations by Walter et al. [25] who documented significant differences in leaf morphology among Senecio ecotypes – one of the primary features used to distinguish the putative subspecies of J. erucifolia in Poland.

The concepts of species and subspecies are very important for understanding the evolution and ecology of living organisms and are now equally important for ensuring their protection and management [26,27]. Creating a subspecies category is also important for understanding geographic variation and speciation. The taxonomic position of subspecies has been the subject of long-standing disputes, mainly due to frequent inconsistencies between genetic results and subspecies identification based on morphological data [25,26,28–35] which raises additional questions regarding the validity of a number of subspecies and may lead to changes in taxonomy. Until now, the widely accepted definition for subspecies was that of Mayr & Ashlock [30]: “A subspecies is an aggregate of phenotypically similar populations of a species inhabiting a geographic subdivision of the range of that species and differing taxonomically from other populations of that species.” However, a kind of molecular revolution has changed the traditional methods (morphological features) of subspecies classification through the use of genetic analyses [26]. Currently, it is assumed that the designation of a subspecies should result from the consistent distribution of many independent genetic characteristics. Thus, subspecies are identified as geographically and genetically distinct populations that exhibit distinct phenotypic differences [27].

Morphological characters for a long time have been practically the main criterion used in plant systematics. However, we must to take in to account that they may be influenced by environmental conditions, and then the morphological differences may be due to differences in habitat rather than being a sign of genetic differences between the groups analyzed [7]. Different species and subspecies, especially those sharing similar habitats, often hybridize with each other, which may lead to phenotypic changes and frequently results in challenges or inaccuracies in taxonomic identification. The occurrence of hybrids with intermediate morphological appearance may lead to the definition of an excessive number of species names within the genus, nomenclature problems and discrepancies in taxonomic approaches [4,19,36]. In such a situation, ambiguities in the species complex can be clarified by molecular investigations [37]. The use of environmentally independent DNA markers can provide more accurate information on genetic diversity, provide insight into speciation processes, and in combination with the analysis of morphological characters is fundamental for taxonomy at the species level.

The most common molecular methods include analyses of organelle genomes (cpDNA and mtDNA), sequence comparison of nuclear ribosomal DNA internal transcribed spacers (ITS) regions, and clustering-based methods that incorporate multilocus genotype data, such as randomly amplified polymorphic DNA (RAPD) [38] or amplified fragment length polymorphism (AFLP) [39]. AFLP, in particular, is a robust and versatile tool for investigating genetic variation. It is widely utilized in genotyping studies, as well as in assessments of population differentiation, and genetic diversity across a variety of organisms, including members of the genus Senecio [17]. High levels of polymorphism and high degrees of discrimination are the main advantages of AFLP in distinguishing closely related accessions and assessing genetic diversity within populations. Crucially, the estimation of diversity with this method depends on the number and genomic distribution of markers used and is entirely independent of morphological characteristics. Compared to alternative approaches, AFLP generates a higher number of products (polymorphic fragments and unique bands) in a single reaction and is highly reproducible. All of these methods have already been used to assess relationships, genetic differentiation, between individuals, to demonstrate introgression, and have been successfully applied in hybridization studies within different Senecioneae species [4,11,40–44]. The current study will employ these methods to address the research objectives.

Environmental factors can influence many aspects of plant morphology, and some of the observed differences may arise from various non-genetic causes such as local habitat conditions or introgressive hybridization. Plants exhibit high plasticity in response to environmental stimuli, and many taxa below species rank are described solely based on their morphology. From a taxonomic standpoint, studying such variation is crucial. Moreover, according to Chevin et al. [45], the persistence of species in new or changing environments depends on the ability to adjust phenotypes in response to environmental variability. Importantly, such phenotypic plasticity can generate distinct morphological forms from a single genotype depending on the conditions [46,47]. Understanding phenotypic plasticity is also important for explaining how some taxa—including those with uncertain taxonomic status—may establish in new or disturbed habitats, as is the case for several Jacobaea populations observed in altered environments.

Widespread changes in the natural environment, including variations across gradients of latitude, longitude, and altitude, significantly influence the selection of vegetation types and the species found within them [48]. Locally, changes in the plant environment may concern soil type, humidity, or temperature, and influence the adaptation of local types, which adapt to environmental changes through phenotypic plasticity or adaptive genetic divergence. According to Mayr [49], adaptation along an environmental gradient, with or without gene flow, may be the first step leading to the emergence of fully reproductively isolated forms, known as biological species.

Given the need to understand how phenotypic traits respond to environmental variation, especially in species showing morphological divergence without clear genetic structure, it becomes important to investigate how genotypes relate to phenotypes across different habitats, as well as the processes influencing such patterns [25]. Equally important is understanding the interactions between phenotypic plasticity and natural selection, and how and when populations specialize in particular environments. The genus Senecio is an ideal model, with species that naturally interbreed, inhabit different environments, and undergo phenotypic differentiation and adaptation [25]. Within the genus, there are several examples of radiation mechanisms. Some of these have been described as introgressive hybridization [50], local adaptation [51], or hybrid speciation [52,53]. For example, introgressed lowland species, such as S. chrysanthemifolius and S. ovatus, have been observed to move to higher altitudes, possibly in response to recent global warming, displacing their higher-altitude counterparts from these areas [48].

Genetic variation is essential for populations to persist in the face of environmental pressures, including those driven by human activity. Habitat fragmentation, agricultural intensification, and the introduction of invasive species can reduce effective population size and gene flow, leading to a loss of genetic diversity. This, in turn, limits a population’s ability to adapt to new or changing environmental conditions. In this context, evaluating the genetic structure of populations is crucial for understanding both their evolutionary potential and their vulnerability.

In the face of global climate change and human activity, the significant role of hybridization cannot be overlooked when considering adaptation and the occupation of new habitats. On one hand, hybridization can lead to the formation of new species and adaptive traits (such as the ability to occupy new niches and expand range) [54,55]. On the other hand, it may also lead to the replacement of native species by invasive ones [56,57], potentially resulting in a loss of biodiversity. In the absence of reproductive isolation, hybridization can reduce or eliminate differentiation (introgressive hybridization; [58], leading to the genetic erosion of parental gene pools [59]. In the case of local endemic species and small isolated populations, introgressive hybridization can pose a significant threat. Human activity also plays a significant role in promoting hybridization [60], which, driven by anthropogenic factors, is one of the key problems for conservation. In light of this, maintaining existing biodiversity has become a major challenge [60–62]. Moreover, environmental changes and the introduction of exotic species can lead to demographic or genetic swamping, threatening the integrity of small populations or rare species [60,63]. Due to increasing global changes, the rate of hybridization appears to be rising [62,64].

Available data from literature, herbarium materials, and field inspections have shown that the process of Jacobaea. erucifolia’s (hoary ragword) decline is progressing, largely due to the intensification of agriculture and the elimination of fallow lands, which are currently the most frequently recorded habitat of this species. Habitat loss is further exacerbated by the spread of invasive species, such as Solidago gigantea and Calamagrostis epigejos, which have been extensively observed in areas previously occupied by J. erucifolia. as well as the effects of urbanization in areas where the species once occurred in Poland [22]. The species is now rare at the national level and deserves special protection [22,23].

In this study, we employed a combination of molecular tools— AFLP fingerprinting, cpDNA restriction site variation, and sequence analysis to investigate the genetic status of Polish populations of J. erucifolia in the context of their variability within the species’ European range.. We aimed to test whether the morphological variation observed in these populations corresponds to genetically distinct subspecies or instead reflects adaptation to environmental conditions or introgressive hybridization. This approach reflects broader questions in plant taxonomy and conservation regarding the reliability of morphological criteria for subspecies recognition under anthropogenic pressure. Our findings also contribute to conservation planning by clarifying the genetic distinctiveness of declining Polish populations. Specifically, the objectives of this study were: (1) to assess the level of genetic variation in Polish populations of J. erucifolia in comparison with populations from other countries within the continuous distribution range of the species, and (2) to re-evaluate the taxonomic validity of the subspecies tenuifolia in Poland, using comparative material. from Slovakia Hungary and the Netherlands.

Materials and methods

Plant material

A total of 127 individuals of J. erucifolia (representing both putative subspecies) were included in this study, originating from 15 populations in total (Table 1). The sampling covered ten populations from Poland, four from Hungary, and one from the Netherlands. Among the Polish populations, four represented J. erucifolia subsp. erucifolia (Western Pomerania) and six were described as J. erucifolia subsp. tenuifolia. Each Polish population was represented by ten individuals, selected to ensure wide geographic and ecological coverage, based on morphological classification and site accessibility. The Hungarian and Dutch populations have not been formally assigned to subspecies; however, based on morphology, they most likely correspond to subsp. tenuifolia (tentatively referred to as J. erucifolia in Table 1). From each of the 15 populations, ten individuals were sampled (although in a few cases only three or five individuals were available). All of these samples were collected during fieldwork. Additionally, six individuals of J. erucifolia subsp. tenuifolia originating from six different Slovakian regions were included, based on herbarium material deposited in the Herbarium NI, Department of Botany, Faculty of Agrobiology and Food Resources, Slovak University of Agriculture, Nitra, Slovakia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. List, origin and number of individuals in the populations studied.

https://doi.org/10.1371/journal.pone.0332808.t001

Due to the small number of individuals (single specimens) from particular locations, the samples from Slovakia were not treated as populations but as a reference group (in most statistical analyses the required number of individuals is at least three or more). Twenty-nine individuals of J. vulgaris (2 populations from Poland and one from the Netherlands) served as an out-group. In our abbreviation system, the first letter denotes the country of origin for foreign populations, while the subsequent letters refer to a specific locality, typically a nearby town (both capital letters). Polish populations are labeled directly by sampling site names. The same material has been previously used in morphological analyses as well as in our earlier study on hybrids between Jacobaea erucifolia and J. vulgaris. In the current study, we intentionally did not use the letter “P” to denote Polish populations in order to maintain consistency in the naming and localization system applied to all samples. Details of the specimens tested are provided in Table 1.

The choice of J. vulgaris as an out-group was related to the results of previous studies documenting the occurrence of spontaneous hybridization, between J. vulgaris and J. erucifolia, especially in areas where both species occurred. Moreover, our analyses showed that some individuals in populations previously described as pure were probably also of hybrid origin, backcrosses to J. erucifolia and J. vulgaris [65]. The fact that both species are genetically different and frequently interbreed facilitates the assessment of the studied J. erucifolia populations in terms of introgression.

Chloroplast DNA haplotypes

The usefulness of universal primers for assessing chloroplast polymorphism among plant populations was demonstrated for many species. These universal primers are located in coding regions of cpDNA, separated by more variable intergenic regions [79,80]. One of the popular methods useful in population analyses is the PCR-RFLP method (combining both the PCR and RFLP techniques), which allows for the distinction of genotypes based on the presence or absence of restriction sites in the amplified DNA.

The total DNA of individual plants was used in PCR reaction, the variation of the chloroplast genome was tested using a set of universal primers. In a first step, four pairs (LF, CS, HK, and K1K2) of chloroplast primers were used to identify and screen polymorphism on individuals from the populations investigated. Three of the pairs of primers mentioned above were selected because of their good amplification pattern. They were included the non-coding region between trnL and trnF genes of cpDNA, CS region psbC and trnS and HK region trnH trnK. The first fragment was amplified using universal primers described in Taberlet et al. [81], while the CS and HK regions were amplified with pairs of universal primers described in Demesure et al. [79]. (Table 2)

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Chloroplast DNA (cpDNA) universal primer pairs used in this study.

https://doi.org/10.1371/journal.pone.0332808.t002

PCR reactions were carried out in a final volume of 25 μL containing 20 pmol each of primer and about 40 ng of template DNA using the KAPA3G Plant PCR Kit (Sigma Aldrich). Reactions were performed in a PTC-100 thermal cycler (MJ Research), according to the previously described program [65]. Annealing temperature and the extension time were dependent upon the primers used and the length of the PCR product. The efficiency and quality of amplification were verified by electrophoretic separation of PCR products in 1.5% agarose gel. The size of the amplified fragments was estimated using a 100 bp ladder DNA marker (Thermo Fisher Scientific). PCR products (5µl) obtained for several samples representing both putative subspecies of J. erucifolia were digested with 18 restriction endonucleases: AluI, HinfI, RsaI, TruI(MseI), EcoRI, BsuRI, HhaI, HindIII, HpaII, TagI, Sau3a, MboI, EcoRV, KpnI, SacI, PstI, DraI and Alw26I under conditions recommended by the enzyme supplier (Thermo Fisher Scientific). The enzymes used for digestion were selected randomly. Based on their ability to cut and reveal polymorphism in cpDNA regions, four chloroplast fragment-enzyme combinations (CS–Alw26I, CS–BsuI, HK–RsaI, and HK–HinfI) that exhibited readable polymorphism were selected. In a second step, only these combinations were used to characterize a total of 127 individuals from 15 populations and six individuals representing subspecies tenuifolia from Slovakia. Restriction digests were separated by electrophoresis in 8% polyacrylamide gels using Tris-borate EDTA buffer (0.5) at 70 V overnight and visualized by silver staining. 100-bp ladder marker (Thermo Fisher Scientific) was used as a size marker on each gel To determine relative size differences, after the first size measurement, the different variants for each polymorphic locus were re-ran on the same gel. cpDNA variants at each locus were numbered, the numbers increase from the highest to the lowest molecular weight fragments. A binary matrix was created based on the presence or absence of each restriction fragment at each polymorphic site. The similarity matrix (estimated using the simple matching coefficient from the binary matrix) was used to construct a dendrogram by unweighted pair group method average (UPGMA). Variation in the restriction patterns was interpreted as a haplotype and the haplotype nomenclature was used as described by Petit et al. [82].

Results

AFLP analyses

To evaluate patterns of genetic diversity and structure among populations, we first analyzed AFLP data, which provides multilocus genome-wide information independent of sequence assumptions. The study included 156 samples from 18 populations (including three populations of J. vulgaris) from Poland and several European countries. Fifteen of them were represented by both putative subspecies of J. erucifolia. Six individuals from Slovakia, referred to as subspecies tenuifolia were used as the reference group.

Among 15 AFLP primer combinations prescreened, four were selected. AFLP fingerprinting using the 4 selective primer combinations yielded 687 clearly resolved and unambiguously scored fragments including those characteristic for populations of J. vulgaris consisting an out-group (54 products). In populations of J. erucifolia (both subspecies), 625 out 633 (98%) were polymorphic across all populations (127 accessions). The number of polymorphic bands per primer combination ranged from 149 to 172, with an average of 158 bands. The percentage of polymorphic bands varied among the different primer combinations. The highest percentage of AFLP polymorphisms was obtained for EcoACA/MseCAT and EcoACA/MseCCG combinations, where 149 out of 151 and 163 out of 169 (98.7 and 96.4%), respectively, bands were polymorphic. The size of the AFLP amplified fragments ranged from 50 bp to 500 bp. In general, the number of polymorphic AFLP fragments characterizing intra-populational variation in Polish populations of J. erucifolia regardless of the subspecies was higher than in other populations and averaged 92,5%. The lowest values: 290 polymorphic fragments and a polymorphism of 73% were observed for the NZ population from the Netherlands (Table 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Genetic diversity estimates within the J. erucifolia populations (both subspecies) based on 633 AFLP fragments.

https://doi.org/10.1371/journal.pone.0332808.t003

In order to confirm the morphological differences between subspecies, the genotypic profiles of these taxa were also examined and checked whether there were any differences in DNA fragments between them. The search for polymorphic private AFLP fragments per population as well as per taxon was carried out. In the EcoACA/MseCAG combination, one fragment was noted that met the criteria for defining it as private for the P population, the 399 bp fragment was present in all 10 individuals and absent in the other populations. The presence of a similar fragment 396 bp in EcoACA/MseCCG combination, was also demonstrated for the S population (7 out of 10 specimens). Profile analysis revealed a number of unique products usually, occurring in single individuals only in one or several populations in the group including the erucifolia subspecies. In the S population, 14 of them were noted (some were shared with other populations described as erucifolia subspecies). A total of 59 unique products were found in the Polish population profiles. However, such fragments also frequently appeared in the Slovak reference group. In comparison with Polish populations, only 11 unique fragments were recorded in the profiles of populations recognized as subspecies tenuifolia. As with the previous group, fragments most often occurred in single individuals or in single populations. According to the commonly accepted definition, diagnostic markers were considered to be those that are present in more than 95% of individuals of one taxon and absent in another taxon. None of the mentioned unique fragments met these requirements and could not be classified as a species-specific diagnostic marker.

Cluster analyses were performed based on data from polymorphic AFLP bands including specimens from the out-group. The estimated similarity coefficient based on the combination of 687 markers. The pairwise Dice’s similarity coefficient values between all possible pairs of genotypes ranged from 0.172 for P3 and Pu3 to 0.805 between two J. vulgaris accessions. The principal coordinate analysis (PCoA) divided all accessions into three groups between the two first coordinates, which accounted for 24.9% of the total genetic variation (15.7% and 9.2% for the first and second axes, respectively) (Fig 1). The first well-separated group was formed by two Polish populations of J. vulgaris (Pu and G served as an out-group), the largest group encompassed individuals of J. erucifolia identified on the basis of morphological features as both subspecies in Poland. Most of them formed a loose grouping without a clear division into subspecies or population. The third group, located between them included accessions represented by J. erucifolia from Hungary and the Netherlands as well as 6 individuals from Slovakia representing J. erucifolia subsp. tenuifolia and J. vulgaris individuals from the Netherlands. Nevertheless the latter (specimens NP1-NP10) formed a subgroup loosely connected to this group.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Principal coordinates analysis (PCoA) plot based on the individual genetic distance calculated with 687 AFLP markers.

https://doi.org/10.1371/journal.pone.0332808.g001

The first two axes explain 15.7% and 9.2% of the variation, respectively. Individuals are represented by dots in colors, (red corresponds to J. vulgaris, blue to J. erucifolia subsp. erucifolia, green to J. erucifolia subsp. tenuifolia and sea blue for individuals/populations not assigned to a subspecies during collection, probably subsp. tenuifolia).

Both the UPGMA dendrogram and the Dice Bio NJ network (Fig 2) showed good agreement and confirmed the results of the principal coordinate analysis indicating no separation between J. erucifolia subspecies in Poland. Compared to PCoA, although the major clustering pattern was found to be similar, changes in minor grouping were observed. NJ tree revealed three clusters corresponding to the main groups in PCoA. (Fig 1). The Polish individuals of J. vulgaris formed a homogeneous group I. As before, individuals from all foreign populations (HT, HI, HL, HP, NZ) and most individuals constituting the reference group (SS, SB, SU, SL, and SV) from Slovakia grouped together and formed group II. The subclaster consisting of individuals of J. vulgaris from the Netherlands, previously located within this group, now separated clearly and formed a branch within the J. vulgaris group, (Polish representatives of this species). All remaining individuals clustered together within the largest cluster III, further subdivided into several smaller subclusters (A, B, C, D). Three of them were heterogeneous with accessions scattered randomly, independently from their taxonomic position and apart from the P (Pęczelice) and B (Brwinów) populations also of their geographic origin. Interestingly, the smallest subclaster IIIC was represented by 3 individuals from Kłosów (K), one from Szczecin Skolwin (S) populations corresponding to the erucifolia subspecies, previously described as probably backcrosses [65] and SZ individual from Slovakia representing tenuifolia subspecies. Also noteworthy were five individuals not assigned to any group, which were located between the large group IIID and the branches formed by J. vulgaris. Three of them (Sn1, Sn4, Brz6) formed a small group, while the remaining two (HP4 and HP5) occupied separate positions without forming one group with them (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Neighbor-Net derived from AFLP binary matrices of 156 individuals representing Jacobaea erucifolia (both subspecies) and J. vulgaris (out-group).

https://doi.org/10.1371/journal.pone.0332808.g002

Individuals are represented by dots in colors, (red corresponds to J. vulgaris, blue to J. erucifolia subsp. erucifolia, green to J. erucifolia subsp. tenuifolia and sea blue for individuals/populations not assigned to a subspecies during collection, probably subsp. tenuifolia). It should be noted that the Dutch individuals morphologically representing J. vulgaris form a separate group, a cluster which, although related to the branch of this species from Poland, is located closer to the cluster formed by the J. erucifolia subsp. tenuifolia. Accessions codes as in Table 1.

A similar pattern of genetic diversity was observed in the dendrogram and PCoA based on pairwise genetic distances among populations when the analysis was performed at the population level. The NJ tree revealed two well-supported clusters (bootstrap value of 100%). One group included Polish populations (both putative subspecies), while the second group consisted of populations from Hungary, the Netherlands, and Slovakia (subsp. tenuifolia). In both analyses, the position of two populations, P and B, is noteworthy, as they clearly separate from the others (Fig 3). The population from Pęczelice (P) was characterized by significantly lower similarity coefficient values compared to the other populations. The lowest similarity was recorded in relation to the Hungarian HI population (0.666). The remaining populations showed relatively high similarity coefficient values, with a maximum of 0.879 between the Pa and Sn populations.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Grouping of populations of J. erucifolia subspecies into clusters based on Dice distances calculated from AFLP binary matrices.

(A) Scatter plot of PCoA of studied populations. Two first coordinates explain 52.4% of the total observed variation. (B) An unrooted neighbour-joining (NJ) tree, numbers at the nodes indicate bootstrap support. Names of populations refer to Table 1 (SK-code for reference group). Individuals are represented by dots in colors, (blue to J. erucifolia subsp. erucifolia, green to J. erucifolia subsp. tenuifolia and sea blue for individuals/populations not assigned to a subspecies during collection, probably subsp. tenuifolia).

https://doi.org/10.1371/journal.pone.0332808.g003

The analysis of molecular variance (AMOVA) of J. erucifolia populations was performed twice, using geographic origin and taxonomic position (subspecies) as the grouping criteria. In both cases, AMOVA attributed most of the genetic variation to within-population variation (83–84%), rather than between populations or subspecies, regardless of the type of analysis (Table 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 4. Analysis of molecular variance (AMOVA) for J. erucifolia populations based on AFLP markers.

https://doi.org/10.1371/journal.pone.0332808.t004

Prior to interpreting the STRUCTURE results, the most likely number of clusters (K) (S2 Fig) for each dataset was estimated using the ΔK method of Evanno [77]. This approach provides a more robust justification than raw likelihood values, which are associated with large variance among replicates. In the STRUCTURE analysis (Fig 4A), which included all Polish populations and individuals from Slovakia, the best-supported number of genetic groups was found to be four (K = 4). Although the studied populations differed in their allele composition, the occurrence of similarly colored segments indicated the presence of shared alleles. The proportion of these shared alleles varied across populations, but did not allow for the identification of distinct groups. However, two exceptions were noted: the nearly homogeneous population P, and two populations (B and Brz), which exhibited an exceptionally high level of admixture (Fig 4B). This analysis step, restricted to Polish and Slovakian populations, provided a direct comparison of Polish material with the only available reference samples of J. erucifolia subsp. tenuifolia.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Clustering analysis of AFLP data for J. erucifolia (both subspecies) and J. vulgaris populations, and their distribution in Poland and other countries.

(A) A map with pie charts showing the proportion of membership in each cluster, corresponding to the bottom barplot with STRUCTURE results for K = 3. (B) Clustering analysis using Bayesian analysis (STRUCTURE) for Polish populations and the reference group from Slovakia, for K = 4. (C) Analysis after adding the remaining populations representing subsp. tenuifolia from Hungary and the Netherlands, for K = 2. (D) Clustering analysis including all J. erucifolia together with populations of J. vulgaris, for K = 3. Individuals are represented by columns, with colors representing the proportion of their genome assigned to the inferred clusters in the model-based admixture analysis. Symbols identify the different populations investigated. Additional details for each population are given in Table 1. Map created in QGIS [QGIS Development Team, 2025); land, sea, and administrative boundary layers were sourced from Natural Earth (naturalearthdata.com).

https://doi.org/10.1371/journal.pone.0332808.g004

When the remaining populations from Hungary and the Netherlands were added, the best-supported number of clusters decreased to K = 2. The observed pattern of differentiation corresponded to the division of the samples into two groups (Fig 4C). The first group, encompassing all Hungarian populations, the population from the Netherlands, and individuals from Slovakia (morphologically defined as J. erucifolia subsp. tenuifolia), was relatively homogeneous. In contrast to most Polish populations, these populations showed minimal variation. The second group included Polish populations represented by both subspecies (defined on the basis of morphology), and showed a clear overlap of the gene pools of both groups, except for populations P and B which exhibited very slight admixture. Based on the Mantel test, no significant correlation was found between genetic (PhiPTP) and geographic distances (R² = 0.026, P = 0.132) (Fig 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Scatterplot showing the results of correlation (Mantel test) between Φ_PT coefficient values and geographic distance for J. erucifolia (both subspecies) populations.

https://doi.org/10.1371/journal.pone.0332808.g005

Given the previously demonstrated spontaneous hybridization and evidence of introgression between J. erucifolia and J. vulgaris [65], STRUCTURE analysis was also performed after including different J. vulgaris populations. When these populations were added, Bayesian clustering revealed that the best-supported number of clusters was K = 3, dividing the samples into three main groups (Fig 4D). Although the results for two of the groups (J. erucifolia) were similar to those found for K = 2, slight admixture of J. vulgaris was observed in most of the populations. In the third clade, encompassing J. vulgaris populations (G, Pu, and Nz from Poland and the Netherlands, respectively), an overlap of J. erucifolia gene pools was also observed. Moreover, in the latter population (assigned to J. vulgaris), the cluster typical of J. erucifolia dominated clearly. The genetic affinity revealed by the STRUCTURE analysis was consistent with the results of the NJ network analysis. All analyses conducted using AFLP markers clearly indicated that the populations from Hungary and the Netherlands are very similar to the representatives from Slovakia, which are classified as the tenuifolia subspecies, and were treated as such in subsequent analyses. Some of the Polish populations, which were also assigned to this subspecies based on morphological features, consistently formed a separate group in each analysis, alongside the populations defined as the erucifolia subspecies. Furthermore, as these populations exhibited no significant differences, they were collectively referred to as Polish populations of J. erucifolia in the subsequent sections.

Analysis of chloroplast haplotypes

To complement the genome-wide data, we next assessed plastid diversity through cpDNA RFLP analyses, which reveal maternal lineage patterns and are useful for inferring phylogeographic structure. Initially, four universal primer pairs of (LF, CS, HK, and K1K2) were used for cpDNA amplification. However, due to their high amplification efficiency, three primer pairs (trnC/trnS, trnH/trnK, and trnL/trnF) were selected for this study (Table 2). The estimated sizes of the fragments amplified with the CS, HK, and LF primer pairs were approximately 1600, 1800, and 960 bp, respectively. No length polymorphisms were detected in the PCR products using any of the primer pairs tested.

PCR products were digested by six selected restriction enzymes (Alw26I, BsuI, RsaI, HpaI, ThruI(MseI), and HinfI) to identify polymorphisms, as described by Dumolin-Lapegue et al. [80]. Of the fragment-enzyme combinations, four (CS–Alw26I, CS–BsuI, HK–RsaI, and HK–HinfI) yielded clear polymorphic patterns. The LF region, while initially amplified, showed minimal variation among individuals when digested with four enzymes (LF–HpaI, LF–RsaI, LF– ThruI(MseI) and LF–Alw26I). Therefore RFLP analysis was primarily focused on the CS and HK regions. The LF region, however, was analyzed based on sequencing.

The selected four combinations generated 12 polymorphic fragments and gave 33 length mutations primarily short length variants, presumably indels. For a single locus the number of alleles ranged from two to four. The CS–Alw26I combination proved to be the most polymorphic, with 12 distinct patterns.The other three combinations (CS–BsuI, HK–RsaI and HK–HinfI) yielded 9, 7, and 5 different patterns, respectively (S4 Table).

Analysis of the polymorphism of plastid DNA markers (all four combinations together) revealed 99 different haplotypes across the 127 accessions studied. Nineteen haplotypes were shared by two or more accessions, while the remaining 80 were unique (private) haplotypes for the respective accessions. Among these, private haplotypes, 65 were found in Polish populations of J. erucifolia, and 15 were found in individuals of J. erucifolia subsp. tenuifolia. This high number of unique haplotypes reflects high intra-population variability.

Using all combinations of fragment-restriction enzyme for haplotype identification provides resolution that is too high to draw broader conclusions. Therefore, in the subsequent analysis, we focused on the CS-Alw26I combination, which generated 27 haplotypes (Table 5). Most populations were polymorphic and eight of them had unique haplotypes. Populations P (Pęczelice, Poland) as well as HT (Tápiószentmárton, Hungary) were monomorphic and represented by only one haplotype H1 and H5 respectively. The most common haplotype was H1, occurred in nine populations from Poland and individuals from Slovakia (20 individuals, frequency = 0.157), and was absent in other countries. Equally abundant was the H2 haplotype (6 populations, 12 individuals, frequency = 0.150), which occurred exclusively in Polish populations and was present regardless of subspecies affiliation. The third group of the most frequent haplotypes were haplotypes H3 and H19 (each present in 5 populations, in 11 and 12 individuals respectively). The four most frequent haplotypes (H1, H2, H3, and H19) were found in 62 individuals from various populations accounted for 48.8% of all individuals.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Haplotype frequencies and composition of the populations of Jacobaea erucifolia (both subspecies) based on combination CS-Alw26I.

https://doi.org/10.1371/journal.pone.0332808.t005

The remaining haplotypes were less common and unique to specific populations (11 haplotypes in total). Sixteen of the identified haplotypes were shared by two or more populations. The frequency of the private haplotypes ranged from 0.008 to 0.031 (Table 5). In the Sn, Brz, and St populations, private haplotypes were particularly abundant (specifically: H8, H11, and H20, respectively). The St and Pa populations exhibited the greatest haplotype richness, with as many as seven haplotypes identified in each. Moreover, the maximum number of private haplotypes (two) was also present in the St population and two other HL (Hungary) and Pa (Poland) populations. Reference individuals for the tenuifolia subspecies (SS, SB, SU, SL, SV, and SZ) shared haplotypes with Polish populations. In all populations representing the tenuifolia subspecies, the haplotype diversity was relatively low. Most populations in this group contained two haplotypes, with the exception of the HL population, where four haplotypes were found (two of which were unique). In Polish populations, haplotype diversity was much higher (5–7 in one population), and some haplotypes were not found in the populations of the tenuifolia subspecies. There was no haplotype common to all populations. Aside from the H2 haplotype, which occurred only in Polish populations (mostly), there was no clear pattern of haplotypes occurring exclusively in populations assigned to a given subspecies. When subspecies-specific haplotypes did occur, they were present in only 1–2 populations per group, and most were unique haplotypes (for details, see Table 5).

Detectable variation was also found in two other combinations: CS–BsuRI, and HK–RsaI + HK–HinfI, (analyzed together). The analysis revealed the presence of 20 and 33 different haplotypes, respectively, with a significant number of them being unique (7 and 22, respectively). As before, most populations were polymorphic. Of the two monomorphic populations, both were represented by the H1 haplotype found in Hungarian populations. As previously noted, certain haplotypes appeared with higher frequency. In both combinations, the St population once again proved to be the most diverse. Information regarding haplotypes generated using the above combinations is provided in the supplementary materials (S1 and S2 Table).

Both the UPGMA dendrogram and PCoA analysis, based on a binary matrix of restriction fragment presence/absence, divided the populations into three groups. The smallest group consisted of two Hungarian populations (HT and HI). The two larger groups, further subdivided into smaller subgroups, included populations representing different subspecies. Notably, the reference group itself was split, with its members distributed across two distinct clusters (Fig 6).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Grouping of populations of J. erucifolia.

https://doi.org/10.1371/journal.pone.0332808.g006

Clustering based on principal component analysis of data on presence and proportions of chloroplast DNA haplotypes in populations (A) and UPGMA dendrogram (B). Names of populations refer to Table 1.

Individuals are represented by dots in colors, (red corresponds to J. vulgaris, blue to J. erucifolia subsp. erucifolia, green to J. erucifolia subsp. tenuifolia and sea blue for individuals/populations not assigned to a subspecies during collection, probably subsp. tenuifolia).

In summary, no taxonomic distribution of haplotypes was observed, dendrogram analysis did not reveal any relationship between the haplotypes present in a given population and its subspecies affiliation. The composition of the clusters formed suggests that the populations are more similar in terms of the habitats they occupy rather than their taxonomic classification (Table 1). Similarly, no clear geographical pattern of haplotypes distribution was observed in the study area.

Discussion

Genetic diversity is the basis of the evolutionary process, and in the face of global climate change, it has become one of the most critical factors for monitoring the effects of climatic alterations on intraspecific genetic diversity. According to May [85], genetic variation is a fundamental component of biodiversity. Global climate change is widely expected to affect biodiversity, species distribution, and population growth in numerous ways. These effects include changes in population and species ranges [86], as well as impacts on phenotypic plasticity and adaptation in response to changing environmental conditions. Such changes may reduce genetic diversity within populations and species, thereby decreasing population viability and potentially leading to extinction [87,88]. Additionally, anthropogenic factors, such as land use changes, have a significant impact on migration patterns and adaptation to environmental conditions [87]. In light of this, more attention should be given to species whose populations are particularly vulnerable to these impacts, especially those at risk of extinction. Assessing the genetic diversity of species to determine the current patterns of population distribution and genetic variation is especially important considering the long-term consequences of current global environmental changes. Molecular methods, such as AFLP and cpDNA analyses, are invaluable tools for studying population structure and inter-population genetic diversity.

Jacobaea erucifolia is a rare species in Poland, occupying anthropogenically modified habitats such as roadsides, pastures, abandoned vineyards, and waste areas [17,89]. However, habitat disturbances can also create new ecological niches fostering the persistence of diverse forms, such as stabilized introgressants and homoploid hybrid species [89,90].

Two main subspecies (erucifolia and tenuifolia) have historically been recognized in Poland, primarily based on leaf shape and ecological preferences. J. erucifolia subsp. tenuifolia exhibiting a broader ecological range. The greatest difference between the two subspecies is the index of continentalism: J. erucifolia subsp. erucifolia is an Atlantic taxon, while the second subspecies is subcontinental. In terms of morphology, aside from leaf variability, the most notable differences (micromorphological) concern the hairs present on the surface of the achenes [91]. Although morphological and micromorphological analyses have shown the presence of both subspecies in Poland [21], no studies have yet demonstrated that these subspecies are genetically distinct. Furthermore, individuals classified as tenuifolia subspecies have not been compared with native populations occurring in neighboring regions.

RAPD-based analyses previously indicated that morphological subspecies in Poland are not genetically distinct. RAPD markers divided populations into two well-defined groups, with mixed clusters formed by individuals from both subspecies. Cluster analysis suggested that the grouping was influenced more by habitat similarity than genetic differentiation between the two taxa (S1 Fig). This justified the use of higher-resolution molecular markers such as AFLP and cpDNA, supported by additional sampling from other regions.

AFLP fingerprinting analysis revealed high polymorphism within populations and no consistent separation between subspecies. However, Polish populations (regardless of subspecies) exhibited a slightly higher degree of polymorphism compared to the others. Overall, the level of genetic diversity within the subspecies did not differ significantly, suggesting that their gene pools are not separate. Variation in sample sizes, especially in foreign populations, reflects differences in local population sizes and accessibility at the time of collection. While Polish populations were consistently represented by 10 individuals selected for broad geographic and ecological coverage, the number of individuals from Hungary, the Netherlands, and Slovakia was limited by available material. Although smaller sample sizes may influence some statistical outcomes, particularly estimates of diversity and structure, the consistency of clustering patterns across multiple methods supports the reliability of the overall genetic signal. Among the total amplified polymorphic products, none distinguished the two subspecies, and no unique products were classified as diagnostic or species-specific.

An AMOVA attributed most of the genetic variation to within populations (83–84%) rather than among populations or subspecies (15–16%), regardless of the type of analysis. The analysis indicated that most of the variation was due to differences between individuals within populations. This pattern aligns with previous findings for J. vulgaris and J. aquatica populations [17,92–94] where differences among individuals within populations also accounted for most of the variation. Further studies incorporating additional diversity metrics (e.g., Nei’s gene diversity, allelic richness) and confidence intervals could provide valuable information to better understand population-level variability and conservation potential. We consider this a promising direction for future research.

All analyses conducted using AFLP data (PCoA, UPGMA dendrogram, and NJ network) showed consistent results and divided all accessions into three main groups. One well-separated group consisted of J. vulgaris populations, which served as the out-group. The largest group encompassed individuals of J. erucifolia from Poland, forming a loose and mixed grouping without a clear subdivisions by subspecies or populations. The third group consisted of all foreign populations, which clustered together and notably included most individuals of the subsp. tenuifolia (reference group from Slovakia). Based on these results, we can assume that the populations from Hungary and the Netherlands (initially not assigned to a subspecies rank) also represent the tenuifolia subspecies.

Bayesian analyses using STRUCTURE grouped the accessions similarly to the previous analyses. Most Polish populations exhibited a high degree of admixture, with the exception of two populations (P and B) that stood out from the others and showed a low level of admixture. No significant differences were detected between populations tentatively classified as subspecies based on morphology, suggesting that their gene pools are not separate. Overall, population clustering revealed a consistent pattern: one group representing Polish J. erucifolia populations and another encompassing populations from the remaining regions. Moreover, unlike the foreign populations (from the Netherlands, Hungary, and Slovakia), most Polish J. erucifolia populations showed a noticeable overlap of gene pools. The STRUCTURE-defined clusters and admixture patterns suggest that Polish populations represent a genetically unified unit, despite morphological variation. This undermines the case for recognizing two subspecies.

The three-step STRUCTURE analysis was designed to address one of the main objectives of the study: clarifying the status of Polish populations, which until now had been assigned to subspecies solely on morphological grounds. Slovakian samples represented the only reference material with a confirmed taxonomic status (J. erucifolia subsp. tenuifolia). Therefore, analyzing Polish populations together with Slovakian material (Fig 4B) was a critical step for testing their genetic distinctiveness relative to a verified reference group. In contrast, material from Hungary and the Netherlands, although morphologically resembling subsp. tenuifolia, has not been formally assigned to subspecies and is referred to in our study only as J. erucifolia. For this reason, the intermediate analysis (Fig 4B) provided the necessary reference-based framework for interpreting the subsequent results when all foreign populations were included (Fig 4C).

Previous research [65] documented natural hybridization between J. vulgaris and J. erucifolia in Poland and the presence of hybrid populations. It also showed that introgression processes occur in both directions in the parental populations. The admixture of gene pools in both species is further confirmed by the current results. Both J. erucifolia populations show a slight admixture of the J. vulgaris gene pool, and several individuals from the J. vulgaris populations exhibited a slight admixture of the J. erucifolia gene pool. The most surprising result came from the Dutch population, which was morphologically classified as J. vulgaris. In previous analyses (PCoA, NJ), this population formed a subcluster (IB) related to the J. vulgaris group, but it occupied a position between this group and the J. erucifolia group formed by individuals from areas outside Poland. Half of the individuals in this population (classified as J. vulgaris) exhibited admixture, with both gene pools contributing to their genotypes, though the J. erucifolia cluster predominated. The remaining individuals were almost pure J. erucifolia. This might indicate that they are likely later-generation backcrosses containing a high proportion of J. erucifolia. In light of the above information, a more detailed and targeted analysis of hybridization—including the use of additional markers, quantification of introgression levels, and comparison of morphological traits between hybrids and non-hybrids would provide valuable insights. Moreover, such an integrated approach, combined with expanded geographic sampling, could significantly enhance our understanding of population dynamics and inform conservation strategies for J. erucifolia in Poland. Hybridization, especially under anthropogenic pressure, may play both adaptive and threatening roles for rare taxa. Our data are consistent with hybrid zones, asymmetrical gene flow, and potential for cryptic introgression.

All AFLP-based analyses mentioned above did not distinguish individuals belonging to particular subspecies within the Polish populations. Nevertheless, foreign populations assigned to subsp. tenuifolia have always formed a distinct, separate cluster. Therefore, this study challenges the status of subsp. tenuifolia in Poland.

According to some authors, the reliability of studies on complex genetic processes requires the collection of evidence from many sources [95–97]. For example, Boecklen and Howard [98] suggested that, in hybridization analyses, more than 70 diagnostic markers are needed to distinguish parent species from advanced backcrosses with sufficient certainty. On the other hand, highly backcrossed hybrids can be difficult to distinguish from their parents, as such individuals often cluster with pure individuals. Therefore, in addition to nuclear markers (AFLP), maternally inherited plastid markers were also analyzed using the PCR-RFLP method.

The highly conserved nature of the chloroplast genome is considered to be an ideal marker for population studies [99,100]. Although the rate of cpDNA evolution is slow, molecular analysis has shown that insertions, deletions, and point mutations cause detectable variation when using PCR-RFLP-based methods. In recent decades, plastid markers either species-specific or universal have been used to resolve phylogenetic relationships and species delimitations. The fact that they are maternally inherited makes them especially useful in research related to hybridization and introgression. They are also frequently used in analyses of representatives of the genus Senecio [101].

To characterize the haplotype diversity, its distribution in the studied populations and species, and to compare Poland with other parts of Europe, three fragments (CS, HK, and LF) were amplified using universal primer pairs developed by Demesure et al. [79] and Taberlet et al. [81], respectively. The resulting fragments were of the expected length, and no size differences were noted between the individuals in the studied populations. After digestion, the selected four fragment-enzyme combinations generated 12 polymorphic fragments, allowing the detection of 33 length mutations. The mutations detected in this study were mainly short-length variants, presumably indels. According to Clegg et al. [102]), the frequency of indels was higher than that of point mutations.

Analysis of the polymorphism of plastid DNA markers revealed 99 different haplotypes, the majority of which (80.8%) were unique. This result indicates significant diversity within individual populations, while also making it difficult to clearly interpret the similarities and differences between populations. Since each fragment-enzyme combination was sufficient to determine the haplotypes in the studied populations, the analysis focused on the most diverse combination, CS-Alw26I (supported by the results of the other combinations). Most populations were polymorphic, and in each fragment-enzyme combination, at least one haplotype was dominant and occurred at a higher frequency. These common haplotypes were found regardless of the region of origin of the individual populations or subspecies. However, no haplotype was common to all populations. Polish populations showed greater haplotype diversity compared to populations representing the tenuifolia subspecies. In conclusion, haplotypes did not exhibit a clear pattern of differentiation among the subspecies, and no clear regularity was observed in the geographic distribution of haplotypes.

In the case of the LF locus, the differences in the size of the fragments after digestion with restriction enzymes were so slight that it was uncertain whether they truly existed or were merely due to imperfections in the electrophoretic separation. To precisely determine the type of mutation within the theoretical variants, which may differ by only a few base pairs, sequencing was performed. Sequence comparison showed that the sequences of all analyzed individuals of J. erucifolia (from Poland, Slovakia, Hungary, and the Netherlands), regardless of subspecies, were identical. According to available data, chloroplast genomes of Asteraceae species do not show significant differences in size, gene content, or their arrangement [103]. However, this is largely not applicable to non-coding regions, where the differences can be significant. We demonstrated this in our previous publication [65], where sequence analysis of the LF fragment allowed the confirmation of hybrid presence in Jacobaea populations due to clear differences in the cpDNA sequences of the parent species. Furthermore, literature reports suggest that, for example, subspecies of Senecio vulgaris have different sequences of both cpDNA and rDNA [104]. In light of these data, the ITS sequences obtained in this study for reference individuals of the tenuifolia subspecies were also compared with those previously obtained for J. erucifolia (both subspecies; [65]). Again, no differences were found; the sequences were identical.

Finally, to resolve the ambiguous classification of individuals from population NP (identified as J. vulgaris in the field; however, STRUCTURE analysis suggested its affinity with J. erucifolia), LF fragments were sequenced and compared with those typical of both species (J. vulgaris and J. erucifolia). The sequences of some individuals were consistent with J. vulgaris, while two showed sequence differences typical of a hybrid origin. The AFLP results (see above) clearly indicated the probability that this population contained a mixture of J. erucifolia-like individuals, introgressed to different degrees.

The aim of our research was primarily to answer questions regarding both the condition of Polish populations of J. erucifolia in comparison with populations from other regions, as well as the taxonomic status of subspecies described solely based on morphological features.

Geographically and genetically distinct populations that exhibit clear phenotypic differences are assumed to be identified as subspecies [27,105] (see Introduction). Since intraspecific differentiation itself is often considered sufficient to designate a subspecies, there remains a significant open question regarding the patterns of variation that should be expected to confirm whether a given taxon truly merits subspecies rank.

In summary, detailed analysis of AFLP banding patterns allowed us to distinguish two distinct groups of individuals: one consisting of 10 Polish populations (both putative subspecies), and the other including the tenuifolia subspecies (Slovakia), as well as genetically and morphologically similar populations from Hungary and the Netherlands, currently also considered as the same subspecies. The Polish populations exhibited considerable intra-population diversity (a mixed group of both subspecies) and shared the same haplotypes. No diagnostic bands were found that would allow distinguishing the two subspecies within this group. Moreover, the lack of genetic divergence between these two subspecies was also confirmed by the results of studies based on nuclear and chloroplast DNA sequence analyses. In light of these results, we suggest that, at least in Poland, the subspecies status of J. erucifolia subsp. tenuifolia (genetically indistinct) is questionable.

The individuals constituting the second group of populations were clearly distinguished genetically from those from Poland. The populations from Hungary and the Netherlands, initially defined as J. erucifolia without indicating their subspecies affiliation, consistently grouped together with the subspecies tenuifolia from Slovakia. This grouping may confirm that they also represent the same subspecies. The fact that in all analyses based on AFLP markers, these populations were very clearly separated from the Polish representatives of J. erucifolia, proves their genetic distinctiveness. This, in addition to morphological differences, is one of the required conditions for recognizing them as a subspecies. Importantly, aside from a small number of fragments specific/unique to this subspecies (1.7%), no diagnostic products were found that would allow for distinguishing individuals forming both main groups (potential subspecies). Moreover, the sequencing results do not confirm the distinctness of these populations. When compared with the sequences obtained for representatives of the Polish populations, no differences were found in the cpDNA sequence, and in the case of individuals treated as reference, also in the rDNA sequence. So, the question arises: what are we dealing with?

Numerous studies have shown significant genetic differences, both at the sequence level and based on chloroplast and nuclear markers, between different subspecies [106–111]. According to Wang et al. [111], the degree of separation of the gene pools of subspecies may be related to the level of genetic diversity between them. The authors also observed that, although there was substantial genetic divergence among the subspecies studied, it varied from locus to locus across the genome. Furthermore, they found that the most information on subspecies delimitation came from a small part of the genome, while the rest was only slightly differentiated, indicating uneven gene flow and incomplete isolation. Since genetic distinctiveness has already been confirmed by AFLP analysis (foreign populations), the lack of clear differences within the analyzed sequences for subsp. tenuifolia does not necessarily undermine its taxonomic classification. It may, however, indicate the need for further analysis of other sequences.

Polish populations show significant phenotypic differentiation. Based on morphological features, two groups were distinguished, defined as two subspecies: erucifolia and tenuifolia (details above). However, analyses based on molecular markers did not support this. Although the main cluster formed by Polish populations was divided into smaller subclusters, both putative subspecies co-occurred in each of them, and the similarity of the habitats occupied seems to be the basis for their formation. The relationship with the environment was also indicated by earlier morphological studies. Examination of the geographic dependence of morphological differentiation in the taxa (using the same material) showed that population heterogeneity likely results from environmental factors [21]. Similarly, in molecular analyses, the Mantel test showed no significant relationship. Moreover, the comparison of Polish populations with populations representing the tenuifolia subspecies also excluded the occurrence of this subspecies in Poland. Therefore, the question arose: what could be the basis of this differentiation if it is not genetically determined? J. erucifolia, like other Jacobaea species, inhabits ruderal habitats, often altered by humans (urbanization), which certainly provokes adaptation processes that we can observe at various levels.

It is widely recognized that most organisms exhibit different phenotypes in response to different environmental factors. According to Sultan [46], each genotype has its own genetically determined level of phenotypic plasticity, with some traits being more conserved than others. Phenotypic plasticity, described as the ability of a single genotype to produce different phenotypes under variable environmental conditions, enables species persistence in novel or changing environments [46,112]. It is often linked to a mechanism that may facilitate invasions [47]. Unlike diagnostic traits for a given taxon, differences arising from phenotypic plasticity are not stable and disappear when plants are grown under the same conditions in common garden experiments [113]. In changing environments, phenotypic plasticity allows organisms to maintain their genotypes and continue to produce alternative phenotypes, in contrast to local adaptations that involve changes at the genetic level, creating long-lasting phenotypes and resulting in higher fitness. Local adaptation results in the formation of ecotypes, ecological races with distinct characteristics that provide them an environment-specific fitness advantage [114]. It typically requires changes at multiple loci (different alleles are favored in different environments), so the observed phenotypic differences between populations from different eco-regions are not accidental. They result from genetic differences arising from adaptation to the habitat [115]. Different ecotypes are often geographically close and interfertile, but there is no evidence of gene flow between populations of divergent ecotypes [116]. Ecotypes show strong differences in leaf morphology [117]. Some researchers perceive ecotypic differentiation as the first step towards the formation of new species [25,118,119]. However, we must remember that changes taking place in the environment not only result in speciation but also pose a threat to many species.

In our study, the populations of J. erucifolia from regions outside Poland (Slovakia, Hungary, and the Netherlands) are represented by the subspecies tenuifolia. Despite their clear genetic distinctiveness in relation to the Polish populations, some doubts arise from the fact that no diagnostic markers were found in their AFLP profiles. Such markers were also absent in the profiles of the Polish population group, which theoretically represents different subspecies. Therefore, it cannot be ruled out that we are dealing with the occurrence of ecotypes. Both groups seem to meet the criteria described above. Such varieties may occur within the same geographic region, where distinct habitats provide ecological niches. In the case of the Polish populations, the situation is more complex due to the internal morphological division. Therefore, we lean toward the interpretation that the morphological differences that led to the separation of the tenuifolia subspecies in Poland arose as a result of phenotypic plasticity. Both of these hypotheses require further research to confirm their validity.

In our analysis, we cannot overlook one of the most important causes of phenotypic differences: naturally occurring hybridization processes, which may go unnoticed or be incorrectly interpreted. This was pointed out by Abbott [48]. Naturally occurring hybrids are typically recorded based on morphology. There is a high probability that only the early generations of hybrids will be recognized as such, while later generations of backcrosses resembling one or both parents might be overlooked. Therefore, based solely on morphological analysis, hybrid zones that primarily include backcrosses and parental types will not be recognized in the wild. A good example of this is the Dutch population, whose individuals, designated as J. vulgaris, turned out to be a group of advanced backcrosses with a predominance of J. erucifolia admixture in our molecular studies. Furthermore, hybridization is believed to occur primarily between closely related, genetically similar species. Under normal circumstances, many of these species are isolated from one another by various barriers that prevent interbreeding and the formation of hybrid zones. However, natural or anthropogenic disturbances can disrupt these natural isolation mechanisms. According to literature data collected by Abbott [48], about one-third of hybrid zones occur in areas with disturbed environments.

Europe, like other continents, is influenced by global climate change, which, as we know, favors hybridization. Moreover, hybrids tend to adapt better to changing conditions. Among Jacobaea species, hybridization is common, and hybrids are not rare. Our previous research unequivocally confirmed the presence of hybrid populations between J. erucifolia and J. vulgaris [65], occurring in areas close to those currently studied. Polish populations are relatively small and scattered, and the species is considered endangered. Thus, the observation of mixed genotypes in almost all populations (as indicated by the STRUCTURE analysis) may raise concerns regarding biodiversity. Hybridization can threaten the existence of a rare species [120] by causing genetic assimilation of a rare taxon by a closely related and more common one. Arnold [121] defined this phenomenon as the loss of genotypes or phenotypes of a rare taxon due to asymmetric gene flow. It is worth emphasizing that our previous research showed a clear asymmetry toward J. erucifolia and a significant share of backcrosses in the group of analyzed hybrids. On the other hand, hybridization can prevent extinction by maintaining or increasing population size. This phenomenon is known as the “rescue effect” of hybridization [64,122] and is stronger when there is an asymmetry in the direction of introgression (backcrossing; [123]. Muhlfeld et al. [124] defined hybridization as an absorbing process that eventually leads to the point where all apparently phenotypically pure individuals (parental lines) may have hybrid origins, potentially resulting in the extinction of pure genomes. However, this process takes time and may remain unnoticed for an extended period. Moreover, according to O’Brien & Mayr [105], most subspecies are likely to be monophyletic; however, they may also arise from the hybridization of ancestral subspecies.

In the STRUCTURE analysis, the Polish populations formed a separate group. Apart from two populations (P and B), which were almost homogeneous, there was a clear overlap of two gene pools. One gene pool belonged to a quite homogeneous group corresponding to the subspecies tenuifolia. A slight admixture of the J. vulgaris gene pool was also present in most individuals from both groups comprising the J. erucifolia populations. It is known that genetically similar individuals tend to cluster together to form a species, subspecies, or population, but individuals can also form clusters if they are similarly mixed with other species.

To sum up, our research has shown that the morphological differences suggesting the presence of two subspecies in Poland are not determined by genetic differentiation and are likely the result of phenotypic plasticity and the tenuifolia subspecies does not occur in Poland. Our analyses revealed genetic differences (AFLP profiles) between Polish populations and those described as subspecies tenuifolia from neighboring countries. This differentiation, however, is not confirmed by the sequences analyzed (no differences). In combination with the lack of diagnostic markers, this is more consistent with their classification as ecotypes.

The presence of significant admixture in Polish J. erucifolia populations could be explained by alternative scenarios, such as, for instance, ancient hybridization events with other gene pools. If we assume that the traits considered diagnostic for both subspecies are largely the result of phenotypic plasticity, then we can question either their actual existence or the reliability of distinguishing them solely based on morphology. However, verification of any of these hypotheses requires further research, including a broader geographic sampling as well as the analysis of larger areas of the genome.

Supporting information

Acknowledgments

We sincerely thank everyone who contributed research material to this study. Special thanks to Marek Podsiedlik (Poznań) for collecting samples from all Polish populations, Király Gergely (University of Sopron) for providing samples from four populations in Hungary, Rense Haveman (Zetten, De Ronde & Haveman Research and Consultancy Agency for Geobotany and Landscape) for a population in the Netherlands, and Prof. Pavol Eliáš (Slovak University of Agriculture in Nitra) for herbarium specimens collected from six sites in Slovakia. Publication of this article was supported by the Ministry of Research and High Education Research Capacity Grants for Poznan University of Life Sciences, No. 506.181.11.00 and was financed by the Polish Minister of Science and Higher Education as part of the Strategy of the Poznan University of Life Sciences for 2024–2026 in the field of improving scientific research and development work in priority research areas.