Genetic Diversity in Invasive Populations of Lupinus polyphyllus Lindl. and Heracleum sosnowskyi Manden.

Simple Summary Aggressive-invasive species often interact with native ones, thus considerably changing the biological communities, with ecological, economic, and even social effects. It is a challenge to evaluate the direction and the rate of microevolution in native and introduced populations. One of the ways to do this is to estimate the genetic diversity. An introduction often imposes a reduction in population size (genetic drift, bottleneck, founder effect), which has the potential to reduce genetic diversity. However, after a lag, the genetic diversity can be restored due to repeated invasions (multiply introductions), hybridization between individuals from two different subspecies or species in the invaded ranges, as well as during rapid genetic changes under selection pressures in the novel environment. The purpose of this study was to determine the level of genetic diversity in successful invasive species Lupinus polyphyllus Lindl. and Heracleum sosnowskyi Manden. from Russia and Ukraine, and whether it may be associated with the strategy of their further expansion. Abstract In our study, two aggressive-invasive species, Lupinus polyphyllus Lindl. and Heracleum sosnowskyi Manden. from Russia and Ukraine, were investigated. The success in naturalization of both species is associated with human activities, since they have been used in agriculture and floriculture and both have qualities such as environmental tolerance, high fertility and phenotypic plasticity. The purpose of this study was to determine the level of genetic diversity of both species. For Heracleum sosnowskyi Manden., genetic diversity was compared in invasive and native populations. For Lupinus polyphyllus Lindl., the genetic diversity was compared in variety, feral and invasive populations. A genetic diversity was formulated using RAPD, ISSR and REMAP. For Heracleum sosnowskyi Manden., the average genetic diversity within the invasive population was similar (0.432), but slightly less (0.502) than within the native Caucasian population. This may suggest the successful naturalization of invaders and almost complete reconstruction of their genetic diversity. For Lupinus polyphyllus Lindl., the genetic diversity for the invasive population was the highest, with an average of 0.294, while for variety, it was the lowest, with an average of 0.194. The feral population had an intermediate place with an average of 0.248, which could suggest an increase of diversity in the process of naturalization.


Introduction
Invasive-alien species (IAS) can be considered the second most important threat to biodiversity, after habitat destruction [1]. When introduced into the ecosystem, IAS directly or indirectly affect human health [2]. Warnings from the scientific community for this serious environmental problem have led to the emergence of special international agreements and programs operating under the auspices of several trustworthy organisations. of the USSR. The consequence then was that hogweed began to spread especially actively. Thus, H. sosnowskyi is called "an indicator of mismanagement" [19]. H. sosnowskyi increased in its growth area by 10% annually, with its invaded ranges including the countries of the Baltic region (Denmark, Poland, Estonia, Lithuania, Latvia), Belarus, Ukraine, and Russia. In Russia, hogweed has spread widely in the European part of Russia, penetrating the Urals, the south of Western Siberia, and the Far East. It grows on the outskirts of roads, fields and forest belts, wastelands, forests, and floodplain meadows [20]. An active search for methods to control H. sosnowskyi began in the 2000s, with one of the first projects being the Giant Alien Project (2002)(2003)(2004)(2005), dedicated to three giant hogweeds (H. mantegazzianum Sommier & Levier, H. sosnowsky, and H. persicum Desf. ex Fisch.). More than 40 scientists from seven countries participated, proposing, methods to combat all three species [20,21].
Upon introduction of a new species in an invaded range, the level of genetic diversity within populations is often reduced [22]. However, after a lag the genetic diversity can be restored [23][24][25]. Therefore, by understanding the level of genetic diversity it should be possible to predict the further spread of invasion. One of the most accessible, fast, and inexpensive techniques for detecting genetic diversity at the DNA level is a variant of PCR, multiple arbitrary amplicon profiling (MAAP), which includes Random Amplified Polymorphic DNA (RAPD) [26], Inter Simple Sequence Repeats (ISSR) [27], Retrotransposon-Microsatellite Amplified Polymorphisms (REMAP) [28]. Thus, our objectives were to estimate the genetic diversity in populations of L. polyphyllus and H. sosnowskyi by means of RAPD, ISSR, and REMAP. The genetic diversity of invasive populations of H. sosnowskyi was compared with its native population. The genetic diversity of invasive population of L. polyphyllus was compared with its variety and feral population, as examples at the beginning of the invasive process.

Plant Material
Plant material used in this study included 84 L. polyphyllus and 99 H. sosnowskyi samples (Tables 1 and 2 Figures S1 and S2). Leaf material was collected in the summers of 2019 and 2020 from individual plants at a minimum of 10 m apart. Leaves were desiccated in silica gel until completely dried. 'Minaret' variety samples were grown in the greenhouse of K.A.Timiryazev Institute of Plant Physiology, Moscow. Feral population was obtained from a variety that had been running wild for about 5 years in the Main Botanical Garden. Feral population consisted of naturalized plants or established plants. Pyshek characterized such plants as self-replacing populations sustained for at least 10 years without direct intervention by people capable of independent growth [29]. They are already not variety, but are not yet invasive.

DNA Isolation
DNA was isolated according to Dellaporta et al. [30] with the procedure scaled down for 50 mg dry leaves. As a negative control, C1-, water was added to the lysis buffer instead of plant material.

DNA Amplification
Twenty-five microlitres of the reaction mixture contained 0.25 µM primer (Lytech, Moscow, Russia), 0.2 mM each dNTP, 2U of Taq DNA polymerase, standard 10× PCR buffer (700 mM Tris-Cl pH 8.6, 166 mM (NH 4 ) 2 SO 4 and 25 mM MgCl 2 ) (Silex, Moscow, Russia) and 30 ng of DNA. The reaction mixture was overlayed by 30 µL of mineral oil. Following optimization, the amplification conditions were: denaturation at 94 • C for 2 min, 5 cycles as follows: denaturation at 94 • C for 20 s, annealing at t • C for 10 c, elongation at 72 • C for 10 c; 35 cycles as follows: denaturation at 94 • C for 5 c, annealing at t • C for 5 c, elongation at 72 • C for 5 c; 1 cycle as follows: elongation for 2 min at 72 • C; t • C was 37 • C for RAPD primers, 37-55 • C for ISSR primers and 50-60 • C for REMAP primers. Amplifications were done in a thermocycler MC2 (DNA-Technology, Moscow, Russia). To control the purity of reagents, water was added to the reaction mixture instead of DNA (reactions C1-and C-). The experiments were done in triplicate.

Gel Electrophoresis of the Amplification Products
Fifteen microlitres of the reaction products were analysed in 2% agarose-TBE gels with EtBr. The DNA molecular size marker M 100 bp+2 Kb+3 Kb (12 fragments from 100 bp to 3000 bp, Sibenzyme, Novosibirsk, Russia) was used to measure the sizes of DNA fragments.

Quantitative Estimates of Genetic Diversity
To quantitatively estimate the genetic diversity, the data are presented as binary trait matrices, where the presence or absence of the same size PCR bands were assigned values of 1 or 0, respectively. The value of 1 was assigned only to bands of high intensity consistently detected in all experiments. To estimate the diversity, the binary data for each population were summed, and the most common values were used to build the matrices.
The binary trait matrices were used to derive the difference matrices, with the Nei and Li genetic diversity (Gd) calculated as follows: Gd(xy) = 1 − 2Nxy/(Nx + Ny), where Nx is the number of fragments present in profile x, but absent in profile y; Ny is the number of fragments present in profile y, but absent in profile x; Nxy is the number of fragments present in both profiles [31].
The matrices obtained were used to build phylogenetic trees. The trees were built by the Unweighted Pair Group Method with Arithmetic Mean (UPGMA) [32] using Treecon 1.3b software [33]. To evaluate the confidence intervals of the trees, the bootstrap method with 100 samples was used [34].

Results
All primers yielded PCR bands amplified from plant genomic DNA. A total of 39 RAPD and 17 ISSR primers were used. Depending on the primer, the number of Biology 2021, 10, 1094 5 of 16 amplified fragments varied from 3 to 16, with their sizes varying from 200 to 2000 bp; the optimal annealing temperature was determined experimentally for each primer. The optimal primer pairs were designed for REMAP. Five REMAP pairs gave sharp and consistent profiles for lupin and seven REMAP pairs for hogweed (Supplementary Table S1). The genetic diversity between populations and the 'Minaret' variety varied from 0.196 to 0.341 and was 0.260, on average (Table 3).  Table S2). Depending on the method (RAPD, ISSR, or REMAP), the mean values of genetic diversity varied, but the general trend was the same: the invasive population had The genetic diversity between populations and the 'Minaret' variety varied from 0.196 to 0.341 and was 0.260, on average (Table 3). Additionally, the genetic diversity within invasive and feral populations and the 'Minaret' variety was calculated. To evaluate the genetic diversity within invasive population, the populations from Moscow, Kaluga, Smolensk, and Kostroma regions were pooled (34 samples). The amplified fragment profiles from 34 plant samples were presented as binary matrices and used to determine genetic diversity values (Supplementary Table S2). Depending on the method (RAPD, ISSR, or REMAP), the mean values of genetic diversity varied, but the general trend was the same: the invasive population had the highest values, and the 'Minaret' variety had the lowest values. With all the methods combined, the mean value of genetic diversity was 0.294 for the invasive population and 0.248 for the feral MBG population, which was 1.2-times lower than that for the invasive population. For the 'Minaret' variety, genetic diversity was 0.194, which was 1.5 times lower than that for the invasive population (Table 4). Notably, genetic diversity among populations (0.260) was less than the genetic diversity within invasive population (0.294). Table 4. The mean values of genetic diversity (GD) for L. polyphyllus invasive and feral populations and the 'Minaret' variety obtained by RAPD, ISSR, REMAP and the combination thereof; (RATIO), the ratio of the genetic diversity of the invasive population over the genetic diversity of the given population. Phylogenetic trees were built by using individual methods (RAPD, ISSR, REMAP) and all methods combined ( Figure 2).  Phylogenetic trees were built by using individual methods (RAPD, ISSR, REMAP) and all methods combined ( Figure 2).  The genetic diversity values within the feral MBG population (21 samples) and the 'Minaret' variety (29 samples) were calculated similarly (Supplementary Tables S3 and  S4, respectively), and phylogenetic trees were built (data not shown).

Molecular Analyses of H. sosnowskyi
In hogweed, in contrast to lupin, only a few primers amplified monomorphic fragments typical for all plants and some populations (RAPD: QR-1-900 bp, QR-2-590 and 800 bp; ISSR: MS1-1500 bp; REMAP: MS6 + Thv19-580 bp). Interestingly, with primers MS6 + Thv19, a 440 bp fragment was detected, which is characteristic of plants growing in the centre of Zhytomir but absent in plants of the industrial zone of Zhytomyr ( Figure  4).  (Supplementary Tables S3 and S4, respectively), and phylogenetic trees were built (data not shown).

Molecular Analyses of H. sosnowskyi
In hogweed, in contrast to lupin, only a few primers amplified monomorphic fragments typical for all plants and some populations (RAPD: QR-1-900 bp, QR-2-590 and 800 bp; ISSR: MS1-1500 bp; REMAP: MS6 + Thv19-580 bp). Interestingly, with primers MS6 + Thv19, a 440 bp fragment was detected, which is characteristic of plants growing in the centre of Zhytomir but absent in plants of the industrial zone of Zhytomyr ( Figure 4).  The populations from industrial zones and the centre of Zhitomir were pooled (33 samples), and the genetic diversity between the populations of hogweed was significant, varying from 0.438 to 0.614, and was 0.539, on average (Table 5).  The populations from industrial zones and the centre of Zhitomir were pooled (33 samples), and the genetic diversity between the populations of hogweed was significant, varying from 0.438 to 0.614, and was 0.539, on average (Table 5). The greatest genetic diversity was between populations from Moscow and Caucasus (0.614). This could be explained by the significant remoteness of the Moscow population and its evolution in the process of adaptation to a new habitat. The least genetic diversity was found between the Zhitomir and Peskovka populations (0.438). This could be explained by anthropogenic dispersal from one region to another. This is hardly a possible result of gene flow, since the distance between the Zhitomir and Peskovka is about 100 km (Supplementary Figure S2 There were no significant differences in genetic diversity in populations from different regions of Zhytomyr. The genetic diversity within the native Caucasian population was compared with that within the invasive population but was not significantly higher (Table 6). Unlike lupin, the genetic diversity between hogweed populations was greater than the genetic diversity within populations.  Phylogenetic trees were built using individual methods (RAPD, ISSR, REMAP) and all methods combined ( Figure 5).  Regardless of the method used, the Zhytomyr and Peskovka populations clustered together with high bootstrap values, indicating their close relationship ( Figure 5). Hogweed from Zhitomir may have been introduced to Peskovka or vice versa. ISSR and REMAP clustered populations from Kiev, Zhitomir, and Peskovka that corresponded to their close geographical locations (Figure 5b,c). The use of RAPD combined Moscow and Kiev populations, but the dendrogram obtained by the combination of RAPD, ISSR, and REMAP revealed only one cluster with Zhitomir and Peskovka, although the Kiev population was close to them (Figure 5d). The Caucasian population was separated in all the dendrograms, suggesting a significant divergence between the native and invasive populations ( Figure 5).

Discussion of the Genetic Diversity within Populations of Lupin
RAPD, ISSR, and REMAP techniques were used to study the genetic diversity of various populations of L. polyphyllus. All techniques gave reproducible results with an adequate number of polymorphic and monomorphic fragments necessary for estimating the genetic diversity both within and between populations. As there are few investigations of L. polyphyllus by means of PCR markers, especially where genetic diversity was calculated [35,36], we considered studies of other species of Lupinus.
Qiu et al. [37] used short eight-nucleotide primers for studying L. albus L. Wolko et al. [38] identified polymorphisms using RAPD among four cultivars of L. albus, seven cultivars of L. angustifolius L., and twelve cultivars of L. luteus. Gilbert et al. [39] used ISSR to reveal genetic variability within and between L. albus accessions from a collection of lupin germplasm. Yorgancilar et al. [40] successfully used RAPD and ISSR to determine the genetic relationships among 20 Old World lupin genotypes of three lupin species (L. albus, L. angustifolius, and L. luteus). Using ISSR, Artyukhova [41] developed a technique for the identification of varieties of three lupin species: L. angustifolius, L. hiteus, and L. albus. Clements et al. [42] investigated 10 L. angustifolius genotypes using ISSR markers. Similarly, Guilengue et al. [43] analysed L. mutabilis using ISSR. Khapilina et al. [44] succeeded in using REMAP and other retrotransposon analysis methods to estimate the genetic diversity of L. angustifolius.
The reliability of ISSR, REMAP and, especially, RAPD is supported by the results obtained. Firstly, three phylogenetic trees of the analysed populations built upon RAPD, ISSR, and REMAP data showed the 'Minaret' variety is considerably divergent from the other populations (presumably due to its independent origin) ( Figure 2). Secondly, the tree of the invasive population built on the combined data had four clusters that almost coincided with the locations of the respective populations, namely Moscow, Kaluga, Smolensk, and Kostroma regions (Figure 3). Thirdly, the genetic diversity values obtained by RAPD, ISSR, and REMAP for each population tightly correlated ( Table 4).
The genetic diversity for the invasive population was the highest, and varied from 0.045 to 0.411 (Supplementary Table S2), with an average of 0.294 ( Table 4). The genetic diversity among populations varied from 0.196 to 0.341 (Table 3) and was 0.260, on average, which was less than the genetic diversity within invasive population. Similar results have been previously published for other species of lupin. Vysniauskiene et al. [35] analysed 10 invasive populations of L. polypyllus in Lithuania with RAPD and found that Nei's average genetic distance between populations was 0.148 ± 0.021, which was also less than the average genetic distance between individuals in the forest (0.290 ± 0.062) and field populations (0.229 ± 0.037). The authors observed significant genetic differentiation among populations in genetically heterogeneous seed stock, low gene flow between populations, and, possibly, in local adaptations [35]. Li et. al. [36] investigated 51 L. polypyllus populations in Finland using 13 polymorphic microsatellite loci. The authors found that the genetic variation among populations was 0.05; it was significantly higher (0.25) within populations. Pairwise F ST values among populations ranged from 0.02 to 0.25 and the global F ST value was 0.19, suggesting moderate levels of genetic differentiation. The authors suggested that it was the result of human-mediated dispersal with multiple introductions from different sources rather than the natural spread of L. polyphyllus from a single or few sources in Finland [36]. Oumer et al. [45] used ISSR to study four L. albus populations from two zones of Ethiopia and found that the genetic diversity was 0.223 for Merawi, 0.198 for Addis Kidam, 0.189 for Sekela, and 0.167 for Wembera: larger genetic diversity was found within rather than between populations. The authors supposed that high genetic diversity between populations might be the result of moderate gene flow and within the populations could be caused by the presence of preferential or diverse adaptive genes [45]. Similar data on the level of genetic distances were described by Mahfouze et al. [46], who estimated the genetic similarity among seven genotypes of L. albus using RAPD and ISSR, where the Nei genetic similarity index ranged from 0.74 to 0.88 (diversity from 0.12 to 0.26) [46].
In our study, the high level of genetic diversity within populations (among individual plants) was most likely associated with genetically heterogeneous seed stock and crosspollination. The lower level of genetic diversity between populations was hardly a possible result of gene flow, since the distance between the Smolensk and Kostroma regions was about 700 km (Supplementary Figure S1), but more likely the result of seed source. Since lupin was cultivated for green mass or as an intermediate crop, the same standardised seed stock approved for cultivation was imported to different regions and could be the reason for the clustering of Smolensk and Kostroma populations on the dendrogram (Figure 2d).
There are studies where the values of genetic diversity within species are much higher than ours. Al Rawashdeh et al. [47] studied L. pilosus using RAPD and found its genetic similarity to range from 0.02 to 0.450 (diversity 0.550-0.980). The low similarity could have been due to a high, long diverging process in non-coding regions [47]. There are studies where the values of genetic diversity within species are more variable. Sbabou et al. [48] revealed the genetic similarity between and within Moroccan germplasm (L. albus, L. angustifolius, L. cosentinii, and L. luteus) by means of AFLP and ISSR markers. Similarity values were 0.70-0.82 (diversity 0.18-0.30) among accessions of L. albus, 0.24-0.61 (diversity 0.39-0.76) among accessions of L. cosentinii, and 0.68-0.88 (diversity 0.12-0.32) among accessions of L. luteus [48]. There are also studies where the values of genetic diversity within species are very low. Talhinhas et al. [49] used RAPD, ISSR, and AFLP to detect interspecific and intraspecific polymorphism in Lupinus spp. The intraspecific polymorphisms were revealed by AFLPs with similarity values of 0.908-0.955 (diversity 0.045-0.092) among L. albus accessions, 0.913-0.954 (0.046-0.087) in L. angustifolius, 0.920-0.988 (0.012-0.080) in L. hispanicus, 0.867-0.940 (0.060-0.033) in L. luteus, and 0.957-1.000 (0.043-0.000) in L. mutabilis [49]. Unfortunately, it was not possible to compare the genetic diversity of the invasive population in our study with the native population of L. polypyllus due to the lack of samples. Thus, it is unclear whether the invasive population has recovered after invasion and adaptation, or whether its genetic diversity is increasing. As far as we are aware, there are no published data on the genetic diversity or similarity of the native population of L. polypyllus.
In our research, in addition to the invasive and feral populations, the 'Minaret' variety was analysed. That this variety is not the source of the feral population is confirmed by all the dendrograms (Figure 2; unfortunately, information about the original variety has been lost). Although 'Minaret' cannot be a negative control for this feral population, low genetic diversity was expected as nearly all varieties have specific genotypes with selected characteristics which reduces their genetic diversity. Indeed, 'Minaret' had the lowest genetic diversity, varying from 0.099 to 0.288 (Supplementary Table S4), with 0.194, on average ( Table 4).
The genetic diversity of the feral population (MBG) varied from 0.130 to 0.363 (Supplementary Table S3), with an average of 0.248 (Table 4), and was greater when compared to 'Minaret', but lower when compared to the invasive population. We believe that genetic diversity could have increased in the process of naturalization. The feral population, represented by the escaped variety, could be regarded as the beginning of the invasion. Additionally, the blue flowers characteristic for the wild species began to prevail. This may be due to gene flow by cross-pollination with other lupin plants growing nearby.
Tkacheva [50] compared the biomorphological traits of feral (5 years of naturalisation) and Smolensk populations (30 and 40 years of naturalisation). The author found that the number of leaves and lateral flowering shoots per plant increased with a longer naturalisation period, but the number of seeds and plant height did not differ significantly. Although Ramula and Kalske [51] revealed that L. polyphyllus from the introduced populations were larger in size, they flowered less frequently and with fewer flowering shoots than plants from the native populations. Tkacheva [50] also studied the density of populations growing in forests and fields (Moscow and Smolensk regions). No significant differences in both populations were noted, indicating the ability of lupin to occupy biotopes with different levels of illumination. The author supposed that the success of L. polyphyllus as an invasive species could be explained by the increased competitive ability (EICA) hypothesis, which proposes that a larger size is selected due to intense intraspecific competition or reduced herbivore pressure in a new area [52].
The intermediate level of genetic diversity in the lupin feral population and high level in the invasive population discovered in our research, together with the other data revealed for the same populations (the increase of some biomorphological traits, and the ability to occupy new biotopes [50]) suggests the further expansion of L. polyphyllus invasion.

Discussion of the Genetic Diversity within Populations of Hogweed
Unfortunately, as with L. polyphyllus, there are few published investigations of H. sosnowskyi by means of PCR markers, especially where the values of genetic diversity have been calculated [53]. Therefore, again we analysed studies of other Heracleum spp. A study of H. sosnowskyi was performed previously using RAPD [54,55]. Strygina et al. [56] analysed the genus Heracleum using RAPD and ITS analysis, while noting its high genetic and morphological heterogeneity.
We compared published data with our study, where the genetic diversity among the populations varied from 0.438 to 0.614 with an average of 0.539 (Table 5). Niinikoski and Korpelainen [57] investigated H. mantegazzianum from eight populations in Finland using microsatellite markers. The authors found that the range of pairwise F ST values between populations was high (0.052-0.797), and the mean F ST among all populations was 0.498, showing a significant difference. The authors observed differentiation between populations, resulting from multiple introductions combined with limited gene flow [57].
It is difficult to say why the average genetic diversity for lupin (0.260) was two times lower than for hogweed (0.539). Both species are diploid, L. polyphyllus with 2n = 48 [58] and Heracleum spp. with 2n = 22 [59]. It is probable that the hogweed genome contains more non-coding repeating regions, where as a rule, MAAP primers are annealing.
Unlike lupin, the genetic diversity among hogweed populations was greater (0.539) (Table 5) than that within the populations (0.398-0.502) ( Table 6). This greater diversity might be explained by the large spread of values in genetic diversity within the population; for example, for the Caucasus from 0.244 to 0.739 (Supplementary Table S10), while the spread of values between populations was not so large 0.438 to 0.614 (Table 5). The lower genetic diversity within the populations also could be explained by the ability of H. sosnowskyi to self-pollinate. Walker et al. [60] investigated 13 H. mantegazzianum populations in northeast England using four nuclear microsatellite and one plastid marker. Similar to our research, the genetic differentiation between populations from different river catchments was higher (F ST mean pairwise 0.28 ± 0.1) than that within populations from the same area (F ST mean pairwise 0.11 ± 0.08). The overall F ST mean pairwise was 0.24 ± 0.13. The authors suggested that the high genetic differentiation between populations may indicate a large initial founder population or multiple introductions. Within population, seeds of this species are dispersed in water, resulting in a relatively small dispersal area and the relatively small dispersal range of likely pollinators [60].
The genetic diversity within invasive populations varied from 0.398 to 0.457 with an average of 0.432 (Table 6); diversity was similar, but slightly less within the native Caucasian population with an average of 0.502 (Table 6). This suggests the successful naturalisation of invaders and reconstruction of their genetic diversity, probably by means of hybridisation and rapid evolution. In our study, similar genetic diversity values for invasive and native populations were observed. However, according to the literature, different cases have been described. Invasive populations may display higher genetic diversity than native populations and vice versa. Jahodová et al. [53] analysed three invasive Heracleum species using AFLPs. The genetic similarity (Pairwise Dice's similarity coefficient) in H. sosnowskyi among all populations was 0.891, on average (diversity 0.109); within Europe, it was 0.901, on average (diversity 0.099), and within the native range, it was 0.908 (diversity 0.092). The authors noted that more within-taxon variation was detected in the invaded range (Europe) than in the region of the native distribution, probably driven by hybridisation and inbreeding. The results also indicate that multiple introductions, rapid evolution and drift are likely to have occurred [53]. Henry et al. [61] investigated 49 H. mantegazzianum populations from the western Swiss Alps and 11 Caucasian populations using eight nuclear microsatellite loci together with plastid DNA markers and sequences. F ST mean pairwise genetic differentiation for nuclear microsatellite loci between invasive populations varied from 0.180 to 0.396 with an average of 0.315 and between native populations from 0.092 to 0.202 with an average of 0.162. The authors indicated that the high population differentiation observed in the invasive range compared to the native range could have been generated by sequential founder events. The species may have been introduced multiple times, possibly from disparate source populations [61]. Rijal et al. [62] studied 50 native and introduced populations of H. persicum with 27 microsatellite markers and estimated that F ST mean pairwise genetic differentiation (averaged over the population) was lowest between England and Sweden (0.267 ± 0.006), and highest between Norway and Denmark (0.552 ± 0.005). The mean F ST was lower in the native (0.25) compared to the introduced range (0.30), but the difference was marginal and nonsignificant [62].
Analysing other data such as the percentage of polymorphic loci, observed and expected heterozygosity, allelic richness and inbreeding coefficient, Henry et al. [61] and Rijal et al. [62] noted the loss of genetic diversity in the introduced ranges compared to native ranges. In our investigation, MAAP markers detected differences throughout the entire genome. As a rule, these markers are dominant and can detect only two alleles (present or absent fragment). At the same time, nuclear microsatellite markers, plastid, ribosomal, and mitochondrial DNA markers are locus specific and can detect several alleles, since their PCR fragments are sequenced. Therefore, the definition of 'genetic diversity' used for these techniques could be different. For our purposes, 'genetic differentiation' could be similar to 'genetic diversity. ' According to the literature, invasive populations often have lower genetic diversity compared to the native population due to bottlenecking, genetic drift during colonisation, and founder effects [63]. However, in the territory of the former USSR, H. sosnowskyi was widely introduced at the state level. The Institute of Biology of the Komi Republic received H. sosnowskyi seeds from many botanical gardens and from native ranges (Nalchik). In the decade 1940-1950, accumulation of seeds and cultivation of this plant was of great importance in five main centres: Murmansk, Moscow, and Leningrad regions, the Republic of Komi, and Kabardino-Balkaria. For example, 140 hectares of H. sosnowskyi were planted in the Rzhevsky region, and 7600 kg of its seeds were sent to 60 addresses [64].
A rapid local adaptation and evolution in invasive ranges is characteristic of H. Sosnowskyi. The sizes of all parts of invasive plants of H. sosnowsky are larger than plants of native populations: the height of the stem (1-1.5 m in the Caucasus and 2-4 m in Europe) and the size of the seeds (in Caucasian species, 9 mm long and 6 mm wide; in Europe, 15 mm long and 8 mm wide). In addition, Caucasian plants grow in forests and meadows, whereas plants of the invaded range are photophilous and mainly found in open habitats [65].
A high level of genetic diversity in hogweed invasive populations, similar to the native Caucasion population, was found in our research, suggesting reconstruction of their genetic diversity. This high level of genetic diversity, together with data suggesting the increase of some biomorphological traits, the ability to occupy new biotopes [65], and to hybridize with the native H. sibiricum and H. spondyllium and other introduced hogweeds [66] reported in the literature, could drive the further expansion of H. sosnowskyi invasion.

Conclusions
Both L. polyphyllus and H. sosnowskyi have been used in agriculture (lupin since 1922 [11], hogweed since 1947 [16]). Many seeds of both crops were imported from different ranges, including their natural habitats, so that multiple introductions resulted in a large number of founders. It is probable that the genetic diversity did not initially decrease. Moreover, the genetic diversity could have increased as a result of genetic exchanges between populations originating from different regions. Therefore, in this case we cannot speak to the genetic paradox of invasion, because genetic variation in populations is not lower than the native source population and has not passed through a bottleneck [67]. Where varieties were bred, genetic diversity was likely reduced, but after naturalization to a new habitat, diversity in feral populations began to increase again. Therefore, our results suggest that anthropogenic dispersal may be a major factor contributing to the successful invasion of L. polyphyllus and H. sosnowskyi, together with their individual characteristics, which include environmental tolerance, high fertility, phenotypic plasticity, cold resistance, and the ability to occupy new biotopes. Both species have turned into transformers and supplanted the native species, significantly reducing the diversity of biogeocenoses [65,68]. Thus, these species require further study to successfully control their invasion.
Supplementary Materials: The following materials are available online at https://www.mdpi.com/ article/10.3390/biology10111094/s1, Figure S1: Map of the lupin sampled populations, Figure S2: Map of the hogweed sampled populations, Table S1: The primers used in the experiments, Table S2: Genetic diversity within lupin invasive population obtained by combination of RAPD, ISSR and REMAP data, Table S3: Genetic diversity within lupin feral MBG population obtained by combination of RAPD, ISSR and REMAP data, Table S4: Genetic diversity within the 'Minaret' variety obtained by combination of RAPD, ISSR and REMAP data, Table S5: Genetic diversity within hogweed Moscow invasive population obtained by combination of RAPD, ISSR and REMAP data, Table S6: Genetic diversity within hogweed Peskovka invasive population obtained by combination of RAPD, ISSR