Intra- and interspecific variability of Mentha arvensis L. and M. canadensis L

The identification of plants of the genus Mentha is often difficult due to significant intraspecific polymorphism, intense interspecific hybridization, and ploidy changes. An attempt was made to apply an integrated approach to the study of different parameters of two species: Mentha arvensis L. and M. canadensis L. Eight geographically dispersed populations of Mentha in different regions (European Russia, Khakassia and Far East, Western Ukraine, and Indochina) were studied. Diagnostic morphological characters and compositions of essential oil components were examined, and DNA was analyzed with ISSR markers. The data obtained were statistically processed by cluster, principal component, and principal coordinate analyses. The European and Asian groups of samples were clearly distinguished by the analysis of quantitative parameters of the calyx and leaves, but different methods of data processing produced different results in determining the belonging of the Far Eastern plants to a particular group. Therefore, their taxonomic positions can hardly be determined on morphological grounds. According to the composition of essential oil and ISSR fragments, a group of the genetically, morphologically, and phytochemically closest plants was identified, which included representatives of the populations of the Moscow oblast, Vladimir oblast, Kaluga oblast, the Komi Republic, and Khakassia. All these plants belonged to M. arvensis. Plants collected in the natural flora of the Russian Far East showed a greater resemblance in essential oil composition and ISSR markers to the European group of M. arvensis than to plants from Indochina, which, according to the data obtained, belonged to M. сanadensis. It was shown that a comprehensive study of plant morphological characters, the compositions of essential oil, and ISSR fragments allows one to clarify the species identity and to assess their polymorphism and the degree of kinship between populations. A certain correlation between the data of molecular analysis and the composition of essential oil and, to a lesser extent, their correlation with morphological characters of plants was revealed.


Introduction
Mentha L. species have a significant potential due to the content of essential oil in the aerial parts of plants. Mint is used in medicine as an antispasmodic, sedative, gastric, and choleretic herb. In addition, it is employed in the food and perfume industries. In southern regions of Asia, it is cultivated to obtain technical essential oil. Special treatment of this oil yields menthol, which is a component of many important medicines.
Various sources report from 18 Mentha L. species and 11 hybrids (Tucker, Naczi, 2007) to 27 species and 15 hybrids (WCSP, 2019). There are different opinions as to the volumes and boundaries of morphologically similar but polymorphic species. A distinctive feature of this genus is a significant polymorphism within its species and, for all that, the simila rity of a number of morphological features among the species. In this regard, it is advisable to use the composition of the essential oil as an additional feature for the characterization of species. The composition of essential oil in all Mentha species is genetically determined. Under favorable growing conditions, component synthesis proceeds to the full set of terpenoid compounds characteristic of the species, but un favorable conditions result in arrest of the synthesis of these substances at early stages, which leads to the appearance of simpler essential oil components (Gouyon et al., 1986). It raises difficulties in the taxonomical discrimination of related species, such as Mentha arvensis L. and M. сanadensis L.
Mentha arvensis grows in Europe and Asia. This species is common in European Russia, Ciscaucasia, and Siberia (Gubanov et al., 2004). The geographic range of M. сanadensis in Russia includes Siberia and the Far East (Doron'kin, 1997); outside Russia, North America, and East Asia (Tucker, Naczi, 2007).
Mentha arvensis L. (field mint) is a widespread and very variable morphological species, and the habit of plants can vary significantly depending on the growing conditions. In shady damp places, tall plants with ascending or recumbent stems, large green or light green leaves, and pale lilac, some times almost white flowers are formed. Plants growing in drier sites are undersized, with erect stem, close internodes, small leaves with redpurple anthocyanin shade, and bright lilac color flowers.
Mentha сanadensis L. (Canadian mint) was described by C. Linnaeus (Linne, 1753). It is related to M. arvensis and difficult to identify. M. сanadensis differs from M. arvensis in having a higher, unbranched, densely pubescent stem; one half as wide, sharp, and deeply serrated leaves; and a specific calyx shape. Different morphological criteria are used for the discrimination of M. arvensis and M. сanadensis. These spe cies are proposed to be divided according to the shape of the calyx and the structure of the calyx teeth (Doro'nkin, 1997). M. arvensis calyces are bellshaped; the teeth are short and widetriangular, whereas M. canadensis calyces are tubular with long pointed teeth. Other researchers identify the two species by leaf form and flavor; however, these traits do not provide a reliable criterion (Tucker, Naczi, 2007).
Such morphological features as the type and degree of branching, the shape and size of the leaf, and shoot pubes cence cause difficulties in the identification of these two species. These traits are so variable within a species that they can overlap between species. All these features vary greatly depending on growing conditions (especially in M. arvensis); nevertheless, there are important in identifying species.
Researchers often disagree in recognizing these two spe cies. In particular, M. сanadensis was considered a form or subspecies of M. arvensis. Holmes (1882)  M. canadensis is probably an amphidiploid resulting from interspecific hybridization of M. arvensis and M. longifolia (Tucker, Chambers, 2002). There is evidence from chloroplast DNA sequences that in this interspecific hybridization process M. arvensis may have been the maternal parent (the source of female gametes) for the resulting species, M. сanadensis (Bunsawat et al., 2004). A comprehensive study involving flow cytometry and ISSRanalysis suggested that M. сanadensis was an allopolyploid resulting from cytomixis between repre sentatives of parental species with different ploidies (Jedrzej czyk, Rewers, 2018).
Our work concerns the intra and interspecific variability and differentiation of M. arvensis and M. сanadensis of several populations distant from each other. Previously, we studied polymorphism in M. arvensis populations on the base of mor phological and molecular data (Shelepova et al., 2016b) and attempted to differentiate these species using morphological and phytochemical analysis (Shelepova et al., 2016a). How ever, the resulting notion was still incomplete. A number of investigators successfully undertook a combined assessment of plants by morphological characteristics, essential oil com position, and ISSRanalysis data to study interspecies and interpopulation variation in the genus Mentha (Hua et al., 2011;Rodrigues et al., 2013;Shelepova et al., 2017). There fore, to accomplish our task; we comprehensively examined morphological features, essential oil compositions, and ISSR markers of DNA in two mint species.

Materials and methods
Sampling in populations. Plants of M. arvensis and M. сanadensis to be examined belonged to eight populations. They were collected from four regions distant from each other ( Table 1).
Five to eight shoots from typical plants collected at 10-200m intervals depending on the population size in the nature and 1-3 shoots of plants cultivated in experimental plots were taken for the study of morphological features and ISSR analysis of DNA. In our preliminary study, samples from natural populations demonstrated a wide variety of amplified fragments (significant diversity of genotypes within a popula tion), and the plants from the collection did not differ in the spectrum of ISSR fragments (belonged to a single clone). Mentha spicata L. was used as an external standard. One to five plant samples were taken to study the composition of es sential oil in populations concurrently with the collection of herbarium images (in the phase of mass blossoming) from a site of 0.2 m 2 . Herbarium specimens are stored at the Labora tory of Plant Physiology and Immunity and in the Herbarium of GBS RAS (MHA).
Study of morphological features. The collected plants were identified to the species level. Most of the samples were attributed to M. arvensis, but plants from the natural flora of the Russian Far East and collection samples from the Indochina flora were previously identified as M. сanadensis. The following quantitative indices were studied: calyx length, calyx tooth length, calyx tooth width at the base, the tooth length : width ratio, the calyx length : tooth length ratio, the calyx tube length : tooth length ratio; the length and maximum width of the leaf, the distance from the basis to the maximum width of leaf; the density of secretory granules on the lower and upper surfaces of the leaf per 1 cm 2 , the ratio between the numbers of secretory granules on the lower and upper surfaces of the leaf. The structure of the calyx was examined with a light stereomicroscope at 400× magnification. Images for comparative analysis were taken with a Lumenera Infinity 2 video camera and processed with the Infinity Analyses 5.0.2 program. Secretory granules were counted as in (Shelepova et al., 2012).
Methods of studying the essential oil composition. Es sential oil was isolated from an average sample of the above ground mass (a mixture of inflorescences and leaves) of plants. The oil was obtained by hydrodistillation of crushed airdry material (Ginsberg, 1932). Plant essential oil consists of a mix ture of natural compounds, mainly isoprenoids (monoterpenes, diterpenes, hemiterpenes, sesquiterpenes, and their oxides). The most detailed report on their classification, biosynthesis, and detailed study of each class of compounds was provided by V.V. Plemenkov (2007). The qualitative composition of the oil was determined by gas chromatography in the "Biotechno logy" Shared Access Center, RAS (RFMEFI62114X0002) in a Shimadzu GS 2010 gas chromatograph with a GCMsQP 2010 mass detector as previously described in (Shelepova et al., 2017).
DNA extraction and PCR. DNA was isolated from dry herbarium specimen leaves by the CTAB method (Doyle J.J., Doyle J.L., 1987). DNA polymorphism was analyzed by the ISSR approach. Primers for PCR analysis were synthesized and purified by PAAG by Syntol Company (Moscow, Russia). Eight primers were chosen after pilot tests (Table 2). Two series of PCR were carried out for 46 and 42 samples, and some of the samples were retested by PCR. ISSRPCR and the separation of PCR products were carried out by previously reported methods (Shelepova et al., 2016b(Shelepova et al., , 2017. Analysis of molecular data and statistical data proces sing. The band profiles of the ISSR fragments were compared visually and with the CrossChecker program. Only bright and distinct fragments were taken into consideration, and unclear bands were discarded. Each band in an electrophoretic gel was considered a countable character and used as a binary code in the matrix of the presence/absence of fragments. Then the results were analyzed in the PAST program using cluster analysis, principal coordinate method, and principal component method (Hammer et al., 2001). Bootstrap analysis was carried out with 1000 replicas to assess the stability of the dendrograms obtained.

Results
The analysis of quantitative morphological features by the principal component method shows that the samples can be di vided into two main groups (Fig. 1, a). The first group includes all samples from European Russia, the Republic of Komi, Khakassia, the Russian Far East, and Ukraine and one plant from Indochina (Ind2). Plants from the Far East are grouped together to form part of a large group, and other samples form a mixed cloud. An individual cloud is formed by plants from Indochina, which were identified as M. сanadensis. Two main clusters with somewhat different compositions of samples therein were obtained when processing the same data by cluster analysis using the Gower distance ( Fig. 1, b). The first cluster includes all plants from the Russian Far East and Indochina, and the second, plants from Europe and Khakassia.
Essential oil composition can be an independent criterion for the identification of Mentha plants. About 47 components were found in Mentha essential oil. All components constitut ing more than 0.1 % of the total amount were easily identi fied by retention time and mass spectra. The proportions of individual components in the essential oil vary significantly among plants from different regions.
The distribution of samples according to essential oil composition proved to be consistent with morphological data. However, the plant group from the Russian Far East shows a more distinct and stable position among others regardless of the statistical methods applied (Fig. 2). When using the principal components method (see Fig. 2, a), the samples are divided into three main groups: 1, M. spicata (the most abun dant component is menthone); 2, IndoAsian plants (menthol); and 3, a large group of samples, which included plants from Moscow, Kaluga, and Vladimir regions, the Russian Far East, Komi, Khakassia and Ukraine (the predominant components are trans-and cis-βocimenes, γterpinene, 1,8cineole, α and βpinenes and pulegone).
Intra-and interspecific variability of Mentha arvensis L. and M. canadensis L.
Euclidean distancebased cluster analysis of the essential oil composition divided samples into two clusters with a high degree of support (bootstrap support level 84-87 %) (see Fig. 2, b). M. spicata, taken as an external group, assumes a basal position in relation to two clusters, the predominant com pound of its essential oil being menthone (67.9 %) ( Table 3). The first cluster includes plants from Indochina, which are characterized by the accumulation of menthol (46.9-67.7 %) and its derivatives: menthone (3.11-30.3 %) and isomenthone (2.33-22.7 %) (see Table 3).
The second cluster includes all other plants (from Central Russia, Komi, Khakassia, and the Far East). This cluster has a bootstrap support below 84 %, and it is divided into three well supported subclusters. Within the second cluster, a subcluster is distinguished, consisting of plants of M. arvensis from Moscow (M6), Kaluga (K1), and Vladimir (F) regions and Ukraine (U2). The mints from the Russian Far East (RFE),  as well as the Republic of Komi (Komi) and Khakassia (Kh) formed two independent sister subclusters within it. It appears from data in Table 3   European Russia, and Western Ukraine with M. spicata as an external standard. Samples from the Far East flora were added to this group in the second experiment. A matrix including 117 polymorphic fragments was con structed in the first scheme of PCR. Plants were divided into two main groups when analyzing the data by the principal coordinate method involving the Dyce index (Fig. 3, a). One of these groups includes plants from Indochina, and the other, all plants from European Russia and Khakassia. M. spicata and one sample of Asian flora (Ind1) occupied a separate position. Plants from Ukraine formed a group hardly separated from other European and Khakassian samples. According to the results of cluster analysis (Fig. 3, b) there are two large groups of plants with a medium level of bootstrap support (41-56 %) and a fairly high degree of similarity (0.6). The first group included IndoAsian plants, and the second, European and Khakassian.
A matrix based on 80 polymorphic fragments was compiled according to the results of the second scheme of PCR. As a re sult of processing the data by the principal coordinate method and cluster analysis with the Dice index (Fig. 4), the plants were also divided into two groups. The first group included plants from Indochina, and the second, all the other plants. Some of them formed a common cloud/cluster, and others formed separate subgroups adjoining the general. Plants from Indochina contained specific amplicons, absent from other plants, including samples from the Russian Far East.
Plants of the second group fall into several subgroups. Plants from the Moscow, Vladimir, and Kaluga regions, as well as from the Republic of Komi and Khakassia, form a common cloud in which samples are practically inseparable. This indicates a significant genetic similarity between these populations. Two relatively separate groups of plants adjoin

Discussion
On the base of quantitative calyx and leaf characters, the Euro pean and IndoAsian groups of plants are clearly recognized, but the assignment of Far Eastern plants to one or the other group is ambiguous. Accordingly, it is difficult to determine their taxonomic position solely by morphological features. An additional criterion for the identification of M. arvensis and M. сanadensis might include such qualitative indicators as the ultrastructure of the seed surface (Shelepova et al., 2016a), but seeds were not available from plants of Indochi nese origin, as seeds rarely ripen when these accessions are grown in Central Russia. According to essential oil composition in the aerial parts of the plants, all the samples were distributed into differ ent chemotypes related to the genetic characteristics of the plants. The literature mentions many chemotypes for both M. arvensis and M. canadensis isolated on the basis of com ponents with shares above 10 % (Tucker, Chambers, 2002). Nine chemotypes are mentioned for M. arvensis: limonene, pulegone, 1,8cineole and/or βpinene, and others. Trans and cisβocimenes and γterpinene dominated in the essential oils of M. arvensis plants examined. These are acyclic limonene precursors and cyclic terpenoid intermediates. Limonene is    Ind1  Ind1-1a  Ind2  Ind3  Ind5-1  Ind5-2  Ind5-3  Ind3-1  Ind3-2  Ind3-3  Ind4-1  Ind5-2  Ind5-3  Ind4-2  Ind4-3  Ind5-2  Kh-1  Kh-3  Kh-4  Kh- known to be transformed through a series of biochemical reactions into isopulegone, which, in turn, is a precursor of pulegone (Bugaenko, 2011). This is recorded in M. arvensis plants from Komi and Khakassia (provisionally pulegone chemotype).
There have been identified, among others, 1,8cineole and βocimene chemotypes in M. canadensis from North America, and numerous chemotypes with menthol and menthone as major components in M. canadensis from Asia (Tucker, Chambers, 2002). Similar data were obtained in our stud ies: the menthol chemotype was identified in plants M. canadensis from Indochina and the 1,8cineole and trans-and cisβocimene in plants from the Russian Far East.
Analyzing the composition of ISSR fragments, we infer that M. arvensis plants from different populations of Euro pean Russia have significant genetic similarity and are almost indistinguishable from each other. On the contrary, Ukrainian and Far Eastern populations clearly differ from others. Far Eastern plants form a distinct group different from European populations, although they belong to the same cluster/cloud according to the ISSR analysis. Plants from Indochina differ significantly from other samples and belong to another species according to the results of ISSRPCR.
A similar pattern is observed for M. arvensis and M. сana densis plant distribution according to molecular analysis, essential oil composition, and, somewhat, morphological characters. Thus, there is a correlation between the molecular, phytochemical, and, to some extent, morphological data. On the basis of our results, a group of the genetically, morphologi cally, and phytochemically closest plants can be recognized. It includes representatives of the populations of the Moscow, Vladimir, and Kaluga regions; the Komi Republic; and Khakassia. All of them belong to M. arvensis.

Conclusion
According to our data on the compositions of essential oils and ISSR fragments, the plants collected in the natural flora of the Far East were more similar to M. arvensis from European Russia than to the collection samples from Indochina. With all that, a certain similarity to M. сanadensis was noted in plants of the Far Eastern population according to some morphological features (lengths of the calyx teeth were intermediate between M. arvensis from European Russia, and M. сanadensis from Indochina) and essential oil components, which were charac teristic of M. сanadensis from North America. Also, genetic isolation from the rest of the M. arvensis plants was detected. Perhaps the Far Eastern population includes plants of hybrid origin and has a closer relationship to M. arvensis than to plants from Indochina, which, according to the data obtained, belong to M. сanadensis. It follows from our results that a comprehensive analysis of morphological, phytochemical, and molecular data allows the most complete view of spe cies polymorphism and clarifies the species identification of plants.