The genus Populus L. is distributed in the northern hemisphere from subtropical to boreal forests (DiFazio et al., 2011) and is one of the most economically and ecologically important genera of forest trees (Stettler et al., 1996). Poplar wood is used to produce plywood, pulp, paper, and biofuels (Kobolkuti et al., 2019). Species and interspecific hybrids of the genus Populus are widely used in landscaping cities and creating protective plantings because poplars have such valuable characteristics as fast growth, ecological plasticity, and ease of vegetative propagation (Bakulin, 2005). Poplars show high dust, smoke and gas resistance, and a huge leaf surface purifies the air of harmful impurities and absorbs a significant amount of CO₂ (Kulagin, 1974; Jansson et al., 2010). In addition to their commercial value, poplars are ecologically important pioneers, colonizing disturbed and riverine ecosystems and contributing to the formation of natural forests in temperate and boreal regions, as well as in mountainous areas in the northern hemisphere (Schroeder et al., 2012).

The genus Populus, according to the Eckenwalder classification, includes 29 species in six sections: Abaso, Turanga, Populus (synonym: Leuce), Leucoides, Aigeiros, and Tacamahaca (Eckenwalder, 1996). Species of the genus Populus show extensive hybridization within sections as well as between closely related sections (Roe et al., 2014; Jiang et al., 2016). In this regard, the systematics of the genus Populus is rather complicated, and so far the researchers have not come to a consensus on the classification criteria, and the establishment of the taxonomic status of poplars of hybrid origin (both natural and cultivated) is particularly difficult (Cronk, 2005; Slavov and Zhelev, 2010; Schroeder et al., 2012; Mayorov et al., 2020). The question of the number of sections that should be distinguished in the genus Populus is also unresolved. For example, molecular genetic studies have shown that sections Leucoides, Aigeiros, and Tacamahaca should be considered not as separate sections but as a single subgenus, “Eupopulus” (Wang et al., 2020; Wang et al., 2022). Indeed, species of sections Aigeiros and Tacamahaca easily interbreed with each other, are connected by a powerful gene flow, and, in fact, have a common gene pool (at least, within Eurasia). Moreover, a common range of Eurasian poplars spreads from Western Europe to the Far East. Contacts between species in nature occur only at the borders of the habitats, i.e. “mixing” of genes occurs slowly. In this case, natural selection, despite gene flow, manages to maintain specificity of adaptive traits of each species, due to which it is possible to easily distinguish natural species morphologically (Nasimovich et al., 2019). In the study of cultivated poplars, identification is more difficult, but even in this case there is a factor supporting the discreteness of cultivars – the work of nurseries that propagate cultivars vegetatively. In this case, however, the control of varietal identity of seedlings is extremely important. When propagating poplars for use in urban landscaping, it is preferable to use only male clones that do not produce fluff (downy seeds) (Melnikova et al., 2019; Huang et al., 2022). However, it is impossible to determine the sex of a 1-2-year-old seedling by its phenotype; this can only be done after a few years when an adult tree begins to flower. In the cities of Central Russia, poplar was actively used in landscaping, but many female clones were mistakenly planted, which leads to annual problems with poplar fluff, because it collects and carries allergens, clogs drains and ventilation systems, is a fire hazard, and so on (Janković et al., 2021; Melnikova and Dmitriev, 2022). Genetic testing of seedlings to determine the sex would help not to repeat previous mistakes. In addition, when creating poplar hybrids for urban landscaping or industrial wood use, it is advisable to properly select parental species considering their adaptation to certain habitats (for example, representatives of the section Aigeiros are mostly plain poplars, and the section Tacamahaca includes mostly mountain poplars) (Nasimovich et al., 2019). Thus, understanding the evolutionary relationships of Populus species is important for basic research and breeding work (Wang et al., 2014).

A significant number of studies have been devoted to the search for reliable molecular markers for establishing phylogenetic relationships between poplar species – an analysis of various regions of nuclear DNA as well as sequences of the chloroplast genome was carried out (Schroeder et al., 2012; Wang et al., 2014; Alexandrov and Karlov, 2018; Zhou et al., 2018; Wang et al., 2022). The informativity for phylogenetic studies of poplars was noted for the sequences of non-transcribed intergenic spacers (NTS) of the 5S ribosomal RNA genes (rDNA) (Negi et al., 2002; Wilson, 2013; Alexandrov and Karlov, 2018) and internal transcribed spacers of rRNA genes (ITS) (Leskinen and Alström-Rapaport, 1999). At the same time, there are discrepancies between the phylogenetic trees obtained by analyzing the chloroplast and nuclear genome sequences of Populus species, which are usually attributed to their polyphyletic or hybrid origins (Hamzeh and Dayanandan, 2004; Wang et al., 2014; Liu et al., 2016; Zhou et al., 2018; Wang et al., 2022). In the analysis of chloroplast genome sequences, some accessions of the same species (a prime example is P. nigra) did not cluster together, but there was a significant association with geographic origin, which can be explained by recent hybridization (Wang et al., 2022). Analysis of different sequences allows clarification of various issues: chloroplast genome sequences showed great informativity for determining interspecific relationships in the genus Populus, including taxa difficult to systematize, while nuclear genome sequences were more effective for studying the deep phylogeny of Populus (Wang et al., 2022).

Poplars are dioecious plants, and another promising sequence for clarifying the systematic relationships of species and hybrids of the genus Populus is the sex-determining region (SDR). Sex chromosomes are actively used in phylogenetic studies (Parker et al., 2020; Navarro-Lopez et al., 2021). The species of sections Aigeiros and Tacamahaca are characterized by an XY sex-determination system (Gaudet et al., 2008; Pakull et al., 2009; Pakull et al., 2011; Kersten et al., 2014; Geraldes et al., 2015; Pakull et al., 2015; McKown et al., 2017), which allows the examination of poplar relationships in the male lineage. A series of studies made it possible to identify sex-associated DNA polymorphisms, determine the location and sequence of the SDR in various species of the genus Populus, and establish that sex is determined by the ARABIDOPSIS RESPONSE REGLATOR 17 (ARR17) gene – when ARR17 is turned on, female flowers develop, and when it is turned off – men ones (Gaudet et al., 2008; Yin et al., 2008; Pakull et al., 2009; Paolucci et al., 2010; Kersten et al., 2014; Geraldes et al., 2015; Pakull et al., 2015; McKown et al., 2017; Melnikova et al., 2019; Muller et al., 2020; Xue et al., 2020; Zhou et al., 2020; Kim et al., 2021; Melnikova et al., 2021; Yang et al., 2021). In species of sections Aigeiros and Tacamahaca, the SDR is localized on the end of chromosome 19, has a length of about 120 thousand nucleotides, and contains several genes, including T-complex protein 1 subunit gamma (TCP), Chloride channel protein CLC-c (CLC), and DNA-methyltransferase 1 (MET1). Only in the SDR of male plants, there are partial repeats of the ARR17 gene, from which small RNAs are synthesized, leading to the inactivation of ARR17, which is localized on the opposite to the SDR end of chromosome 19 (Xue et al., 2020; Zhou et al., 2020; Melnikova et al., 2021). Although the ARR17 gene plays a key role in sex determination in the genus Populus, species of this genus differ in the sex-determination system and structure of the SDR. P. euphratica and P. tremula have an XY sex-determination system, however, in P. tremula, the SDR containing partial ARR17 repeats is located in the peritelomeric region of chromosome 19, while in P. euphratica, it is located in the peritelomeric region of chromosome 14. Moreover, the SDR of these species is several hundred thousand nucleotides, differing significantly from the SDR of poplars of sections Aigeiros and Tacamahaca (Muller et al., 2020; Zhang et al., 2022). In P. alba, three copies of the ARR17 gene are located in the peritelomeric region of chromosome 19 in female plants, whereas in male plants, ARR17 is absent, indicating that P. alba has an ZW sex-determination system (Muller et al., 2020).

Previously, we showed that SDR polymorphisms can be effectively used in assessing the genetic proximity of Aigeiros and Tacamahaca species and hybrids based on whole genome sequencing data (Melnikova et al., 2021). The aim of this work was to evaluate the efficiency of using the SDR to establish the relationships of poplar species and hybrids of sections Aigeiros and Tacamahaca in combination with the nuclear and chloroplast genome sequences that have proven themselves in the study of the genus Populus systematics based on the target deep sequencing data.

To study the genetic relationship of poplar species and hybrids from sections Aigeiros and Tacamahaca, we chose an approach based on targeted deep sequencing of amplicons. This method drastically reduces the cost of analysis per sample compared to whole genome sequencing, has incomparable accuracy in the study of heterozygous genotypes and multicopy sequences, as well as less laboriousness in the analysis of large sample sets for a significant number of loci compared to Sanger sequencing (McCormack et al., 2013; Cruaud et al., 2017; Borkhert et al., 2018; Melnikova et al., 2019; Dmitriev et al., 2020; Novakovskiy et al., 2020; Povkhova et al., 2022). However, effective application of targeted deep sequencing requires the selection of the most informative loci.

Compared to the previous studies on Populus species based on the analysis of multicopy and single-copy sequences of nuclear and chloroplast genomes (Schmidt et al., 1994; Leskinen and Alström-Rapaport, 1999; Page, 2000; Huson and Scornavacca, 2011; Schroeder et al., 2012; Huang et al., 2014; Wang et al., 2014; Alexandrov and Karlov, 2018; Borkhert et al., 2018; Zhou et al., 2018; Wang et al., 2019), we sequenced regions of the SDR and the ARR17 gene, which is the key gene in sex determination in the genus Populus (Muller et al., 2020). The SDR allowed us to reveal the origin of the examined genotypes in the male lineage and significantly complemented the results obtained in the analysis of the DNA sequences traditionally used in poplar studies.

We studied interspecific hybrids, including those created by breeders, actively used in landscaping, as well as hybrids of natural origin and “pure” species from sections Aigeiros and Tacamahaca, which have an XY sex-determination system and interbreed easily. In poplars of these sections, the SDR is localized on the peritelomeric region of chromosome 19 and includes male-specific region containing partial repeats of the ARR17 gene, and a region that is present in plants of both sexes but differs significantly in DNA polymorphisms between males and females (Muller et al., 2020; Xue et al., 2020; Zhou et al., 2020; Melnikova et al., 2021; Pushkova et al., 2021), which is probably associated with the suppression of recombination. In this regard, the SDR could be promising for use in phylogenetic studies of species and hybrids of sections Aigeiros and Tacamahaca. Previously, based on whole-genome sequencing data, we showed the informativity of the SDR polymorphisms for determining putative parental forms of poplar hybrids (Melnikova et al., 2021). However, this approach requires significant costs for the analysis of one sample, so the analysis of large sample sets is complicated. Therefore, in this work, we designed primers for the amplification of the SDR loci of poplars of sections Aigeiros and Tacamahaca containing sex-specific polymorphisms ( Supplementary Data 3 ) for subsequent targeted deep sequencing.

Analysis of the SDR sequences made it possible to divide the studied samples of sections Aigeiros and Tacamahaca into male and female, and within groups of the same sex, the samples were predominantly clustered according to belonging to a particular hybrid or species ( Figure 4 , Supplementary Data 7E , 8R ). In addition to the SDR, we also sequenced DNA sequences traditionally used in poplar phylogenetic studies, namely NTS 5S rDNA, ITS, DSH 2, DSH 5, DSH 8, DSH 29, 6, 15, 16, X18, trnG-psbK-psbI, rps2-rpoC2, rpoC2-rpoC1 (Schmidt et al., 1994; Schroeder et al., 2012; Huang et al., 2014; Wang et al., 2014; Zhou et al., 2018; Wang et al., 2019). Clustering based on multicopy (NTS 5S rDNA and ITS) and single-copy (DSH 2, DSH 8, DSH 29, 6, 15, 16, X18) sequences of the nuclear genome and regions of the chloroplast genome (trnG-psbK-psbI and rps2-rpoC2) made it possible to separate a number of samples into groups according to belonging to a particular hybrid (often) or species (less often), however, none of the sequences allowed us to obtain a clear picture corresponding to the understanding of the similarity of the studied accessions based on morphological data. The most complete information on the phylogenetic relationships of the studied poplar accessions was achieved through the joint analysis of data for multicopy and single-copy sequences of nuclear DNA, loci of chloroplast DNA, and the SDR – clustering based on all obtained data of targeted deep sequencing most closely corresponded to the ideas about systematics of poplars based on morphological characteristics ( Figure 5 , Supplementary Data 7G , 8T ). Since the results of genetic analysis corresponded to the known facts, we can assume that they can be used to test hypotheses about the origin of poplars, as well as to determine the species identity of accessions whose identification on the basis of morphological features was complicated.

The analysis of sequencing data for the SDR and chloroplast genome loci made it possible to determine the origin of P. × petrovskoe [P. laurifolia (female tree) × P. × canadensis (male tree)] and P. × rasumovskoe (P. nigra (female tree) × P. suaveolens (male tree). The original description of P. × rasumovskoe given by M. Wolkenstein is as follows: “A hybrid between P. nigra, fertilised with the pollen of P. suaveolens” (Wolkenstein, 1882). It agrees with the results obtained by us on the basis of genetic analysis. However, for P. × petrovskoe, the original description by M. Wolkenstein is as follows: “A hybrid between P. canadensis fertilised by the pollen of P. suaveolens” (Wolkenstein, 1882). The indication of P. suaveolens as the parental species for P. × petrovskoe is a probable mistake by M. Wolkenstein. The Russian poplar species of the section Tacamahaca, of which there are two, were primarily involved in hybridization in Russia, and the “balsamic” component in P. × petrovskoe is probably represented by P. laurifolia, which is also proved morphologically by the ribbing of shoot axes (Mayorov et al., 2020). The sequencing data also confirm the error in the original description of P. × petrovskoe. The first parental species, P. × canadensis, was indicated correctly by Wolkenstein, although it went to the hybrid in the male lineage.

We analyzed the relationships of the studied poplar species and hybrids of sections Aigeiros and Tacamahaca based on molecular genetic (obtained results of targeted deep sequencing of NTS 5S rDNA, ITS, DSH 2, DSH 5, DSH 8, DSH 29, 6, 15, 16, X18, trnG-psbK-psbI, rps2-rpoC2, rpoC2-rpoC1, ARR17, and SDR – Figure 5 ), morphological, and ecological-geographical data. Below is our interpretation of the results obtained for particular poplar species and hybrids.

It is known that not universally recognized P. koreana and P. maximowiczii species are morphologically very close to P. suaveolens or even coincide with it (Skvortsov and Belyanina, 2006). In Figure 5 and dendrograms based on individual loci ( Supplementary Data 6 , 7 ), P. koreana and P. maximowiczii accessions always appear together with P. suaveolens ones. Thus, the obtained results indicate not only genetic proximity but also the identity of P. suaveolens, P. koreana, and P. maximowiczii, confirming the controversial assumption of Skvortsov and Belyanina (Skvortsov and Belyanina, 2006), which is still not recognized by most botanists, that P. koreana and P. maximowiczii should not be distinguished as species, but should be assigned to P. suaveolens. If these taxa are understood separately, we can speak only about geographical races, which are strictly confined to certain regions (P. koreana to Korea, P. maximowiczii to the Far East, etc.). It is unfounded to consider, as it is usually done, that all three species grow simultaneously, for example, in the Far East south of Amur. In this case, we can speak only about the external deviation to the appearance peculiar to a certain race.

All P. × canadensis accessions were arranged compactly in Figure 5 , with only P. deltoides and its hybrid with P. × moscoviensis falling into their subcluster, and this is natural, since P. × canadensis, according to the well-established notions, is a hybrid of P. deltoides and P. nigra. The distinction between P. × canadensis and P. deltoides is sometimes difficult; in the past, they were generally mixed (Skvortsov, 2010), and even errors in identification are possible. In the neighboring subclusters, exclusively P. nigra and P. pyramidalis appeared, and in the next ones by distance, almost exclusively P. × petrovskoe was observed. P. × petrovskoe, according to our recent assumption (Mayorov et al., 2020), is a hybrid of P. × canadensis and P. laurifolia, and the obtained molecular genetic results confirm this hypothesis. This result is not trivial at all, because previously the putative parental species of P. × petrovskoe were indicated as P. deltoides s.l. × P. suaveolens (Regel, 1889; Wolkenstein, 1882); P. deltoides × P. jackii (Dippel, 1892); P. × canadensis [or also P. deltoides?] × P. suaveolens (Bogdanov, 1965); P. deltoides × P. laurifolia (Koltzenburg, 1999; Tsvelev, 2001). Only Skvortsov (Skvortsov, 2010) indicated the same P. × canadensis and P. laurifolia, but in relation to P. × berolinensis K.Koch, for which this researcher mistakenly took P. × petrovskoe.

Most P. nigra accessions in Figure 5 were compactly arranged in cluster III together with P. pyramidalis. This group is adjacent to P. × canadensis, P. deltoides, and P. × petrovskoe. Several more P. nigra accessions were in other clusters, and for 5 of them, P. × irtyschensis accessions were the closest neighbors. Another close neighbor of P. nigra is P. laurifolia, which, according to our conceptions (Nasimovich et al., 2019), is a hybridogenic species that originated with the involvement of P. nigra. These results are quite natural, they support theoretical insights, but even more so, they support the ability of the analysis based on the targeted deep sequencing of the SDR and other nuclear and chloroplast loci used in phylogenetic studies to examine poplars. Importantly, samples of P. nigra and closely related species (primarily P. × irtyschensis) formed joint compact groups simultaneously in three different clusters (II, III, V). This suggests that the distribution of accessions, in addition to species affiliation, is influenced by the general polymorphism of P. nigra [also shown by Wang et al. (Wang et al., 2022)] and closely related species, and this must be taken into account when working with poplars. It is also important that accessions of P. nigra and closely related species did not fall into cluster I, formed primarily by P. suaveolens (including P. koreana and P. maximowiczii). This confirms the idea expressed earlier (Nasimovich and Vasilyeva, 2019), that P. nigra and P. suaveolens are geographical, biotopic, and morphological antipodes, occupying as if opposite “poles” in the unified system of Eurasian poplars. It turns out that they are also genetic “poles” of this system.

All but one P. laurifolia accessions in Figure 5 formed a compact group of subclusters in cluster IV, isolated from P. nigra and closely related species, and isolated from P. suaveolens and closely related species (cluster I). This probably implies the species independence of P. laurifolia, although we have argued that P. laurifolia emerged as a hybridogenic species at the intersection of gene flow from P. nigra and P. suaveolens (Nasimovich et al., 2019). Based on the molecular genetic analysis, P. laurifolia is slightly closer to P. nigra than to P. suaveolens. P. laurifolia is one of the “balsam” poplars, but partially shares a common habitat with P. nigra, which turns out to be more important. In the same or close subclusters, together with P. laurifolia, there were its putative hybrids: P. × irtyschensis (P. laurifolia × P. nigra – 1 sample) and P. × wobstii (P. laurifolia × P. longifolia – 2 samples), which confirms their hybrid nature and the putative composition of parental species. For P. × wobstii, this result is not quite trivial, because previously the putative parental species of P. × wobstii were indicated as P. suaveolens (Regel, 1889); P. balsamifera (Dippel, 1892); P. suaveolens and P. jackii (Sokolov et al., 1951); P. simonii and P. suaveolens (Karhu and Hamet-Ahti, 1992); P. laurifolia and P. tristis (Rehder, 1949; Koltzenburg, 1999); P. laurifolia and P. tristis (or P. longifolia) (Ascherson and Graebner, 1908-1913; Tsvelev, 2001).

Most P. longifolia accessions in Figure 5 were in cluster I together with P. suaveolens and closely related species, but they were absent in clusters III (together with P. nigra) and IV (together with P. laurifolia). This indirectly confirms our hypothesis on the origin of P. longifolia from P. suaveolens (Mayorov et al., 2012), although for complete certainty it is necessary to investigate in the same way the similarity with the American species P. balsamifera and P. trichocarpa, which are also considered as ancestors of P. longifolia (Tsinovskis, 1977; Skvortsov, 2008).

P. suaveolens and closely related species constitute the vast majority in cluster I ( Figure 5 ) and, in addition, are present in cluster II (3 accessions) together with P. × rasumovskoe, while in clusters III and IV, where P. nigra and P. laurifolia dominate among “pure” species, they are absent. This means a molecular genetic opposition with respect to other (more western) “pure” species, as already mentioned above. P. longifolia, P. talassica, P. afghanica, and the hybrid P. × rasumovskoe (P. nigra × P. suaveolens), which form clusters I and II, are also close to P. suaveolens (together with P. koreana and P. maximowiczii). P. talassica has even been combined with P. suaveolens into one species in the past; P. afghanica is geographically close to P. suaveolens; P. longifolia may have separated from P. suaveolens; and the situation with P. × rasumovskoe is even clearer (see above). In general, the position of P. suaveolens on the dendrogram is completely logical and shows great opportunities for the use of targeted deep sequencing of chloroplast and nuclear sequences, including the SDR, in the study of poplars. Analysis of sequencing data revealed the genetic proximity of P. talassica and P. afghanica. These results support the hypothesis of hybridogenic origin of P. afghanica – this species is close to P. nigra by the traits of leaves and shoots, but there are also traits characteristic of poplars of the section Tacamahaca, probably derived from P. talassica ( Supplementary Data 2 ). It can also be seen that the geographical proximity of species and, therefore, the possibility of hybridization between them is sometimes more important in terms of genetics than the sectional position of species: P. suaveolens and P. afghanica belong to different sections, but are geographically close and therefore alike.

Accessions of the complex hybrid P. × petrovskoe, together with its putative parental species, P. deltoides and P. nigra, form the core of neighboring clusters III and IV ( Figure 5 ), which proves their kinship. Another putative parental species of P. × petrovskoe, P. laurifolia, is also a part of cluster IV. It is also important that P. × petrovskoe does not overlap with P. suaveolens (clusters I and II) in Figure 5 , and it is the only mass hybrid of Moscow landscaping whose parental species, according to our knowledge (Mayorov et al., 2020), does not include P. suaveolens. Thus, the molecular genetic analysis fully confirmed our ideas about P. × petrovskoe.

Based on the results of targeted deep sequencing and on the analysis of morphological and ecological-geographical data (see Supplementary Data 2 ), we present a scheme of relationships between species and hybrids of sections Aigeiros and Tacamahaca ( Figure 6 ). It should be noted that relationships between taxa from different sections are traced. P. afghanica (Aigeiros) is mountain poplar and morphologically close to P. nigra (Aigeiros), but also has some features similar with “balsam” poplars. We supposed that P. afghanica can be a species of hybrid origin that arose at the intersection of gene flow from P. nigra and P. talassica (Tacamahaca). The obtained genetic data support this hypothesis. P. laurifolia (Tacamahaca), which is a fairly typical representative of “balsam” poplars and has all diagnostic features of this section, is genetically close to P. nigra (Aigeiros), but distant from another “balsam” poplar, P. suaveolens (Tacamahaca). In addition, a significant number of poplar hybrids, including P. × petrovskoe, P. × rasumovskoe, P. × sibirica, and P. × irtyschensis, come from crosses of representatives of different sections.

In the group of “black” poplars (Aigeiros), there are similarities of P. × canadensis with P. nigra and P. deltoides. The genetic data are consistent with the known information that P. nigra and P. deltoides are the parental species of P. × canadensis. In the group of “balsam” poplars (Tacamahaca), it is inexpedient to distinguish P. suaveolens, P. koreana, and P. maximowiczii as separate species – they are a single species according to both morphological (Skvortsov and Belyanina, 2006) and genetic data. P. suaveolens is a prominent representative of “balsam” poplars growing in mountain conditions. It is probably the parental species for P. laurifolia, P. talassica, and possibly P. longifolia based on the analysis of both morphological and genetic data. P. × wobstii probably comes from crosses between P. laurifolia and P. longifolia. According to genetic data, P. simonii is close to P. suaveolens, which can be explained by the ecological-geographical relationship of these species.

Considering the status of sections Aigeiros and Tacamahaca, it should be noted that their representatives freely interbreed in both cultivation and nature. Meanwhile, sections Aigeiros and Tacamahaca differ in a large set of traits (more details in Materials and Methods), but all these traits are adaptations to either plain (“black” poplars) or mountain conditions (“balsam” poplars) (Nasimovich et al., 2019). The traits of “black” poplars are most pronounced in P. nigra growing on the vast Russian plain, while the traits of “balsam” poplars are most expressed in P. suaveolens growing in the harsh mountain conditions of the Eastern Siberia (Nasimovich and Vasilyeva, 2019). Our study revealed significant genetic differences between P. nigra and P. suaveolens. Considering a number of other poplar species (e.g., P. laurifolia and P. afghanica), the genetic unity within sections is not revealed. Geographical proximity and the possibility of gene exchange turn out to be more important in terms of genetics than the sectional position. Therefore, sections Aigeiros and Tacamahaca are most likely ecological (plain and mountain poplars) and have no phylogenetic or genetic meaning. However, discarding these sections would make the taxonomy of poplars less convenient, as we discussed earlier (Nasimovich et al., 2019).

The performed genetic analysis also made it possible to establish belonging to the common Moscow hybrids, P. × rasumovskoe, P. × petrovskoe, and P. × sibirica, for a number of accessions with unknown or doubtful origins, as well as to find several errors in identification of accessions. To control fluff in Russian cities, poplars are often subjected to cardinal pruning leading to changes in the traits of leaves, shoots, and crowns, which makes it difficult to identify poplar species and hybrids on the basis of morphological characteristics, and for such accessions, genetic analysis can clarify the situation.

Our study allowed us to evaluate the polymorphism of the three hybrids most commonly used in Moscow landscaping: P. × rasumovskoe, P. × petrovskoe, and P. × sibirica. The analysis based on the data for all sequenced regions at once provided the clearest picture of the separation of samples of these hybrids into their respective groups, but even analyses by individual loci in most cases allowed us to separate the P. × rasumovskoe, P. × petrovskoe, and P. × sibirica groups ( Figures 1 - 5 , Supplementary Data 6 , 7 ).

P. × rasumovskoe was represented by male accessions that were genetically very close – the average genetic distance between accessions was 0.2 ( Supplementary Data 5 ). P. × rasumovskoe is one of the spontaneous hybrids discovered before 1882 by R.I. Schroder at the Moscow Agricultural Academy (Wolkenstein, 1882). Probably, all accessions of P. × rasumovskoe studied by us are descendants of one clone, which was quite reasonably used in landscaping of Moscow – male plants do not form fluff and have an attractive weeping crown shape.

The studied samples of P. × petrovskoe were both male and female. Most male accessions were genetically close (the average genetic distance between the accessions was 0.2) and were probably the descendants of one clone. However, two other small groups of P. × petrovskoe male accessions were genetically different from the main group (the average genetic distance between them and the main group was 0.8) ( Supplementary Data 5 ). The main group of female trees of P. × petrovskoe also probably came from one clone (the average genetic distance between the accessions was 0.2), and the three genetically different female accessions most likely belong not to P. × petrovskoe but to P. × sibirica. Thus, two clones of P. × petrovskoe were predominantly used in Moscow landscaping – male one and female one. While the use of the male clone is justified, the planting of female plants was a mistake that led to the appearance of poplar fluff in the city during the ripening of seeds (poplar fluff contains seeds).

For P. × sibirica, the picture was different from that obtained for P. × rasumovskoe and P. × petrovskoe. P. × sibirica was represented predominantly by female trees, and not as genetically similar as in the case of P. × rasumovskoe and P. × petrovskoe. It can be assumed that several predominantly female P. × sibirica clones were used in the landscaping of Moscow. Such a choice was unreasonable and led to a massive problem with poplar fluff.

It is worth noting that among Moscow citizens an opinion about the sex change of poplars is widespread, including due to the media. There is a belief that originally only male plants were planted, and then under the influence of adverse factors they have changed sex to a female, and this has led to the problem of poplar fluff, which is formed only on female plants. Our work shows that the cause of the mass appearance of fluff is errors in the choice of initial material for landscaping, and not at all the sex change of poplars. Only in the case of P. × rasumovskoe, descendants of a male clone were used. For P. × petrovskoe, both male and female descendants were planted. And for P. × sibirica, descendants of several female clones were mostly used in landscaping. We observed complete correspondence between the presence of female or male generative organs on P. × rasumovskoe, P. × petrovskoe, and P. × sibirica trees and the division of all samples of these hybrids into two sexes based on the sequences of the SDR. In this regard, the assumption of mass sex change in Moscow poplars is erroneous. The possibility of formation of male, female, and hermaphroditic generative organs on one tree and the change of their ratio during tree development was shown for hybrids of white poplar (Populus alba L.) and aspen (Populus tremula L.) – gray poplar (Populus × canescens (Aiton) Sm.) (Sabatti et al., 2020). This is due to the fact that P. alba has a ZW system of sex determination and P. tremula has an XY system of sex determination (Cronk and Muller, 2020; Muller et al., 2020). P. × canescens, as well as P. alba and P. tremula, occurs in Moscow an order of magnitude less frequently than intersectional hybrids of poplars of sections Aigeiros and Tacamahaca (Nasimovich et al., 2019; Mayorov et al., 2020); they can also be identified using genetic markers (Kim et al., 2021).

Thus, the use of the SDR loci considerably complements the sequences traditionally used for phylogenetic studies to clarify poplar systematics, makes it possible to divide accessions into male and female and determine the ancestral species in the male lineage. We shed light on a number of controversial hypotheses on the division of poplar accessions into species and on the origin of a number of species and hybrids. A detailed analysis of the obtained data on the SDR sequences and the regions of nuclear and chloroplast genomes traditionally used in phylogenetic studies of poplars with involvement of morphological data will make it possible to clarify the origin and systematics of other studied species of the genus Populus as well.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, BioProject PRJNA943410.

AD and NM conceived and designed the work. EB, EP, YN, MK, NV, RM, RN, LP, DZ, AT, ES, AS, and NB performed the experiments. EB, EP, YN, MK, RM, ED, AK, NB, GK, AD, and NM analyzed the data. EB, EP, YN, DZ, GK, AD, and NM wrote the manuscript. All authors contributed to the article and approved the submitted version.