Micropropagation as a Tool for the Conservation of Autochthonous Sorbus Species of Czechia

Members of the genus Sorbus are the only endemic tree species that occur in Czechia. They are important components of endangered plant communities. Their natural regeneration is usually problematic because of their mode of reproduction and because they can survive in rare populations with small numbers of individuals. The aim of this study was to develop a successful micropropagation protocol for selected Sorbus species, of which two are endemic (S. gemella and S. omissa) and two are hybrid (S. × abscondita and S. × kitaibeliana). We found significant differences in shoot induction and rooting ability between the Sorbus species under study. With the exception of S. × abscondita, N6-benzyladenine had a significantly greater effect on shoot regeneration, both in terms of shoot number and total shoot length, than meta-topolin. Root induction was key to the successful micropropagation of the Sorbus species studied. Our results show that four Sorbus species can be successfully rooted under ex vitro conditions, without a rooting powder treatment in a steamed peat-perlite substrate. Auxin-untreated microcuttings of S. gemella, S. × kitaibeliana and S. omissa, but not S. × abscondita, rooted better than ones treated with indole-3-butyric acid. This is the first time a micropropagation protocol for S. omissa, S. × abscondita and S. × kitaibeliana has been published.


Introduction
In the flora of Europe, the genus Sorbus L. is one of the most taxonomically challenging groups of woody plants. Taxonomical complications stem chiefly from the vast morphological variability and huge genetic diversity of the genus, which is generated by frequent interspecific hybridization, genome duplication and subsequent stabilization of offspring by apomixis [1][2][3][4][5]. Furthermore, diverse opinions on taxonomic treatment [6,7] and a tangled nomenclature still cause disagreements and misinterpretations.

In Vitro Shoot Proliferation
The occurrence of explant contamination was low, ranging from 0 to 15%. Satisfactory reactivity of the explants to the in vitro cultures was achieved and reached a total of 79%. In our experiments, initial cultures of Sorbus species were established from axillary buds originating from the lower branches of donor trees. Better regeneration of shoots and roots could be expected in the subsequent stages of in vitro culture. This effect was observed in S. aucuparia, when microshoots derived from the lower branches of mature trees exhibited better shoot proliferation and rooting response than from the top branches [43]. After 6 months of cultivation, in vitro cultures of Sorbus species were stabilized.
The application of cytokinins during multiplication significantly increased shoot regeneration compared to the control (1.03; Figure 1C). A higher shoot production (4.17) was observed on the multiplication medium with N 6 -benzyladenine (BA) than the one with meta-topolin (mT, 1.58). The combination of each of the cytokinins with the auxin indole-3-butyric acid (IBA) had no statistical effect on the mean number of shoots.
Another growth parameter was total shoot length, in which significant differences were found between Sorbus species ( Figure 1B); the response was similar to that for the mean number of shoots per explant ( Figure 1A). Sorbus × abscondita (61.48 mm) and S. × kitaibeliana (53.72 mm) produced significantly longer shoots than S. omissa (38.47 mm) and S. gemella (36.07 mm).
The cytokinins also had a positive effect on total shoot length compared to the control (20.67 mm; Figure 1D). The shoots on the multiplication medium with BA were significantly longer (77.89 mm) than that with mT (39.61 mm). The combination of each of the cytokinins with the auxin (IBA) had no statistical effect on total shoot length.
The interaction between the genotype and the PGR treatment showed a different response in shoot production in three Sorbus species (Table 1). In S. × abscondita, the type of cytokinin played a significant role, and shoot formation was significantly higher on the medium with mT (3.72) or mT + IBA (3.47) than on that with BA (2.97). In S. gemella, shoot formation was significantly higher on the medium with mT + BA (2.69) compared to the control (1.11) and the IBA treatment (1.03). In S. × kitaibeliana, shoot formation was significantly higher on the medium with BA (3.42) compared to the control (1.28) and IBA treatment (1.36).   Regarding the interaction between the genotype and the PGR treatment on total shoot length, significant differences were observed in two Sorbus species (Table 2). In S. × abscondita, shoot elongation was significantly better in the mT + IBA treatment (89.22 mm) than in the combined BA + IBA treatment (54.94 mm) and the IBA treatment (39.99 mm). In S. × kitaibeliana, the combination of mT + IBA (65.50 mm) was also effective, as was the BA treatment (71.00 mm), where the total length of the shoots was significantly higher compared to the control (31.69 mm) and the IBA treatment (36.31 mm).

Ex Vitro Rooting and Acclimatization
The effects of different factors (genotype, auxin treatment/commercial rooting powder and microcutting size) on rooting were examined. All observed factors had significant effects on the rooting percentage and the mean number of roots per microcutting at different significance levels. The effects of interactions between two factors were found to be significant for genotype × auxin treatment for all growth characteristics and genotype × microcutting size for rooting percentage (Table 3). Rhizogenesis was observed in all Sorbus species. There were statistically significant differences in rooting between Sorbus × abscondita, S. gemella, S. × kitaibeliana and S. omissa (Figures 2A and 3). A high rooting percentage was found in S. × abscondita (74%) and S. omissa (62%), whereas a low root capability was observed in S. × kitaibeliana (27%) and S. gemella (14%). A similar trend was also observed for the mean number of roots per microcutting (5.78, 3.18, 1.98 and 2.51, respectively; Figure 2B). *, **, ***, ns-Tukey's range test significant at p = 0.05, p = 0.01, p = 0.001 or not significant; df means degrees of freedom, and F is the F-statistic.    Root production also depended on the presence and the concentration of IBA in the rooting powder ( Figure 2C). Auxin treatments (1% and 2% IBA) reduced the rooting percentage significantly (40% and 32%) compared to the control (61%). On the other hand, the auxin significantly increased the mean number of roots per microcutting, both in the 1% Plants 2023, 12, 488 9 of 20 IBA treatment (4.66) and in the 2% IBA treatment (5.12.) compared with the control (2.91), see Figure 2D.
The size of the microcutting was also an important factor affecting Sorbus rooting ( Figure 2E). We found that microcuttings that were 15-25 mm long rooted significantly better (54%) than microcuttings that were 26-35 mm long (34%). The opposite effect was observed for the mean number of roots per microcutting; in longer microcuttings, the formation of roots was significantly higher (4.66) than in shorter microcuttings (3.56), see Figure 2F.
The interaction between the genotype and the auxin treatment showed a different response in the rooting percentage of Sorbus × abscondita compared to other Sorbus species ( Figure 2G). Only in this species was there a significant difference in rooting percentage between the control (81%) and the 2% IBA treatment (90%). The effect of auxin on the mean number of roots per microcutting was demonstrated conclusively in three Sorbus genotypes ( Figure 2H). The auxin treatment increased the mean number of roots in S. × abscondita (8.0 and 6.59 with 1% and 2% IBA, respectively) and in S. omissa (1% IBA, 3.32) and had an antagonistic effect in S. gemella (2% IBA, 0.0) compared to the control. The effect of the interaction between the genotype and the size of the microcuttings on rooting percentage was significant only in S. omissa, where shorter microcuttings rooted better (84%) compared to longer microcuttings (40%; Figure 2I).
The plant acclimation process started during ex vitro rooting and was successfully completed after transplanting the plants to a greenhouse (Figures 3 and 4).

Discussion
In preservation programs that specialize in the propagation of endangered tree species, often only adult trees are available as a source plant. The success of micropropagation is strongly influenced by the ontogenetic and chronological age of source plants. According to [27], ontogenetic aging (physiological) refers to the phases of development that the seedling plant undergoes from embryonic to juvenile to intermediate to mature (adult), and chronological aging continues through the life of an individual plant whether as a seedling or vegetatively propagated. In vitro cultures initiated from source plants that are physiologically mature or less vital (old) or both may exhibit growth depression during micropropagation.
In vitro cultures can be successfully obtained from an adult tree in several ways: establishing cultures from plant parts showing juvenile characteristics, from mature parts after rejuvenation or from serial subcultures of explants in a hormone medium [44][45][46].
Plant growth regulators play a key role in adventitious shoot regeneration in woody species of the temperate zone [47]. The presence of a cytokinin is necessary for de novo shoot formation in woody species, and very often a cytokinin is combined with auxin at a low concentration to increase the number of shoots.
BA is an important aromatic cytokinin that is very effective at shoot induction in many plant species, including Sorbus species, and it is affordable compared to other cytokinins. On the other hand, it can cause growth abnormalities in some plant species, reduce

Discussion
In preservation programs that specialize in the propagation of endangered tree species, often only adult trees are available as a source plant. The success of micropropagation is strongly influenced by the ontogenetic and chronological age of source plants. According to [27], ontogenetic aging (physiological) refers to the phases of development that the seedling plant undergoes from embryonic to juvenile to intermediate to mature (adult), and chronological aging continues through the life of an individual plant whether as a seedling or vegetatively propagated. In vitro cultures initiated from source plants that are physiologically mature or less vital (old) or both may exhibit growth depression during micropropagation.
In vitro cultures can be successfully obtained from an adult tree in several ways: establishing cultures from plant parts showing juvenile characteristics, from mature parts after rejuvenation or from serial subcultures of explants in a hormone medium [44][45][46].
Plant growth regulators play a key role in adventitious shoot regeneration in woody species of the temperate zone [47]. The presence of a cytokinin is necessary for de novo shoot formation in woody species, and very often a cytokinin is combined with auxin at a low concentration to increase the number of shoots.
BA is an important aromatic cytokinin that is very effective at shoot induction in many plant species, including Sorbus species, and it is affordable compared to other cytokinins. On the other hand, it can cause growth abnormalities in some plant species, reduce rooting and worsen the subsequent acclimatization of plants [30,48,49]. For these reasons, new aromatic cytokinin derivatives have been the subject of intensive research [50]. One of them is mT, a natural highly active aromatic cytokinin occurring in poplar leaves [51]. Meta-topolin has been successfully used in the micropropagation of many plant species, including both herbs [49,52,53] and trees [54][55][56]. Ördögh et al. [57] published the effects of different types of cytokinin, including mT, on shoot proliferation in S. borbasii Jáv.
In our study, we compared the effectiveness of two cytokinins, BA and mT, singly or in combination, with auxin in four Sorbus species. In agreement with the results of Nikolaou et al. [37] and Jeong and Sivanesan [58], the presence of a cytokinin was essential for shoot proliferation in Sorbus species. In general, BA had a significantly higher effect on shoot regeneration, both in terms of shoot number and total shoot length, than mT ( Figure 1C,D). This finding that BA increases the number of shoots more effectively than mT is consistent with the results obtained by Ördögh et al. [59] for S. redliana Kárpáti, by Malá et al. [30] for S. torminalis, by Ördögh et al. [57] for Sorbus borbasii and by Meyer et al. [60] for Hypericum L. species. On the other hand, some Sorbus species behaved differently: In S. × abscondita, shoot formation was significantly more vigorous on the multiplication medium with mT than the one with BA (Tables 1 and 2). A similar stronger reaction to mT in shoot proliferation was observed in Ribes [61]. Likewise, Hlophe et al. [62] noticed that each Brachystelma R. Br. species differed in its response to specific cytokinins.
The combination of each of the cytokinins with the auxin IBA or the auxin alone had no significant effect on growth parameters in the Sorbus species studied (Figure 1), although a slight improvement in shoot production was observed in S. gemella and S. omissa on the medium with mT + IBA (Table 2). Similar results were obtained with S. domestica, where the addition of IBA to BA had no effect on shoot proliferation [37]. Many authors routinely used a combination of a cytokinin and the auxin IBA for proliferation without further analyses [30,41,43,57,59]. In contrast to our results, it has been reported that the combination of a cytokinin (BA) and auxin (IAA or NAA) had a synergistic effect on shoot multiplication in Sorbus commixta Hedl. [58].
The juvenile phase of most plants inherently has a higher rooting potential than the mature phase [27]. Shoots from juvenile seedlings had a higher rooting ability than those from mature material in Diospyros kaki L. [63]. Physiological status also played a significant role in the rooting ability of Sorbus species [34,37,41,43].
In vitro culture can conduce rejuvenation (or reinvigoration), but not always, and it can depend on the number of subcultures in some species. In vitro serial subculture improved rooting in Diospyros kaki [63], Eucalyptus urophylla S. T. Blake [64], apple and cherry rootstocks [65].
Root formation in microcuttings depends on the plant species (genotype) and the cultivation method applied. Even within a species, the rooting ability of microcuttings varies among different cultivars and clones [66]. The effect of genotype was also observed in four Sorbus species; significant differences in rooting ability were found between S. × abscondita, S. gemella, S. × kitaibeliana and S. omissa (Figure 2A). Regardless of the age of the donor plants, Sorbus × abscondita rooted best even when a donor tree was about 50 years old, while S. gemella had the lowest rooting ability of all Sorbus species even when the in vitro culture was established from a 10-15-year-old tree. The opposite situation was found in Sorbus × kitaibeliana, where rooting ability was low (an approximately 80-year-old tree), while S. omissa had a high rooting ability (a 10-year-old tree). These results suggest that genotype plays a large role in rooting ability. Sorbus × abscondita is a member of S. subgen. Soraria Májovský and Bernátová [67] originating from the hybridization of some members of S. subgen. Aria and S. aucuparia. In contrast, S. gemella, S. × kitaibeliana and S. omissa come from the S. subgen. Tormaria Májovský and Bernátová [67], where instead of S. aucuparia, the second parent is S. torminalis. Since the two parent species are quite different ecologically and genetically, the newly created species can show significantly different characteristics accordingly.
One positive aspect of in vitro cultures is that woody species that do not root easily via conventional propagation methods can easily be rooted as microcuttings [68][69][70]. In general, the problems of rooting are more pronounced with woody species [71]. Physiological status played a significant role in the rooting ability of some Sorbus species. Microcuttings derived from juvenile material had a higher rooting ability than those from mature material [34,37,41].
The rooting of microcuttings can take place either in vitro or ex vitro. Rooting in ex vitro environments has two main benefits: it is economic and functional. Rooting under non-sterile conditions should reduce the cost of plant production by microcutting the steps of aseptic manipulation [44]. In vitro rooting of micropropagated plants is an expensive process that can double the final price of these plants [72].
The formation of roots in a tissue culture environment, especially under high humidity and in the presence of sugar, results in morphological and physiological differences compared to roots that develop under normal conditions [73][74][75]. Therefore, in vitro roots must adapt in non-sterile conditions to function properly. By contrast, ex vitro rooted plantlets did not require any additional acclimatization prior to transplanting to regular greenhouse conditions [76].
Ex vitro rooting includes two basic methods to root microcuttings. Both the induction and expression of roots are performed either ex vitro in a greenhouse medium, or they take place in agar or a liquid culture within an ex vitro environment [27].
In previous experiments, we tested the rooting potential of Sorbus × abscondita and S. omissa on a half-strength MS agar medium containing IBA or NAA (α-naphthalene acetic acid) at concentrations of 0, 0.5 and 3.0 mg·L -1 . The highest rooting (55.6%) was observed in S. × abscondita at 3 mg·L −1 of IBA. In S. omissa, no roots were formed, regardless of the type of auxin and its concentration [78]. When these genotypes were rooted ex vitro, there was a significant increase in rooting ability in S. × abscondita (74%) and in S. omissa (62%- Figure 2A). Similar results were observed in mulberry [68].
Auxin is a key growth regulator in adventitious root formation in plants. Many plants require the presence of auxin for efficient root regeneration [79]. Some of the most important factors affecting in vitro adventitious rooting are the choice of auxin, its concentration and the duration of tissue exposure [80]. IBA is the most often-used agent for rooting microcuttings of a wide variety of species [27]. When microcuttings are rooted in vitro, not only does the level and the duration of auxin treatment differ from ex vitro rooting, but so does the rate of gas exchange by the section of the stem in which the new roots are formed. The presence of auxin in the rooting medium increases the synthesis of ethylene and can have a negative effect on root regeneration [81,82].
The use of various auxins for in vitro root formation in Sorbus species has been reported as follows: Rhizogenesis was induced with IBA in S. commixta and in S. × abscondita [58,79]; in S. domestica, a high percentage of rooting was achieved with both IBA and NAA. A combination of IBA with NAA proved effective in S. torminalis and S. aucuparia [32,43], and a combination of IBA with IAA was found to have an effect in infrageneric hybrid × Malosorbus florentina [41] and NAA has been used with success in seven rare endemic species [42].
In our experiments, we tested the possibility of ex vitro rooting (performing both induction and expression) in four Sorbus species. Rooting was achieved in all the species tested. Rooting ability was high in S. × abscondita and S. omissa but lower in S. gemella and S. × kitaibeliana (Figure 2A). Regarding the effect of auxin treatment on rooting ability, very interesting results were achieved: in S. gemella, S. × kitaibeliana and S. omissa, better results were obtained in rooting without auxin treatment, excluding S. × abscondita, where the highest rooting percentage was observed in the 2% IBA treatment ( Figure 2G). A reduction in rooting in the presence of auxin in an agar medium compared to the control (no auxin) has been reported for Anemone L. [83]. The balance of endogenous growth hormones is a key factor in successful organogenesis in plant tissue cultures. Hormonal levels are significantly affected by the addition of exogenous PGRs to the culture medium [84]. The application of PGRs during multiplication can have a negative effect on subsequent rhizogenesis, which has been demonstrated, for example, in S. torminalis [30].
Another factor that can influence rooting in woody plants is the size of the cutting/microcutting [85,86]. Our results showed that determining the optimal length of microcuttings in Sorbus species had a positive effect on rooting. In general, shorter shoots were thinner than longer shoots. The leaves were also less mature. In S. × abscondita, the biggest differences were observed visually in the maturity of the leaves; in addition, shorter shoots were more sensitive to dying.
In conclusion, a successful micropropagation protocol for four Sorbus species was developed. The most difficult phase of the protocol was root induction. The results showed that Sorbus species can be rooted non-sterile in a peat substrate without the presence of auxin. Compared to the other species, Sorbus × abscondita had the highest regeneration and rooting ability. The significantly better results we achieved may have been caused by the influence of a different parental combination of the mentioned species. This protocol will be used in future studies to evaluate the regeneration potential in other species of Sorbus subgen. Tormaria and Sorbus subgen. Soraria [67].

Plant Material and In Vitro Culture Initiation
Characteristics of Sorbus species used in our study ( Figure 5) and their localities are as follows: Sorbus × abscondita is a rare hybrid between S. aucuparia and S. danubialis [14] that is treated as a member of S. subgen. Soraria [67]. To this day, it has been recorded at eleven localities in Bohemia (historical part of Czechia) [14], but five of them have gone extinct. Sorbus gemella is a triploid species [87] regarded as a member of S. subgen. Tormaria [67], with the supposed parentage of S. danubialis and S. torminalis. Its distribution is limited to the Džbán tableland hills located mostly in the Louny district in the north-central part of Bohemia [14]. The species inhabits edges of marlite plateaus and upper parts of steep slopes on the same bedrock predominantly in oak woodlands. Several subpopulations consist of approximately 1500 individuals. The approximate GPS coordinates of the type locality in Konětopy are 50 • [14]. There is one known adult, a 12 m high tree, and one 1.5 m high seedling.
Sorbus omissa is another triploid species [88] of the same subgenus and supposed parentage. It is a stenoendemic species occurring in central Bohemia, in the vicinity of the towns of Roztoky and Libčice nad Vltavou (Praha-západ district) in the valley of the lower reach of the river Vltava, where it grows primarily in oak woodlands on humus soils on a substratum of Proterozoic slate. Its two known populations consist of approximately 150 individuals [14,88]. The GPS coordinates of the type specimen of the species are 50 • 10 06.9 N, 14 • 21 09.5 E. The areas of species with larger distribution were delineated with the "Sample by Buffered Local Adaptive Convex-Hull" tool [89], which combines the creation of a wrapper zone and a minimal convex polygon. Data on Sorbus species with larger distribution were derived from the Pladias database [14,90].
Plant material of the four Sorbus species, with one clone from each of the selected samples (S. × abscondita-Kladno-Švermov, an approximately 50-year-old donor tree; S. gemella-Konětopy, an approximately 10-15-year-old tree; S. × kitaibeliana-Trmice, an approximately 80-year-old tree and S. omissa-Roztoky, an approximately 10-year-old tree; Figure 6), was identified and collected from natural localities in April or May. Shoots were taken from the lower part of the crown; epicormic sprouts were not available. To obtain in vitro cultures, stem cuttings were taken (40-60 mm in length, Figure 7A,B) and surfacesterilized in 2.5% sodium hypochlorite solution (50% Savo ® , Bochemie a.s., Bohumín, The Czech Republic) for 15 min with one drop of Tween-20. Finally, the segments were washed with sterile distilled water three times for 10 min each. Cultures were initiated from apical stem segments with one apical bud and one pair of axillary buds (10-15 mm in length, Figure 7C) and transferred onto a full-strength solid Murashige and Skoog (MS) basal medium with vitamins [91] (Duchefa Biochemie B.V., RV Haarlem, The Netherlands) supplemented with 0.5 mg·L −1 BA (N 6 -benzyladenine), 0.1 mg·L −1 IBA (indole-3butyric acid), 20 g·L −1 sucrose, 2 mL·L −1 PPM (Plant Preservative Mixture, Washington, DC, USA) and 7 g·L −1 agar (Sigma-Aldrich Inc., St. Louis, USA). The pH was adjusted to 5.7 using NaOH prior to autoclaving. Explants were cultured in 100 mL Erlenmeyer flasks containing 25 mL of a MS medium. The cultures were maintained in a growth room under a 16 h photoperiod with a photosynthetic photon flux density (PPFD) of 60 µmol·m −2 ·s provided by cool-white fluorescent tubes (Tungsram, General Electric Company, Boston, USA) at 22 ± 1 °C. Newly grown shoots were divided into two or three parts by subculturing to a fresh medium every three to four weeks. Multiplication and rooting experiments were performed on in vitro cultures that were at least one year old (Figure 8).

Figure 5.
Map showing a distribution of Sorbus taxa in Czechia except for Sorbus aucuparia (it is widespread throughout the country). The areas of species with larger distribution were delineated with the "Sample by Buffered Local Adaptive Convex-Hull" tool [89], which combines the creation of a wrapper zone and a minimal convex polygon. Data on Sorbus species with larger distribution were derived from the Pladias database [14,90].
Plant material of the four Sorbus species, with one clone from each of the selected samples (S. × abscondita-Kladno-Švermov, an approximately 50-year-old donor tree; S. gemella-Konětopy, an approximately 10-15-year-old tree; S. × kitaibeliana-Trmice, an approximately 80-year-old tree and S. omissa-Roztoky, an approximately 10-year-old tree; Figure 6), was identified and collected from natural localities in April or May. Shoots were taken from the lower part of the crown; epicormic sprouts were not available. To obtain in vitro cultures, stem cuttings were taken (40-60 mm in length, Figure 7A,B) and surfacesterilized in 2.5% sodium hypochlorite solution (50% Savo ® , Bochemie a.s., Bohumín, The Czech Republic) for 15 min with one drop of Tween-20. Finally, the segments were washed with sterile distilled water three times for 10 min each. Cultures were initiated from apical stem segments with one apical bud and one pair of axillary buds (10-15 mm in length, Figure 7C

In Vitro Shoot Proliferation
We evaluated the effect of two cytokinins (BA, mT) alone at a concentration of 0.5 mg·L −1 and in combination with an auxin (IBA) at a concentration of 0.1 mg·L −1 , and of IBA alone at the same concentration on the induction of shoot formation in four Sorbus species. The plant growth regulator-free medium (PGR) was used as a control. Single shoots (2-3 expanded leaves, ≥ 1.5 cm long) were excised from the stock cultures and multiplied on an MS-medium containing vitamins, 20 g·L −1 sucrose and 7 g·L −1 agar. Each treatment was repeated three times, with twelve explants per treatment. The cultivation conditions for this experiment were the same as above. The mean number of shoots per explant and total shoot length were recorded after five weeks.

Ex Vitro Rooting and Acclimatization
Shoots of the four Sorbus species were harvested from an 8-10-week-old in vitro culture maintained on a basal MS medium with 0.5 mg·L −1 BA and 0.1 mg·L −1 IBA, but without PPM. For rooting experiments, microcuttings of two different shoot lengths, 15-25 mm or 26-35 mm, were used. The microcuttings were treated with rooting powder (Rhizopon ® AA, Rhizopon BV, Rijndijk, The Netherlands) at IBA concentrations of 1% or 2%.
The microcuttings were cut at the base with a sharp scalpel and immersed in 0.15% antifungal Previcur Energy (Bayer S.A.S., Lyon, France) for 1 min before the application of the rooting powder product (excluding the control). They were then inserted into a plastic dish (14 × 8.5 × 5 cm) with four holes for drainage of excess water, containing a steamed peat-perlite substrate (1:1, v/v) and watered with tap water. Each dish contained

In Vitro Shoot Proliferation
We evaluated the effect of two cytokinins (BA, mT) alone at a concentration of 0.5 mg·L −1 and in combination with an auxin (IBA) at a concentration of 0.1 mg·L −1 , and of IBA alone at the same concentration on the induction of shoot formation in four Sorbus species. The plant growth regulator-free medium (PGR) was used as a control. Single shoots (2-3 expanded leaves, ≥ 1.5 cm long) were excised from the stock cultures and multiplied on an MS-medium containing vitamins, 20 g·L −1 sucrose and 7 g·L −1 agar. Each treatment was repeated three times, with twelve explants per treatment. The cultivation conditions for this experiment were the same as above. The mean number of shoots per explant and total shoot length were recorded after five weeks.

Ex Vitro Rooting and Acclimatization
Shoots of the four Sorbus species were harvested from an 8-10-week-old in vitro culture maintained on a basal MS medium with 0.5 mg·L −1 BA and 0.1 mg·L −1 IBA, but without PPM. For rooting experiments, microcuttings of two different shoot lengths, 15-25 mm or 26-35 mm, were used. The microcuttings were treated with rooting powder (Rhizopon ® AA, Rhizopon BV, Rijndijk, The Netherlands) at IBA concentrations of 1% or 2%.
The microcuttings were cut at the base with a sharp scalpel and immersed in 0.15% antifungal Previcur Energy (Bayer S.A.S., Lyon, France) for 1 min before the application of the rooting powder product (excluding the control). They were then inserted into a plastic dish (14 × 8.5 × 5 cm) with four holes for drainage of excess water, containing a steamed peat-perlite substrate (1:1, v/v) and watered with tap water. Each dish contained

In Vitro Shoot Proliferation
We evaluated the effect of two cytokinins (BA, mT) alone at a concentration of 0.5 mg·L −1 and in combination with an auxin (IBA) at a concentration of 0.1 mg·L −1 , and of IBA alone at the same concentration on the induction of shoot formation in four Sorbus species. The plant growth regulator-free medium (PGR) was used as a control. Single shoots (2-3 expanded leaves, ≥1.5 cm long) were excised from the stock cultures and multiplied on an MS-medium containing vitamins, 20 g·L −1 sucrose and 7 g·L −1 agar. Each treatment was repeated three times, with twelve explants per treatment. The cultivation conditions for this experiment were the same as above. The mean number of shoots per explant and total shoot length were recorded after five weeks.

Ex Vitro Rooting and Acclimatization
Shoots of the four Sorbus species were harvested from an 8-10-week-old in vitro culture maintained on a basal MS medium with 0.5 mg·L −1 BA and 0.1 mg·L −1 IBA, but without PPM. For rooting experiments, microcuttings of two different shoot lengths, 15-25 mm or 26-35 mm, were used. The microcuttings were treated with rooting powder (Rhizopon ® AA, Rhizopon BV, Rijndijk, The Netherlands) at IBA concentrations of 1% or 2%.
The microcuttings were cut at the base with a sharp scalpel and immersed in 0.15% antifungal Previcur Energy (Bayer S.A.S., Lyon, France) for 1 min before the application of the rooting powder product (excluding the control). They were then inserted into a plastic dish (14 × 8.5 × 5 cm) with four holes for drainage of excess water, containing a steamed peat-perlite substrate (1:1, v/v) and watered with tap water. Each dish contained eighteen microcuttings. Three dishes were placed in a Minipa plastic box covered with a clear plastic cover with ventilation (tall lid model, Fima, Brno, The Czech Republic) and transferred to a growth room lit with cool-white fluorescent tubes (Tungsram, General Electric Company, Boston, MA, USA) and a photosynthetic photon flux density of 60 µmol·m −2 ·s −1 for a 16 h photoperiod at 24/19 ± 1 • C (day/night). Each treatment was repeated at least three times, with eighteen microcuttings, i.e., a total of 54 microcuttings per treatment. Root induction (%) and the mean number of roots per microcutting were recorded after six weeks. Then, the rooted plants were replanted into pots (Teku ® 10 cm diameter) with a peat substrate Remix-D (Rékyva, Siauliai, Lithuania) and transferred to a greenhouse at 23 • C under natural photoperiod conditions.

Statistical Analysis
Two-factor analysis of variance and Tukey's multiple comparison method were used to evaluate differences between the explants. The factors in this case were the Sorbus species and the type of treatment. The dependent variables were the mean number of shoots per explant and the total length of all shoots in the explant. The mean number of shoots per explant could be considered a variable with a Poisson distribution, so the following transformation: was applied to it [92]. The total length of all explant shoots could be considered a variable with a normal distribution with sufficient accuracy, so its transformation was not necessary. A three-factor analysis of variance (with the factors Sorbus species, treatment variant and microcutting size) followed by Tukey's multiple comparison method was used to assess differences between microcuttings. For some dependent variables, the assumption of normal distribution was violated, and various transformations were applied. Specifically, the proportion of rooted microcuttings was a variable with a binomial distribution, so the following transformation: was applied to it [92]. Again, the mean number of shoots per microcutting could be considered a variable with Poisson distribution, so it was transformed using (1). The results are presented using homogeneous groups at the 0.05 level of significance.