**4. Discussion**

In conventional breeding work including field experiments, the test for stress tolerance of breeding lines is expensive, takes a long time [46], and the results obtained in different experiments can be variable because the reaction of the plants is strongly influenced by environmental conditions [47]. That is why several new approaches are applied to evaluate traits that can be associated with drought tolerance, including laboratory mo-dels and biotechnological tools. Responses of several potato varieties to osmotic stress under in vitro conditions matched well with the drought tolerance results obtained in the field experiments [33,34]. In addition, if we take into account the very large number of offspring produced by crosses to be screened for the desired traits, the importance of alternative experiments cannot be ignored.

In field and greenhouse and also pot experiments, drought affected all physiological and agronomic traits of the potato cultivars studied [35,36]. Osmotic stress applied to model drought stress under in vitro conditions also resulted in significant changes in all studied traits in potato, e.g., survival rate, shoot length, fresh and dry weight, and root number and length [37,38]. However, no difference could be revealed between genotypes if the evaluation was based only on the measured parameters, even though significant inhibition of the growth of in vitro shoots and roots as an effect of osmotic stress in *Solanum* species was observed [32,39]. The development of a multi-parameter stress index (SI), when the results were expressed as a percentage of the values for the controls, allowed a differentiation of genotypes according to their stress tolerance [32,40]. Accordingly, a stress index was formed from our results for each treatment examined, and the examined genotypes were evaluated based on their SI values.

In our experiments, the responses of potato breeding lines were compared both to each other and to two referent lines. We observed that the osmotic stress resulted in explant deaths in varying proportions in the breeding lines and the referent lines, most often without shoot development. High survival rates were found in C2, C8, C20, C30, and C63 in the treatments with D-mannitol, in C8 and C30 in the treatments with PEG 6000, and in C8, C12, C20, C57, C58 and C63 in the treatments with PEG 600. After ranking the breeding lines based on their SI values, nine genotypes (in descending order: C8, C63, C14, C2, C5, C58, C12, C11 and C30) were found to be valuable breeding material (Tables S1–S4). However, the C103 referent line was the 1st in the ranking, while C107 was the 11th.

Considering the responses of referent genotypes to osmotic stress, they showed (sometimes significantly) different SI results. In general, SI results of C103 were higher than those of C107 in almost all treatments and for almost all traits studied.

It could be supposed from these results that drought tolerance of referent lines could be based on a different mechanism. Although osmotic adjustment ability in potato was found to be restricted to 0.16 MPa [48], its role in drought tolerance was reported for several crops [49]. Osmotic stress tolerance expressed at the tissue level may play a greater role in referent genotype C103, while other factors should be considered in the case of the C107 genotype. As drought tolerance is rather complex in nature, other factors, for example, regulation of stoma closure [16,50] and/or phenological properties (especially at time of maturity) could be relevant [14,51,52]. However, considering our results, morphological factors including root developmental characters formed under stress conditions can be of great importance [53]. All of these traits can help prevent dehydration of tissues either via reduction of water loss or increase in water uptake [14,52].

Simple morpho-physiological traits were proven to be suitable parameters for distinguishing genotypes according to their osmotic stress tolerance [32,35,41–43]. Although several biochemical markers [46,53,54] and QTLs [42] were found to be exact tools for the selection of drought-tolerant genotypes, when laboratory infrastructure and/or budget are limited, researchers are forced to use simple parameters. Therefore, we observed shoot and root length, number of roots, and survival rate and found that morpho-physiological responses of potato shoot cultures to osmotic stress included both the adaptive and—most frequently—the damage responses. Phenomena of reduced growth of shoot and root were mild physiological damage responses, while strong osmotic stress often resulted in explant death. However, several breeding lines showed increased root length and/or root number, maybe as an adaptive response that could play an important role in drought to-lerance and lead to plant escape from water stress [14].

The underground part of the plant (the root system), in addition to providing a site for fixation and water and nutrient uptake, plays a significant role in evolving abiotic stress responses [17,53]. Prolonged drought can induce adaptive responses in plants, and it can be manifested by modified structure and function of roots [53], which can play an important role in coping with water stress [19,35,55].

Although the morphology, function and even histology of roots can be highly varied in plants developed in the field, in growing containers, in hydroculture systems or in vitro [53,56], the separation of genotypes based on their rooting characters in laboratory tests can be applied in breeding work [57]. The relationship between root growth and water uptake was demonstrated by Iwama [58] under both in vitro and field conditions.

In our experiments many breeding lines responded to osmotic stress with longer and/or more root development, while others showed varying degrees of inhibited growth. In general, the lower concentrations of osmotic agent resulted in longer roots and, most frequently, more roots developed on shoots compared to the control culture. The root length SI results for the C103 referent line were very high under several osmotic treatments. Both referent lines developed the most roots under osmotic stress conditions at each level of PEG 6000, and very few breeding lines were able to outperform the referent genotypes in the PEG 600 treatments. However, SI results for root number obtained from several breeding lines were higher than those of the referent genotypes in the D-mannitol treatments.

Changes in rooting parameters under stress conditions can vary; for example, each observed rooting trait (root number and root length) decreased as D-mannitol concentration increased [32], but the root number of cv. Boró (a highly drought-tolerant variety) increased at the 0.2 M D-mannitol level. Similarly, the root dry mass increased by 25% on average in 43 potato genotypes [34]. Moreover, when Zaki and Radwan [33] tested the osmotic stress tolerance of 21 potato cultivars on media containing sorbitol at three concentration levels (0.1, 0.2 and 0.3 M), they also found that tolerant genotypes developed greater or slightly decreased root mass under stress compared to the sensitive cultivars, which suffered from strong inhibition in their root development. They also observed that stimulated root growth occurred frequently at the lowest level of stress, and sometimes it could be detected at the mid-level of stress.

Opposite results observed and published in terms of root parameters such as root length, root dry mass, and root number [19,20] may be attributed to the fact that some tolerant—varieties can respond to drought stress with increased root length, while root length does not change or decrease in more sensitive varieties compared to referent lines [36]. In addition, different experimental settings and interactions between genotypes and environment can also contribute to the variability of results [19].

In general, drought had a greater effect on the above-ground than under-ground growth in potato crop [36,55], and in the case of other crops [59–61]. In vitro experiments with potatoes yielded similar results in osmotic stress-induced changes in shoots and root systems [32,34]. In our experiments, shoot growth was also more inhibited than the growth of the root system. SI values for all treatments and all breeding lines averaged 40.06, 60.79, 76.02 and 82.61 for SL, RL, SR and RN, respectively (Table 1).

Despite a significant reduction detected in shoot length, this alone did not appear to be an appropriate trait for grouping genotypes by their osmotic stress tolerance, because high SI values of shoot lengths were not regularly accompanied by higher survival rates in breeding lines. A similar conclusion was reached by other researchers who did not recommend the use of changes in shoot length as a suitable parameter for selection, for example, in wheat [62] or in potato [32].

However, potato clones did not show the same reactions to different osmotic agents. Responses observed on media supplemented with PEG 6000 and D-mannitol were sometimes similar, but most frequently the performance of potato shoot cultures varied on different media considering the type of osmotic material as well as its concentration. In addition, the evaluation of potato genotypes on media with PEG 600 seemed to be difficult because of the very high levels of inhibition detected in breeding lines. Gangopadhyay et al. [63] also found that considering the growth, viability and proline content of tobacco (*Nicotiana tabacum* L. *var*. Jayasri), the responses of callus lines depended on physicochemical characters of the used osmotic agents (PEG 6000, D-mannitol and NaCl).

The various responses of genotypes may be due to different effects of the three osmotic agents (PEG 6000, D-mannitol and PEG 600) on the growth and developmental characters of in vitro shoot cultures. In fact, we used PEG 6000, which is a non-ionic, non-penetrating osmotic substance due to its high molecular weight [63]. In contrast, D-mannitol is a sugar alcohol that is also a non-ionic but penetrating osmotic agent [31]. PEG 600, being less than 1000 in molecular weight, is a non-ionic osmotic, but due to its lower molecular weight, it is more likely to be absorbed by plants and may have toxic effects [64].

As we observed, they all inhibited shoot growth, and the already quite strong inhibition further increased with increasing concentrations of these osmotic agents. In contrast, the root length was inhibited and also stimulated when explants were grown on media supplemented with the lowest levels of PEG 6000 or D-mannitol, whereas an inhibitory effect was observed in the case of PEG 600. Moreover, the number of roots were often significantly increased by PEG 6000 and D-mannitol treatments, but this was not true for treatments with PEG 600. Survival rates were also decreased by each treatment.

Even though PEG can result in serious stress to plants, it is used frequently as an osmotically active agent [51]. In our experiments the significantly strongest inhibition on studied characters was observed on shoot cultures grown on media with PEG 600. PEG 6000 and D-mannitol had a broadly similar inhibitory effect on plantlets, except for the number of roots, where the presence of D-mannitol resulted in a significantly lower SI value compared to PEG 6000 in the mean of all concentrations and breeding lines.

According to results reported by Thimann et al. [65], a very small amount of exogenous D-mannitol was able to enter potato disc tissues. In contrast, Trip et al. [31] found that potato leaf discs absorbed D-mannitol in a very large proportion (99%), although only 1.3% was metabolized. In potato tissue cultures, Lipavská and Vreugdenhil [66] revealed that in vitro potato shoots can readily absorb D-mannitol from the medium, and its transport to shoots was unobstructed as well. In spite of these results, in experiments involving potatoes, D-mannitol was also used to induce osmotic stress. The level that could be applied to distinguish genotypes was higher (0.8 M) for calli culture than that for shoot culture (0.2–0.4 M) [67,68]. The researchers detected strong inhibition in growth and survival with 0.4 M D-mannitol, even in the case of tolerant genotypes. Thus, we tested the effect of D-mannitol at concentrations of 0.1, 0.2 and 0.3 M, the last of which also led to strong inhibition of shoot length but had a weaker inhibitory effect on survival. Evers et al. [39] found the same tendencies when 0.2 and 0.3 M D-mannitol were applied in experiments with *Solanum phureja* and *S. tuberosum* clones.

In addition to variations in the tolerance of breeding lines to osmotic stress, there might be some differences in their ability to absorb D-mannitol, and/or tolerate the toxic effect of PEG 600; thus, interactions between genotypes and osmotic agents added to medium could lead to various responses. Significant interactions between genotypes and the degree of osmotic pressure also should be considered [40].
