Iron (Fe) is the fourth most abundant element in the Earth's crust, but about 30% of the world's cultivated soils are classified as calcareous and deficient in Fe for plant growth (Guerinot and Yi, 1994; Barton and Abadía, 2006; Hansen et al., 2006; Kim and Guerinot, 2007). The principal cause of Fe deficiency is that Fe occurs mainly in the form of water-insoluble hydroxides and oxides and/or carbonates–bicarbonates complexes that are not bioavailable for uptake by plant roots (Guerinot and Yi, 1994; Hell and Stephan, 2003; Zuo and Zhang, 2011). Other factors have also been linked to Fe deficiency. These include soil salinity, low temperature, high pH, carbonates, high moisture content, low water drainage, soil bulk density, high nitrate concentration, interactions of Fe with other soil minerals, landscape position across and within the same farmland (Hansen et al., 2006; Zuo and Zhang, 2011; Vasconcelos and Grusak, 2014). Reduced Fe²⁺availability results in chlorosis, stunted growth, reduction in crop quality and yields and a decrease in nutritional value of many crops (Naik et al., 1990; Abadia et al., 2011; Hindt and Guerinot, 2012; Bert et al., 2013; García-Mina et al., 2013).

Significant progress has been made in deciphering key players of the transcriptional networks that control Fe homeostasis, and in understanding the genetic regulation of Fe uptake mechanisms in plants (Vert et al., 2003; Briat et al., 2007; Ivanov et al., 2012). The iron-regulated transporter (IRT1) and ferritin (FER) are proposed to be involved in iron transport, storage and protection against oxidative stress (Eide et al., 1996; Robinson et al., 1999). Genes transcriptionally regulated under conditions of low-Fe bioavailability synchronize iron uptake, translocation and storage in plant cells (Legay et al., 2012; Reyt et al., 2015). Iron-associated genes seem to play a vital role in the molecular responses to Fe deficiency-associated stress and tolerance in plants (Li et al., 2004; Ivanov et al., 2012; Legay et al., 2012; Darbani et al., 2013). Changes in the root transcriptome and morphology of Strategy I plants, for example, potatoes (Bienfait et al., 1987; Ivanov et al., 2012; Legay et al., 2012) in response to iron deprivation have been reported. Also, genes involved in iron acquisition are upregulated in response to Fe deficiency (Hindt and Guerinot, 2012; Kobayashi et al., 2012).

To reduce the agricultural and economic impact of iron deficiency stress, it has become necessary to develop an iron-efficient plant that is highly capable of optimum use of Fe for its metabolism. Such a plant can take up a smaller amount of Fe compared to a less efficient plant and produce the same yield (García-Mina et al., 2013; Vasconcelos and Grusak, 2014). Selection for Fe-efficiency in plants could enhance the use of marginalized calcareous soils while limiting Fe fertilizer usage and subsequently limit the economic losses associated with Fe deficiency. Biotechnological tools can be applied to develop plants that are adapted to Fe deficiency stress. In vitro plant cell selection is a cost-effective and practical tool for the selection of stress-tolerance in plants. The direct in vitro selection strategy prevents the development of epigenetically adapted cells (Chandler and Vasil, 1984; McHughen and Swartz, 1984; Tal, 1994). In vitro selection by indirectly inducing Fe deficit with incorporation of CaCO₃ in tissue culture medium resulted in the generation of Fe-efficient lines of sugarcane plants (Naik et al., 1990). Palombi et al. (2007) employed in vitro regeneration to obtain and establish somaclones highly tolerant to calcareous soils. Two somaclonal variants recovered from shoots regenerated from quince leaves exposed to limited Fe supplies were found to be Fe-efficient (Dolcet-Sanjuan et al., 1992). Vasconcelos and Grusak (2014) identified Fe-efficient soybean plants and confirmed their tolerance to Fe-deficiency-induced chlorosis (IDC) under laboratory and field conditions. Comparative studies of iron-efficient (EF) and iron-inefficient (IFN) plants including investigations in relation to genes linked to iron homeostasis should lead to a better understanding of the molecular mechanisms governing stress responses to Fe-deficiency.

The generation of novel Fe-efficient potato variants in vitro is beneficial for crop improvement and can be used to complement conventional breeding. In vitro selection for Fe-efficiency trait in potatoes is valuable given the economic importance of potatoes and the prevalence of Fe-deficiency worldwide (Hindt and Guerinot, 2012; King and Slavin, 2013). However, there is a paucity of research to develop Fe-efficient potatoes based on in vitro plant cell selection. In this study, in vitro plant cell selection was applied to obtain EF and IFN Solanum tuberosum (cv “Iwa”) plant lines. The aim of this study was to gain insights from a comparison of the chlorosis symptoms and morphological characteristics of these novel plants and the parental biotype grown under in vitro and Fe deficiency conditions. Further insights were gained with an investigation into the transcriptional responses of the novel plant lines with differential tolerance to IDC. In particular, the expression of two key marker genes related to iron homeostasis, irt1 and fer3, were investigated using RT-qPCR. It was hypothesized that the expression levels of irt1 and fer3 would be dissimilar in EF, IFN potato plant lines and the parental biotype.

Fe²⁺ deficiency can restrict plant growth and yield. Development of Fe-efficient plant varieties is suggested to be the best practical means to correct IDC and avert plant growth restrictions due to Fe²⁺ deficiency in soils (Hansen et al., 2006; Helms et al., 2010; García-Mina et al., 2013; Vasconcelos and Grusak, 2014; Zhu et al., 2015). This is the first report on in vitro selection of Fe-efficient potato plant lines. In the present study, low concentrations of Fe in culture medium were used as selective pressures. This direct selection scheme was designed to recover the preferential survival of somaclonal variants adapted and tolerant to the Fe-deficiency stress applied. Therefore, the advantage of the present approach over the use of CaCO₃ as a selection agent in other previous studies was to avoid uncovering variations related only to CaCO₃ per se (Naik et al., 1990; Dolcet-Sanjuan et al., 1992; Palombi et al., 2007; Kabir et al., 2015).

The initial selection by visual screening enabled the identification of IDC-tolerant callus lines over sensitive ones based on differences in pigmentation and cell survival. When cells experience trauma or stress, they may reset genome expression (Phillips et al., 1994). Such reprogramming can give rise to somaclones with altered gene expression in plants arising from point mutations, gene amplification, chromosomal rearrangements and recombination, DNA methylation, ploidy change, altered sequence copy number, activation of transposable elements, and chromatin loss (Phillips et al., 1990; Rao et al., 1992; Kaeppler et al., 2000; Kumar and Mathur, 2004; Pérez-Clemente and Gómez-Cadenas, 2012).

The putative Fe-efficient potato plant lines obtained from selected callus cultures were subjected to Fe-deficiency conditions in vitro. In this way, the stability of the trait was determined so that the most promising Fe-efficient plant lines could be selected for further plant germplasm evaluation and breeding. The main advantage of testing the plant lines under Fe-deficiency conditions was that IDC-tolerant plant lines could be identified and distinguished from IDC-susceptible plants. Similar to the study of Vasconcelos and Grusak (2014), plants with an average VCR of ≤2.42 or ≥3.50 were designated as chlorosis-tolerant (i.e., Fe-efficient: EF) or chlorosis-sensitive (Fe-inefficient: INF) lines, respectively. Here, the chlorosis tolerance trait was found to persist in the selected EF plant lines over long-term subculturing under Fe-deficiency stress conditions. The selected Fe-efficient plant lines had mild to no symptoms of chlorosis and grew better on Fe-deficient medium than chlorosis-susceptible plants (IFN and controls). Likewise, in M. ciliaris, Fe deficiency had a more pronounced impact (severe chlorosis) on IDC-sensitive lines than tolerant ones (M'sehli et al., 2014).

The morphological characteristics of the Fe-efficient potato plants obtained were similar to those observed in other studies. For example, Fe-efficient soybean plants were observed to be shorter compared to inefficient ones (Elmstrom and Howard, 1969; Vasconcelos and Grusak, 2014). Similarly, the stem height of both chlorosis-sensitive and tolerant chickpea varieties exposed to Fe deficiency was markedly lower than the respective controls (Mahmoudi et al., 2009). Under Fe deficiency conditions, shoot height, fresh weight and growth were reduced significantly in chlorotic compared to non-chlorotic plants (M'sehli et al., 2008; Kabir et al., 2015). Morphological variations in leaf development and size have been linked to Fe deficiency chlorosis. Chlorotic leaves exhibited significantly reduced leaf expansion, size, fresh, and dry weight (Larbi et al., 2006; Fernández et al., 2008) compared to non-chlorotic leaves under Fe-sufficient conditions (Maldonado-Torres et al., 2006). Fernández et al. (2008) proposed that IDC can prevent leaf development via changes in leaf cuticle and cell wall. They showed that green leaves had thick and homogeneous cell walls while cells in chlorotic leaves appeared as thin, discontinuous and heterogeneous. Fe-deficiency can modify the barrier properties of the leaf surface, which can have a marked effect on leaf water relations, solute permeability and pest and disease resistance (Fernández et al., 2008). Fe shortage results in reduced leaf growth, leaf number, leaf surface area, leaf biomass, loss of turgor and the inhibition of the formation of new leaves (Kosegarten and Koyro, 2001; Mahmoudi et al., 2009; Kabir et al., 2015). Poor leaf growth and suppressed leaf formation are characteristic symptoms of IDC response due to the high sensitivity of the meristematic apex to low iron availability (Gruber and Kosegarten, 2002; Vasconcelos and Grusak, 2014).

In strategy 1 plants (e.g., potato), Fe-deficiency is associated with inhibition of root elongation, increase in diameter of apical roots, subapical swelling with abundant root hairs and formation of transfer cells (Bienfait et al., 1987; Broadley et al., 2012). The most efficient way of increasing absorbing root surface area is in the formation of root hairs. In the current study, the increased number of lateral roots and root hairs identified in the potato plant lines especially the Fe-efficient ones are proposed to be necessary for a more efficient exploration of the medium for the acquisition of Fe. Also, most EF plant lines had reduced root lengths than IFN lines. This is contrary to findings in previous reports on other plants which were evaluated under different conditions. Increase and decrease in root biomass was observed in M. ciliaris chlorosis tolerant and sensitive ecotypes, respectively (M'sehli et al., 2008). Similarly, a significant decline in root length and fresh weight was observed in a chlorosis-susceptible okra variety compared to a tolerant variety under low Fe availability (Kabir et al., 2015). Although root biomass was unaffected by Fe in medium, reduction in dry weight and root biomass was more severe in chlorosis-susceptible chickpea varieties than tolerant ones in comparison to their respective controls under bicarbonate-induced Fe deficiency (Mahmoudi et al., 2009). Fe-deficient plants developed more lateral roots and abundant root hairs (Licciardello et al., 2013). Improvement in root hair formation in Arabidopsis mutant was regulated by the Fe concentration in the growth medium (Schikora and Schmidt, 2001). In the present study, the Fe-efficient plant lines that exhibited increased lateral roots, root hairs and decreased length could be of interest for further evaluation in the field. In tomato (Zamboni et al., 2012) and potato (Bienfait et al., 1987), enhanced root hair proliferation is a morphological adaptation linked to suboptimal Fe conditions. When plants growing under Fe deficient conditions were resupplied with Fe, morphological adaptations exhibited by the root were diminished and transfer cells became degenerated (Broadley et al., 2012).

Plant ferritin plays an essential role in the maintenance of iron buffering, oxidative stress prevention and adaptation to adverse environmental situations (Ravet et al., 2009; Briat et al., 2010; Parveen et al., 2016). According to Zok et al. (2010), increased production of ferritin improved abiotic stress tolerance in transgenic grapevine plants. In response to excess Fe supply there was an up-regulation of ferritin (fer1) transcript in chickpea seedlings (Parveen et al., 2016) and fer3 in potato leaves (Legay et al., 2012). In transgenic chickpea plants, overexpression of ferritin (fer1) was found to confer enhanced growth (Parveen et al., 2016). An interesting finding from the present in vitro cell line selection study is that the expression of fer3 was significantly higher in the leaves of predominantly IDC-tolerant potato plant lines compared to IDC-sensitive plants under Fe deficient conditions. A plausible explanation for this result could be that the induction of fer3 expression due to low Fe availability possibly serves as a response mechanism to improve Fe acquisition in IDC- tolerant plants. The tissue culture-derived variant IDC-tolerant plants might have this as an adaptive mechanism to enable survival under Fe-deficiency stress conditions. This remains to be elucidated further, however. Ferritin and Fe are co-localized primarily in chloroplasts of leaves storing Fe where photosynthesis is active (Briat et al., 2006, 2010; Darbani et al., 2013). It is not surprising that more pronounced ferritin expression was found in leaves than in roots. Elevated fer3 expression in the putative Fe-efficient lines may be connected to reduced chlorosis observed in EF lines which had high leaf chlorophyll contents (Boamponsem, 2017).

In the root, the functional role of ferritin or its involvement in oxidative stress responses has been studied less thoroughly and some findings are debatable (Vigani et al., 2013; Reyt et al., 2015). Santos et al. (2015) have reported that Fe-inefficient plants (at an early growth stage) had higher induction levels of the ferritin gene in the roots than Fe-efficient plants. Vigani et al. (2013) reported that ferritin was abundant in mitochondria of the roots of cucumber plants grown under Fe-excess. However, lower ferritin levels were detected in roots of the cucumber plants grown in medium with sufficient Fe. Reyt et al. (2015) also reported that the ferritin gene, fer1, was most highly expressed in Arabidopsis roots in response to excess Fe. Low-iron availability led to the repression of fer3 expression in the roots of potato plantlets whereas in high Fe medium, fer3 expression was significantly higher (Legay et al., 2012). In the present study, the level of fer3 expression was strongly increased in the roots of some of the Fe-efficient and Fe-inefficient lines and control plants. However, fer3 expression was markedly reduced in other lines. It appears, therefore, that differential tolerance to IDC in potato plants may not be linked to fer3 expression levels in the root.

The literature on irt1expression in leaves is scanty. Earlier studies by Vert et al. (2002) have shown that irt1 expression is noticeable in leaves and not restricted to the roots of Arabidopsis. The expression of irt1 gene was detected in the rosette leaves and found to be expressed in the basal part of flowers, suggesting its role in Fe acquisition in aerial tissues. Here, irt1 expression in leaves was generally low, but significantly high levels were identified in the A1 line, the A2 Fe-inefficient line and the parental control plant. Similarly, Santos et al. (2015) indicated that IRT1-like gene expression was high in leaves of both EF and IFN soybean plants under Fe-liming conditions. Contrarily, in examining plants' response to Fe supply, it was observed that irt1 mRNA was reduced in leaves of plants grown in iron-limited conditions (Barberon et al., 2011; Legay et al., 2012).

Vert et al. (2002) observed severe leaf chlorosis typical of iron-deficient plants in the Arabidopsis knockout mutant (irt1-1). In the present work, not enough evidence has been found to imply that high chlorophyll content in leaves of potential Fe-efficient potato lines may be directly related to the high expression of irt1 in leaves. This pinpoints a possible lack of correlation between visual chlorosis symptoms and the molecular mechanisms associated with leaf iron transport. Legay et al. (2012) found a strong increase in irt1 expression in roots of potato plants grown in low Fe-containing medium. In contrast, irt1 expression was decreased when the potato plants were grown under high Fe excess. In the present study, expression levels of irt1 were significantly higher in roots than in leaves. Fe deficiency induced changes in irt1 transcript levels largely in roots of potato plant lines with differential tolerance to Fe-deficiency. This is consistent with the notion that irt1 has a role in Fe uptake in potatoes (Figure 8).

In the roots of 50% of the potential Fe-efficient plant lines derived from in vitro selection here, irt1 expression was enhanced as a consequence of Fe deficiency (Figure 8). Additionally, irt1 was expressed at increased levels in the roots of both IDC-tolerant and -sensitive potato plant lines with a more pronounced effect detected in the IDC-tolerant line, A1. These findings are consistent with earlier reports on irt1 expression in okra plants with differential IDC tolerance (Kabir et al., 2015). Irt1 seems to play a vital role in the response of potato plants exposed to Fe-deficiency conditions by enhancing Fe-transport capacity, mainly in the roots in agreement with previous reports (Vert et al., 2002; Legay et al., 2012; Kabir et al., 2015).

There have been conflicting reports and divergent conclusions concerning the molecular responses of EF and IFN plants under Fe limiting conditions. Some studies have shown that the expression of genes associated with Fe-deficiency was increased in EF plants (Kabir et al., 2015). However, IFN plants have also been found to exhibit enhanced expression of these genes (Martínez-Cuenca et al., 2013; Santos et al., 2015). In the present study, it was also not possible to establish specifically that Fe-efficiency might be attributable to higher irt1 expression in the root as about the same number of EF lines exhibited either significantly increased or decreased irt1 expression. Furthermore, some chlorosis- sensitive lines (A2 and C5) showed greater irt1 expression levels than other tolerant lines (E1-2 andE7). This is similar to the findings that irt1 expression in chlorosis-sensitive citrus rootstocks was at least two times higher than the tolerant ones (Martínez-Cuenca et al., 2013). Furthermore, Santos et al. (2015) noted that the expression of irt1-like gene was higher in the roots of IFN plants than in the EF ones. The considerably higher gene expression levels (mainly irt1) measured in control plant roots and leaves compared to 75% of the plant lines may have been due to a rapid or shock response of control plants to Fe deficiency stress. In contrast, the selected plant lines did not exhibit changes in expression of the genes as they might have adapted to Fe deficiency conditions.

In conclusion, a novel set of EF and IFN potato plant lines has been developed which can be useful in further research, serve as germplasm and/or aid in breeding programs aiming at selecting IDC tolerant potato genotypes. Taken together, the results showed that 37.5% of the putative EF plant lines (A1, B2, and B9) have the capacity to elicit improved morphological attributes including enhanced formation of lateral roots and root hairs under low iron supply conditions. In addition, the EF plant lines were found to exhibit enhanced expression of two Fe homeostasis-related genes, fer3 in the leaf and irt1 in the root. These morphological and molecular changes seem to be important for resistance to Fe deficiency stress. The use of these novel plant lines could contribute to enhancement of crop productivity in calcareous soils.

GB carried out the experiments and wrote the initial drafts of the paper. DL was the senior supervisor and CL was the associate supervisor. DL also wrote and edited the final draft of the paper.