Acidic soils represent 50% of the earth's arable land (Inostroza-Blancheteau et al., 2012), and in those soils aluminum (Al) toxicity is the main factor restricting crop productivity (Kochian, 1995). In acidic soils with pH lower than 5, Al-containing minerals (e.g., aluminosilicates) are solubilized in the phytotoxic form Al³⁺ (Kochian, 1995; Seguel et al., 2013). The root is the first organ feeling Al³⁺ toxicity, thus conditioning stress sensitivity and hampering crop productivity (Delhaize and Ryan, 1995; Rengel et al., 2015). For many plant species, Al concentration in the range of 1–2 mg L⁻¹ can inhibit root elongation by damaging the cell structure of the root apex and thus influencing water and nutrient uptake (Kochian et al., 2015; Rengel et al., 2015). Al also interferes with cell membrane stability, enzyme function, and the synthesis of chlorophyll (Simon et al., 1994; Rouphael et al., 2015). Several authors reported that exposure to Al³⁺ in the rooting medium strongly inhibit Ca²⁺ and Mg²⁺ fluxes across the plasma membrane of root cells (Rengel and Elliott, 1992; Horst et al., 2010; Bose et al., 2011). Similarly, Al toxicity can decease or even block the uptake of K⁺, since it interacts with various different plasma-membrane channel proteins (Kochian et al., 2005). This may decrease cations uptake leading to nutrient deficiencies, thus affecting metabolism and productivity of crops. However, plant species have different degrees of adaptation to Al; these differences are often related to the specific mechanisms developed for mitigating stress (Panda and Matsumoto, 2007). The mechanisms conferring tolerance to Al have been classified into mechanisms of exclusion and those of intracellular tolerance (Kochian, 1995). Exclusion mechanisms not allowing Al entering the symplast can take place by exudation of Al-chelating compounds (e.g., organic acids), raising the pH in the rhizospheric environment and by binding of Al in the cell wall itself (Inostroza-Blancheteau et al., 2012; Brunner and Sperisen, 2013; Kochian et al., 2015). Internal tolerance mechanisms include those that bind the Al entering the root cells as well as the sequestration of Al in subcellular compartments (e.g., vacuoles).

To overcome the limitations of Al toxicity several solutions have been proposed. Amendments such as gypsum, lime, and phosphate fertilizers are common agricultural practices used to raise the soil pH causing the Al to become insoluble (Nawrot et al., 2001). Liming is only efficient in the topsoil, while it does not remedy the subsoil acidity (>1 m), since deep incorporation of lime is difficult to realize and very costly (Choudhary and Singh, 2011; Yang et al., 2013). Another possible solution to alleviate the negative effects of acid soils would be the use Al-tolerant cultivars obtained by breeding and/or biotechnological approaches (Choudhary and Singh, 2011). However, the long time needed for the breeding of Al-tolerant cultivars as well as the complexity of the Al-trait make this task extremely difficult. In order to overcome this situation more rapidly, grafting elite vegetable cultivars onto rootstocks tolerating higher Al concentrations and low pH could be a key tool in alleviating the effects of acidity and Al toxicity on crop productivity.

By selecting suitable rootstocks, grafting can control scion morphology (Albacete et al., 2015), increase productivity and fruit quality (Colla et al., 2008; Proietti et al., 2008; Kyriacou et al., 2016), improve nitrogen use efficiency (Colla et al., 2010a, 2011), and also induce tolerance against several abiotic stresses, among them nutrient toxicity, heavy metals, and alkalinity (Edelstein et al., 2005, 2007; Rouphael et al., 2008a,b; Colla et al., 2010b; Savvas et al., 2010, 2013; Kumar et al., 2015a). Therefore, we hypothesized that the effectiveness of root genotypes to enhance the uptake of nutrients and to limit the Al accumulation in aerial parts may be improved by grafting onto suitable rootstocks.

To verify the above hypothesis, two hydroponic greenhouse experiments were conducted (1) to assess the effects of grafting combinations on morphological traits of cucumber at early development stage in response to different nutrient solution pH and Al concentrations; and (2) to elucidate the agronomical, physiological, and mineral composition changes of cucumber mediated by grafting under acidity and Al toxicity conditions using long-term treatments.

Under acidic soils, reduced plant growth and consequently productivity are induced by different morphological, biochemical, and physiological alterations (Kochian et al., 2015; Rengel et al., 2015). Rapid damage and growth inhibition of the root system is the primary symptom of Al toxicity at concentrations in the μM range, which may be associated with interference in the elongation and division of meristematic cells (Rengel and Zhang, 2003; Kochian et al., 2004; Sivaguru et al., 2013). This interference reduces the absorption of water and nutrients, thus conditioning stress sensitivity and limiting plant biomass production (Inostroza-Blancheteau et al., 2012). In agreement, in both experiments a significant reduction in root, shoot biomass, and yield was observed at low pH (3.5) and even more, when Al was added (Figure 1; Tables 1, 2), suggesting that these two stresses differ in their inhibition on plant growth (Rangel et al., 2005; Bose et al., 2010). Moreover, the reduction in plant biomass production depended on the Al concentration. The constrained crop growth and fruit yield observed under pH 3.5 and under pH 3.5 + Al has been reported in previous open field and greenhouse studies on tomato (Nogueirol et al., 2015) and zucchini squash (Rouphael et al., 2015). Nevertheless, decreasing pH and increasing Al concentration in the nutrient solution improved fruit quality characteristics such as fruit DM and TSS contents (Table 3). These phenomena have been found before for zucchini and explained by a decrease in water accumulation by the fruit without any effect on the synthesis and accumulation of organic solutes (Rouphael et al., 2012, 2015).

When cucumber cv. Ekron were grafted onto pumpkin rootstock “P360” compared with non-grafted and grafted onto figleaf gourd, the suppression of growth and production suppression under Al-supply was mitigated indicating a genetic diversity in Al-tolerance among cucumber rootstocks. The Al-tolerance of cucumber grafted onto pumpkin “P360” may be attributed to a greater uptake and translocation of K, Ca, Mg, Mn, and Zn to the aerial parts (leaves and stems). The enhanced nutritional status and abiotic stress-tolerance of grafted vegetables has often been associated with an enlarged and more vigorous root system (Colla et al., 2010c; Savvas et al., 2010). The root is the first organ sensing abiotic stresses in soil or substrates, such as acidity and Al toxicity. Therefore, the enhanced root growth (Table 1), length, and density (data not shown) of a pumpkin compared with a cucumber rootstock are important aspects for cucumbers cultivated in acidic conditions with or without Al toxicity.

Al accumulates in the epidermis and in the outer cortex of the root (Delhaize et al., 1993; Ryan et al., 2001). Many studies demonstrated that the major fraction of absorbed Al (30–90% of total Al) is localized in the apoplast (Yang et al., 2000; Ryan et al., 2001; Pereira et al., 2010). In Experiment 2, Al concentration in the root tissues was 50 times higher than in the cucumber leaves and even higher than in stems or fruits (Table 4). This indicates that the endodermis of roots represents a barrier to the transport of Al to the shoot (Dogan et al., 2014). Interestingly, this ratio was independent of the type of rootstock used although the absolute Al concentration differed between grafted and non-grafted root systems. The different Al accumulation in plant tissues highly depends on the mechanism (exclusion or internal tolerance) used by plants to confer tolerance to Al toxicity (Delhaize and Ryan, 1995; Kochian, 1995). Our results showed that Al concentration in cucumber roots was similar between grafted (E/C and E/F) and non-grafted plants grown under Al-stress (Table 4). This indicates that neither the figleaf gourd nor the pumpkin rootstocks were able to exclude Al from the root apex by exudation of organic acids (citrate, malate, and oxalate) into the rhizosphere (Brunner and Sperisen, 2013). On the other hand, the lowest Al accumulation in leaves with pH 3.5 + Al was recorded in the E/C combination compared to the other grafting combinations (Table 4) indicating that another mechanism was involved in mitigating the Al toxicity effect (Barceló and Poscheinrieder, 2002; Brunner and Sperisen, 2013). It appears that grafting cucumber onto pumpkin rootstock can restrict the Al root-to-shoot translocation (Table 4), throughout “the sequestration of this toxic element into less sensitive parts of the plant and cell compartments (e.g., vacuoles)” (Inostroza-Blancheteau et al., 2012; Brunner and Sperisen, 2013; Kochian et al., 2015). This suppression of Al uptake and translocation to shoots of grafted plants is consistent with previous reports indicating that appropriate rootstocks of cucurbitaceous and solanaceous species may limit the uptake of toxic elements and heavy metals (Edelstein et al., 2005; Rouphael et al., 2008a; Kumar et al., 2015b). Authors show that grafting annual fruit crops, such as melon, cucumber, and tomato onto vigorous rootstocks (i.e., pumpkin for cucurbits and “Maxifort” for tomato) may significantly restrict trace elements' (B and Cu) and heavy metal (Ni) concentrations in leaf and fruit tissue, thereby alleviating their detrimental effects on crop productivity and human health via the food chain (Savvas et al., 2010).

Nutrient uptake is a crucial aspect for the maintenance of homeostasis and plant growth under unfavorable soil conditions (i.e., edaphic stress; Seguel et al., 2013). Tolerance of Al is mostly reflected in limited alterations of macro and micronutrient's acquisition and translocation (Andrade et al., 2009). In the present study, decreasing pH level and increasing Al concentration in the nutrient solution depressed concentrations of macro- (N, P, K, Ca, and Mg) and micro-elements (Mn and Zn) in shoots (leaves and/or stems) and to a lesser degree in roots in both grafted and non-grafted plants (Tables 5, 6). Therefore, excessive Al accumulation affected uptake, translocation and accumulation of these nutrients in plant tissues, and hence was responsible for mineral imbalances and deficiencies as well as for the depression of plant growth and yield (Rouphael et al., 2015). Studies have demonstrated that Al interferes directly with several different plasma membrane channel proteins, thus reducing the uptake of mono- and divalent cations such as K and Ca (Gassmann and Schroeder, 1994; Piñeros and Tester, 1995). Al-stress has been reported to reduce Ca through three mechanisms: “(1) inhibition of Ca transport via symplasm by Al, (2) disruption of Ca homeostasis in cytoplasm by Al, and (3) Ca displacement by Al from apoplasm” (Delhaize and Ryan, 1995; Kochian, 1995). In addition to the reduction in K and Ca uptake, Al can also decrease Mg uptake in aerial parts (Rengel and Robinson, 1989), as we can confirm with our current data. The competition between these two cations (Al and Mg) was demonstrated for membrane transporters and metal binding sites in enzymatic reactions (Rengel and Robinson, 1989; Pécsváradi et al., 2009).

However, grafted and non-grafted plants diverged in their tissue nutrient concentration in relation to acidity and Al toxicity. Grafting cucumber plants onto the pumpkin rootstock (E/C) increased K, Ca, Mg, Mn, and Zn in shoot and/or root (Tables 5, 6) suggesting that selecting appropriate rootstocks can maintain nutrient homeostasis under Al-stress. The lower reduction in cation uptake observed in the E/C combination under Al-stress, could be associated to the lower inhibition of root growth and elongation since cation (especially K) accumulation contributes to the expansion of cell elongation (Cristancho et al., 2011). Similarly, the shoot and root Mg concentrations were higher in E/C compared to the other grafting combinations. It remains to be elucidated whether higher Mg uptake in E/C combination is a consequence of a greater root growth and/or greater Al tolerance of Mg-uptake systems (Rengel et al., 2015). From above we could suggest that an efficient metabolism system exists in E/C combination under Al toxicity, indicating that Al tolerance is correlated with enhanced mineral nutrient concentrations in plant tissues (Rouphael et al., 2015). These results are consistent with several studies (Rouphael et al., 2008a; Kumar et al., 2015a,b) reporting that rootstocks used in grafting of annual fruit crops were able to enhance the uptake of some macro- and micro-elements under heavy metal (e.g., Cd, Ni, and Cu) toxicity.

High Al concentrations in soil or substrates hamper plant development at a physiological and biochemical level, affect its photosynthetic rate, total chlorophyll content, and also inhibit electron transport in PSII (Chen, 2006; Inostroza-Blancheteau et al., 2012). This was confirmed with our measurements of the SPAD index, widely used as a non-destructive estimate of chlorophyll content. It dropped sharply in leaves of Al stressed non-grafted plants compared to both E/C and E/F plants (Table 7). This suggests the occurrence of chlorophyll degradation and early senescence, likely due to the harmful effects of reactive oxygen species on chloroplasts (Chen, 2006; Rouphael et al., 2015). Contrarily, the grafted cucumber plants in particular E/C combination were able to maintain a higher chlorophyll content in both stress treatments, thus exhibiting the highest yield and biomass. In addition to a reduced chlorophyll content, reductions in total leaf area at low pH and especially at Al-supply could be caused by a premature leaf senescence and ion toxicity (Seguel et al., 2013). Maintenance of a large leaf area upon Al stress in E/C combination (Table 7) may be crucial to guarantee production and translocation of photosynthates to the fruits, thus increasing the final yield in cucumber.

Adverse pH conditions and Al-stress induced impairment of membrane integrity and affect all physiological activities linked to membrane functioning (Rouphael et al., 2015). This is related to the fact that Al-enhanced oxidative stress caused by an increased production of reactive oxygen species leads to lipid and protein oxidation (Boscolo et al., 2003; Jones et al., 2006). The results of Experiment 2 indicate that in cucumber the electrolyte leakage percentage (i.e., degree of cell membrane injury) was significantly increased, and depended on the nutrient solution pH and the Al concentration (Table 7). The degree of cell membrane injury induced by Al-stress has often been related to calcium concentration, an important factor in increasing structural stability of cell membrane (Borer et al., 2005). Increasing Al concentration in the nutrient solution reduced leaf and root calcium uptake resulting in a reduction of cell membrane integrity (Pereira et al., 2010). However, E/C plants reduced the amount of ion leakage in Al-stress treatment by improving the Ca uptake in shoot and root tissues indicating that grafting cucumber onto pumpkin rootstocks has facilitated the membrane functions (i.e., semipermeability).

Our findings indicate that in both grafted and non-grafted plants, agronomical, physiological, and mineral composition responses were negatively affected by acidity and Al concentration, with Al-stress being more phytotoxic than low pH treatment. The results of both experiments were able to verify our hypothesis that grafting onto suitable rootstocks (i.e., pumpkin) may limit the Al accumulation in the aerial parts, improve the uptake of nutrients (K, Ca, Mg, Mn, and Zn), enhance chlorophyll synthesis as well as the root genotypes to control the cell membrane stability; thus alleviating the impacts of adverse pH level and Al toxicity on crop productivity. These results might be useful to assist the selection of tolerant rootstocks in breeding programs particularly in acidic soils, where Al-toxicity is a major agronomic constraint.

YR performed the long term experiment with agronomical and physiological analysis and he was involved in writing the manuscript (Experiment 2); ER and MC performed the mineral analysis in both experiments and they collaborated in manuscript preparation; MB and DS were involved in the short term experiment (Experiment 1) and in manuscript preparation; GC defined the scientific hypothesis, set up the experimental protocols, made the statistical analysis of experimental data, and he was involved in the manuscript preparation.