*3.8. Principal Component Analysis*

The first two principal components gave eigenvalues equal to 2.75 (PC1) and 0.21 (PC2), and together accounted for 98.9% of the total variance (Table 8). The As accumulation into the different plant fractions positively contributed to PC1, with correlation coefficient ranging from 0.556 to 0.589, while PC2 was strongly and positively correlated to Asroot (0.830) and negatively correlated to Asshoot and Aspulp (−0.367 and −0.419, respectively). The resulting PCA scatterplot showed a clear separation among plants grown in As− (on the negative side of PC1) and those grown in As+ (on the right side of PC1). The first group clustered together, with no separation between different plant organs, while the second group showed a further separation in two sub-clusters, in the second and fourth quadrant of the centroid, respectively (Figure 1). The first sub-cluster grouped on the positive side of PC2 the grafting combinations (self-grafted "Proteo", "Proteo"/"Magnus", and "Proteo"/"Dinero") accumulating As mainly in root, while the second sub-cluster grouped those combinations ("Proteo"/'Strong Tosa', "Proteo"/"Shintoza", and "Proteo"/"RS841"), accumulating As in the aerial parts of the plant, namely shoot and pulp (Figure 1).

**Table 8.** Correlation coefficients for each trait with respect to the first two principal components, eigenvalues, and relative and cumulative proportions of the explained variance.


**Figure 1.** Principal Component Analysis scatter-plot based on total As concentration in root, shoot and pulp of melon plant "Proteo" as a function of the grafting combination and As concentration of the nutrient solution. Dotted lines group together treatments at different concentrations of As in the nutrient solution, while solid lines group together the grafting combinations.

#### **4. Discussion**

The environmental As contamination is a worldwide health threat due to the toxic and carcinogenic nature of this metalloid [29]. From an agricultural viewpoint, plants exposure to As− stress can cause morphological, physiological, and biochemical changes, leading to altered photosynthesis, stunted growth, reduced crop productivity, and worsened toxicological profile of the edible fractions [30]. In our experiment, the bioaccumulation factors values revealed that, even at a low As concentration in the nutrient solution, melon behaves as an hyperaccumulator plant, following the gradient root > shoot > fruit generally reported in literature [31]. However, when exposed to an As+ solution, all the grafting combinations acted to limit the As entrance into the plant, as can be inferred from the average reduction of BAFroot (−94.4%), BAFshoot (−98.1%), and BAFpulp (−99.6%). Nonetheless plants exposure to As stress highlighted evidences of systemic stress, consisting in a decrease of the whole aboveground dry biomass per plant, which mainly mirrored the reduction of both, stem and leaf dry biomass per plant. The reduced plant growth in response to As toxicity originates from complex alterations involving enzymes activity, induction of oxidative stresses, or altered nutrient uptake and balance into the plant [32,33]. In the present experiment melon plants subjected to As stress showed a decreased ability to maintain the normal growth equilibrium among plant fractions, by partially losing (~20%) the ability of photosynthates investment into leaf biomass, so indicating the main photosynthetic organs as an elective target of As-induced alterations. Leaf area, leaf area ratio, and leaf weight ratio were predominantly affected by the As toxicity, more than leaf number, indicating leaf cells proliferation and elongation as primarily affected by the As stress, more than leaf cells differentiation. According to Koyama and Kikuzawa [34] and Ropokis et al. [35], leaf area is a trait having a central role in determining the level of nutrient uptake, via the rate of whole plant photosynthesis and transpiration. In this sense, the decreased leaf area we recorded suggests the triggering of morpho-physiological modifications reducing the As entrance inside the plant, probably by modulating the plant's transpiration and nutrient demand.

Regarding photosynthesis, As accumulation in leaf tissues is responsible for key physiological events such as chlorophyll degradation and leaf necrosis, decreased activity of the enzymes involved in photosynthetic metabolism, disturbance of photosynthates transport, and stomatal behavior [36,37]. This, apparently, was not reflected in our experiment where, besides the absence of any leaf necrosis (data not shown), there was a general increase in leaf relative chlorophyll content, as can be inferred from SPAD readings. This result is consistent with those of Zu et al. [37], which have reported that As is able to promote chlorophyll content in leaf tissues, becoming harmful only above a threshold concentration.

The analysis of net photosynthetic rate revealed different degrees of sensibility to the As toxicity among grafting combinations, with self-grafted "Proteo" experiencing a slight stimulation under As stress, a feature opposite to that displayed by "Proteo"–"RS841" combination. Both trends were associated to correspondingly similar patterns of water use efficiency, but not to any correlated change in stomatal conductance, indicating that mechanisms beyond stomatal behavior are involved in determining the photosynthetic response of As-stressed melon plants. Interestingly, under severe As stress, the AN variations induced by the different rootstocks were not correlated to any corresponding variation of the aboveground biomass per plant, so that AN and plant growth were apparently not related to each other. Indeed, under As+ treatment, both whole plant and fruit biomass resulted generally higher in "Proteo" grafted onto the *C. maxima* × *C. moschata* hybrids (namely "RS841", "Shintoza" and "Strong Tosa"; i.e., those rootstocks conferring no AN enhancement under As-enriched nutrient solution). This seems to suggest a differential alteration of the energy balance of the plants, likely attributable to different energy dissipation pathways involved in the As stress response [38]. The data of As speciation and partitioning we found seem to confirm such hypothesis. Indeed, when exposed to the As+ treatment, the intraspecific rootstocks showed the highest total As concentration in roots, but the lowest one under its inorganic form (from 31% to 49%), meaning that up to 69% of their Asroot was present in organic and complexed form. This is the As form whose translocation from root is hampered, because of bonds with thiol groups of root-synthetized phytochelatins and subsequent sequestration into root cells vacuole [39]. On the contrary, the interspecific rootstocks subjected to the As+ treatment, showed the lowest total As concentration in their root, demonstrating their superior ability to buffer the As entrance into the plant from the nutrient solution. However, they showed also the highest inorganic Asroot incidence (up to 81%; i.e., the most mobile As form from the root organ) [39,40]. Inorganic As(III) is the main species suitable for transport trough xylem vessels and/or root efflux, while As(V) is rapidly reduced in roots to As(III) and then transported from the root cortical cells to the xylem vessels [40,41]. Accordingly, when growing under conditions of As stress, all the interspecific rootstocks had the highest TFshoot and TFpulp values, with "Shintoza" and "Strong Tosa" having also the highest concentration of As(III) in their roots. On the contrary, "Magnus" and "Dinero" confirmed the lowest TFshoot and TFpulp values, respectively, since the highest incidence of organic As into their roots.

Several authors have reported that the As uptake and translocation kinetics are mainly dependent on its concentration in the root environment and plant species [42,43], with the diversity of the root systems having a central role in determining differences among genotypes [44]. The PCA scatter plot highlighted that the As concentration in the root environment had a pivotal role in triggering different behaviors among rootstocks. Indeed, when exposed to the standard nutrient solution, all the grafting combinations clustered together on the negative side of PC1 (indicating lower values of As accumulation in plant tissues), without differences in terms of As partitioning inside the plant fractions. The opposite was noticed under conditions of imposed As stress, so that the grafting combinations clustered on the positive side of PC1, with a further distinction along the PC2 between rootstocks promoting the As accumulation mainly in roots (i.e., the intraspecific rootstocks) and those promoting its accumulation in shoot and pulp, namely "Shintoza" and "Strong Tosa". "RS841" slightly diverged from these latter owing to its ability to contain the As accumulation into the pulp.

The present study took into account the uptake and translocation of P too, given its peculiar interaction with As, as well as for its role in improving the conversion of solar energy into new plant biomass [45]. The complex interrelation between arsenate and phosphate in the substrate–plant systems has brought, up to now, no univocal results in literature, since their interaction can be either

synergistic or antagonistic, depending on the growing conditions [46]. In our experiment, both P concentration and partitioning inside the plant were significantly modified by the As level in the nutrient solution, so that in As-stressed plants prevailed an increased P concentration in root and pulp, while a lowered one in shoot. This tendency was more marked when "Proteo" was grafted onto the intraspecific combinations. To this end, the protective role of P against the As-induced lipid peroxidation of cellular membranes has been suggested [47]. In this view, the selective increase in P concentration inside the plant fractions could be the result of a melon response aimed at protecting both root (the plant organ firstly exposed to the As pollution) and reproductive structures from the As injury, this latter feature having been formerly described in rise [48]. On the other hand, the decreased P concentration in shoot tissues seems well related to the decreased LA, LAR, and LWR per plant recorded under conditions of As stress, so suggesting a possible role of P metabolism in buffering the transpiration-driven As entrance inside the plant.
