Do rootstock-scion combinations and irrigation water salinity affect the growth and yield of grafted tomatoes under root-restricted conditions?

Salinity is one of the most detrimental environmental factors affecting the growth and productivity of cultivated crops. High salt concentrations in the soil and irrigation water lead to a progressive build-up of sodium (Na⁺) and chlorine (Cl⁻) in the plant cells. This causes dehydration, nutrient imbalances, and oxidative stresses^[1,2] which ultimately restricts the growth and yield of cultivated crops^[3−5]. Tomato (Solanum lycopersicum L.) is among the most widely produced and consumed vegetables. Although it is considered a moderately salt-sensitive crop^[6,7], a 10% decrease in yield has been reported for each unit of soil electrical conductivity (EC) above a threshold value of 2.5 dS·m^–1[8].

Unfortunately, excess soil salinity is an increasing constraint for crop production worldwide. Excessive use of chemical fertilizers, the high salt content of irrigation water, and climate change are considered the main factors deteriorating the situation. Thus, 25% of irrigated agricultural land in the Mediterranean regions is significantly affected by high salinity^[9]. This significantly limits their production capacities and farming effectiveness.

Replacing highly saline soils with substrates, or simply growing plants in artificially prepared nutrient solutions, is considered a reliable alternative. Soilless culture systems have been recently developed and are gradually expanding^[10]. However, the spatial confinement of roots into a specific, well-defined root zone is a big constraint for these production systems^[11]. In container-grown plants, roots grow very dense to compensate for limited root zone volume^[12] and are much more exposed to environmental changes and human-imposed mistakes that can impact their morphology and nutrient uptake capabilities^[13]. Furthermore, the poor quality of irrigation water remains a primary concern for soilless cultivation systems^[14,15].

To relieve salinity pressure, different agricultural practices are applied in both soil-based and soilless production systems. In addition to the development of salt-tolerant cultivars^[16], plant inoculation with arbuscular mycorrhizal fungi^[17,18]—alone, or in combination with plant growth-promoting bacteria^[19]—and plant growth enhancement through high nitrogen doses^[20,21] are among commonly used agricultural practices.

In addition, grafting is considered one of the most effective agronomic interventions to alleviate the negative effects of salinity on cultivated crops^[22−24]. Although the published results are often contradictory, grafting tomatoes is often considered a powerful tool to increase the yield^[25,26]. However, the result depends on the characteristics of the scion and rootstock, their interaction, and stress intensity^[6].

As increased market demands and climate changes expose vegetable producers to new challenges, breeding new high-yielding, resilient, resource-use efficient, and high-quality cultivars is a common attitude of the industry^[27]. In addition, new rootstocks may help introduce other sources of genotypic and phenotypic variability into the crop^[28]. Testing and evaluating their effectiveness under commercial production systems and finding the most appropriate rootstock-scion combination under diverse production systems and growing conditions is a continuously ongoing task.

Although numerous studies have been conducted, the underlying mechanisms by which grafting improves the salt tolerance of vegetable crops remain largely unknown^[29]. Recently, we reported that root-restricted tomato plants benefit from appropriate rootstock-scion combinations. Increased root density, steadier growth under fluctuating salinity levels in the irrigation water due to a more stable uptake of nutrient elements, and higher sodium compartmentalization capabilities of the rootstocks were the most distinguished benefits^[30]. The question that follows is whether these advantages will be reflected in a higher yield or improved fruit quality of the grafted tomato plants under root restriction saline conditions. Hence, this research aimed to evaluate the effectiveness of several tomato rootstocks, most recently introduced, on the growth and yield of a specific tomato cultivar under root-restricted, saline conditions.

The experiment was conducted in a plastic greenhouse in Tirana, Albania, in the spring (March 24−June 23) of 2022. Six different rootstocks (Arnold, Dohkko, Emperador, He-man, Kaiser, and Suzuka; details are presented in Supplementary Table S1) and a tomato cultivar (Izmir) were included in the study. The grafted and self-grafted seedlings were produced by a commercial seedling producer nursery. Grafting was performed using the splice grafting method. Immediately after grafting, all plants were placed in a growth chamber. In the first 3 d, the air temperature was maintained at 26 °C, whereas the relative humidity was 100% at the start and then gradually decreased to 90%. A photosynthetic photon flux density (PPFD) of 100 μmol·m⁻²·s⁻¹(white, fluorescent lamps) was maintained, with a photoperiod of 12 h. On the fourth day after grafting, plants were moved to a heated greenhouse (air temperatures 18−24 °C, relative humidity 80%, natural lighting) until the experiment started. 65 d after sowing, the grafted and self-grafted seedlings were arranged in a factorial experiment conducted in a non-heated greenhouse. For that purpose, the plants were transplanted in 5 L plastic pots, one plant per pot, filled up with a mixture of peat (Klasmann-Deilmann GmbH, Geeste, Germany) and vermiculite (Dupré Minerals Ltd, Staffordshire, England) (2:1, v/v). The total number of plants of each graft combination was split in half and treated, respectively with the standard nutrient solution (EC 2.2 mS·cm⁻¹), or a 50 mM (EC 8 mS·cm⁻¹) saline solution obtained by adding sodium chloride (NaCl, 3 g·L⁻¹) to the nutrient solution. A mixture of commercial chemical fertilizers containing N, P₂O₅, K₂O, S, and micronutrients was diluted in tap water (pH 7.6, EC 1.2 mS·cm⁻¹) to prepare the nutrient solution (N: 100 mg·L⁻¹, P₂O₅ 25 mg·L⁻¹, K₂O 140 mg·L⁻¹). Since the water already contained considerable amounts of Ca (80 mg·L⁻¹) and Mg (35 mg·L⁻¹), no extra amounts of these elements were added to the nutrient solution. The pH of nutrient solutions was 7.6 and 8, respectively; no acidification was performed. A total of eight plants were used in each experimental plot-treatment. The plants were irrigated daily with equal amounts, equivalent to ETp values (3 mm·day⁻¹ at the start of the experiment, 5 mm·day⁻¹ during the harvesting time) corrected with the respective crop coefficient factors, of the normal or saline nutrient solution. To avoid salinity build-up in the root zone, a 20%−25% drainage rate coefficient was applied to all variants. At the beginning of the experiment, pH and EC values measured in the drainage water were similar to the respective values of the fresh nutrient solution. Later, pH of the drainage water remained almost constant, but slightly increased EC values (2.4 mS·cm⁻¹ for the non-saline, and 8.3 mS·cm⁻¹ for saline variants) were measured.

All plants grew up with a single (primary) stem, and plant height and leaf number were measured and recorded sequentially during the experiment. Further, the plant relative growth rate concept^[31] was adapted to calculate the relative stem elongation rate (RSER, cm·cm⁻¹·d⁻¹) and relative new leaf emerging rate (RLER, leaf·leaf⁻¹·d⁻¹) in the period from day 1 after transplanting, until day 42 after transplanting.

Three randomly selected plants were harvested 42 d after transplanting for biomass assessment and nutrient analyses. The roots were dissected from the shoots and washed off adhering soil particles, dried (65 °C, 72 h), and the dry matter of roots and shoots of each plant was determined separately to an accuracy of ± 1 mg (TP 303; Denver Instruments GmbH, Göttingen, Germany). Dried shoots (stems + leaves) of the self-grafted plants and two hetero-graft combinations (Dohkko/Izmir, and Suzuka/Izmir) were separately ground with a planetary mill (Pulverisette5; Fritsch, Idar-Oberstein, Germany) for mineral analysis, and nutrient contents (Ca, K, Mg, Na, and P) were analyzed by Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES) (Optima 8300; Perkin Elmer, Waltham, USA). For that purpose, the samples were first dissolved in nitrohydrochloric acid to ensure that all minerals were transferred in a liquid solution. The liquid sample was transferred into an argon plasma at a high temperature (6,000−10,000 K), which excites the atoms and ions in the sample, causing them to emit light at characteristic wavelengths. The emitted light is passed through a spectrometer, which separates it into component wavelengths.

The remaining (5) plants were topped up after the fifth cluster was established. Fruits were harvested at full maturity and weighed separately for each experimental unit. The number of fruits was counted for each harvesting date, and the average fruit weight was calculated by dividing the weight of harvested fruits by the respective fruit number. Five randomly selected fruits at the third harvesting date were used to measure soluble solids content (SSC, Brix) and fruit acidity using a digital refractometer (Atago Co. Ltd., Tokyo, Japan).

Statistical analyses

A factorial arrangement of 14 treatments (seven rootstock-scion combinations × two levels of salinity), eight replicates each, was employed in a factorial randomized complete block design.

The contribution of different growth and yield parameters under non-saline and saline conditions to the diversity of different rootstock-scion combinations was assessed by the principal component analysis (PCA). To produce a graphical evaluation of their relationships, the respective heat map was generated via the ClustVis (https://biit.cs.ut.ee/clustvis/, accessed on 20 October 2024) online program package.

Furthermore, to determine the influence of growth rate on plant yield, the respective correlograms of RSER and LRER with yield were accomplished via STHDA (www.sthda.com), separately for saline and non-saline conditions.

Residuals of all variables were tested for equality of variances and normality using Brown-Forsythe and Shapiro-Wilk tests, respectively. Differences in DM, RSER, RLER, shoot nutrient concentration, yield, and yield quality parameters were tested by two-way ANOVA using the PC program SigmaPlot 13 (Systat Software Inc., San Jose, CA, USA). Each significant ANOVA result (p < 0.05) was followed by a Holm-Sidak test at p < 0.05 as a post hoc test. Values given throughout the text are means ± standard errors (SE).

Only slight differences were found among the self-grafted and hetero-grafted variants regarding plant height under normal, non-saline conditions. Although the self-grafted variant (Izmir/Izmir) was among the tallest, the difference with most of the hetero-grafted variants was not significant (Table 1). Significantly fewer leaves (LN) were found in Arnold/Izmir vs the self-grafted variant, which led to a significantly higher internode length (IL). Some variations regarding LN and IL did exist among other grafted variants, but were less important. No differences were found among the self-grafted and hetero-grafted variants under non-saline conditions regarding relative stem elongation rate (RSER), although some variation was found regarding relative new leaves emerging rate (RLER).

Raised salinity has significantly reduced growth parameters; plant height (p < 0.001), leaf number (p < 0.001), internode length (p 0.04), relative stem elongation rate (p < 0.001), and relative new leaves emerging rate (p 0.004) to all variants (Table 1). Under such conditions, almost all hetero-graft combinations recorded higher RSER than the self-grafted variant. Kaiser/Izmir and Suzuka/Izmir recorded significantly higher RLER values than the self-grafted plants. Interestingly, hetero-grafted variants were able to sustain RSER despite the rise of salinity, whereas the self-grafted variant did not. In addition, despite the overall reduction in the number of leaves per plant, Suzuka/Izmir recorded the highest value of RLER (Table 1).

Under non-saline conditions, the self-grafted seedlings (Izmir/Izmir) showed a significantly higher root dry matter (DM_root) than the hetero-grafted variants. Suzuka/Izmir was the only exception. However, except for Emperador/Izmir, no differences were found between the self-grafted and hetero-grafted variants regarding the dry matter of shoots (DM_shoot) (Table 2). Consequently, the self-grafted plants had a higher ratio of root to shoot dry matter (DM_root : DM_shoot) than most of the hetero-grafted variants (Arnold/Izmir, Dohkko/Izmir, Emperador/Izmir, and He-man/Izmir). The rest—Kaiser/Izmir and Suzuka/Izmir—were included in the same statistical group as Izmir/Izmir (Table 2). The rise of salinity in the irrigation water reduced DM_root and DM_shoot to all variants, but the self-grafted variant, together with Suzuka/Izmir and Kaiser/Izmir, recorded the highest DM_root. Almost no differences were found between the self-grafted and hetero-grafted plants regarding DM_shoot (Table 2).

Rootstock-scion combinations and salinity effects on plant's nutrient concentration

Large variations in nutrient concentrations were found among the grafted plants. Lower Ca concentrations, but significantly higher P concentrations were found in the self-grafted plants compared with hetero-grafted variants under non-saline conditions. No difference was found regarding Na concentration. Although the rise of salinity in the irrigation water multiplied shoot Na concentration several times, there was no difference between the self-grafted and hetero-grafted variants (Table 3). The increase of salinity in the irrigation water did not affect Ca concentration, but significantly reduced the shoot concentration of K and P. On the contrary, Mg concentration was significantly increased alongside an immense increase in Na concentration. Ca, K, Mg, and P concentrations were also significantly affected by Rst/Sc combinations. Higher Mg and P concentrations were found in the shoots of the self-grafted variant compared to both hetero-grafted ones. However, the hetero-grafted variants were able to maintain higher K (Suzuka/Izmir, Dohkko/Izmir) and Ca (Dohkko/Izmir) shoot concentrations under saline conditions.

Rootstock-scion combinations and salinity effects on yield

Hetero-grafting did not provide any increase in the yield versus the self-grafted variants. No difference in the number of fruits per plant between the self-grafted and hetero-grafted variants was observed under either non-saline or saline conditions. The average fruit size of the self-grafted variant was generally higher under non-saline conditions, but no differences existed under saline conditions (Table 4). There were cases of hetero-grafted variants (Armold/Izmir under non-saline conditions) with a lower yield than the self-grafted variant. No correlations were found between RSER and RLER with yield in non-saline conditions, but some mild positive correlations were found between RLER and yield under saline conditions (Fig. 1a, b).

Figure 1. 

(a) Yield-RSER-RLER correlograms of hetero-grafted and self-grafted tomato plants (cv. Izmir F1), under non-saline conditions. RSER (relative stem elongation rate), RLER (relative new leaf emerging rate). (b) Yield-RSER-RLER correlograms of hetero-grafted and self-grafted tomato plants (cv. Izmir F1), under saline conditions. RSER (relative stem elongation rate), RLER (relative new leaf emerging rate).

To all variants, the yield was drastically reduced due to the rise of the salinity level in the irrigation water (Table 4). Mostly, the yield reduction was a consequence of significantly reduced (more than halved) average fruit weight. However, a significant reduction in the number of fruits per plant due to raised salinity was observed in several hetero-grafted variants (Arnold/Izmir, Dohkko/Izmir, He-man/Izmir) (Table 4). No effects of grafting were found regarding soluble solids content (Brix) and acidity under non-saline conditions, whereas some variations were found regarding acidity under saline conditions. Under such conditions, the self-grafted variant (Izmir/Izmir) appeared with the lowest acidity level (1.277), significantly lower than the rest of the grafted variants (Table 4).

The principal component analyses (PCA) have shown that the most contributing factors to the variability were the plant height (PH, −8.97) and average fruit weight (−6.15). The rest of the factors were grouped but still split into two subgroups with either positive (RSER (2.52), RLER (2.53), DM_root : DM_shoot (2.51), fruit acidity (2.36), yield (2.24), DM_root (1.93), Brix (1.6)), or negative (fruit number (−1.51), DM_shoot (−0.7), LN (−0.43)) values (Table 5). Most of the variability was represented by the first principal component (PC1, 96%), and the next principal component (PC2) accounted for nearly 4%. The non-saline and saline variants were clearly distinguished (Fig. 2), whereas the plant traits most affected by raised salinity were stem length and average fresh fruit weight.

Figure 2. 

The heatmap of growth, yield, and fruit quality traits of self-grafted (Izmir/Izmir) and heterografted tomato variants (Arnold/Izmir, Dohkko/Izmir, Emperador/Izmir, He-man/Izmir, Kaiser/Izmir, Suzuka/Izmir) under non-saline (0 mM NaCl) and saline (50 mM NaCl) conditions. Rows are centered; unit variance scaling is applied to rows. PH Plant height, LN leaf number, IL internode length, RSER relative stem elongation rate, RLER relative new leaf emerging rate, DMroot dry matter weight of roots, DMshoot dry matter of shoots, DM_root : DM_shoot the ratio of root dry matter with shoot dry matter. Both rows and columns are clustered using correlation distance and average linkage. Fourteen rows, 13 columns.

Interesting conclusions can be drawn from the heatmap chart regarding the variability of different rootstock-scion combinations under both non-saline and saline conditions. The non-grafted (Izmir/Izmir) variant was the most distinguished under non-saline conditions. Closely followed were Emperador/Izmir and He-man/Izmir. The remaining variants (Arnold/Izmir, Dohkko/Izmir, Kaiser/Izmir, Suzuka/Izmir) were grouped. On the contrary, He-man/Izmir was the most distinguished variant under saline conditions. Closely followed were Suzuka/Izmir and Izmir/Izmir. All remaining variants were grouped (Fig. 2).

It is commonly reported that grafting enhances growth and produces more vigorous plants^[23,26]. Although often seen as an advantage of grafted plants, this might be a concern for greenhouse tomatoes. The vegetative plants increase the chances of foliar fungi diseases^[32], induce assimilate partitioning imbalances between fruits and vegetative organs in the plants^[33], and delay fruit setting and maturity^[34]. Therefore, the rootstocks that produce less vigorous, compact plants fit better into a greenhouse environment^[27] and might potentially allocate more carbohydrates toward fruit formation^[27,35]. No significant differences between the self-grafted and hetero-grafted plants was found regarding stem height. Some of the rootstocks were able to shorten the length of plant internodes which satisfies the demand for compact plants in greenhouse production. No difference exists between the self-grafted and hetero-grafted plants regarding dry matter of shoots (DM_shoot). Conversely, the self-grafted plants had a higher dry matter of roots (DM_root) than most of the grafted variants which grants them the highest DM_root : DM_shoot ratio. Previous reports confirm that the reduced growth of the root system due to grafting is compensated by an enhancement in root enzyme activity^[36].

Although plant response to salinity is highly variable^[37], commonly, salinity reduces plant growth^[19] by inhibiting both cell division and elongation^[38]. Yet, little is known about salt-induced changes in shoot growth rate^[4]. Interestingly, the study found significant differences between self-grafted and hetero-grafted plants regarding relative stem elongation rate (RSER) and relative new leaf emerging rate (RLER). The enhanced shoot growth and the increased ability of certain rootstock-scion combinations to produce new leaves indicate that they are more capable of alleviating the salinity stress by improving the osmotic tolerance of grafted plants^[37]. Maintaining an appropriate stem growth rate was considered essential to achieving a good yield in cucumber plants^[39]. In contrast, the study did not find any significant correlation between relative stem elongation rate, or relative new leaf emerging rate with tomato yield in non-saline conditions, but good correlations were found between the relative new leaf emerging rate and yield under saline conditions.

Grafting is often considered a tool to increase yield in several vegetable crops^[17,22,40]. On the contrary, the rootstocks involved in this experiment were not able to increase tomato yield in the condition of reduced root volume, either in non-saline or saline conditions. No advantage of grafted variants was found versus the non-grafted plants, either regarding the number of harvested fruits per plant, or the average fruit weight. In contrast, in a previous publication of Djidonou et al.^[40] an increase in the yield of grafted tomatoes was supported by a higher number of larger fruits. Still, similar to us, di Gioia et al.^[25] did not report any significant change in yield components of a heirloom tomato variety grafted onto two interspecific (S. lycopersicum × S. habrochaites) rootstocks. More recently, Parisi et al.^[41] reported an increase in harvested yield by only three out of seven rootstocks used in grafting with an Italian tomato landrace, and Gong et al.^[42] found varying results depending on specific rootstock-scion combinations and planting season. Considering a large number of different, even contradictory, evidence, the study concludes that the wide range of rootstock-scion combinations leads to mixed results. Huge differences regarding specific production systems and the interactions of specific genotypes with production systems make the performance of grafted tomato plants even more complicated^[43].

The higher yield in grafted plants is often linked with a more vigorous rootstock root system, its enhanced root hydraulic conductance, and nutrient uptake^[44]. However, if the growing system can maintain a constant concentration of nutrient elements at the root surface, the restricted root system is no longer a limiting factor^[11]. Under these circumstances, frequent fertigation fulfills plants' requirements through the continuous replenishment of nutrients in the vicinity of the root interface^[45]. Indeed, despite the differences between self-grafted and hetero-grafted plants, shoot concentrations of Ca, K, Mg, and P to all variants varied within the range of respective recommended optimum concentrations^[46]. We discussed in detail the differences between the self-grafted and heterografted variants regarding the root and shoot nutrient concentrations in a previous publication^[30]. Even under salinity conditions where the shoot concentration of some elements (K, Mg, P) falls, it remained within the recommended range of optimum concentration. Therefore, improved nutrient uptake in several rootstocks was not enough to distinguish the grafted and non-grafted plants regarding the harvested yield. To increase the efficiency of grafted plants under such production systems, further studies regarding the appropriate management of limited root systems to ensure adequate nutrition, optimum irrigation frequency, proper root zone temperature, and oxygenation level are needed^[13].

In both the self-grafted and hetero-grafted plants, yield was drastically reduced by increased salinity. A slight reduction in fruit number was noticed, but the real factor that led to severe yield reduction was the drop in average fruit weight by more than half. No difference was found between the self-grafted and hetero-grafted variants regarding soluble solids content (SSC) and fruit acidity under non-saline conditions. In line with the study, Gong et al.^[47] concluded that scion cultivars have a much more pronounced impact on fruit quality attributes than rootstocks. On the contrary, SSC and acidity were significantly increased by the rise of salinity in the irrigation water. Under such conditions, no differences were found between the self-grafted and hetero-grafted variants regarding SSC, but a distinctive difference was found regarding fruit acidity. The self-grafted plants showed a significantly lower acidity level than each of the hetero-grafted variants. As such, the results conform with a previous report by Turhan et al.^[48] who reported that grafting tomatoes with interspecific rootstocks (S. lycopersicum × S. habrochaites) increased titratable acidity relative to the non-grafted controls. Contrary to the study, Djidonou et al.^[40,49] and Parisi et al.^[41] concluded that either beefsteak or cherry tomatoes grafted with interspecific rootstocks did not affect fruit SSC, acidity, and SSC/TA ratio. A clear answer regarding this issue is further complicated by the fact that rootstocks with similar genetic backgrounds could exert different impacts on tomato fruit sensory attributes depending on the scion genotype and growing conditions^[48].

Overall, the results of this experiment prove that there is not always a clear distinction between the non-grafted and grafted tomato plants under root-restricted conditions. Only carefully selected rootstock-scion combinations might make a distinction. Therefore, the effects of grafting itself and the role of specific rootstock-scion combinations on tomato growth, yield, and the fruit quality remain an open question for further elaboration.

Grafted plants demonstrated improved shoot growth and new leaf emerging rate, indicating enhanced capabilities to alleviate the salinity stress. Good correlations were found between the new leaf emerging rate and yield under saline conditions. Considering the low variability found between the self-grafted and grafted tomato variants under root-restricted volume conditions, only carefully selected rootstock-scion combinations might make a distinction. The study's result suggest that He-man can be considered an appropriate rootstock for root-restricted grafted tomatoes under salinity conditions.