Solanum lycopersicum L. (cv Robin) seedlings were obtained from a local vendor. Sowing was done in pots with an automatic seeder to ensure uniform germination and growth up to the stage of two true leaves (BBCH scale 12) in a greenhouse. Only seedlings of the same size that were well-rooted and disease-free were used for the experiments. After reaching the desired size, plant trays were placed in a growth chamber with controlled conditions: temperature 24°C ± 2, photoperiod 16-8h (light/dark), humidity 60-75%, and irradiance 300 µmols m^-2 s^-1.

Solanum lycopersicum L. seeds were sterilized using 50% commercial bleach plus 0.02% Tween-20 for 1 hour followed by 5x5 minutes washes with sterile water. Seeds were sown on 150 mm petri plates with 4 layers of sterile filter paper previously wetted with 15 mL of control (0.1% DMSO), ABA, or AMF4 solution. 30 seeds per plate were sown in duplicate for each treatment (n=60). The plates were placed in a growth chamber under long day conditions and 23°C, and radicle emergence was scored on day 8.

Plants with two true leaves were subjected to a growth experiment under well-watered (WW) or water deficit (WD) conditions, essentially following the procedure described by Jiménez-Arias et al. (2022b). Thus, WW plants received 10 ml/pot (full field capacity) whereas WD plants received 5 ml/pot each day for 14 days. Symptoms of WD were evident after 5 days and photosynthesis showed a dramatic decline after 11 days. Briefly, four groups of plants were established to test the effect of AMF4- versus mock-treatment, either under well-watered or water deficit conditions, and each experimental condition was tested in twenty plants ( Figure 1 ). Plants designated as well-watered were irrigated to full field capacity, while those grown under water deficit conditions were irrigated with 50% less water. To ensure good nutrition, in all cases, plants were irrigated with a half-strength Hoagland solution (Hoagland and Arnon, 1950). Foliar agonist- or mock-treatments were carried out in 10 mM MES at pH 5.7 with 0.1% DMSO and 0.05% Tween 20, and additionally, 50 µM AMF4 was applied in agonist-treated plants. The treatment consisted of spraying 2.5 ml directly onto the tomato leaves to ensure that the entire leaf was soaked. The mock- or 50 µM AMF4-treatments were applied four days after water deficit imposition ( Figure 1 ). Measurements of gas-exchange parameters were performed at 1, 3, 5, 7 and 9 days after foliar agonist- or mock-treatment.

Two independent experiments were conducted with 23 plants for each treatment, and the plants were finally collected 14 days after the experiment start. Plants were dried in an oven at 60 °C for two days and the shoots and roots were weighed separately. Different indices were calculated using the weight of the plants at 14 days, such as the stress susceptibility index (SSI) (Ganança et al., 2015), stress tolerance index (TSI) (Farshadfar et al., 2013), relative growth rate (RGR), plant water use efficiency (WUEp), and relative water content (RWC) (Jiménez-Arias et al., 2022a; Jiménez-Arias et al., 2022b).

Gas exchange analyses were carried out on the fully developed leaves (N = 30). Photosynthesis (Pn), intracellular CO₂ (Ci), stomatal conductance (gs) and transpiration rate (E) were measured on the attached leaves using a portable infrared gas analyzer (LCPro, BioScientific Ltd., Hoddesdon, UK). The values for instantaneous water use efficiency (iWUE) and intrinsic water use efficiency (intWUE) are the ratios between Pn/E and Pn/gs, respectively (Seibt et al., 2008). The ratio between Pn and Ci was also calculated. The measurements were carried at ambient CO₂ concentration, photosynthetic photon flux density (PPFD) of 1000 μmol m-2 s-1 (optimized with a light curve) and cuvette airflow of 500 mL min-1. Finally, water use efficiency at the total plant stage (WUEp) was measured as the ratio between biomass increment and the amount of water consumed (Medrano et al., 2015).

The proline concentration at each experimental time-point was calculated as the average of 6 plants. Proline content was determined as described by Bates et al. (1973) with minor modifications (Jiménez-Arias et al., 2015). Samples of 20-50 mg of dry tissue (leaves) were ground and extracted with 4 ml of 3% sulphosalicylic acid. The plant extract was centrifuged at 15000 g for 30 minutes, and 2 μl of the supernatant was mixed with 2 ml of acid ninhydrin and incubated at 100°C for 60 minutes. This reaction was stopped in an ice bath. After extraction with 4 ml toluene, the absorbance of the organic phase was measured at 520 nm in an Aquarius CE7200 double-beam spectrophotometer (Cecil Instruments, Cambridge, England). The proline concentration was calculated from a standard curve and normalized to dry weight.

Phosphatase activity was measured using p-nitrophenyl phosphate (pNPP) as a substrate. The assays were performed in a 100 μL solution containing 25 mM Tris HCl [pH 8.0], 10 mM MnCl2, and 25 mM pNPP. The assays contained 1 μM phosphatase (ΔN-HAB1), 2 μM receptor and the indicated concentrations of ABA or AMF4. Phosphatase activity was recorded with a ViktorX5 reader at 405 nm every 60 s over 20 min, and the activity obtained after 20 min is indicated in the graphs. The His-tagged Sl08g076960, Sl06g061180 and Sl03g007310 ABA-receptor proteins were purified as described by González-Guzmán et al. (2014).

Six tomato plants were mock- or 50 μM AMF4-treated and epidermal peels were taken 24 hours after the treatment from abaxial and adaxial leaf surfaces and mounted in glycerin. Observations and photomicrographs were carried out using an Optika B-350 light microscope with a computer image capture system Moticam 2500. The stomatal aperture analysis was performed on the computer using Motic Images Plus 2.0 program, and 120 stomata from 6 different plants were measured for each treatment condition.

Six plants were mock- or 50 μM AMF4-treated and analyzed 24 hours after treatment. Two leaves were cut into small pieces (3 mm²), fixed in 2.5% glutaraldehyde in PB 0.05 M (pH 7), dehydrated in graded ethanol series, dried using 1,1,1,3,3,3-hexamethyldisilazane (HMDS) and coated with gold (15 nm). Observations and photomicrographs were made using a ZEISS EVO15 (ZEISS) scanning electron microscope (General Services for Supporting Research, University of La Laguna). Representative images observed in more than 90% of stomata were used in the figure.

Analyses of Macronutrients (Ca, K, Mg and P) were conducted using leaf samples from each treatment harvested at the end of the water deficit experiment. Samples were dried at 80 °C and then ground using an IKA M20 mill. The samples were then placed in an oven at 105 °C for five hours and then transferred to a desiccator to obtain the dry weight. Next, 500 mg of ground powder was taken from each tomato sample and after conversion to ash, the samples were treated with 6 N hydrochloric acid in a muffle furnace at 480 °C. Mineral content was determined using Avio^® 500 ICP-OES (Perkin Elmer) and interpolation of the data into a standard curve. N and S measurements were carried out on 1 mg dry leaf samples using the CHNS TruSpec Micro (LECO Corporation). All measurements were carried out in triplicate.

One-way ANOVA tests (Duncan’s post hoc) were applied to analyze the differences between treatments in all measures studied. A two-way ANOVA was then run to test the differences among factors. Irrigation deficit was the first factor, AMF4 treatment the second, and interaction the third. All statistical studies were performed using IBM-SPSS24 statistical package.

Ten-day-old tomato seedlings (cv. Moneymaker) were mock- or 10 μM AMF4-treated for 3 h. Total RNA was extracted using a NucleoSpin RNA plant kit. Synthesis of cDNA and quantitative real-time PCR (qRT-PCR) analyses were performed as described by González-Guzmán et al. (2014). Amplification of the ABA-responsive Sl02g084850 (SlRAB18), Sl06g067980 (SlLEA), and Sl06g019170 (SlP5CS1) genes was done using the primers described by González-Guzmán et al. (2014). Expression was normalized using the values obtained with Sl06g009970 (SlEF1a).

AMF4 (N-(2-oxo-1-propyl-1,2,3,4-tetrahydroquinolin-6-yl)-1-(2,3,5,6-tetrafluoro-4-methylphenyl) methanesulfonamide) was synthesized from commercial 3,4-Dihydroquinolin-2(1H)-one following the protocols specified in literature (Cao et al., 2017).

In order to characterize the activity of the ABA-receptor agonist AMF4 in tomato, firstly, we tested the in vitro activity of AMF4 using two tomato ABA receptors that are orthologs of AtPYL1, namely Sl06g061180 and Sl08g076960. The cloning and purification of the corresponding recombinant proteins were described previously (González-Guzmán et al., 2014). As a result, we found that both PYL1-like tomato receptors were activated by AMF4, resulting in the inhibition of the PP2C HAB1 ( Figure 2A ). AMF4 does not activate PYL8, and indeed, the tomato ortholog of PYL8, namely Sl03g007310, did not inhibit the activity of the PP2C HAB1 even at 10 μM AMF4 ( Figure 2A ). These results confirm that AMF4 can activate certain tomato ABA receptors in vitro.

ABA plays an important role in the maintenance of seed dormancy and exogenous application of ABA leads to inhibition of germination (Cornforth et al., 1965). To test the in vivo activity of AMF4 compared to ABA, we performed seed germination assays on filter paper wetted with ABA or AMF4. ABA- or AMF4-treated seeds showed a similar percentage of germination inhibition at 1 µM ( Figure 2B ). Moreover, 10 µM AMF4 inhibited tomato seed germination by more than 80%, thus confirming AMF4’s activity in tomato ( Figure 2B ). AMF4 regulates ABA-responsive genes in arabidopsis plants and reduces transpiration both in arabidopsis and soybean plants as a result of AMF4-induced stomatal closure (Cao et al., 2017). In order to test the biological activity of AMF4 in tomato seedlings, 10-day-old plants were treated with 10 μM AMF4 for 3 h and subsequently, we analyzed transcript levels of three tomato genes by qRT-PCR, namely RESPONSIVE TO ABA 18 (RAB18), LATE EMBRYOGENESIS ABUNDANT (LEA) 6g067980 and the Δ1-PYRROLINE-5-CARBOXYLATE SYNTHASE (P5CS1) genes. These genes were upregulated 26, 16 and 2-fold by AMF4 treatment, respectively ( Figure 2C ). Given that P5CS1 is a rate-limiting enzyme for the accumulation of proline, its upregulation suggests that AMF4 treatment might have a positive effect to increase the levels of this metabolite (Bhaskara et al., 2015; see below). Osmotic adjustment and regulation of transpiration can improve plant adaptation in water-limited environments (Bhaskara et al., 2015). Therefore, we also measured the stomatal aperture of tomato leaves using whole-leaf imaging before and after 50 μM AMF4-treatment. AMF4 induced stomatal closure in tomato and the stomatal pore area of AMF4-treated leaves was 28% and 32% lower in the adaxial and abaxial sides of the leaf, respectively, compared to mock-treated plants ( Figure 2D ). Additionally, we obtained stomatal images by both light and scanning electron microscopy, which show a reduction in stomatal aperture after AMF4 treatment (top and bottom panels, respectively, Figure 2E ).

In nature, water is possibly the most limiting factor for plant growth and development. Water deficit negatively affects photosynthesis (Luo et al., 2016) and plant nutrient uptake (Coelho et al., 2023), and overall reduces plant growth and final crop yield (Shao et al., 2008). In order to test the potential use of AMF4 to increase drought tolerance we subjected tomato plants to a water deficit regime for 14 days as described in methods and Figure 1 . The negative impact on tomato growth after a 50% reduction in irrigation for two weeks was measured, resulting in a 42% reduction in plant growth at the end of the experiment ( Figure 3 ; Supplementary Table 1 ). In WW plants, treatment with 50 μM AMF4 had a negative effect on plant growth and resulted in an 18% growth reduction compared to mock-treated WW plants ( Figure 3 ). The growth-repressing effect of ABA signaling is well known and it represents a trade-off for stress adaptation (Miao et al., 2018). Thus, in the absence of a water deficit, the use of AMF4 needs further evaluation. For example, proper use of AMF4 might lead to a reduction of vegetative growth but with minimum effect on the yield of some crops, thus increasing the harvest index. In contrast, it is noteworthy that AMF4-treated plants that were exposed to water deficit showed a significant 17% increase in plant weight compared to mock-treated ( Figure 3 ). We also tested the effect of a second AMF4 treatment (AA-WW or AA-WD), applied 6 days after the first one, but the results showed no significant difference with the plants treated only once ( Figure 3 ). Therefore, AMF4 application under water deficit conditions improves plant growth of tomato plants compared to mock-treated plants ( Figure 3 ).

In addition to a direct measurement of plant weight to evaluate plant tolerance to water deficit, we have used the stress susceptibility index (SSI) (Fischer and Maurer, 1978), stress tolerance index (STI) (Rosielle and Hamblin, 1981) and relative growth rate (RGR) (Hunt et al., 2002). SSI was proposed by Fischer and Maurer (1978) to measure yield stability, which includes changes in potential and actual yields in variable environments. The stress tolerance index (STI) was defined by Rosielle and Hamblin (1981) as the differences in yield between plants grown under water deficit or well-watered conditions. Finally, RGR shows how the plant increases weight in a certain period (Hunt et al., 2002). The increase in plant tolerance to water deficit by AMF4 treatment was evident when we used the mentioned indices ( Table 1 ). Thus, according to the RGR, A-WW plants have a daily growth reduction of 1% compared to the mock-treated WW plants, while A-WD plants show the opposite behavior compared to mock-treated in WD conditions. WUEp has been defined as the ratio between plant biomass produced and the water used in a period of time (Stanhill, 1986). In WD conditions, A-WD plants showed a 25% increase in WUEp compared to mock-treated, indicating that AMF4 improves biomass production per unit of water consumed by tomato seedlings ( Table 1 ). Concerning AMF4 treatment in the absence of water deficit, although RGR was reduced, proper use might benefit agronomic yield as indicated above. An important indicator of plant water status is the RWC (Soltys-Kalina et al., 2016). A comparison of RWC values in WD and A-WD plants indicated that AMF4 treatment led to improved water status of the plant under water deficit ( Table 1 ). Moreover, A-WD plants reached a similar RWC as WW plants, whereas mock-treated WD plants showed a significant 20% lower RWC value compared to WW plants. In agreement, A-WD plants showed a lower SSI index compared to WD plants, indicating higher stress tolerance, because lower values of the SSI index are positively correlated with higher yields under stress (Farshadfar et al., 2013; Zdravkovic et al., 2013). The higher values of the STI index in A-WD plants indicated that AMF4 significantly increased tolerance to water deficit stress. Finally, the effect of WD and AMF4 treatment was tested either individually or considering their mutual interaction through a two-way ANOVA analysis, in which we compared the dry weight of whole plants, shoots or roots ( Supplementary Table 1 ). The WD had a significant effect on the weight of whole plants, shoots or roots, but the AMF4 treatment did not produce a significant effect on the rootweight, compared to WW conditions, suggesting that the agonist had the greatest influence on shoot growth. Interestingly, the interaction showed differences between A-WD in all variables studied, and the F-values and statistical significance showed that the interaction is higher in shoots, showing that the treatment has a direct and positive interaction in plant acclimation under water deficit stress ( Supplementary Table 1 ).

ABA treatment has direct effects on leaf gas-exchange parameters by promoting stomatal closure, which directly affects water balance (Li and Liu, 2021). Plants benefit from ABA-induced stomatal closure to reduce transpiration during water deficit stress (Lee and Luan, 2012). As expected for an ABA-receptor agonist that mimics the ABA phytohormone, AMF4 had a clear effect on gas-exchange parameters, as shown by the E, gs and Pn values ( Figure 4 ). Firstly, AMF4 treatment in WW plants led to a reduction of gs that was persistent throughout the experiment ( Figure 4B ). Initially, this reduction was similar to that imposed by WD, but 5 days after foliar treatment, the reduction of gs and E by WD was higher than that achieved by AMF4 treatment in WW plants ( Figures 4A, B ). Interestingly, the dramatic decrease in photosynthesis induced by WD conditions was partially mitigated in AMF4-treated plants subjected to WD, see for example Pn values at 5, 7 and 9 days ( Figure 4C ). These results can explain the higher growth under WD conditions of AMF4-treated plants compared to mock-treated plants ( Figure 3 ), as well as the higher RWC and the improved stress tolerance indexes under WD conditions ( Table 1 ).

Tomato plants under WD conditions showed a dramatic decrease in gs and E, as well as in Pn. The activation of ABA signaling by AMF4 was able to mitigate the drop in Pn and led to a reduction of gs, which increased iWUE and intWUE values ( Figures 5A, B ). These parameters are crucial to maintaining crop yield under WD field conditions (Condon, 2004). Thus, AMF4-treated plants under WD conditions showed higher water use efficiency compared to mock-treated WD plants as shown by intWUE and iWUE values ( Figures 5A, B ). These ratios are not always directly correlated with WUEp (Medrano et al., 2015; Yang et al., 2019); however, in our WD conditions, AMF4-treated plants also showed higher WUEp compared to mock-treated WD plants ( Table 1 ). It is interesting to point out that the A-WW plants had a lower intWUE value compared to the WW plants. This ratio, which reflects CO₂ assimilation versus water loss, shows again that the AMF4 treatment negatively affects CO₂ assimilation when the plant is not under stress conditions. The Pn/Ci ratio also provides a simple method for assessing the efficiency of CO₂ assimilation according to the intracellular CO₂ concentration (Long, 2003). The AMF4 treatment led to plant protection in plants exposed to WD, as shown by the higher value of Pn/Ci compared to mock-treated WD plants ( Figure 5C ). In summary, AMF4 treatment improved water stress avoidance, therefore tomato plants showed better water balance and CO₂ assimilation; therefore, might be a promising approach to maintain yield under drought conditions.

Osmotic adjustment is a plant adaptive response to water deficit and requires the intracellular accumulation of compatible solutes to facilitate water uptake, protect the integrity of cellular structures and prevent the denaturation of soluble enzymes (Alvarez et al., 2022). Proline is accumulated in a wide range of taxonomically diverse plants, including tomato (Alvarez et al., 2022). Either ABA or ABA-receptor agonists such as quinabactin are able to upregulate the expression of the P5CS1 gene, which encodes a rate-limiting enzyme for proline biosynthesis (Szekely et al., 2008; González-Guzmán et al., 2014). AMF4 also upregulated P5CS1 expression in tomato ( Figure 2C ). In order to evaluate the effect of AMF4 treatment on proline levels, we measured proline at 24, 48 and 72 h in WW, A-WW, WD and A-WD plants ( Figure 6 ). The combination of water deficit and AMF4 treatment enhanced proline levels compared to mock-treated plants subjected only to WD ( Figure 6 ). The additional boosting of proline biosynthesis might reflect an additive effect of both osmotic stress and AMF4 on P5CS1 upregulation.

Water deficit stress leads to impairment of nutrient uptake, also affecting transport and distribution, which finally downregulates mineral nutrition and impairs nutrient partitioning (Sánchez-Rodríguez et al., 2010; D’Oria et al., 2022). We analyzed the nutrient content of Ca, K, Mg P, N and S, both in shoots and roots of the four groups established in our experimental conditions. WD significantly reduced the content of Ca, K and Mg in shoots, and uniquely of K in roots ( Table 2 ). AMF4 treatment significantly prevented this reduction in nutrient content in both tissues. Thus, the treatment maintains potassium levels, which are crucial for osmotic adjustment. Proper K levels are also required for stomatal regulation and to maintain dry matter accumulation under drought stress (Wang et al., 2013). On the other hand, adequate Ca levels are required for different processes involved in plant response to water stress, such as plasma membrane function, cell wall structure and signaling (Tonetto de Freitas et al., 2014). Specifically, in tomato Ca deficiency leads to disorders in fruit such as blossom-end rot (BER), and foliar application of ABA prevented BER disorder (Tonetto de Freitas et al., 2014). Spraying with ABA increased apoplastic Ca concentrations and reduced BER incidence (Tonetto de Freitas et al., 2014). Interestingly, AMF4 treatment recovered Ca levels in shoots of plants that were subjected to water deficit ( Table 2 ). Finally, it is interesting to note the significant increase observed in nitrogen concentration in shoots after the application of the agonist under WW conditions, while the comparison of WD and A-WD plants did not reveal a significant difference ( Table 2 ).