Two sweet potato genotypes, namely ‘Japanese Yellow’ (WD tolerance; positive check) and ‘Tainung 57’ (WD sensitive) obtained from Agricultural Extension Group, Phichit province, Thailand, were used as master stock materials (Yooyongwech et al., 2013). Single vine cuttings (15 ± 1 cm in length) without leaf blades were propagated and planted into plastic pots (∅ = 20 cm) containing 2 kg garden soil (EC = 2.687 dS m^-1; pH = 5.5; organic matter = 10.36%; total nitrogen = 0.17%; total phosphorus = 0.07%; and total potassium = 1.19%). The cuttings planted in the pot culture were incubated in a greenhouse under 500-1,000 μmol m^-2 s^-1 photosynthetic photon flux density with a 10 h d^-1 photoperiod, 28 ± 2°C ambient temperature and 80 ± 5% relative humidity, for 4 weeks. The ‘Tainung 57’ plants were treated with exogenous foliar applications of PBZ at 0, 17, 34, and 51 μM (25 mL plant^-1). Each treatment consisted of well-watered (WW; 47.5% soil water content), mild water deficit (MWD; 32.3% soil water content, by withholding water for 7 days), and severe water deficit (SWD; 17.5% soil water content, by withholding water for 14 days) conditions (Supplementary Figure S1). Growth characteristics (vine length and number of leaves), soluble sugar, free proline content, leaf osmotic potential, photosynthetic pigments, chlorophyll fluorescence, net photosynthetic rate (P_n), stomatal conductance (g_s), and transpiration rate (E) in each treatment was measured. In addition, number of storage roots per plant, storage root yield, vine fresh weight, root fresh weight, vine dry weight, and root dry weight in the harvesting period (150 days after cutting) were recorded.

Sucrose, glucose, and fructose content in the second fully expanded leaf from shoot tip were measured following the method of Karkacier et al. (2003). In brief, sweet potato leaves were collected and freeze-dried using a freeze-dryer. Fifty-milligram sample was ground in a mortar with liquid nitrogen. Following this, 1 mL of nanopure water was added and centrifuged at 12,000 rpm for 15 min. The supernatant was collected and filtered through a 0.45 μm membrane filter (VertiPure^TM, Vertical^®). Twenty microliters of the filtrate was injected into Waters HPLC equipped with a MetaCarb 87C column and a guard column. Deionized water was used as the mobile phase at a flow rate of 0.5 mL min^-1. The online detection was performed using a Waters 410 differential refractometer detector and the data was analyzed by Empower^® software. Sucrose, glucose, and fructose (Fluka, United States) were used as the standards.

Free proline in the second fully expanded leaf from shoot tip was extracted and analyzed according to the method of Bates et al. (1973). In brief, 50 mg of fresh material was ground with liquid nitrogen in a mortar. The homogenate powder was mixed with 1 mL aqueous sulfosalicylic acid (3%, w/v) and filtered through Whatman #1 filter paper (Whatman, England). The extracted solution was reacted with an equal volume of glacial acetic acid and ninhydrin reagent (1.25 mg ninhydrin in 30 mL glacial acetic acid and 20 mL 6 M H₃PO₄) and incubated at 95°C for 1 h. The reaction was terminated by placing the container in an ice bath. The reaction mixture was mixed vigorously with 2 mL of toluene. After cooling to 25°C, the chromophore was measured at 520 nm by spectrophotometer (HACH DR/4000; Model 48000, HACH Company, Loveland, CO, United States) using L-proline as a calibration standard.

Osmotic potential in the leaves of sweet potato was measured, according to Lanfermeijer et al. (1991). In brief, 100 mg of fresh leaf tissue was chopped into small pieces, transferred to 1.5 mL micro tube, and then crushed using a glass rod. The 20 μL of extracted solution was dropped directly onto a filter paper in an osmometer chamber (5520 Vapro^®, Wescor, Logan, UT, United States) and subsequently the data were collected. The osmolarity (mmol kg^-1) was converted to osmotic potential (MPa) using conversion factor of osmotic potential measurement.

Chlorophyll a (Chl a), chlorophyll b (Chl b), total chlorophyll (TC), and total carotenoid content in the second fully expanded leaf from shoot tip were analyzed according to the method of Shabala et al. (1998), whereas total carotenoid (C_x+c) content was assayed following Lichtenthaler (1987) method. One hundred milligram leaf tissue was homogenized in glass vials using 10 mL of 99.5% acetone, and blended using a homogenizer. The glass vials were sealed with Parafilm^® to prevent evaporation, and then stored at 4°C for 48 h. Chl a and Chl b concentrations were measured at 662 and 644 nm, whereas C_x+c concentration measured at 470 nm using UV-VIS spectrophotometer against acetone (99.5%) as a blank.

Chlorophyll fluorescence emission was measured from the adaxial surface of the second fully expanded leaf from the shoot tip using a fluorescence monitoring system (model FMS 2; Hansatech Instruments Ltd., Norfolk, United Kingdom) in the pulse amplitude modulation mode (Loggini et al., 1999). A leaf, kept in dark for 30 min was initially exposed to the modulated measuring beam of far-red light (LED source) with typical peak at 735 nm. Original (F₀) and maximum (F_m) fluorescence yields were measured under weak modulated red light (<0.5 μmol m^-2 s^-1) with 1.6 s pulses of saturating light (>6.8 μmol m^-2 s^-1 PAR) and calculated using FMS software for Windows^®. The variable fluorescence yield (F_v) was calculated using the equation: F_v = F_m - F₀. The ratio of variable to maximum fluorescence (F_v/F_m) was calculated as the maximum quantum yield of PSII photochemistry. The photon yield of PSII (Φ_PSII) in the light was calculated as: Φ_PSII = (F_m′ -F)/F_m′ after 45 s of illumination, when steady state was achieved (Maxwell and Johnson, 2000).

Net photosynthetic rate (P_n; μmol m^-2 s^-1), stomatal conductance (g_s; mmol H₂O m^-2 s^-1), and transpiration rate (E; mmol m^-2 s^-1) of the second fully expanded leaf from shoot tip were measured using a Portable Photosynthesis System with an Infra-red Gas Analyzer (Model LI 6400, LI-COR^® Inc., Lincoln, NE, United States). Transpiration rate was measured continuously by monitoring the content of air entering and exiting the IRGA headspace chamber, according to Cha-um et al. (2007). Flow rate of IRGA headspace chamber was set as 500 μmol s^-1 with ambient CO₂ concentration (350 ± 10 μmol CO₂ mol^-1) and 1,000 μmol m^-2 s^-1 photosynthetic photon flux density provided by 6400-02 B LED Red/Blue light source.

Vine length, leaf length, leaf width, number of leaves, storage root yield, vine fresh weight, root fresh weight, vine dry weight, and root dry weight of sweet potato were measured. Vine and roots were dried at 80°C in a hot-air oven for 2 days, and then incubated in desiccator before the measurement of dry weight. In addition, the leaf toxic symptoms, i.e., leaf wilting and leaf chlorosis, were observed (Supplementary Figure S2).

The experiment was arranged as 4 × 2 factorial in a Completely Randomized Block Design (CRBD) with eight replicates (n = 8). Two-way analysis of variance (ANOVA) in each parameter was validated using SPSS software (Supplementary Tables S1–S3). The mean values obtained were compared using Tukey’s HSD and analyzed with SPSS software. Pearson correlation between Chl a content and F_v/F_m, TC and Φ_PSII, Φ_PSII and P_n, and P_n reduction and plant dry weight was analyzed.

Morphological characteristics of the storage roots of sweet potato at the harvest stage are shown in Figure 1. There was no significant difference observed in the storage root yield of ‘Japanese Yellow’, the positive check, under WW and WD conditions (only 5.4% reduction). ‘Tainung 57’ plants grown under WW conditions and pretreated with PBZ had a higher storage root yield than the non-pretreated plants (Figure 1). In WD treatment, storage root yield was significantly improved especially in 37 μM PBZ pretreated plants (Figure 1) Storage root yield of untreated ‘Tainung 57’ was significantly decreased by 47.55% when exposed to 17.5% SWC (SWD) compared to WW conditions. However, when the plants were treated with 17, 34, and 51 μM of PBZ, the yield decreased up to 37.68, 35.77, and 34.95%, respectively (Table 1).

Vine fresh weight, root fresh weight, vine dry weight, and root dry weight in PBZ-pretreated plants raised under WW conditions were greater than those in water stress. In WD stress conditions, vine fresh weight, root fresh weight, vine dry weight, and root dry weight in non-pretreated plants declined by 40.10, 65.47, 67.86, and 66.91%, respectively; whereas, these parameters improved with the 34 μM PBZ treatment (19.64, 42.80, 44.21, and 48.06%, respectively). In addition, under WD conditions the root fresh weight (35.10% reduction) and vine dry weight (31.73% reduction) in 17 μM PBZ-pretreated plants were maintained. In the positive check cv. ‘Japanese Yellow,’ the above traits in WD stressed plants showed higher performance than the ‘Tainung 57’ plants (Table 1).

Leaf wilting and chlorosis were evidently observed in the older leaves when sweet potato plants were exposed to SWD, especially in non-treated plants (Supplementary Figure S2). Vine length was significantly reduced by PBZ treatment, whereas leaf length, leaf width, and number of leaves remained same. In WW conditions, plant height or vine length of sweet potato pretreated with 34 μM PBZ was reduced by 18.78% when compared with control. Under MWD conditions (32.3% SWC), vine length, leaf area, leaf width, and number of leaves of sweet potato cv. ‘Japanese Yellow’ were maintained (with only 6.96, 2.54, 4.25, and 6.04% reduction, respectively) with a lower reduction rate as compared to the cv. ‘Tainung 57’ (with 31.62, 10.01, 13.78, and 19.36% reduction, respectively). The degree of reduction in each parameter was amplified with the extent of water deficiency and genetic variations, especially in cv. ‘Tainung 57’ (Table 2). In PBZ-pretreated plants, overall growth performances under WD conditions were alleviated. Interestingly, vine length, leaf area, leaf width, and number of leaves in 34 μM PBZ pretreated plants of cv. ‘Tainung 57’ were maintained in both MWD (10.52, 3.09, 0.79, and 3.13% reduction) and SWD (23.02%, 13.00%, 20.10% and 34.92% reduction) (Table 2).

Free proline in sweet potato cv. ‘Japanese Yellow’ grown under MWD was enriched by 1.87-fold (2.78 μmol g^-1 FW) over the WW control. However, the proline levels of PBZ pretreated and untreated cv. ‘Tainung 57’ were not significantly different when exposure to MWD (Figure 2A). In SWD, free proline peaked (2.81 μmol g^-1 FW) in the positive check cv. ‘Japanese Yellow’, while it was maintained in cv. ‘Tainung 57’ without PBZ treatment (Figure 2B). In addition, free proline content in PBZ-pretreated cv. ‘Tainung 57’ grown under SWD was significantly enriched by 2.94- (2.03 μmol g^-1 FW), 3.84- (2.15 μmol g^-1 FW), and 2.28-fold (2.14 μmol g^-1 FW), when pretreated with 17, 34, and 51 μM PBZ, respectively (Figure 2B).

Total soluble sugar in MWD stressed plants increased by 2.87- (26.85 mg g^-1 DW) and 1.85-folds (27.61 mg g^-1 DW) over the control in cvs. ‘Japanese Yellow’ and ‘Tainung 57,’ respectively. The increase was most prominent at 34 μM PBZ where the total soluble sugar concentration was 49.33 mg g^-1 DW (3.69-fold over control) (Figure 2C). Under SWD conditions, total soluble sugar was enriched in both the positive check cv. ‘Japanese Yellow’ (3.63-folds over control) and the treated cv. ‘Tainung 57’ (1.86-folds over control). Under SWD conditions, there was a significant increase in total soluble sugar content in ‘Tainung 57’, by 3.91-, 2.39-, and 1.53-folds over well irrigated conditions when sprayed by 17, 34 and 51 μM PBZ, respectively (Figure 2D). In WD conditions, the enhancement in sugar classes followed a fructose > glucose > sucrose trend in the leaf tissues and this may play a key role in osmotic adjustment in sweet potato. Sucrose content (6.93 mg g^-1 DW) in 34 μM PBZ pretreated plants grown under MWD was enriched by 6.08-folds over control. Glucose and fructose in cvs. ‘Japanese Yellow’ and ‘Tainung 57’ increased in relation to the degree of WD stress. Glucose and fructose contents peaked at 11.32 mg g^-1 DW (2.87-folds over control) and 31.08 mg g^-1 DW (4.21-folds over control) (Table 3). In SWD, sucrose, glucose, and fructose in positive check cv. ‘Japanese Yellow’ were enriched by 3.97-folds (3.02 mg g^-1 DW), 1.99-folds (13.44 mg g^-1 DW), and 5.19-folds (35.71 mg g^-1 DW), respectively, over the control. In contrast, sucrose content in PBZ-pretreated plants cv. ‘Tainung 57’ in both well irrigated and SWD stress was unchanged (Table 3). However, glucose and fructose in the leaf tissues of 34 μM PBZ-pretreated plants of cv. ‘Tainung 57’ peaked at 13.35 mg g^-1 DW (2.29-folds over control) and 26.94 mg g^-1 DW (2.49-folds over control), respectively (Table 3).

Osmotic potential (Ψ_s) in WD condition was decreased depending on the degree of stress. Leaf Ψ_s in 34 μM PBZ-pretreated plants of cv. “Tainung 57” was similar to the positive check (cv. “Japanese Yellow”) when subjected to MWD (Figure 3A). In SWD, leaf Ψ_s in all the treatments significantly declined (Figure 3B). It was confirmed that free proline and soluble sugar enrichment in the PBZ-pretreated sweet potato had a potential to control the osmotic potential in the leaf tissues, preventing the water loss via stomatal closure. In WD conditions, g_s in 34 μM PBZ-pretreated plants of cv. “Tainung 57” was stabilized in the same pattern as the positive check (Figures 3C,D).

Photosynthetic pigments, i.e., Chl a, Chl b, and C_x+c, in the positive check cv. ‘Japanese Yellow’ under WD conditions were maintained, whereas those in cv. ‘Tainung 57’ were significantly degraded, especially in SWD. Chl a and Chl b in WD stressed plants of cv. ‘Japanese Yellow’ were maintained (with only 3.84 and 3.65% degradation, respectively) under MWD and declined by 5.25 and 2.27%, respectively, under SWD. Total carotenoid (C_x+c) content in leaf tissues of plants under SWD conditions significantly dropped by 14.32%, but was unaffected under MWD conditions (Table 4). In cv. ‘Tainung 57’, Chl b and C_x+c contents significantly declined by 8.30 and 27.41%, respectively, whereas Chl a was stable. Chl a, Chl b, and C_x+c in plants grown under SWD conditions were significantly reduced by 13.97, 6.64 and 34.63, respectively. The Chl b and C_x+c in 34 μM PBZ (5.32 and 7.98%) and 51 μM PBZ-pretreated plants (4.61 and 9.48%) were alleviated when subjected to MWD. In SWD, Chl a and Chl b in 34 μM PBZ (11.71 and 2.82%, respectively) and 51 μM PBZ-pretreated plants (9.59 and 4.90%, respectively) were maintained, whereas C_x+c, a sensitive indicator, was significantly reduced (Table 3).

Maximum quantum yield of PSII (F_v/F_m) in cv. ‘Tainung 57’ grown under MWD was significantly dropped (13.40% diminution), whereas it was maintained in cv. ‘Japanese Yellow’ (only 2.01% diminution) (Figure 4A). In MWD, F_v/F_m in PBZ-pretreated plants was retained. Similarly, F_v/F_m in 34 and 51 μM PBZ-pretreated plants grown under SWD was maintained (only 2.95 and 1.34% diminution, respectively) (Figure 4B). Photon yield of PSII (Φ_PSII) in MWD was maintained, whereas it was declined by 6.8% in 51 mM PBZ-pretreated plants (Figure 4C). In contrast, Φ_PSII in 51 mM PBZ-pretreated plants was maintained (1.58% diminution) similar to cv. ‘Japanese Yellow’. A significant diminution of Φ_PSII (13.08% reduction) was revealed in cv. ‘Tainung 57’ plants without PBZ-pretreatment (Figure 4D). P_n in cv. ‘Tainung 57’ was lower than that of cv. ‘Japanese Yellow’ and significantly dropped when plants exposed to MWD and SWD (Figures 5A,B). In MWD, the P_n in cv. ‘Japanese Yellow’ was maintained (12.6% reduction), whereas it was decreased by 72.3% in cv. ‘Tainung 57’ plants without PBZ-pretreatment (Figure 5A). Moreover, P_n in cv. ‘Japanese Yellow’ grown under SWD was reduced by 37.09% and it was sharply declined by 93.36% in cv. ‘Tainung 57’ non-treated plants. Interestingly, P_n in 34 mM PBZ-pretreated plants under MWD and SWD was alleviated (with only 41 and 47.5% reduction, respectively) when compared with other treatments (Figures 5A,B). In addition, transpiration rate (E) in cv. ‘Japanese Yellow’ and cv. ‘Tainung 57’ pretreated with 34 mM PBZ subsequently grown under MWD was sustained (with only 6.46% reduction) (Figure 5C). In SWD, E was a sensitive parameter that has significantly declined (Figure 5D). A positive correlation between Chl a content and F_v/F_m (R² = 0.60), TC content and Φ_PSII (R² = 0.58), Φ_PSII and P_n (R² = 0.73), P_n reduction and plant dry weight (R² = 0.45) was demonstrated (Figure 6).