Soybean leaves are consumed in Asian countries including China as a seasonal vegetable or as preserved leaves [1,2]. Consumption of soy leaves or their extracts has been implicated in preventing type 2 diabetes, obesity, heart disease and cancer through a multitude of mechanisms such as enhancing pancreatic β-Cell function and suppressing hepatic lipid accumulation [1], downregulating adipogeneic transcription [3], inhibiting α-glucosidase [4], decreasing non-HDL to HDL cholesterol ratio [5], relaxing carotid arteries [6], inhibiting fatty acid synthase [7], cancer cell specific cytotoxicity [8], and reducing DNA damage [9]. Much of the bioactivity of soy leaves is thought to be linked to isoflavonoids, kaempferol glycosides and pterocarpans [10], which are absent or found in extremely low levels in soy seed [10,11]. Kaempferols’ antioxidative, free radical scavenging and anti-inflammatory properties [12,13] also play a synergistic role in the aforementioned health benefits.

In plants, while ~350 known kaempferol glycosides are distributed ubiquitously from Bryophytes to Anthophytes, their specific functions are still poorly understood [14]. Their suggested or observed functions include UV protection [15], antioxidant activity [16], phytoalexins against pathogens [17,18], role in infestation of N fixing bacteria [19], attraction of feeding animals [20,21], flower color [22,23], seed production [24] and plant development [25].

In soybean, the kaempferol aglycone has shown inhibitory effects on photosynthesis [26,27] but its glycosides were less influential [26]. However, mesophyll cells of mature soy leaves had no kaempferols or their glycosides indicating that they may not directly affect photosynthesis [28]. Epidermis appears to be the primary leaf tissue of kaempferol accumulation in soy also as is the case in pea leaves [29] consistent with their UV protection and anti-pathogen functions. Soybean leaf tissue undergoes a programmed shift from isoflavone to flavonoid metabolism 3 days after germination, and mature soybean leaves are composed mainly of glycosides of kaempferol [30]. Kaempferol glycosides have been shown to vary in the leaves of different soybean cultivars [31], and certain combinations of kaempferol glycosides were associated with photosynthetic rate [32].

How plants balance the allocation of resources derived from carbon assimilation between the growth-related primary metabolism vs the defense-related secondary metabolism is an intriguing and complex phenomenon. Among the explanations on how plants achieve this balance, the “growth-differentiation balance model” [33-35], considered to be the most integrative [36], stipulates that the defenserelated secondary metabolism is increased under the conditions of lower than maximum gross productivity or at least moderate stress. How kaempferol levels in soy leaves change under the stress treatments that affect carbon assimilation, and the correlation between kaempferol levels and specific gas exchange variables under such treatments is not well-understood. Given the variety of useful bioactivities of soy kaempferols, finding stress treatments that increase kaempferol yield and are environmentally sound is important. Thus, the main objectives of this study were to:

a) determine the physiological responses of leaf carbon assimilation and photosystem function to two concentrations of exogenously applied ABA and two types of mechanical wounding in young soy leaves and,

b) determine the concentrations of major kaempferol glycosides in the leaf closest to the leaf used for physiological measurements.

Plant material, growth conditions and treatments

Soybean (Glycine max L. Merr., cultivar IA 2032) seeds were obtained from USDA, New Orleans. Presoaked seeds were overplanted in Sungro Metromix 380 in 5 gal (30 cm diam) pots and thinned to one plant per pot at V2 stage. Plants were grown in greenhouse without supplemental light (~1200 μmol m^-2 s^-1 of photon flux during day, and 22-30°C min-max temperature) during Fall. All plants were watered daily to field capacity, and fertilized weekly with Scotts Peters Professional 20-20-20 (10 g/8 L, 250 mL per pot first four weeks and 500 mL thereafter - Scotts Sierra Horticultural Products Co., Marysville, OH). Leaf wounding and ABA (abscisic acid, Sigma Cat. #A1049) spray treatments were done three times at V6, V9 and R1 stages before measurements. For leaf wounding, each fully expanded leaflet of each trifoliate was either pierced at 15 places with a dissecting pin or holed with a paper whole puncher at five places on either side of midrib avoiding major veins at each treatment time. Plants receiving ABA treatments were separated from other plants and sprayed either 100 or 200 μM aqueous ABA solution with a drop of Tween 20 using hand spray bottles to complete wetness at each treatment time. Plants receiving other treatments were sprayed with DI water. In plants with leaf wounding, the third fully expanded leaf and leaves above were left unwounded for measurements taken during R1 stage.

Leaf gas exchange and fluorescence measurements

Middle leaflet of the third fully expanded leaf from top on the main stem was used for measurements with LICOR 6400-40 photosynthesis system with leaf fluorescence chamber attached (LICOR Inc. Lincoln, NE). Measurement conditions were 1300 μmol m^-1 s^-1 photon flux (based on light saturation measured), 200 μmol s^-1 flow, 400 μmol mol-1 CO₂, 23°C block temperature and ~40% sample RH. Fluorometer settings were, 5 intensity, 20 kHz modulation, 10 gain, 0.8 duration, 8 intensity and 20 kHz modulation for flash. Readings of F´m, ΦPSII and ETR were adjusted to compensate for the Rectangular Single Flash used during measurements as per Loriaux et al. [37]. Gas exchange and leaf fluorescence measurements of each leaflet were taken when gs and Pnet stabilized on the digital display.

A/C_i curves and stomatal limitations

Data were collected using A/Ci fluorescence autoprogram feature with same measurement conditions as above. Sample CO₂ concentrations of 400, 300, 200, 100, 0, 400, 400, 600, 800, 1000, 1500 and 2000 μmol mol-1 were used. Resulting Ci (converted to Pa units) and A values were analyzed using Photosyn Assistant ver. 1.2 (Dundee Scientific, Dundee, UK) for rubisco carboxylation rate V_cmax, electron transport driving RuBP regeneration rate Jmax, triose phosphate utilization rate TPU, day respiration Rd and CO₂ compensation point Γ. Stomatal limitation in each plant was computed as l=(A”–A’)/A” for assimilation [38] and as l=(ΦPSII” – ΦPSII’)/ΦPSII” for quantum yield, where l=stomatal limitation, A” or ΦPSII”=reading at 390 μmol m^-2 s^-1 of C_i, and A’ or ΦPSII’=reading at 390 μmol m^-2 s^-1 of sample CO₂ concentration. Second order polynomial equations of Microsoft Excel trend line function (R² ≥ 0.85) were used to derive the respective A and ΦPSII values.

Pigment assays

Two leaf discs (0.6 cm diam) taken from the same leaflet used for above measurements were left in 2 mL methanol at 4°C in dark for 24 h. Extract was centrifuged for 5 min at 10,000 rpm and absorbance was read at λ =665.2, 652.4 and 470 nm using a BioTek Synergy HT microplate reader (Winooski, VT). Concentrations of chl a, b and carotenoids were measured according to Lichtenthaler [39].

Total antiradical activity assay

Antiradical activity was assayed according to modified Schwarz et al. and Yu et al. [40,41]. Briefly, five leaf discs from the same leaflet used for above measurements were ground with liquid N2 and stirred in 800 μL 95% ethanol for 30 min. Homogenate was centrifuged for 10 min at 10,000 rpm. To assay antiradical activity of samples (S), 50 μL of supernatant and 150 μL of 100 μM DPPH (2,2-diphenyl- 1-picrylhydrazyl, DPPH•, Sigma Cat. #D9132) in 95% ethanol were combined in triplicate wells on a 96 well plate, with 50 μL of supernatant and 150 μL of 95% ethanol in triplicate wells as blank. For control (C), 200 μL of DPPH was blanked with 200 μL of 95% ethanol. Ascorbic acid (100 μM) in 95% ethanol was used as standards. Plate was incubated at 37°C while shaking for 45 min before absorbance was read at 515 nm at 37°C in the same plate reader as above in pigment assays. Antiradical activity was measured as ((C-S)/C) × 100, where C and S are absorbance of control and sample, respectively.

Leaf kaempferol assay

The leaf immediately above the leaf that was used for above measurements was used for the kaempferol assay. Approximately 200 mg of the lyophilized leaf was homogenized in 1 mL methanol followed by sonication for 1 h. Samples were filtered (0.4 μm) before injecting 10 μL on to a Waters HPLC system using a 4. 6 × 150 mm C18 column (10 μm particle size). The system included a Waters 600E System Controller combined with a Waters UV-VIS 996 detector. Elution was carried out at a flow rate of 1.0 mL/min with the following solvent system: A=acetonitrile, B=water 0.1% TFA; 5% A for 5 min, then 5% A to 90% A in 24 min followed by holding at 90% A for 6 min. The solvent acetonitrile (HPLC grade Sigma Cat. #271004) was purchased from Sigma Chemical Company. Water was obtained using a Millipore system and used during sample preparation procedures and HPLC analyses. Samples were injected in triplicate and measured using peak areas at 260 nm. Kaempferol glycosides were quantified by comparing peak area responses with those of kaempferol (Cat.#60010 Sigma Chemical Company).

Mass spectral analysis of kaempferol glycosides

The mass spectrometer used was a Finnigan MAT LCQ ion trap (San Jose, CA, USA) equipped with a heated nebulizer atmospheric pressure chemical ionization (APCI) interface. HPLC effluent at 1 mL/ min was introduced directly into the interface without splitting using a source temperature of 500°C. Positive ion mode was used with a sprayer needle voltage of 4 kV. The capillary temperature was 210°C. The full scan mass spectra of the flavones from m/z 100-1000 were measured using 500 ms for collection time and three micro scans were summed. The instrument was set to measure total ion chromatograms (TICs) in full scan MS mode to measure protonated (M+H)⁺ ions.

Statistical analysis

Statistical significance of the treatments for each response variable was tested using general linear model on SPSS version 19.0.0.1 (International Business Machines, [42]). Pairwise comparisons after significant ANOVA were performed with Tukey’s HSD.

Long standing efforts to increase yields of crops including soybean have often involved genetic improvement and development of cultural practices that promote photosynthetic primary metabolism. However, these methods often targeted no or little improvement in the crop’s nutritional or medicinal value resulting from secondary phytochemicals generally elicited under suboptimal abiotic or biotic environmental conditions. Thus, finding agronomic methods that can impart health-promoting phytochemical quality and understanding how these methods impact the balance between carbon assimilatory primary metabolism and the accumulation of the secondary phytochemicals of interest remain important research goals. We focused on ABA and mechanical wounding treatments for several reasons. First, both abiotic and biotic stresses involve ABA as a natural plant signaling molecule and leaf wounding as a result of pest, wind or hail damage. Secondly, these stress treatments do not severely reduce the total leaf area so that the plant’s contribution of photosynthetic primary metabolism toward crop yield, either leaf or seed, is only mildly affected. Thirdly, both treatments are eco-friendly. However, plants may respond differently to exogenously applied vs endogenous ABA [43] and mechanical wounding vs pest damage using distinct pathways for defense [44,45]. Analyses of physiological and phytochemical responses to ABA or wounding treatments in our study, therefore, were attempts to find cultural practices that can influence the balance between the photosynthetic primary metabolism and kaempferol accumulation rather than to directly understand the defense orchestrated by endogenously synthesized ABA through kaempferol biosynthesis under stress or how leaf damage by natural causes would affect kaempferol yield.

Effects of ABA treatments on physiological variables

Both ABA 100 and 200 μM applications reduced the gas exchange variables such as Pnet, gs, E and Ci. However, only the reductions by ABA 200 μM, >20% in each of these variables compared to control, were statistically significant (Table 1). Instantaneous water use efficiency increased 32% in ABA 200 μM-treated plants due to their larger reduction in E than in Pnet compared to control. Furthermore, ABA 200 μM application reduced the variables of energy harvest and transport namely, F´m, ΦPSII and ETR by 16%, 10% and 17%, respectively, compared to control (Table 2). Stomatal limitations of net photosynthesis and quantum yield were 55% and 100% greater, respectively, in ABA 200 μM-treated plants than control (Figure 1). ABA’s role in stomatal closure through a multitude of molecular and physiological mechanisms within the guard cells including ABA receptor-involved activation of genes that encode enzymes [46,47] and other proteins involved in cellular dehydration tolerance [48], collaborative signaling using reactive oxygen species [49], Ca²⁺ channel activation [50], anion efflux causing membrane depolarizationdependent inhibition of inward K⁺ channels [51], effects on aquaporins [52], etc. is well-known. Li et al. [1,3] found that exogenously applied ABA slowed photosynthetic dark reactions generating H₂O₂ and a glut of Ci which then promoted further stomatal closure whereby mainly regulating plant’s water balance. In soybean, Ward and Bunce [53] also found combined and proportional reductions in leaf gas exchange variables and ΦPSII in response to exogenously applied ABA. However, RuBP content increased but RuBP carboxylation efficiency decreased in ABA-treated plants in their study. They also concluded that this decrease in carboxylation efficiency was not due to a change in activation status or total activity of rubisco as ABA treatment had no effect on them [53]. In our study, both V_cmax and Jmax were increased by ABA 200 μM treatment with the quickest response of carbon assimilation to early increases in C_i compared to other treatments (Table 3, Figure 2). However, effects of ABA on V_cmax and Jmax were fairly equal as seen by the same V_cmax:J_max ratio as in control. Also, Rd was relatively high under ABA 200 μM treatment (Table 3) which likely caused higher Γ. Increased biochemical variables of assimilation and Rd of ABA-treated plants in our study are consistent with the observations that exogenously applied ABA increased shoot and root growth, size of young leaves and pod yield in field grown soybean [54].

Figure 1: Stomatal limitations of net photosynthesis (P_net) and quantum yield of photosystem II (ΦPSII) under the treatments. Error bar=SE, N=5. Different letters above the error bar indicate statistically significant difference across treatments for the given variable (P<0.05).

Figure 2: Second degree polynomial A/C_i curves for the plants under different leaf treatments (N=5). Control, y=-4E-05x²+0.0772x-5.1435, R²=0.8274; ABA 100 μM, y=-4E-05x²+0.0761x-4.1612, R²=0.7441; ABA 200 μM, y=-7E- 05x²+0.1052x-9.2175, R²=0.7898; Leaf piercing, y=-4E-05x²+0.0736x-3.6489, R²=0.7016; Leaf hole punching, y=-3E-05x²+0.0667x-5.0474, R²=0.8169.

Table 1: Gas exchange characteristics and leaf temperature under treatments.

Thus, in our study, the main nonstomatal limitation to photosynthesis observed in ABA-treated plants was the reduced ΦPSII. Reduced ΦPSII under ABA treatments has been attributed to impaired trans-thylakoid proton motive force [55]. Consistent with the reduced ΦPSII, F´_m and ETR were also lower in ABA 200 μM-treated plants (Table 2). Most of the chlorophyll fluorescence in a healthy leaf originates from the light harvesting antennae complex of photosystem II [56]. However, leaf chlorophyll and carotenoid concentrations were unaffected by ABA treatment under the conditions of our study (Figure 3). Thus, reduced F´_m likely indicates a transient, rather than a chronic, weakening or underutilization of the antennae strength of photosystem II in ABA-treated plants. This lower energy capture likely caused the reduced ETR as well. However, ETR/P_net ratio was the same (data not presented) in ABA-treated plants and control indicating that thylakoid redox status adjusted in ABA-treated plants to match the carbon assimilation. Thus, alternative electron sinks such as photorespiration may not have been a major reason for the reduced carbon assimilation in ABA-treated plants as also shown by Li et al. [57].

Figure 3: Leaf chlorophyll and total carotenoid contents affected by treatments. Error bar=SE, N=5. Different letters above the error bar indicate statistically significant difference across treatments for the given variable (P<0.05).

Table 2: Leaf photochemical properties under the treatments.

Leaf antiradical activity increased 25% under ABA 100 μM treatment compared with the control (Figure 4). However, the 9% increase in antiradical activity by ABA 200 μM treatment was insignificant (P=0.12). Part of this increased antiradical activity may be attributed to the increased leaf kaempferol levels (see below) which are known free radical scavengers [12,13]. However, given that ABA 200 μM treatment had generally greater leaf kaempferol levels than ABA 100 μM treatment, although statistically insignificant, ABA 200 μM treatment is expected to produce higher antiradical activity. One probable reason for lower antiradical activity in ABA 200 μM application is that this ABA concentration may have decreased other radical scavenging metabolites such as glutathione (GSH) as reported by Okuma et al. [58]. Furthermore, Mittler and Blumwald [49] found that ABA itself can induce reactive oxygen species as part of its signaling network. Therefore, leaf extracts from ABA 200 μM-treated plants may have had more free radicals than the extracts from ABA 100 μM-treated plants.

Figure 4: Leaf antiradical capacity affected by treatments. Error bar=SE, N=5. Different letters above the error bar indicate statistically significant difference across treatments (P<0.05).

Effects of leaf mechanical wounding on physiological variables

Leaf-pierced plants had the highest Pnet and gs although statistically significant only compared to ABA 200 μM-treated plants (Table 1). Similarly, they had the highest ΦPSII, F´m and ETR, significant differences compared to ABA 200 μM-treated plants (Table 2). Leafpierced plants also showed statistically insignificant increases in V_cmax and J_max with an A/C_i response similar to control (Table 3, Figure 2). This strong photosynthetic capacity in leaf-pierced plants was also supported by their 39% higher total chlorophyll and 38% higher total carotenoids compared to control plants (Figure 3). Besides conferring light harvesting complexes and photoprotection, this pigment enrichment may add to the nutritional quality of the young leaves in leaf-pierced plants. Furthermore, leaf-pierced plants had the lowest stomatal limitation to either P_net or ΦPSII (Figure 1) among all treatments. For instance, leaf-pierced plants had 49% and 77% less stomatal limitations to Pnet and ΦPSII, respectively, compared to ABA 200 μM-treated plants although only the 53% less stomatal limitation to ΦPSII was significantly different from the control. Both leaf-pierced and -holed plants had similar levels of antiradical capacity (Figure 4) compared to control suggesting that an ROS boost may not have occurred and a redox status conducive to carbon assimilation was still maintained following wounding. Leaf-holed plants, however, didn’t show above indicators of an elevated photosynthetic capacity but had gas exchange and quantum use variables that were more comparable to control. Furthermore, leaf-holed plants had a reduced slope and lower A_max in A/C_i curves than the other treatment groups resulting in non-significantly lower V_cmax and J_max compared to all other treatments or control. In addition to the greater loss of tissue, total kaempferol levels were slightly lower in leaf-holed plants (Table 4) than leafpierced plants. Liu et al. [59] found that treatment of apple leaves with quercetin or kaempferol inhibited the ABA-induced stomatal closure by decreasing levels of ROS in the guard cells. In our study, leaf-pierced plants had slightly higher leaf kaempferol levels (next section) which along with less tissue loss likely contributed to higher stomatal activity as seen in their lower stomatal limitations to both gas exchange and ΦPSII compared to leaf-holed plants. Antiradical activities (Figure 4) did not parallel the leaf kaempferol levels likely reflecting the different degrees of contribution by non-kaempferol antioxidants to free radical scavenging under the two types of leaf wounding.

Table 3: Biochemical variables of assimilation derived from A/C_i curves.

Peterson et al. [60] reported that leaf mechanical wounding by clipping a piece of each leaflet in a leaf with scissors in six legume crops had no significant effect on photosynthesis 24 hrs after injury. In soybean, previous works found the similar results when leaflets were wounded by either cork-borer or hole-puncher [61]. We were not able to find reports on the effect of soy leaf piercing on physiological variables. However, the differential physiological response that we observed between leaf piercing and holing is consistent with the concept that injury types can be classified into guilds (groups or types) based on the within-group-homogeneities of plant response to those guilds [62].

Table 4: Concentrations of leaf kaempferol glycosides (µg g^-1 DW) detected by HPLC.

Enhanced or sustained P_net and related structural and physiological variables that were observed in response to mechanical wounding can also be attributed to reduced source/sink ratio as observed by others in soybean [63] and other species [64]. In our study, plants were treated through early vegetative stages to early flowering when all the measurements were taken. Thus, wound healing rather than woundinduced leaf senescence was observed during the experiment which probably caused a reduced source/sink ratio requiring more carbon assimilation to provide row materials for wound healing. Furthermore, soluble carbohydrates from current photosynthesis and woundinduced jasmonic acid (methyl jasmonate) were found to co-regulate the expression of genes encoding vegetative storage proteins (VSPs), a group of glycoproteins that accumulate in young shoot tissue as a temporary sink for carbon and nitrogen usable for wound healing in leaf-wounded soy plants [65,66]. When current photosynthesis was inhibited in leaf-wounded soy plants VSP accumulation was negated [66].

Effects of treatments on kaempferol glycoside yield

Six leaf kaempferol glycosides were identified based on UV and MS spectra as shown in the HPLC chromatogram in Figure 5. Using mass spectrometry, K1 was identified as a kaempferol diglycoside (m/z 611, 449, 287), K2 as a kaempferol triglycoside (m/z 757, 741, 595, 449, 287), K3 as a kaempferol triglycoside (m/z 741, 595, 449, 287), and K4, K5, and K6 as kaempferol monoglycosides (m/z 449, 287). Concentrations of all leaf kaempferol glycosides quantified increased under each ABA treatment causing the total kaempferol concentration to rise 42% in ABA 100 μM and 68% in ABA 200 μM-treated plants compared to control (Table 4). However, the difference between the kaempferol glycoside concentrations under the two ABA treatments was insignificant. Similarly, all six kaempferol glycoside concentrations were significantly higher in leaf-wounded plants compared to control except for the 21% increase of K1 in leaf-pierced plants. Total kaempferol glycoside concentration was 55% greater in leaf-pierced plants and 43% greater in leaf-holed plants compared to control. Two leaf wounding treatments also had statistically similar concentrations of kaempferol glycosides (Table 4).

Figure 5: HPLC chromatogram of the kaempferol glycosides in soybean leaves. K1 through K6 were the identified kampferol glycosides.

According to the growth-differentiation balance model, soy kaempferol glycoside concentrations in our study were expected be negatively correlated with photosynthesis and related variables. However, since photosynthesis and related physiological responses to ABA treatments were different from those of wounding treatments and, as far as known to us, correlations between soy leaf kaempferol glycosides and photosynthesis-related physiological responses have not been reported we studied the correlations of kaempferol glycosides to P_net and g_s (Table 5) under the leaf treatments. We found that these correlations differed depending on the leaf treatment and also on the specific physiological variable, P_net vs g_s. For instance, kaempferol glycoside concentrations and P_net showed a more positive correlation in ABA 100 μM-treated plants compared to control with the exception of K1. A similar positive correlation of kaempferol glycoside levels was observed with g_s rather than with P_net in ABA 200 μM-treated plants. In contrast, a negative correlation between kaempferol glycoside concentrations and P_net was seen in leaf-pierced plants while there were no clear correlations between kaempferol glycoside concentrations and photosynthesis-related variables in leaf-holed plants (Table 5).

Table 5: Pearson correlation coefficients between each kaempferol glycosides and two key gas exchange variables (Pnet, net photosynthesis; gs, stomatal conductance) under different treatments.

Fairly positive correlation between kaempferol glycoside levels and gas exchange variables in ABA-treated plants suggests that carbon skeletons feeding kaempferol biosynthetic pathway were more directly dependent on current photosynthetic carbon assimilation compared to leaf wounded plants. This is also consistent with the growthdifferentiation balance model [33-35] which stipulates that primary metabolism parallels secondary metabolism or synthesis of defense compounds under relatively low or restricted resource availability. Under ABA treatments this restriction was likely a physiological condition, the low gs. The greater stomatal limitation to both ΦPSII and Pnet in ABA 200 μM-treated plants may explain their more positive correlation between kaempferols and g_s compared to ABA 100 μMtreated plants. Reasons for the differences in correlation coefficients of kaempferol glycosides with P_net vs with g_s under the two ABA treatments, though not obvious, may be due to the different degrees of association between P_net and g_s and slight differences in the nonstomatal regulation of P_net between the two treatments.

As discussed earlier under the effects of leaf mechanical wounding on physiological variables, leaf-pierced plants had slightly elevated photosynthetic capacity boosted by a stronger pigment bed and an elevated gs especially compared to ABA treatments. Leaf-pierced plants had 5% greater gs than even the control plants though this increase was statistically insignificant. Thus, the negative correlations between the yields of kaempferol glycosides and Pnet under leaf piercing likely corresponds to the more optimum resource availability or conducive conditions for photosynthetic primary metabolism of the growthdifferentiation balance model during which secondary metabolism toward kaempferol accumulation has probably passed its peak but is still higher relative to control. Furthermore, kaempferol glycosides may play a role in the systemic defense against wounding [24] or in wound healing itself. In leaf-holed plants, however, the generally negative correlation between kaempferol glycosides and P_net seen in leafpierced plants was absent, and no discernible pattern of correlation between kaempferol glycoside yields and gas exchange physiological variables was observed (Table 5). Based on the physiological variables discussed earlier and due to greater loss of leaf tissue, leaf-holed plants experienced a different type and severity of stress compared to the leafpierced plants. Thus, the different wound healing and antioxidative demands in leaf-holed plants may attribute to the different patterns of the associations of kaempferol glycoside yields and gas exchange variables compared to leaf-pierced plants. These differential associations of primary vs secondary metabolic responses between the two leaf wounding treatments are also consistent with the withingroup- homogeneities of plant response to different guilds [62].

Kaempferols are produced by the actions of five, namely, glycolytic, pentose phosphate, shikimate, phenylpropanoid and kaempferol flavonoid pathways. Glycolytic and pentose phosphate pathways produce phosphoenolpyruvate and erythrose-4-phosphate which feed shikimate pathway producing phenylalanine. Phenylpropanoid metabolism uses phenylalanine to produce p-Coumaroyl CoA which then is used to produce chalcone isomers that feed kaempferol biosynthesis by falavonol synthase and flavanone 3 hydrolase [67,68]. Buttery et al. [32] reported that certain combinations of the nonallelic flavonol glycoside genes in soybean, especially the genotype that produces kaempferol 2G-glucosylgentiobioside, named K9, was associated with low chlorophyll levels, photosynthesis, stomatal density and specific leaf weight. They suggested that such genotypes are eliminated or reduced in soy breeding programs since photosynthesis is highly correlated with seed yield in soybean. Therefore, the cultivar used in this study, a high yielding IA 2032, is unlikely to have been a K9 line. While the inhibitory effects of kaempferol aglycone on photosynthesis in vitro has been well-known [26,27] glycosides were less effective [26]. However, Cosio and McClure [28] found that kaempferol has no direct inhibitory effect on photosynthesis in soybean since neither kaempferol, nor its glycoside including K9 was found in mesophyll cells but only in epidermis. Thus, inhibitory effects of kaempferol on photosynthesis were suggested to be indirect through its effects on formation of stomata and other features of leaf development [69,70]. Furthermore, aglycones of kaempferol and other related flavonoids such as quercetin which is an early product of flavonoid biosynthetic pathway are found to inhibit polar transport of auxin causing localized auxin accumulation. Auxin may play a role in controlling stomatal opening and resource allocation under stress [71-73]. Thus, the patterns of correlation between kaempferol glycoside levels and gas exchange variables in our study also likely resulted from differential effects of leaf treatments on the growth and resource allocation associated with source-sink balance rather than solely the direct effects of kaempferols themselves [74].

In conclusion, elicitation of kaempferols or their glycosides is an important step toward improving nutritional and pharmaceutical potential of soy leaves whether used as vegetable, ingredients in foodstuff and traditional medicine or raw material for extraction. However, increasing phytochemical quality without compromising the total plant productivity is a challenge as phytochemicals such as kaempferols are produced as part of plant’s stress response. We found that kaempferol-rich young soy leaves can be produced by mild stress treatments by way of foliar ABA spray and wounding to already mature plants thus avoiding major growth effects. Although ABA and leaf wounding influenced photosynthetic primary productivity differently they both increased leaf kaempferol glycoside yields. While ABA-elicited kaempferol accumulation occurred with reduced Pnet it still showed a positive correlation to a given gas exchange variable depending on the concentration of ABA applied. Kaempferol accumulation elicited by leaf wounding occurred with no reduction in Pnet showing different patterns of correlation to gas exchange variables depending on the type of the wounding treatment. Results of this study show that appreciable increase in kaempferol yield and thus phytochemical or nutritional quality of soy leaves is achievable with environmentally friendly foliar treatments.

Authors thank Louisiana Cancer Research Consortium and USDA-Tulane- Xavier-Toledo Research Collaborative for financial support, and Shannon Combe for technical assistance.