Synthesis, phloem mobility and induced plant resistance of synthetic salicylic acid amino acid or glucose conjugates

Abstract BACKGROUND The growing demand for food, combined with a strong social expectation for a diet produced with fewer conventional agrochemical inputs, has led to the development of new alternatives in plant protection worldwide. Among different possibilities, the stimulation of the plant innate immune system by chemicals represents a novel and promising way. The vectorization strategy of an active ingredient that we previously developed with fungicides can potentially extend to salicylic acid (SA) or its halogenated analogues. RESULTS Using the click chemistry method, six new conjugates combining SA or two mono‐ or di‐halogenated analogues with L‐glutamic acid or β‐D‐glucose via a 1,2,3‐triazole nucleus have been synthesized. Conjugate 8a, which is derived from SA and glutamic acid, showed high phloem mobility in the Ricinus model, similar to that of SA alone despite a much higher steric hindrance. In vivo bioassays of the six conjugates against two maize pathogenic fungi Bipolaris maydis and Fusarium graminearum revealed that, unlike SA, the amino acid conjugate 8a with good phloem mobility exerted a protective effect not only locally at the application site, but also in distant stem tissues after foliar application. Moreover, compounds 8a and 8b induced up‐regulation of both defense‐related genes ZmNPR1 and ZmPR1 similar to their parent compounds upon challenge inoculation with B. maydis. CONCLUSION The vectorization of salicylic acid or its halogenated derivatives by coupling them with an α‐amino acid can be a promising strategy to stimulate SA‐mediated plant defenses responses against pathogens outside the application site. © 2022 The Authors. Pest Management Science published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.


INTRODUCTION
Over the last seven decades, the advent of organic chemistry and the emergence of plant protection products contributed to an increase in crop yields that was previously technically impossible. However, undesirable effects of agrochemicals on human health and the environment have soon become apparent. 1,2 In addition, repeated use of single-site acting compounds caused pesticide resistance in pests, which in turn resulted in raising the economic costs of agrochemical use. 3 Fungicide resistance in some pathogens was detected after only 2 years of product introduction. 4 Hence, there is an urgent necessity for the innovation of green pest management technologies, 5,6 including new and safer agrochemicals. 7 In the long-term competitive host-pathogen coevolution, plants have formed multilevel defense mechanisms to combat pathogen infection, including increasing cell wall strength, synthesizing antibacterial secondary metabolites, expressing pathogenesisrelated proteins and hypersensitive response. 8,9 One particular inducible systemic immune response, known as systemic acquired resistance (SAR), mediates long-lasting broad-spectrum resistance to a wide range of pathogens in uninfected tissue to prevent second infection. 10,11 At present, the development of certain natural and synthetic chemical inducers that can trigger similar plant defense systems has become an important way to promote green plant protection. [12][13][14] Exogenous application of defense-related hormone salicylic acid (SA) and SA analogs has been reported to induce SAR-like responses. 12 SA showed fungistatic effect for limiting the growth of parasitic fungi in vitro 15 or in vivo when it is brought exogenously to the plants. 16 However, it is rapidly metabolized, 17,18 which can limit its practical use. To overcome this problem, synthetic analogues have been proposed such as acibenzolar-Smethyl 19 and 2,6-dichloroisonicotinic acid (INA). [20][21][22] The strategy proposed here is different and involves the prodrug concept. In previous works, it was shown that by coupling an active ingredient to a nutrient, mainly an ⊍-amino acid, the resulting conjugates could be recognized and manipulated by the plasma membrane transporters of the sieve element-companion cell complex and thus be translocated to new growth via the phloem sap. [23][24][25][26] The nutrient promoiety, amino acid or sugar, 27,28 as well as the structure of the spacer arm that connects the active substance to the nutrient, 29 play key roles in the rate of conjugate translocation and bioactivation. In some cases, it has been shown that the conjugates released the active substance progressively and thus exhibited prodrug behavior. 24,25,30 The objective of this work was to synthesize new conjugates associating salicylic acid or two mono-or dichlorinated analogues with a nutrient promoiety. Regarding the latter, L-glutamic acid was first chosen because its additional carboxylic acid function was necessary to ensure the binding with the active ingredient while keeping free ⊍-amino acid function, 31 thus allowing the recognition by the amino acid transporters of the plasma membrane. In order to study the influence of the nutrient on the biological properties, a similar series with ⊎-D-glucose was also prepared. The binding between salicylic acid or its halogenated analogues and the nutrient was achieved by a click chemistry procedure. The phloem systemicity of the six resulting conjugates was then estimated using the Ricinus model. Finally, the defenseinducing activities of the resulting conjugates were evaluated on maize seedlings inoculated with two pathogens that are responsible for southern corn leaf blight and maize stalk rot, respectively. The expression profiles of two defenserelated genes ZmNPR1 and ZmPR1 were also studied after treatment with the conjugates.

Synthesis
Some reactions were carried out under nitrogen. All reactions were monitored by TLC analysis using Merck silica gel 60F-254 thin-layer plates. Column chromatography was carried out on silica gel Merck 60 (0.015-0.04 mm). Melting points were determined on a Büchi B-540 melting point apparatus and are uncorrected. 1 H and 13 C NMR spectra were performed in CDCl 3 or DMSO-d 6 using a Bruker AVANCE 400 MHz or a Bruker Ultra-shield™ 500 Avance NEO spectrometer, respectively at 400 or 500 MHz frequencies for 1 H experiments and 101 or 126 MHz for 13 C experiments. DEPT (90 and 135) and 2D experiments ( 1 H 13 C and 1 H 1 H) were used to confirm the NMR peak assignments. Chemical shifts are reported as ∂ values in parts per million (ppm) relative to tetramethylsilane (TMS) as internal standard and coupling constants (J) are given in hertz (Hz). The following abbreviations are used to describe peak patterns when appropriate: s (singlet), d (doublet), t (triplet), q (quartet), m (multiplet). High-resolution mass spectra were obtained on a Bruker Q-TOF Impact HD spectrometer using an electrospray ionization source (ESI).

Plant material and fungal strain 2.3.1 Systemicity test
Castor bean seeds (Ricinus communis L. cv Sanguineus), obtained from Zibo Academy of Agricultural Sciences (Shandong, China) were placed in wet cotton wool for 24 h at 27°C ± 1°C prior to sowing in vermiculite watered with tap water. Seedlings were grown in a humid atmosphere (80% ± 5%) at 28°C ± 1°C.

Activity on plant defense response
Two pathogens were used in this study. Bipolaris maydis, the causal agent of southern-corn leaf blight disease, was isolated in Gongzhulin, Jilin, China. Fusarium graminearum is a soil-born pathogen that causes stalk rot of maize, which was also isolated in Gongzhulin, Jilin, China. B. maydis and F. graminearum were grown on oat-meal agar (OA) and potato dextrose agar (PDA) medium plates at 28°C before collecting conidia for inoculation, respectively. Maize B73 inbred line was grown in Conviron growth chamber at 28°C with a 16 h photoperiod for 2-3 weeks.

Phloem sap collection and analysis
The sap collection method was similar to that previously described. 32 The phloem sap was analyzed using an Agilent Technologies 1260 high-performance liquid chromatography after dilution with UHQ grade water (1:9; v/v). An Agilent SB-C18 reversed-phase column (length 250 mm, internal diameter 4.6 mm, 5 μM) was used at a flow rate of 0.8 mL min −1 in accordance with the procedure set out in Table 1. The injection volume was 10 μL.

Evaluation of activity on plant defense response
The in vitro fungicidal activities of SA conjugates against B. maydis were evaluated on OA plates. The compound was dissolved in 1 mL of absolute ethanol and diluted with 5 mL of sterile water containing 0.1% Tween 80 at the concentration of 10 mM. Then, 1 mL of mixture solution was added to 9 mL OA medium. The final concentration of each compound was 1 mM in Petri dishes. The same ethanol solution without tested compound was served as control. The plates were inoculated with agar plugs (5 mm diameter) from a 6-day-old growing colony of B. maydis strain and incubated in the dark at 28°C. The colony diameter was measured 6 days after inoculation. There were three replications for each treatment.
Maize infection assays were performed using inbred line B73. Tested conjugate was sprayed (2 mL per plant) on the leaves of maize seedling at a concentration of 1 mM. The conjugate was solubilized in ethanol (5% of final volume of application solution) and diluted with 10 mM MES buffer solution (pH 5.0) containing 0.15% Silwet-L77. The same application solution without tested compound served as a control. After 2 days, the two pathogens were individually inoculated to pretreated maize seedlings. Detached-leaf-spotting assays were used for the inoculation of B. maydis according to a previous described method. 33 The conidial suspension of B. maydis was prepared at a final concentration of 1 × 10 5 conidia/mL by flooding 12-day-old B. maydis culture plates with sterile distilled water containing 0.1% (v/v) Tween-20. Two droplets (10 μL each) of conidial suspension were spotted on to a detached maize leaf (10 cm in length) from conjugate pretreated plants. The inoculated leaf was kept in a Petri dish that contained 0.1% 6-benzylaminopurine sterile water and then transferred to a chamber at 28°C and 90% relative humidity for 4 days. The lesion size was determined using ImageJ software after photographing. Artificial inoculation of F. graminearum on maize seedling stem was performed as described previously by Sun et al. 34 10 μL of conidial suspensions of F. graminearum (1.0 × 10 6 conidia/mL in 0.1% Tween-20) was dropped to the wounded point of seedling stem using a pipette. The seedlings were maintained under a condition at 28°C and 90% relative humidity for 3 days without moving. Disease severity assessments were performed using a scale ranging from 1 to 5 described by Sun et al. 34 The disease severity (%) = ∑ (Number of plants in that rating × rating)/(total number of plants assessed × maximum rating) × 100. There were three replications for each treatment.
For the analysis of defense-related gene expression, maize seedlings were also pretreated with 1 mM SA conjugates following foliar application. Plants treated with blank application solution were used as control. The expression of the non-expressor of pathogenesis-related genes 1 (NPR1) and the defense marker gene pathogenesis-related-1 (PR-1) was measured using quantitative real-time PCR (qRT-PCR) with or without pathogen challenge. In the absence of challenge, the leaves at the same positions of seedlings were harvested for RNA extraction 24 h after chemical treatment. In B. maydis challenged group, the abovementioned conidial suspension of B. maydis (1 × 10 5 conidia/mL) was sprayed on the maize plantlets (0.5 mL per plant) 48 h after chemical treatment, then the inoculated leaves at the same positions were collected 24 h post-inoculation. Total RNA was extracted using Plant Total RNA Isolation Kit (FOREGENE, Chengdu, China). First-strand cDNA was subsequently synthesized using a One- Step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech). Primers for ZmNPR1 were 5 0 -ACAAACTGCTGATGGTGA-TACG-3 0 (forward primer) and 5 0 -GCAACTTCACAAGCTCAACATC-3 0 (reverse primer). Primers for ZmPR1 were 5 0 -CCACTACGGGGAGAA-CATCT-3 0 (forward primer) and 5 0 -AGTTGCAGGTGATGAAGACG-3 0 (reverse primer). qRT-PCR was performed by a 7500 Real-Time PCR System (Applied Biosystems, CA, USA) using the kit of RealStar Green Fast Mixture (Genestar). The Expression values were normalized to the expression of ACTIN1 using the primers Actin1-F (5 0 -GATTCCTGG-GATTGCCGAT-3 0 ) and Actin1-R (5 0 -TCTGCTGCTGAAAAGTGCTGAG-3 0 ). The transcription level was calculated using the 2 -ΔΔCT method. 35 The experiments were repeated three times.

Physicochemical properties
Physicochemical properties and descriptors were predicted using ACD/Labs Percepta 2015 release (Build 2726) software from Advanced Chemistry Development, Inc. (Toronto, Canada). The calculated properties (Table 2) were chosen according to their influence on passive membrane transport in plants. 36

Statistical analyses
To examine the in vitro activity of the different products on the radial growth of B. maydis, three assays were performed for each compound. The Kruskal-Wallis nonparametric test coupled with Dunn's multiple comparison test was used to assess statistically significant differences between the control and the nine treated sets, assuming significance at P ≤ 0.05.
To evaluate the protective effect of parent compounds and the conjugates against B. maydis and F. graminearum on maize, the assays using 12 detached leaves or three pots with five plants were performed for every product, respectively. A one-way ANOVA followed by Tukey's HSD test was performed to assess statistically significant differences between the control and the treated sets or among the different compounds tested for B. maydis, assuming significance at P ≤ 0.05.
To investigate the expression of the defense-related genes ZmNPR1 and ZmPR1 in maize leaves, a representative experiment with three samples was selected from three replicates showing the same trend. A one-way ANOVA followed by a Dunnett's multiple comparison test was performed to assess statistically significant differences between the control and the treated sets.

RESULTS AND DISCUSSION
3.1 Synthesis of L-glutamic acid and ⊎-D-glucose conjugates In previous work, we described the three-step synthesis of conjugates that associated fenpiclonil, a fungicide from the phenylpyrrole family, with a nutrient that could be an amino acid or a sugar. While an active transporter recognizes these compounds, they are manipulated in different ways. The fenpiclonil-glutamic acid conjugate showed a much more favorable phloem mobility than fenpiclonil-glucose conjugate. 27 Similarly, the structure of the spacer arm that connects fenpiclonil to the nutrient greatly influences the phloem mobility of the conjugate. 29 Surprisingly, the latter has proven to be an excellent tool to study the properties of sucrose carriers in plants. 28 In this work, we have developed a four-step method for the synthesis of new conjugates associating salicylic acid or two chlorinated analogues with L-glutamic acid or ⊎-D-glucose: (i) the first step consisted of obtaining azides from glutamic acid or glucose whose reactive functions were protected to avoid unwanted side reactions in the following stages; (ii) the second step allowed to obtain propargyl derivatives from salicylic acid and its halogenated analogues that will be coupling partners for subsequent click chemistry reactions; (iii) in the third step, the two derivatives obtained during the previous stages were condensed by a click chemistry process to give the protected conjugates with a spacer arm containing a 1,2,3-triazole ring and finally, (iv) the ⊍-amino acid function of glutamic acid or hydroxyl groups of glucose were deprotected to give the desired conjugates. The series of six compounds thus obtained is coherent for further structure-activity relationship studies: the active part of the conjugate is represented by salicylic acid or by two monoor dichlorinated analogues, the part related to the vectorization of the conjugate is a nutrient which can be an ⊍-amino acid or a sugar, while in all cases the spacer arm is linked to the active part of the molecule by an amide bond and includes a 1,2,3-triazole ring in its structure.
3.1.1.1. Experimental procedure for the synthesis of derivative 3. To a solution of 2,3,4,6-tetra-O-acetyl-⊍-D-glucopyranosyl bromide (2.02 g, 4.91 mmol, 1 equiv.) in an acetone/water mixture (60:10 mL), sodium azide (1.78 g, 27.4 mmol, 5.6 equiv.) was  The second part of the synthesis was the preparation of salicylic acid alkynes 5a-c in a one (5a) or two steps (5b, 5c) from acetylated salicylic acid or its halogenated analogues. The first step was the acetylation reaction of 5-chloro and 3,5-dichlorosalicylic acids in the presence of a catalytic amount of sodium acetate in acetic anhydride under heating to afford compounds 4b and 4c in 77% and 85% yields, respectively. This hydroxyl group protection step was necessary to avoid the formation of the mesomeric form that occurs under basic conditions and significantly reduces the activation ability of (1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (EDCl) on the carboxylic acid function. Next, aspirin and its chlorinated analogues 4b-c were condensed with propargylamine in a basic medium (DMAP) and in presence of the coupling agent EDCl to afford propargylamides 5a-c in 50%, 40% and 25% (c) yields, respectively. The presence of one or two chlorine atoms was clearly correlated with lower yields, probably due to their electron withdrawing effect, which decreases the reactivity of the carboxylic acid function.
3.1.2.1. Acetylation of 5-chloro or 3,5-dichlorosalicylic acids. To a suspension of 5-chlorosalicylic acid (4.00 g, 23.2 mmol, 1 equiv.) or 3,5-dichlorosalicylic acid (5.60 g, 27.0 mmol, 1 equiv.) in acetic anhydride (10 mL, 105.8 mmol, 4.6 or 3.9 equiv., respectively), sodium acetate (0.30 g, 3.66 mmol, 0.16 or 0.14 equiv., respectively) was added. The suspension was stirred under heating (60°C) for 3 h. The mixture was then cooled to room temperature and washed with cold water. The resulting precipitate was isolated on a büchner funnel, washed with cold water and dried in an oven to obtain 2-(acetyloxy)-5-chlorobenzoic acid 4b (3.85 g, 17.9 mmol, 77% yield) or 2-(acetyloxy)-3,5-dichlorobenzoic acid 4c (5.75 g, 23.1 mmol, 85% yield).    (5b). Compound 5b was obtained using the same procedure as described for compound 5a, except for DMAP which was used at 0.2 equivalent relatively to the product 4b. The crude product was purified by column chromatography on silica gel gel using pentane/ethyl acetate (19:1) as eluent to afford 5b (0.79 g, 3.77 mmol, 40% yield) as a white solid. Rf = 0.61 (CH 2 Cl 2 ); mp = 167°C. 1 Compound 5c was obtained using the same procedure as described for compound 5b, except for reaction conditions where, after 1 h at 0°C as described above, an additional 5 h reflux heating was required followed by cooling to room temperature overnight. The crude product was purified by column chromatography on silica gel gel using pentane/ethyl acetate (19:1) as eluent to afford 5c as a white solid in 25% yield. Rf = 0.66 (CH 2 Cl 2 ); mp = 182°C. 1   3.1.4 Coupling azido derivatives 2 and 3 with propargyl derivatives of salicylic acid or its chlorinated analogues 5a-c by click chemistry (Figs 1 and 2; compounds 6a-c and 7a-c) The third step of the synthesis pathway was to react the azido derivatives 2 and 3 with the propargylic amides 5a-c via a click chemistry process. This 1,3-dipolar cycloaddition reaction leads to the introduction of a spacer arm between the parent molecule and the nutrient that includes a 1,2,3-triazole ring in its structure. The reaction is catalyzed by active Cu(I) generated in situ by reducing Cu(II) salts (copper sulfate) with sodium ascorbate in a heated tert-butanol/water medium. These reaction conditions allowed to get specifically the 1,4-disubstituted regioisomers 6a-c and 7a-c. The L-glutamic conjugates 6a-c were obtained with 41%, 44% and 74% yields, respectively. The ⊎-D glucose derivatives 7a-c were obtained with 83%, 47% and 42% yields, respectively.

2-[4-[(3,5-Dichloro-2-hydroxybenzamido)methyl
]-1H-1,2,3-triazol-1-yl]-6-(ethanoyloxymethyl)tetrahydro-2H-pyran-3,-4,5-triyl triethanoate (7c). Compound 7c was synthesized using the same procedure as described for compound 7b, but with a heating temperature raised to 60°C for a six-hour period in order to have a complete conversion. The purification procedure was also the same as described above for 7b and this resulted in compound 7c as a white solid in 42% yield. Rf = 0.43 (CH 2 Cl 2 /EtOAc 9:1); mp = 225-226°C. 1   The last step in the synthesis leading to the desired conjugates is the removal of the protecting groups from the ⊍-amino acid function on the 6a-c products and from the glucose hydroxyl groups on the 7a-c products. The tert-butyl protecting groups (ester and carbamate) of the amino acid function of the products 6a-c were removed under acidic conditions with trifluoroacetic acid in dichloromethane to afford the final amino acid conjugates 8a-c in 70%, 42% and 41% yields, respectively. The acetyl protecting groups of the hydroxyl functions of the ⊎-D glucose derivatives 7a-c were removed with sodium methoxide generated in situ. Then, the resulting alkoxides were neutralized using an acidic cation exchange resin (Amberlite IRN 77), giving the final glucose conjugates 9a-c in 80%, 38% and 74% yields, respectively.  ((2R,3R,4S,5S,6R)-3,4,5-trihydroxy-6-(hydroxymethyl)tetrahydro-2H-pyran-2-yl)-1H-1,2,3-triazol-4-yl)methyl)benzamide (9c). Compound 9c was obtained in an identical manner to that used to get 9a, as a white solid in 74% yield. Rf = 0.01 (EtOAc); mp = 91-93°C. 1  , H e ), 3.21 (t, 3 J = 9.0 Hz, 1H, H g ). 13  3.2 Comparison of the phloem mobility of compounds 8a-c and 9a-c using the Ricinus model After the cuticle, the plasma membrane is the second physical barrier that must be passed for the uptake and translocation of xenobiotics in plants. Systemic products must cross the cell plasma membrane at least once to enter the conducting cells of phloem. 25 Therefore, membrane permeation mechanisms are key factors for the long-distance transport and distribution of xenobiotics in plants. In this context, it became possible to modulate and control the distribution of an active ingredient within the plant by associating it with a vector, which is called the vectorization process. In the case of molecular vectorization used in this work, a vector group (⊍-amino acid, sugar) is associated with a biologically active molecule (i.e., salicylic acid), allowing the resulting conjugate to be recognized and manipulated by membranebased active nutrient transporters. 25 For example, the accumulation of Lys-2,4D, a synthesized ⊍-amino acid conjugate, in roots was 5-to 10-times higher than that of its parent compound 2,4D after foliar application. 24 Given these considerations, the phloem mobility of conjugates was estimated on the Ricinus model, which is a widely used plant model to evaluate phloem systemicity of xenobiotics. 29 As shown in Fig. 3, SA exhibited very good phloem mobility in Ricinus seedling. The concentration factor in phloem sap (the ratio of the concentration in phloem sap/the concentration in the incubation medium) was about 5.7, which was consistent with our previous study (about 6.9). 32 The phloem mobility of 5-ClSA and 3,5-diClSA was much lower than that of SA, showing a concentration factor of 2.7 and 0.6 in phloem sap, respectively. The differences in mobility between the three parent compounds can be attributed to their ability to cross the plasma membrane and to evaluate this, different chemical descriptors or physicochemical properties are commonly used to predict the diffusion of small molecules across human membranes. This approach was later extended to the plant field. 36 The chemical descriptors that are considered in the Lipinski 37 (Molecular weight <500 Da; Hydrogen bond donors ≤5; Hydrogen bond acceptors ≤10; Log D ≤ 5.0) or Weber 38 (Free rotatable bonds ≤10; Polar surface area-≤140 Å 2 ) rules differ little between the three compounds, with the exception of the distribution coefficient Log D that increases with the number of chlorine atoms in the molecule (Table 2). For the parent compounds, it is thus clear that their mobility in Ricinus is clearly negatively correlated with the number of chlorine atoms in their structure.
Among all of the conjugates, 8a and 8c showed almost the same level of phloem mobility with their parent compound, and the concentration factor in phloem sap was 3.4 and 0.52, respectively (Fig. 3). The phloem mobility of glucose conjugates (9a-c) displayed much lower levels (concentration factors 0.04; 0.05; 0.14, respectively) than amino acid conjugates 8a-c, suggesting that the amino acid promoiety was more favorable to phloem mobility than that of glucose promoiety. This may be due to the higher substrate specificity of hexose transporters than that of amino acid transporters. 27,28 Furthermore, it has been shown that SA could be converted in the cytoplasm into several metabolites such as salicylic acid 2-O-ß-D-glucoside (SAG), which is often the major metabolite. SAG is then compartmentalized in the vacuole by two active transport mechanisms, either through an ATPbinding cassette transporter mechanism in soybean cells, 39 either through an H + -antiport mechanism in tobacco cells. 40 Therefore, it cannot be excluded that the glucose conjugates 9a-c were also readily compartmentalized in the vacuole in the same way.
Considering all the conjugates, their chemical descriptors are very similar (Table 2), with the main exception of Log D as before. Contrary to the parent compounds, no correlation between the number of chlorine atoms in the structure and mobility could be observed. Similarly, unlike SA and its chlorinated derivatives, two parameters of Lipinski rule were violated (HBD ≤5; HBA ≤10) as well as one for Veber rule (PSA ≤140 Å 2 ). Therefore, the results suggest that these six conjugates are unlikely to diffuse across the plasma membrane and carrier-mediated processes may contribute to phloem transport, in particular for the conjugate 8a associating salicylic acid with an amino acid, which shows a remarkable mobility in the Ricinus model.

Increased resistance to maize foliar and root pathogen following treatments
The in vitro and in vivo bioassays were conducted to evaluate the fungicidal and defense-inducing activity of SA conjugates against foliar pathogen B. maydis. The mycelial growth of B. maydis on the OA plate was not significantly inhibited by SA, amino acid conjugates 8a and 8b and glucose conjugate 9b, indicating that these compounds have no direct fungicidal effect against B. maydis at the concentration of 1 mM (Fig. 4). In contrast, the growth of the fungus was inhibited by 43% by 5-ClSA, 47% by 3,5-di-ClSA, 56% by the amino acid conjugate 8c and 39% by the glucose conjugates 9a and 9c. Thus, these products have a direct antifungal activity at 1 mM concentration on the growth of B. maydis in vitro (Fig. 4).
In in vivo experiments, all compounds resulted in significantly increased resistance of maize to prevent B. maydis infection when comparing to the ethanol control (Fig. 5). The three parent compounds showed the highest antifungal activity in vivo, especially for 3,5-diClSA exhibiting 66% reduction of the lesion size. The decreasing rates of the lesion size for each conjugate were 42% for 8a, 43% for 8b, 32% for 8c, 31% for 9a, 28% for 9b and 37% for 9c, respectively, which indicated that amino acid or glucose conjugates had almost the same activity level (no significant difference in each separate group; Fig. 5). Conjugates 8a and 9a exhibited the same level of antifungal activity with their parent compound SA, demonstrating that the addition of amino acid or glucose promoiety did not affect the defense-inducing activity of SA against B. maydis. However, the lesion size was found to be significantly increased when comparing amino acid conjugate 8c and glucose conjugates 9b and 9c to their corresponding parent compounds 5-ClSA and 3,5-diClSA, suggesting that the amino acid or glucose promoiety had negative impact on the activity of chlorinated SA analogues in vivo. For example, 8c showed the highest in vitro antifungal activity, but did not exhibit the same trend in vivo. It should take into account the possibility that the conjugates cross the plasma membrane into cells easier, then they may not be available in the extracellular spaces where much of the fungal infection would occur. Another possibility would be that the chlorine atoms of the parent compounds reduce the affinity for the active sites of the enzymes that hydrolyze the amido bond of the conjugates, thus affecting the release of the chlorinated analogs of salicylic acid. It has been reported that SA can be conjugated with L-aspartate (SA-Asp) as the dominant form in grape (Vitis vinifera) 41 and bean (Phaseolus vulgaris). 42 SA-Asp is an inactive form of SA, 43 thus there are certain enzymes existing in plant to release SA from its amino acid conjugates. Future metabolic studies will be conducted to clarify this point.
As the same leaves were treated with the tested compound and inoculated with B. maydis, it was difficult to discriminate whether the antifungal effect against B. maydis was the result of the induced systemic resistance in plants or the direct fungicidal effect of compound. Thus, further bioassays in vivo were conducted to investigate the protective effect of SA conjugates on maize stalk rot caused by F. graminearum (Fig. 6(A)). The spore suspensions of F. graminearum were inoculated on the stem base of maize seedling 2 days after chemical treatment on foliar tissues. As shown in Fig. 6, parent compounds SA and 3,5-diClSA did not show protect effect against F. graminearum under the present experimental conditions, and 5-ClSA slightly reduced the disease severity of inoculated plants by 13.4%. Conjugates 8a and 8b treatment significantly reduced the disease severity by 18.7% and 24.0%, respectively, while there was no beneficial effect for other conjugates (Fig. 6(B)). This is consistent with a previous study in which exogenous application of SA to foliar tissues did not activate defense gene expression in the roots of Arabidopsis 44 even though SA exhibited very good phloem mobility. This may be due to the fact that SA can be quickly metabolized in plants. 17 Considering the phloem mobility of the conjugates (Fig. 3), the protective effect of conjugates 8a and 8b in the stem could be due to their basipetal transport through phloem. With poor phloem mobility, glucose conjugates 9a-c cannot exert their defense-inducing activity outside the application sites. It is known that biological activity of pesticides can be significantly affected by the uptake and translocation of pesticide within plants, especially for the foliar-applied pesticides, of which the sites of action may be distant from the point of application. 45 3.4 SA conjugates induced the expression of defenserelated gene ZmNPR1 and ZmPR1 upon pathogen challenge The expression analysis of two defense-related genes was performed during the plant defense responses following the treatments of SA conjugates. NPR1 and PR1 play important roles in SA signaling pathway in plants. In Arabidopsis, NPR1 was found to be the receptor for SA, and the binding of SA to NPR1 was a prerequisite to the transcription of PR1. 46 As shown in Fig. 7, in the absence of pathogen challenge, only parent compounds directly , Radial growth measurements of the colonies; n = 3 assays, mean ± CI 95%. The Kruskal-Wallis test was used to assess statistically significant differences between the control set and the product sets at the 5% probability level. S, significant; NS, non significant. enhanced the expression of ZmNPR1 and ZmPR1 24 h after treatment, while almost no significant change of gene expression was induced by conjugates ( Fig. 7(A), (C)). However, following exposure to B. maydis, both parent compounds and conjugates increased the expression of ZmNPR1 and/or ZmPR1 in maize leaves at 24 h post-inoculation, demonstrating that SA conjugates were capable of inducing SA-mediated defense responses upon pathogen attack, except for conjugates 8c and 9a. A 3.5-fold significant increase in ZmNPR1 expression was observed compared to control after conjugate 9b treatment (Fig. 7(B)). The plants treated with conjugate 8b had highest expression (2.3-fold) of the SAresponsive marker gene PR1 (Fig. 7(D)). Amino acid conjugates 8a and 8b induced up-regulation of both genes similar to their parent compounds at 24 h post-inoculation, while glucose conjugate 9b and 9c only increased the expression of ZmNPR1 showing different patterns with their parent compounds at the same time.
These results may be due to the penetration ability of the conjugates across the plasma membrane. The absorption and translocation of glucose conjugates in plant tissues were slower than that of amino acid conjugates according to their physicochemical properties and phloem mobility. Thus, the expression profile induced by glucose conjugates may result from a delayed response. NPR1 was well known as a key molecule involved in SA signaling, and the npr1 mutant exhibited enhanced disease susceptibility. 47 The transcriptional level of NPR1 increases 2-to 3-fold following infection with pathogens and SA treatment. 48 SA and its two halogenated analogues directly triggered stronger defense response than that of their conjugates over short time (24 h) at a concentration of 1 mM. However, conjugates 8a and 8b markedly increased the gene expression of two defenserelated genes until the plants were exposed to a challenge infection, suggesting that these conjugates may trigger the plants into a priming phase. Defense priming is a physiological state in which a plant can display longer-lasting activation or attenuated repression of defense upon pathogen or pest attack than unprimed plants. 49 SA can directly induce defenses when applied at high doses, but at low doses it can trigger the establishment of defense priming. 50 Based on prodrug strategy, SA conjugate may act as a novel chemical priming agent, which could be safer to the plant than SA by minimizing fitness costs of resistance. , The nine treated groups were compared with the untreated control by performing an ANOVA followed by Tukey's HSD test at the 5% probability level. S: significant. Different lower-case letters (a, b, c, d) indicated significant differences within each group (parent compounds, amino acid conjugates, glucose conjugates) at the 5% probability level by ANOVA followed by Tukey's HSD test. Different upper-case letters (E, F, G, H, I) indicated significant differences regarding the chlorinated status of the compounds (0, 1 or 2 chlorine atoms) at the 5% probability level by ANOVA followed by Tukey's HSD test. n = 12 samples. (B) Disease severity in the nine treated groups, which were compared with the untreated control by performing an ANOVA followed by Tukey's HSD Test. Bars represent means ± SE (n = 3 pots of 5 plants). Asterisks indicate significant differences from control (*P < 0.05, **P < 0.01, ***P < 0.001, NS: not significant).

CONCLUSION
In conclusion, six conjugates associating salicylic acid, or its analogues monochlorinated in position 5 or dichlorinated in positions 3,5, with glutamic acid or glucose were synthesized, the two moieties being in all cases separated by a spacer arm containing a 1,2,3-triazole ring. Depending on the conjugates, the syntheses were performed in four to six steps with generally satisfactory yields. Phloem mobility assays using the Ricinus model showed that the conjugates with an ⊍-amino acid promoiety were more favorable to phloem systemicity than those with glucose promoiety. In addition, conjugates 8a and 8c retained a concentration factor in the phloem sap of the same order of magnitude as the parent compounds. Conjugate 8a exhibited the best phloem mobility among all conjugates with a concentration factor of 3.4. Moreover, the amino acid conjugates 8a and 8b, as well as the glucose conjugate 9b, had no direct action on the in vitro growth of B. maydis, the agent of southern corn leaf blight. In contrast, conjugates 8c, 9a and 9c showed a moderate inhibitory effect on mycelial growth of B. maydis. All six conjugates showed a statistically significant protective effect on the size of necroses induced by B. maydis on leaves, even the effect appeared to be weaker than the parent compounds especially for the chlorinated molecules. It should be noted that conjugate 8a reduced necrosis size by 42%, showing similar activity with SA. When tested in vivo against F. graminearum, conjugates 8a and 8b also showed protective effect on stem-inoculated maize seedlings after foliar application. The expression of defense-related gene ZmNPR1 and ZmPR1 were up-regulated by SA conjugates upon pathogen challenge. Conjugates 8a and 8b increased almost the same level of gene expression as their parent compounds when maize plants were exposed to challenge inoculation with B. maydis. Future metabolism studies will be performed to clarify whether SA could be released from the conjugate. Anyway, combining the results of the different bioassays, it seems that the conjugate 8a is the most promising candidate of this series to stimulate plant defenses, with very good phloem mobility in plant, a preventive effect against pathogens similar to that of salicylic acid and an inducing activity of defense gene at the same level as the parent molecule. The results were from one representative replicate among three independent experiments showing the same trends. Bars represent means ± SE (n = 3 samples). Asterisks indicate significant differences from control in a Dunnett's multiple comparison test following a one-way ANOVA (*P < 0.05, **P < 0.01, ***P < 0.001, NS, not significant).