Wild-type Arabidopsis thaliana L. ecotype Columbia-0 (Col-0) and mutant plants, whose parental ecotype is an ecotype Col-0, grew in an environmentally controlled growth chamber (21±1°C, 50-60% relative humidity, 12 h day and 12 h night). phytoalexin deficient4 (pad4), salicylic acid induction deficient2 (sid2), nonexpressor of pathogenesis-related proteins1 (npr1), ethylene receptor1 (etr1), jasmonate response1 (jar1) and jasmonate insensitive1 (jin1) mutants were used in this study (Bleecker et al., 1988; Staswick et al., 1992; Berger et al., 1996; Cao et al., 1997; Jirage et al., 1999; Nawrath and Métraux, 1999). Seeds of wild type and mutant plants were sterilized with 50% bleach solution, and kept at 4°C for 3 days until sowing. Plants were watered twice a day until experiments.

Tryptic soy agar media (20 L) (15 g tryptone, 5 g soytone, and 5 g NaCl/L) incubated with B. vallismortis BS07 at 28°C for 7 days were harvested with methanol (20 L × 2) twice and the extracts were concentrated under vacuum. The concentrate was subsequently dissolved in 50% aqueous methanol (500 mL) followed by extractions with n-butanol (4 L × 3). The n-butanol phase was evaporated to dryness, affording 43.5 g of a brown gum, which was chromatographed on a C18 flash column (90 id × 70 mm, C18, 40 × 63 μm, Merck) using a 10% stepwise gradient elution of increasing methanol concentration in H₂O (1 L each) to afford eleven fractions (A to K: 28.0 g, 3.0 g, 2.0 g, 3.0 g, 2.0 g, 3.0 g, 1.5 g, 4.5 g, 2.0 g, 1.5 g, and 1.0 g, respectively). Fractions C (2.0 g) was fractionated on a normal phase MPLC (50 id × 220 mm, Silica Redi Sef R_f: Teledyne Isco) eluted with ethyl acetate/methanol/H₂O (80:20:0 to 30:63:7) to give eighteen sub-fractions (C1 to C18) from fraction C. Sub-fraction C2 (130 mg) was purified by preparative C18 HPLC (Cosmosil, 5C₂₂-AR-II, 20 id × 250 mm, 25-30% aqueous methanol) to [6.0 mg, t_R 25.0 min, cyclo(LAla-L-Ile)], [12.0 mg, t_R 28.0 min, cyclo(L-Ala-L-Leu)], and [2.5 mg, t_R 39.0 min, cyclo(L-Leu-L-Pro)]. The cyclic dipeptides were synthesized according to the previously reported method (Thajudin et al., 2010). Briefly, N-Boc protected amino acid (1 mmol) was dissolved in dimethyl formamide. N,N-diisopropylethylamine (1.1 mmol) and 2-(1H-benzotriazol-1-yl)-1,1,3,3,-tetramethyluronium hexa-fluorophosphate (1.1 mmol) were added at room temperature, followed by addition of amino acid methyl ester HCl salt (1 mmol). The mixture was stirred for 3 hours. The reaction mixture was diluted with ethyl acetate, washed with water, 1 M NH₄Cl solution, saturated sodium bicarbonate solution, and saline. The resulting linear dipeptide was dissolved in water and autoclaved at 130°C for 4 hours and water was evaporated under vacuum. The remaining residue was purified by silica gel column chromatography (hexane/ethyl acetate/MeOH or CH₂Cl₂/MeOH).

Different concentrations of cyclo(Gly-L-Pro), cyclo(L-Ala-L-Ile), cyclo(L-Ala-LLeu), and cyclo(L-Leu-L-Pro) were dissolved in distilled water before treatment. Both 1 and 10 ppm (part per million) of each CDP solution were infiltrated in leaves of Arabidopsis with a needless syringe and 100 ppm of them was applied onto plants with a sprayer 1 day prior to pathogen infection. The CDP-treated plants were cultivated in a typical growth conditions as described above.

A virulent Pseudomonas syringae pv. maculicola ES4326 was employed in this study. The strain was cultured in King’s B liquid media (10 g proteose peptone, 1.5 g K₂HPO₄, 15 g glycerol and 5 mM MgSO₄/L) supplemented with 50 μg/ml streptomycin at 28°C. The freshly cultured strain was diluted to OD₆₀₀=0.0001 (to test disease resistance) or OD₆₀₀=0.01 (to test gene expression) in 10 mM MgSO₄. We infiltrated the diluted bacteria into leaves of wild type and mutant plants with a needleless syringe. After infection, the infected plants were covered with a transparent plastic dome to maintain high humidity, and incubated in a growth chamber during 3 days (21±1°C, 12 h day and 12 h night). Number of bacteria was counted on day 3 after infection.

One ppm of these CDP solutions were infiltrated into leaves of wild-type Arabidopsis. One day after treatment, P. syringae pv. maculicola ES4326 (OD₆₀₀=0.01) was infiltrated into the pretreated leaves with each CDP. During infection, plants were kept in a growth chamber. The infiltrated leaves were collected at the indicated time after infection. Total RNAs were isolated with the TriZol^® reagent according to the manufacturer’s instruction (Thermo Fisher Scientific, MA, USA). To minimize DNA contamination, TURBO DNase (Thermo Fisher Scientific) was treated in the total RNAs at 37°C for 30 min. About 1 μg of total RNA was used to synthesize first-strand cDNAs with Superscript II reverse transcriptase (Thermo Fisher Scientific). qRT-PCR was performed with SYBR Green PCR mixture (Takara Bio, Japan) using the cycling program as following: 95°C for 5 min. followed by 40 cycles at 95°C for 15s, 60°C for 15s and 72°C for 35s (Bio-Rad, CA, USA) (Jung et al., 2009; Lee et al., 2016). Actin2 (At3g18780) was used as the internal control. Oligo-nucleotide sequences used in this study will be provided under a request.

In order to examine whether or not CDPs could inhibit microbial growth in Arabidopsis, first, we synthesized cyclo(Gly-L-Pro) (compound A), that is one of well-documented CDPs from Aspergillus spp, according to the literature (Thajudin et al., 2010; Ortiz-Castro et al., 2011; Li et al., 2012). Additionally, we also recently isolated 17 different CDPs from B. vallismortis BS07 strain. By means of initial screening to identify bioactive compounds, we selected three different CDPs, cyclo(L-Ala-L-Ile) (compound B), cyclo(L-Ala-LLeu) (compound C), and cyclo(L-Leu-L-Pro) (compound D) conferring disease resistance on several crop plants (data not shown). The molecular structures of these CDPs was described in Fig. 1A.

In order to characterize function of these CDP in plants, we used Arabidopsis plants. Wild-type Arabidopsis Col-0 plants were cultivated in potting mixture (Dongbu Farm Hannong, Korea), in an environmentally controlled growth chamber. Different concentrations of each CDP (1, 10, and 100 ppm) were prepared in distilled water. Each solution was treated on plants 1 day prior to inoculation, as mentioned above. A virulent P. syringae pv. maculicola ES4326 strain was infiltrated in the pretreated leaves with the each compound (OD₆₀₀=0.0001). Treatment of 1 ppm of the compounds inhibited Pseudomonas growth in plants, as compared with mock-treatment (Fig. 1B, left). Direct infiltration of 10 ppm of the compounds, however, was not effective, except compound C (Fig. 1B, left), indicating that concentration of the CDPs may be important for controlling bacterial disease. Additionally, number of bacteria was also reduced in plants applied with 100 ppm of each CDP (Fig. 1B, right). Next, to test whether or not these CDPs had a direct antimicrobial activity to phytopathogenic fungi, and oomycetes, we performed a paper disc-agar diffusion assay of the compound. Alternaria alternata, C, acutatum, Fusarium oxysporum, Rhizoctonia solani, and P. capsici were co-incubated with paper discs containing different concentration of the compound. Mycelial growth of these microorganisms was not inhibited by these compounds in vitro (data not shown), which indicate that these CDPs do not have direct antifungal and antioomycete activity. Thus, we hypothesize that biological activity of these CDPs is due to the activation of plant’s built-in defense response, rather than direct inhibition of microbial growth in plants.

As described above, it is known that disease resistance induced by PGPRs usually depends on JA-related defense response (Pieterse et al., 1998; Zamioudis and Pieterse, 2012). On the other hand, systemic resistance initiated by primary local infection with avirulent or virulent pathogens relies on SA-related events both in local and distal tissues and generation of long distance mobile signals in local tissues (Dempsey and Klessing, 2012). However, it is not yet clear which defense-related signaling pathways are activated by these CDPs. To address this question, first we prepared 100 ppm of each CDP and applied them on various Arabidopsis mutants lacking a SA-dependent event as following; pad4 (SA regulation) (Jirage et al., 1999), sid2 (SA biosynthesis) (Nawrath and Métraux, 1999), and npr1 (SA responsiveness) (Cao et al., 1997). Twenty six-day old plants were used for the experiments. One day after treatment, a virulent P. syringae pv. maculicola ES4326 strain (OD₆₀₀=0.0001) was infiltrated into leaves of plants, and number of bacteria was counted in the infected leaves 3 day after inoculation. Bacterial growth was only reduced in sid2 mutant, but not pad4 and npr1 mutants, as compared with that in mock-treated plants, after treatment of compound A, which indicates that compound A requires PAD4 and NPR1 for inducing disease resistance (Fig. 2A). In case of compound B, PAD4, SID2, and NPR1 are inessential to induce resistance response against Pseudomonas infection, because these mutant plants were still sensitive to application of compound B (Fig. 2A). Exogenous treatment of compound C effectively inhibited bacterial growth in leaves of pad4 mutant, but not sid2 and npr1 (Fig. 2A). The result proposes that the compound C requires SA biosynthesis and responsiveness for establishing disease resistance in Arabidopsis. On the other hand, mutation of SID2 and NPR1 did not affect compound D-induced resistance in plants, whereas PAD4 was necessary for the resistance (Fig. 2A). Based on these results, we conclude that each CDP tested in this study can differentially influence SA-dependent defense response in Arabidopsis.

Exogenous treatment of B. vallismortis strain BS07 induces PR1 expression, but not PDF1.2, in Arabidopsis (Park et al., 2013). In general, however, JA-dependent signaling is important for induced-resistance triggered by PGPRs. Since these CDPs were produced by Aspergillus sp. (compound A), and B. vallismortis (compound B, C, and D), we expected that key components of JA-dependent events might also be required for the resistance response induced by these CDPs. To test this, we applied 100 ppm of these CDPs on etr1 (ethylene perception) (Bleecker et al., 1988), jar1 (JA-Ile production) (Staswick et al., 1992), and jin1 (JA responsiveness) (Berger et al., 1996) mutants 1 day prior to P. syringae pv. maculicola ES4326 inoculation (OD₆₀₀=0.0001). Treatment of both compound A and B failed to suppress bacterial growth in etr1 and jin1 mutants, whereas bacterial growth was decreased in leaves of jar1 mutants, as compared with those in mock-treated plants (Fig. 2B). These indicate that both compound A and B require ETR1 or ETR1-regulated proteins and JIN1 to trigger disease resistance against Pseudomonas infection. In case of compound C and D, numbers of bacteria in etr1, jar1, and jin1 mutants were comparable with those in mock-treated plants, which proposes that ethylene and JA-related events are crucial for compound C- and/or D-induced disease resistance in Arabidopsis.

There are two different mechanisms of action of exogenous stimuli to execute induced-resistance in plants. Benzothiadiazole (BTH), a well-known elicitor for plant disease resistance, directly activates SA-related defense response after treatment (Lawton et al., 1996). Priming of defense response is another mechanism of action for induced-resistance (Conrath et al., 2002). Exogenous treatment of chemical primers does not cause any detectable dramatic shift of defense response before pathogen infection. However, immune response is strongly activated in primed plants more than in non-primed plants (Jung et al., 2009; Conrath et al., 2015). Indeed, expression levels of a few defense-related genes in Arabidopsis treated with compound A, B, C, and D were similar to those in mock-treated plants in the absence of pathogen infection (data not shown). The observation allowed us to examine whether or not these chemical inducer acted as primers for establishing disease resistance response in Arabidopsis. To test the possibility, 100 ppm of each CDP was applied on wild-type Arabidopsis. At 24 h after treatment, P. syringae pv. maculicola ES4326 (OD₆₀₀=0.01) were infiltrated into leaves of CDP-treated plants, and the infected leaves were taken 12 h and 18 h after inoculation (Fig. 3A). To test whether or not the treatment primed plants to rapidly express defense-related genes, we examine mRNA expression of PR1 (At2g14610) and PR4 (At3g04720, Hevein-like protein, a marker for JA-dependent defense response) after subsequent pathogen infection (Fig. 3B) (De Vos et al., 2005; Lorenzo and Solano, 2005). Expression of PR1 in compound A, B, and D-treated plants were comparable with that in mock-treated plants after Pseudomonas infection, whereas PR1 expression level in compound C-treated plants was higher than mock-treated plants 18 hours after infection. This strongly suggests that compound C may have an ability to prime SA-dependent defense responses in Arabidopsis after pathogen infection (Fig. 3B, left panel). Interestingly, exogenous treatment of these CDPs triggered strong mRNA expression of PR4 in Arabidopsis leaves 12 h after subsequent pathogen infection (Fig. 3B, right panel). However the expression did not strongly continue until 18 h after infection in the CDP-treated plants, as compared with that in mock-treated plants. These results strongly suggest that CDPs tested in this study mainly trigger expression of JA-responsive genes during pathogen infection in Arabidopsis.

In conclusion, CDPs from B. vallismortis BS07 strain had different mode of action to induce disease resistance in Arabidopsis. Especially, the facts that most of compounds require JA-responsiveness and enhanced PR4 expression at early infection phase demonstrate that the CDPs tested in this study can prime JA-dependent defense response during pathogen infection. How do plants recognize these CDPs from soil-born bacteria? We are confident that the perception mechanism will be elucidated using genetic analysis in the near future.