Introduction

In response to pathogen challenge, plant cells undergo dramatic transcription reprogramming to favor immune responses over normal cellular functions (Spoel et al. 2009). Multiple methods have been developed to enhance the plant resistance to viruses. However, most of the approaches including transgenic processes (not broadly acceptable to consumers; Slade et al. 2004) or some drugs (toxic for human beings and leaving residuals to environment) are plant-species-limited or virus-species-limited, and the inhibitory efficiency to virus replication are usually <50% (Vlot et al. 2009). Therefore, most methods available so far cannot be used in agricultural industry directly (Vlot et al. 2009).

Here we are trying to look for some other ways, and to obtain the best innate immune response of plant cells. Salicylic acid (SA) and jasmonic acid (JA) are important natural phytohormones, working at systemic immunity in plant defense responses, although the role of JA is not well-established (Cui et al. 2005; Attaran et al. 2009). The contribution of these signaling molecules in plant defense differs depending on the invading pathogens. SA-dependent defenses are in general effective against biotrophic pathogens (which includes biotrophic fungi and oomycetes, bacteria and viruses), while JA-dependent defenses are generally effective against necrotrophic pathogens (which also includes bacteria and fungi). SA- and JA-dependent defense pathways have been shown to cross-communicate, providing the plant with a regulatory potential to fine-tune the defense reaction depending on the type of attacker encountered (Takahashi et al. 2004). Park et al. (2007) found that SA could be converted to methyl salicylate, and could act as a critical mobile signal for plant resistance to Tobacco mosaic virus (TMV). However, Truman et al. (2007) found that JA but not SA rapidly accumulates in phloem exudates of leaves challenged with an avirulent bacterial strain of Pseudomonas syringae and suggested that JA may be the key signaling molecule to systemic immunity. Attaran et al. (2009) indicated that neither methyl salicylate nor JA is essential for systemic immunity but emphasize the crucial role of SA in Arabidopsis infected with P. syringae. On the other hand, a large number of reports suggested that SA and JA defense pathways are mutually antagonistic. However, evidences of synergistic interactions also have been reported (Van Wees et al. 2000; Cui et al. 2005; Mur et al. 2006; Mishina and Zeier 2007; Clarke et al. 2009). Despite all these controversies, we noticed that SA always accumulates 1 day after JA accumulation, in natural plant’s innate-immune-response to viruses, at least for CMV- and TMV-infected tobaccos (See Supplementary Fig. S1; Park et al. 2007). Therefore, we simulated the changes of JA followed by SA in natural innate-immune-response of plants by exogenous application of the phytohormone. By this method, the inhibitory efficiency to virus replication could achieve 80–90%. The application has a broad-spectrum and is effective for Arabidopsis, Nicotiana benthamiana, Nicotiana glutinosa, Nicotiana tabacum, hot pepper and tomato, and for CMV, TCV, TMV and Tobacco necrosis virus (TNV) as well.

Materials and methods

Plant culture and pathogen inoculation

All plants were grown in a temperature-controlled growth chamber with a 12 h-light (100 μmol m−2 s−1) 12 h-dark cycles at 21°C for 2 weeks. All Arabidopsis mutants were in the Col-0 background. The seed stock numbers (Ohio State University, Columbus, OH, USA) are SALK_034157C, CS876251, CS833488, SALK_096040C, CS3726, CS6358, CS803227, CS871051, SALK_063847C, CS875096, CS877584, SALK_129907, CS31099, CS879510, CS800040, SALK_140054C, SALK_127284, CS331868, SALK_084332C and SALK_053805 for wrky53, wrky70, mpk4, pdf1.2, npr1, pr1, pr2, gst, mpk1, mpk2, mpk3, mpk5, mpk6, mpk12, mpk14, mkk1, mkk2, mkk4, mkk6 and opr3 mutants, respectively. Plants homozygous for the T-DNA insertion were identified based on the PCR analysis. opr3 × NahG transgenic plants were acquired by performing a genetic crossing, and the progeny were analyzed by PCR. Only homozygous plants were used in our experiments.

Virus isolates of Cucumber mosaic virus (CMV), Turnip crinkle virus (TCV), and Tobacco necrosis virus (TNV) were acquired form Horticulture Institute, Sichuan Academy of Agricultural Sciences, China. The inoculation with a virus was carried out as described previously (Shang et al. 2009). In brief, carborundum was evenly applied to the surface of leaves, and the virus RNA at a concentration of 1 μg/ml in 10 mM Hepes buffer, pH 7.0, was rubbed onto the leaves with cheese cloth.

Before 24 h of virus inoculation, the seedlings were sprayed with JA (Sigma, St. Louis, MO, USA; methyl jasmonate could also be used and with the same concentration) followed by SA (Sigma) at different time-intervals and different concentrations. Then the virus replication levels were detected by quantitative real-time PCR analysis at 10 days post inoculation (dpi).

RNA analysis

Total RNAs were isolated from plant tissues by a Trizol reagent (Invitrogen Corp., Carlsbad, CA, USA). The RNAs were separated on 1.0% denaturing agarose gel. For RT-PCR, the first-strand cDNA was prepared using the ReverTra Ace kit (Toyobo Co., Ltd., Osaka, Japan). To further assay the expression levels of genes, quantitative real-time PCR analysis was performed on a Bio-Rad iCycler (Bio-Rad, Beijing, China). The primer sets corresponding to Supplementary Table S1. The cDNA was amplified by using SYBR Premix Ex Taq (TaKaRa, Otsu, Shiga, Japan). The amplification of the target genes was monitored every cycle by SYBR-green I fluorescence. The C t (threshold cycle), defined as the PCR cycle at which a statistically significant increase of reporter fluorescence was first detected, was used as a measure for the starting copy numbers of the target transcripts. Relative quantitation of the target gene expression level was performed using the comparative C t method (Shang et al. 2009). Three technical replicates were performed for each experiment. ACTIN1 gene was used as an internal control (Shang et al. 2009).

Superoxide and H2O2 staining and cell death determination

In situ superoxide and H2O2 were detected with nitroblue tetrazolium (NBT) and 3,3-diaminobenzidine (DAB), respectively, as described previously (Yang et al. 2004; Shang et al. 2010). Tobacco leaves were excised at the base with a razor blade and supplied through the cut ends with NBT (0.5 mg ml−1) or DAB (2 mg ml−1) solutions for 8 h. Leaves were then decolourized in boiling ethanol (95%) for 15 min. Staining of dead cells with trypan blue (1.25 mg ml−1, Sigma) was the same as previously reported (He et al. 2007). Wilt of the seedling (cell death) was characterized by the relative water content (RWC) (Shang et al. 2010).

TCV–GFP vector construction and GFP detection

Turnip crinkle virus (TCV) vector expressing the green fluorescent protein (GFP) was constructed. GFP imaging was performed using UV illumination by a long wavelength UV lamp (380–420 nm), and photographs were taken using an Olympus Camedia E10 digital camera (Liu et al. 2005).

JA and SA hormone measurements

Liquid nitrogen-frozen material was ground to a fine powder and extracted with ice-cold solvent (60% methanol in 0.1% acetic acid) by using about 10 ml/g FW tissue. Internal standards (10 ng each per gram FW; 13C2 JA and 2H4 SA) were added and after 10 min on ice the slurry was centrifuged. The pellet was re-extracted with 5 ml of solvent and both supernatants pooled and evaporated to near dryness. The solution was made to 3 ml in 0.1% acetic acid and loaded onto a 500-mg C18 Bond-Elut cartridge (Varian, Beijing, China). The column was washed with 6 ml of 0.1% acetic acid and the hormones eluted in 80% methanol/0.1% acetic acid (v/v). Samples were dried under vacuum and analyzed by liquid chromatography–mass spectrometry (LC/MS) using the Multiple Reaction Monitoring (MRM) mode and negative ion Turboion Spray source on an Applied Biosystems Q-TRAP 2000 (Applied Biosystems, Beijing, China). Quantitation was based on appropriate MRM ion pairs for labeled and endogenous SA and JA: 13C2-JA 211/61, JA 209/59, 2H4 SA 141/97, and SA 137/93 by using Analyst 1.4.1 (Applied Biosystems). For phloem hormone determination, exudates were made to 3 ml in 0.1% acetic and loaded onto Bond-Elut cartridges as above (Truman et al. 2007).

Statistical analysis

An independent (unpaired) Student’s t-test (two-tailed) was chosen to test the significance of differences among means of small ‘n’ sample sets. A difference was considered to be statistically significant when P < 0.05.

Results

Symptom and virus replication of Nicotiana glutinosa leaves infected with CMV

Nicotiana glutinosa and CMV compose a model system for plant–virus interactions. N.glutinosa leaves were pretreated with JA followed by SA at different time-intervals with different concentrations before CMV infection (Fig. 1). Then the CMV replication levels were detected by quantitative real-time PCR analysis at 10 days post inoculation (dpi). JA followed by SA (JA → SA) had the highest inhibitory efficiency to CMV replication (over 80%), higher than JA and SA simultaneous co-pretreatment (JA + SA), and higher than JA or SA single pretreatment. The optimal concentrations of JA and SA were 0.06 and 0.1 mM, respectively, and the optimal time for SA was 24 h after JA. Extraordinary high level of JA or SA does great harms to plant seedlings (chlorosis or wilting but not affecting agronomical traits) (Rakwal and Komatsu 2001; Yuan and Lin 2008; Luo et al.2009). The low levels (below 0.06 mM) would not work either. The length of delay between JA and SA is very important. Too short or too long delay may change the relationship between JA and SA from synergism to antagonism.

Fig. 1
figure 1

Symptoms of Nicotiana glutinosa systemic leaves infected with CMV and CMV-CP gene expression levels. N. glutinosa leaves were pretreated with JA followed by SA at different time-intervals with different concentrations before CMV infection. Then the CMV replication levels were detected by quantitative real-time PCR analysis at 10 days post inoculation (dpi). There were 14 treatment groups as shown in the figure. Below each treatment group, the symptom is shown. At the bottom of the figure shows CMV-CP (coat protein) gene expression levels. CK+ means seedlings infected with CMV, but without JA or SA pretreatment. The expression level of CMV-CP in CK+ is normalized to 100 units. Error bars show standard deviations (n = 3)

Full size image

The applicability of JA → SA pretreatment in different plants infected by different viruses

The applicability of JA → SA pretreatment was further defined. For interactions of tomato (Lycopersicon esculentum)–CMV, tomato–TCV, hot pepper (Capsicum frutescens)–TCV, N. benthamiana–CMV, N. tabacum–CMV, N. tabacum–TMV and N. tabacum–TCV, the virus replication could be inhibited by 80–95% with JA → SA pretreatment (Figs. 2, 3). However, JA → SA pretreatment was less effective for the severe interaction (N. benthamiana–TCV, Fig. 4a) and not effective for the lethal interaction (N. benthamiana–TNV, Fig. 2 and Fig. 4b). In the N. benthamiana–TCV interaction, virus-induced H2O2 and superoxide accumulation, cell death and virus replication could be half inhibited by the JA → SA pretreatment (Fig. 4a). However, only a slight alleviation to reactive oxygen species (ROS) accumulation (data not shown), cell death, virus replication, and water content declination could be observed for the lethal N. benthamiana–TNV interaction (Fig. 4b). These were not due to inappropriate JA and SA concentrations or inappropriate time delay, as changes to JA and SA levels or JA → SA time delay resulted in worse resistance to TNV (data not shown).

Fig. 2
figure 2

Symptom and virus accumulation of virus-infected plants pre-treated with JA and SA. a Symptoms of tomato (Lycopersicon esculentum) leaves infected with CMV and CMV-CP gene expression levels. b The heights of tomato seedlings infected with TCV and TCV-CP gene expression levels. TCV inhibited tomato seedling growth, which could be alleviated by JA and SA pretreatments. c Symptom of hot pepper (Capsicum frutescens) leaves infected with TCV and TCV-CP gene expression levels. d Symptom of Nicotiana benthamiana leaves infected with TNV and TNV-CP gene expression levels. “J → S” indicates 0.06 mM JA followed by 0.1 mM SA pretreatment at the 24-h interval and before the virus inoculation at 24 h; “J + S” means JA and SA simultaneous co-pretreatment; “SA” indicates solo SA pretreatment; “JA” indicates solo JA pretreatment. CK+ means seedlings infected with CMV, but without JA or SA pretreatment. Virus accumulation levels were detected by quantitative real-time PCR analysis at 10 dpi. Error bars show standard deviations (n = 3)

Full size image
Fig. 3
figure 3

Symptom and virus accumulation of virus-infected plants pre-treated with JA and SA. a Symptoms of Nicotiana benthamiana infected with CMV and CMV-CP gene expression levels. b Symptoms of Nicotiana tabacum infected with CMV and CMV-CP gene expression levels. c Symptom of Nicotiana tabacum leaves infected with TMV and TMV-CP gene expression levels. d Symptom of Nicotiana tabacum leaves infected with TCV and TCV-CP gene expression levels. “J → S” indicates 0.06 mM JA followed by 0.1 mM SA pretreatment at the 24-h interval and before the virus inoculation at 24 h; “J + S” means JA and SA simultaneous co-pretreatment; “SA” indicates solo SA pretreatment; “JA” indicates solo JA pretreatment. CK+ means seedlings infected with CMV, but without JA or SA pretreatment. Virus accumulation levels were detected by quantitative real-time PCR analysis at 10 dpi. Error bars show standard deviations (n = 3)

Full size image
Fig. 4
figure 4

Symptoms, ROS levels and virus replication levels of Nicotiana benthamiana leaves infected with TNV or TCV. a N. benthamiana leaves infected with TCV. b N. benthamiana leaves infected with TNV. Cell death, relative water content or superoxide and H2O2 staining are shown. The virus replication levels were detected by quantitative real-time PCR analysis at 10 dpi. “JA → SA” indicates 0.06 mM JA followed by 0.1 mM SA pretreatment at the 24-h interval and before the virus inoculation at 24 h. Error bars show standard deviations (n = 3)

Full size image

The applicability of JA → SA post-treatment

What would happen if the JA → SA treatment was applied several days after the virus inoculation? JA → SA treatment to Arabidopsis at 3 dpi resulted in 45% inhibition of CMV replication. At 7 dpi, it resulted in no apparent effect (See Supplementary Fig. S2). Our data suggest that JA → SA treatment is a valuable precaution method for plant viruses, rather than a therapeutic approach. Induced immunity was activated in JA → SA pretreated plants, and therefore they were resistant to the subsequent virus-infection.

Simulating field experiment

JA and SA are natural plant hormones, and are safe to human beings in low levels (Chan et al. 1995). Therefore, the JA → SA virus-control method should be practiced directly in agricultural industry. Corresponding simulating experiments were performed. 5 pot × 5 pot N. benthamiana seedlings were placed side by side, touching each other with leaves. All the seedlings were sprayed with or without 0.06 mM JA and sprayed with 0.1 mM SA after 1 day later. On the third day, the centric seedling was removed and replaced by one that was infected with TCV–GFP and 50 aphids (Myzus persicae, vector for the viruses) infested N. benthamiana seedling. The TCV–GFP RNA level of each seedling was detected 10 days later. Figure 5 is a schematic diagram showing the spread of TCV–GFP. No GFP fluorescence could be observed for JA → SA pretreated seedlings (except the originally centric one). In contrast, the virus had spread all around to seedlings without JA → SA pretreatment. The maximum virus accumulation level in the outer layer of JA → SA pretreated seedlings was about 10% of the centric one. For the controlling seedlings without any treatment, the level was about 60%. As mentioned earlier, JA → SA pretreatment is not very effective for N. benthamiana–TCV interaction. Therefore, we repeated the simulating field experiments with CMV, TMV, and TNV. Interestingly, almost no CMV or TMV RNA could be detected in the outer layer of JA → SA pretreated seedlings (data not shown). However, for TNV-infected N. benthamiana, the JA → SA virus-control method does not work as well as found for other viruses (data not shown). Nevertheless, we like to conclude that JA → SA pretreatment has a great potential to be used as crop protection against most viruses in the field.

Fig. 5
figure 5

Simulating field experiment of TCV–GFP spread in Nicotiana benthamiana seedlings. 5 pot × 5 pot N. benthamiana seedlings were placed together, touching each other with leaves. All the seedlings were sprayed with or without 0.06 mM JA, followed by 0.1 mM SA after 1 day later (JA → SA, marked by pink color). Then at the third day, the centric seedling was removed and replaced by a TCV–GFP infected and 50 aphids (Myzus persicae) infested N. benthamiana seedling. The TCV–GFP RNA level of each seedling was detected 10 days later. The left part of the figure is a schematic diagram showing the spread of TCV–GFP. The depth of the green color in each pot indicates the approximate TCV–GFP RNA level. The virus did not spread around uniformly. Therefore, the green color of the each layer only shows the maximum virus accumulation level in the layer. The GFP fluorescence in the outer layers of seedlings is marked with white cycles

Full size image

Putatively molecular mechanism

The fact that both JA and SA are related to plant immunity has been further proven in JA-deficient mutant (opr3) (Truman et al. 2007), SA-deficient transgenic plants (NahG) (Delaney et al. 1994), and JA and SA double-deficient transgenic plants (opr3 × NahG, See Supplementary Fig. S3). Either JA deficiency or SA deficiency could result in decreased CMV resistance. opr3 × NahG plants had the weakest CMV resistance. In general, the wild-type plants (Col-0) had the highest levels of resistance gene transcripts, especially for MPK4 (Petersen et al. 2000), PDF1.2 (Li et al. 2004) and NPR1 (Spoel et al. 2003, 2009; Durrant and Dong 2004) genes (Fig. 6). JA or SA deficiency down-regulates some resistance proteins and therefore hampers plant immunity to RNA viruses.

Fig. 6
figure 6

CMV-CP accumulation and resistance gene expression in wild-type Arabidopsis plants, opr3 mutant, NahG transgenic plants, and opr3 × NahG hybrid plants. Gene expression levels were detected by quantitative real-time PCR analysis at 10 dpi. The maximum expression level of each gene is normalized to 100 units. Error bars show standard deviations (n = 3)

Full size image

Possible molecular mechanisms were investigated and some putative signaling factors have been defined. Transcripts of two transcription factors WRKY53 (Miao and Zentgraf 2007) and WRKY70 (Li et al. 2004), MAP Kinase 4 (MPK4), JA-responsive resistance protein PDF1.2 (Manners et al. 1998), SA-responsive resistance protein NPR1, PR1 and PR2 (Spoel et al. 2003, 2009; Durrant and Dong 2004), and ROS-related resistance protein glutathione S-transferase (GST) (Uquillas et al. 2004) were quantified in Arabidopsis seedlings inoculated with CMV (Fig. 7a). Among them, WRKY53, WRKY70, and PDF1.2 showed the maximum expression in the JA → SA pretreated seedlings at 10 dpi, compared with the other treatments, therefore corresponding to the optimal effect of JA → SA. MPK4 transcript increased in all the seedlings inoculated with CMV. However, the increasing amplitude was minimal in the JA → SA pretreated seedlings, compared with the other treatments.

Fig. 7
figure 7

Resistance gene expression levels in wild-type Arabidopsis seedlings and CMV-CP replication levels in 19 Arabidopsis mutants. a Resistance gene expression levels of wild-type Arabidopsis seedlings infected with CMV. “J → S” indicates 0.06 mM JA followed by 0.1 mM SA pretreatment at the 24-h interval and before the virus inoculation at 24 h; “J + S” means JA and SA simultaneous co-pretreatment; “SA” indicates solo SA pretreatment; “JA” indicates solo JA pretreatment. CK+ means seedlings infected with CMV, but without JA or SA pretreatment. b CMV-CP gene expression levels of 19 Arabidopsis mutants. Gene expression levels were detected by quantitative real-time PCR analysis at 10 dpi. The expression level of CMV-CP in CK+ is normalized to 100 units. Error bars show standard deviations (n = 3)

Full size image

The roles of WRKY53, WRKY70, PDF1.2, and MPK4 in JA → SA pathways have been confirmed further with the corresponding mutants (Fig. 7b). Except mpk4, all other mutants have declined or unchanged resistance to CMV. The virus replications in some of the mutants could not be inhibited by JA → SA pretreatment, and thus some putative signaling factors mediating the JA → SA pathways have been identified. Besides WRKY53, WRKY70, PDF1.2, and MPK4, they are MPK2, MPK3, MPK5, MPK12, MPK14, MKK1, MKK2, and MKK6. WRKY 53 and WRKY70 are the nodes of convergence for JA-mediated and salicylate-mediated signals, which are induced by SA, but repressed by JA (Li et al. 2004; Miao and Zentgraf 2007). mpk4 mutants exhibit constitutive systemic acquired resistance (SAR) with elevated SA levels (Petersen et al. 2000). MKK2 may work upstream of MPK4 (Teige et al. 2004). It seemed that both JA-related genes and SA-related genes were both important to the synergism of JA and SA.

Discussion

In this paper, we show that JA acts with SA and both confer optimal virus resistance when applied in appropriate concentrations and time delay. The method is effective, has a broad-spectrum, is green and safe, and easy to be performed. Single JA followed by SA treatment does not affect agronomical traits of the crops or plant yield. However, all the plants we tested in these studies are herbaceous plants. For woody or vine plants (such as grape), the concentrations and time delay should be adjusted. Even for different herbaceous plants, the JA and SA concentrations and the time delay should also be adjusted finely and differently. Especially for rice, the adjustment must be done since rice has a very high endogenous SA level (Yang et al. 2004; Yuan and Lin 2008).

H1N1 and H5N1 are known forms of the common influenza viruses and cause strong and early inflammatory responses, which contribute to the severe lung pathology and lethality (White et al. 2009). Enhancement of immunity may be helpful for controlling the common influenza viruses. But it is unlikely that it can alleviate symptoms or virus replications of the people who are infected with mortal influenza viruses (Kohlmeier and Woodland 2009). JA → SA pretreatment utter-mostly prompts the plant immunity. However, it cannot stop the lethal symptom development or lethal virus replication in TNV-infected N. benthamian leaves. In other words, JA → SA pretreatment is a quite effective method for non-lethal plant virus curation, but not an appropriate approach for controlling of rapid lethal viruses. For these mortal viruses, animals or plants may choose sacrifice of individuals to prevent virus spreading in the community (Kohlmeier and Woodland 2009). Symptom alleviation is more important than plant lethal virus control. Both of them may be considered for plant lethal virus cures (data not shown).

Previous studies showed that mpk4 mutants exhibit temperature-dependent cell death and constitutive activation of defense responses (Petersen et al. 2000; Teige et al. 2004; Qiu et al. 2008; Gao et al. 2009). MPK4 itself is a negative regulator of plant resistance (R) protein signaling (Petersen et al. 2000; Teige et al. 2004; Qiu et al. 2008; Gao et al. 2009). Thereby JA → SA pretreatment may adjust MPK4 transcript to a moderate level to facilitate the strongest systemic immunity. NPR1, PR1, PR2, and GST seemed not to be related with the optimal effect of JA → SA. JA-marker gene PDF1.2 is known as a SA-repressible gene; however, its expression level was higher in the SA-pretreated CMV-infected seedlings than in the CMV-infected seedlings without any treatment. This may be due to the complex cross-talk between JA and SA. For the long-term interaction (10 dpi), PDF1.2 might become a SA-inducible gene (or JA levels were prompted later after SA treatment), although it should be repressed by SA instantaneously (Spoel et al. 2003), which requires further investigations. The protection was still intact in the SA-insensitive npr1, pr1 or pr2 mutant, which does not suggest that the protective mechanism is unlikely systemic immunity, but that NPR1, PR1, and PR2 are involved in neither the cross-talk between JA and SA nor the optimal effect of JA → SA. The cross-talk between JA and SA pathways is very complex. More factors and detailed mechanism need more studies in the future.

At present, both JA and SA can be synthesized artificially. However, JA is relatively expensive (500$ per gram). Benzothiadiazole (BTH) was used commercially instead of SA because of its reduced phytotoxicity and higher effectiveness (Katz et al. 1998). BTH could be used as well as SA with the same concentration (data not shown). JA’s price is about one thousand times of SA’s (0.05$/g). Methyl jasmonate (MeJA) could be used as well with the same concentration (data not shown). However, MeJA is still relatively expensive (6$/g). Cheaper synthetic JA or its bioactive derivants should be exploited, before the JA → SA approach could be used in agricultural industry widely.