Background

Eukaryotes have developed highly effective immune mechanisms for protection against microbial pathogens using pattern-triggered immunity (PTI) as the first line of defense. Pathogen-associated molecular patterns (PAMPs), also referred to as microbe-associated molecular pattern (MAMPs), are highly conserved and essential molecules common to entire classes of microbes but are absent from the host. Host cells utilize pattern recognition receptors (PRRs) to recognize PAMPs as non-self to initiate a large number of signaling responses that contribute to growth restriction of microbial pathogens [1–3]. To evade these host immune responses, pathogenic microbes express and deliver effector molecules into host cells to interfere with PTI [4]. For example, the virulent model bacterium Pseudomonas syringae pathovar tomato (Pto) DC3000 translocates 28 or more effector proteins into plant cells via the type III secretion system (T3SS), some of which are known to suppress PTI [5–7]. Some effector proteins, however, betray the pathogen due to their direct or indirect recognition by cytosolic host resistant proteins resulting in Effector-Triggered Immunity (ETI) [1, 5–8]. Non-pathogenic strains lacking functional T3SS such as Pto DC3000 hrcC- (Pto hrcC-) do not suppress PTI because of their inability to deliver effectors into host cells [6, 8].

In plants, only few PAMP/PRR pairs involved in PTI have been identified and characterized [1, 2, 9, 10]. In the model plant Arabidopsis thaliana, the best studied PTI-system is perception of bacterial flagellin by Flagellin Sensing 2 (FLS2), a plasma membrane localized PRR [2, 11]. Flagellin is the main proteinaceous component of extracellular flagellum filaments essential for the mobility and ability of bacteria such as Pto to infect host plants [12, 13]. Binding of flagellin or flg22, a 22-amino acid peptide derived from the consensus sequence for the most highly conserved region in the N-terminus of eubacterial flagellin [14], to the extracellular domain of FLS2 occurs within the plant apoplast and leads to a variety of early and late signaling responses [11, 15, 16], all of which are dependent on BRI1-Associated Receptor Kinase 1 (BAK1) [17, 18]. One of the best characterized and robust early PAMP-signaling events is the rapid and transient accumulation of apoplastic reactive oxygen species (ROS). Assessing PAMP-elicited ROS has proven to be a valuable tool in identifying and characterizing novel PAMP-signaling components and specific amino acids necessary for their function [18–23]. Production of rapid apoplastic ROS in response to PAMP peptides is solely dependent on the plasma membrane localized NADPH respiratory burst oxidase homolog D (RbohD) [22, 24, 25]. Although the exact role of ROS production in innate immunity remains unclear, ROS initiates a plethora of downstream signaling events, some of which are essential in establishing defense mechanisms to prevent the spread of bacterial pathogens [26, 27].

Over the past decade, studies utilizing commercially synthesized flg22 peptide have greatly aided in increasing our insight into early and late signaling events and in identifying required signaling components and their contribution to PTI [2, 28]. Other efforts exploited boiled bacterial extract to investigate various PAMP-induced responses in leaves or cultured plant cells [14, 29–31]. Boiling bacterial cells results in the release of both extra- and intracellular PAMPs, thus making it difficult to determine the biologically relevant order of PAMP recognition by specific host PRRs. The disadvantage of utilizing cultured cells as opposed to leaf tissue is that because of the unavailability of Arabidopsis mutant cell culture lines, cell culture limits the ability to assess the genetic requirement of the plant host for the response of interest. In contrast, the genetic plant host-pathogen interplay between Arabidopsis plants and Pseudomonas can be interrogated in planta due to the availability of large number of Arabidopsis ecotypes and mutant lines. When investigating responses induced by living bacteria in plant tissue (in planta), most efforts have focused on later responses such as accumulation of the defense hormone salicylic acid (SA) [12-24 hours post infection (hpi)], transcriptional changes of the late gene marker PR1 (24 hpi) or changes in resistance to bacterial infection measured 3 days post-infection (dpi). Only more recently, attention has been given to identifying early signaling events and their genetic requirements induced by living bacterial pathogens on plant host leaves [32–35], the tissue that serves as the primary source for bacterial pathogen infection.

Here, we describe advancement of a fast and convenient in planta bioassay that allows for quantitative assessment of early ROS production in Arabidopsis leaf tissue, the primary site of Pto infection, induced by living Pto strains. Importantly, we provide necessary control experiments showing that in planta, early ROS production was dependent on the presence of Pto cells. By utilizing Arabidopsis mutants previously shown to be affected in flg22-induced ROS production, we demonstrate that early ROS produced in response to live Pto strains was fully dependent on RbohD and FLS2 and partially dependent on BAK1. No statistical differences were observed between ROS induced by Pto DC3000 and hrcC- cells, thus the virulence-promoting T3SS does not appear to have an influence on early ROS production. Because of the ease in setting up ROS assays in a 96-well plate assay, this quantitative analysis is highly suitable to screen within a relatively short period of time large populations of Arabidopsis accession lines in combination with diverse Pto mutant strains to define the genetic requirements of host and bacterial pathogen.

Results and discussion

Our goal was to establish a rapid and robust in planta method to quantitatively measure early ROS production in response to live Pto strains in Arabidopsis leaves, a tissue commonly used for bacterial infection assays. To this end, we adapted the luminol-based ROS assay used for PAMP-elicitation [18, 21, 36, 37] in that we substituted the synthetic PAMP peptide with bacterial cells of Pto strains that were freshly grown for 36-48 hours at 22°C on King’s B media (KBM) plates containing appropriate antibiotics [Figure 1; see Methods for details]. In brief, Arabidopsis leaf discs obtained from 5 week old plants were cut in half with a sharp razor blade to increase surface area, and each half was placed into a distilled water (dH2O)-containing well of a 96-well titer plate for at least 20 hours to reduce wounding effect [Figure 1A]. Immediately prior to elicitation, bacteria were harvested directly from plates [38] and washed twice in sterile dH2O. After the second wash step, bacterial cells were resuspended in sterile dH2O [38], and their cell densities (OD600) were adjusted between OD600 = 0.001 to OD600 = 0.1 [Figure 1B]. The OD600 can be used as a rapid means to provide an estimation of bacterial cell density (as colony-forming units (cfu)/mL) [38, 39]. After Luminol and Horseradish Peroxidase (HRP) was given to the Pto-containing solution, this Elicitation Solution was added to Arabidopsis leaf halves to measure luminol-based ROS production between 0 and 80 minutes using a luminometer [Figure 1C].

Figure 1
figure 1

Outline for plate-setup and preparation of bacterial elicitation solution to measure luminol-based ROS production in leaf disc halves of Arabidopsis wildtype (Col-0) and mutant lines in response to live bacterial Pto cells. A: Preparation of a 96-well assay plate containing leaf disc halves of wildtype (WT, white) and mutant Arabidopsis lines (mut 1, gray and mut 2, black). B: Preparation of Elicitation Solution containing live Pto cells. C: Elicitation of wild-type and mutant tissue with Elicitation Solution to measure ROS response. WT, wild-type; mut, mutant; OD, optical density.

Full size image

First, we established that virulent live Pto DC3000 cells induced ROS-production in wild-type Arabidopsis Col-0 leaves. As shown in Figure 2A, ROS production after elicitation with Pto DC3000 (OD600 = 0.1) was transient, in that increases in ROS was first detected around 20 minutes, peaking around 35-40 minutes and returning to near-basal levels around 70 minutes post-elicitation. Thus, it was delayed compared to that of PAMP peptide flg22-induced ROS which can be detected within 2-4 minutes, peaks at 10-14 minutes and returns to basal levels around 30-35 minutes [Additional file 1]. An explanation for the delayed ROS response to Pto DC3000 cells may be that in contrast to pure and short PAMP peptide(s), the PAMP(s) derived from living bacterial cells may not be readily accessible but may need to be released and/or processed for perception by the plant PRR(s) as suggested for flagellin-FLS2 system [40]. The reduced amplitude in response to Pto DC3000 cells may be an indication that living bacteria possibly release lower concentrations of PAMP(s) into the host apoplast compared to the concentration of applied synthesized flg22 peptide.

Figure 2
figure 2

Transient ROS production in response to live Pto DC3000 and Pto hrcC-in wildtype Arabidopsis Col-0. A: Time-course of ROS production in response to Pto DC3000 (n = 48/treatment). B: Time-course of ROS production in response to Pto hrcC- (n = 48/treatment). C: ROS production induced by Pto DC3000 (black bar) and Pto hrcC—(white bar) shown at peak ROS production (35 minutes post-elicitation) and as total Relative Light Units (RLUs) over 80 minutes following elicitation with indicated bacteria (n = 48/treatment). D: Specificity of ROS production based on presence of Arabidopsis tissue and Pto DC3000 cells (n = 12/treatment). E: Specificity of ROS production based on presence of Arabidopsis tissue and Pto hrcC- cells (n = 12/treatment). To allow direct comparisons, ROS experiments in (A, B and C), in D or in E were performed in the same 96-well plate at the same time. Values are mean ± SE, means with different letters denote a significance difference while similar letters denote no significance (Two tailed student’s t-test, P < 0.0001). For all experiments, bacterial solutions were used at OD600 = 0.1. Experiments were repeated more than 3 times with similar results. RLU, Relative Light Units; ns, no significance; DC, Pto DC3000; hrcC, Pto hrcC-; Lum, Luminol; HRP, Horseradish Peroxidase.

Full size image

It is also possible that the reduced Pto DC3000-induced ROS amplitude may be due to suppression of ROS by bacterial effectors delivered into host cells. To address this hypothesis, we measured ROS production in response to non-pathogenic Pto hrcC- cells (OD600 = 0.1) that lacked functional T3SS and are defective in effector delivery [Figure 2B]. Importantly, ROS induced by Pto hrcC- was similar in the time of initiation, amplitude and duration compared to that by virulent Pto DC3000 [Figure 2B; also see Figure 2C]. ROS production at their peaks (at 35 minutes) and total ROS productions over the 80 minute time-course did not show any statistically significant difference between these two Pto strains [Figure 2C], indicating that this Pto-elicited ROS production is independent of a functional T3SS. These results also suggest that compared to PAMP-induced ROS, the lower level ROS amplitude in response to Pto cells is unlikely due to interference of ROS by bacterial effectors. Similar ROS results were obtained over 75 min time-course after treatment with Pto avrRpm1 and Pto avrRps4, two avirulent bacterial strains known to inject the avirulence proteins AvrRpm1 and AvrRps4, respectively, into host cells resulting in ETI-dependent responses [5–8]. Comparing Pto avrRpm1 and Pto avrRps4 to Pto DC3000, no difference in ROS initiation, amplitude and attenuation as well as total ROS production was observed over the 80 min time-course [Additional file 2]. Taken together, our results using virulent, non-pathogenic and avirulent Pto strains are in agreement that the observed ROS response was due to PTI-dependent events and are consistent with Pto effector delivery into host cells occurring at significantly later times post-infection (> 3 hours) [41] than the ROS response investigated in our study.

In control experiments, ROS production was measured using Elicitation Solutions that contained or lacked Pto strain (Pto), luminol, HRP and Col-0 leaf discs (Col-0) in different combinations. As evident in Figure 2D and E, ROS was produced only when the Pto strain, luminol, HRP and the leaf discs were present. Lack of any of these components did not result in any significant ROS production. No difference in the ROS production was observed in Pto strains resuspended in dH2O or 10 mM MgCl2 [Additional file 3].

Next, we showed that Pto-dependent ROS production was dose-dependent by eliciting Col-0 leaf tissue with different bacterial cell densities between OD600 = 0.001 and 0.1 [Figure 3]. Either strain, Pto DC3000 or Pto hrcC-, elicited a ROS response in a dose-dependent manner [Figure 3A or B, respectively]. Consistent with Figure 2C, we did not observe any statistically significant difference when comparing ROS production elicited by Pto DC3000 and Pto hrcC- at comparable optical densities [Figure 3C]. Furthermore, Pto optical densities of OD = 0.001 or below (data not shown) did not elicit any significant detectable ROS. Pto cells with an OD600 = 0.1 yielded the most consistent and robust ROS response, thus we used this bacterial cell density for all subsequent experiments.

Figure 3
figure 3

Dose-dependent ROS production in response to live Pto DC3000 and Pto hrcC-in wildtype Col-0. A: Time-course of ROS production in response to Pto DC3000 cells with an OD600 of 0.001, 0.01 and 0.1. B: Time-course of ROS production in response to Pto hrcC- cell with an OD600 of 0.001, 0.01 and 0.1. C: Comparison of peak ROS production for each concentration of Pto DC3000 (black bar) or Pto hrcC- (white bar) shown at the point of maximum ROS production 35 minutes post-elicitation). For ROS production, Col-0 leaf disk halves were elicited with DC3000 (filled shape) or hrcC- (open shapes) with the indicated bacterial cell density at 0 min (n = 16/treatment). To allow direct comparisons, all ROS experiments (A-C) were performed in the same 96-well plate at the same time. Values are mean ± SE, means with different letters denote a significance difference while similar letters denote no significance (Two tailed student’s t-test, P < 0.005). Experiments were repeated more than 3 times with similar results. RLU, Relative Light Units; OD, Optical density.

Full size image

To investigate the genetic requirement of the plant host for Pto-induced ROS, we made use of previously characterized Arabidopsis mutants functioning in or upstream of apoplastic ROS production. A rbohD null mutant plant line in Col-0 background [22, 25] was used to investigate dependency on RbohD, the plasma membrane localized protein previously shown to be the sole NADPH oxidase required for rapid induction of apoplastic ROS after elicitation with PAMP peptides [22, 24, 26, 33]. In response to either live Pto DC3000 or hrcC- cells, no ROS production was detected in rbohD compared to wild-type Col-0 for the 80 min time-course [Figure 4A or B, respectively; see also Figure 4C for peak comparison at 35 min post-elicitation]. We conclude that similar to the synthetic flg22 peptide, Pto DC3000 or hrcC- elicited ROS production was fully dependent on RbohD. These result also suggested that the ROS measured in our assays was unlikely due to cell death-associated ROS.

Figure 4
figure 4

RbohD and BAK1 -dependency of ROS production in response to live Pto DC3000 and Pto hrcC-cells. A and B: Time-course of ROS production in response to Pto DC3000 (A; closed symbols) or Pto hrcC- (B; open sympols) in Col-0 (square) and rbohD null mutant (triangle) leaf disc halves (n = 24/treatment). C: Comparison of peak ROS production between Pto DC3000 or Pto hrcC- in Col-0 and rbohD shown at the point of maximum ROS production (35 minutes post-elicitation from A and B). D and E: Time-course of ROS production in response to Pto DC3000 (D; closed symbols) or Pto hrcC- (E; open sympols) in Col-0 (square) and bak1-4 null mutant (triangle) leaf disc halves (n = 24/treatment). F: Comparison of peak ROS production between Pto DC3000 or Pto hrcC- in Col-0 and bak1-4 shown at the point of maximum ROS production (35 minutes post-elicitation from D and E). To allow for direct comparisons, all ROS experiments shown in (A, B and C) and (D, E and F) were performed in the same 96-well plate at the same time. Values are mean ± SE, means with different letters denote a significance difference while similar letters denote no significance (Two tailed student’s t-test, P < 0.0001). For all experiments, bacterial solutions were used at OD600 = 0.1. Experiments were repeated more than 3 times with similar results. RLU, Relative Light Units.

Full size image

Next, we determined whether Pto-elicited ROS was dependent on BAK1, the receptor-like kinase known to be required very early in initiating signaling after flg22-elicitation [18, 20, 42]. Underlining its crucial role in PAMP-signaling and PTI, BAK1 forms a ligand-induced receptor complex with FLS2 within seconds [18, 20, 42, 43]. In response to Pto DC3000 or hrcC-, we observed an increase in ROS production over time in bak1-4 null mutant leaf discs; but importantly, the ROS amplitude was significantly reduced compared to Col-0 [Figure 4D or E, respectively]. No statistical significant difference in ROS production was observed in bak1-4 mutant tissue in response to Pto DC3000 or Pto hrcC- [Figure 4F]. Taken together, these results suggest that Pto-elicited ROS was only in part dependent on BAK1. Our studies are consistent with previous reports showing that full signaling responses to the bacterial PAMP peptides flg22 require other proteins in addition to BAK1 [18, 20, 42].

Previous studies have shown that Arabidopsis mutants lacking functional FLS2 receptor are unable to perceive flg22 which results in a complete loss of any flg22/FLS2-dependent signaling event including lack of apoplastic ROS production [16, 31, 36]. In addition, FLS2-dependent responses contribute to restriction of Pto DC3000 growth [31]. However when treated with crude boiled Pto DC3000 extracts, fls2 null mutants are still able to elicit signaling responses including ROS production [29, 31]. These results are consistent with the idea that in addition to flagellin (flg22), Pto DC3000 contains multiple extra- and intracellular PAMPs that are exposed upon boiling of bacterial strains. To determine whether FLS2 contributed to ROS production induced by intact living Pto DC3000 or hrcC-, we utilized previously characterized fls2 null mutants [18, 31, 36]. Interestingly, fls2 null mutants in a Col-0 background did not elicit any ROS in response to either Pto strain [Figure 5A-C]. To confirm this observation, we challenged Wassilewskija (Ws-0), an Arabidopsis ecotype that is flagellin-insensitive and considered a natural fls2 mutant due to an early stop mutation in the FLS2 gene [31], with Pto DC3000 or Pto hrcC-. Similar to the fls2 null mutant in Col-0, Ws-0 displayed a complete lack of ROS production in response to either Pto strains [Additional file 4]. Based on our studies utilizing two different Arabidopsis ecotypes lacking functional FLS2, the flagellin/flg22 receptor FLS2 appeared to be the predominant host receptor responsible for initiating early ROS production in response to intact, living Pto DC3000 and hrcC-. Our results are consistent with observations that in response to intact Pto DC3000 and DC3118 at 1 to 2 hours post-treatment, stomatal closure of epidermal leaf peels is entirely dependent on the presence of functional FLS2 [33, 34]. Thus, these and our present studies suggest that not all potential Pto PAMPs appear to be exposed simultaneously during the initial bacterial infection. Our results further underline the advantages of utilizing living intact Pto cells in contrast to crude boiled bacterial extract [14, 29, 31] when studying early signaling responses induced by bacteria in planta. Boiling of bacteria likely releases both extra- and intracellular localized PAMPs simultaneously. In contrast, utilizing intact live bacterial cells enables researchers to determine the biologically relevant contribution and/or order of PAMP recognition by specific host PRRs.

Figure 5
figure 5

FLS2 -dependency of ROS production in response to live Pto DC3000 and Pto hrcC-cells. Time-course of ROS production in response to Pto DC3000 (A; closed symbols) or Pto hrcC- (B; open sympols) in Col-0 (square) and fls2 null mutant (triangle) leaf disc halves (n = 24/treatment). C: Comparison of peak ROS production between Pto DC3000 or Pto hrcC- in Col-0 and fls2 shown at the point of maximum ROS production (35 minutes post-elicitation from A and B). To allow for direct comparisons, all ROS experiments shown in (A, B and C) were performed in the same 96-well plate at the same time. Values are mean ± SE, means with different letters denote a significance difference while similar letters denote no significance (Two tailed student’s t-test, P < 0.0001). For all experiments, bacterial solutions were used at OD600 = 0.1. Experiments were repeated more than 3 times with similar results. RLU, Relative Light Units.

Full size image

Conclusions

In this study, we report the advancement of a rapid and convenient bioassay allowing the quantitative assessment of ROS production between Arabidopsis leaf tissue (the primary site of Pto infection) and living Pto bacterial strains. Because of the ease in setting up ROS assays in a 96-well plate assay, this quantitative analysis is highly suitable to screen large populations of Arabidopsis accession lines in combination with diverse Pto mutant strains to define the genetic requirements of host and bacterial pathogen. In future experiments, this ROS bioassay may also allow addressing which PAMP/PRR pair quantitatively contributes to early signaling in response to bacterial pathogens in leave tissue. Furthermore, the utility of this bioassay can be easily extended beyond Arabidopsis and Pto to diverse model or crop species and their cognate microbial pathogen to define components required for early ROS responses.

Methods

Plant material and growth conditions

Arabidopsis seeds were sterilized with 10% bleach + 1% Triton X-100 for 1 hour, rinsed with water and plated aseptically on 0.5% agar containing 2.14 g L-1 Murashige and Skoog (MS) salts (Sigma Chemical Company, St. Louis, MO, USA, http://www.sigmaaldrich.com/) + 1% sucrose, pH 5.7. Following stratification for 2 days at 4°C, seedlings were germinated in Percival CU-36 L4 growth chambers (Percival, Perry, IA) under continuous light at 22°C [18]. Seven day-old seedlings were transplanted in soil and grown in an 8-h light/16-h dark photoperiod at 82 μmol m-2 s-1. Fully expanded rosette leaves were used from 4-5 week old plants for all ROS experiments. The Ws-0, fls2Δ (Col-0), bak1-4 (Col-0), and rbohD (Col-0) mutants have been previously described [18, 22, 25].

Chemicals

Synthetic flg22 peptide [21] was made by GenScript (Scotch Plains, NJ) and used at indicated concentrations. Horseradish Peroxidase (HRP; Sigma, catalog # P6782) was prepared as a 500x HRP stock solution by dissolving 10 mg/mL in sterile H2O. Aliquots (10-30 μL) of the 500x HRP stock solution were stored at -20˚C and used at a final concentration of 20 μg/mL. For the 500x Luminol stock solution, 17 mg Luminol (≥ 97% purity-HPLC; Sigma; catalog # A8511) was completely dissolved in 1 ml of 200 mM KOH and used at a final concentration of 0.2 μM. For example, 10 μl of each of the 500x stocks of luminol and HRP were added to 5 ml of resuspended bacterial solution. Because Luminol is light-sensitive, all solutions containing Luminol must be protected from light by wrapping tubes in aluminum foil. The 500x Luminol stock solution is made fresh prior to use and discarded daily.

Bacterial preparation

Two days prior to ROS experiments, Pseudomonas syringae pv. tomato strain DC3000 (Pto DC3000) or Pto hrcC- was streaked from glycerol stocks (stored at -80°C) onto King’s B medium (KBM) agar plates containing 50 μg ml-1 kanamycin and 30 μg ml-1rifampicin (Pto DC3000) or 30 μg ml-1rifampicin (Pto hrcC-) [38]. KMB plates containing bacteria were incubated for 36 to 48 hours at room temperature. Prior to elicitation, bacteria were scraped from plates and washed twice in sterile dH2O by repeated centrifugation at 10,000×g for 5 minutes. After the second wash step, a 1:10 dilution of the bacterial solution was made and its optical density (OD600) was measured using a spectrophotometer as a means to provide an approximate quantification of bacterial cell density [38, 39]. The final bacterial elicitation solution was adjusted to an OD600 between 0.001 and 0.1, which under our conditions equated to 1 × 106 to 108 colony forming units (cfu)/mL based on serial dilution plating. The OD600 of the final bacterial elicitation solution was measured again to ensure accuracy of the dilution. To accurately compare the dose-dependency of the ROS response [Figure 3], bacteria were serially diluted from a stock solution.

Measurement of apoplastic ROS production

One day before the ROS assay, leaf disks (1.1 cm2) from 4-5 week old plants were cut into two equal halves with a sharp razor blade to increase the cellular surface area exposed to elicitation solution, an important step for obtaining reproducible responses with less variability within and between experiment. Each leaf disc half was floated adaxial side up in an individual well of a 96-well microtiter plate (Costar; Fisher Scientific, catalog # 3912) containing 150 μl dH2O and then incubated overnight at 22°C in continuous light for 20 to 24 hours to reduce the wounding response. Prior to elicitation, the Elicitation Solution was prepared containing bacteria, Luminol and HRP. For a 10 ml Elicitation Solution, 20 μl of 500x HRP stock solution and 20 μl of the 500x Luminol stock solution is added to 10 ml of bacterial cells that have been already diluted to the desired concentration. For flg22-induced ROS production, flg22 peptide was used instead of bacteria in the Elicitation Solution. All Elicitation Solutions were kept at room temperature. Immediately prior to the elicitation, the incubating dH2O solution was carefully removed from each well avoiding any tissue damage or desiccation. Then using a multichannel pipetman, 100 μl of the Elicitation Solution was quickly added to each well containing leaf disc half. For Luminol-based ROS production, the plate was placed without delay into a GloMax® 96-well microplate luminometer (Promega, Madison, USA) to measure Pto-induced ROS production between 0 and 80 minutes.

Statistical analysis

Each experiment was done at least 3 independent times with similar results. Statistical significances based on unpaired two sample t-test were determined with Graph Pad Prism4 software (La Jolla, CA).