Bacteria communicate with each other in a population density-dependent process known as quorum sensing. N-acyl-homoserine lactones (HSLs) are the autoinducers of Gram-negative bacteria and the best-studied quorum sensing signals so far. HSLs induce various responses in plants, including systemic resistance and root development.
Here, we used different methods, including tritium labelling, sensor strain assays and monoclonal antibodies (mAbs), to analyse the uptake and translocation of C8- and C10- homoserine lactones into barley (Hordeum vulgare cv Barke).
Both HSLs were already systemically transported into the shoot at 2 h after application. HSL uptake could be inhibited by orthovanadate, demonstrating that ABC transporters are involved in the uptake. Root transport occurs predominantly via the central cylinder, which was shown by transport inhibition via KCl application and autoradiography of root cross-sections. Furthermore, a newly established detection method with mAbs allowed the first detection of a systemic transport of long-chain HSLs in plants.
The coupled use of different HSL detection methods demonstrated that the uptake and transport of HSLs into barley does not occur passively, but relies, at least partially, on active processes in the plant.
For more than four decades, it has been known that bacteria are able to communicate with each other via so-called autoinducers (Nealson et al., 1970) in a population density-dependent process known as quorum sensing (QS; Fuqua et al., 1994). The first autoinducer that was extracted and analysed was N-(3-oxohexanoyl)-l-homoserinelactone from Vibrio fischeri (Eberhard et al., 1981) and to date N-acyl-homoserine lactones (HSLs) produced by Gram-negative bacteria are the best-studied QS molecules so far. The basic structure of HSLs is a lactone ring coupled to an acyl-side chain. HSLs are ranked as intraspecies signals, because the interaction between the bacterial HSL-receptors to their corresponding HSL is very specific (Taga & Bassler, 2003; Huse & Whiteley, 2011). The specificity of the HSL molecule is conferred by the length of the acyl-side chain (between four and 18 C-atoms) and the substitution at the C-3 position of the side chain, which can be hydroxylated or oxidized to a keto-group (Stevens et al., 2011). QS-controlled bacterial processes include bioluminescence, motility, conjugation, generation of biofilms, production of antibiotics, and the expression of virulence factors (reviewed in Taga & Bassler, 2003; Chhabra et al., 2005). Today it is known that QS functions beyond an exclusive measurement of population density. In accordance with the efficiency sensing model, bacteria measure their own number and that of other bacterial species, as well as their spatial distribution and the efficiency of excreted molecules (Hense et al., 2007).
Since the first plants settled on land c. 450 million yr ago, they were confronted with bacteria and their QS molecules, especially in the rhizosphere. From an evolutionary point of view, it is advantageous for plants to eavesdrop and eventually interfere with the bacterial communication system. This phenomenon, known as quorum quenching, is observed in bacteria and plants. To date, two different enzymes are known to degrade HSLs. HSL-lactonases open the lactone ring and hydrolyse the molecule, while HSL-acylases cleave the lactone-ring from the acyl-side chain (Dong et al., 2001; Xu et al., 2003; Dong & Zhang, 2005). HSL-lactonases were found in Delisea pulchra, where the enzyme prevents the formation of biofilms on the surface of the algae (Manefield et al., 1999; Rasmussen et al., 2000). Lotus corniculatus is able to enzymatically degrade HSLs in the medium as well (Delalande et al., 2005). Plants react to the application of HSLs in a discrete manner. In Medicago truncatula, HSL application led to changes in protein expression, with consequences for the plant immune reaction, general stress reaction, primary metabolism, and phytohormone responses (Ortíz-Castro et al., 2009; Teplitski et al., 2011). In tomato, inoculation with the HSL-producing strain Serratia liquefaciens MG1 led to systemic acquired resistance against Alternaria alternata, while inoculation with the HSL-negative mutant Serratia liquefaciens MG44 failed to induce resistance (Schuhegger et al., 2006). Schikora et al. (2011) demonstrated that oxo-C14-HSL induces resistance against biotrophic but not necrothrophic pathogens in Arabidopsis thaliana. Growth effects and changes in root architecture were found in A. thaliana after inoculation with short-chain HSLs (Ortíz-Castro et al., 2008; von Rad et al., 2008; Schikora et al., 2011). Additionally, regulation of auxin- and cytokinin-dependent genes was found (von Rad et al., 2008). Interestingly, a class of plant secondary metabolites, namely N-acylethanolamides and alkamides, which are both structurally related to HSLs, have been found to induce similar growth responses in plants, including altered root architecture (López-Bucio et al., 2006; Morquecho-Contreras et al., 2010). Moreover, an A. thaliana mutant with reduced reaction to both N-isobutyl-decanamide and C10-HSL was identified, indicating a potentially similar mechanism behind the plants' perception of alkamides and HSLs (Morquecho-Contreras et al., 2010). Alkamide signalling requires a functional cytokinin-signalling pathway, which seems to be independent of auxin signalling (López-Bucio et al., 2007; Morquecho-Contreras et al., 2010).
Interestingly, Pseudomonas spp. that are found predominantly in the rhizosphere produce HSLs more often than species that are found in bulk soil (Elasri et al., 2001; DeAngelis et al., 2007). It is argued that the increased occurrence of HSL-producing bacteria in the rhizosphere is used for inter-kingdom communication between plants and bacteria (Bais et al., 2006). An uptake of bacterial autoinducers into plants could represent a first step in the communication pathway. Recently this has been confirmed by demonstrating that HSL signalling modulates the interaction between Pseudomonas aeruginosa and A. thaliana (Ortiz-Castro et al., 2011). Several studies have shown that HSLs are transported into the root of A. thaliana (von Rad et al., 2008; Schikora et al., 2011), Hordeum vulgare, and Pachyrhizus erosus (Götz et al., 2007). Whereas short chain HSLs such as C6- or C8-HSL were also detected in the shoots (Götz et al., 2007; von Rad et al., 2008; Schikora et al., 2011), a systemic transport of long-chain HSLs such as C10-HSL (Götz et al., 2007) or oxo-C14-HSL (Schikora et al., 2011) could not be confirmed. These studies primarily measured HSLs in plants after a relatively long period of incubation. In the present study, we were particularly interested in the early uptake and transport of HSLs in plants, as it could be a great evolutionary advantage for a plant to be able to react as quickly as possible to any inter-kingdom signal. As yet, it is still unknown whether HSL uptake is an active process of the plant or if whether occurs passively, and in which tissues HSLs are transported. We used different HSL detection methods to address and critically discuss these questions.
Materials and Methods
Plant growth conditions
Seeds of Hordeum vulgare L. cv Barke 2003 (Josef Breun GdbR, Herzogenaurach, Germany) where surface-sterilized according to standard protocols: After incubating the seeds in 1% Tween80 solution for 2 min and 70% ethanol for 5 min, seeds where washed with sterile dH2O and put into 12% sodium hypochlorite solution for 20 min. Seeds were kept in sterile dH2O for 4 h to swell, immersed again in 12% sodium hypochlorite solution for 10 min, thoroughly washed and placed on nutrient broth-agar plates. Three days after germination in the dark at room temperature, seedlings without visible contamination were transferred to a sterile vial system consisting of two Duran vials with a diameter of 30 mm. The lower vial contained 50 g glass beads (diameter, 1.7–2 mm; Carl Roth GmbH, Karlsruhe, Germany) and 10 ml MS medium (Duchefa Biochemie BV, Haarlem, the Netherlands). Each seedling was placed 0.5–1 cm deep into the glass beads and the vials were adjusted together with Parafilm. Plants were grown for 14 d in a climate chamber with a day : night cycle of 14 : 10 h and day : night temperatures of 15 : 12°C.
Roots for PITMAN-chamber experiments were taken from seedlings growing on nutrient broth-agar plates for 5–7 d.
Plant sap sampling
Plant sap was collected for sensor strain and enzyme-linked immunosorbent assay (ELISA) with a method modified after Djordjevic et al. (2007).
Surface-sterilized seeds were germinated for 3 d on NB agar plates and seedlings were transferred to a sterile glass vial system. The lower vial (500 ml) contained a 3 cm layer of glass beads and 30 ml of MS medium. Ten seedlings per vial were placed c. 0.5 cm deep into the glass beads and the system was sealed with Parafilm after adjusting a second vial on top. After 8 d the vial system was carefully opened and C8-HSL or C10-HSL (Sigma-Aldrich) was applied via a syringe to a final concentration of 10 μM. HSL stock solutions were always freshly prepared in ethanol, and controls were treated with corresponding amounts of ethanol. All media and nutrient solutions were buffered to ensure HSL stability.
Plant stalks were harvested after 4 h, cut into pieces of 1 cm in length, placed into apo-solution (20 mM sodium-ascorbate, 20 mM calcium chloride), and vacuum-infiltrated on ice in a desiccator for 1 h after which the vacuum was slowly released. The apo-solution is slightly acidic. This, together with working at low temperatures, increases HSL stability. Stalk pieces were briefly dried on tissue paper, cautiously put into 0.5 ml Eppendorf tubes and centrifuged at 1500 g for 30 min at 4°C. The plant sap collected was stored in threaded glass vials (1.5 ml; Neolab, Heidelberg, Germany) at −20°C until further use.
The stalk pieces (further referred to as ‘dry’ stalk material) were ground in liquid nitrogen and sonicated for 15 min in 25% acetonitrile. After centrifugation at 2000 g for 10 min the supernatant was collected into a separate vial and the extraction was repeated. Supernatants were combined, vacuum-evaporated and re-dissolved in distilled water, because remainders of acetonitrile would interfere with the bacterial sensor strain test. Samples were stored in amber glass vials at −20°C until further use, but never longer than 3 wk to ensure HSL stability.
Hordeum vulgare plants were inoculated with [3H]-C8- or [3H]-C10-HSL via a sterile syringe through a silicon septum in the lower glass vial to a final concentration of 10 μM and a radioactivity of 100 μCi. Radiochemical purity of both substances was > 98%. Plants were harvested 1, 2.5, 4, and 16 h after HSL treatment and cut into four parts: root, and shoots 1–3. Shoot 1 consisted almost solely of stalk material, and shoots 2 and 3 of leaves. Shoots 1 and 2 were cut to 5 cm length, whereas the length of shoot 3 varied somewhat. Plant material was extracted in a similar manner to the stalk material as described in the previous section. One millilitre of the supernatants was collected into separate vials and 0.5 ml of the combined supernatants were mixed with 5 ml of scintillation cocktail Rotiszint eco plus (Carl Roth GmbH). Measurements were performed in a scintillation vial for 5 min in a scintillation counter (LS 6500; Beckman, Brea, CA, USA). Two replicates were measured from each sample.
Fourteen-day-old barley plants were fixed to a rack and roots were spread on a glass Petri dish so that only root tips were attached to the glass surface. To each root tip, 20 μl of 10 μM and 10 μCi [3H]-C8- or [3H]-C10-HSL solution were applied. Roots were harvested after 15, 30, and 60 min, rinsed with 25% acetonitrile, dried and immediately placed on a tritium-sensitive film. Films were incubated for 27 h at −20°C. Radioactivity was visualized with a Fujifilm BAS-5000 Scanner. Pictures were scanned with a sensitivity of 1000 and a pixel size of 50.
For root cross-sections, maize seedlings with c. 4-cm-long main roots were positioned in a glass vial so that only the root tip touched the bottom of the vial. To the root tip, 40 μl HSL-solution with a concentration of 10 μM and a radioactivity of 20 μCi was applied for 1 h. Cross-sections were done using a cryotome (Mikrom HM 560; Mikrom, Walldorf, Germany). Therefore, small root pieces cut directly behind the root tip, from the middle of the root and from the root base, were embedded in cryogel (Instrumedics Inc., Richmond, IL, USA) on a metal disc. Cuts were made at an angle of 5°. Working temperature in the cryotome was set to −19°C and specimen temperature to −21°C. Cross-sections were c. 30 μm thick and placed onto tritium-sensitive film, which was incubated at −20°C for c. 65 h and then scanned with a sensitivity of 30 000 and a pixel size of 25.
Root transport studies in a Pitman chamber
Detached roots were transferred into a Pitman chamber (Pitman, 1971), which consists of acrylic glass and flexible bars, so it can be divided into variable compartments. For this experiment the chamber was divided into three compartments. Root tips were located in the first compartment and cut root ends in the third. To seal up all gaps between the compartments, Vaseline was injected with a syringe between the chamber and the bars and into the notches. Residual Vaseline was carefully removed. HSL solution was applied to the first at a concentration of 10 mM and a radioactivity of 100 μCi. Samples were taken right after application and after 2, 4, 8, and 24 h from the third compartment; a magnetic stirrer ensured an even distribution of the transported HSL. The second compartment served as a buffer against diffusion from the first to the third compartment. To inhibit HSL transport through the root, KCl was applied to the second compartment after 4 h at a final concentration of 110 mM. To inhibit uptake of HSL, sodium-orthovanadate was applied to the first compartment after 4 h at a final concentration of 1 mM.
Sensor strain assay
C8-N-acyl-homoserine lactone was measured in plant sap using the HSL sensor strain Serratia liquifaciens MG44 (Eberl et al., 1996), which carries a Green Fluorescent Protein under the control of an HSL-inducible promoter on the plasmid pBAH9. To ensure the stability of the test, a standard curve with C8-HSL was done in parallel. The test was more robust when bacteria were taken out of glycerol stocks instead of agar plates. S. liquifaciens MG1 and S. liquifaciens MG44 were grown at 30°C. Overnight cultures were grown from glycerol stocks in LB media containing 100 μg ml−1 ampicillin, 50 μg ml−1 kanamycin, and 10 μg ml−1 tetracycline (all Sigma-Aldrich), pH 7. After overnight cultivation at 30°C, cultures were diluted 1 : 4 with LB media and grown for another 2.5 h. To prove that the assay was functional, C8-HSL was diluted in LB media to 0.1–5 μM for a standard curve. Plant sap samples were sterilized via a centrifuge filter system (mini) with a pore size of 0.2 μm (Carl Roth GmbH) by centrifuging at 7400 g for 10 min. The setup was done in Duran glass vials (diameter 1.6 cm, height 16 cm; Carl Roth GmbH) containing 1.15 ml LB medium, 250 μl standard or plant sap, and 100 μl of the S. liquifaciens MG44 culture diluted 1 : 4 with LB medium. Controls contained 250 μl LB medium instead of sample and cultures were grown for another 6 h. Each setup was placed in four replicates of 200 μl onto FluoroNunc™ plates (black; Thermo Fisher Scientific, Schwerte, Germany) and instantly measured in a fluorimeter (Spectra Max Gemini EM; Molecular Devices GmbH, Ismaning, Germany).
Immunoassays with monoclonal antibodies (mAbs) were performed in the coating antigen format in a similar manner to that described in Chen et al. (2010) (Fig. 1). Optimal dilutions of all reagents were determined by two-dimensional titration. Nunc MaxiSorp™ 96-well microtitre plates (Thermo Fisher Scientific) were coated overnight with 100 μl HSL3-OVA1/4 in 50 mM carbonate buffer (pH 9.6) at a final concentration of 0.5 μg ml−1 at 4°C. Nunc MaxiSorp™ plates were washed after each incubation step with an automated microtitre plate washer (Bio-Tek Instruments, Bad Friedrichshall, Germany) using a washing buffer containing 4 mM PBS (pH 7.6) and 0.05% Tween20. After overnight incubation and washing, the plates were blocked with 200 μl 1% skimmed milk powder (casein sodium salt; Carl Roth GmbH) for 1 h at room temperature to prevent any unspecific binding. Plant sap was diluted 1 : 5 in PBS. A standard curve of C10-HSL was set up in plant sap (1 : 5 diluted with 40 mM PBS, pH 7.6) with concentrations ranging from 1 to 50 000 μg l−1 (3.9 nM–0.2 mM) in protein low-bind tubes (Eppendorf, Hamburg, Germany). All probes and standards were hydrolysed by adding 10% (v/v) 1 M NaOH, incubated for 15 min at room temperature on a shaker and then readjusted to pH 7 by adding 8% (v/v) 1 M HCl. Thereafter all samples were centrifuged at 8500 g for 10 min to discard the Ca2+ that precipitated when plant sap and PBS buffer were mixed. Supernatants were incubated 1 : 1 (v/v) with the mAb HSL4-6D3 (culture supernatant diluted 1 : 2000 with 40 mM PBS, pH 7.6) in Greiner 96-well U-shaped medium binding plates (Greiner Bio-One, Solingen-Wald, Germany) shaking on a Heidolph incubator 1000 (Metrohm, Herisau, Switzerland) for 2 h at room temperature. A 100 μl quantity of the antibody–analyte mixture was transferred from the U-shaped plate to the washed Nunc MaxiSorp™ plate and incubated for 1 h at room temperature after another incubation with 100 μl of the second antibody GAR-HRP (Dianova, Hamburg, Germany; 1 : 8000 diluted in 40 mM PBS, pH 7.6). After 1 h, 100 μl of the substrate (0.4 mM tetramethyl-benzidine (Sigma-Aldrich), 1.3 mM H2O2 in 100 mM sodium acetate buffer, pH 5.5) was placed on the washed plate and incubated for 15 min at room temperature in darkness. The reaction was stopped by adding 50 μl 2 M H2SO4 without prior washing, and absorbance of the completely oxidized form of tetramethyl-benzidine was measured at 450 nm (reference 650 nm) with a multi-detection reader Spectra Max M5e (Molecular Devices, Palo Alto, CA, USA, now part of Danaher Corporation, Washington, DC, USA).
All statistical analyses were performed with SigmaPlot 11.0 (SyStat, San Jose, CA, USA), Microsoft Office Excel 2007 (Redmond, WA, USA), and Softmax Pro 4.6 (Molecular Devices Corp., Sunnyvale, CA, USA) software.
Exposure of roots to radioactively labeled C8-HL and C10-HL and time course of accumulation in shoots
To investigate systemic HSL transport in barley, [U−3H]C8- or [U−3H]C10-HSL was applied to the roots at a concentration of 10 μM or a radioactivity of 100 μCi. Both, C8- and C10-HSL are transported from the root into the shoot of barley linearly (linear regression for shoots 1, 2, and 3 taken together: C8-HSL, R² = 0.9957; C10-HSL, R² = 0.9934; n =6; Fig. 2). The highest radioactivity was always detected in the stalks independently of the HSL used. Less radioactivity was detected in shoot 2 and the lowest in shoot 3, albeit the difference between shoot 2 and shoot 3 was only marginal and not statistically significant. After 16 h of incubation, the values of radioactivity g–1 FW in shoots 2 and 3 together were only 26% (C8-HSL) and 30% (C10-HSL) of that in shoot 1. However, the difference between shoot 1 and shoot 2 + 3 was not as pronounced at earlier time points (Fig. 2). With C8-HSL the difference between shoot 1 and shoot 2 + 3 was statistically significant at all time points (1 h, P <0.05; 2.5 h, P <0.0001; 4 h, P <0.0001; 16 h, P <0.0001, t-test, n =6). When C10-HSL was applied, the difference between shoot 1 and shoot 2 + 3 was statistically significant after 4 h (P <0.01, t-test, n =6) and after 16 h (P <0.001, t-test, n =6). C8-HSL entered the shoot at a higher rate than C10-HSL. There was 20% more C8-HSL g–1 FW transported into the whole shoot after 2.5 h than C10-HSL (t-test: P <0.05; n =4) and 26% after 4 h (t-test: P <0.05; n =6). After 16 h the difference between transported C8-HSL and C10-HSL was 27%; however, this was not statistically significant (t-test: P =0.16; n =6).
Time course analyses of HSL-accumulation at the outer root surface and HSL-transport through the roots
Linear transport rates of HSLs through the root were never observed, with the experimental setup used to examine transport into the shoot. Radioactivity measured in the roots had already reached saturation after 2–4 h and was > 100-fold higher at the end of the experiment than that measured in the shoot (Supporting Information, Fig. S1). Contrary to the results obtained from the shoot measurements, the detected radioactivity in roots after C10-HSL application was higher than after C8-HSL application, although this was mostly not statistically significant (1 h, P =0.08; 2.5 h, P =0.17; 4 h, P =0.01; 16 h, P =0.44; t-test, n =6). Similar results were obtained by Götz et al. (2007), although in that work saturation was not reached until 50 h after inoculation.
One explanation of these findings is that both tested HSLs attach tightly to the root surface, thus camouflaging any ongoing transport through the root. To test this hypothesis [U−3H]C8- and [U−3H]C10-HSL were applied to the root tips of H. vulgare and root strands were cut from the plant 15, 30, and 60 min after treatment and rinsed with 25% acetonitrile to ensure that all loosely attached HSLs were removed. Acetonitrile was used because it was found to be the best solvent for AHLs in plant tissues that could be used for liquid scintillation without further problems. Roots were then positioned on a tritium-sensitive film and scanned. Tritium is a relatively weak β-emitter, and thus radiation does not penetrate tissues thicker than 2 μm. Consequently, the radiation detected on the film probably originated solely from HSLs attached to the outer root surface and not from HSLs transported through the roots. Indeed, after 15 min of incubation, both C8- and C10-HSL were found to be tightly attached to the root surface (Fig. 3). After 60 min of treatment, at least 2% of the applied C8-HSL and 4% of the applied C10-HSL were attached to the root surface. At every time point, twice as much C10-HSL attached to the root surface as C8-HSL. After 60 min, the shoots of the treated plants were harvested and analysed for HSL uptake via liquid scintillation. The total uptakes of C8- and C10-HSL into the shoot after 1 h were 1.26 and 0.34%, respectively.
The amount of surface-bound HSLs after 1 h was relatively high compared with the amount measured in the shoots, especially in the case of C10-HSL (11 times higher) and the attachment to the root surface was rather strong. We concluded that root uptake and transport cannot be detected with conventional methods, because surface-bound HSLs would overlay it. Therefore, we examined HSL uptake and transport in barley roots via a Pitman chamber (Pitman, 1971). Roots were cut from the seedling and clamped into the chamber, creating three compartments. [U−3H]C8- or [U−3H]C10-HSL was applied to the root tips in the first compartment at a concentration of 10 μM and a radioactivity of 100 μCi. Samples were taken from the third compartment where the root bases ended (Fig. S2). Proper leak tightness was ensured by sampling compartment 2: if radiolabelling occurred in this compartment, the experiment was rejected. This setup ensured that only a minimal amount of root tip surface was in direct contact with the HSL solution. Both, C8- and C10-HSL were transported through barley roots following a linear regression (C8-HSL, R² = 0.9994; C10-HSL, R² = 0.9962, n =5) (Fig. 4a). C10-HSL was transported at a slightly lower rate than C8-HSL, although the difference was not significant (2 h, P =0.36; 4 h, P =0.18; 8 h, P =0.47; 16 h, P =0.64; t-test, n =5). After 24 h, 0.2% of the initially applied concentration of C8-HSL (c. 20 nM) and 0.1% of C10-HSL (c. 10 nM) was transported through the root. This is equal to about one-tenth (C8-HSL) to one-fortieth (C10-HSL) of the amount that was bound to the root surface.
Application of inhibitors of HSL uptake and transport in roots
To examine whether the uptake and transport of HSLs occurs passively, if it requires specific transporters or other active plant processes, roots were placed into the Pitman chamber and [U−3H]C8-HSL was applied to the root tip compartment as described earlier. Samples were taken at the beginning of the experiment and after 2, 4, 6, 8, and 24 h from the third compartment. Transport was monitored for 4 h before orthovanadate was applied to the root tip compartment in a final concentration of 1 mM. Orthovanadate acts as a phosphate analogue, thus blocking ATP-dependent ABC transporters. After adding ortho-vanadate to the root tip compartment, an increase of radioactivity in the third compartment was detected for another 4 h. From that time point until the end of the experiment, there was no further increase (Fig. 4b). For the first 4 h before vanadate treatment, the increase of radioactivity in compartment three was virtually identical, and from t = 4 h to t = 8 h, rates were not statistically significantly different (27 Bq h−1 in controls and 20 Bq h−1 in vanadate treatment; P =0.65; t-test; n =5). However, the radioactivity in compartment three between t = 8 and t = 24 h in the vanadate treatment (1 Bq h−1) was significantly lower than in the control treatment (28 Bq h−1, P <0.05; t-test; n =5). It is reasonable to assume that HSLs were already taken up into the roots before the orthovanadate treatment and were further transported to the root end, thus creating a delay until no further increase of radioactivity in the third compartment was detected. Consequently, by blocking the HSL transport mechanism instead of the uptake, one would see an immediate inhibition of detected radioactivity in the third compartment.
To test whether HSLs are transported in the apoplast or in the symplast via the central cylinder, we disrupted symplastic transport by the addition of 110 mM KCl. Incubation with 110 mM KCl leads to closure of plasmodesmata and disrupts symplastic transport (Diekmann, 2005). KCl was applied to the second compartment because here the main part of the root was located, and thus it was here that the transport through the roots could be best manipulated. In that way, it was also ensured that there was no direct contact between the applied HSL and the KCl that could interfere with HSL stability. In order to establish and observe an undisturbed HSL transport in the first 4 h, KCl was applied 4 h after the start of the experiment. The increase of radioactivity in compartment three between t = 0 h and t = 4 h was 28 Bq h−1 in the control treatment and 28 Bq h−1 in the KCl treatment without not statistically significant difference between the treatments during that period (P =0.98; t-test; n =5; Fig. 4b). After the addition of KCl to the second compartment of the Pitman chamber, there was no further increase of radioactivity detected in the third compartment. However, at the end of the experiment the detected radioactivity increased slightly, but did not reach the values of an undisturbed transport (Fig. 4b). The measured increase of radioactivity in compartment three between t = 4 h and t = 24 h was 28 Bq h−1 in the control treatment and 6 Bq h−1 in the KCl treatment, with a statistical significance of P =0.02 (t-test; n =5). This indicates that the HSL transport through the roots occurs predominantly in the symplast, although a slower bypass through the apoplast cannot be excluded. To further verify these results, root cross-sections of seedlings treated with tritium-labelled AHLs were made using a cryotome and developed on a tritium-sensitive film. In this experiment, maize roots had to be used, because barley roots are too filigreed to produce high-quality cross-sections. Therefore [3H]-C8- or [3H]-C10-HSL (10 μM, 20 μCi) was applied to the root tip of the seedlings and incubated for 1 h. Both C8-and C10-HSL could be detected in cross-sections of maize roots (Fig. 5). The radioactive signals in cross-sections from the area shortly behind the root tip were more intense than in cross-sections from the middle of the root. Notably the highest radioactivity was located in the middle of all cross-sections in the area of the central cylinder. In contrast to cross-sections from the middle of the root, in cross sections from the root tip, signals of radioactivity could also be detected outside the central cylinder. This further indicates that HSLs are taken up behind the root tip and transported predominantly through the central cylinder, whereas a transport through the apoplast seems to occur, albeit to a lower extent.
HSL-detection in plant sap using the HSL-sensor strain Serratia liquefaciens MG44
To test whether HSL transport in the shoot also proceeds via the central cylinder and to verify that transported HSLs are unmetabolized, plant sap of barley shoots was collected 4 h after treatment with 10 μM C8-HSL or C10-HSL. The HSL sensor strain Serratia liquefaciens MG44 was used to detect C8-HSL. Bacterial sensor strains do react very specifically to their respective HSLs, but not to HSL metabolites. Thus any detected fluorescence is attributed to the original compound. Standards could be measured up to a concentration of 0.1 μM C8-HSL. C8-HSL could be detected with a statistically significant difference from control treatments in the plant sap samples (P <0.0001; t-test; n =3) and in the ‘dry’ tissue of the stalk material of H. vulgare (P <0.01; t-test; n =3; Fig. 6a). The standard curve of C8-HSL was not linear, nad hence it is not possible to quantify the concentration of C8-HSL in the stalk samples; however, it is < 1 μM. C10-HSL could not be detected with S. liquefaciens MG44, because the sensitivity of the sensor strain against C10-HSL was not high enough (data not shown).
HSL detection in plant sap using mAbs
To detect transported C10-HSL, mAbs were used. The test was done in the coating antigen format with the mAb HSL4-6D3 and the coating antigen HSL3-OVA1/4. Currently available mAbs against HSLs are far more sensitive against their corresponding homoserines, which is probably because of the fact that the HSLs are hydrolysed during the conjugation and immunization process when the mAbs were fabricated (Chen et al., 2010). As the expected concentration of C10-HSL in the plant sap of barley was very low, all samples and standards were also hydrolysed according to the procedure mentioned in the Materials and Methods section. In addition, all samples, including the negative controls, were HSL-spiked and the spiking was subsequently subtracted. With the test optimized this way C10-HSL could be detected in the plant sap of H. vulgare with a statistically significant difference from control treatments (P <0.5; t-test; n =4; Fig. 6c). The concentration amounted to 38 μg l−1 after the subtraction of the spiking, which, by adding the dilution factor, resulted in a concentration of 190 μg l−1or 0.7 μM of C10-HSL in the plant sap. The corresponding standard curve had a fit of R2 = 0.999 (n =3; Fig. 6b).
This study shows that HSLs are taken up actively by barley roots and systemically transported into leaf tissues primarily through the central cylinder. H. vulgare was chosen for the present study, because HSLs in its rhizosphere are stabilized through the excretion of organic acids (Götz et al., 2007). Moreover, barley does not possess any lactonases (Götz et al., 2007), in contrast to some legumes (Delalande et al., 2005), which would rapidly inactivate the applied HSLs.
Instead of using classic analytical methods, the HSL-detection in this study was performed with: radiolabelled compounds; HSL sensor strains that are able to detect intact HSLs, but no metabolites; and specific mAbs. This was necessary because analytical methods require a rather large sample volume, which was impossible to obtain with plant sap-sampling, and ultra performance liquid chromatography (UPLC) and Fourier-transform ion cyclotron resonance mass spectrometry (FTICR-MS) have a detection limit for HSLs of c. 390–440 nM (Malik et al., 2009; Chen et al., 2010). In particular, the initial transport through the root was far below these limits.
The use of radioactively labelled HSLs enabled us to detect HSLs in the pM range and is, to date, the most sensitive method available. HSL uptake and transport occurs in a linear manner within the first 24 h after application. First signals in the shoots could already be detected after 2 h. C8-HSL was transported at a higher rate than C10-HSL. Although this was not always significant, there is a firm pattern. This is explained by the longer acyl-side chain of C10-HSL which makes the molecules more lipophilic than those of C8-HSL. However, it was problematic that both HSLs attached strongly to the outer root surface as a result of hydrophobic interaction. Using a Pitman chamber is one suitable method to circumvent these problems, because the contact between root surface and HSL-solution was minimized and only HSLs that were transported through the root from tip to root base are measured. In tobacco plants bioengineered to produce short- and long-chain HSLs, it could be shown that the distribution of short-chain HSLs apparently took advantage of transport mechanisms already operative in the plant (Scott et al., 2006). Correspondingly, in our system, the uptake of C8-HSL into the root could be inhibited with orthovanadate. Orthovanadate as a phosphate-analogue is able to block ABC transporters, indicating that plant ABC transporters might be involved in the transport of HSLs, and thus the uptake and distribution rely at least partially on active parts of the plant (Rea, 2007). Interestingly, in human T-cells the preincubation with HSLs in the μM range results in the decrease of the binding capacity of an ABC transporter known to play a crucial role in multidrug resistance (Davis et al., 2010). These authors suggest that the decrease of the binding capacity of this specific ABC transporter is caused by the partitioning of AHLs into biological membranes and an associated decrease of the membrane dipole potential. However, direct interaction between the ABC transporter and AHLs could not be proven. The partitioning into membranes as well as the decrease of the dipole potential and further associated immunosuppressive effects were observed after the use of long-chain HSLs, such as 3-oxo-C12-HSL and 3-oxo-C14-HSL. Accordingly, the use of C8-HSL in our study should have smaller effects on membrane dipole potentials.
Considering the present state of knowledge it is reasonable to assume that the interaction between HSLs and different members of the ABC transporter superfamily is highly diverse, yet ABC transporters might play an important role in the transport and indirectly mediate HSL-based inter-kingdom signalling between bacteria and eucaryotes.
Root transport of C8-HSL could be blocked through the application of KCl almost entirely. In root cross-sections, the radioactive signal was predominantly located in the central cylinder, but in cross-sections from the root tip, radioactivity was also visible outside the central cylinder, indicating an uptake close to the youngest tissues. Thus it is reasonable to assume that HSLs are transported via the central cylinder across the root, although some apoplastic transport might also occur, but to a lower extent, as in P. aeruginosa where short-chain C4-HSL and OC-6-HSL have been described to diffuse freely into and out of the cells, whereas longer-chain HSLs exhibited membrane partitioning and required efflux pumps (Pearson et al., 1999). The symplastic transport via the central cylinder enables a faster systemic transport across the plant. As a by-product of the autoradiography experiments, we could show that the uptake of C8-HSL was enhanced by transpiration. However, the exact role of transpiration forces have to be examined further. Interestingly, Joseph & Phillips (2003) detected an enhanced transpiration rate in bean plants after application with homoserines up to 30%. It is speculative whether it is important to the plant to systemically transport HSLs faster and at higher rates. The fact that radioactive signals in the shoot were mainly associated with the stalks can eventually be attributed to the experimental setup. Since the plants grew within 100% humidity there was a lack of transpiration force, and consequently HSLs were not effectively transported into the fine leaf vessels. Additionally, it has to be taken into account that the hydroponic system used in this setup is artificial and leads to altered root structure, although glass beads were used to mimic some physical soil conditions.
Since the tritium label was evenly distributed across the molecule, it is not possible to discriminate between the original substance and any kind of HSL-metabolite solely by detecting the radioactive signal. The plant sap collected from the stems was therefore analysed using the bacterial sensor strain S. liquefaciens MG44 for C8-HSL-treated plants. This was done to verify that the systemic transport of C8-HSL into the shoot is also mediated through the central cylinder and that the substance is, at least in part, still the original, nonmetabolized molecule. With this system we could show that the majority of C8-HSL was located in the plant sap. Little was associated with the stalk material after extracting the sap, and thus we verified the results of Götz et al. (2007), who detected C8-HSL in barley shoots via UPLC and FTICR-MS. C10-HSL could not be detected with S. liquefaciens MG44, because of a lack of sensitivity against this HSL. A test with mAb against C10-HSL was therefore established and C10-HSL could in fact be detected in the plant sap of barley. As the plant sap of barley is slightly acidic, it is reasonable to assume that the detected HSLs were chemically unaltered and not hydrolysed. The concentration of C10-HSL in the plant sap amounted to 0.7 μM. It is very likely that C10-HSL bound to tissue inside the stalk even more than did C8-HSL. To our knowledge, this is the first demonstration that longer-chained HSLs are transported systemically and in nonmetabolized form from root to shoot inside plants.
Götz et al. (2007) had previously detected tritium-labelled C8- and C10-HSL in the shoots of barley plants, but the verification via UPLC and FTICR-MS had only been successful in the case of C8-HSL. The authors therefore speculated that C10-HSL might enter the shoot in metabolized forms. This could also be true for other long-chain HSLs like oxo-C14-HSL, which likewise could be detected only in the root, but not in the shoot of A. thaliana. The detection of HSLs in roots in all previous studies was done by extracting them from the root material (von Rad et al., 2008). We have shown that HSLs attach tightly to the outer root surface and that the amount that is bound to the root surface exceeds the amount that is transported into the root by far. It is very likely that most of the HSL detected by extraction from root material originated from HSLs bound to the root surface. Long-chain HSLs attach to surfaces even more and are transported less than short-chain HSLs, as a result of their higher hydrophobicity. Consequently, the error that is made by measuring root uptake of long-chain HSLs by conventional methods is even higher. The amount of HSLs that are taken up and transported into the root should only be measured by methods that take the surface binding into account.
Previous studies showed that HSLs induce various responses in different plant species, including systemic resistance (reviewed in Hartmann & Schikora, 2012). If, and to which extent, the systemic transport of HSLs is needed to establish systemic resistance is still not known; nor is it known if there is any kind of HSL receptor in plants. Davis et al. (2010) proposed that partitioning of long-chain HSL into T-cell membrane domains containing high proportions of cholesterol might modulate susceptibility or trigger defense of a host, and thus influence transmembrane proteins. The results obtained in this study suggest that among membrane-located systems, ABC-transporter proteins might be one important target that could at least provide rapid and widespread signal transport. In light of this, the present paper provides strong evidence that an active transport system for HSLs exists in plants. More research has to be done to understand how and whether transport can be modulated by HSL of different chain lengths, and how HSLs and other QS molecules contribute to trans-kingdom interactions and cellular signalling.