Arabidopsis histidine kinase 5 regulates salt sensitivity and resistance against bacterial and fungal infection


Author for correspondence:
Radhika Desikan
Tel: +44 20 7594 3895


  • The ability of plants to adapt to multiple stresses imposed by the natural environment requires cross-talk and fine-tuning of stress signalling pathways. The hybrid histidine kinase Arabidopsis histidine kinase 5 (AHK5) is known to mediate stomatal responses to exogenous and endogenous signals in Arabidopsis thaliana. The purpose of this study was to determine whether the function of AHK5 in stress signalling extends beyond stomatal responses.
  • Plant growth responses to abiotic stresses, tissue susceptibility to bacterial and fungal pathogens, and hormone production and metabolism of reactive oxygen species were monitored in a T-DNA insertion mutant of AHK5.
  • The findings of this study indicate that AHK5 positively regulates salt sensitivity and contributes to resistance to the bacterium Pseudomonas syringae pv. tomato DC3000 and the fungal pathogen Botrytis cinerea.
  • This is the first report of a role for AHK5 in the regulation of survival following challenge by a hemi-biotrophic bacterium and a necrotrophic fungus, as well as in the growth response to salt stress. The function of AHK5 in regulating the production of hormones and redox homeostasis is discussed.


Plants are challenged by many different organisms and, because of their sessile lifestyle, must defend themselves not only against parasites, pathogens and herbivores (biotic factors) but also against stresses imposed by the environment (abiotic factors). These stresses and stimuli can act in succession or in combination. As a result, plants have developed complex signalling networks with common components acting to integrate responses to multiple stimuli.

Cross-talk and overlap between signalling pathways allow plants to mediate rapid responses, either when multiple stimuli are likely to occur together or where different stresses require a similar physiological response. By integration of signalling pathways based on the physiological effects of the stress, plants are able to coordinate and fine-tune responses without the need for separate signalling pathways for every stimulus or combination of stimuli that they may encounter.

An important mechanism used for intracellular signalling is the phosphorylation of proteins mediated by protein kinases. Of recent interest are the plant histidine kinases (HKs) of the two-component systems which have also been identified in yeast, bacteria, amoeba and plants (Urao et al., 2000; Hwang et al., 2002; Grefen & Harter, 2004). In Arabidopsis, 11 HKs have been identified, of which nine are of the hybrid type, the general structure of which consists of an input domain, an HK domain and a receiver domain, whereas the archetypal nonhybrid HKs lack the receiver domain. Signal transduction occurs via a phosphorelay between the HK, a histidine-containing phosphotransfer protein (AHP) and a response regulator (ARR), leading to changes in target gene expression or protein activity (Urao et al., 2001; Hwang et al., 2002; Grefen & Harter, 2004).

The nine hybrid HKs of Arabidopsis are known to function in ethylene signalling (ETR1 ETHYLENE RESPONSE1, ETR2 ETHYLENE RESPONSE2 and ERS1 ETHYLENE RESPONSE SENSOR1) (Chang et al., 1993; Schaller et al., 2008), as osmosensors (Arabidopsis histidine kinase 1 (AHK1)/AtHK1, AHK2, AHK3, AHK4/CRE1/WOL CYTOKININ RESPONSE1/WOODEN LEG and ETR1) (Urao et al., 1999; Zhao & Schaller, 2004; Tran et al., 2007), as cytokinin receptors (AHK2, AHK3 and AHK4/CRE1/WOL) (Higuchi et al., 2004), in megagametophyte development (CKI1 CYTOKININ INDEPENDENT1) (Hejátko et al., 2003), in temperature perception (AHK2 and AHK3) (Jeon et al., 2010) and more recently as integrators of environmental and endogenous signals (AHK5/CKI2) (Desikan et al., 2008).

Of these kinases, the least is known about AHK5, the only hybrid HK initially predicted to have a cytoplasmic location and subsequently shown to be localized in both the cytoplasm and the plasma membrane (Desikan et al., 2008). AHK5 was first shown to function as a negative regulator of root growth inhibition mediated by ABA/ethylene (Iwama et al., 2007). Subsequently, we have shown that AHK5 integrates abiotic and biotic stimuli in stomatal guard cells through regulation of H2O2 homeostasis (Desikan et al., 2008).

Here, we demonstrate that the function of AHK5 in abiotic and biotic stress signalling is not restricted to stomata. Using the ahk5-1 T-DNA insertion line (Desikan et al., 2008) and ahk5-1 complemented with full-length AHK5, we show that AHK5 function is required for full immunity to the bacterial pathogen Pseudomonas syringae pv. tomato DC3000 (PstDC3000) and the necrotrophic fungus Botrytis cinerea. In addition, loss of function of AHK5 also confers tolerance to high salinity, suggesting that AHK5 acts to integrate multiple stress responses. Data are also presented to suggest a regulatory function for AHK5 in redox and hormone balance.

Materials and Methods

Plant lines and growth conditions

All plants were sown on Levington’s F2 compost + sand (Scotts, Ipswich, UK) and grown under short-day conditions of 10 h light : 14 h dark cycles with a light intensity of 120–150 μmol m−2 s−1 at 23°C and 55–65% relative humidity. Plants used for experiments were 5–6 wk old, unless otherwise stated.

The ahk5-1 mutant containing a T-DNA insertion in the receiver domain of the AHK5 histidine kinase was originally obtained from Syngenta (Minnetonka, Minnesota, USA) (SAIL 50_H11). Two complemented lines, PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4, were generated by complementation of the ahk5-1 mutant with full-length AHK5 including 3205 bases upstream of the ATG start codon. The plasmid pMKC111 containing this construct was generated as described by Desikan et al. (2008) and used for complementation of ahk5-1 mutant plants. pMKC111 was transformed into Agrobacterium tumefaciens strain GV3101 and subsequently transformed into ahk5-1 plants by vacuum transformation. T3 lines were isolated by selection on 50 μg ml−1 hygromycin B (Duchefa, Haarlem, the Netherlands). Confirmation of expression of AHK5 in the mutant background was obtained by RT-PCR (Supporting Information Fig. S1).

Salinity treatment of seedlings in tissue culture

For seedling root growth assays, seeds were surface-sterilized and plated onto half-strength Murashige and Skoog medium (Duchefa), pH 5.7, with 1.5% agar and supplemented with 0, 25, 50, 100 and 150 mM NaCl. Plated seeds were stratified at 4°C for 2 d before transfer to a growth chamber and maintained vertically under a 16-h photoperiod with a light intensity of 120–150 μmol m−2 s−1 at 23°C. After 7 d of growth, root length was measured with a ruler to an accuracy of ± 0.5 mm. In addition, the number of germinated seeds was also recorded for determination of percentage germination in response to salinity.

For the seedling survival assay, seeds were surface-sterilized, stratified and grown on half-strength Murashige and Skoog medium as above. After 7 d of growth, seedlings were transferred to plates containing half-strength Murashige and Skoog medium which contained 200 mM NaCl and the number of bleached seedlings three, five, seven, nine and 11 days after transfer were monitored.

Salinity stress in mature plants

Three-wk-old plants were watered with 250 mM NaCl solution or water as controls. At the start of the experiment, plants were watered with 30 ml of salt solution directly into each pot. Plants were then watered in the same fashion three times a week for 2 wk, resulting in a total of six watering events. The aerial portion of the plant was then harvested on the third day after the last watering dose for determination of fresh shoot weight.

Bacterial growth and surface inoculation of plants

Pseudomonas syringae pv. tomato DC3000 (PstDC3000) used in this study was provided by M. Grant (University of Exeter, Exeter, UK). Bacteria were maintained on solid Pseudomonas Agar F Base medium (Merck, Darmstadt, Germany) grown at 25°C supplemented with 50 μg ml−1 rifampicin.

For inocula, PstDC3000 was grown overnight in liquid Luria Bertani (LB) broth supplemented with 50 μg ml−1 rifampicin at 25°C. Bacteria were pelleted, re-suspended in 10 mM MgCl2 and diluted to the appropriate density by estimating absorbance at 600 nm (A600). Silwet L-77 (Lehle Seeds, Round Rock, Texas, USA) was added to a concentration of 0.04% (v/v) and the inoculum gently coated onto both sides of the leaf. Plants were maintained in a growth chamber for the duration of the experiment and covered with a transparent propagator lid to increase humidity. To determine in planta population counts of bacteria after surface inoculation, leaves were weighed and surface-sterilized as described by Katagiri et al. (2002) and ground in quarter-strength Ringer’s solution (Merck, Darmstadt, Germany), and the number of colony-forming units (cfu) was counted after 2 d and expressed as cfu g−1 FW of leaves.

Estimation of chlorophyll content

Leaves inoculated as described in the previous section were excised and chlorophyll was extracted from leaves in 100% methanol at 50°C for 1 h in a heated block. A665 and A652 of the methanolic extracts were measured for calculation of chlorophyll content. Concentrations of chlorophylls a and b in the methanolic extracts were calculated as described by Porra et al. (1989).

Inoculation with Botrytis cinerea and assessment of disease progression

Botrytis cinerea (obtained from K. Denby, Horticulture Research International, Warwick, UK) was cultured on Potato Glucose Agar (PGA) (Sigma-Aldrich, UK) supplemented with 500 μg ml−1 spectinomycin and incubated at 18–20°C under Mini Black Light Blue fluorescent lamps emitting long-wave UV light (Phillips, Guilford, UK) for 7–10 d. Spores were harvested and adjusted to a concentration of 2 × 105 spores ml−1 in one-eighth strength Potato Dextrose Broth (PDB) (Sigma-Aldrich, UK).

Symptom development was assessed on detached leaves kept in clear plastic boxes lined with moist tissue with the petioles wrapped in tissue paper. The abaxial side was inoculated with 10-μl droplets of the spore suspension. Boxes were kept at 23°C under short-day conditions with a 10-h photoperiod at a light intensity of 25 μmol m−2 s−1. For assessment of the effect of diphenyleneiodinium chloride (DPI) (Sigma-Aldridge) and 2-(4-carboxyphenyl)-4, 4, 4, 5-tetramethylimidazoline-oxyl-3-oxide (cPTIO) (Sigma-Aldrich) treatment on lesion development, leaves were vacuum-infiltrated with 1 μM DPI or 50 μM cPTIO for 5 min and floated in the same solution for 2 h before inoculation with the B. cinerea spore suspension as described earlier in this section.

The severity of disease was assessed by visual inspection and based on the spread of the lesions. Symptoms were scored on a scale of 0–4, with each score denoting the severity of the lesions present as follows: 0, no lesions; 0.5, multiple small lesions within the inoculum droplet; 1, single large lesion within the inoculum droplet; 2, single large lesion spreading beyond the inoculum droplet; 3, tissue collapse; 4, collapse of tissue and sporulation.

For measurement of lesion area, images of individual lesions were captured using a stereomicroscope (Leica MZ16F; Leica, Wetzlar, Germany) with an attached Leica DFC300FX camera and lesion size was measured using Image J software (Abramoff, 2004).

Histological observations

Fungal hyphae on leaves were visualized by trypan blue staining as described by Xiao et al. (2003). Leaves were decolourized in an 8 : 1 : 1 : 1 mixture of ethanol : phenol : lactic acid : glycerol for 24 h, during which the solution was changed twice. Leaves were then stained for 30 min in 0.025% (w/v) trypan blue (Sigma-Aldrich) in a 1 : 1 : 1 mixture of lactic acid : glycerol : water, cleared in saturated chloral hydrate and mounted in 60% glycerol.

The presence of reactive oxygen species (ROS) was visualized by staining with 3,3-diaminobenzidine (DAB) (Sigma-Aldrich) at a concentration of 1 mg ml−1 in water (pH 3.8). Excised leaves were vacuum-infiltrated with DAB solution for 5 min. Leaves were then incubated in DAB in the dark overnight, destained in 100% ethanol and mounted in 60% glycerol.

To assess the effect of DPI on germination of B. cinerea spores in vitro, 10-μl droplets of a 2 × 105 spores ml−1 suspension in one-eighth PDB supplemented with 1 μM DPI or 50 μM cPTIO were deposited onto glass slides which were placed in clear plastic boxes as described in the previous section. Before microscopic observation, the hyphae were stained by adding a drop of lactophenol cotton blue solution (Pro-Lab Diagnostics, Cheshire UK) to the inoculum droplet; a coverslip was then applied and sealed with clear nail varnish.

Stained material was examined either with a stereomicroscope (Leica MZ16F), with images being captured with an attached Leica DFC300FX camera, or with a Zeiss AxioSkop2 Plus Microscope, with images being captured with an attached Zeiss AxioCam with AxioVision 3.1 software (Zeiss). Hyphal length and area of cellular DAB staining was quantified using Image J software.

Extraction and analysis of camalexin

Individual lesions were excised with a razor blade and freeze-dried, and camalexin was extracted from each individual lesion in 200 μl of 30% methanol with a steel ball using a Tissue Lyser (Qiagen) at 30 Hz for 1 min and left to soak for 15 min. The extraction procedure was performed three times on each sample and supernatants from the successive extractions were collected and pooled for analysis.

Analysis of samples was by high-pressure liquid chromatography (HPLC) using an Agilent 1200 HPLC system (Agilent Technologies, Cheshire, UK). The compounds were separated on a Phenomenex C-18 column (100 mm × 2 mm; 3 μm) (Phenomenex, Torrance, California, USA) using an isocratic solvent system of H2O : CH3CN (7 : 3) with a flow rate of 0.25 ml min−1 at a temperature of 35°C. Typically, 20-μl injections were used. Camalexin was detected with a Shimadzu RF535 Fluorescence detector (Shimadzu, Milton Keynes, UK) set to camalexin’s characteristic fluorescent spectra (318 nm excitation and 385 nm emission). Identification and quantification were with reference to an authenticated standard from B. A. Halkier (University of Copenhagen, Copenhagen, Denmark); camalexin eluted with a retention time of 13.5 min. The data were acquired and integrated using Agilent’s Chemstation software.

Hormone and coronatine extraction and analysis

Leaves for hormone extraction were inoculated with PstDC3000 or mock-inoculated with 10 mM MgCl2. For each biological sample, three leaves from three plants were pooled. Harvested leaves were snap-frozen in liquid nitrogen, subsequently freeze-dried and milled with a steel ball at 25 Hz for 3 min using a tissue lyser. Hormones were extracted from 10 mg of freeze-dried material and analysed by LC-MS/MS using an Agilent 1100 LC coupled to an Applied Biosystems Q-TRAP LC/MS/MS system (Applied Biosystems, California, USA) fitted with a Turbo Ion Spray source operating in negative mode as described by Forcat et al. (2008). For coronatine, a calibration curve was constructed from data obtained by injection of known quantities of coronatine (purchased from S. Abrams, NRC, Ontario, Canada) and the amount was quantified by monitoring the coronatine mass transition pair 318 > 163.

Statistical analysis

To test for statistical significance, data were analysed with Student’s t-test (for comparison of wild-type and ahk5-1 mutant responses), one-way ANOVA with Tukey’s post hoc test (for comparison of the responses of the wild type, the ahk5-1 mutant and the complemented lines) or Kruskal–Wallis one-way ANOVA with Dunn’s multiple comparison test for nonparametric data (for comparison of the responses of the wild type, the ahk5-1 mutant and the complemented lines).


AHK5 contributes to salt sensitivity

To determine whether AHK5 is involved in modulating reactions to abiotic stress, phenotypic responses of wild-type and mutant seedlings to heat, cold, salt and osmotic stress were tested. As AHK5 was shown to be most highly expressed in the roots (Iwama et al., 2007; Desikan et al., 2008), we focused on the root growth of seedlings under diverse abiotic stresses. Under the conditions tested, a markedly increased inhibition of seedling root growth was observed in response to salinity (Fig. 1a). This inhibitory effect of salinity was greater in wild-type plants, with the root length of the ahk5-1 mutant being significantly greater at all concentrations of NaCl tested, suggesting that AHK5 positively regulates salt-induced root growth inhibition.

Figure 1.

The AHK5 mutant is more resistant to salt stress. (a) Root length of wild-type (open bars) and ahk5-1 mutant (closed bars) seedlings after 7 d of germination and growth on half-strength Murashige and Skoog medium supplemented with 0–150 mM NaCl. Measurements are the mean ± SEM for c. 140 seedlings for each line from three independent experiments. Asterisks denote statistically significant differences between wild-type and mutant responses as determined by Student’s t-test. Control root lengths were as follows: Columbia (Col-0), 14.83 ± 0.78 mm; ahk5-1, 16.19 ± 0.67 mm. (b) Root length of wild-type, ahk5-1 mutant and complemented lines PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4 after germination and 7 d of growth on half-strength Murashige and Skoog medium supplemented with 100 mM NaCl. Root length of seedlings grown in the presence of 100 mM NaCl is presented as a percentage of control treatments (no NaCl) from an average of at least 130 seedlings for each line from three independent experiments. Asterisks denote statistically significant differences between mutant responses compared with the wild type and the complemented lines as determined by one-way ANOVA with Tukey’s post hoc test. Control root lengths were as follows: Col-0, 17.54 ± 0.48 mm; ahk5-1, 16.19 ± 0.67 mm; PAHK5-AHK5/ahk5-1-1, 18.93 ± 0.28 mm; PAHK5-AHK5/ahk5-1-4, 16.42 ± 0.27 mm. (c) Fresh weight of mature soil-grown plants watered with a 250 mM solution of NaCl over the course of 2 wk. Data are the mean ± SEM for the fresh weights of at least six plants for each line tested. Fresh weights of untreated control plants were as follows: Col-0, 3.62 ± 0.20 g; ahk5-1, 2.24 ± 0.13 g; PAHK5-AHK5/ahk5-1-1, 2.94 ± 0.13 g; PAHK5-AHK5/ahk5-1-4, 2.88 ± 0.13 g. Asterisks denote statistically significant differences between wild-type and mutant responses as determined by one-way ANOVA with Tukey’s post hoc test. *, < 0.05; **, < 0.005; ***, < 0.0001.

The root growth of seedlings of two independent complemented lines, PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4 (see Fig. S1 for genotyping of these lines), on media containing 100 mM NaCl was similar to that of wild-type Columbia (Col-0) seedlings, confirming that the insensitivity of the ahk5-1 mutant in response to salinity was caused by loss of AHK5 function (Fig. 1b). At a higher concentration of NaCl (200 mM), survival of the ahk5-1 mutant was similar to that of the wild type and the complemented lines, suggesting that AHK5 is involved in tolerance to, rather than survival under, saline conditions (Fig. S2). Seeds of ahk5-1 also showed a significantly higher percentage of germination on 150 mM NaCl, with 90% of mutant seeds germinating compared with 72% of Col-0 seeds (< 0.05 using Student’s t-test). In the absence of NaCl, all seeds sown for both Col-0 and ahk5-1 germinated, suggesting that the difference in germination of wild-type and ahk5-1 seeds in the presence of NaCl was caused by differential sensitivity of the seeds to salinity.

The investigation of salt sensitivity was subsequently extended to mature soil-grown plants, for which plants were irrigated with 250 mM NaCl for a period of 2 wk, after which the fresh weights of the plants were measured. The relative shoot fresh weight (as a percentage of the shoot fresh weight of the respective control-treated plants) was determined and found to be significantly greater in ahk5-1 mutant plants compared with wild-type plants and the complemented lines PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4 (Fig. 1c), demonstrating that AHK5 positively contributes to salt sensitivity in both seedlings and mature plants.

Salt is known to cause ionic toxicity as well as osmotic stress; build-up of ions in the cytoplasm inhibits enzymatic activity, whereas build-up in the cell wall causes dehydration and osmotic stress (Munns, 2002; Munns & Tester, 2008). No difference was seen between the responses of Col-0 and ahk5-1 seedlings to osmotic stress with either sorbitol or KCl as the osmoticum (Fig. S3), suggesting that the differences seen in root length and germination may be attributable to the ionic component (i.e. Na+) of salt stress.

AHK5 is involved in defence against Pseudomonas syringae pv DC3000

Symptom development and bacterial growth  It was previously shown in our laboratory that the ahk5-1 mutant is defective in stomatal responses to PstDC3000 and to bacterial flagellin (Desikan et al., 2008). Here we set out to determine whether the defect in stomatal responses to bacterial pathogens in the ahk5-1 mutant correlated with an increased susceptibility to infection.

When PstDC3000 was inoculated onto the leaf surface, disease progression in the ahk5-1 mutant was accelerated compared with wild-type plants, as shown by increased chlorosis of leaves (Fig. S4). No obvious senescence phenotype was seen in ahk5-1 mutant plants in the absence of bacterial inoculation (Fig. S4). Quantification of the extent of chlorosis via chlorophyll measurements revealed a significantly lower amount of chlorophyll in the ahk5-1 mutant at 6 d post-inoculation (dpi) compared with wild-type Col-0 and the complemented lines PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4 (Fig. 2a). In planta bacterial populations also showed a clear increase in PstDC3000 growth in the ahk5-1 mutant compared with wild-type plants and the complemented lines (Fig. 2b). The increased susceptibility of ahk5-1 plants indicated by higher bacterial numbers was most significant at 6 dpi, correlating with the decrease in chlorophyll content observed at this time-point. The delayed increase in susceptibility indicates that the effects of the ahk5-1 mutation extended well beyond the early stages of stomatal penetration.

Figure 2.

The AHK5 mutant is more susceptible to infection with Pseudomonas syringae pv. tomato DC3000 (PstDC3000). Leaves were surface-inoculated with 4 × 108 cfu ml−1PstDC3000. (a) Chlorophyll content of infected leaves expressed as a percentage of that of mock-inoculated (10 mM MgCl2) leaves. Values are mean ± SEM of measurements from six plants per genotype. Asterisks denote a statistically significant difference between the ahk5-1 mutant response and the responses of wild-type Columbia (Col-0) and complemented PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4, as determined by one-way ANOVA with Tukey’s post hoc test. (b) In planta growth of PstDC3000. Values are mean ± SEM obtained from six plants per genotype. Asterisks denote a statistically significant difference between the ahk5-1 mutant response and the responses of wild-type Col-0 and complemented PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4 plants, as determined by one-way ANOVA with Tukey’s post hoc test. *, < 0.05; ***, < 0.0001.

Hormones and infection  Resistance to PstDC3000 involves changes in concentrations of the phytohormones salicylic acid (SA), abscisic acid (ABA) and jasmonic acid (JA) (de Torres Zabala et al., 2009; Verhage et al., 2010). SA is required for basal resistance against biotrophic pathogens such as PstDC3000, whereas manipulation of ABA can promote virulence of the bacterium through inhibition of SA biosynthesis and responses (Block et al., 2005; de Torres Zabala et al., 2007, 2009). However, an intact ABA signalling pathway is required for stomatal closure in response to bacteria on the leaf surface, and the ability of PstDC3000 to manipulate ABA concentrations and to overcome stomatal defences requires the presence of the bacterial phytotoxin coronatine (Brooks et al., 2004; Melotto et al., 2006; de Torres Zabala et al., 2007, 2009).

In order to determine whether changes in hormone concentrations may account for the altered ahk5-1 disease phenotype, hormone concentrations following surface inoculation of leaves were measured. No significant differences were observed in the concentrations of hormones in mock-inoculated leaves of wild-type and ahk5-1 plants (SA: 1.14 μg g−1 DW in Col-0 and 1.53 μg g−1 DW in ahk5-1; ABA: 62.20 ng g−1 DW in Col-0 and 61.95 ng g−1 DW in ahk5-1; JA: 1.07 μg g−1 DW in Col-0 and 0.88 μg g−1 DW in ahk5-1). Following inoculation, similar patterns of increase in hormone concentrations were observed in both genotypes, but concentrations were significantly lower in ahk5-1 leaves (Fig. 3). Differences were most striking for SA and JA after 2 and 6 dpi, respectively, but were consistently higher in wild-type plants during the course of infection (Fig. 3a,c). ABA concentrations were also lower in the mutant ahk5-1 plants compared with wild-type plants, and this difference was statistically significant at 6 dpi (Fig. 3b). Although no overlap in error bars was seen for ABA concentrations at 2 dpi in wild-type Col-0 and the ahk5-1 mutant, non-overlap of error bars does not necessarily indicate a significant difference (Cumming et al., 2007). In agreement with the observations of de Torres Zabala et al. (2009), JA was seen to increase later than ABA and SA during infection (Fig. 3c). On the other side of the interaction, coronatine accumulation was observed by 2 dpi (Fig. 3d); however, the concentrations of the phytotoxin were fourfold lower at 2 dpi and twofold lower at 4 and 6 dpi in ahk5-1 compared with wild-type plants (Fig. 3d).

Figure 3.

Concentrations of (a) salicylic acid (SA), (b) abscisic acid (ABA), (c) jasmonic acid (JA) and (d) coronatine (COR) in leaf tissue surface inoculated with 4 × 108 cfu ml−1Pseudomonas syringae pv. tomato DC3000. Measurements are an average of two independent experiments. Asterisks denote statistically significant differences between wild-type (open bars) and mutant (closed bars) plant responses as determined by Student’s t-test. *, < 0.05; **, < 0.005; ***, < 0.0005.

AHK5 is involved in defence against the necrotroph Botrytis cinerea

Symptom development and camalexin production  Infection with PstDC3000 has been shown to increase susceptibility to fungal pathogens (Spoel et al., 2007), demonstrating a link between resistance pathways operating against bacterial and fungal pathogens. Here, the response of the ahk5-1 mutant was tested to infection with the necrotroph B. cinerea, which is able to directly penetrate the host plant cuticle (Williamson et al., 2007).

Symptom development/severity of disease was assessed by monitoring lesion spread and the amount of tissue collapse. This revealed that the ahk5-1 mutant was markedly more susceptible to infection by the necrotrophic pathogen; significantly higher lesion scores were observed at 3 and 5 dpi in the ahk5-1 mutant compared with leaves of the wild type and the complemented lines PAHK5-AHK5/ahk5-1-1 and PAHK5-AHK5/ahk5-1-4, with higher lesion scores representing more severe disease symptoms (Fig. 4a). Measurement of lesion size at 3 dpi confirmed that lesion size was significantly greater in the ahk5-1 mutant, approximately double the size of lesions seen in leaves of wild-type plants and the complemented lines (Fig. 4b).

Figure 4.

The AHK5 mutant is more susceptible to infection with Botrytis cinerea. Leaves were inoculated with 2 × 105 spores ml−1. (a) Lesion scores (mean ± SEM of 37–108 lesions) observed at 3 and 5 d post-inoculation (dpi); experiments were repeated three times. Asterisks denote statistically significant differences as determined by the Kruskal–Wallis test. (b) Lesion area as measured at 3 dpi (mean ± SEM of 37–108 lesions). Asterisks denote statistically significant differences as determined by one-way ANOVA with Tukey’s post hoc test. (c) Mean concentrations of camalexin extracted from a total of 20 individual lesions caused by B. cinerea, on wild type (open bars) and ahk5-1 mutant (closed bars) plants per time-point. *, < 0.05; **, P < 0.005; ***, < 0.0001.

The phytoalexin camalexin has been shown to contribute to resistance to a number of necrotrophic fungi, including B. cinerea (Ferrari et al., 2003; Denby et al., 2004; Lazniewska et al., 2010). Concentrations of camalexin in individual lesions from wild-type and mutant plants infected with B. cinerea were measured at 1, 2, 3 and 4 dpi. Camalexin concentrations were seen to increase as lesion development progressed but were not significantly different between wild-type and mutant leaves (Fig. 4c). This shows that the increased susceptibility of the ahk5-1 mutant is not attributable to a defect in camalexin production.

Differences in reactive oxygen species production and fungal growth in planta ROS production during infection with B. cinerea is known to occur in both plant cells and fungal structures, and has been shown to enhance fungal growth and symptom development (Govrin & Levine, 2000). To determine whether ROS production was altered in the mutant plant or in the fungus during the infection process, ROS accumulation was monitored using DAB. At 8, 12 and 24 h post-inoculation (hpi), an apparent increase in DAB staining at the site of inoculation was seen, with more DAB staining recorded in the ahk5-1 mutant earlier than in wild-type plants (Fig. 5a). On closer examination, it appeared that the DAB staining seen at 8 and 12 hpi was associated with fungal structures and not with the leaf cells (Fig. S5). However, it was unclear whether the increased amount of DAB associated with the fungus on the ahk5-1 mutant was caused by increased production of fungal ROS or an increase in the actual mass of fungus present. To determine whether the latter was the case, fungal hyphae were stained with trypan blue and hyphal length was measured. A significant difference in the length of fungal hyphae was seen as early as 8 hpi, with hyphal length an average of 45% greater on the ahk5-1 mutant leaves than that measured on wild-type plants (Figs 5b, S6).

Figure 5.

Botrytis cinerea growth is greater on leaves of AHK5 mutant than wild-type leaves. Leaves were inoculated with 2 × 105 spores ml−1 in 1/8th potato dextrose broth (PDB). (a) 3,3-Diaminobenzidine (DAB) staining of plant cells and fungal hyphae on wild-type and ahk5-1 leaves. Representative images are shown for each time-point. Arrows indicate the position of the inoculation droplet. Bar, 3 mm. (b) Length of fungal hyphae on wild-type and ahk5-1 leaves measured at 8 h post-inoculation using trypan blue staining. Values are mean ± SEM of 237–271 measurements. Asterisks denote a statistically significant difference between Col-0 and ahk5-1 as determined by Student’s t-test. (c) Hyphal length of B. cinerea spores germinated in vitro in 1/8th PDB supplemented with water (control, open bars) or 1 μM diphenyleneiodinium chloride (DPI; closed bars) for various durations. Asterisks denote a significant difference between DPI-treated and control samples as determined by Student’s t-test. Values are mean ± SEM of at least 120 measurements for each treatment and time-point. ***, < 0.0001.

To further investigate the role of ROS in fungal growth, the effect of the NADPH oxidase inhibitor DPI on hyphal growth of B. cinerea was also investigated. Botrytis cinerea contains two NADPH oxidases, BcnoxA and BcnoxB, which have been shown to contribute to the virulence and pathogenicity of the fungus (Segmuller et al., 2008). The hyphal length of spores germinated in vitro in the presence of 1 μM DPI was measured at 4, 6 and 8 hpi. At all three time-points, hyphal length was significantly shorter in spores germinated in the presence of DPI compared with control spores germinated without DPI (Fig. 5c), suggesting that fungal-derived ROS produced by NADPH oxidase(s) contributes to hyphal growth.

Whereas DAB staining in fungal hyphae on leaves was observed at all experimental time-points (Fig. 5a), plant-derived ROS in response to infection with B. cinerea was not seen until 24 hpi and was noticeably less extensive in the ahk5-1 mutant compared with wild-type leaves (Fig. 6a). Analysis of the proportion of DAB staining at the site of inoculation revealed that ahk5-1 displayed 28% less staining than wild type, suggesting that the ahk5-1 mutant is attenuated in ROS production in response to infection with B. cinerea at 24 hpi (Fig. 6b). At later time-points (up to 3 dpi), the extent of DAB staining corresponded to lesion size, that is, in the mutant there appeared to be more ROS at the lesion sites as a result of the larger lesions formed. Interestingly, pretreatment of wild-type Col-0 leaves with DPI 2 h before inoculation with B. cinerea resulted in measurably smaller lesions at 3 dpi compared with mock-treated, inoculated leaves (Fig. S7).

Figure 6.

The AHK5 mutant produces less reactive oxygen species (ROS) in response to infection with Botrytis cinerea. Leaves were inoculated with a spore suspension at a concentration of 2 × 105 spores ml−1 in 1/8th potato dextrose broth. (a) Wild-type Columbia (Col-0) and ahk5-1 mutant leaves stained with 3,3-diaminobenzidine (DAB) at 24 h post-inoculation. Representative pictures at the site of inoculation are shown. Bar, 3 mm. (b) Quantification of the amount of DAB staining measured at the site of inoculation 24 h post-inoculation. Values are an average obtained from Image J analysis of 20 lesions per plant. The asterisk denotes a statistically significant difference as determined by Student’s t-test. *, < 0.05.


In response to the changing environment, plants have the ability to sense and respond to single or multiple stimuli via activation of unique signalling cascades. The two-component systems in bacteria and yeast act as unique sensing and signalling systems responding to osmotic (Forst & Roberts, 1994; Posas et al., 1996; Paithoonrangsarid et al., 2004), temperature (Aguilar et al., 2001; Suzuki et al., 2005) and oxidative stimuli (Singh, 2000; Kanesaki et al., 2007). More recently, some plant HKs have been shown to function in response to environmental stimuli such as drought/osmotic (AtHK1, AHK2, AHK3 and AHK4) (Urao et al., 1999; Tran et al., 2007, 2010) and salinity stress (AHK2, AHK3 and ETR1) (Zhao & Schaller, 2004; Tran et al., 2007; Wang et al., 2008; Tran et al., 2010). Additionally, a role for AHK2 and AHK3 in modulating disease resistance in response to infection with PstDC3000 has been reported (Choi et al., 2010). Previously, our laboratory has shown that AHK5 mediates stomatal responses to endogenous and environmental cues (Desikan et al., 2008). Here we demonstrate the ability of AHK5 to function in response to both abiotic and biotic stimuli to affect the growth and survival of Arabidopsis.

AHK5 regulates tolerance to salinity

The present study shows that, of the various abiotic stresses tested, AHK5 positively regulates growth inhibition to salinity stress. This is the first demonstration of a function for AHK5 in salinity-induced growth responses in Arabidopsis. ABA is known to play a role in drought and salinity stress tolerance (Zhu, 2002). However, the finding that AHK5 negatively regulates ABA-induced root growth inhibition (Iwama et al., 2007) is hard to reconcile with the finding that it positively regulates salt-induced root growth inhibition. It is possible that ABA concentrations induced following salt stress are insufficient to enable AHK5-mediated signal transduction to act in an inhibitory manner. This possibility is supported by the promotion of root growth seen with low concentrations of ABA in both wild-type and ahk5-1 mutant plants (Ghassemian et al. (2000) and Fig. S8). In addition, there is evidence suggesting that ABA may function to maintain rather than to inhibit root growth under conditions of stress (Sharp & LeNoble, 2002).

Salinity stress imposes both osmotic and ionic stress on the plant (Munns, 2002; Munns & Tester, 2008). Here, we have obtained evidence to suggest that the ahk5-1 mutant is less sensitive to the ionic component of salt stress. Under salt stress, the increased uptake of Na+ disrupts homeostasis of ions and inhibition of cellular processes, through toxicity to cellular enzymes and disruption of K+ uptake by the plant (Munns, 2002; Zhu, 2002). Tolerance to salt stress in Arabidopsis is mediated in part through efflux of Na+ by the Na+/H+ antiporter SOS1 (SALT OVERLY SENSITIVE1) at the plasma membrane, which is activated by the myristolated calcium-binding protein SOS3 and the Ser/Thr kinase SOS2 as part of the SOS pathway (Zhu, 2002, 2003). The phenotype we observed in the ahk5-1 mutant appeared to be specific to the presence of Na+ (Figs 1a, S3), raising the possibility that AHK5 may interact with the SOS pathway. Whether AHK5 directly regulates uptake of Na+/K+ is not yet known. However, a link with the SOS pathway via ROS is plausible.

ROS are known to accumulate to varying concentrations in response to salinity treatment in different subcellular compartments (Miller et al., 2010) and the SOS1 protein has been shown to interact with RCD1 (RADICAL-INDUCED CELL DEATH1), a regulator of oxidative stress responses, to regulate ROS concentrations (Katiyar-Agarwal et al., 2006). Preliminary observations indicate that ahk5-1 seedlings also accumulate higher concentrations of H2O2 following salinity stress than treated wild-type Col-0 seedlings (2.3-fold and 1.4-fold increases over untreated controls, respectively). Given the previously identified function for AHK5 in regulating H2O2 concentrations (Desikan et al., 2008), it is possible that AHK5 acts to regulate redox balance following salinity stress challenge to mediate growth responses.

Loss of AHK5 function increases susceptibility to the bacterial pathogen PstDC3000

Stomata form an active part of plant defences against bacterial pathogens (Melotto et al., 2006; Underwood et al., 2007; Zeng et al., 2010) and previous work has shown that stomata of the ahk5-1 mutant are unresponsive to the presence of PstDC3000 on the leaf surface (Desikan et al., 2008). Here, we report that the ahk5-1 mutant is more susceptible to bacterial infection, when surface-inoculated with PstDC3000. These observations were similar to those reported by Zipfel et al. (2004) with the fls2 (flagellin sensitive2) mutant, which is defective in the perception of bacterial flagellin and does not close stomata in response to either flg22 or PstDC3000 at the leaf surface (Melotto et al., 2006; Zeng & He, 2010). In addition to PstDC3000, the ahk5-1 mutant is also defective in stomatal closure in response to flg22, ethylene, H2O2 and darkness (Desikan et al., 2008). AHK5 may therefore act to integrate responses at the leaf surface to diverse exogenous stimuli.

Further analysis of metabolic changes following PstDC3000 challenge revealed quantitative differences between wild-type and ahk5-1 mutant plants in the hormones SA, ABA and JA. To our knowledge, this is the first report of changes in plant hormones following surface inoculation of bacteria onto Arabidopsis leaves, with the timing of hormone changes correlating with the spread of symptoms. Concentrations of all hormones were maintained at low levels following infection of ahk5-1 leaves. It is clear from recent studies that the balance between the concentrations of the different hormones influences the outcome of disease (Block et al., 2005; de Torres Zabala et al., 2009). During the Arabidopsis–PstDC3000 interaction, antagonism of SA defences through manipulation of ABA production promotes virulence of the bacterium (de Torres Zabala et al., 2007, 2009). In this study, AHK5 differentially affected the temporal accumulation of SA, JA and ABA in response to PstDC3000. In the absence of a stomatal closure response to bacteria, as in ahk5-1, PstDC3000 would enter at an accelerated rate into the apoplastic space and multiply to higher numbers than that seen in wild-type plants. This is likely to suppress basal defence more efficiently, and have prolonged effects on hormonal imbalance, as seen in ahk5-1, throughout the course of the infection.

Another obvious cause for increased virulence in the mutant could be increased production of the bacterial phytotoxin coronatine, also known to suppress basal defences and to manipulate hormone signalling to promote disease (Brooks et al., 2005; Uppalapati et al., 2007; Ishiga et al., 2010; Zeng et al., 2010). Surprisingly, however, PstDC3000 produced lower concentrations of coronatine in ahk5-1 plants. This might result from host defences (e.g., hormones) already being suppressed to a large extent, thereby not necessitating the production of coronatine by PstDC3000. In support of this, Block et al. (2005) suggest that coronatine concentrations do not necessarily correlate with the growth of bacteria. Rather, PstDC3000 might be reallocating its resources to other virulence mechanisms, such as expression of the type-three secretion system and effector proteins, and expression of genes involved in nutrient assimilation or adaptation to the apoplastic environment (Boch et al., 2002; Rico & Preston, 2008).

An alternative explanation for the reduced synthesis of coronatine in ahk5-1 plants is that AHK5 indirectly regulates coronatine biosynthesis. There have been some early reports in the literature of plant-derived factors from the shikimate pathway regulating coronatine biosynthesis in PstDC3000 (Li et al., 1998). In this study, we observed a metabolite unique to ahk5-1 mutant leaves which appeared to be independent of bacterial infection (Fig. S9). Preliminary mass spectrometry analysis is indicative of this metabolite belonging to the indolic class of compounds (part of the shikimate pathway). Is it possible that AHK5 normally suppresses the synthesis of this unique metabolite, and lack of AHK5 removes this suppression, leading to altered susceptibility to PstDC3000? Further work to identify this compound will address this question.

AHK5 contributes to resistance to the necrotrophic fungal pathogen B. cinerea

In addition to increased susceptibility to the hemibiotroph PstDC3000, the ahk5-1 mutant was also strikingly susceptible to the necrotrophic fungal pathogen B. cinerea. Although differences in lesion formation between wild-type and mutant plants were not apparent until 3 dpi in response to B. cinerea, close inspection of infected leaves revealed an increase in fungal growth in ahk5-1 as early as 8 h after contact of the fungal spores with the leaf surface. The cause of this early difference in fungal growth could be due to differences in the properties of the leaf surface structure and composition and/or compounds exuded onto the leaf surface affecting fungal growth (Rossall et al., 1977; Doss et al., 1993; Calo et al., 2006; Bessire et al., 2007; Chassot et al., 2008; Curvers et al., 2010). DAB staining of leaves at early time-points showed more fungal structures being stained, and from the images obtained it appeared that individual hyphae on the mutant showed more DAB staining (Fig. S5). Although there is an increase in fungal hyphal length in the mutant, the experiments here cannot reveal clearly whether or not the mutant leaves somehow enhance fungal production of ROS, thereby enhancing fungal growth. As our data also reveal a requirement for ROS in in vitro hyphal growth, this is a strong possibility.

In addition to the early increase in fungal growth observed on the ahk5-1 mutant, in planta ROS production in ahk5-1 leaf cells was significantly lower than that in wild-type tissue at 24 hpi at the site of inoculation. Early ROS production in tomato (Solanum lycopersicum) and bean (Phaseolus vulgaris) has previously been linked to resistance to infection with B. cinerea (Unger et al., 2005; Asselbergh et al., 2007), and the early attenuation in ROS production by the ahk5-1 mutant may be linked to a dampened immune response resulting in more severe disease symptoms seen in this mutant. However, plant-derived ROS is also required for virulence of the fungus, as ROS-triggered cell death, although effective against biotrophic pathogens in the form of the hypersensitive response (HR), enhances infection by Botrytis cinerea (Govrin & Levine, 2000). This is in agreement with the DPI inhibition of lesions seen here at 3 dpi (Fig. 5a). Moreover, after 24 hpi, the presence of ROS correlated to the area covered by the lesions appearing later (i.e. there appeared to be more extensive ROS production in ahk5-1 leaves as a consequence of larger lesions). Clearly, the role of ROS in the outcome of necrotroph–plant interactions is complex and is determined by factors such as the timing of induction, and the cellular location in the host and the pathogen, as well as the source and concentration of ROS generated. In response to infection with B. cinerea, whereas early ROS production regulated by AHK5 contributes to triggering defence mechanisms against B. cinerea infection, at the later stages of infection, loss of AHK5 function results in increased ROS production associated with cell death, which further facilitates disease progression.

Aside from ROS, nitric oxide (NO) is also known to contribute to resistance to B. cinerea (Asai & Yoshioka, 2009). Additionally, a recent study found that NO produced in B. cinerea was able to diffuse into the surrounding growth medium, raising the possibility that fungal-derived NO may influence plant signalling (Turrion-Gomez & Benito, 2011). However, we found that pretreatment of leaves with the NO scavenger cPTIO did not significantly affect the size of lesions caused by B. cinerea in wild-type Col-0 (Fig. S7). Similarly, treatment of spores with cPTIO did not affect hyphal elongation in vitro (Fig. S10), suggesting that neither a defect in NO production in the ahk5-1 mutant nor alteration in NO production in B. cinerea contributes to the increased susceptibility of the ahk5-1 mutant.

Another factor that contributes to resistance to B. cinerea is the phytoalexin camalexin, which is known to accumulate at the site of infection (Denby et al., 2004; Kliebenstein et al., 2005). In this study, the increased susceptibility of the ahk5-1 mutant was not attributed to a defect in camalexin production, indicating that factors other than camalexin determine host susceptibility or that a mutation in AHK5 results in the secretion of other compounds that influence the sensitivity of B. cinerea to camalexin.

In response to necrotrophic pathogens such as B. cinerea, the plant hormones ethylene (ET), SA and JA are known to positively contribute to resistance (Thomma et al., 1998, 1999, 2000; Zimmerli et al., 2001; Ferrari et al., 2003; van Baarlen et al., 2007; Vicedo et al., 2009). Although the endogenous concentrations of the plant hormones measured were similar in ahk5-1 and wild-type leaf tissue, it would be interesting to see how concentrations of these hormones change over the course of infection with B. cinerea in the ahk5-1 mutant.

Interestingly, pretreatment of leaves with flg22 has been shown to increase resistance to infection with B. cinerea (Ferrari et al., 2007; Galletti et al., 2008), and the kinase Botrytis-induced kinase 1 (BIK1) which is involved in resistance to B. cinerea infection (Veronese et al., 2006) has also been shown to interact with the flagellin receptor FLS2 and BAK1 (BRI1 ASSOCIATED RECEPTOR KINASE1) to mediate flagellin responses (Lu et al., 2010). Given the link with AHK5 and flagellin signalling (Desikan et al., 2008), it is possible that AHK5 acts to integrate basal defence signalling pathways activated by diverse pathogens.

In conclusion, we have shown here that the hybrid histidine kinase AHK5 functions in salinity tolerance and resistance to PstDC3000 and B. cinerea. An underlying theme linking AHK5 to the signalling pathways of the responses tested here is phytohormones and ROS. It is possible that, depending on the stimulus, AHK5 functions differentially to perturb hormone accumulation/signalling and redox balance, resulting in the ahk5-1 mutant being more tolerant to salinity stress, yet less resistant to infection with bacterial and fungal pathogens. In previous studies, AHK2 and AHK3 were found to positively contribute to resistance to PstDC3000 but to negatively regulate cold, salinity and drought stress (Tran et al., 2007, 2010; Choi et al., 2010; Jeon et al., 2010). The opposing functions of both AHK2 and AHK3 in biotic and abiotic stress responses are similar to that found for AHK5 in this study, whereby AHK5 is required for full resistance to the plant pathogens PstDC3000 and B. cinerea but loss of function increases resistance to salinity stress. This study provides further evidence for the importance of hybrid HK function in regulation of growth and survival responses to both abiotic and biotic stresses. Further studies to identify AHK5 interactors will be important for determining the mechanism of AHK5 function. Other key genes which appear to integrate abiotic and biotic stress responses via redox changes and hormones include the DELLA proteins (Achard et al., 2008), UPS1 (UNDERINDUCER AFTER PATHOGEN AND STRESS1) (Denby et al., 2005), BOS1 (BOTRYTIS SUSCEPTIBLE1) (Mengiste et al., 2003) and ATAF1 (ARABIDOPSIS NAC DOMAIN CONTAINING PROTEIN2) (Wu et al., 2009). Thus, single genes controlling multiple stress responses do exist and are undoubtedly important for the phenomenon of cross-tolerance and acclimation to stress. Our novel findings highlight the importance of identifying key nodes in the signalling pathways that mediate multiple stress responses in plants.


This research was funded by a Biotechnology and Biological Sciences Research Council (BBSRC) DTG to J.P. We thank Emmy McGarry (BSPP summer student) for preliminary work phenotyping plant responses to B. cinerea. We also thank Yvonne Stewart for her comments on the manuscript.