ABSTRACT
The hypersensitive response and pathogenicity (hrp) type III secretion system (T3SS) is a key pathogenicity factor in Erwinia amylovora. Previous studies have demonstrated that the T3SS in E. amylovora is transcriptionally regulated by a sigma factor cascade. In this study, the role of the bacterial alarmone ppGpp in activating the T3SS and virulence of E. amylovora was investigated using ppGpp mutants generated by Red recombinase cloning. The virulence of a ppGpp-deficient mutant (ppGpp0) as well as a dksA mutant of E. amylovora was completely impaired, and bacterial growth was significantly reduced, suggesting that ppGpp is required for full virulence of E. amylovora. Expression of T3SS genes was greatly downregulated in the ppGpp0 and dksA mutants. Western blotting showed that accumulations of the HrpA protein in the ppGpp0 and dksA mutants were about 10 and 4%, respectively, of that in the wild-type strain. Furthermore, higher levels of ppGpp resulted in a reduced cell size of E. amylovora. Moreover, serine hydroxamate and α-methylglucoside, which induce amino acid and carbon starvation, respectively, activated hrpA and hrpL promoter activities in hrp-inducing minimal medium. These results demonstrated that ppGpp and DksA play central roles in E. amylovora virulence and indicated that E. amylovora utilizes ppGpp as an internal messenger to sense environmental/nutritional stimuli for regulation of the T3SS and virulence.
IMPORTANCE The type III secretion system (T3SS) is a key pathogenicity factor in Gram-negative bacteria. Fully elucidating how the T3SS is activated is crucial for comprehensively understanding the function of the T3SS, bacterial pathogenesis, and survival under stress conditions. In this study, we present the first evidence that the bacterial alarmone ppGpp-mediated stringent response activates the T3SS through a sigma factor cascade, indicating that ppGpp acts as an internal messenger to sense environmental/nutritional stimuli for the regulation of the T3SS and virulence in plant-pathogenic bacteria. Furthermore, the recovery of an spoT null mutant, which displayed very unique phenotypes, suggested that small proteins containing a single ppGpp hydrolase domain are functional.
INTRODUCTION
Erwinia amylovora causes a devastating fire blight disease of apples and pears, which results in severe economic losses to growers around the world (1, 2). E. amylovora is closely related to members of the Enterobacteriaceae family, including many important human pathogens, such as Escherichia coli, Yersinia pestis, and Salmonella enterica (3). Studies have revealed that the hypersensitive response and pathogenicity (hrp) type III secretion system (T3SS) is a major pathogenicity factor in E. amylovora (4–7). The hrp T3SS genes are carried on a pathogenicity island (8), and the alternative sigma factor HrpL, a member of the ECF subfamily of sigma factors, serves as the master regulator to control the expression of the structural and effector genes by binding to a consensus sequence known as the hrp box (9–13). In turn, expression of hrpL is positively regulated by the sigma 54 (σ54) protein RpoN, its modulation protein YhbH, and HrpS, a member of the NtrC family of σ54 enhancer binding proteins (EBPs) (4, 9, 11). However, the molecular mechanism that triggers the T3SS or activates the sigma factor cascade in E. amylovora remains unknown.
Upon initiating plant infection, plant-pathogenic bacteria undergo tremendous stresses, especially limiting nutrient stress and oxidative stress. T3SS genes are believed to be expressed rapidly under conditions such as limited nutrition (minimal medium), low pH, and relatively low temperature and are induced in planta or by iron but repressed in rich media (14–16). These observations suggest that nutrient limitation and/or oxidative stress may be one of the primary factors that activate the sigma factor cascade and trigger the expression of the T3SS. However, the exact environmental/host signal(s) remains elusive.
As one of the global regulatory systems in bacteria, the stringent response often results in swift and massive transcriptional reprogramming in response to various nutrient limitation conditions (17, 18). During the stringent response, bacterial cells accumulate high levels of the linear nucleotide second messengers, i.e., guanosine tetraphosphate (ppGpp) and guanosine pentaphosphate (pppGpp) [collectively known as (p)ppGpp; referred to here as ppGpp] (19). In bacteria, the RelA-SpoT homologue (RSH) proteins are responsible for (p)ppGpp biosynthesis and degradation in response to nutrient starvation, e.g., lack of amino acids, phosphates, fatty acids, carbon, or iron, similar to the conditions activating the T3SS (17, 20). In E. coli, RelA is a ribosome-associated monofunctional protein that synthesizes ppGpp by using ATP and GTP in response to amino acid starvation and is sensed by the presence of uncharged tRNA molecules in the A site of a ribosome (21, 22). In contrast, the cytoplasmic SpoT protein has both ppGpp synthase and hydrolase activities and is activated in response to a lack of fatty acids, carbon, phosphorus, or iron, as well as hyperosmotic shock and oxidative stress (22, 23).
In E. coli, the ppGpp-mediated stringent response redirects the global transcriptional capacity of a cell from genes responsible for growth and reproduction toward those responsible for survival (17). Interactions among RNA polymerase (RNAP), ppGpp, and its partner transcription factor DksA result in downregulation of highly expressed stable RNA, DNA replication, ribosome and protein synthesis, and simultaneous upregulation of stress and starvation genes as well as virulence genes (20, 24, 25). In general, ppGpp directly interacts with the interface of the β′ and ω subunits of RNAP to orchestrate fine-tuning of cellular processes through direct inhibition and activation of genes (26–28), whereas the DksA protein, which binds to the RNAP secondary channel, greatly enhances the effect of ppGpp by modulating the direct interaction between RNAP and ppGpp (18, 28). In addition, ppGpp inhibits RNAP binding to σ70-dependent stringent promoters, thus indirectly allowing RNAP to bind to alternative sigma factors, such as RpoN, and promoting the expression of alternative sigma factor-dependent genes (19, 29). Importantly, inhibition of σ70 binding by ppGpp is transient and reversible, thus enabling rapid and reversible control of stress response genes (17).
In S. enterica, accumulated ppGpp induces HilA, a master regulator of Salmonella pathogenicity island 1 (SPI1). Furthermore, ppGpp directly interacts with SlyA, a transcriptional activator of pathogenicity island 2 (SPI2), to facilitate the intracellular virulence program of S. enterica (30, 31). In E. coli, accumulation of ppGpp activates LEE gene expression and increases bacterial adherence (32). In plant-associated pseudomonads and rhizobia, ppGpp affects epiphytic fitness, biocontrol activity, biofilm formation, and hydrogen peroxide and antibiotic tolerance, as well as nodulation (33–36). In plant-pathogenic bacteria, ppGpp is required for cell wall-degrading enzyme production, quorum sensing signal degradation, and Ti plasmid transfer (37–39). However, it remains unknown whether ppGpp regulates the T3SS and virulence in plant-pathogenic bacteria.
The goal of this study was to determine whether ppGpp regulates T3SS gene expression and virulence in E. amylovora. Our results demonstrate that ppGpp and DksA play central roles in E. amylovora virulence and suggest that E. amylovora utilizes ppGpp as an internal messenger to sense environmental/nutritional signals for regulation of the T3SS.
MATERIALS AND METHODS
Bacterial strains and growth conditions.Bacterial strains and plasmids used in this study are listed in Table 1. LB broth was used for routine growth of E. amylovora and E. coli strains. An hrp-inducing minimum medium (HMM) [1g (NH4)2SO4, 0.246 g MgCl2 · 6H2O, 0.1 g NaCl, 8.708 g K2HPO4, 6.804 g KH2PO4] with 10 mM galactose as the carbon source was used to induce T3SS gene expression (4, 40). MBMA minimal medium [3 g KH2PO4, 7 g K2HPO4, 1 g (NH4)2SO4, 2 ml glycerol, 0.5 g citric acid, 0.03 g MgSO4] supplemented with 1% sorbitol (7, 41) was also used. When required, antibiotics were used at the following concentrations: 50 μg ml−1 kanamycin, 100 μg ml−1 ampicillin, and 10 μg ml−1 chloramphenicol. Primer sequences used for mutant construction, mutant confirmation, quantitative real-time PCR (qRT-PCR), and cloning are available upon request.
Bacterial strains and plasmids used in this study
Generation of single and double mutants by λ Red recombinase cloning.E. amylovora mutant strains were generated using the λ phage recombinase method as described previously (6, 7, 42). Briefly, overnight cultures of E. amylovora strains harboring pKD46 were inoculated into LB broth containing 0.1% arabinose and grown to exponential phase (optical density at 600 nm [OD600] = 0.8). Cells were harvested, made electrocompetent, and stored at −80°C. These cells were electroporated with recombination fragments containing a cat or kan gene with its own promoter flanked by a 50-nucleotide homology region from the target gene(s). Recombination fragments were obtained by PCR amplification from the pKD32 or pKD13 plasmid, respectively. To confirm relA, spoT, and dksA mutations, PCR amplifications from the internal cat or kan primers to the external region of the target genes were performed. The coding regions of the relA, spoT, and dksA genes were absent from the corresponding mutant strains, except for the first and last 50 nucleotides. Double mutants were generated by using single mutants as the background.
Construction of plasmids.For complementation of the mutant strains, the genomic regions containing the promoter and gene sequences of the relA, spoT, and dksA genes were PCR amplified, gel purified, cut with KpnI and SacI, and ligated into the pWSK29 plasmid digested with the same enzymes. Standard molecular procedures were performed (43). Plasmid verification was performed by sequencing at the UIUC Core Sequencing Facility. The resulting plasmids were designated pRelA, pSpoT, and pDksA and transformed by electroporation into the relA, spoT, and dksA single and double mutant strains. For Western blot assay, the hrpA gene with a six-His tag coding sequence at the C terminus was cloned into the pWSK29 plasmid by use of KpnI and EcoRI restriction sites. The resulting plasmid was verified by sequencing and designated pHrpA-6His.
ppGpp quantification.ppGpp measurements were performed as previously described (44, 45), with some modifications. Briefly, overnight cultures of wild-type (WT) and mutant strains were washed three times with HMM and inoculated into 50 ml of HMM to a final OD600 of 0.2. Cells were incubated for 2 h with shaking at 18°C. Cultures were harvested by centrifugation, and 3 ml of 100% methanol was added to the pellets and mixed by vortexing for 50 s. To remove cell debris, suspensions were centrifuged and supernatants were freeze-dried at −50°C overnight. After drying, samples were dissolved in double-distilled water (ddH2O) and mixed with a fluorescent chemosensor [pyrene (Py) plus bis(Zn2+-dipicolylamine) (DPA) (PyDPA)] to a final concentration of 10 μM, and fluorescence was measured in a microplate reader with excitation at 365 nm and emission at 470 nm. ppGpp concentrations were calculated by comparison to a standard curve created with purified ppGpp (TriLink Biotechnologies Inc., CA). This experiment was performed twice.
Epifluorescence microscopy.Cell sizes of E. amylovora strains in rich and minimal media were determined by epifluorescence microscopy. Overnight cultures of bacterial strains constitutively expressing green fluorescent protein (GFP) were harvested, washed, and transferred to LB broth or HMM. Following 4 h of incubation, 3 μl of bacterial suspension was mixed with 5 μl of Aqua-Polymount (Polysciences, Warrington, PA), mounted on a coverslip, and immediately observed under an Axiovert 200M fluorescence microscope (Carl Zeiss, Jena, Germany), using a fluorescein isothiocyanate (FITC) filter set with absorbance at 490 to 494 nm and emission at 517 nm. Images were captured with an AxioCam MPc digital camera. ImageJ software was used to analyze the cell lengths of at least 200 individual cells from 10 different images.
Virulence and bacterial growth assays.Virulence assays on apple trees were performed as described previously (46). Briefly, overnight cultures of E. amylovora WT and mutant strains were harvested by centrifugation and suspended in 0.5× phosphate-buffered saline (PBS). Cell suspensions were adjusted to an OD600 of 0.1 in PBS and inoculated onto seven actively growing cv. ‘Gala’ apple shoots by pricking the tip with a sterile needle and pipetting 5 μl of bacterial inoculum onto the tip. Symptom development was recorded at 7 days postinoculation (dpi), and the experiment was performed at least two times.
For virulence and bacterial growth assays, immature pear fruits (Pyrus communis L. cv. ‘Bartlett’) were surface sterilized with 10% bleach, pricked with a sterile needle, and inoculated with 2 μl of a 100× dilution of bacterial suspension at an OD600 of 0.1 in PBS (12, 13). The tissue surrounding the inoculation site was excised with a no. 4 cork borer and homogenized in 1 ml of 0.5× PBS. Bacterial growth within the pear tissue was monitored at 1, 2, and 3 dpi by dilution plating on LB medium with appropriate antibiotics. For each time point and strain tested, fruits were assayed in triplicate. Symptom development was recorded at 4 and 8 days postinoculation, and the experiment was performed three times.
Hypersensitive response (HR) assay.Overnight cultures of E. amylovora WT, mutant, and complementation strains were harvested by centrifugation and suspended in 0.5× PBS to an OD600 of 0.1. Bacterial suspensions were infiltrated into tobacco leaves (Nicotiana tabacum) by use of a needleless syringe. HR symptoms were recorded at 24 h postinfiltration, and the experiment was repeated three times.
RNA isolation.Bacterial strains grown overnight in LB medium with appropriate antibiotics were harvested by centrifugation and washed three times before being inoculated into 5 ml of HMM, to a final OD600 of 0.2. After 3 h of incubation at 18°C in HMM, 2 ml of RNAprotect reagent (Qiagen) was added to 1 ml of bacterial cell culture, mixed by vortexing, and incubated at room temperature for 5 min. Cells were harvested by centrifugation, and RNA was extracted using an RNeasy minikit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. DNase I treatment was performed with a Turbo DNA-free kit (Ambion, Austin, TX), and RNA was quantified using a NanoDrop ND100 spectrophotometer (NanoDrop Technologies, Wilmington, DE). For in vivo conditions, overnight cultures of bacterial strains were harvested by centrifugation, washed three times, and resuspended in PBS. Immature pear fruits were cut in half and inoculated with bacterial suspensions. After 3 h of incubation at 28°C in a moist chamber, bacterial cells were collected by washing pear surfaces with RNAprotect reagent (Qiagen) mixed 2:1 with water, and RNA was extracted as described above.
qRT-PCR.One microgram of total RNA was reverse transcribed using Superscript III reverse transcriptase (Invitrogen, Carlsbad, CA) following the manufacturer's instructions. One microliter of cDNA was used as the template for qPCR, using a StepOnePlus real-time PCR system (Applied Biosystems, Foster City, CA). Power SYBR green PCR master mix (Applied Biosystems) was used to detect the expression of selected genes amplified with primers designed using Primer3 software. Amplifications were carried out by incubation at 95°C for 10 min followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. Dissociation curve analysis was performed after the program was completed to confirm the amplification specificity. Three technical replicates were performed for each biological sample. Relative gene expression was calculated by the relative quantification (ΔΔCT) method, using the rpoD gene as an endogenous control.
Western blotting.E. amylovora cells grown in HMM at 18°C for 6 h were harvested, and equal amounts of cell lysates were separated in sodium dodecyl sulfate-polyacrylamide gels. Proteins were transferred to a polyvinylidene difluoride membrane (Millipore) and blocked with 5% milk in PBS. To detect HrpA-6His, membranes were probed with rabbit anti-His antibodies (GeneScript, Piscataway, NJ) that were diluted to 1.0 μg/ml with PBS containing 0.1% Tween 20 (PBST). Immunoblots were then developed with horseradish peroxidase-linked anti-rabbit IgG antibodies (Amersham Biosciences), diluted 1:10,000 in PBST, followed by enhanced chemiluminescence reagents (Pierce). Images of the resulting blots were acquired using an ImageQuant LAS 4010 charge-coupled device (CCD) camera (GE Healthcare).
Flow cytometry analysis.Bacterial strains containing promoter-GFP fusion plasmids were grown overnight in LB medium, washed three times with HMM, and inoculated into HMM containing 10 mM galactose to a final OD600 of 0.2 (40, 47). dl-Serine hydroxamate (SHX) and α-methylglucoside (αMG) were added to cells, to final concentrations of 0.1 mM and 0.5%, respectively. GFP intensities were measured by flow cytometry (BD Biosciences, San Jose, CA) after incubation at 18°C for 18 h. Flow cytometry was performed on a BD LSRII 10 parameter multilaser analyzer (BD Bioscience, San Jose, CA). Data were collected for a total of 100,000 events and analyzed statistically by gating using the flow cytometry software FCS Express V3 (De Novo Software, Los Angeles, CA). The geometric mean was calculated for each sample. Each treatment was performed in triplicate, and each experiment was repeated three times.
RESULTS
Growth of Erwinia amylovora is regulated by ppGpp and DksA.Based on the genome sequence of Erwinia amylovora (48), two annotated genes were found to encode RelA-SpoT homologue (RSH) proteins, i.e., Eam_2706 (relA) and Eam_0043 (spoT), whereas Eam_0793 encodes DksA. During mutant construction, growth of the spoT mutant was observed to be very slow in rich medium, which is consistent with observations in E. coli, where the spoT deletion mutant is lethal due to accumulation of high levels of ppGpp (23). An spoT deletion mutant has also been reported for Pseudomonas fluorescens (35).
Growth rates of the WT and five mutant strains in rich (LB) and minimal (MBMA) media were determined. In LB medium, growth rates of the WT and four mutants (the relA, dksA, relA/spoT, and relA/dksA mutants) were similar, with the spoT mutant exhibiting slower growth at the beginning but reaching a level similar to that of the WT at 24 h (Fig. 1A and C). When the spoT gene was provided in trans, slow growth of the spoT mutant was restored (Fig. 1A). In contrast, growth rates of the spoT and relA mutants were similar to that of the WT in MBMA medium (Fig. 1B), but growth of the dksA mutant was much reduced in MBMA medium, and its growth could be complemented by expressing the dksA gene in trans (Fig. 1B). Furthermore, both the relA/spoT and relA/dksA mutants were unable to grow in MBMA medium (Fig. 1D). Complementation of the relA/spoT mutant with the spoT gene in trans recovered its growth, whereas complementation with the relA gene was unsuccessful, possibly due to ppGpp accumulation. Complementation of the relA/dksA mutant with the dksA gene in trans recovered its growth, whereas complementation with the relA gene in trans partially recovered its growth (Fig. 1D). These results suggest that DksA and ppGpp play roles in regulating cell growth under both nutrient-rich and nutrient-limited conditions.
Growth of Erwinia amylovora is affected by DksA and ppGpp. The graphs show the growth of the E. amylovora WT strain, the relA, spoT, and dksA single mutants, and their corresponding complementation strains in LB (A) and MBMA (B) media and the growth of the E. amylovora WT strain, the relA/spoT (ppGpp0) and relA/dksA double mutants, and their corresponding complementation strains in LB (C) and MBMA (D) media. The experiments were repeated at least two times with similar results.
Intracellular ppGpp levels in E. amylovora WT and mutant strains were quantified in HMM by using the selective and sensitive fluorescent chemosensor PyDPA for ppGpp as reported previously (44, 45). PyPDA contains pyrene (Py) and bis(Zn2+-dipicolylamine) (DPA), the latter of which is well known for its strong binding to pyrophosphate groups in water (45). We found that the ppGpp level was slightly increased (10% higher) in the spoT mutant, while the ppGpp level was about 7 times lower in the relA mutant than in the WT. Accumulation of ppGpp in the dksA mutant was about half that in the WT (Fig. 2). Furthermore, ppGpp levels in the relA/dksA and relA/spoT mutants (the latter is referred to as ppGpp0 from this point on) were undetectable (Fig. 2). These findings indicate that both RelA and SpoT are required for ppGpp synthesis in HMM and also suggest that RelA plays a major role in ppGpp synthesis.
ppGpp measurement. Intracellular ppGpp levels in Erwinia amylovora WT and relA, spoT, dksA, relA/spoT, and relA/dksA mutant strains were quantified in HMM by using the fluorescent chemosensor PyDPA as reported previously (44, 45). One-way analysis of variance (ANOVA) and Student's t test (P = 0.05) were used to analyze the data. Values marked with the same letter were not significantly different (P < 0.05). NI, not included in statistical analysis. This experiment was performed twice.
Mutations in relA and spoT of Erwinia amylovora result in altered cell lengths.When E. coli ppGpp mutant cells were grown under isoleucine starvation conditions, they were found to be considerably longer than WT cells (49, 50). The cell sizes/lengths of E. amylovora WT and mutant strains constitutively expressing GFP and grown in LB medium or HMM were observed by epifluorescence microscopy (Fig. 3A and B). When cells were grown in nutrient-rich LB medium, the average length of WT cells was about 2.23 μm, whereas the average lengths of the relA and ppGpp0 mutant cells were slightly or much longer, reaching 2.65 μm and 3.87 μm, respectively. In contrast, the average length of the spoT mutant cells was 1.94 μm, which is slightly shorter than that of the WT (Fig. 3). However, the lengths of the majority of the spoT mutant and WT cells were the same (Fig. 3B). When cells were grown in nutrient-limited HMM, they were shorter than those grown in LB medium, except for the spoT mutant (Fig. 3A and B). While the average lengths of the relA and ppGpp0 mutant cells reached 2.37 μm and 3.47 μm, respectively, the average lengths of WT and spoT mutant cells were about 1.37 and 2.17 μm, respectively (Fig. 3). However, the lengths of the majority of the spoT mutant cells were the same in both media and were also identical to those of the relA mutant cells in HMM (Fig. 3B). These findings indicate that the reduced lengths of WT cells and increased lengths of relA and ppGpp0 mutant cells might be due to ppGpp accumulation in these cells, thus suggesting that the ppGpp-mediated stringent response controls the cell size of E. amylovora, and also suggesting that E. amylovora might require certain levels of ppGpp to control cell size in vitro.
ppGpp controls cell size in Erwinia amylovora. (A) Epifluorescence microscopy images of E. amylovora WT and relA, spoT, and relA/spoT (ppGpp0) mutant strains constitutively expressing GFP and grown in LB medium or HMM for 4 h. Magnification, ×200. (B) Distributions of sizes and average cell lengths of WT and relA, spoT, and relA/spoT (ppGpp0) mutant strains constitutively expressing GFP and grown in LB medium or HMM. The experiments were repeated at least two times with similar results.
Both DksA and ppGpp are required for virulence, elicitation of the hypersensitive response (HR), and bacterial growth.To determine the role of ppGpp/DksA in E. amylovora pathogenesis, virulence assays were performed on apple shoots and immature pear fruits. Necrosis around the point of inoculation was visible at 3 dpi and moved quickly into the apple shoots, and the length of necrotic diseased shoots reached more than 30 cm after 7 days for the WT strain (Table 2). While disease symptoms were not visible for the dksA, relA/dksA, and ppGpp0 mutant strains, disease severity was strongly reduced for the relA and spoT mutants, with average lengths of necrotic shoots of 10 and 19 cm, respectively. Complementation of the relA, spoT, and dksA mutants with their respective genes restored disease severity to the WT level, while complementation of the relA/dksA and ppGpp0 double mutants with a single gene rescued the ability to cause disease to the level of a single mutant (Table 2). Similarly, the dksA, relA/dksA, and ppGpp0 mutant strains were nonpathogenic on immature pear fruits, while the relA and spoT mutants caused disease similarly to the WT strain (Fig. 4A). Complementation of these mutant strains resulted in disease recovery similar to that described above for apple shoots (Fig. 4B).
Comparison of disease severities with Erwinia amylovora strain Ea1189, ppGpp mutants, and complementation strains
Pathogenicity and HR assays. (A and B) Symptoms caused by the WT strain, the relA, spoT, dksA, relA/spoT (ppGpp0), and relA/dksA mutants (A), and their complementation strains (B) on immature pear fruits. Immature pears (cv. ‘Bartlett’) were surface sterilized, pricked with a sterile needle, and inoculated with 2 μl of bacterial suspension. Symptoms were recorded and photos were taken at 4 and 8 dpi. (C and D) HR assay on tobacco leaves. The E. amylovora WT strain, the relA, spoT, dksA, relA/dksA, and ppGpp0 mutants (C), and their complementation strains (D) were allowed to infiltrate into 8-week-old tobacco leaves at a concentration of 108 CFU ml−1. PBS was used as a negative control. Photographs were taken at 24 h postinfiltration.
When infiltrated into tobacco leaves, the WT, the relA and spoT mutants, and the relA, spoT, and dksA complementation strains elicited typical HR cell death in tobacco leaves (Fig. 4C and D). However, no HR was observed for the dksA, relA/dksA, and ppGpp0 mutant strains (Fig. 4C and D), indicating that both DksA and ppGpp are required for eliciting HR in tobacco.
To determine whether disease symptoms were correlated with bacterial growth, growth of the five mutants on immature pears was compared to that of the WT. At 1 and 2 dpi, bacterial growth of the relA and spoT mutants, as well as the relA, spoT, and dksA complementation strains, was slightly reduced compared to that of the WT strain, but growth of these strains was similar to that of the WT at 3 dpi (Fig. 5A). In contrast, bacterial growth of the dksA, relA/dksA, and ppGpp0 mutant strains as well as the relA/dksA mutant complemented with the relA gene did not increase at all from 1 to 3 dpi, representing 100- to 1,000-fold less growth than that of the WT (Fig. 5A and B). At 3 dpi, the ppGpp0 mutant complemented with the spoT gene and the relA/dksA mutant complemented with the dksA gene had restored growth on immature pears, and the bacterial growth was similar to that of the WT (Fig. 5B). These findings indicate that ppGpp and DksA are required for virulence, HR elicitation, and bacterial growth of E. amylovora and also suggest that both RelA and SpoT might be involved in ppGpp biosynthesis in planta.
Both DksA and ppGpp are required for bacterial growth in planta. (A) Growth of the E. amylovora WT strain, the relA, spoT, and dksA single mutants, and their corresponding complementation strains in immature pears. (B) Growth of the E. amylovora WT strain, the relA/dksA and relA/spoT (ppGpp0) double mutants, and their corresponding complementation strains in immature pears. Immature pears (cv. ‘Bartlett’) were surface sterilized, pricked with a sterile needle, and inoculated with 2 μl of bacterial suspension. Tissue surrounding the inoculation site was excised with a no. 4 cork borer and homogenized in 1 ml of 0.5× PBS. Bacterial growth within the pear tissue was monitored at 1, 2, and 3 days postinoculation by dilution plating on LB medium with appropriate antibiotics. d, day.
Both DksA and ppGpp are required for T3SS gene expression.Since both DksA and ppGpp are required for virulence and HR, the effects of DksA and ppGpp on T3SS gene expression were determined using qRT-PCR. We found that expression of T3SS regulatory and effector genes in HMM, including hrpL, hrpA, hrpN, and dspE, was abolished in the dksA, relA/dksA, and ppGpp0 mutants (Fig. 6A). In the relA mutant, expression of hrpL, hrpA, hrpN, and dspE was about 2- to 5-fold lower than that in the WT. In contrast, expression of hrpL, hrpA, hrpN, and dspE in the spoT mutant was 1.7- to 2.5-fold higher than that in the WT (Fig. 6A). Similarly, expression of T3SS genes in the dksA, relA/dksA, and ppGpp0 mutants was strongly downregulated on immature pears, but expression levels were similar in the relA mutant and the WT (Fig. 6B). While expression of hrpN and dspE was upregulated >2-fold in the spoT mutant, the hrpL and hrpA genes in the spoT mutant were not differentially expressed compared to those of the WT on immature pears (Fig. 6B). These results indicate that both ppGpp and DksA are required for T3SS gene expression.
Both DksA and ppGpp activate T3SS gene expression in Erwinia amylovora. (A) Expression of T3SS regulatory and effector genes (hrpL, hrpA, hrpN, and dspE) in the relA, spoT, dksA, relA/dksA, and relA/spoT (ppGpp0) mutant strains compared to the WT strain in HMM, as determined by qRT-PCR. (B) Expression of T3SS regulatory and effector genes (hrpL, hrpA, hrpN, and dspE) in the relA, spoT, dksA, relA/dksA, and ppGpp0 mutant strains compared to the WT on immature pear fruits. Relative gene expression of selected T3SS genes was calculated by the 2−ΔΔCT method, utilizing the rpoD gene as an endogenous control. Fold changes are the means of results for three replicates. Each experiment was performed at least two times with similar results. Error bars indicated standard deviations.
Moreover, an abundance of HrpA protein in the WT and four mutants grown in HMM was detected by Western blotting (Fig. 7). Only 4 and 10% protein signals were detected for the dksA and ppGpp0 mutants, respectively, but about 96 and 89% protein signals were detected for the relA and spoT mutants, respectively, compared to the signal of the WT strain (Fig. 7). These results indicate that both ppGpp and DksA are required for protein accumulation.
Accumulation of HrpA protein is controlled by DksA and ppGpp. The HrpA-His6 protein in the WT and relA, spoT, dksA, and relA/spoT (ppGpp0) mutant strains was detected by Western blotting using an anti-histidine protein antibody after growth in HMM at 18°C for 6 h. Relative protein abundances were calculated by using ImageJ software, utilizing the average pixel value of the signals and considering the abundance of the WT sample to be 100%.
RelA is mainly responsible for T3SS activation in vitro.To assess the roles of RelA and SpoT in T3SS gene expression in response to different starvation signals, promoter activities of the hrpL and hrpA genes were measured by flow cytometry analysis of bacterial strains grown in HMM supplemented with serine hydroxamate (SHX) or the nonmetabolizable glucose analog α-methylglucoside (αMG), both of which induce the stringent response (51, 52). The GFP intensities from both the hrpL and hrpA promoters in the WT and the relA mutant exhibited little change in HMM plus SHX or αMG compared to those in HMM alone (Table 3). In contrast, the GFP intensities from both the hrpL and hrpA promoters in the spoT mutant were higher in HMM plus SHX and 2- to 4-fold higher in HMM plus αMG compared to those in HMM alone (Table 3). These findings indicate that RelA is mainly responsible for activating T3SS gene expression in vitro and might respond to both SHX- and αMG-induced stresses in HMM. These results also suggest that SpoT does not respond to these starvation signals or that the hydrolase activity of SpoT is dominant under in vitro conditions.
Promoter activities of hrpL and hrpA genes in Erwinia amylovora WT and ppGpp mutant strains
DISCUSSION
E. amylovora utilizes the hrp T3SS to deliver effector proteins into host plants to suppress host defense, acquire nutrients, and cause disease (1, 2). In the regulatory networks of E. amylovora, a sigma factor cascade (RpoN-HrpL) quickly activates the T3SS in response to inducing signals, including limited nutrients and oxidative stress (4, 9, 11). These signals are similar to those sensed by RSH proteins, which are responsible for bacterial alarmone (p)ppGpp biosynthesis and degradation. In this study, we demonstrated that the DksA/ppGpp-mediated stringent response is essential for T3SS gene expression and virulence in E. amylovora through activation of the sigma factor cascade. These findings suggest that signals triggering ppGpp biosynthesis are most likely to be responsible for activation of the T3SS and also indicate that activation of the sigma factor cascade by the ppGpp-mediated stringent response might allow the pathogen to integrate diverse host/environmental signals encountered during the infection process (4).
It has been reported that the null mutation of the spoT gene in the relA+ background in E. coli is lethal, suggesting that high levels of ppGpp may be toxic (23, 53–55). On the other hand, an spoT mutant of E. coli was first reported and later found to contain a secondary spontaneous point mutation in the relA gene (H354Y), which reduces RelA activity to about 20% of the wild-type level (56, 57). In this study, an spoT null mutant was generated in E. amylovora, as reported for P. fluorescens, both of which accumulated only slightly higher levels of ppGpp (35). To exclude the possibility of any secondary mutations in the relA gene in the spoT null mutant, the relA gene was sequenced, and no mutations were found (data not shown). Bioinformatic analysis revealed that many bacterial genomes encode additional single-domain, ppGpp-synthesizing or -hydrolyzing RSHs (58, 59). A gene (Eam_3399) annotated a hypothetical gene in E. amylovora (48) encodes a small RSH protein (paSpo) containing a single hydrolase domain (58). A similar RSH protein (pbcSpo2), also containing a single hydrolase domain, is encoded in the genome of P. fluorescens (35, 58). Although these small RSH proteins have not been characterized functionally (59), the recovery of an spoT single deletion mutant for both E. amylovora and P. fluorescens indirectly suggests that both paSpo and pbcSpo2 might be functional as ppGpp hydrolases, and this might also explain why deletion of the spoT gene is not lethal (59).
Characterization of the spoT null mutants of both E. amylovora and P. fluorescens yielded very interesting results. For P. fluorescens, despite higher levels of ppGpp, the spoT mutant showed an attenuated level of antibiotic activity, like the relA mutant, which produced almost undetectable levels of ppGpp (35). In this study, despite higher levels of ppGpp, the length of the spoT mutant cells was much longer than that of the WT cells in HMM. However, the lengths of the majority of the spoT mutant cells were the same in both media and were identical to those observed for the WT in LB medium and for the relA mutant in HMM (Fig. 3B). Growth rates of the relA and spoT mutants were slightly reduced compared to that of the WT in planta, suggesting that accumulation of ppGpp, which may depend on both RelA and SpoT within plants, is required for bacterial multiplication in a plant environment. Furthermore, both relA and spoT mutants caused reduced disease symptoms in apple shoots, suggesting that both RelA and SpoT are required for full virulence of E. amylovora in planta. An earlier report also found that low levels of ppGpp are sufficient for full virulence in Enterococcus faecalis (60). Although we still lack plausible explanations for the nonlinear relationship between ppGpp levels and various observed phenotypes, it is possible that quantitative differences in the intracellular ppGpp level determine the precise expression of genes controlling various phenotypes in different environments, as reported recently (61). These findings also suggest that E. amylovora might require certain levels of ppGpp to cause disease in vivo and to control cell size in vitro.
It is well known that both DksA and ppGpp bind to RNAP and that DksA enhances the effect of ppGpp by modulating the direct interaction between RNAP and ppGpp (18, 26–28). Although growth of the dksA and ppGpp0 mutants was normal in rich medium, growth of the dksA and ppGpp0 mutants on immature pears and in HMM was impaired, and these mutants were unable to cause disease or elicit HR, suggesting that both DksA and ppGpp strongly influence the growth of E. amylovora in planta and in HMM. In addition, many studies have reported that nutrient availability adversely affects cell size (49, 62–67). Upon inoculation of Pseudomonas syringae onto a bean leaf surface, the length of these cells was rapidly reduced, suggesting that the leaf surface is a habitat with limited nutrients (66). Similarly, ppGpp mutant cells of E. coli were considerably longer than WT cells when cultured under starvation conditions (49). It has also been reported that cells of smaller sizes become increasingly resistant to abiotic stresses, including osmotic and oxidative stresses, thus enhancing their ability to survive under harsh environmental conditions (33, 36, 68–70). This may provide an explanation for why dksA and ppGpp0 mutants fail to grow in HMM, suggesting that E. amylovora might require both ppGpp and DksA to coregulate cell size and resistance to environmental stresses, thus contributing to overall survival (59). It is interesting that the ppGpp level in the dksA mutant was reduced, suggesting that a positive-feedback regulation may exist for the biosynthesis of ppGpp, as reported previously (71). It is also possible that a growth defect of the dksA mutant in HMM contributes to reduced ppGpp accumulation. Furthermore, it is tempting to speculate that DksA may control gene expression independently of ppGpp, since DksA acts as a transcriptional factor (59). These findings suggest that both DksA and ppGpp are required for virulence, bacterial growth, and T3SS gene expression, and thus survival, in E. amylovora.
Although we observed a linear relationship between ppGpp levels and T3SS gene expression in vitro, Western blotting showed that HrpA protein accumulations in the spoT and relA mutants were not significantly different and were about 5 to 10% less than that in the WT. This result was consistent with the virulence assay data showing that both relA and spoT mutants caused reduced disease on apple shoots. The discrepancy between T3SS gene expression, protein accumulation, and disease-causing ability may be due to the T3SS in E. amylovora also being regulated at the posttranscriptional or translational level, as reported for other plant-pathogenic bacteria. In addition, both SHX and αMG could induce the stringent response in E. coli by mimicking amino acid and carbon starvation, respectively (17, 51, 52). Interestingly, when either SHX or αMG was added to HMM, promoter activities of hrpL and hrpA were strongly induced only in the spoT mutant, not in the WT and the relA mutant, suggesting that RelA is mainly responsible for activating T3SS gene expression in vitro and might respond to both SHX- and αMG-induced stresses. Initially, it appears that our results contrast with the notion that SpoT responds to carbon starvation. However, our results are consistent with those of a previous report indicating that amino acids and fructose act synergistically in activating the T3SS in P. syringae (14). Since HMM contains galactose, a sugar used to induce the T3SS (15), it is possible that αMG acts synergistically with galactose to activate T3SS expression. Furthermore, these findings also suggest that the hydrolase activity of SpoT is dominant and that RelA may act as a main ppGpp synthase under in vitro conditions. However, it will be interesting to investigate why only the spoT mutant responds to both inducers.
Based on our results and previously reported data, the following model is proposed for how E. amylovora incorporates host and environmental signals in regulating T3SS gene expression (Fig. 8). Upon arrival on a plant surface, E. amylovora cells experience stress conditions, such as limited nutrients and oxidative stress, which trigger the activation of the RelA/SpoT system, leading to the accumulation of ppGpp. Both DksA and ppGpp directly bind to RNAP, indirectly promoting binding to alternative sigma factors, such as RpoN and HrpL. RpoN, along with HrpS and YhbH, binds to the hrpL promoter to trigger hrpL transcription (4). HrpL then recognizes a conserved “hrp box” at the promoter regions of HrpL-dependent operons or genes, leading to the expression of other T3SS structural and effector genes. However, the exact signals sensed by RelA and SpoT in the plant environment are still not clear. In the future, research should focus on investigating the global effects of ppGpp both in vitro and in planta, as well as the signals that activate ppGpp accumulation and the T3SS. Given that the ppGpp0 strain is unable to survive and cause disease, further research on targeting ppGpp in E. amylovora for development of control strategies is warranted.
Working model illustrating the role of ppGpp in Erwinia amylovora in response to plant and environmental stimuli. This model is based on findings obtained in this study as well as those reported in previous studies (4, 5, 8, 11). Symbols: ↓, positive effect; ⊥, negative effect; IhfA and -B, integration host factors α and β; RNAP, RNA polymerase; OM, outer membrane; IM, inner membrane.
ACKNOWLEDGMENTS
We thank Raymond Zielinski at the University of Illinois for his help with epifluorescence microscopy.
This project was supported by Agriculture and Food Research Initiative competitive grant 2010-65110-20497 from the USDA National Institute of Food and Agriculture and by USDA-Hatch/Multistate Project grants ILLU-802-913 and ILLU-802-396 (to Y.Z.).
FOOTNOTES
- Received 5 December 2014.
- Accepted 2 February 2015.
- Accepted manuscript posted online 9 February 2015.
- Address correspondence to Youfu Zhao, zhao888{at}illinois.edu.
↵* Present address: Veronica Ancona, Texas A&M University-Kingsville Citrus Center, Weslaco, Texas, USA.
Citation Ancona V, Lee JH, Chatnaparat T, Oh J, Hong J-I, Zhao Y. 2015. The bacterial alarmone (p)ppGpp activates the type III secretion system in Erwinia amylovora. J Bacteriol 197:1433–1443. doi:10.1128/JB.02551-14.
REFERENCES
- Copyright © 2015, American Society for Microbiology. All Rights Reserved.
