Área de Genética, Facultad de Ciencias, Instituto de Hortofruticultura Subtropical y Mediterrnea “La Mayora”, Universidad de Málaga-CSIC (IHSM-UMA-CSIC), Málaga, Spain
Correspondence: Isabel Pérez-Martínez, Área de Genética, Facultad de Ciencias, Universidad de Málaga, Campus de Teatinos, Boulevard Louis Pasteur s/n, E-29010-Málaga, Spain. Tel.: +34 952131676; fax: +34 952132001; e-mail: email@example.com
Indole-3-acetic acid (IAA) is a widespread phytohormone among plant-associated bacteria, including the tumour-inducing pathogen of woody hosts, Pseudomonas savastanoi pv. savastanoi. A phylogenetic analysis of the iaaM/iaaH operon, which is involved in the biosynthesis of IAA, showed that one of the two operons encoded by Pseudomonas savastanoi pv. savastanoi NCPPB 3335, iaaM-1/iaaH-1, is horizontally transferred among bacteria belonging to the Pseudomonas syringae complex. We also show that biosynthesis of the phytohormone, virulence and full fitness of this olive pathogen depend only on the functionality of the iaaM-1/iaaH-1 operon. In contrast, the iaaM-2/iaaH-2 operon, which carries a 22-nt insertion in the iaaM-2 gene, does not contribute to the production of IAA by this bacterium. A residual amount of IAA was detected in the culture supernatants of a double mutant affected in both iaaM/iaaH operons, suggesting that a different pathway might also contribute to the total pool of the phytohormone produced by this pathogen. Additionally, we show that exogenously added IAA negatively and positively regulates the expression of genes related to the type III and type VI secretion systems, respectively. Together, these results suggest a role of IAA as a signalling molecule in this pathogen.
Production of the phytohormone indol-3-acetic acid (IAA) is common among plant-associated bacteria, and this phytohormone can interfere with plant development by disturbing the auxin balance in plants. Moreover, IAA has been described as a signalling molecule that can influence microbial gene expression, including that of bacterial phytopathogens (Spaepen & Vanderleyden, 2010; Patten et al., 2012). The best characterized IAA biosynthetic pathway in phytopathogenic bacteria is the indole-3-acetamide pathway. In this pathway, the genetic determinants involved in the conversion of tryptophan (Trp) to IAA are Trp monooxygenase (encoded by the iaaM gene), which converts Trp to indoleacetamide (IAM), and IAM hydrolase (encoded by the iaaH gene), which catalyses the conversion of IAM to IAA. Furthermore, the oleander pathogen Pseudomonas savastanoi pv. nerii (Psn) is also able to convert IAA to a less biologically active compound, the amino acid conjugate IAA–lysine (IAA–Lys), through expression of the iaaL gene (Glass & Kosuge, 1986).
The iaaM/iaaH genes, which are best characterized in phytopathogenic bacteria such as Agrobacterium spp. and Pseudomonas savastanoi, have been proposed to have originated from common ancestor genes in these two species based on their amino acid sequences (Gielen et al., 1984; Yamada et al., 1985, 1986; Inze et al., 1987). In Psn, these two genes are encoded on plasmids. They are co-transcribed in the same transcriptional unit, and they are essential for gall formation in oleander and olive (Comai & Kosuge, 1980; Yamada et al., 1985; Palm et al., 1989; Gaffney et al., 1990). In contrast, P. savastanoi pv. savastanoi (Psv, olive isolates) usually carries two copies of both IAA genes on its chromosome (Pérez-Martínez et al., 2008). Psv NCPPB 3335 causes olive knot disease and is a model bacterium in which to study the molecular basis of pathogenesis and tumour formation in woody hosts (Matas et al., 2012). This strain contains two chromosomally encoded copies of the iaaM/iaaH genes (Rodríguez-Palenzuela et al., 2010), which are organized into two operons called iaaM-1/iaaH-1 and iaaM-2/iaaH-2. Additionally, the iaaM-2 gene has been suggested to be a pseudogene that contains a premature termination codon yielding a product of 142 amino acids (Ramos et al., 2012). However, the specific role of each of these two IAA operons in the pathogenicity and virulence of Psv strains has not been reported to date.
In addition to the contribution of IAA in bacterial phytopathogens in circumventing plant defence responses, IAA can also directly affect bacterial physiology and survival during plant infection (Spaepen & Vanderleyden, 2010). Examples of the signalling role of IAA in plant-associated bacteria include the upregulation of the type III secretion system (T3SS) and the type VI secretion system (T6SS) in the bacterial phytopathogen Dickeya dadantii (Yang et al., 2007) and the rhizosphere bacterium Azospirillum brasilense (Van Puyvelde et al., 2011), respectively. However, the role of IAA as a signalling molecule in the Pseudomonas syringae complex, which also includes P. savastanoi, has only been reported for the synthesis of the virulence determinant syringomycin (Mazzola & White, 1994).
In the present study, we investigated the functionality of the two iaaM/iaaH operons present in the genome of Psv NCPPB 3335 and their individual roles in virulence using single and double ΔiaaMH mutants constructed by gene replacement. Moreover, we report the effect of exogenously added IAA on the transcription of virulence-related genes, including the T3SS (hrpA and hrpL) and the T6SS (vgrG).
Materials and methods
Bacterial strains and culture conditions
Strains and plasmids used in this study are listed in Table 1. The strains of Psv NCPPB 3335 and Escherichia coli were grown at 28 and 37 °C, respectively, in Luria–Bertani (LB) medium (Miller, 1972) or super optimal broth (SOB; Hanahan, 1983). When required, antibiotics were added at the following concentrations. For E. coli: ampicillin 100 μg mL−1, kanamycin 50 μg mL−1 and tetracycline 10 μg mL−1. For P. savastanoi: ampicillin 300 μg mL−1, kanamycin 10 μg mL−1, and tetracycline 10 μg mL−1. When relevant, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) was added to a final concentration of 0.02%.
pGEM-T derivative containing c. 1.2 kb on each side of the iaaM/iaaH-1 operon from NCPPB 3335 interrupted by the kanamycin resistance gene nptII (ApR, KmR).
pGEM-T derivative containing c. 1.2 kb on each side of the iaaM/iaaH-2 operon from NCPPB 3335 interrupted by the kanamycin resistance gene nptII (ApR, KmR).
Contains a fragment of 270 bp corresponding with the hrpL promoter from NCPPB 3335 directionally cloned in pMP220 using EcoRI
Contains a fragment of 245 bp corresponding with the hrpA promoter from NCPPB 3335 directionally cloned in pMP220 using EcoRI
Construction of Psv mutants
To construct the single mutants ΔiaaMH-1 and ΔiaaMH-2, a fragment of c. 2 kb was removed from the iaaM/iaaH operon. To construct the plasmids pGEM-T-ΔiaaMH1 and pGEM-T-ΔiaaMH2, the upstream and downstream fragments of the region to be deleted were cloned into the pGEM®-T Easy Vector (Table 1) as described by Matas et al. (2014). Later, the nptII kanamycin resistance gene obtained from pGEM-T-KmFRT-XhoI and pGEM-T-KmFRT-HindIII (Table 1) was introduced into these plasmids, yielding pGEM-T-ΔiaaMH1-Km and pGEM-T-ΔiaaMH2-Km, respectively (Table 1). Finally, for marker exchange mutagenesis (Supporting Information, Fig. S1), each plasmid was transformed by electroporation into NCPPB 3335 as described previously (Pérez-Martínez et al., 2007). Screening and verification of the mutants was carried out as previously described (Matas et al., 2014).
The double mutant ΔiaaMH-1.2 was generated using the ΔiaaMH-1 single mutant, in which the kanamycin gene was removed using the pFLP2 plasmid (Zumaquero et al., 2010). Then, a ΔiaaMH1-KmS clone was transformed with the pGEM-T-ΔiaaMH2-Km plasmid for marker exchange mutagenesis to obtain the double mutant.
Transcription initiation mapping by 5′cRACE
The transcription start point of the iaaM-1/iaaH-1 operon was determined using a system for rapid amplification of cDNA ends (5′cRACE; Filiatrault et al., 2011). Total RNA from wild-type Psv NCPPB 3335 was obtained as previously described (Matas et al., 2014). One microgram of this RNA was used as a template to synthesize the first-strand cDNA, using the synthesis kit SMART™ RACE cDNA Amplification (Clontech, Mountain View, CA) and iaaM-1-specific primers designed to anneal within the coding region of this allele and not iaaM-2. All reactions were performed according to the manufacturer's instructions. The amplification products were cloned into the pGEM®-T Easy Vector and sequenced.
Quantification of IAA production
For the IAA extraction for HPLC, 2 mL of the supernatant from liquid cultures grown in LB medium without added tryptophan was collected and acidified to pH 2.5–3.0 with 1 M HCl. Then, the supernatant was mixed with the same volume of ethyl acetate and incubated under shaking conditions for 30 min at room temperature. Finally, the extracted ethyl acetate fraction was evaporated at room temperature, and the pellet was resuspended in 1 mL MeOH/H2O (10 : 90; v/v), filtered (0.2 μm) and injected into the HPLC machine. The analyses were carried out using a SunFire HPLC system with C18 analytical column, 5 μm particle size, 2.1 × 100 mm (Waters). The chromatographic system was interfaced to a triple quadrupole mass spectrometer (TQD, Waters, Manchester, UK). masslynx nt version 3.4 (Micromass) was used to process the quantitative data from the calibration standards and bacterial samples. IAA quantities were normalized using the dry weight (DW) of a 2-mL bacterial pellet, which was obtained by complete drying for 1 h at 80 °C.
Plant infection and isolation of bacteria from olive knots
Olive plants were micropropagated, rooted and maintained as previously described (Rodríguez-Moreno et al., 2008). Micropropagated plants were infected in the stem wound with a bacterial suspension (c. 103 CFU) and incubated for 30 days in a growth chamber as described by Rodríguez-Moreno et al. (2008). The morphology of the olive plants infected with bacteria was visualized using a stereoscopic microscope (Leica MZ FLIII; Leica Microsystems, Wetzlar, Germany). Knot volume was quantified using a three-dimensional scanner, and minimagics 2.0 software. The competitive index (CI) between the wild-type and the iaaMH mutants was determined as described previously (Rodríguez-Moreno et al., 2008; Matas et al., 2012). To analyse the pathogenicity of P. savastanoi isolates in 1-year-old olive explants (lignified plants), the plants were maintained and inoculated according to previously described methods (Penyalver et al., 2006; Pérez-Martínez et al., 2007; Matas et al., 2012). Morphological changes, scored at 90 days post-infection (d.p.i.), were captured with a high-resolution digital camera (Canon D600; Canon Inc., Tokyo, Japan). Knot volume was calculated as previously reported (Moretti et al., 2008; Hosni et al., 2011). Statistical data analysis was performed by anova followed by Student's t-test (P ≤ 0.05).
The transcriptional fusions of the T3SS-related genes, hrpA (structural gene encoding Hrp pilin) and hrpL (a regulatory gene encoding an alternative sigma factor) promoters, to lacZ were constructed by PCR amplification using NCPPB 3335 genomic DNA and the forward and reverse primers shown in Supporting Information, Table S1. The resulting DNA fragments were EcoRI-digested and cloned into pMP220 to generate the plasmids pMP220-PhrpA and pMP220-PhrpL (Table 1). PCR using the forward primer specific for each promoter and pMP220-R was carried out to confirm that the promoters were directionally cloned in pMP220. Preinocula of bacterial strains harbouring the relevant plasmids were grown overnight in LB. The next day, the cells were diluted to an OD600 of 0.05 and grown to an OD600 of 0.5. At this point, total cells were pelleted, washed twice with 10 mM MgCl2 and resuspended in 5 mL of Hrp-inducing medium (Huynh et al., 1989) in the presence or absence of 1 mM IAA. No changes in the pH of this medium (pH 5.7) were observed after addition of IAA. β-Galactosidase activity was measured after 2, 6 and 24 h of incubation in this medium, as previously described (Miller, 1972).
Quantitative RT-PCR assays
Quantitative real-time PCR (qRT-PCR) assays in Psv NCPPB 3335 were performed in the same conditions as those used for the β-galactosidase assays. RNA extraction was carried out after 6 h of incubation in the Hrp-inducing medium. The cells were pelleted and processed for RNA isolation using TriPure Isolation Reagent (Roche Applied Science, Mannheim, Germany) as described previously (Matas et al., 2014). cDNA synthesis and qRT-PCR assays were as described by Matas et al. (2014). Transcriptional data were normalized to the housekeeping gene gyrA. qRT-PCR reactions were performed in triplicate.
Results and discussion
Horizontal transfer of the iaaM/iaaH operon in the P. syringae complex
IAA is synthesized in P. savastanoi through the indole-3-acetamide pathway, which involves the genetic determinants iaaM and iaaH (Fig. 1a). Two different alleles of each of these genes were found in the chromosome of Psv NCPPB 3335, organized in two operons (iaaM-1/iaaH-1 and iaaM-2/iaaH-2; Rodríguez-Palenzuela et al., 2010; Fig. 1c), which are located at a distance of about 670 kb (data not shown). A nucleotide sequence alignment of the complete iaaM/iaaH operons (from the start codon of the iaaM gene to the stop codon of the iaaH gene) encoded by NCPPB 3335 revealed 92% identity between them (the percentage identity between iaaM-1 and iaaM-2 and between iaaH-1 and iaaH-2 was 92% and 93%, respectively). In addition, a comparative phylogenetic study of the same nucleotide fragment was performed using all available bacterial genomes. Those sequences showing an identity lower than 40% with NCPPB 3335 were not included in this analysis. The iaaM/iaaH operon was found in five different groups of plant-pathogenic bacteria, i.e. the Pseudomonas syringae complex, Pantoea spp., Burkholderia spp., Dickeya spp. and Agrobacterium spp. In fact, all these bacteria have recently been included in Group I of iaaM-like sequences (Patten et al., 2012). According to these authors, Group I can be clearly distinguished from Group II, which includes many other bacterial species encoding a divergent Trp-monooxygenase. In general, the phylogeny of the iaaM/iaaH sequences is largely congruent with the phylogeny deduced from housekeeping genes (Ramos et al., 2012), suggesting that this operon is ancestral within these bacterial groups. However, the clustering of iaaM-2/iaaH-2 from Psv NCPPB 3335 (genomospecies 2) with P. syringae pv. aceris M302273PT and P. syringae pv. syringae B728a (genomospecies 1; Fig. 1b) is consistent with the possibility of horizontal gene transfer. This result is not surprising given that the iaaM/iaaH operon is often found in several copies and located in plasmids (Caponero et al., 1995; Pérez-Martínez et al., 2008). In contrast, the clustering of the NCPPB 3335 iaaM-1/iaaH-1 operon with Psn and P. syringae pv. glycinea (genomospecies 2) is consistent with the phylogeny of the iaaL-1 gene reported by Ramos et al. (2012). This gene is located upstream of the iaaM-1/iaaH-1 operon in NCPPB 3335 (Matas et al., 2009). Co-evolution and recent stabilization of the plasmid-encoded Psn fragment containing all three genes (iaaM, iaaH and iaaL) in the genome of Psv have been proposed (Matas et al., 2009). In agreement with this hypothesis, the transcription start site of the iaaM-1/iaaH-1 operon in Psv NCPPB 3335, determined by 5′cRACE, was located within the ORF of an IS4 transposase gene 401 bp upstream of the iaaM-1 start codon (Fig. 1c). This site is almost coincident (1 bp further upstream) with the transcription initiation site predicted for this operon in Psn (Gaffney et al., 1990). Moreover, the −10 and −35 promoter regions previously described in Psn (Gaffney et al., 1990) were found upstream of the +1 site identified for the iaaM-1/iaaH-1 operon in Psv NCPPB 3335. Taking into account that the iaaM-2 gene has been shown to be a pseudogene (Ramos et al., 2012), the transcription start site of the iaaM-2/iaaH-2 operon was not identified in this study.
Production of IAA in Psv NCPPB 3335 mainly depends on the iaaM-1/iaaH-1 operon
To determine the individual roles of the iaaM/iaaH operons in the biosynthesis of IAA by Psv NCPPB 3335, single (ΔiaaMH-1 and ΔiaaMH-2) and double (ΔiaaMH-1.2) knockout mutants were constructed by gene replacement (Fig. 1c). Psv NCPPB 3335 and all three mutants were grown on LB medium to the stationary phase, and IAA was extracted from the culture supernatants and quantified by HPLC. The level of IAA in the culture supernatants of the ΔiaaMH-1 and ΔiaaMH-1.2 mutants was c. 40 times lower than the level in the wild-type-strain (5565 μg g−1 DW). However, a residual amount of IAA was found in the culture supernatants of these two strains. Moreover, no significant differences were found between the ΔiaaMH-2 mutant and the wild-type strain (Fig. 2), suggesting that the iaaM-2/iaaH-2 operon does not contribute to the pool of IAA produced by Psv NCPPB 3335. In fact, transcription analysis of the iaaM-2 gene performed by qRT-PCR indicated that this gene is not expressed in Psv cells grown in LB medium (data not shown). Furthermore, the residual amounts of IAA (120 and 167 μg g−1 DW) detected in cultures of ΔiaaMH-1 and ΔiaaMH-1.2, respectively, might indicate the existence of an additional IAA biosynthetic pathway in Psv NCPPB 3335. Redundancy in IAA biosynthetic pathways has been previously predicted in other microorganisms after the inactivation of a single pathway (Patten et al., 2012).
Contribution of the iaaM-1/iaaH-1 operon to the virulence and fitness of Psv NCPPB 3335 in olive plants
To determine the role of the Psv NCPPB 3335 operons in the virulence and fitness of this pathogen in olive plants, we analysed the ability of the single mutants (ΔiaaMH-1 and ΔiaaMH-2) and the double mutant (ΔiaaMH-1.2) to cause olive knot symptoms and maintain fitness in the olive plant. Fitness attenuation and knot size reduction have been commonly used to describe hypovirulent mutants in Psv (Rodríguez-Moreno et al., 2008; Matas et al., 2012). Mixed infections using the wild-type strain and each of the mutants were prepared to calculate the CI on in vitro olive plants at 30 d.p.i. The results showed that the CI values obtained for the wild-type strain and each of the ∆iaaMH-1 and ∆iaaMH-1.2 mutants were significantly lower than 1, indicating that these mutants were outcompeted by the wild-type strain in planta. Conversely, this value was not significantly different from one in the competition assay between the ∆iaaMH-2 mutant and the wild-type strain, revealing that the iaaM-2/iaaH-2 operon is not required for the full fitness of Psv NCPPB 3335 (Fig. 3).
All the strains were inoculated independently on in vitro-grown and 1-year-old olive plants, and disease severity was evaluated after 30 or 90 d.p.i., respectively. In both plant systems, the symptoms developed upon infection with the ΔiaaMH-1 and ΔiaaMH-1.2 mutants were visually less severe than those induced by the wild-type strain and the ΔiaaMH-2 mutant (Fig. 4a and b). Additionally, the knot volumes in olive plants infected with the ΔiaaMH-1 and ΔiaaMH-1.2 mutants were approximately eight times smaller (in vitro-grown olive plants, Fig. 4a, bottom) and 16 times smaller (lignified olive plants, Fig. 4b, bottom), respectively, of the knot volume of plants infected with the wild-type strain. Together, these results show that the iaaM-2/iaaH-2 operon encoded by Psv NCPPB 3335 does not contribute to IAA biosynthesis or the induction of symptoms in olive plants. Moreover, these results suggest that the redundancy of IAA biosynthetic genes in some Psv strains (Pérez-Martínez et al., 2008) does not affect the total amount of IAA produced or the full virulence and fitness of the pathogen in olive plants. Nevertheless, a comparative virulence analysis with other Psv strains and a functional analysis of their iaaM/iaaH operons is necessary to verify this hypothesis.
Exogenous IAA affects the expression of virulence-related genes in Psv
IAA has been described as a signalling molecule that affects the expression of virulence factors, such as the T3SS and T6SS, and IAA biosynthesis in several plant-associated bacteria (Spaepen & Vanderleyden, 2010; Van Puyvelde et al., 2011; Patten et al., 2012). Previous studies on Psv NCPPB 3335 demonstrated that the T3SS-related genes, hrpA and hrpL, are required for infection establishment and knot formation on olive plants (Pérez-Martinez et al., 2010; Matas et al., 2014). To investigate a possible link between IAA and the T3SS in Psv, we monitored expression of the hrpA and hrpL genes using transcriptional fusions of their corresponding promoter regions to the lacZ gene. Psv NCPPB 3335 cells grown to the exponential phase expressing each of these fusions were transferred to Hrp-inducing medium lacking IAA or containing 1 mM IAA. The expression of T3SS-related genes in P. syringae and related pathogens is usually analysed in this medium, which is widely believed to mimic the apoplastic environment (Rico & Preston, 2008). However, the production of IAA by Psv NCPPB 3335 under these conditions was negligible (data not shown). Thus, a comparison of the expression of the T3SS promoters in the wild-type strain vs. the Δiaa mutants could not be performed. To circumvent this limitation, we analysed the expression of the promoters in the wild-type strain transferred to Hrp-inducing medium lacking IAA or supplemented with IAA. In agreement with expression analysis previously reported for other Psv genes (Matas et al., 2014), the highest differences were observed 6 h after transfer to Hrp-inducing medium. In the presence of IAA, the expression levels of hrpA and hrpL were c. 77 and seven times lower, respectively, than those in the absence of this compound, indicating that IAA had a negative effect on the expression of the T3SS genes (Fig. 5a). Similar differences in the β-galactosidase activities of these promoters were observed after 24 h (data not shown). Additionally, qRT-PCR experiments were carried out under the same culture conditions used for the β-galactosidase assays. In these assays, we also analysed the expression of hrpA and hrpL, as well as vgrG, a gene related to the T6SS, and iaaL, a gene involved in the biosynthesis of IAA-Lys in Psn (Glass & Kosuge, 1988), which contains an hrp box promoter sequence recognized by the alternative sigma factor HrpL (Fouts et al., 2002). In the presence of IAA, and in agreement with the β-galactosidase assays (Fig. 5a), the total amount of hrpA and hrpL transcripts was reduced 40- and 25-fold, respectively, compared with their levels in the absence of IAA (Fig. 5b). Interestingly, the level of expression of the iaaL gene was similar in the presence and absence of IAA (Fig. 5b). Taking into account the possible dependency of the expression of this gene on HrpL (Fouts et al., 2002), the low level of transcripts of the hrpL gene observed in the presence of IAA does not seem to affect the levels of iaaL transcripts. Nevertheless, the expression and function of the iaaL gene remain to be investigated in Psv. Conversely, the transcript levels of vgrG doubled in the presence of IAA, indicating a positive effect of IAA on the expression of the T6SS in Psv. A similar effect on the transcript levels of T6SS genes has been reported in Azospirillum brasilense by the addition of exogenous IAA (Van Puyvelde et al., 2011).
The phytohormone IAA plays an essential role in the virulence of bacterial phytopathogens, including the olive pathogen Psv. Here we show that the biosynthesis of IAA in this pathogen mainly depends on one of its two chromosomally encoded iaaM/iaaH operons (iaaM-1/iaaH-1), a gene cluster horizontally transferred among bacteria belonging to the P. syringae complex. We also demonstrate that the full fitness and virulence of Psv in olive plants depends on the functionality of this operon. Moreover, we show that IAA acts as a signalling molecule in Psv, affecting the expression of other virulence-related genes such as the T3SS and the T6SS.
This research was supported by the Spanish Plan Nacional I+D+I grants AGL2011-30343-CO2-01 and AGL-2012-39923-CO2-02, as well as grants ref. P08-CVI-03475 and P10-AGR-05797 from the Junta de Andalucía. I.M.A. was supported by an FPU fellowship (Ministerio de Ciencia e Innovación/Ministerio de Economía y Competitividad, Spain). We thank Manuela Vega (SCAI, UMA) for assistance with the image analysis of knot volumes. We are grateful to A. Barceló (IFAPA, Churriana, Spain) and J. M. Sanchez Pulido (CSIC-La Mayora, Málaga, Spain) for the plant material. M. Castillo-Lizardo is acknowledged for help with plasmid constructions. Pablo García is gratefully acknowledged for technical help. We thank the SCIC (Universitat Jaume I) for its technical support.