Plant-derived nitric oxide (NO) triggers defence, priming the onset of the hypersensitive response and restricting pathogen ingress during incompatibility. However, little is known about the role of pathogen-produced NO during pre-infection development and infection. We sought evidence for NO production by the rice blast fungus during early infection.
NO production was measured using fluorescence of DAR-4M and the role of NO assessed using NO scavengers. The synthesis of NO was investigated by targeted knockout of genes potentially involved in NO synthesis, including nitric oxide synthase-like genes (NOL2 and NOL3) and nitrate (NIA1) and nitrite reductase (NII1), generating single and double Δnia1Δnii1, Δnia1Δnol3, and Δnol2Δnol3 mutants.
We demonstrate that Magnaporthe oryzae generates NO during germination and in early development. Removal of NO delays germling development and reduces disease lesion numbers. NO is not generated by the candidate proteins tested, nor by other arginine-dependent NO systems, by polyamine oxidase activity or non-enzymatically by low pH. Furthermore, we show that, while NIA1 and NII1 are essential for nitrate assimilation, NIA1, NII1, NOL2 and NOL3 are all dispensable for pathogenicity.
Development of M. oryzae and initiation of infection are critically dependent on fungal NO synthesis, but its mode of generation remains obscure.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Nitric oxide (NO) is a free radical gas that can diffuse rapidly through biological membranes, allowing it to act as a transient, local, intra- and intercellular signalling molecule (Ignarro et al., 1987; Palmer et al., 1987). In mammals it is a pivotal messenger in the immune, nervous and cardiovascular systems (Anbar, 1995; Grisham et al., 1999; Pfeiffer et al., 1999; Lundberg et al., 2008), while in plants it has been implicated in several processes, including germination and leaf and lateral root development, but has been most extensively studied in abiotic stress responses and plant immunity (Besson-Bard et al., 2008; Wilson et al., 2008; Moreau et al., 2010; Gupta et al., 2011). Indeed, there is considerable evidence that plant-derived NO is important in initiating plant responses to pathogens or elicitors (Delledonne et al., 1998, 2001; Conrath et al., 2004; Van Baarlen et al., 2004; Zeier et al., 2004; Prats et al., 2005; Zaninotto et al., 2006; Floryszak-Wieczorek et al., 2007). Evidence is also emerging that NO is an important regulatory molecule in fungi, including plant pathogens, although there are few papers published, and these are spread over a wide range of different species and developmental stages. Thus, NO influences germination in Colletrotrichum coccodes (Wang & Higgins, 2005), conidiation in Coniothyrium minitans (Gong et al., 2007) and sporangiophore development in Phycomyces blakesleeanus (Maier et al., 2001), and affects the formation of the appressorium in the obligate biotrophic powdery mildew fungus Blumeria graminis (Prats et al., 2008).
This presents an interesting challenge: fungi may use NO as a signalling molecule to control development, but, concurrently, NO may prime the host plant and activate defence. Interestingly, some degree of cross-talk between plant host and fungal NO signalling systems has been reported in the necrotrophic fungus Botrytis cinerea (Turrion-Gomez & Benito, 2011), suggesting the non-cell autonomous activity of NO provides potential for complex interplay in species interactions.
Most data in fungal systems are derived from NO measurements in vivo or through exogenous application of mammalian nitric oxide synthase (NOS) inhibitors, NO scavengers or NO donors (Gong et al., 2007; (Conrath et al., 2004; Prats et al., 2008; Turrion-Gomez & Benito, 2011). However, the mechanism of NO synthesis has not yet been described in fungi, and, thus far, it is unclear from analysis of published genomes that sequences homologous to the canonical mammalian nitric oxide synthesis (mNOS) enzymes are present. By analogy with plants, where there is similar controversy regarding the mechanism of NO synthesis, there may be a number of different routes for NO formation, including both oxidative and reductive pathways.
The dominant oxidative NO-synthesis pathway in mammals is through oxidation of L-arginine to give NO and citrulline, using NADPH and O2, by varying isoforms of mNOS, although evidence is also emerging for a reductive pathway from nitrite under low oxygen tensions (Lundberg et al., 2008). Structurally, mNOSs operate as homodimers with an N-terminal oxygenase with binding sites for L-arginine, haem and BH4 (tetrahydrobiopterin), linked, by a short calmodulin binding hinge that confers calcium sensitivity, to a C-terminal reductase domain with binding sites for flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN) and NADPH, which shows some similarity to cytochrome p450 reductases (Alderton et al., 2001; Gorren & Mayer, 2007).
A similar mammalian-like NOS activity in plants was reported initially, but has not been substantiated, as the original putative NOS candidate, Arabidopsis AtNOA1 (Guo et al., 2003), was subsequently shown to be indirectly associated with NO production, and not directly involved in NO synthesis (Moreau et al., 2008). Indeed, there are no genes in higher plant genomes with significant homology to the canonical mNOS enzymes. Nevertheless, there are several lines of indirect evidence that L-arginine-dependent NO generation can occur in plants, particularly in peroxisomes and plastids, even if it is not generated by mNOS activity (Corpas et al., 2004, 2009; Gas et al., 2009). A key diagnostic feature of these pathways is sensitivity to arginine-substrate analogues, such as L-NG-nitroarginine methyl ester (L-NAME).
In addition, a range of other oxidative, reductive and nonenzymatic NO-synthesis pathways have been proposed for plants, but with no clear consensus on their relative importance (Moreau et al., 2010; Gupta et al., 2011). For example, other potential oxidative NO-producing systems, in addition to those exploiting L-arginine, may use polyamines (Tun et al., 2006; Wimalasekera et al., 2011) or hydroxylamine (Rumer et al., 2009) as substrates. Meanwhile, the best-characterized reductive pathway involves NO formation from the reduction of nitrite by cytosolic nitrate reductase (Yamasaki et al., 1999; Yamasaki, 2000; Rockel et al., 2002). In particular, analysis of single and double nitrate reductase (NR) mutants in Arabidopsis reveals that nitrate reductase 1 (NIA1) alone is the source of NO during ABA signalling (Bright et al., 2006). Conversely, antisense nitrite reductase (NiR) tobacco (Nicotiana tabacum) plants show increased levels of nitrite and consequently increased levels of NO production (Morot-Gaudry-Talarmain et al., 2002). There is also evidence for NO production by a distinct plasma-membrane bound nitrite–NO reductase (NiNOR) activity (Stohr & Stremlau, 2006), the mitochondrial electron transport chain under anoxia (Planchet et al., 2005), or nonenzymatic reduction of nitrite during apoplastic acidification to pH 3–4 which can be enhanced by phenolics (Bethke et al., 2004) or reductants, such as ascorbate or glutathione (Yamasaki, 2000).
The situation in fungi is even more ambiguous than in plants as a consequence of the paucity of papers published to date. On the oxidative side, mNOS or NOS-like sequences have been alluded to in Aspergillus oryzae (Gorren & Mayer, 2007), Aspergillus spp. and Glomerella graminicola (Turrion-Gomez & Benito, 2011). On the reductive side, nitrate reductase and nitrite reductase genes are present in all filamentous fungal genomes analysed thus far, but their potential role in NO synthesis has not, hitherto, been addressed in fungi. The molecular identity of the other potential NO-synthesis pathways in fungi is unknown.
Here, we report evidence for production of NO by germinating conidia and during early development in the hemibiotrophic ascomycete Magnaporthe oryzae using fluorescent probes. This fungus is a devastating pathogen of rice (Oryza sativa) (Fisher et al., 2012) that attacks through formation of an appressorium, which develops within a few hours of germination at the tip of the germ tube distal from the conidium. The melanized appressorium allows the build-up of sufficient turgor pressure to drive entry into the host via a penetration peg (Bourett & Howard, 1990; Wilson & Talbot, 2009). This elaborate process is triggered by perception of host-derived cues, including a hard, hydrophobic surface, cutin monomers and low levels of nutrients (Ebbole, 2007; Skamnioti & Gurr, 2009; Wilson & Talbot, 2009), and is co-ordinated with cell-cycle progression and programmed cell death of cells in the conidium and germ tube (Veneault-Fourrey et al., 2006). The early stages of infection-related development, including formation of melanized appressoria, can be initiated on artificial hard hydrophobic surfaces, greatly facilitating chemical and genetic dissection of signal cascades involved in germling differentiation (Wilson & Talbot, 2009).
We demonstrate a regulatory role for NO during germination and appressorium formation, using DAR-4M fluorescence measurements and NO scavengers. Notably, NO scavengers delayed germination and early development on artificial surfaces and dramatically reduced lesion formation on barley (Hordeum vulgare). We tested likely NO-generating enzymes, by creating knockout strains of candidate genes. We revealed that neither nitrate nor nitrite reductase is responsible for NO generation, and that both are dispensable for pathogenicity on rice and barley. Likewise, knockout of candidate members of the most closely related mNOS-like gene family does not affect NO production or produce an obvious defect in pathogenicity. We show that NO is not produced by other arginine-dependent systems or polyamine oxidases in M. oryzae. We conclude that nitric oxide is a critical signalling molecule in early development and has a major impact on plant–pathogen interactions, but its mode of synthesis is unresolved.
Materials and Methods
Fungal strains and growth conditions
Wild-type rice-pathogenic Magnaporthe oryzae (M. grisea (T.T. Herbert) M.E. Barr) strain Guy11 and NHEJ Δku70 and mutant strains were cultured at 24°C, 14 h : 10 h, light: dark cycle. Strain maintenance and medium composition were as described by Talbot et al. (1993).
Growth and biomass assays
Plate growth assays assessed radial colony growth on complete medium (CM) or minimal medium (MM in the presence/absence of 300 mM potassium chlorate), inoculated with 20 μl of 2.5 × 105 conidia ml−1 harvested from 10-d-old cultures and incubated at 24°C for 10–14 d.
Fungal biomass was determined using 20 μl of inoculum (as above) in 20 ml of MM, dark-incubated at 24°C, and shaken at 150 rpm for 14 d. The cultures were filtered onto pre-dried, weighed glass microfibre papers (Whatman), oven-dried at 80°C and weighed. There was a minimum of three biological replicates per experiment, and two-tailed pairwise Student's t-test was used to assess the statistical significance of differences in growth.
Mutant strain generation
Guy11 and Δku70 strains were used in DNA-mediated protoplast transformation (Talbot et al., 1993). Putative transformants were selected on MM supplemented with 300 μg ml−1 hygromycin B (Calbiochem, Merck, Darmstadt, Germany) or defined complex medium (DCM) with 60 μg ml−1 Bialophos (Goldbio, St Louis, MO, USA), and subjected to PCR to confirm the presence of the antibiotic resistance marker, its correct site of integration, and native gene replacement and to Southern blot analysis (in Supporting Information Methods S1, S2) to confirm single targeted gene replacement (Fig. S5). PCR primers used to generate mutant strains are detailed in Table S2. Standard techniques (Ausubel et al., 1999) were used to prepare constructs; details of the generation of single Δnia1, Δnii1, Δnol2, Δnol3 and double Δnia1Δnii1, Δnol2Δnol3, Δnia1Δnol3 strains are given in Methods S1.
RNA was extracted from Guy11 harvested at 0, 0.5, 1, 2, 5, and 12 hours post inoculation (hpi) from detached barley epidermal peels (Skamnioti & Gurr, 2007). First-strand cDNA was synthesized from total RNA using the RETROscript First Strand kit (Ambion, Applied Biosystems, Paisley, UK). RT-PCR was performed on cDNAs, with primers summarized in Table S3, for nitric oxide synthase-like 1 (NOL1) (P24 and P25), NOL2 (P26 and P27), NOL3 (P28 and P29) and NOL4 (P30 and P31). The transcript abundance of NOL genes, relative to constitutively expressed normalizer genes, β-tubulin (MGG_00604, P32 and P33) or ElongationFactor-1α (MGG_03641, P34 and P35), was quantified, using the Pfaffl method (Pfaffl, 2001), taking account of primer efficiencies, and calibrated to expression at 1 hpi.
Real-time quantification was performed in MicroAmp Optical 96-Well Reaction Plates using the 7300 Real-Time PCR System (Applied Biosystems). PCR conditions were: 50°C for 2 min, one cycle; 95°C for 10 min, one cycle; 15 s at 95°C, followed by 1 min at 60°C, 40 cycles. Reactions with no cDNA monitored for the presence of primer dimers and no reverse transcriptase controls were included for each cDNA sample. PCRs were carried out in triplicate and mean values determined.
Spores of Guy11 (50 μl; 2.5 × 105 ml−1) were inoculated onto hydrophobic glass slides and germinated in the presence of 2 μM DAR-4M-AM. Samples were viewed using the C-Apochromat ×40/1.2 water immersion lens of a Zeiss LSM 510 Meta microscope, with excitation at 543 nm from a HeNe laser attenuated to 6 μW at the objective, and emission at 590 ± 25nm. Simultaneous nonconfocal transmission 4-D (x,y,z,t) images were collected with a pixel spacing of 0.23 μm × 0.23 μm × 3 μm as z-stacks of 9–12 optical sections, repeated at 60 s intervals, for up to 120 time-points. Images were smoothed with a 3 × 3 × 3 kernel and displayed as maximum projections along the z-axis over time for fluorescence signals and minimum projections for the simultaneous (nonconfocal) transmission images. Images were analysed using a custom software suite written in MatLab (The Mathworks, Natick, MA, USA), available from MF.
Fluorescent plate reader assay
NO production was measured using the NO-sensitive fluorescent dyes DAR-4M (non-cell permeable) and DAR-4M-AM (cell permeable) in a FLUOstar Galaxy (BMG Labtech, Aylesbury, UK) fluorescence plate reader using NUNC 96-well optical bottom plates (Thermo Fisher Scientific, Langenselbold, Germany).
Conidia were harvested from 10-d cultures, washed via centrifugation and re-suspended in demineralized water three times (to remove extracellular esterases). 1 mM DAR-4M (AM) stock in dimethyl sulphoxide (DMSO) were diluted to 2 μM DAR-4M (AM) in 10 mM HEPES, pH 7, on ice. Suspensions were dark-incubated for 30 min at room temperature to allow dye loading, washed twice and re-suspended in 10 mM HEPES, pH 7, and the spore concentration was adjusted to 2.5 × 105 spores ml−1. Two hundred microlitres of conidia suspension was inoculated into each well and fluorescence (λex = 544 nm; λem = 590 nm) was recorded for 12–16 h at 20°C, unless otherwise stated. Each experiment contained a minimum of three biological replicates and was replicated independently on at least three separate occasions. 4,4,5,5-Tetramethylimidazoline -l-oxyl3-oxide (PTIO) or carboxy-PTIO (cPTIO) was added, as described in the figure legends. A significant instantaneous drop in fluorescence was observed with increasing concentrations of PTIO, caused by an absorption or quenching effect of PTIO on the DAR-4M triazole (DAR-4M-T) fluorescence signal (Fig. S1a). The quench magnitude was estimated from the instantaneous drop at the start of each experiment, or by adding PTIO at the end of the time-course (Fig. S1b). The concentration-dependent quench response was fitted with a mono-exponential curve (Fig. S1c), inverted and used as a concentration-dependent PTIO correction factor.
Pathogenicity and infection-related morphogenesis assays
Germling and appressorium development was assessed at 1, 2, 4, 8 and 16 hpi by following differentiation on hydrophobic glass cover-slips (Gerhard Menzel, Glasbearbeitungswerk GmbH & Co., Braunschweig, Germany). One hundred and twenty germlings were counted in three independent experiments.
Cuticle penetration was assessed by scoring the frequency with which appressoria formed penetration pegs and intracellular infection hyphae on onion epidermis, after incubation at 24 hpi at 24°C. One hundred germlings were counted in three independent experiments.
Detached leaf and whole-plant barley (Hordeum vulgare L.) and rice (Oryza sativa L.) leaf infection assays are detailed in Methods S3. To test the effect of NO scavenger PTIO on host lesion development, 2.5 × 105 spores ml−1 were re-suspended in 0, 250 or 500 μM PTIO, with the addition of 0.2% (w/v) gelatine, and spray-inoculated onto cut barley leaves.
Iterative hidden Markov model searches
Iterative hidden Markov model searches were performed in the search for NOL sequences (Kelly, 2011). The protein family (PFAM) (Finn et al., 2010) seed domain for nitric oxide synthase (PF02898) was converted to a hidden Markov model (HMM) and used to search 937 fully sequenced genomes (Table S1), using the Hmmer program (Eddy, 1998). The hits were filtered based on an e-value threshold of 1 × 10−10 and aligned using mafft (Katoh et al., 2005). Columns that contained > 50% gaps were removed to prevent species-specific or clade-specific amino acid insertions biasing the models (Collingridge & Kelly, 2012). The gap-parsed alignments were re-parsed for > 95% identity to other sequence within the alignment, to prevent biasing of HMM towards any particular group of organisms, which may be overrepresented as a result of the presence of paralogues or uneven taxon sampling. The gap- and identity-parsed alignment was used to generate the HMM for the next search, being terminated when no further hits passing the e-value threshold were identified. The final sets were clustered on the basis of all pairwise BLAST similarity scores using DendroBLAST (S. Kelly & P. K. Maini, unpublished). Hidden Markov model searches were performed using either M. oryzae nitrate reductase (MGG_06062) or nitrite reductase (MGG_00634) sequences and aligned as described.
Phylogenetic tree inference
Sequences were aligned using MergeAlign-91 (Collingridge & Kelly, 2012) and trees constructed using a 100 bootstrap maximum likelihood, inferred with RAxML (Stamatakis, 2006), employing the LG model of sequence evolution (Le & Gascuel, 2008) and CAT rate heterogeneity (fixed number of rate categories). Fifty per cent majority-rule consensus trees were calculated from the 100 bootstrap replicates using the python module dendropy (Sukumaran & Holder, 2010).
Results and Discussion
NO is produced during germling development in M. oryzae
A number of different techniques are available to monitor NO production (Wardman, 2007; Vandelle & Delledonne, 2008; Nagano, 2009; Mur et al., 2011), although measurements are challenging as NO is active at low concentrations, has a high diffusion coefficient, and exists transiently in the cell environment before it reacts to give NO2, N2O3, N2O4, N2O, HNO, peroxynitrite or GSNO (Brown et al., 2009; Baudouin, 2011). We chose to use in vivo fluorescent assays to measure intracellular synthesis rates in M. oryzae as these are highly sensitive to NO and can be imaged at the cellular level. A range of fluorescent probes developed by Nagano and co-workers react with N2O3, an auto-oxidation product of NO, to give a fluorescent triazole product (Kojima et al., 1998, 1999, 2001). Of the commercially available probes, DAR-4M is reported to have greater specificity for NO, improved photostability, reduced pH sensitivity, and lower cytotoxicity in comparison to the earlier fluorescein-based probes (Kojima et al., 2001; Lacza et al., 2005, 2006), and was selected for use in M. oryzae.
Conidia were left to germinate on a hydrophobic coverslip for 0.5 h post inoculation to ensure that they were adherent, loaded with DAR-4M as the membrane-permeant acetoxymethyl (AM) ester derivative, and imaged using 4-D (x,y,z,t) confocal microscopy at different stages of development (Fig. 1a–f). Fluorescent DAR-4M-T increased steadily over 2 hpi during germination of the apical cell, but reached a peak and then declined in the mid and basal cell (Fig. 1c,e). Furthermore, if the DAR-4M-AM loading solution was replaced by perfusion, there was a relatively rapid partial loss of signal over 5 min from cells and background, and the subsequent rate of fluorescence increase was reduced (Fig. 1c,e). At later stages of development, during appressorium formation (Fig. 1b), signal loss was also observed from the apical cell cytoplasm, which was matched, to some degree, by increased labelling in the cell walls (Fig. 1d,f). Perfusion also reduced the signal from all three cells and the cell wall (Fig. 1 d,f). We infer that either the DAR-4M-T is sufficiently membrane-permeant to diffuse out of the cells, or that it is actively transported by an unknown plasma-membrane xenobiotic detoxification system in M. oryzae. The overall increase in fluorescence is consistent with NO production in developing conidia of M. oryzae. It is, however, challenging to make quantitative measurements of NO production as the dynamics of fluorophore localization are more complex.
To allow high-throughput measurements with multiple treatment conditions over an extended time-course, the DAR-4M fluorescence assay was adapted to a 96-well plate format. Conidia were competent to germinate in optical-bottom well plates and the majority (c. 60%) formed melanized appressoria (Fig. 1g), although the germ tubes were typically slightly longer than those grown on inductive glass coverslips (compare Fig. 1g with Fig. 1b). None of the commercially available 96-well plates screened were fully inductive for M. oryzae; nevertheless, sufficient germlings progressed through to the appressorial stage to allow measurements of NO formation during early development. The first measurement, taken 30–60 min after the start of loading with DAR-4M-AM, reflected the time required to wash out excess DAR-4M-AM and set up the plate. The fluorescence showed a small initial peak, then a plateau relative to the control without dye, over the first 180 min, then increased almost linearly for the next 16 h (Fig. 1h).
The transient response observed in the first few hours is somewhat unexpected, as formation of triazole is irreversible, so the fluorescence signal should only increase (or remain constant) over time, rather than decrease. However, consistent with the observation of a decrease in signal following perfusion in the confocal imaging, we hypothesized that the fluorescent product might be released from the conidia into the medium where the detection efficiency for dye was lower compared with that in the germling adhered to the base of the well. External release of the triazole was confirmed in the plate reader system, as replacement of the buffer caused a c. 30% decrease in fluorescence (Fig. 1i). We confirmed that there was still an excess of active dye present in the germlings at the end of the experiment through the addition of the NO donor detanonoate ((Z)-1-[2-(2-Aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate, 3,3-Bis(aminoethyl)-1-hydroxy-2-oxo-1-triazene, 2,2′-(Hydroxynitrosohydrazino)bis-ethanamine) at 16 h (Fig. 1j).
NO scavengers have a complex effect on fluorescent NO measurements
The increase in fluorescence is consistent with the production of NO during germination and early development in M. oryzae. However, DAR-4M and other diamine probes can react with other molecules, such as dehydroascorbate (Nagata et al., 1999; Zhang et al., 2002; Ye et al., 2008), so increased fluorescence cannot be unequivocally attributed to NO production without additional supporting evidence. We therefore used the NO scavengers PTIO and cPTIO (Akaike et al., 1993) to deplete levels of NO by oxidizing it to NO2 (Eqn ). PTIO is more lipophilic than cPTIO and might be expected to permeate the plasma membrane more readily, while cPTIO is regarded as being more reactive (Akaike et al., 1993; Nakatsubo et al., 1998). At low scavenger concentrations, the rate of NO2 formation by (c)PTIO does not immediately consume all available NO, leading to a situation where both NO and NO2 are present at comparable concentrations and are able to react to form N2O3 (Eqn ). As N2O3 is the substrate for the diamine probes (Eqn ), this gives a characteristic stimulation of triazole fluorescence at low PTIO concentrations (Nakatsubo et al., 1998; Vitecek et al., 2008; Mur et al., 2011). At higher PTIO concentrations, all NO is rapidly converted to NO2 and the decrease in fluorescence expected for an NO scavenger is observed.
When conidia loaded with DAR-4M-AM were exposed to increasing concentrations of (c)PTIO, increases in fluorescence were observed with a maximum around 5 μM for PTIO (Fig. 2a) or cPTIO (Fig. 2b), consistent with Eqns . Higher (c)PTIO concentrations caused a decrease in fluorescence (Fig. 2a,b) once the data were corrected for absorption or quenching on the DAR-4M-T fluorescence signal (see Materials and Methods; Fig. S1). The varying impact of PTIO was summarized by integrating the area of the curve between the control in the absence of PTIO and increasing concentrations of PTIO over the first 8 h. PTIO and cPTIO gave c. 60% and 30% stimulation at 5 μM compared with controls, and c. 60% inhibition at 250 μM. There was a component of the increase in fluorescence that was not inhibited by PTIO, even at high concentrations, which may therefore be attributable to reaction with other molecules, such as the fungal antioxidant erythroascorbate (e.g. Georgiou & Petropoulou, 2001; Baroja-Mazo et al., 2005), leading to a fluorescence product (Zhang et al., 2002). The difference between the maximum fluorescence observed and the inhibition with 250 μM PTIO was regarded as the PTIO-sensitive component of the fluorescence signal that is most likely to be specific for NO. This PTIO-sensitive component rose to a peak around 1 hpi that was maintained for c. 4 h before declining towards the baseline (Fig. 2d).
To determine whether NO produced within the cells was detectable externally, we repeated the PTIO titration in the presence of cell-impermeant fluorophore DAR-4M (Fig. 2e). A small stimulation of fluorescence was observed at PTIO concentrations up to 5–10 μM over the first 3–4 h in the quench-corrected data, while higher concentrations gave the expected reduction in fluorescence. The overall response tended towards a plateau at 6h in the absence of PTIO (Fig. 2e), whereas signal from DAR-4M released internally from hydrolysis of DAR-4M-AM continued to show an increase throughout the time series (Fig. 2a,b). The absolute magnitude of the fluorescent signal from external DAR-4M was 3–4-fold higher than that observed for internal DAR-4M (compare Fig. 2a, b with Fig. 2e). The external PTIO-sensitive component increased more slowly, reaching a peak at around 5 hpi before declining (Fig. 2f).
The most parsimonious explanation for the distinctive fluorescence profiles observed in response to PTIO would be by stimulation of N2O3 production, as predicted, at low concentrations of PTIO, when stoichiometric amounts of NO2 are likely to be produced, followed by inhibition at higher PTIO concentrations. Taken together, these measurements can be used as a diagnostic indicator for NO production, rather than reaction with other biomolecules, as the PTIO-sensitive component is most likely to be specific for NO. We infer, therefore, that germinating spores produce NO. We found a stronger stimulation response and at lower concentrations with PTIO compared with cPTIO. As the former is more hydrophobic, this would be consistent with an internal source of NO. However, the absolute magnitude of the fluorescence signal is greater in the presence of a cell-impermeant form of the dye, suggesting that considerable amounts of NO either diffuse out of the germling or are synthesized externally. Interestingly, intracellular NO detection continues to increase during development, while external detection tends to plateau around the time at which melanized appressoria appear.
NO scavengers delay early development in M. oryzae
To investigate the significance of NO production by M. oryzae during early development, we quantified the impact of the NO scavenger PTIO on germination, germ tube elongation and appressorium formation on a surface inductive to formation of fully melanized appressoria in wild-type M. oryzae. Typically, c. 60% of untreated spores germinated by 1 hpi, and nearly 100% by 2 hpi (Fig. 3a). Under normal conditions, germ tube elongation proceeded rapidly over the next 2–3 h, with c. 50% of germlings starting to form appressoria within 4 hpi, which became fully melanized by 8 hpi (Fig. 3a). The addition of 10 μM PTIO was sufficient to cause a 60% reduction in germ tube emergence at 1 hpi, but by 4 hpi, the developmental profile of these germlings had almost recovered to the control, untreated levels. Increasing concentrations of PTIO caused longer delays in germination and slowed progression through the developmental pathway. Thus, while 95% of spores had germinated at 8 hpi in the presence of 200 μM PTIO, only 10% had formed melanized appressoria (Fig. 3a). If the primary mode of action of PTIO is to deplete NO, we infer that endogenous NO is required as part of the normal developmental sequence, possibly to initiate germination in contact with an inductive surface and/or to co-ordinate subsequent development between the different germling cell types.
Depletion of NO produced by M. oryzae abolishes pathogenicity on barley
We asked whether the delay in germination and development observed with NO scavengers on an artificial surfaces was manifest in vivo during M. oryzae infection of a susceptible host and would reduce the level of infection in this compatible plant–pathogen interaction. Spores were sprayed onto barley leaves in the presence and absence of 250 and 500 μM PTIO, and the number of lesions scored after 5 d. There was a significant reduction in the number of lesions in the presence of PTIO (Fig. 3b,c), so providing evidence that pathogen-derived NO plays an important role in the infection process and is required for successful host colonization. By contrast, the literature has focussed largely on the impact of PTIO in plant-derived NO signalling during incompatibility in other plant–pathogen interactions. Plant-derived NO production is known to stimulate plant defence reactions during the hypersensitive response (HR) and race-specific resistance (Hong et al., 2008). For example, PTIO scavenging of plant-derived NO led to increased penetration frequencies by the biotroph Blumeria graminis and to a reduced host response on a barley isoline manifesting HR (Prats et al., 2005). Thus, production of fungal NO in the context of a compatible interaction would seem counterintuitive, as NO might be expected to prime host defence. Nevertheless, Prats et al. (2008) demonstrated a pivotal role for a transient burst of NO in appressorium maturation in B. graminis, that is, during pathogen development before host penetration, but probably at a stage that is too early to prime defence. This interplay between a biotrophic pathogen, showing extreme host specificity, differs markedly from the exchange between the necrotroph Botrytis cinerea and its broad range of hosts, where HR plays a critical role in fungal infection (Govrin & Levine, 2000). Here, NO facilitates fungal infection, but is also required for plant defence. Indeed, Floryszak-Wieczorek et al. (2007) recorded a strong and immediate host NO burst by a resistant plant cultivar upon challenge with B. cinerea, but a weak and slower burst in a susceptible cultivar upon infection. Clearly, the role of plant-derived NO is complex and varies with host response, while the role of pathogen-derived NO is important to infection-related development.
Genetic approaches to characterizing the mechanism of NO production in M. oryzae
A number of different pathways have been invoked for NO production in animals and plants, including NOS (Gorren & Mayer, 2007), NOS-like enzymes (Corpas et al., 2009), nitrate reductase (Rockel et al., 2002), NiNOR (Stohr & Stremlau, 2006) and polyamine oxidases (Tun et al., 2006; Yamasaki & Cohen, 2006). We therefore set out to systematically test for the presence of each of these systems in M. oryzae.
NOS and NOS-like enzymes as potential generators of fungal NO
To determine whether fungi have homologues of the canonical metazoan nitric oxide synthase gene, we performed an iterative hidden Markov model search (Eddy, 1998; S. Kelly, unpublished) with PFAM seed alignment of the nitric oxide synthase domain (PF02898) passed 385 times over 937 completed metazoan, plant, fungal, eubacterial and archaebacterial genomes (Table S1; e-value cut-off of 1 × 10−10). These sequences were clustered based on their pairwise BLAST similarity scores (DendroBLAST; Fig. S2) and analysed for the presence of domains necessary for NO production. This revealed six clusters of sequences (Fig. 4a).
In addition to NOS sequences identified in the green alga Ostreococcus sp. (Foresi et al., 2010), the slime mould Physarum polycephalum (Werner-Felmayer et al., 1994; Golderer et al., 2001; Messner et al., 2009) and possibly Aspergillus oryzae (acc XP-001825673), we identified three further sequences in the ascomycetes Colletrotrichum graminicola (acc 10854T0) and Mycosphaerella graminicola (acc 42401 and 28714, short sequence). These fungi formed a strongly supported group separate from the amoeba and metazoan, but within the NOS cluster (Fig. S3). Interestingly, the M. graminicola sequence (42401) and A. oryzae sequence carry a reduced complement of residues characteristic of the arginine binding pocket in the mNOS oxygenase domain (Fig. 4b). The presence of this sequence in this monophyletic subgroup of fungi suggests that their common ancestor acquired this by lateral gene transfer. However, the tree is insufficiently resolved to identify the donor organism.
The iterative search revealed that there are no sequences in the M. oryzae genome that contain an NOS oxygenase domain (Fig. 4b), although multiple groups of sequences contain the reductase-associated domains found in canonical NOS proteins (Fig. S2). However, all these groups, with one exception, contain multiple members that have been functionally characterized in metazoa, yeast or both, and shown not to be NOSs (Fig. 4a, Fig. S2). It is highly unlikely that M. oryzae sequences lying in these groups are NOS.
The only group that does not have functionally characterized homologues in metazoa or amoebazoa contains a domain structure that could be consistent with NO synthesis. This fungal-specific group is composed of proteins, labelled as putative bifunctional p450: NADPH-P450 reductases, each of which contains an N-terminal cytochrome p450 domain and FAD, NAD and flavodoxin binding domains. Members represent the best candidates for oxidative production of NO in M. oryzae by an NOS-like mechanism. Within the fungal NOL cluster, we refer to the genes as NOL1 (MGG_01925), NOL2 (MGG_05401), NOL3 (MGG_07953) and NOL4 (MGG_10879).
Selection of candidate NOL genes
Of the four NOL genes, qRT-PCR, normalized, independently, against β-tubulin and ElongationFactor-1α (Fig. 4c,d), suggested that the NOL2 transcript was 4–5 times more abundant at 0.5–1 hpi (germination) compared with 0 hpi, and NOL3 showed a 24–28-fold uplift in transcript activity at 12 hpi, coincident with mature appressorium formation and initiation of host infection. These two genes were targeted as the most likely candidates for NOS activity in M. oryzae. To test for NOS activity directly, we expressed NOL3 by heterologous expression in Escherichia coli and Pichea pastoris. However, we were not able to purify the protein further. Crude cell extracts did not show NOS enzymatic activity by ultrasensitive colorimetric NOS assay (Oxford Biomedical Research Inc., Rochester Hills, MI, USA), with L-arginine as the substrate (data not shown). Furthermore, attempts to isolate interacting partners that might provide a canonical substrate binding site by yeast two-hybrid assay did not yield positive results, even after extensive optimization (data not shown). At this stage, therefore, the evidence that these are potential NOS enzymes is limited to sequence data.
Nitrate and nitrite reductases as potential generators of fungal NO
Given the relatively weak sequence homology of the NOL genes, we considered nitrate and nitrite reductases as potential sources of NO, by analogy with the plant systems. The domain architecture of nitrate reductase includes binding domains for a molybdenum cofactor, cytochrome b5, FAD and NAD(P)H (Campbell & Kinghorn, 1990) and is closely conserved in all fungal sequences (Fig. S4a). This was revealed using the fungal sequences listed in Table S1, by hidden Markov model search (Eddy, 1998; S. Kelly, unpublished), using acc. MGG_06062, clustered by BLAST similarity scores (DendroBLAST) and by domain analysis. Likewise, there is a high degree of architectural conservation for nitrite reductase proteins (using acc. MGG_00634), with all members containing cysteine, FeS-siroheme, nitrite/sulphite reductase and ferrodoxin-like domains (Fig. S4b). Magnaporthe oryzae carries single copies of nitrate reductase (MGG_06062; NIA1) and nitrite reductase (MGG_00634; NII1), which were therefore selected as targets for gene knockout.
Analysis of mutants deficient in nitrate reductase, nitrite reductase and nitric oxide synthase-like enzymes
To test whether nitrate reductase, nitrite reductase or the two most abundant NOLs might contribute to NO production in M. oryzae, single (Δnia1, Δnii1, Δnol2, and Δnol3) and double (Δnia1Δnii1, Δnia1Δnol3, and Δnol2Δnol3) knockouts were constructed by homologous recombination in Δku70 background NHEJ strain, derived from M. oryzae Guy 11 (Wilson & Talbot, 2009). All knockouts were verified by PCR analysis with internal and flanking primers, and Southern blot analysis confirmed a single targeted replacement event (Fig. S5).
Plate growth assays revealed no differences in growth morphology or colony diameter among strains Guy11, Δku70, Δnia1, Δnii1, Δnia1Δnii1, Δnol2, Δnol3, Δnol2Δnol3, and Δnia1Δnol3 on CM (Fig. 5a). On MM (containing 70.6 mM sodium nitrate) the strains showed differential growth. This was quantified by biomass determination in liquid medium and showed a significant reduction in growth of Δnia1, Δnii1, Δnia1Δnii1 and Δnia1Δnol3 as compared with the wild type (Fig. 5b), as these strains were unable to use nitrate as the sole nitrogen source in the absence of NR. By contrast, only the NR-deficient strain Δnia1 survived exposure to 300 mM chlorate (Fig. 5c), which is metabolized to toxic chlorite in strains with functional NR (Cove, 1976).
If the gene knockout mutants impact on NO production, we might expect them to phenocopy the effects of PTIO on germination, early development and pathogenicity. However, all strains formed melanized appressoria on glass slides within 8 hpi (Fig. 6a), and were able to initiate penetration pegs, form invasive hyphae on onion epidermis (Fig. 6b), and make lesions in a cut leaf assay on both barley and rice susceptible to wild-type M. oryzae infection, with similar frequencies to the parental strains (Fig. 7). Some variation was noted in the detached rice leaf bioassay (Fig. 7c), so assays were repeated on intact plants, revealing no statistically significant differences between lesion numbers in wild-type and mutant strains (Δnol2Δnol3 P = 0.021; others P > 0.2; data not shown).
There was no significant difference in NO production in the mutants compared with the wild type, as determined by the concentration-dependent PTIO profile of triazole formation. All mutants produced NO with the characteristic stimulation at low concentrations of PTIO and inhibition at high concentrations (Fig. 8). Finally, we were not able to detect any inhibition in triazole formation with L-NAME in Guy11 or the Δnia1 and Δnii1 backgrounds (Fig. 9), or any difference in response compared with the inactive stereoisomer D-NAME up to 500 μM in Guy11. If anything, at higher (mM) concentrations, both compounds stimulated fluorescence from DAR-4M.
We infer that none of these putative NO-synthesizing enzymes are individually responsible for the observed NO production in M. oryzae strain Guy11. Furthermore, the absence of any strong phenotype in the double knockouts tested and in the presence of the general NOS inhibitor L-NAME further indicates the lack of functional redundancy in NO production between these pathways.
To test for the presence of polyamine oxidase activity that might produce NO (Tun et al., 2006; Wimalasekera et al., 2011), we analysed NO production in the presence of the polyamines spermine and spermidine (Fig. 10), but did not observe any increase in triazole fluorescence. In mammalian systems, the mNOS isozymes are stimulated by the substrate arginine and inhibited by arginine substrate analogues, such as L-NMMA and inhibitor 1400W (Garvey et al., 1997) which compete for the arginine binding site. We found no evidence for inhibition of NO production by L-NMMA, D-NMMA and 1400W or with the mNOS reaction product citrulline (Fig. 10), similar to the results obtained with L-NAME (Fig. 9). Finally, we do not consider that nonenzymatic NO production, akin to plant apoplastic NO synthesis (Bethke et al., 2004), is significant, as our measurements were conducted in medium buffered at pH 7.
In summary, NO is produced by M. oryzae during germination and early development, and is critically required to progress through to appressorium formation. Thus, removal of NO by NO scavengers slows down development on artificial surfaces and abolishes infection in vivo. The in vivo fluorescent assays provide some evidence that NO diffuses out of the germlings, at least until the formation of melanized appressoria. It is possible that the role of NO is to co-ordinate behaviour between different cells in the conidium which go on to have very different fates, as a non-cell autonomous signal. We have not investigated signalling events downstream from NO, but the targets are unlikely to include soluble guanylyl cyclase (sGS), as in mammals, as fungal genomes lack these sequences (Schaap, 2005). In other systems there is the potential for extensive chemical modification, particularly by S-nitrosylation or tyrosine nitration (Mur et al., 2006). These pathways are inextricably linked to the dynamics of other reactive oxygen species (ROS) (Winterbourn, 2008; Moreau et al., 2010). It is likely that there will be a complex spatial and temporal interplay between NO and ROS signalling systems, such as M. oryzae plasma-membrane NADPH oxidases (Egan et al., 2007), which are known to be critical determinants of infection.
Despite the importance of NO identified here, the synthesis pathway in M. oryzae is not clear. We found no evidence for NO production by reductive nitrate reductase or nitrite reductase pathways or oxidative NOS-like enzymes, arginine-dependent NO systems, polyamine oxidases, or by low pH, that is nonenzymatically. While investigations on NO in fungi are at an early stage, this cautions against uncritical adoption of an NO-signalling paradigm from either animal or plant systems. Indeed, phylogenetic comparisons provide limited support for canonical NOS or NOS-like enzymes across all fungi, with the possible exception of a few sequences that appear to reflect horizontal gene transfer events. The mechanism by which fungi generate NO remains elusive.
This work was funded by a BBSRC award to S.G. and M.F. to support M.S. (BB/G00207x/1) and studentship funding for J.J. and M.I. We are grateful to Luis Mur (Aberystwyth) who provided impetus for this work, Mick Kershaw (Exeter) for help with rice pathogenicity assays and Nick Talbot (Exeter) for his sage advice.