Silicon influences cytological and molecular events in compatible and incompatible rice-Magnaporthe grisea interactions

Physiological and Molecular Plant Pathology 66 (2005) 144–159 www.elsevier.com/locate/pmpp

Silicon influences cytological and molecular events in compatible and incompatible rice-Magnaporthe grisea interactions ´ . Rodrigues1, Wayne M. Jurick II, Lawrence E. Datnoff, Fabrı´cio A Jeffrey B. Jones, Jeffrey A. Rollins* Department of Plant Pathology, 1453 Fifield Hall, University of Florida, Gainesville, FL 32611-0680, USA Accepted 3 June 2005

Abstract Many monocot plants grown in soils amended with silicon (Si) exhibit increased levels of resistance to fungal diseases. Silicon-mediated resistance to rice blast disease, caused by Magnaporthe grisea, has been hypothesized to be the result of a mechanical barrier produced from Si polymerization in planta. The present study examined the cytological and molecular influences of Si amendment during genetically defined incompatible (resistant) and compatible (susceptible with no major resistance genes) interactions of rice with M. grisea. Differential accumulation of glucanase, peroxidase, and PR-1 transcripts were associated with limited colonization by M. grisea in epidermal cells of SiC plants of the susceptible cultivar M201. Katy, a resistant cultivar, responded to an avirulent race of M. grisea through the development of a hypersensitive response along with a strong induction of PR-1 and peroxidase transcripts independent of Si amendment. These findings support an active participation, distinct from single-gene-defined resistance, for Si in the defense of rice against M. grisea. While not discounting a physical barrier role for Si, new insights into a potential active role for this ubiquitous element during non-genetically defined rice-blast resistance is suggested. q 2005 Elsevier Ltd. All rights reserved. Keywords: Silicon; Autofluorescence; Lignin; Soluble phenolics; Cytoplasmic granulation; PR-proteins; Oryza sativa; Magnaporthe grisea

1. Introduction Rice (Oryza sativa L.) is one of the worlds most important food crops and blast disease, caused by the hemibiotrophic fungus Magnaporthe grisea (T. T. Hebert) Yaegashi & Udagawa) Barr (anamorph

Abbreviations: Si, Silicon; AVR, avirulence gene; R, resistance; LRR, leucine-rich repeat; HR, hypersensitive response; PAL, phenylalanine ammonia-lyase; hai, hours after inoculation; kb, kilo base pair; bp, base pairs; CHS, chalcone synthase; FDI, fungal development index; TSP, total soluble phenolics; LTGA, lignin-thioglycolic acid; chit, chitinases; Glu, b-1,3-glucanase; POX, peroxidase; l.w., lyophilized weight. * Corresponding author. Tel.: C1 352 392 3631!235; fax: C1 352 392 6532. E-mail addresses: [email protected] (F.A Rodrigues), [email protected] (W. M. Jurick), [email protected] (L.E. Datnoff), [email protected] (J.B. Jones), [email protected] (J.A. Rollins). 1 Present address: Department of Plant Pathology, Vic¸osa Federal University, Vic¸osa, Minas Gerais 36570-000, Brazil. 0885-5765/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.pmpp.2005.06.002

Pyricularia grisea (Cooke) Sacc.), is a major constraint to its production worldwide [44]. Rice blast disease follows a gene-for-gene interaction as proposed by Flor [15], in which a race of M. grisea expressing an avirulence gene (AVR) triggers the corresponding resistance (R) gene-mediated defense [7,33,54]. Two of the major blast R genes, Pi-ta and Pi-b, have been cloned and characterized at the molecular level. These genes are predicted to encode cytoplasmic proteins with a centrally located nucleotide binding site and a carboxy terminal leucinerich repeat (LRR) region [2,61]. The cultivar Katy contains the Pi-ta2 R gene or both Pi-ta and Pi-ta2 R genes [38] while the cultivar M201 has no known major or minor gene(s) for resistance to race IB-49 of M. grisea [53]. The hypersensitive response (HR), defined as a rapid and localized death of host cells to restrict pathogen colonization [17], is a well known hallmark for race-specific resistance of rice cultivars to M. grisea [21]. The HR is characterized by plasma membrane depolarization, detachment of plasma membrane from the epidermal cell wall, alteration of ion channel

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

activities, change in calcium homeostasis, degeneration of the nucleus, lack of cytoplasmic streaming, cytoplasmic granulation and protoplast collapse [17,31,60]. In addition to the above cellular events, other mechanisms of the rice defense response such as oxidative burst; formation of papillae; fortification of cell walls by phenolics and lignin deposition; and induction of phytoalexins and pathogenesis-related proteins [3,16,39, 45,52] are also triggered in the M. grisea AVR-Pi-ta-rice interaction [24,43]. Several defense-related PR-like genes encoding PR-1, b-1,3-glucanases (PR-2), chitinases (PR-3), peroxidases (PR-9), and phenylalanine ammonia-lyase (PAL) have been cloned from rice [39,56]. PR-1 is a reliable marker for resistance in many pathosystems including rice-M. grisea [34]. Chitinases and b-1,3-glucanases are capable of degrading chitin and b-1,3-glucan, respectively, in the hyphae of M. grisea [13]. Phenylalanine ammonia-lyase is the key enzyme in determining the rate of phenolic production through the phenylpropanoid pathway [25] while peroxidases participate in the biosynthesis of lignin [18]. The flavonoid biosynthetic gene, chalcone synthase, is involved in the production of many flavonoids such as the rice phytoalexin sakuranetin [30]. Autofluorescence of epidermal and mesophyll cells is another marker of rice defense against M. grisea attack. Koga [31] induced death of rice leaf cells by heat shock and sodium arsenite, a fatty acid synthesis inhibitor, and observed that the growth of M. grisea within the penetrated cell was not affected unless the dead cells exhibited strong autofluorescence. Several physical and biochemical studies have demonstrated that compounds accumulating in autofluorescent epidermal cells of barley and oat plants attacked by Erysiphe graminis f.sp. hordei have ultraviolet absorption and emission characteristics of phenolics [5,6,36]. Although the temporal and spatial production of these autofluorogen compounds appears to be affected by Si [5], its positive or negative effect in the production of phenolics depends on each pathosystem. Carver et al. [4] observed a strong autofluorescence of epidermal cells in contact with germ tubes of Blumeria graminis and an increase in phenylalanine ammonia-lyase activity. Although autofluorescence of oat epidermal cells infected by B. graminis appeared to be independent of Si accumulation, these two events coincided spatially [4]. The massive accumulation of Si within autofluorescent barley epidermal cells that responded to an avirulent race of B. graminis suggests that monosilicic acid forms complexes with organic hydroxy compounds present in those cells [31]. The objectives of this study were to investigate the effects of silicon amendment on these dynamic cytological processes and determine if they correlate with specific molecular processes in compatible and incompatible rice-M. grisea interactions.

145

2. Materials and methods 2.1. Growth of rice plants, silicon amendment, and inoculation with M. grisea Plastic pots (12 cm in diameter) were filled with 2 kg of peat Fafard No. 2 (soilless medium) (Conrad Fafard Inc.; Agawan, MA) and amended with calcium silicate slag (22% available soluble Si; Calcium Silicate Corp., Lake Harbor, FL) at the rates of 0 and 10 g potK1 five days before sowing [49,50]. The native Si concentration in the substrate was 4 mg Si lK1. Rice seeds were sown at the rate of eight seeds per pot and, at three days after emergence, each pot was thinned to two seedlings. Seedlings were then fertilized by adding 100 ml of a nutrient solution to each pot containing (in milligrams per kilogram of peat Fafard No. 2): 100 N, 300 P, 150 K, 85 Ca, 70 Mg, 40 S, 0.81 B, 1.33 Cu, 3.66 Mn, 0.15 Mo, and 4.00 Zn [49]. The nutrient solution was prepared using silicon-free water. A second application of this nutrient solution was made 25 days later. Iron deficiency was avoided by adding 10 ml of a solution containing 300 g of FeSO4 lK1 per pot after thinning. Plants were kept under flooded conditions until the end of the experiments. Plants from rice cultivars of Katy (P.I. 527707) and M201 (C.I. 9980) were inoculated at the time of emergence of the seventh leaf from the main tiller [35] with a suspension of 4!105 conidia mlK1 of race IB-49 of M. grisea (isolate 793). The rice cultivar M201 has no known major or minor genes for resistance to race IB-49 of M. grisea resulting in a compatible interaction [53]. Race IB-49 is avirulent on Katy because it carries the Pi-ta2 R gene, or both Pi-ta and Pi-ta2 R genes [38]. Katy also contains a tightly linked cluster of at least seven R genes that map in the same region as Pi-ta and Pi-ta2 [8,38]. The suspension of conidia was applied as a fine mist to the adaxial leaf blades of two plants per pot until runoff using an aerosol sprayer (Crown Spra-Tool; Fisher Scientific Co., Pittsburgh, PA). Gelatin (1%, w/v) was added to the sterile water used to prepare the conidial suspension to aid conidial adhesion to the leaf blades. The non-inoculated plants from each cultivar were sprayed with a solution of gelatin-sterile water (1%, w/v) until runoff. Inoculated and non-inoculated plants were covered with plastic bags and transferred to a growth chamber configured to the following conditions: constant temperature of 25 8C and photoperiod of 12 h dark and 12 h light with approximately 162 mE mK2 sK1 provided by coolwhite fluorescent lamps. Pots were placed in water-filled plastic trays to maintain the relative humidity inside the growth chamber at approximately 85% for the remainder of the experiment. Inoculated and non-inoculated plants were submitted to an initial 24 h dark period and 48 h after inoculation the plastic bags were removed from the plants.

146

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

2.2. Cytological observations A 2!2!7 factorial experiment, consisting of leaves collected from cultivars Katy and M201, amended (SiC) or not (SiK) with Si, at 0, 12, 24, 36, 48, 72, and 96 h after inoculation (hai) with M. grisea, was arranged in a completely randomized design. The experiment was conducted twice with three replicates of each treatment. A total of 80–100 leaf pieces, approximately 0.5 cm2 in size, were randomly collected from leaves of two plants per replication and treatment. Since rice susceptibility to blast is affected by leaf age [44], samples were collected only from the fourth, fifth, and sixth leaves in the main tiller of two plants per replication to insure the use of leaves of the same age among treatments. Leaf pieces were fixed and decolorized in boiling 95% ethanol (v/v) for approximately 15 min before being cleared for three weeks in saturated chloral hydrate solution (30 g mlK1) (SigmaAldrich, St. Louis, MO). Cleared leaf pieces were mounted adaxial side up on glass slides containing 2–3 drops of modified Hoyer’s mounting medium [11]. Fifty appressorial sites per replication and treatment were randomly examined in detail to determine the fungal development index (FDI) within the epidermal cell(s); the temporal outcome of M. grisea-rice cultivar interactions; and the number and intensity of browning of epidermal cells using a Leica Model DMR microscopy (Leica Microsystems Inc., Bannockburn, IL) equipped with differential interference contrast optics. The FDI within epidermal cell(s) of each appressorial site was determined according to Takahashi [58]. The FDI for each infected cell ranged from 0 to 4 where 0Zconidium has formed an appressorium, but the infection hyphae has not been observed within the epidermal cell, 0.5Zthe infection hyphae within the epidermal cell has a length shorter than the diameter of the appressorium, 1Zinfection hyphae has a length greater than two times the diameter of the appressorium, 2Zinfection hyphae has a length greater than five times the diameter of the appressorium, but without any branching, 3Zinfection hyphae has elongated within the epidermal cell forming a few branches, and 4Z fully developed infection hyphae within epidermal cell without extension to neighboring epidermal cells. A value greater than four corresponded to the sum of infection indexes observed in the first penetrated epidermal cell and secondarily colonized neighboring cells. FDI values at 50 independent sites were summed and the average FDI value from each time interval was plotted The temporal outcomes of M. grisea-rice cultivar interactions were grouped into three categories: category A—absence of infection hyphae within epidermal cell underlying the appressorium; category B—successful penetration (infection hyphae within the epidermal cell and absence of cytoplasmic granulation); and category C— successful penetration (infection hyphae within the epidermal cell associated with intense cytoplasmic granulation).

The frequency of each category of cellular reaction for the fifty appressorial sites examined per replication and treatment was calculated. The number and the intensity of browning of epidermal cells for the fifty appressorial sites observed were also evaluated. The intensity of browning of epidermal cells was grouped into three categories as described: category 1—cell wall of epidermal cells showing no browning; category 2—cell wall of epidermal cells palely yellowed; and category 3—whole-epidermal cells deeply browned with adjacent cells showing slightly brown cell wall. The autofluorescence of epidermal cell walls or the wholeepidermal cells of each appressorial site examined was recognized by incident fluorescence microscopy (Chroma Endow GFP filter set-excitation: 470–510 nm, dichroic beamsplitter: 495 nm, emission: 525–575 nm). Images of the details observed on the fifty appressorial sites examined per each replication and treatment were acquired digitally (Spot Insight Camera; Diagnostic Instruments Inc., Sterling Heights, MI) and further processed with the Spot Insight 3. 2 Software. 2.3. Determination of total soluble phenolics A 2!2!7 factorial experiment, consisting of leaves collected from cultivars Katy and M201, amended with (SiC) or without (SiK) Si, at 0, 12, 24, 36, 48, 72, and 96 hai with M. grisea, was arranged in a completely randomized design. The experiment was conducted twice with three replicates per treatment. The fourth, fifth, and sixth leaves from the main tiller of each of two plants were collected at the seven time-points after inoculation, freezedried (Labconco Corporation, Kansas City, Missouri) for 72 h, and then ground together into a fine powder in a mortar and pestle with liquid nitrogen. The fine powder was transferred to a plastic tube and stored at K80 8C until further analysis. A representative sample of 0.1 g of the fine powder material from each replication and treatment was transferred to an Ependorf tube, homogenized with 1.5 ml of 80% methanol, and extracted overnight on a rotary shaker (150 rpm) at room temperature. The homogenate solution was protected from light oxidation by covering the Ependorf tube with aluminum foil. The dark-green methanolic extract was centrifuged at 12,000!g for 5 min, the supernatant was transferred to a new Ependorf tube, and stored at K20 8C. The residue was kept at K20 8C for further determination of lignin and lignin-like phenolic polymers. The methods developed by Zieslin and Ben-Zaken [66] were used to analyze the total soluble phenolics with a few modifications. A volume of 150 ml of 0.25 N Folin and Ciocalteau’s Phenol reagent (Sigma-Aldrich, St. Louis, MO) was added to 150 ml of methanolic extract and the mixture was homogenized and kept at room temperature for 5 min. Next, 150 ml of 1 M Na2CO3 was added to the mixture, which was homogenized again and stood for 10 min at room temperature.

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

The mixture was further homogenized with 1 ml of distilled water and allowed to stand for 1 h at room temperature. The absorbance of a representative sample (500 ml) of the mixture from each replication and treatment was measured at 725 nm in a SmartSpec 3000 Spectrophotometer (Bio-Rad Laboratories Inc., Hercules, CA). Total soluble phenolics were expressed as mg of phenolics (in terms of catechol) per kg of dried leaf tissue. 2.4. Determination of lignin and lignin-like phenolic polymers A volume of 1.5 ml of sterile distilled water was added to the residue obtained after extraction of total soluble phenolics and, after homogenization, the mixture was centrifuged at 12,000!g for 5 min. The supernatant was discarded and the residue was left to dry at 65 8C overnight. The dried alcohol-insoluble residue, containing both true lignin and phenolic acids esterified to the cell walls, was used for determination of lignin according to the methods of Barber and Ride [1]. A volume of 1.5 ml of a 1:10 solution of thioglycolic acid (Sigma-Aldrich, St. Louis, MO) and 2 N HCl was added to the dried residue. The Ependorf tube was shaken gently to hydrate the residue and then placed in boiling water (approximately 100 8C) for 4 h. The tube was cooled on ice in a cold room (4 8C) for 10 min. The mixture was centrifuged at 12,000!g for 10 min, the supernatant was discarded, the precipitate was washed with 1.5 ml of sterile distilled water, and then centrifuged at 10,000!g for 10 min. After centrifugation, the supernatant was discarded, the precipitate was resuspended in 1.5 ml of 0.5 N NaOH, and the mixture was agitated in a rotary shaker (150 rpm) at room temperature overnight. The mixture was centrifuged at 10,000!g for 10 min and the supernatant was transferred to a new Ependorf tube. After adding 200 ml of concentrated HCl to the supernatant, the Ependorf tube was transferred to a cold room (4 8C) for 4 h to allow the lignin-thioglycolic acid (LTGA) derivatives to precipitate. Following centrifugation at 10,000!g for 10 min, the supernatant was discarded and the orange-brown precipitate was dissolved in 2 ml of 0.5 N NaOH. The absorbance of LTGA derivatives in the supernatant was measured at 280 nm in a SmartSpec 3000 Spectrophotometer (Bio-Rad Laboratories Inc., Hercules, CA). The concentration of LTGA derivatives was expressed as mg kgK1 of dried leaf tissue by using lignin alkali, 2-hydroxypropyl ether (Sigma-Aldrich, St. Louis, MO) as a standard. 2.5. Total RNA extraction, probe preparation, and autoradiography A 2!2!8 factorial experiment, consisting of leaves collected from cultivars Katy and M201, amended with (SiC) or without (SiK) Si, at 0, 12, 24, 36, 48, 60, 72, and 96 hai with M. grisea, was arranged in a completely

147

randomized design. The experiment was conducted twice with three replicates per treatment. The fourth, fifth, and sixth leaves from the main tiller of each of two plants per replication and treatment were collected at eight timepoints after inoculation, immediately frozen in liquid nitrogen, and then stored at K80 8C. Approximately 1 g of frozen leaf tissue was ground to a fine powder in a chilled mortar and pestle using liquid nitrogen. Total RNA was isolated with TRIzolw Reagent (Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s instructions. The RNA pellet was resuspended in 30 ml of deionized formamide (Fisher Scientific Co., Pittsburgh, PA), the concentration determined by UV absorption at 260 nm, and then stored at K80 8C. An aliquot of 15 mg of total RNA (10 ml) from each replication and treatment was added to 13 ml of a master solution (5 ml of DEPC water, 3 ml of 37% formaldehyde, 2 ml of 10! MOPS-EDTA, 2 ml of 5!bromphenol blue, and 1 ml of 0. 1% (v/v) ethidium bromide) and denatured at 65 8C for 10 min. RNA samples were electrophoretically separated on 1.2% (w/v) agarose gel in MOPS-EDTA-formaldehyde [51]. Total RNA was downward transferred to a 20!8 cm MagnaGraph nylon membrane (Osmonics Inc., Westborough, MA) with 20!SSC overnight, and then bound to the membrane by UV irradiation (120 mJ cmK2) within a FB-UVXL-1000 CrossLinker (Fisher Scientific Co., Pittsburgh, PA). Blots were pre-hybridized with 25 ml of a hybridization solution (0.5 M NaHPO4 pH 7.2, 1% BSA, 1 mM EDTA, and 7% SDS) for 1 h at 65 8C, and then hybridized at 65 8C overnight with [32P]dCTP labeled DNA probes using the Random Primers DNA Labeling System (Invitrogen Life Technologies, Carlsbad, CA) according to the manufacturer’s procedure. Two NotI fragments and a XbaI/SalI fragment (1 kilo base pair (kb) each) corresponding respectively to class I chitinases 1, 2, and 3 [40]; a BamI/ SacI fragment (500 base pairs (bp)) from PR-1 (GenBank accession number U89895); a XhoI/SalI fragment (271 bp) from peroxidase [48]; a SphI/XhoI fragment (400 bp) from phenylalanine ammonia-lyase [65]; and two EcoRI PCRamplified DNA fragments of 691 and 916 bp, respectively, from chalcone synthase (CHS) and b-1,3-glucanase were used as labeled DNA probes. The cDNAs inserts were obtained from the plasmid(s) by digestion with specific enzyme(s), and purified from an agarose gel with the QIAquick gel extraction kit (Qiagen Inc., Valencia, CA). Genomic DNA from rice leaves (cultivar M201) was extracted following the protocol of Chen and Ronald [9] and 1 ng of DNA was used as template for PCR amplification of cDNA products from b-1,3-glucanase and CHS genes. PCR amplifications were carried out in a PTC-100e thermocycler (MJ Research Inc., Watertown, Mass). Forward primer (5 0 -GATCGGGGTGTGCTACGGCATGA-3 0 ) and a reverse primer (5 0 -GCTGATGGGGTAGACGTGCTGCA-3 0 ) were used to amplify 916-bp cDNA

148

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

product from b-1,3-glucanase gene [55] using the following cycling parameters: 95 8C for 5 min followed by 30 cycles of 95 8C for 30 s, 60 8C for 30 s, 72 8C for 1 min 15 s, and a final extension of 72 8C for 5 min. A 691-bp cDNA product from the CHS gene (GenBank accession number AB000801) was amplified using the forward primer 5 0 TCCCGCATCACCCACCTCGTCTTCTGC-3 0 and the reverse primer 5 0 -CTTGCGCATCTCGTCGAGGATGAAGAGC-3 0 with the following cycling parameters: 95 8C for 5 min followed by 30 cycles of 95 8C for 30 s, 62 8C for 30 s, 72 8C for 1 min, followed by a final extension at 72 8C for 5 min. The two PCR products were sub-cloned into the pCRw 2.1-Topow vector using the TOPO TA Cloningw Kit according to the manufacturer’s protocol (Invitrogen Life Technologies, Carlsbad, CA). Sequencing of the cDNA clones was carried out at the University of Florida DNA Sequencing Core Laboratory using ABI Prism BigDye Terminator cycle sequencing protocols developed by Applied Biosystems (Perkin–Elmer Corp., Foster City, CA). The fluorescently labeled extension products were analyzed on an Applied Biosystems Model 3100 genetic analyzer (Perkin–Elmer Corp., Foster City, CA). Oligo primers were designed using OLIGO 4.0 (National BioSciences Inc., Phymouth, MN) and synthesized by Gemini biotech. Nucleotide sequences were aligned and assembled using programs in the Sequencher 3.0 Software package (Gene Codes Corp., Ann Arbor, MI). Nucleotide sequences from the cDNA clones were also compared with the original sequences published in GenBank and EMBL databases. After overnight hybridization, each blot was subsequently washed with 25 ml of low stringency wash solution (40 mM NaHPO4 pH 7.2, 0.5% BSA, 1 mM EDTA, and 5% SDS) for 10 min at room temperature; 25 ml of a high stringency wash solution (40 mM NaHPO4 pH 7.2, 1 mM EDTA, and 1% SDS) for 10 min at room temperature; and twice with high stringency wash solution for 10 min at 65 8C. Blots were exposed to X-ray blue sensitive double sided film (Research Products International Corp., Mount Prospect, IL) for autoradiography from 8 to 24 h at K80 8C. Standardization of loaded total RNA was assessed by using a 322-bp BamHI fragment from the maize 5S rRNA cDNA clone [67] as a probe. Blots were analyzed by autoradiography and also scanned by a PhosphorImager (Bio-Rad Laboratories Inc., Hercules, CA) after being exposed to a Storage Phosphor Screen (Bio-Rad Laboratories Inc., Hercules, CA) for approximately 1 h. Hybridization signals from each probe were quantified using the Quantity One 1-D Analysis Software version 4.4.1 (Bio-Rad Laboratories Inc., Hercules, CA), and the relative level of signal for each time point per treatment was corrected based on the level of signal obtained from the maize 5S rRNA probe. Northern blots were conducted twice using RNA samples from two experiments. Results

shown are from one representative hybridization for each gene transcript.

3. Results 3.1. Fungal development index Temporal changes in M. grisea growth within epidermal cells of resistant (Katy) and susceptible (M201) rice cultivars amended with or without Si are represented graphically in Fig. 1 and through photomicrographs in Figs. 2–5. In the genetically resistant cultivar Katy, infection hyphae of M. grisea were observed within the first penetrated epidermal cell at 24 hai regardless of Si treatments (Figs. 1 and 2). Low FDI values were obtained at 24 hai (Fig. 1) indicating that the infection hyphae were shorter than the diameter of appressoria (Fig. 2). This was true also at 36 hai (Fig. 3). After 36 hai, FDI values increased above a value of 1 in both SiK and SiC treatments, but stabilized below a value of 3 between 48 and 96 hai (Fig. 1). FDI values lower than 4 corresponded to fungal hyphae that had formed a few branches and were restricted to the first-invaded epidermal cell (Fig. 4). At

Fig. 1. Fungal development index within adaxial epidermal cells of a resistant (Katy) and a susceptible (M201) rice cultivar amended (SiC) or without (SiK) silicon at different time-points after inoculation with M. grisea (race IB 49). Fifty appressorial sites where a conidium had formed a discernible appressorium were examined per each replication and treatment. FDI values given to all epidermal cells containing fungal cells, at each independent appressorial site, were summed and the average from the 50 independent sites at each time interval was plotted. Each value represents the mean of six replications obtained from two independent experiments. Bars represent the standard error of the mean.

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

149

Fig. 2. Differential interference contrast (DIC) microscopy (top row) and corresponding incident fluorescence microscopy (bottom row) of cleared Katy (resistant cultivar) and M201 (susceptible cultivar) epidermal cells amended with (SiC) or without (SiK) silicon at 24 hai with M. grisea (race IB 49). All images are 1000x. Appressoria (arrows) formed short infection hyphae (arrowheads) in the underlying epidermal cells of M201 in both Si treatments and characterized a type B reaction as showed in the DIC images. Epidermal cells in both Si treatments do not show autofluorescence in response to fungal penetration. Cytoplasmic granulation of epidermal cells of Katy (double arrowheads) beneath appressoria (arrows) in both Si treatments characterizes the type C reaction in the DIC images. Epidermal cell walls of Katy in both Si treatments display a bright autofluorescence (arrowheads). Epidermis (E), Mesophyll (M), and Conidium (C).

96 hai, FDI values for both SiK and SiC treatments were very similar in this incompatible interaction (Fig. 1). In the compatible interaction of M201-M. grisea, FDI values increased from 24 to 96 hai with a steeper slope and end value occurring within the SiK treatment (Fig. 1). At 24 hai, the FDI value for SiC treatment was 40% lower compared to the SiK treatment (Fig. 1). These FDI values indicate that the unbranched fungal hyphae had a length greater than two times the diameter of the appressorium and were restricted to the first-penetrated epidermal cell (Fig. 2). By 36 hai, fungal hyphae were still restricted to the firstinvaded epidermal cell (Fig. 3) as indicated by the lower FDI values (Fig. 1). The greatest difference between SiK and SiC treatments occurred at 72 and 96 hai (Fig. 1). In the SiK treatment, fungal hyphae grew successfully and formed an extensivly branched mycelium in the first-invaded epidermal cell and invaded many neighboring cells

(Fig. 5). By contrast, in the SiC treatment, the growth of fungal hyphae was greatly suppressed and restricted to less than three epidermal cells (Fig. 5). 3.2. Cellular responses of epidermal cells to M. grisea The frequency of three epidermal cellular responses in cultivars Katy and M201 amended with or without Si in response to infection by M. grisea is illustrated in Fig. 6. At 12 hai, all appressorial sites examined exhibited type A reaction (absence of infection hyphae within epidermal cell underlying the appressorium) regardless of cultivar and Si treatments (Fig. 6). In the incompatible Katy-M. grisea interaction, by 24 hai, 21–30% and 29–50% of the appressorial sites began to show, respectively, type B (infection hyphae within the epidermal cell and absence of cytoplasmic

Fig. 3. Differential interference contrast (DIC) microscopy (top row) and corresponding incident fluorescence microscopy (bottom row) of cleared Katy (resistant cultivar) and M201 (susceptible cultivar) epidermal cells amended with (SiC) or without (SiK) silicon at 36 hai with M. grisea (race IB 49). All images are 1000x. An appressorium (arrow) formed an infection hyphae that branched and fully colonized the penetrated epidermal cell (arrowheads) of M201 in the SiK DIC image. The penetrated epidermal cell shows moderate autofluorescence. In the M201 SiC DIC image, a poorly-developed infection hypha originated from an appressorium (arrow) underlying an epidermal cell formed branches (arrowheads) and host cytoplasm shows intense granulation (double arrowheads). The penetrated epidermal cell and the cell walls of two surrounding epidermal cells (arrowheads) exhibit bright autofluorescence. In the DIC images of Katy, the granulation of the cytoplasm (double arrowheads) of epidermal cells beneath appressoria (arrows) is more intense in both Si treatments. Epidermal cells show whole-autofluorescence (arrowheads). Epidermis (E) and Mesophyll (M).

150

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

Fig. 4. Differential interference contrast (DIC) microscopy (top row) and incident fluorescence microscopy (bottom row) of cleared Katy (resistant cultivar) epidermal cells amended with (SiC) or without (SiK) silicon at 72 hai with M. grisea (race IB 49). The extent of epidermal cell browning beneath appressorial sites (arrowheads) in both Si treatments is visualized at 400! in DIC images. The corresponding whole-epidermal cells autofluorescence highlighting the cell walls of neighboring epidermal cells (arrowheads) are presented below the 400! DIC images. A poorly-developed fungal hypha invading an epidermal cell shows a few branches (arrowheads) in the 1000! SiK DIC image. Note the dark browning of the epidermal cell walls and of a few mesophyll cells (double arrowheads). Note the autofluorescence of the penetrated epidermal cell and surrounding epidermal and a few mesophyll cells in the corresponding autofluorescence image. The browning of epidermal cells surrounding the penetrated one partially obscure their autofluorescence. A poorly developed fungal hypha (arrowheads) in the penetrated epidermal cell in the SiC1000! DIC image. Epidermal cell walls show intense browning (double arrowheads). Note the autofluorescence of the penetrated epidermal cell and cell walls of surrounding epidermal cells of the corresponding autofluorescence image. Epidermis (E) and Mesophyll (M).

granulation) and type C (intense granulation of the cytoplasm of the invaded epidermal cell) reactions regardless of Si treatments (Figs. 2 and 6). A decrease in type B reaction from 24 to 96 hai was followed by an increase in the frequency of appressorial sites exhibiting type C reaction regardless of Si treatments (Figs. 3, 4 and 6). The frequency of appressorial sites showing type B and type C reactions was lower in SiC than in SiK treatment because of a great frequency of appressorial sites showing type A reaction (unsuccessful penetration) (Fig. 6). The frequency of appressorial sites exhibiting type A reaction was quite similar from 48 to 96 hai within each Si treatment.

In the M201-M. grisea interaction, the frequency of appressorial sites exhibiting type B reaction increased from 24 to 48 hai in both Si treatments (Fig. 6). A similar frequency of appressorial sites exhibiting type A reaction occurred from 48 to 96 hai within each Si treatment, however, slightly higher frequencies appeared to occur in the SiC treatment in comparison to SiK treatment (Fig. 6). Appressorial sites started to show type C reaction by 36 hai in both Si treatments. Interestingly, the frequencies of appressorial sites showing the type C reaction was slightly higher from 36 to 96 hai in the SiC relative to the SiK treatment (Figs. 3, 5 and 6).

Fig. 5. Differential interference contrast (DIC) microscopy (top row) and incident fluorescence microscopy (bottom row) of cleared M201 (susceptible cultivar) epidermal cells amended with (SiC) or without (SiK) silicon at 72 hai with Magnaporthe grisea (race IB 49). The extent of epidermal cell browning beneath appressorial sites (arrowheads) in both Si treatments is shown at 400! in DIC images. Note that in the SiK treatment, many more epidermal and mesophyll cells show dark browning than in the SiC treatment. The corresponding autofluorescence images of epidermal cells and the cell walls of neighboring epidermal cells (arrowheads) beneath appressorial sites are presented below the DIC images. Note that in the SiK treatment, the autofluorescence of epidermal cells is quenched as the browning develops. Well-developed invaded hyphae in the epidermal cell beneath the appressorium (arrow) producing several branches and colonizing neighboring epidermal cells (arrowheads) is shown in the DIC image of SiK at 1000!. The corresponding incident fluorescence image demonstrating a faint autofluorescence of epidermal and mesophyll cells is presented below the DIC image. In SiC at 1000!, fungal hyphae (arrowheads) within an epidermal cell show limited growth. Note a few browned epidermal and mesophyll cells. Intense autofluorescence of the penetrated and neighboring adjacent epidermal cells is seen in the corresponding incident fluorescence SiC1000! image. Epidermis (E) and Mesophyll (M).

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

151

Fig. 6. Categories of cellular reactions occurring in adaxial epidermal cell(s) of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Cellular reactions were grouped in category A— unsuccessful penetration (absence of infection hyphae within epidermal cell underlying the appressorium); category B—successful penetration (infection hyphae within the epidermal cell and absence of cytoplasmic granulation); and category C—successful penetration (infection hyphae within the epidermal cell associated with cytoplasmic granulation). Each bar represents the mean of six replications obtained from two independent experiments. Fifty appressorial sites, where a conidium had formed a discernible appressorium, were examined per each replication. The standard error of the mean is represented in each bar.

3.3. Analysis of autofluorescence All appressorial sites showing type B and C reactions were autofluorescent independent of the level of resistance of the cultivars used and Si treatments. Autofluorescent appressorial sites, including the whole-epidermal cell and/or cell wall autofluorescence, occurred at 24 and 36 hai, respectively, in Katy and M201 with a slight increase onward regardless of Si treatments (Figs. 2, 3 and 7). The frequency of appressorial sites showing autofluorescence appeared consistently to be slightly greater in SiK than in Si C treatment (Fig. 7). The frequency of autofluorescent appressorial sites stabilized from 36 to 96 h and from 48 to 96 hai, respectively, in Katy and M201 regardless of Si treatments (Fig. 7). The number of

autofluorescent epidermal cells increased from 24 to 96 hai, especially in M201 (Fig. 8). In M201, the number of autofluorescent epidermal cells per appressorial site was dramatically higher in SiK than in SiC treatment while in Katy this difference was not evident (Fig. 8). 3.4. Determination of browned epidermal cells Brown epidermal cells at the appressorial sites examined became visible at 48 hai regardless of cultivar and Si treatments (Fig. 9). The number of brown epidermal cells in M201 dramatically increased after 48 h whereas in Katy this increase was slow and stabilized from 72 to 96 hai (Fig. 9). Interestingly, the lowest number of brown epidermal cells occurred in SiC treatment regardless of the cultivars used.

Fig. 7. Percentage of autofluorescent appressorial sites in adaxial epidermal cells of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Autofluorescence of epidermal cell walls or the whole-epidermal cells of each appressorial site was recognized by incident fluorescence microscopy using a Chroma Endow GFP filter set (excitation 470– 510 nm, dichroic beamsplitter 495 nm, and emission 525–575 nm). Each plotted value represents the mean of six replications obtained from two independent experiments. Fifty appressorial sites, where a conidium had formed a discernible appressorium, were examined per each replication. Bars represent the standard error of the mean.

152

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

Fig. 8. Number of autofluorescent adaxial epidermal cells of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Autofluorescence of epidermal cell walls or the whole-epidermal cells of each appressorial site was recognized by incident fluorescence microscopy using a Chroma Endow GFP filter set (excitation 470–510 nm, dichroic beamsplitter 495 nm, and emission 525–575 nm). Each value represents the mean of six replications obtained from two independent experiments. Fifty appressorial sites, where a conidium had formed a discernible appressorium, were examined per each replication. Bars represent the standard error of the mean.

Representative appressorial sites in the epidermal cells of Katy and M201, with or without Si, at 72 hai is illustrated in Figs. 4 and 5. In Katy, the difference in the development of brown epidermal cells and their autofluorescence between Si treatments was minimal (Fig. 4). However, in M201, the number of brown epidermal cells seemed to be lower in SiC than in SiK treatment (Fig. 5). Indeed, in SiK treatment, many mesophyll cells became dark brown and the autofluorescence of epidermal and mesophyll cells was partially obscured upon the development of the brown color compared to SiC treatment where the autofluorescence was still bright (Fig. 5). The intensity of browning observed in epidermal cell walls was categorized (Fig. 10). At the appressorial sites examined, cell walls became pale yellow or slightly brown (level 2) at 36 hai regardless of cultivar and Si treatments (Fig. 10). A decrease in the frequency of appressorial sites showing browning (level 2) by 36 hai was followed by an increase in the frequency of appressorial sites with wholeepidermal cells deeply brown and adjacent cells with slightly brown cell walls (level 3) at 48 hai regardless of Si and cultivar treatments (Fig. 10). The frequency of appressorial sites showing browning (level 3) was reduced

in SiC relative to the SiK treatment because of the high frequency of appressorial sites exhibiting no browning of the epidermal cell walls (level 1). 3.5. Quantification of total soluble phenolics and lignin–thioglycolic acid derivatives In Katy, regardless of Si treatment, the concentration of total soluble phenolics (TSP) increased from 0 to 12 hai, attained constant values from 12 to 36 hai, and then increased at 48 hai with stable values thereafter (Fig. 11). In M201, the concentration of TSP increased from 0 to 12 hai and remained constant from 24 to 72 hai regardless of Si treatment. Between 72 and 96 hai, TSP decreased in the M201 SiK treatment, but increased in the M201 SiC treatment (Fig. 11). The accumulation of lignin–thioglycolic acid (LTGA) derivatives was biphasic in both Katy and M201 regardless of Si treatment. In Katy, the initial and the second increase in LTGA derivatives occurred at 24 and 36 hai, respectively. Although the final concentration of LTGA appeared nearly equal among all treatments, the rate of accumulation appeared slower in the M201 SiC treatment (Fig. 12).

Fig. 9. Number of brown adaxial epidermal cell(s) of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Each value represents the mean of six replications obtained from two independent experiments. Fifty appressorial sites, where a conidium had formed a discernible appressorium, were examined per each replication. Bars represent the standard error of the mean.

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

153

Fig. 10. Intensity of browning of adaxial epidermal cells of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). The intensity of browning of epidermal cells was grouped in category 1—no browning of cell walls of epidermal cells; category 2—cell walls of epidermal cells pale yellow or slightly brown; and category 3—whole-epidermal cells deeply brown and adjacent cells with slightly brown cell walls. Each bar represents the mean of six replications obtained from two independent experiments. Fifty appressorial sites, where a conidium had formed a discernible appressorium, were examined per each replication. The standard error of the mean is represented in each bar.

3.6. Time-course of transcript accumulations The steady-state levels of chitinases (chit) 1, 2, and 3; b-1, 3-glucanase (Glu); chalcone synthase (CHS); phenylalanine ammonia-lyase (PAL); peroxidase (POX); and PR-1 transcripts in rice leaves of cultivars Katy and M201 amended with (SiC) or without (SiK) silicon at different time-points after inoculation with M. grisea were examined by RNA gel blot hybridization and quantified by phosphorimager analysis to determine if gene transcript accumulation was associated with an increase in resistance to blast disease. Chit 1 transcripts accumulated at a variable basal level from 0 to 96 hai with an apparent increase at 48 hai in all cultivars and Si treatments (Figs. 13 and 14). This peak at 48 hai was very pronounced in the M201 SiK treatment which exhibited significant accumulation of Chit 1 transcripts between 36 and 72 hai. This same pattern of

accumulation was observed with Chit 2 and Chit 3 as well (Figs. 13 and 14). In Katy, the levels of Glu transcripts fluctuated overtime in the SiK treatment with greater accumulation occurring at 24, 48, and 72 hai (Figs. 13 and 14). For the SiC treatment, Glu transcript accumulation was higher at 36 and 48 hai than at other time points. In M201, Glu transcript accumulation in the SiK treatment was constant from 12 to 36 hai with a peak occurring at 48 hai. By contrast, the levels of Glu transcript for the SiC treatment increased from 24 to 36 hai and declined thereafter. Interestingly, in M201, Glu transcripts were strongly induced at 36 hai in the SiC treatment. In Katy, the only major induction of CHS occurred at 12 hai in SiC treatment (Figs. 13 and 14). In M201, CHS transcript accumulation dramatically increased from 24 to 36 hai with a decline onward in the SiK treatment whereas

Fig. 11. Concentration of total soluble phenolics in leaves of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Each value represents the mean of six replications obtained from two independent experiments. Bars represent the standard error of the mean.

154

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

Fig. 12. Concentration of lignin and lignin-like polymers, determined as lignin–thioglycolic acid (LTGA) derivatives, in leaves of a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). Each value represents the mean of six replications obtained from two independent experiments. Bars represent the standard error of the mean.

in the SiC treatment only a major peak was observed at 36 hai. PAL transcripts exhibited variable accumulation at low levels from 0 to 36 hai regardless of cultivar and Si treatment (Figs. 13 and 14). In Katy, high levels of PAL transcripts occurred at 48 and 72 hai regardless of Si treatments. In M201 SiK treatment, accumulation of PAL transcripts showed a small increase from 24 to 48 hai and peaked between 60 and 72 hai. For the SiC treatment, major peaks of PAL transcripts occurred at 60 and 72 hai. Accumulation of POX transcripts in Katy SiK treatment increased from 24 to 48 hai and decreased onward (Figs. 13 and 14). In Katy SiC treatment, POX transcript accumulation exhibited a small peak at 12 hai, decreased slightly at 24 hai then increased to peak again at 60 hai. High levels were still detected at 72 hai with a sharp decrease thereafter. In the M201 SiK treatment, POX transcript accumulation was high at 12 hai, increased further and remained high through 48 hai before decreasing dramatically. In M201 SiC treatment, POX transcript accumulation was very high at 12 hai, and although some fluctuation was observed, remained at high levels through 96 hai.

For Katy, PR-1 transcript dramatically increased from 48 to 60 hai and slightly decreased thereafter with a similar trend for both SiK and SiC treatments (Fig. 13). Induction of PR-1 transcript accumulation in M201 occurred at 60 hai in both Si treatments. Expression of PR-1 transcript was obviously lower in SiK compared to SiC treatment from 60 to 96 hai. Results from the quantification and normalization for equal loading by phosphorimaging of transcripts for each gene (Fig. 14) validated the interpretation of the RNA blots shown in Fig. 13.

4. Discussion Consistent with the trend towards unraveling the mechanism(s) by which Si limits rice blast disease [22,28, 49,50,63], the present study not only supports the concept that Si increases resistance to blast [12,53], but also provides cytological and molecular features associated with this phenomenon. To establish compatibility, it appears important for M. grisea to keep the penetrated epidermal cell of a susceptible cultivar alive in the early stage of

Fig. 13. Northern hybridization analysis of chitinases (chit) 1, 2, and 3, b-1,3-glucanase (Glu), chalcone synthase (CHS), phenylalanine ammonia-lyase (PAL), peroxidase (POX), and PR-1 transcript accumulation in a resistant (Katy) and a susceptible (M201) rice cultivar amended with (SiC) or without (SiK) silicon at different time-points after inoculation with Magnaporthe grisea (race IB 49). RNA was fractionated by agarose gel electrophoresis, transferred to nylon membrane, hybridized with DNA probes, and hybridization signals detected by autoradiography. A maize 5S rRNA was used as a loading control.

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

155

Fig. 14. Quantification of chitinases (chit) 1, 2, and 3, b-1,3-glucanase (glu), chalcone synthase (CHS), phenylalanine ammonia-lyase (PAL), peroxidase (POX), and PR-1 hybridization signals from blots shown in Fig. 13. Hybridization signals were detected by phosphorimaging and quantified relative to the maize 5S rRNA loading control.

infection before changing to a necrotrophic form of parasitism [23]. According to Ohata et al. [42], in an incompatible rice-M. grisea interaction, infection hyphae penetrate the epidermal cells about 12 hai but die 18–24 h thereafter accompanied with host cell death. By contrast, in a compatible interaction, the infection hyphae remain alive for up to 44 h within the penetrated epidermal cell. In the current study, infection hyphae of M. grisea were first observed within epidermal cells of cultivars Katy and M201 at 24 hai independent of their level of resistance to the fungus and Si amendment. At this time point in the incompatible interaction of Katy-M. grisea, epidermal cells reacted to fungal ingress through the development of HR as indicated by the granulation of the cytoplasm and a bright autofluorescence of epidermal cell walls. By 36 hai onward, autofluorescence of whole-epidermal cells occurred, cytoplasmic granulation was intensified, and the growth of fungal hyphae within the invaded epidermal cell was suppressed. These cellular events of HR are associated with race-specific resistance of rice to M. grisea [31,32,45,60] and appear, with

the exception of a slight decrease in the frequency of autofluorescence, to occur independently of Si amendment. In contrast, during the compatible interaction of M201-M. grisea, fungal hyphae grew vigorously within penetrated epidermal cells in the absence of autofluorescence and granulation of cytoplasm until 36 hai. Cytoplasmic granulation, although slightly higher with Si treatment, was observed infrequently in M201 relative to Katy. Considering that all appressorial sites showing fungal penetration with or without granulation (type B and C reactions) were autofluorescent, the slight decrease in the frequency of autofluorescent sites in SiC plants is attributed to their respective frequency of unsuccessful penetrations (type A reaction). Interestingly, the number of autofluorescent epidermal cells per appresorial site was much lower in SiC plants of the M201 cultivar. The number of cells was comparable to the number seen with Katy independent of Si treatment. In resistant rice cultivars, epidermal cells infected by M. grisea die prior to the death of the invading fungus [32].

156

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

Although the exact biological function of autofluorescence and cytoplasmic granulation within epidermal cells cannot be clearly defined, granulation alone appears to be a reliable hallmark of R-gene conditioned cell death associated with limited fungal spread. While early autofluorescence may also serve as an indicator of incompatibility, it is later associated with both compatible and incompatible interactions and involves an increased number of host cells in the latter event. Considerable physical and biochemical evidence suggests that autofluorogens present in epidermal cells of cereals are phenolic compounds probably related to lignin biosynthesis [5]. In this study, epidermal cell walls became autofluorescent 12 h later in M201 reconfirming the biotrophic relationship between M. grisea and the susceptible epidermal cells before cell death is triggered. In Katy, the early occurrence of autofluorescence, one key feature of HR, appears to be an indication of cell death. In contrast, in the interaction with the susceptible M201 cultivar the increased number of autofluorescent epidermal cells and the growth of the fungus in these cells, especially in SiK plants, strengthens the idea that autofluorescence may be the consequence of different cellular activities with dramatically different outcomes. Peng et al. [46] reported that the most notable difference between a compatible and incompatible rice-M. grisea interaction was the reduced growth of the infection hyphae since the frequency of penetration between these two types of interactions was similar. Peng and Shishiyama [45] observed the formation of papilla-like structures at some unsuccessful penetration sites of M. grisea, but the low frequency would not account for the penetration failure of most conidia. The study by Takahashi et al. [59] suggested that epidermal cells might not recognize the appressorium that failed to penetrate or they acquire resistance by receiving signals from neighboring cells undergoing HR. In contrast, appressoria produced by mutants of M. grisea deficient in a mitogen-activated protein kinase did not penetrate the epidermal cells. These appresoria, however, were able to trigger localized autofluorescence probably in response to the mechanical pressure exerted during attempted, but failed, penetration or by soluble factors released from the germinated conidia [62]. Heath et al. [20] showed that M. grisea elicited no ultrastructurally detectable responses of underlying epidermal cells of weeping lovegrass during appressorium development or penetration peg formation. In leaves of the less susceptible goosegrass, the authors observed that successful penetration from appressoria was rare and usually resulted in rapid death of both fungus and epidermal cell. Indeed, fungal growth ceased at various stages during the formation of the penetration peg in association with the deposition of highly electron-opaque material in the underlying epidermal cell which appeared to contain silica. Interestingly, in the current study, the presence of papillae as well as autofluorescence and cytoplasmic granulation were not

observed in epidermal cells underlying appressoria that failed to penetrate regardless of cultivar or Si treatment. The frequency of appressoria failing to penetrate epidermal cells was quite similar between cultivars. However, unsuccessful penetrations appeared slightly more common in SiC plants within cultivar treatments. Hayasaka et al. [19] also observed a higher number of unsuccessful penetrations of M. grisea in rice epidermal cells and a lower frequency of infection sites showing cytoplasmic granulation and browning of both epidermal and mesophyll cells in SiC compared to SiK rice plants. A decrease in the number of blast lesions was attributed to the fact that some appressoria cannot overcome the physical impedance offered by the cuticle-silica double layer [63]. Koga and Kobayashi [32] also observed no difference in the number of unsuccessful penetrations between an incompatible and a compatible rice-M. grisea interaction suggesting that failure of appressoria to penetrate the epidermal cell was independent of the resistance controlled by the major gene Pi-z. Indeed, race-specific resistance in rice is unlikely to impede pre-penetration by M. grisea [26] Necrosis must be triggred by M. grisea to successfully colonize and sporulate in rice tissues. It seems, therefore, that in areas of heavy Si deposition, a delay in fungal ingress and colonization provides the rice plant with enough time to activate mechanism(s) of defense to overcome the infection by M. grisea. By 36 hai, epidermal cell walls of Katy and M201, regardless of Si treatment, became pale yellow, and 12 h later whole-epidermal cells were deeply brown with adjacent cells showing slightly brown cell walls. The genetically resistant cultivar Katy showed a lower number of brown epidermal cells than M201 and the difference between SiK and SiC plants was very small. In M201, however, the number of brown epidermal cells was dramatically higher in SiK plants. Interestingly, epidermal cells became deeply-brown directly after the appearance of autofluorescence in both compatible and incompatible interactions regardless of Si treatment. Additionally, as the browning of epidermal cells developed, their autofluorescence became quenched, most likely due to the oxidation and/or polymerization of phenolics. Taking into consideration that the progressive necrosis of rice cells requires a continued metabolic activity of M. grisea, it is postulated that the higher number of brown epidermal cells in SiK plants of cultivar M201 is the result of an unlimited growth of M. grisea within epidermal cells as compared to SiC plants, where fungal growth was greatly suppressed and restricted to less than three epidermal cells. The reduced hyphal growth within epidermal cells can account for the lower number of brown and necrotic epidermal cells and, consequently, the discrete and smaller lesions on leaf blades of SiC plants of Katy and M201 cultivars. The effect of Si was such that in the experiments of the present study, plants of M201, the susceptible cultivar, amended with Si exhibited a similar visual level (number of brown epidermal

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

cells) of blast severity as plants of the genetically resistant cultivar Katy. There is a significant body of literature describing that application of Si may affect phenolic production upon pathogen attack [4,14,37]. Although Sridhar and Mahadevan [57] reported that rice cultivars susceptible to blast tend to accumulate more phenolics than resistant ones, data from the present study showed that despite the fact that the concentration of total soluble phenolics dramatically increased during the course of infection, the concentration of total soluble phenolics in Katy and M201 within each Si treatment was very similar. Although Zhang et al. [64] found that the expression of PAL transcript was strongly induced in resistant but not in susceptible rice cultivars upon infection by M. grisea, results from this study suggest the opposite. PAL transcripts were detected in non-inoculated plants but accumulated to much higher levels during M201 infection by M. grisea. Support for a close association between the higher levels of PAL transcript observed in SiK plants of cultivar M201 and the subsequent increase in the concentration of total soluble phenolics is illustrated by the higher frequency of appressorial sites showing autofluorescence without granulation (type B reaction) and massive colonization of the epidermal cells by M. grisea. Although background transcript levels from the flavonoid biosynthetic gene CHS occurred in both SiK and SiC plants of cultivar Katy, a strong induction was observed at 12 hai in SiC plants. In M201, however, the accumulation of the CHS transcript in SiK plants was enhanced in a coordinated manner whereas in SiC plants only a single but strong induction occurred at 36 hai. Although Knon et al. [29] observed remarkable changes in the content of flavonoids in resistant rice cultivars upon infection by M. grisea, the weak induction of CHS transcript in Katy suggests that products of the flavonoid biosynthesis pathway, including probably the phytoalexin sakuranetin, might not play a key role in rice resistance against blast. Accumulation of peroxidase transcripts (POX) increased during the course of infection by M. grisea in both compatible and incompatible interactions regardless of Si treatment although the magnitude and kinetics differed. In Katy, POX transcripts accumulated earlier in SiC than in SiK plants. In M201, by contrast, a higher level of POX transcripts accumulated at 12 hai and was sustained in SiC plants through 96 hai, but only until 48 hai in SiK plants. The rate of lignin and lignin-like polymer accumulation, determined as LTGA derivatives, was much slower in SiC than in SiK plants of cultivar M201 whereas in Katy, a difference between SiK and SiC plants was not observed. Accumulation of POX transcripts was associated with an increase in resistance of rice plants to blast [47] presumably due to the participation of peroxidases in the biosynthesis of lignin [18]. Peroxidase activity dramatically increased in roots of cucumber plants amended with Si upon infection by Pythium spp. whereas a comparable level of activity occurred 10 days later in SiK wilted plants

157

[10]. These findings are in agreement with our results showing that the strong induction of POX transcripts following infection by M. grisea corresponded to an increase in the concentration of LTGA derivatives. Although the highest concentration of LTGA derivatives occurred in SiK plants of M201, likely due to the high number of successful penetrations, rice cells were obviously not protected efficiently against M. grisea colonization. By contrast, in SiC plants of M201, the induction of POX transcripts was lengthened until 96 hai and followed the increase in LTGA derivatives. In Katy, the pattern of Glu transcript accumulation was quite similar between SiK and SiC plants. However, in M201, the major difference between SiK and SiC plants was a strong induction of Glu transcript at 36 hai in SiC plants. This strong induction of Glu transcript, likely modulated by Si, may also have contributed to suppression of growth of M. grisea. Transcripts for chitinases 1, 2, and 3 exhibited a similar low level of accumulation in Katy regardless of Si amendment, while in M201 very high levels of chitinase transcripts accumulated in SiK relative to SiC amended plants. Although Nishizawa et al. [41] reported that transgenic rice plants constitutively expressing chitinase genes were resistant to rice blast, the absence and sometimes the weak induction of chitinases in Katy, regardless of Si treatment, and in SiC plants of M201, suggest that chitinases do not play a significant direct role in rice resistance to blast. The strong induction of chitinase transcripts in SiK plants of M201 cannot be linked with resistance because of unlimited growth of M. grisea within epidermal cells. It has been reported that chitinase transcripts accumulate at high levels in a compatible but not in an incompatible rice-M. grisea interaction [34] perhaps because susceptible cultivars need to maintain a continuous retaliation against pathogen colonization even though it can overcome the mechanism(s) of resistance activated by the host. According to Rodrigues et al [49], chitinases may not be an important mechanism of defense in rice against M. grisea because the pattern of chitin localization over fungal cell walls in tissues of SiK and SiC plants of cultivar M201 at 96 hai was very similar in terms of uniformity and density. In Katy, PR-1 transcript was induced earlier and showed a similar trend of accumulation in both SiK and SiC plants. By contrast, in M201, PR-1 transcript was induced 12 h later, but exhibited higher levels of accumulation in SiC compared to SiK plants from 60 to 96 hai. In line with our observations, it has been reported that in an incompatible rice-M. grisea interaction, PR-1 transcript accumulated at 24 hai and preceded the occurrence of HR whereas in the compatible interaction, its accumulation occurred at 48 hai with a peak at 72 hai when visible necrotic lesions started to develop [27]. Based on our observations, we hypothesize that the intensity and kinetics of PR-1 accumulation, modulated by R-genes as well as by

158

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159

Si, indicates the quantitative level of resistance of rice cells to infection by M. grisea. Application of Si is a practice compatible with environmental friendly strategies for sustainable blast disease management and rice production. Results provided in this study indicate that race-specific resistance of Katy against M. grisea in both SiK and SiC plants occurred through the development of HR and with the induction of PR-1 and peroxidase transcripts independently of Si amendment. In SiC plants of the susceptible cultivar M201, differential accumulation of transcripts from glucanase, peroxidase and PR-1 correlated with an inhibition of the spread of the fungus and the damage that it caused to the leaf tissues. Conversely, in SiK plants, M. grisea still inflicts greater damage to host tissue despite the induction of genes encoding chitinases, glucanase, chalcone synthase, and PR1, and the accumulation of high levels of phenolics and lignin–thioglycolic acid derivatives. These findings give support to an active role of Si in the rice-M. grisea interaction and suggests that the mechanisms of Sienhanced resistance may overlap yet contain features unique from major R-gene mediated resistance.

[3]

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11]

Acknowledgements We would like to thank Drs. M. Heath, R. J. Zeyen, and T. L. W. Carver for assistance in the interpretation of cytological results, and Drs. S. T. Talcott and I. A. Dubery, respectively, for their advice on determination of phenolics and lignin. The authors also extend thanks to U. Benny, J. Minsavage and A. Hutchens for technical assistance. We are gratefull to Dr. E. A. Zimmer (Department of Biology, Washington University) for providing the maize 5S rRNA cDNA clone. The cDNA clone from phenylalanine ammonia-lyase was a gift from Dr. W. Y. Song (Department of Plant Pathology, University of Florida) who received it from Dr. C. J. Lamb (Plant Biology Laboratory, Salk Institute for Biological Studies, La Jolla, CA). We are indebted to Dr. Y. Nishizawa (National Institute of Agrobiological Resources, Japan) who provided cDNA clones from PR-1 and class I chitinases 1, 2, and 3. Dr. R. Dudler (Institute of Plant Biology, University of Zurich, Switzerland) kindly provided the peroxidase cDNA clone. This research was supported in part by the Florida Agricultural Experiment Station and approved for publication as Journal Series No. R-10845.

[12] [13]

[14] [15] [16]

[17] [18] [19]

[20]

[21]

[22] [23]

References [1] Barber MS, Ride JP. A quantitative assay for induced lignification in wounded wheat leaves and its use to survey potential elicitors of the response. Physiol Mol Plant Pathol 1988;32:185–97. [2] Bryan GT, Wu KS, Farrall L, Jia Y, Hershey HP, McAdams SA, et al. A single amino acid difference distinguishes resistance and

[24]

[25] [26]

susceptible alleles of the rice blast resistance gene Pi-ta. Plant Cell 2000;12:2033–45. Cartwright DW, Langcake P. Phytoalexin production in rice and its enhancement by a dichlorocyclopropane fungicide. Physiol Plant Pathol 1980;17:259–67. Carver TLW, Robbins MP, Thomas BJ, Troth K, Raistrick N, Zeyen R J. Silicon deprivation enhances localized autofluorescent responses and phenylalanine ammonia-lyase activity in oat attacked by Blumeria graminis. Physiol Mol Plant Pathol 1998;52:245–57. Carver TLW, Zeyen RJ, Bushnell WR, Robbins MP. Inhibition of phenylalanine ammonia lyase and cinnamyl-alcohol dehydrogenase increases quantitative susceptibility of barley to powdery mildew (Erysiphe graminis DC). Physiol Mol Plant Pathol 1994;44:261–72. Carver TLW, Zhang L, Zeyen RJ, Robbins MP. Phenolic biosynthesis inhibitors suppress adult plant resistance to Erysiphe graminis in oat at 20 and 108C. Physiol Mol Plant Pathol 1996;41:149–63. Chao CT, Ellingboe AH. Genetic analysis of avirulence/virulence of an isolate of Magnaporthe grisea from a rice field in Texas. Phytopathology 1997;86:71–6. Chao CT, Moldenhauer KAK, Ellingboe AH. Genetic analysis of resistance/susceptibility in individual F3 families of rice against strains of Magnaporthe grisea containing different genes for avirulence. Euphytica 1999;109:183–90. Chen DH, Ronald PC. A rapid DNA minipreparation method suitable for AFLP and other PCR applications. Plant Mol Biol Report 1999;17:53–7. Che´rif M, Asselin A, Be´langer RR. Defense responses induced by soluble silicon in cucumber roots infected by Pythium spp. Phytopathology 1994;84:236–42. Cunningham JL. A miracle mounting fluid for permanent wholemounts of microfungi. Mycologia 1972;64:906–11. Datnoff LE, Deren CW, Snyder GH. Silicon fertilization for disease management of rice in Florida. Crop Prot 1997;16:525–31. Du LC, Wang J. Activities and distribution of chitinase and b-1,3glucanase in rice induced by Pyricularia oryzae. Acta Phytopathol Sinica 1992;18:29–36. Epstein E. Silicon. Annu Rev Plant Physiol Plant Mol Biol 1999;50: 641–64. Flor HH. Current status of the gene-for-gene concept. Annu Rev Phytopathol 1971;9:275–96. Ge X, Song F, Zheng Z. Active oxygen species production in rice seedlings infected by Magnaporthe grisea is involved in the blast resistance. Acta Phytophysiol Sinica 2000;24:227–31. Goodman RN, Novacky AJ. Hypersensitive reaction in plants to pathogens. St Paul, MN: APS Press; 1994. Grisebach H. Lignin. In: Conn EE, editor. The Biochemistry of plants. vol. 7. New York: Academic Press; 1981. p. 451–78. Hayasaka T, Fujii H, Mayumi H, Ando H, Namai T. How does silicon influence on resistance of rice blast disease? Proceedings of the second silicon in agriculture conference, Tsuruoka, Yamagata, Japan; 2002. p. 243–6. Heath MC, Howard RJ, Valent B, Chumley FG. Ultrastructural interactions of one strain of Magnaporthe grisea with goosegrass and weeping lovegrass. Can J Bot 1992;70:779–87. Ikata S, Matsuura K, Taguchi S. Results of experiments on the control of rice blast, 1. Studies on cultivar resistance to rice blast disease. Noji Kairyo Shiryo 1931;20:1–140. Ito S, Hayashi H. On the relation of silica supply to rice blast. J Sapporo Soc Agric Sci 1931;103:460–1. Ito S, Sakamoto M. Studies on rice blast. Res Hokkaido Univ Bot Lab Fac Agric Rep 1939;1939–43. Jia Y, McAdams SA, Bryan GT, Hershey HP, Valent B. Direct interaction of resistance gene and avirulence gene products confers rice blast resistance. EMBO J 2000;19:4004–14. Jones DH. Phenylalanine ammonia-lyase: regulation of its induction, and its role in plant development. Phytochemistry 1984;23:1349–60. Kaur S, Kaur P, Padmanabhan SY. Differential host response to infection by Pyricularia oryzae. Ind Phytopathol 1983;36:179–80.

F.A´ Rodrigues et al. / Physiological and Molecular Plant Pathology 66 (2005) 144–159 [27] Kim S, Ahn IP, Park CH, Park SG, Park SY, Jwa NS, et al. Molecular characterization of the cDNA encoding an acidic isoform of PR-1 protein in rice. Mol Cells 2001;(11):115–21. [28] Kim SG, Kim KW, Park EW, Choi D. Silicon-induced cell wall fortification of rice leaves: a possible cellular mechanism of enhanced host resistance to blast. Phytopathology 2002;92:1095–103. [29] Knon SH, Oh JH, Song HS. Studies on the relationship between chemical contents of rice plants and resistance to rice blast disease. Korean J Plant Prot 1974;13:33–9. [30] Kodama O, Miyakawa J, Akatsuka T, Kiyosawa S. Sakuranetin, a flavanone phytoalexin from ultraviolet-irradiated rice leaves. Phytochemistry 1992;31:3807–9. [31] Koga H. Hypersensitive death, autofluorescence and ultrastructural changes in cells of leaf sheaths of susceptible and resistant nearisogenic lines of rice (Pi-Zt) in relation to penetration and growth of Pyricularia oryzae. Can J Bot 1994;72:1463–77. [32] Koga N, Kobayashi T. Comparison of the early infection process of Pyricularia oryzae Cav. in rice leaves of compatible and incompatible combinations. Annu Phytopathol Soc Jpn 1982;48:506–13. [33] Lau GW, Chao CT, Ellingboe AH. Interaction of genes controlling avirulence/virulence of Magnaporthe grisea on rice cultivar Katy. Phytopathology 1993;83:375–82. [34] Manandhar HK, Mathur SB, Smedegaard-Petersen V, ThordalChristensen H. Accumulation of transcripts for pathogenesis-related proteins and peroxidase in rice plants triggered by Pyricularia oryzae, Bipolaris sorokiniana and UV light. Physiol Mol Plant Pathol 1999; 55:289–95. [35] Matsuo T, Hoshikama K. Science of the rice plant morphology. Tokyo: Food Agric Policy Res Center; 1993. [36] Mayama S, Shishiyama J. Localized accumulation of fluorescent and U.V.-absorbing compounds at penetration sites in barley leaves infected with Erysiphe graminis hordei. Physiol Plant Pathol 1978;13: 347–54. [37] Menzies JG, Ehret DL, Glass ADM, Samuels AL. The influence of silicon on cytological interactions between Sphaerotheca fuliginea and Cucumis sativus. Physiol Mol Plant Pathol 1991;39:403–14. [38] Moldenhauer KAK, Bastawisi AO, Lee FN. Inheritance of resistance in rice to races IB-49 and IC-17 of Pyricularia grisea rice blast. Crop Sci 1992;32:584–8. [39] Muthukrishnan S, Liang GH, Trick HN, Gill BS. Pathogenesis-related proteins and their genes in cereals. Plant Cell, Tissue Organ Culture 2001;64:93–114. [40] Nishizawa Y, Kishimoto N, Saito A, Hibi T. Sequence variation, differential expression and chromosomal location of rice chitinase genes. Mol Gen Genet 1993;241:1–10. [41] Nishizawa Y, Nishio Z, Nakazono K, Soma M, Nakajima E, Ugaki M, et al. Enhanced resistance to blast (Magnaporthe grisea) in transgenic japonica rice by constitutive expression of rice chitinase. Theor Appl Genet 1999;99:383–90. [42] Ohata K, Goto K, Kozaka T. Observations on the reaction of rice cells to the infection of different races of Pyricularia oryzae. Annu Phytopathol Soc Jpn 1963;28:24–30. [43] Orbach MJ, Farrall L, Sweigard JA, Chumley FG, Valent B. A telomeric avirulence gene determines efficacy for the rice blast resistance gene Pi-ta. Plant Cell 2000;12:2019–32. [44] Ou SH. Rice diseases. 2nd ed. Kew, England: Commonw. Agric. Bur. [45] Peng YL, Shishiyama J. Timing of a cellular reaction in rice cultivars associated with differing degrees of resistance to Pyricularia oryzae. Can J Bot 1985;67:2704–10. [46] Peng YL, Shishiyama J, Yamamoto M. A whole-leaf staining and clearing procedure for analysing cytological aspects of interaction between rice plant and rice blast fungus. Annu Phytopathol Soc Jpn 1986;52:801–8. [47] Rauyaree P, Choi W, Fang E, Blackmon B, Dean R. Genes expressed during early stages of rice infection with the rice blast fungus Magnaporthe grisea. Mol Plant Pathol 2001;2:347–54.

159

[48] Reimmann C, Ringli C, Dudler R. Complementary DNA cloning and sequence analysis of a pathogen-induced putative peroxidase from rice. Plant Physiol 1992;100:1611–2. ´ , Benhamou N, Datnoff LE, Jones JB, Be´langer RR. [49] Rodrigues FA Ultrastructural and cytochemical aspects of silicon-mediated rice blast resistance. Phytopathology 2003;93:535–46. ´ , McNally DJ, Datnoff LE, Jones JB, Labbe´ C, [50] Rodrigues FA Benhamou N, et al. Silicon enhances the accumulation of diterpenoid phytoalexins in rice: a potential mechanism for blast resistance. Phytopathology 2004;94:177–83. [51] Sambrook J, Russell DW. Molecular cloning: a laboratory manual. New York: Cold Spring Harbor Laboratory; 2001. [52] Schaffrath U, Thieron M, Scheinpflug H, Reisener HJ. Biochemical and cytological aspects of genetical and acquired resistance in the rice blast interaction. Cahiers Options Me´diterrane´ennes 1995;15: 125–30. [53] Seebold KW, Kucharek TA, Datnoff LE, Correa-Victoria FJ, Marchetti MA. The influence of silicon on components of resistance to blast in susceptible, partially resistant, and resistant cultivars of rice. Phytopathology 2001;91:63–9. [54] Silue D, Nottenghem JL, Tharreau D. Evidence for a gene-for-gene relationship in the Oryza sativa-Magnaporthe grisea pathosystem. Phytopathology 1992;82:577–80. [55] Simmons CR, Litts JC, Huang N, Rodriguez RL. Structure of a rice b-glucanase gene regulated by ethylene, cytokinin, wounding, salicylic acid and fungal elicitors. Plant Mol Biol 1992; 18:33–45. [56] Song F, Goodman RM. Molecular biology of disease resistance in rice. Physiol Mol Plant Pathol 2001;59:1–11. [57] Sridhar R, Mahadevan A. Physiology and biochemistry of rice plants infected by Pyricularia oryzae, Helminthosporium oryzae, Xanthomonas oryzae and Xanthomonas translucens f.sp. oryzicola. Acta Phytopathol Acad Hung 1979;14:49–82. [58] Takahashi Y. Studies on the mechanism of resistance of rice plants to Piricularia oryzae. II. Pathological changes microscopically observed in host cells in which fungus hyphae do not grow well. Yamagata Univ Agric Sci Bull 1956;2:37–51. [59] Takahashi A, Kawasaki T, Henmi K, Shii K, Kodama O, Satoh H, et al. Lesion mimic mutants of rice with alterations in early signaling events of defense. Plant J 1999;17:535–45. [60] Tomita H, Yamanaka S. Studies on the resistance reaction in the rice blast disease caused by Pyricularia oryzae Cav. 1. The pathological changes in the early infection stage of the inner epidermal cells of leaf sheath. Annu Phytopathol Soc Jpn 1983;49:514–21. [61] Wang Z-X, Yano M, Yamanouchi U, Iwamoto M, Monna L, Hayasaka H, et al. The Pib gene for rice blast resistance belongs to the nucleotide binding and leucine-rich repeat class of plant disease resistance genes. Plant J 1999;19:55–64. [62] Xu JR, Staiger CJ, Hamer JE. Inactivation of the mitogen-activated protein kinase Mps1 from the rice blast fungus prevents penetration of host cells but allows activation of plant defense responses. Proc Natl Acad Sci USA 1998;95:12713–8. [63] Yoshida SS, Ohnishi Y, Kitagishi K. Histochemistry of silicon in rice plant. (III) The presence of cuticle-silica double layer in the epidermal tissue. Jpn J Soil Sci Plant Nutr 1962;8:107–13. [64] Zhang JT, Duan GM, Yu ZY. Relationship between phenylalanine ammonia-lyase (PAL) activity and resistance to rice blast. Plant Physiol Commun 1987;6:34–7. [65] Zhu Q, Dabi T, Beeche A, Yamamoto R, Lawton MA, Lamb C. Cloning and properties of a rice gene encoding phenylalanine ammonia-lyase. J Plant Mol Biol 1995;29:535–50. [66] Zieslin N, Ben-Zaken R. Peroxidase activity and presence of phenolic substances in penduncles of rose flowers. Plant Physiol. Biochem. 1993;31:333–9. [67] Zimmer EA, Jupe ER, Walbot V. Ribosomal gene structure, variation and inheritance in maize and its ancestors. Genetics 1988;120: 1125–36.