Planta (1997) 203: 470±479

Cleavage of chitinous elicitors from the ectomycorrhizal fungus Hebeloma crustuliniforme by host chitinases prevents induction of K+ and Cl) release, extracellular alkalinization and H2O2 synthesis of Picea abies cells Peter Salzer*, Gerhard Hebe, Achim Hager Botanisches Institut, Allgemeine Botanik und P¯anzenphysiologie, UniversitaÈt TuÈbingen, Auf der Morgenstelle 1, D-72076 TuÈbingen, Germany Received: 6 January 1997 / Accepted: 14 March 1997

Abstract. Rapid reactions comprising e‚ux of K+ and Cl), phosphorylation of a 63-kDa protein (pp63), extracellular alkalinization and synthesis of H2O2 are equally induced in cells of Picea abies (L.) Karst. by chitotetraose, colloidal chitin and cell wall elicitors from the ectomycorrhizal fungus Hebeloma crustuliniforme (Bull. ex Fries.) QueÂl. an ectomycorrhizal partner of spruce. Cleavage of fungal cell wall elicitors and of arti®cial chitin elicitors to monomeric and dimeric fragments by apoplasmic spruce chitinases (36-kDa class I chitinase, pI 8.0, and 28-kDa chitinase, pI 8.7; EC 3.2.1.14) equally prevented induction of these rapid reactions. Also, N-acetylglucosamine oligomers and elicitors from the fungal cell walls showed a similar dependence of their activity on the degree of polymerisation. From these results it is suggested that, during ectomycorrhiza formation, only some of the chitin-derived elicitors reach their receptors at the plant plasma membrane, initiating reactions of the hypersensitive response in the host cells. The remaining fungal elicitors will be degraded to varying extents by wall-localized chitinases of the host root, reducing the defence reactions of the plant and allowing symbiotic interactions of both organisms. Key words: Chitinase ± Chitin elicitor ± Ectomycorrhiza ± Hebeloma ± Picea ± Plant defence (suppression)

Introduction Formation of ectomycorrhizas by soil fungi belonging to the most advanced groups of Basidiomycetes and *Present address: Botanisches Institut der UniversitaÈt Basel, Hebelstrasse 1, CH-4056 Basel, Switzerland Abbreviations: DP = degree of polymerisation; HR = hypersensitive response; MS = mineral solution; Pi = inorganic phosphate; pp63 = phosphorylated protein of MW 63 kDa Correspondence to: P. Salzer; Fax: 41 (61) 267 23 30; E-mail: [email protected]

Ascomycetes is essential for survival of most forest trees of the northern hemisphere. Ectomycorrhizas formed by Hebeloma crustuliniforme (Bull. ex Fries.) QueÂl. are frequently found on roots of Picea abies (L.) Karst. During the formation of an ectomycorrhiza the fungus penetrates the root cortex and develops the so-called Hartig net by growing in the plant cell walls. Proper di€erentiation of the Hartig net is crucial for the symbiotic interaction, because in this area the mutualistic exchange of nutrients occurs (Brunner and Scheidegger 1992; Kottke et al. 1997). Surprisingly, the symbiotic fungus H. crustuliniforme releases elicitors constitutively (Salzer et al. 1996) like pathogenic fungi, e.g. Phytophthora species (WaldmuÈller et al. 1992; Zanetti et al. 1992). The elicitors from H. crustuliniforme induce a complex of rapid reactions, including e‚ux of Cl) and K+, in¯ux of Ca2+, phosphorylation of a 63-kDa protein and dephosphorylation of a 65-kDa protein, extracellular alkalinization and synthesis of H2O2 in suspension-cultured cells raised from P. abies roots (Salzer et al. 1996). Several elicitors from pathogenic fungi have been reported to trigger similar reactions in resistant plants: e.g. the proteinaceous elicitor cryptogein from Phytophthora cryptogea in tobacco cells (Viard et al. 1994), the glycolipid elicitors from Pseudomonas syringae (the so-called syringolides) in soybean cells (Atkinson et al. 1996) and the oligopeptide elicitor from Phytophthora megasperma in parsley cells (NuÈrnberger et al. 1994). In particular, release of K+, in¯ux of H+ and Ca2+ and the generation of active oxygen species are assumed to be involved in the hypersensitive response (HR), which is known to be one of the most ecient defence reactions of plants (Atkinson et al. 1990; Naton et al. 1996). This raises the question of how ectomycorrhizas can be established in spite of these reactions. In this context, regulation of plant defence reactions during nodule formation by symbiotic Rhizobium bacteria on legume roots is an interesting example. The Nod factors which function as morphogenetic signals and possess a backbone consisting of four to ®ve N-acetylglucosamine residues (Fisher and Long 1992) can also induce defence

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

471

reactions in plant cells (Mellor and Collinge 1995). But the elicitor activity of the Nod factors can be eliminated by cleaving them by the action of plant chitinases (Staehelin et al. 1994a,b). Also, N-acetylglucosamine oligomers have been shown to act as elicitors (Felix et al. 1993) which can induce plant defence responses, such as ligni®cation and synthesis of pathogenesis-related proteins (Kurosaki et al. 1988; Ride and Barber 1990; Flach et al. 1993), but only, if their degree of polymerisation (DP) is >3. Using spruce cells, we recently demonstrated that constitutively secreted enzymes, which also contained chitinase activity, could inactivate elicitors released from walls of the ectomycorrhizal fungus H. crustuliniforme. This ®nding led us to suggest that, during ectomycorrhiza formation, inactivation by plant chitinases of chitin elicitors released from fungal cell walls might play a decisive role, too (Salzer et al. 1996). In order to obtain more information about the elicitors of H. crustuliniforme and a better understanding of the mechanism of suppression of rapid defence reactions in spruce cells we isolated and characterized elicitors from walls of this ectomycorrhizal fungus and compared their mode of action with that of de®ned N-acetylglucosamine oligomers. On the other hand, we puri®ed two chitinases from the culture medium of spruce cells and studied their e€ects on fungal elicitors and arti®cial chitin elicitors in inducing K+ and Cl) e‚ux, H2O2 synthesis and extracellular alkalinization in spruce cells.

Preparation of elicitors. Fungal cell walls were prepared from hyphae of three- to four-week-old cultures of H. crustuliniforme as previously described (Salzer et al. 1996). Brie¯y: 1 g of hyphae was consecutively washed in 40 ml bidistilled water, in 40 ml 0.1 M and in 40 ml 0.5 M potassium phosphate (KPi) bu€er (pH 7.2). Then the hyphae were homogenized in 0.5 M KPi bu€er, using an Ultraturrax T25 (Janke and Kunkel, Staufen i. Br., Germany) for 4 ´ 30 s (24 000 rpm), and soni®ed 4 ´ 20 s (70 W, Soni®er B12; Branson, Danbury, Conn., USA). The homogenate was centrifuged (4000 á g, 5 min) and the pellet was subsequently washed eight times in 0.5 M, then eight times in 0.1 M KPi bu€er, eight times in bidistilled water, and was ®nally dried at ambient temperature. To release soluble elicitors for HPLC analysis, 0.5 g of the fungal walls was suspended in 20 ml bidistilled water and shaken for 20 min at room temperature. Then the insoluble wall fragments were removed by centrifugation (4000 á g, 5 min) and the supernatant was concentrated by evaporation under vacuum at 40 °C to about 300 ll. For bioassays, 6 mg dry weight (DW) of the ®nely ground fungal walls was agitated for 20 min in 2 ml mineral solution (MS) containing 10 mM NaNO3, 1 mM KCl, 1 mM Na2SO4, 1 mM Mg(NO3)2 and 1 mM Ca(NO3)2, which was bu€ered with 25 mM 2-(N-morpholino)ethanesulfonic acid (Mes)-NaOH at pH 5.2. The soluble elicitors were separated from the insoluble wall particles by centrifugation (4000 á g, 5 min). For measurement of extracellular alkalinization by spruce cells the Mes bu€er was omitted or removed by AG 501 X8. Colloidal chitin was prepared from crab shell chitin according to Berger and Reynolds (1958), adjusted to a concentration of 16.6 mg á ml)1 and stored in bidistilled water at 4 °C. The supernatant of the colloidal chitin contained soluble chitin fragments which were used as elicitors. Aqueous solutions of tetraacetylchitotetraose, triacetylchitotriose, diacetylchitobiose and N-acetylglucosamine were freshly prepared before use.

Materials and methods

Induction of rapid reactions in spruce cells. Extracellular alkalinization, K+ and Cl) e‚ux were measured essentially as previously described (Salzer et al. 1996). For experiments, spruce cells from 5to 8-d-old cultures were separated from the culture medium by ®ltration with a nylon net (mesh 10 lm) and were washed in MS. To measure K+ and Cl) e‚ux, 5 g (wet weight) spruce cells was incubated in 20 ml of MS bu€ered with 25 mM Mes-NaOH (pH 5.2) and shaken for about 60 min at 120 rpm before the elicitors were added. To measure extracellular alkalinization, 1 g cells was suspended in 20 ml MS without bu€er. The Cl) and K+ concentration in the incubation medium was measured with ion-sensitive electrodes (model 94 17B chloride electrode and model 93-19 potassium electrode; Orion research, Boston, Mass., USA) with an Ag/AgCl double-junction reference electrode (Orion; model 90 02). The pH of the cell suspension was measured with half-micro pH electrodes (Ingold, Steinbach, Germany) which were connected to a Research Expandable Ion Analyzer EA 940 (Orion, Cambridge, Mass., USA). Instrument management and data acquisition were controlled by Quick Basic programs developed for IBM-compatible PCs by Dr. H. Stransky (Botanisches Institut TuÈbingen, Germany). To measure H2O2 synthesis, the spruce cells were washed with a bu€er containing 24.7 mM KNO3, 1.1 mM NaH2PO4, 1 mM MgSO4, 1 mM (NH4)2SO4, 1 mM CaCl2 and 2% (w/v) sucrose adjusted to pH 6.6 with a 20 mM Mes-NaOH bu€er. Two grams of P. abies cells was incubated in 20 ml of the bu€er for about 60 min before the elicitors were added. The concentration of H2O2 in the bu€er was determined by the luminol method as described by Schwacke and Hager (1992).

Chemicals. Media for low-pressure chromatography and molecular-weight standards for SDS-PAGE were from Pharmacia Biotech (Uppsala, Sweden). The Aminex HPX 42 A carbohydrate column, de-ashing cartridges, AG 501 X8 ion exchanger and two-dimensional electrophoresis protein standards were obtained from BioRad (Richmond, Cal., USA). Ultra®ltration membranes and Centricons were purchased from Amicon (Witten, Germany). Chemicals for protein sequencing were from Beckman (Munich, Germany), Riedl de HaeÈn (Hannover, Germany) and Applied Biosystems (Munich, Germany); Protogel from HoÈlzel (Manville, N.Y., USA); naphthalene-1-acetic acid from Serva (Heidelberg, Germany); 2,4-dichlorophenoxyacetic acid from Aldrich (Steinheim, Germany). Radiochemicals and Hyper®lm MP were from Amersham Buchler (Braunschweig, Germany). All other chemicals were from Sigma (Deisenhofen, Germany) or Merck (Darmstadt, Germany). If available chemicals of analytical grade were used. All media were prepared with bidistilled water. Suspension culture of fungi and spruce cells. Hebeloma crustuliniforme (Bull. ex Fries.) QueÂl. (strain TuÈ 704) which was isolated from a fruit body growing underneath Picea abies (L.) Karst. was grown on modi®ed Melin Norkrans-agar in 8-cm petri dishes as described by Sirrenberg et al. (1995). To obtain suspension cultures, hyphae grown in two petri dishes were inoculated in 200 ml of liquid modi®ed Melin Norkrans medium and were shaken at 100 rpm at 22±24 °C in the dark. Suspension cultures which were raised from calli developing on P. abies roots were grown in Gamborg's 4X medium as described by Salzer et al. (1996). For bioassays 5- to 8-d-old cultures were used. To isolate chitinases from the culture medium, 40 ml of a 7-dold suspension culture was transferred to 60 ml Gamborg's 4X medium without auxins (9 lM 2,4-dichlorophenoxyacetic acid, 2.7 lM naphthalene-1-acetic acid and 2.7 lM IAA were omitted) and was grown for 10 d.

In-vivo phosphorylation. Spruce cells (0.25 g) were incubated in 1.3 ml MS bu€ered with 25 mM Mes-NaOH at pH 5.2 and were shaken for 60 min on a rotatory shaker at ambient temperature. Then, 30 ll elicitors released from cell walls of H. crustuliniforme or 30 ll elicitors from colloidal chitin or tetraacetylchitotetraose (®nal concentration 3 lM) or triacetylchitotriose (®nal concentration 50 lM) was applied. The phosphorylation reaction was

472

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

stopped after 5 min by adding 260 ll of 30% (w/w) trichloroacetic acid and transferring the cells into liquid nitrogen. Carrier-free [32P]Pi (3370 kBq) was applied as a pulse 3 min prior the reaction stop. Modifying the protocol of Felix et al. (1993), the precipitated proteins were washed in 600 ll of 90% (v/v) acetone (bu€ered with 5 mM Tris-HCl, pH 8.0), then with 600 ll of 80% (v/v) acetone (bu€ered with 10 mM Tris-HCl, pH 8.0). The proteins were solubilized in electrophoresis bu€er by heating to 90 °C for 3 min and SDS-PAGE was performed according to Laemmli (1970) using a 5% acrylamide stacking gel and a 10% separation gel, loading 20 lg proteins per lane. Finally, the gels were dried and the phosphorylated proteins were detected by autoradiography with a Hyper®lm MP. Puri®cation and characterization of chitinases. Spruce cells were removed from the culture medium by ®ltration (nylon net, mesh 10 lm), then the culture ®ltrate (1.5 l) was centrifuged (17 700 á g, 60 min). Centrifugation and all following puri®cation steps were conducted on ice or at 4 °C. Proteins of the supernatant were precipitated with ammonium sulfate (70% saturation, 60 min), centrifuged (17 700 á g, 60 min) and dissolved in 50 ml of a 20 mM NaPi-citrate bu€er, pH 5.2. After removing insoluble precipitates by centrifugation (38 000 á g, 120 min), 100 ml NaHCO3 (20 mM) was added and the pH adjusted to 8.2 by addition of NaOH. Modifying the procedure of Molano et al. (1979), the proteins were loaded on a column (2.6 cm in diameter) of regenerated chitin (freshly prepared from 3.5 g chitosan and autoclaved) at a linear ¯ow rate of 0.3 cm á min)1. Then the column was washed with 20 mM NaHCO3 until the eluate was free of proteins. A second washing step with 20 mM sodium acetate-acetic acid bu€er (pH 5.6) followed for 90 min. Finally, chitin-binding proteins were released with 20 mM acetic acid (pH 3.3) and were transferred to a 20 mM NaPi-citrate bu€er (pH 7.3) using Pharmacia PD 10 columns. The chitin-binding proteins were concentrated by ultra®ltration (Amicon cell; YM 10 Dia¯o membranes) to a volume of about 5 ml and were loaded on a carboxymethyl (CM) Sepharose CL-6B column (1.6 cm in diameter, bed volume 28 ml), which was equilibrated with 20 mM NaPi-citrate bu€er (pH 7.3). After sample loading, the NaCl concentration in the bu€er was linearly increased from 0 to 0.15 M (0 to 180 min), then from 0.15 to 0.4 M NaCl (180 to 540 min). At a linear ¯ow rate of 0.11 cm á min)1, fractions of 2 ml were collected and the chitinase activity was measured. Those fractions with high chitinase activities were separately concentrated by ultra®ltration (Centricon; YM 10) to a ®nal volume of about 60 ll and analysed by SDS-PAGE. For experiments, fractions were used in which only a single protein band at 28 kDa or 36 kDa was detected in the gels by silver staining (Ansorge 1985). Analysis by SDS-PAGE, renaturation of chitinases using Triton X-100 (puri®ed with AG 501 X8) and activity staining with glycol chitin as substrate were performed entirely as described by Trudel and Asselin (1989). The ¯uorescing gels (excitation 302 nm), lying on a blue glass plate, were photographed with a polaroid camera in front of which a Kodak Wratten Nr. 2A Gelatin Filter (Integra Biosciences, Fernwald, Germany) was mounted. For isoelectric focussing, chitinases were dissolved in O'Farrell bu€er (O'Farrell 1975) and samples of 3 ll were focussed on ampholyte gradients ranging from pH 3 to 10, as described by Ho€mann and Hampp (1994). After electrophoresis the gels were ®xed in 10% (w/v) sulfosalicylic acid and stained with Coomassie Brilliant Blue according to Neuho€ et al. (1985). For determination of the isoelectric point two-dimensional protein standards were used. For N-terminal sequencing, spruce chitinases (36 kDa and 28 kDa) bound to Glassybond membranes (Biometra, GoÈttingen, Germany) were washed with water and dried. The proteins were introduced into the reaction cartridge of the amino acid sequencer (3600 LF with on-line PTH analyser; Beckman Instruments, Fullerton, Calif., USA) and sequenced by Edman degradation, as modi®ed by Hunkapiller et al. (1983). The phenylthiocarbamyl (PTC) adducts of each degradation cycle were analysed on-line with a reversed-phase HPLC system (Spherogel Micro PTH; 125 mm

long, 2 mm i.d.; Beckman) following the manufacturer's instructions. The amino acids were identi®ed by comparison of their retention times with those of a standard chromatogram obtained from a mixture of 20 amino acids. Chitinase activity was measured with colloidal chitin, and b-1,3glucanase activity with laminarin as substrate as previously described (Sauter and Hager 1989; Salzer et al. 1996). Activity of acidic proteases was measured as described by Stellmach et al. (1988) with the modi®cation that bovine serum albumin was used as substrate and 25 mM Mes (pH 5.2) as reaction bu€er. Chitinase treatment of elicitors released from fungal cell walls and colloidal chitin. For bioassays, 5 ll of 36-kDa chitinase (0.32 pkat á ll)1) or 28-kDa chitinase (0.36 pkat á ll)1) was added to 500 ll fungal elicitors dissolved in MS bu€ered with 25 mM Mes-NaOH (pH 5.2). After a 90-min incubation period at 37 °C the enzymatic

Fig. 1A,B. Induction of H2O2 synthesis, K+ and Cl) release and phosphorylation of pp63 by tetraacetylchitotetraose. A Spruce cells were treated with 3 lM tetraacetylchitotetraose (tetra) and the concentration of H2O2, K+ and Cl) was measured in the medium. Water was added to control samples instead of the aqueous solution of tetraacetylchitotetraose. B Spruce cells were either treated with tetraacetylchitotetraose 3 lM (tetra), or with elicitors from colloidal chitin (coll. chit.) or elicitors from cell walls of H. crustuliniforme (H. crust.) or with 50 lM triacetylchitotriose (tri ). After a 5-min period the elicitor-induced phosphorylation reaction was stopped and the radioactive labeling of pp63 was analysed by SDS-PAGE and autoradiography. Similar results were obtained in three independent experiments

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases Table 1. Puri®cation of the 36-kDa and 28-kDa chitinase isoforms from the culture medium of spruce cells. Puri®cation started with 1.5 l of culture ®ltrate obtained from cells grown in medium without auxins and fungal elicitors for 10±12 d. The speci®c activity, puri®cation and yield are given as the sum of the 36- and 28 kDa chitinases, because separation of the two isoforms was not complete in all fractions

Puri®cation step

473 Speci®c activity (pkat á lg)1)

Picea abies suspension culture

Puri®cation (-fold)

Yield (%)

2.5

1

100

2.7

1.1

59.5

18.5

7.3

2.4

20.0

7.8

1.6

®ltration, centrifugation ammonium sulfate precipitation

Proteins dissolved at pH 5.2

centrifugation chromatography on regenerated chitin

Proteins released at pH 3.3

gel®ltration (PD 10) ultra®ltration CM Sepharose CL-6B

36-kDa chitinase

reaction was stopped by heating the samples for 2 min to 95 °C. In samples being incubated without chitinases the enzymes were separately heat-inactivated and thereafter added to the samples. Before the extracellular alkalinization response of spruce cells was measured, 25 mg of AG 501 X8 ion exchanger was added to each sample, and shaken for 30 min. For monitoring K+ and Cl) e‚ux and synthesis of active oxygen species, removal of charged compounds was not necessary. Elicitors released from colloidal chitin were treated under identical conditions, except the incubation volume was 230 ll and the incubation time 60 min. Analysis of products released by chitinases from fungal elicitors. The elicitors which had been released from 0.5 g of cell walls from H. crustuliniforme and which had been concentrated to about 300 ll were incubated with 10 mg of AG 501 X8 ion exchanger for 45 min. Then, a 50-ll aliquot of the elicitors was incubated with 3.2 pkat of 36-kDa chitinase or 3.6 pkat of 28-kDa chitinase. Incubation was in a total volume of 140 ll in MS bu€ered with 25 mM Mes-NaOH (pH 5.2) for 16 h at 37 °C. Before HPLC analysis the charged compounds were removed using AG 501 X8. Then, the elicitors were injected in an injection loop (110 ll) which was coupled to a Kontron HPLC system (Kontron HPLC pump 420, Kontron data system 450) and an Aminex HPX 42A carbohydrate column in front of which two de-ashing re®ll cartridges were installed. Chromatography was at 80 °C, at a ¯ow of 0.4 ml á min)1 with bidistilled, degassed water as eluent. The cleavage products were detected with a 2142 LKB di€erential refractometer (range 1) using water as reference. For determination of the DP of the cleavage products and for calibration of the refractometer, N-acetylglucosamine, N,N¢-diacetylchitobiose, N,N¢,N¢¢-triacetylchitotriose and N,N¢,N¢¢,N¢¢¢-tetraacetylchitotetraose were used as standards. Treatment of HPLC-puri®ed elicitors (DP > 10) with chitinases occurred under the same conditions as described above. To analyse cleavage products from particulate substrates, colloidal chitin (0.5 mg) and fungal cell walls (0.75 mg) were treated with 1.5 pkat chitinase in a volume of 250 ll of MS, which was bu€ered with 25 mM Mes-NaOH (pH 5.2) in an Eppendorf

28-kDa chitinase

thermoshaker at 37 °C. The enzymatic reactions were stopped by heating the samples to 95 °C for 2 min. Insoluble particles were removed by centrifugation (4000 á g, 3 min) and charged compounds were removed using AG 501 X8.

Results Rapid reactions induced in spruce cells by tetraacetylchitotetraose, colloidal chitin and elicitors from H. crustuliniforme. Release of K+ and Cl) from spruce cells occurred with a similar time course when induced with

Fig. 2. SDS-PAGE of the 28-kDa and the 36-kDa chitinases puri®ed from the culture medium from spruce cells. Coomassie-stained protein bands coincide with bands obtained by activity staining after renaturation of the chitinases. M, molecular-weight markers

474

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

elicitors from colloidal chitin, from fungal cell walls and tetraacetylchitotetraose. The reactions always commenced with an e‚ux of Cl), followed by a release of K+, an extracellular alkalinization and the synthesis of H2O2 (Salzer et al. 1996). Interestingly, phosphorylation of a 63-kDa protein (pp63) was induced by elicitors from cell walls and colloidal chitin, and by chitotetraose, but not by chitotriose (Fig. 1). These results demonstrate that suspension-cultured spruce cells, which were raised from calli of P. abies roots, respond similarly to chitinderived and fungal elicitors. Chitinases puri®ed from the culture medium of P. abies cells. A 36-kDa chitinase (pI 8.0) and a 28-kDa chitinase (pI 8.7) were puri®ed from the culture medium of spruce cells. The main puri®cation data are summarized in Table 1. The SDS-PAGE of the enzymes and subsequent analysis of parallel lanes by Coomassie Blue and activity staining after renaturing the chitinases resulted in corresponding bands at 28 kDa and 36 kDa (Fig. 2). In addition, the 36-kDa isoform was identi®ed as class I chitinase by N-terminal sequencing. The 28-kDa isoform was blocked and therefore not accessible for direct N-terminal sequencing. Measuring the activities of

Fig. 3. Induction of K+ and Cl) release from spruce cells is prevented by treatment of H. crustuliniforme elicitors with spruce chitinases. Elicitors had been pretreated either with 1.6 pkat of 36kDa chitinase or 1.7 pkat of 28-kDa chitinase for 90 min, at 37 °C. In samples with no chitinase heat-inactivated enzymes were added to the elicitors after the 37 °C incubation period. Soluble elicitors released from cell walls (1.2 mg DW) were added to spruce cells. For calibration an aliquot of 100 mM KCl solution was added to the cells. The bars give the signals caused by addition of 1 lmol K+ or 0.1 lmol Cl) to the samples. The diagram shows a typical result from three independent experiments

Fig. 4. Synthesis of H2O2 by spruce cells is prevented by treatment of H. crustuliniforme elicitors with spruce chitinases. Elicitors released from fungal cell walls (1.2 mg DW) were treated for 90 min with chitinase (1.6 pkat 36-kDa chitinase or 1.7 pkat 28-kDa chitinase). Heat-inactivated chitinase was added to the elicitors after the 90-min incubation period at 37 °C. Elicitors or pretreated elicitors were applied to the cells at time zero. Calibration was done by addition of a de®ned amount of H2O2 to the cells at the end of the incubation period. The diagram shows a typical time course of H2O2 synthesis as found in two independent experiments

Fig. 5. Decomposition of colloidal chitin elicitors by spruce chitinases reduces K+ and Cl) release from spruce cells. A supernatant from colloidal chitin (225 ll) containing active elicitor molecules was treated for 60 min with chitinase (1.7 pkat) at 37 °C and was added to the spruce cells. In control samples inactivated chitinase was added to the chitin elicitors after the 37 °C incubation period. The diagram shows a typical result from three independent experiments

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

475

Fig. 6. Treatment of colloidal chitin elicitors with spruce chitinases reduces induction of H2O2 synthesis by spruce cells. Elicitors released from colloidal chitin (225 ll) were treated for 60 min at 37 °C with 36-kDa (1.6 pkat) or 28-kDa (1.7 pkat) chitinase, then they were added to spruce cells. Heat-inactivated chitinase was added to the elicitors after the 37 °C incubation period. The diagram shows a typical time course from two independent experiments

b-1,3-glucanase using laminarin as substrate and of acidic protease using bovine serum albumin as substrate further indicated that the puri®ed enzymes contained no foreign enzymes capable of cleaving presumptive proteinaceous and glucan elicitors. Inactivation of elicitors released from colloidal chitin and from cell walls of H. crustuliniforme by spruce chitinases. Treatment of elicitors released from cell walls of H. crustuliniforme by the 28-kDa or 36-kDa chitinase prevented the elicitor-induced release of K+ and Cl) (Fig. 3), the synthesis of H2O2 (Fig. 4) and the extracellular alkalinization (not shown) almost completely. A second application of untreated elicitors to spruce cells already treated with inactivated elicitors induced a full response, indicating that the inactivation was not due to the generation of suppressors by the action of the enzymes. In the same way as for fungal elicitors, both chitinase isoforms prevented induction of K+ and Cl) e‚ux (Fig. 5) and synthesis of H2O2 (Fig. 6) by elicitors from colloidal chitin in spruce cells. Analysis of the cleavage products generated by chitinases from the fungal elicitors revealed that monomers and dimers were formed (Fig. 7). Concomitantly with the formation of the cleavage products, elicitor activity vanished, as determined using extracellular alkalinization as a bioassay (not shown). Furthermore, treatment of HPLC-puri®ed cell wall fragments (DP > 10) with 28-kDa chitinase resulted exclusively in the formation of monomeric products (Fig. 8) which

Fig. 7A±C. Formation of monomeric and dimeric cleavage products from H. crustuliniforme elicitors. Soluble components were released from cell wall preparations of H. crustuliniforme (0.5 g DW) and were treated either with 36-kDa (3 pkat) or 28-kDa (3.2 pkat) chitinase for 16 h, at 37 °C or were incubated under identical conditions without enzymes. A Chromatogram of the soluble wall constituents without enzymatic treatment, B after treatment with 28-kDa chitinase, and C after treatment with 36-kDa chitinase. The DP was determined by chromatography of de®ned N-acetylglucosamine oligomers (n = 1±4) as standards. Peaks of DP > 4 were consecutively numbered. The diagram shows a typical result from three independent experiments

were proven to be N-acetylglucosamine by the method of Cabib and Bowers (1971) using dimethylaminobenzoate as dye. Moreover, both enzymes predominantly released monomers, dimers and trimers from colloidal chitin and cell walls of H. crustuliniforme (Fig. 9). There are slight di€erences in the ratio of monomers to oligomers released by the action of chitinases from di€erent substrates. They are probably caused by di€erent solubilities and structures of the substrates and the di€erent reaction times. Activities of H. crustuliniforme elicitors and of N-acetylglucosamine oligomers depend in the same way on their DP. To study the in¯uence of the DP on the e€ectiveness of cell wall elicitors from H. crustuliniforme, as a ®rst

476

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

Fig. 8. Cleavage of elicitors with a DP > 10 to monomeric products by the action of the 28-kDa chitinase. Soluble components were released from cell wall preparations of H. crustuliniforme and were separated according to their DP. The fractions containing the highest elicitor activity (DP > 10) were pooled. Insert, corresponding HPLC separation of the oligomers which had been released from the fungal cell walls. One part of the elicitor sample was treated with 28-kDa chitinase (3.2 pkat) for 16 h at 37 °C, the other part was incubated under the same conditions but without chitinase. Chitinase activity exclusively released monomeric products from the elicitors. The diagram shows a typical result from 3 independent experiments

step elicitors were released from cell walls of the fungus and were separated according to their DP by HPLC using an Aminex HPX 42 A carbohydrate column. Then, the fractions containing fragments with an increasing DP were analysed for their capacity to induce extracellular alkalinization in spruce cells (Fig. 10). Like N-acetylglucosamine oligomers (Fig. 11), cell wall fragments only exerted full activity as elicitors if their DP was >3. Discussion Spruce cells raised from P. abies roots respond to elicitors from colloidal chitin and to chitotetraose very rapidly with an e‚ux of Cl) and K+, the phosphorylation of the protein pp63, an extracellular alkalinization and the synthesis of H2O2. All these reactions, including Ca2+ in¯ux, as well as increased expression of chitinase, b-1,3-glucanase and peroxidase were induced by cell wall elicitors from ectomycorrhizal fungi in spruce cells (Sauter and Hager 1989; Salzer and Hager 1993; Salzer

Fig. 9. Similar products were released from H. crustuliniforme cell walls and colloidal chitin by the action of spruce chitinases. Colloidal chitin and cell wall preparations from H. crustuliniforme were treated either with 36-kDa chitinase (1.5 pkat) or 28-kDa chitinase (1.6 pkat) for 90 min, at 37 °C. Then, the cleavage products were separated according to their DP. The DP was determined with N-acetylglucosamine oligomers as standards. The diagram shows a typical result from three independent experiments

et al. 1996; Salzer et al. in press). These rapid reactions, in particular, show similarities to initial reactions of the HR in other plant cells challenged by incompatible pathogens. For instance, release of K+ accompanied by proton uptake (resulting in extracellular alkalinization; Mathieu et al. 1996) characterizes the HR in tobacco (Atkinson et al. 1990); accumulation of H2O2 was shown to coincide with rapid cell death in a simpli®ed experimental system consisting of single parsley cells infected by Phytophthora megasperma f. sp. glycinea (Naton et al. 1996); Ca2+ in¯ux is involved in signal transduction in spruce cells as well as in signal transduction of the HR in tobacco (Atkinson et al. 1990; Salzer et al. 1996); also, the phosphorylation of proteins seems to be a part of the signal transduction cascade in spruce cells as it was demonstrated for reactions induced by cryptogein, an elicitor of the HR in cells of tobacco (Viard et al. 1994; Tavernier et al. 1995; Salzer et al. 1996). These parallels demonstrate that cells from the host tree and the ectomycorrhiza-forming fungus are capable of reacting like incompatible partners in

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

477

Fig. 10. Extent of the extracellular alkalinization of spruce cells depends on the DP of the H. crustuliniforme cell wall elicitors. Soluble components were released from cell wall preparations of H. crustuliniforme and were separated according to their DP. Elicitor activity of the oligomers was determined using the extracellular alkalinization response of spruce cells as bioassay; leq (DP 1±4) were determined using N-acetylglucosamine, chitobiose, chitotriose, and chitotetraose as standards. The diagram shows results typically obtained in three independent experiments

plant-pathogen interactions. To establish a balanced symbiotic association, therefore, control of these defence reactions is necessary. In many examples it has been demonstrated that the suppression of plant defence reactions is the basis for mycorrhiza formation. The most striking ®nding was that in myc) mutants of pea where fungal strains of Glomus mosseae failed to di€erentiate arbuscular mycorrhizal structures because of an induced barrier formation in the plant tissue underneath the appressoria (Gollotte et al. 1993). Another system in which the suppression of defence reactions during mycorrhizal interactions was unequivocally shown is the reduction of phytoalexin synthesis in alfalfa roots colonized by Glomus intraradix (Volpin et al. 1995). Generally, elicitors are the signal molecules for defence responses in plants. The chemical nature of these elicitors is rather heterogeneous. They belong to di€erent chemical categories, such as oligosaccharides, glycopeptides, peptides, proteins and lipophilic substances (Boller 1995). Three ®ndings strongly indicate that the elicitors released from cell walls of H. crustuliniforme originate from chitin. (i) The action of the fungal elicitors and the elicitors from colloidal chitin in inducing Cl) and K+ e‚ux, as well as H2O2 synthesis and extracellular alkalinization can equally be prevented by treatment with highly puri®ed chitinases. (ii) The chromatographic properties of the fungal elicitors and of de®ned N-acetylglucosamine oligomers on an Aminex HPX 42 A carbohydrate column are identical. (iii) The

Fig. 11. Extent of the extracellular alkalinization of spruce cells depends on the DP of the N-acetylglucosamine oligomers. Commercial N-acetylglucosamine (n = 1), N,N¢-diacetylchitobiose (n = 2), N¢, N¢¢-triacetylchitotriose (n = 3), N,N¢,N¢¢,N¢¢¢-tetraacetylchitotetraose (n = 4) were tested for their ability to induce extracellular alkalinization in spruce cells. Values in parentheses give the ®nal concentrations in the bioassay. Insert, dependence of the alkalinization response of spruce cells on the concentration of tetraacetylchitotetraose. Similar curves were obtained in two independent experiments

action of the cell wall elicitors and N-acetylglucosamine oligomers is dependent on the same DP. In both cases full activity was only found if the DP was >3. Monomers and dimers were inactive, whereas trimers were only active in high concentrations. This correlation of DP and activity of N-acetylglucosamine oligomers was also found for the induction of extracellular alkalinization in tomato cells (Felix et al. 1993). In agreement with this correlation was our ®nding that a signi®cant phosphorylation of pp63 in spruce cells could only be induced with tetraacetylchitotetraose but not with triacetylchitotriose. However, if the release of chitin fragments is an inevitable and permanent process during fungal development, how can the symbiotic partners cope with these elicitors? In ectomycorrhizas, one possibility could be an inactivation of the elicitors by enzymes of the host plant. Constitutively expressed chitinase localized in the apoplasmic space of the root cortex could cleave the chitinderived elicitors during their passage through the plant cell wall before they reach their receptors in the plant

478

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases

plasma membrane (Baureithel et al. 1994; Salzer et al. 1996; Salzer et al. 1997). To prove this hypothesis we puri®ed two constitutively secreted chitinase isoforms from the culture medium of spruce cells, a 36-kDa isoform (pI 8.0) which was shown by N-terminal sequencing to have high homology to class I chitinases of pea (Vad et al. 1991), and a 28-kDa isoform (pI 8.7). Indeed, treatment of elicitors from H. crustuliniforme by both chitinase isoforms prevented the induction of K+ and Cl) release, H+ uptake and H2O2 synthesis in spruce cells almost completely. The chitinases degraded the elicitors to monomeric and dimeric products, which were no longer active as elicitors on the spruce cells. Thus, during ectomycorrhiza formation a similar mechanism could become operative as it is proposed to occur during nodulation processes. Mellor and Collinge (1995) assume in a ``hypothesis paper'' that Nod factors released by symbiotic Rhizobium bacteria might function both as morphogenetic signals and as elicitors. After triggering nodulation processes the Nod factors are thought to be inactivated by chitinases to avoid elicitorinduced defence reactions in the host plant. In fact, chitinases from various legumes have been demonstrated to eliminate elicitor activity of Nod factors by cleaving their N-acetylglucosamine backbone (Staehelin et al. 1994a). The high chitinase activities found in the youngest parts of the spruce roots (Sauter and Hager 1989) favour such an elicitor-destroying role of chitinases during ectomycorrhiza formation. Furthermore, the increased levels of chitinase activity found during di€erentiation of Pisolithus tinctorius ectomycorrhiza on eucalypt roots (Albrecht et al. 1994) could therefore be explained by such a symbiosis-related function of chitinases. We found that the putative formation of suppressor molecules during the degradation of elicitors by chitinases was not the reason for the reduction of the elicitor e€ects. Rather, it was demonstrated that a second application of elicitors to spruce cells could induce K+ release at full intensity irrespective of the chitinase-treated elicitors which were added at ®rst. However, when active elicitors were ®rstly added, the spruce cells did not respond a second time (data not shown). On the other hand, an example of suppressor formation by enzymatic cleavage of elicitors has been reported in the case of the treatment of elicitor-active invertase glycopeptide fragments by endo-b-acetylglucosaminidase H (Basse et al. 1992). Also, the chitinases themselves did not act as suppressors, as demonstrated by simultaneous addition of elicitors and chitinases to the spruce cells. For arbuscular mycorrhizas, too, there are presently no reports on the involvement of suppressors (Lambais and Mehdy 1993; Salzer et al. 1997). In this respect, mycorrhiza-forming fungi di€er from pathogenic microorganisms where many suppressor molecules of plant defence reactions have been identi®ed (Jakobek et al. 1993 and citations therein). Interestingly, the acidic exopolysaccharide EPS I from Rhizobium meliloti is assumed to function as a suppressor of plant defence reactions during nodule formation (Niehaus et al. 1993).

Chitin is a constituent of all ectomycorrhiza-forming fungi, and cells from mycorrhizal host plants respond to elicitors derived from chitin. Therefore, lowering the concentration of these chitinous elicitors below a value which is necessary to activate plant defence responses, might be a general precondition for a compatible interaction between ectomycorrhiza-forming fungi and woody plants. We thank Dr. H. Stransky (Botanisches Institut, TuÈbingen) for provision of his Quick Basic programs and for his help with HPLC problems. We also thank B. HuÈbner (TuÈbingen) for performing isoelectric focussing. This work was supported by the Deutsche Forschungsgemeinschaft. G. H. was ®nanced by a scholarship from the LandesgraduiertenfoÈrderungsgesetz.

References Albrecht C, Burgess T, Dell B, Lapeyrie F (1994) Chitinase and peroxidase activities are induced in eucalypt roots according to aggressiveness of Australian ectomycorrhizal strains of Pisolithus sp. New Phytol 127: 217±222 Ansorge W (1985) Fast and sensitive detection of protein and DNA bands by treatment with potassium permanganate. J Biochem Biophys Methods 11: 13±20 Atkinson MM, Keppler LD, Orlandi EW, Baker CJ, Mischke CF (1990) Involvement of plasma membrane calcium in¯ux in bacterial induction of the K+/H+ and hypersensitive responses in tobacco. Plant Physiol 92: 215±221 Atkinson MM, Midland SL, Sims JJ, Keen NT (1996) Syringolide 1 triggers Ca2+ in¯ux, K+ e‚ux, and extracellular alkalinization in soybean cells carrying the disease-resistance gene Rpg4. Plant Physiol 112: 297±302 Basse CW, Bock K, Boller T (1992) Elicitors and suppressors of the defence response in tomato cells. Puri®cation and characterization of glycopeptide elicitors and glycan suppressors generated by enzymatic cleavage of yeast invertase. J Biol Chem 267: 10258±10265 Baureithel K, Felix G, Boller T (1994) Speci®c, high anity binding of chitin fragments to tomato cells and membranes. Competitive inhibition of binding by derivatives of chitooligosaccharides and a Nod factor of Rhizobium. J Biol Chem 269: 17931±17938 Berger LR, Reynolds DM (1958) The chitinase system of a strain of Streptomyces griseus. Biochim Biophys Acta 29: 522±534 Boller T (1995) Chemoperception of microbial signals in plant cells. Annu Rev Plant Physiol Plant Mol Biol 46: 189±214 Brunner I, Scheidegger C (1992) Ontogeny of synthesized Picea abies (L.) Karst.-Hebeloma crustuliniforme (Bull. ex St. Amans) QueÂl. ectomycorrhizas. New Phytol 120: 359±369 Cabib E, Bowers B (1971) Chitin and yeast budding. J Biol Chem 246: 152±159 Felix G, Regenass M, Boller T (1993) Speci®c perception of subnanomolar concentrations of chitin fragments by tomato cells: induction of extracellular alkalinization, changes in protein phosphorylation, and establishment of a refractory state. Plant J 4: 307±316 Fisher RF, Long SR (1992) Rhizobium ± plant signal exchange. Nature 357: 655±660 Flach J, Jolles P, Pilet PE (1993) Induction of chitinase and b-1,3glucanase in Parthenocissus quinquefolia cells cultured in vitro. Physiol Plant 89: 399±403 Gollotte A, Gianinazzi-Pearson V, Giovanetti M, Sbrana C, Avio L, Gianinazzi S (1993) Cellular localization and cytochemical probing of resistance reactions to arbuscular mycorrhizal fungi in a `locus a' myc) mutant of Pisum sativum L. Planta 191: 112± 122

P. Salzer et al.: Cleavage of H. crustuliniforme elicitors by spruce chitinases Ho€mann EM, Hampp R (1994) Studies on vacuole regeneration in evacuolated tobacco mesophyll protoplasts. Protein analysis by one- and two-dimensional microgel-electrophoresis. Physiol Plant 92: 563±570 Hunkapiller MW, Hewick RM, Dreyer WJ, Hood LE (1983) High sensitivity sequencing with a gas phase sequenator. Methods Enzymol 91: 339±413 Jakobek JL, Smith JA, Lindgren PB (1993) Suppression of bean defence responses by Pseudomonas syringae. Plant Cell 5: 57±63 Kottke I, MuÈnzenberger B, Oberwinkler F (1997) Structural approach to function in ectomycorrhizas. In: Rennenberg H, Eschrich W, Ziegler H (eds). Trees ± contributions to modern tree physiology. SPB Academic Publ, The Hague pp. 357±376 Kurosaki F, Tashiro N, Nishi A (1988) Role of chitinase and chitin oligosaccharides in ligni®cation response of cultured carrot cells treated with mycelial cell walls. Plant Cell Physiol 29: 527± 531 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680±685 Lambais MR, Mehdy MC (1993) Suppression of endochitinase, b1,3-endoglucanase, and chalcone isomerase expression in bean vesicular-arbuscular mycorrhizal roots under di€erent soil phosphate conditions. Mol Plant Microbe Interactions 6: 75±83 Mathieu Y, Lapous D, Thomine S, Lauriere C, Guern J (1996) Cytoplasmic acidi®cation as an early phosphorylation-dependent response of tobacco cells to elicitors. Planta 199: 416±424 Mellor RB, Collinge DB (1995) A simple model based on known plant defence reactions is sucient to explain most aspects of nodulation. J Exp Bot 46: 1±18 Molano J, Polacheck I, Duran A, Cabib E (1979) An endochitinase from wheat germ. Activity on nascent and performed chitin. J Biol Chem 254: 4901±4907 Naton B, Hahlbrock K, Schmelzer E (1996) Correlation of rapid cell death with metabolic changes in fungus-infected, cultured parsley cells. Plant Physiol 112: 433±444 Neuho€ V, Stamm R, Eibl H (1985) Clear background and highly sensitive protein staining with Coomassie Blue dyes in polyacrylamide gels: a systematic analysis. Electrophoresis 6: 427± 448 Niehaus K, Kapp D, PuÈhler A (1993) Plant defence and delayed infection of alfalfa pseudonodules induced by an exopolysaccharide (EPS I)-de®cient Rhizobium meliloti mutant. Planta 190: 415±425 NuÈrnberger T, Nennstiel D, Jabs T, Sacks WR, Hahlbrock K, Scheel D (1994) High anity binding of a fungal oligopeptide elicitor to parsley plasma membranes triggers multiple defence responses. Cell 78: 449±460 O'Farrell PH (1975) High resolution two-dimensional electrophoresis of proteins. J Biol Chem 250: 4007±4021 Reissig JL, Strominger JL, Leloir LF (1955) A modi®ed method for estimation of N-acetylamino sugars. J Biol Chem 217: 959±966 Ride JP, Barber MS (1990) Puri®cation and characterization of multiple forms of endochitinase from wheat leaves. Plant Sci 71: 185±197 Salzer P, Hager A (1993) E€ect of auxins and ectomycorrhizal elicitors on wall-bound proteins and enzymes of spruce (Picea abies (L.) Karst.) cells. Trees 8: 49±55

479

Salzer P, Hebe G, Reith A, Zitterell-Haid B, Stransky H, Gaschler K, Hager A (1996) Rapid reactions of spruce cells to elicitors released from the ectomycorrhizal fungus Hebeloma crustuliniforme, and inactivation of these elicitors by extracellular spruce cell enzymes. Planta 198: 118±126 Salzer P, Mensen R, Hebe G, Gaschler K, Hager A (1997) Control of elicitor-induced reactions in spruce cells by auxin and by enzymatic elicitor degradation. In: Schenk HEA, Herrmann RG, Jeon KW, MuÈller NE, Schwemmler W (eds). Intertaxonic combination and symbiotic adaptation. Proc Sixth Int Colloquium on Endocytobiology and Symbiosis, TuÈbingen 1995, Springer, Heidelberg, in press Salzer P, MuÈnzenberger B, Schwacke R, Kottke I, Hager A (1997) Signalling in ectomycorrhizal fungus-root interactions. In: Rennenberg H, Eschrich W, Ziegler H (eds). Trees ± contributions to modern tree physiology. SPB Publ, The Hague, pp. 339±356 Sauter M, Hager A (1989) The mycorrhizal fungus Amanita muscaria induces chitinase activity in roots and in suspensioncultured cells of its host Picea abies. Planta 179: 61±66 Schwacke R, Hager A (1992) Fungal elicitors induce a transient release of active oxygen species from cultured spruce cells depending on Ca2+ and protein kinase activity. Planta 187: 136±141 Sirrenberg A, Salzer P, Hager A (1995) Induction of mycorrhizalike structures and defence reactions in dual cultures of spruce callus and ectomycorrhizal fungi. New Phytol 130: 149±156 Staehelin C, Granado J, MuÈller J, Wiemken A, Mellor RG, Felix G, Regenass M, Broughton WJ, Boller T (1994a) Perception of Rhizobium nodulation factors by tomato cells and inactivation by root chitinases. Proc Natl Acad Sci USA 91: 2196±2200 Staehelin C, Schultze M, Kondorosi E, Mellor RB, Boller T, Kondorosi A (1994b) Structural modi®cations in Rhizobium meliloti Nod factors in¯uence their stability against hydrolysis by root chitinases. Plant J 5: 319±330 Stellmach B, Gottschick W, Battermann F, Zabel K (1988) Bestimmungsmethoden Enzyme fuÈr Pharmazie, Lebensmittel, Technik, Biochemie, Biologie, Medizin. Steinkop€ Verlag, Darmstadt Tavernier E, Wendehenne D, Blein JP, Pugin A (1995) Involvement of free calcium in action of cryptogein, a proteinaceous elicitor of hypersensitive reaction in tobacco cells. Plant Physiol 109: 1025±1031 Trudel J, Asselin A (1989) Detection of chitinase activity after polyacrylamide gel electrophoresis. Anal Biochem 178: 362±366 Vad K, Mikkelsen JD, Collinge DB (1991) Induction, puri®cation and characterization of chitinase isolated from pea leaves inoculated with Ascochyta pisi. Planta 184: 24±29 Viard MP, Martin F, Pugin A, Ricci P, Blein JP (1994) Protein phosphorylation is induced in tobacco cells by the elicitor cryptogein. Plant Physiol 104: 1245±1249 Volpin H, Phillips DA, Okon Y, Kapulnik Y (1995) Suppression of an iso¯avanoid phytoalexin defense response in mycorrhizal alfalfa roots. Plant Physiol 108: 1449±1454 WaldmuÈller T, Cosio EG, Griesebach H, Ebel J (1992) Release of highly elicitor-active glucans by germinating zoospores of Phytophthora megasperma f. sp. glycinea. Planta 188: 498±505 Zanetti A, Beauvais F, Huet JC, Pernollet JC (1992) Movement of elicitins, necrosis-inducing proteins secreted by Phytophthora sp., in tobacco. Planta 187: 163±170