INTRODUCTION

To defend themselves against microbial attack, plants induce upon perception of microbial signal molecules the production of various antimicrobial substances such as hydrogen peroxide, hydrolytic enzymes, and antimicrobial peptides and proteins. Alternatively, some of these defense molecules are constitutively produced in certain specialized cell types or tissues. One particular family of antimicrobial peptides, occurring in different plant species, is the plant defensins. The plant defensins are small (45-54 amino acids), basic, and have a complex three-dimensional folding pattern, stabilized by eight disulfide-linked cysteines (reviewed in Refs. 1 and 2). They are structurally related to insect and mammalian defensins (1). Plant defensins can inhibit the growth of a broad range of fungi at micromolar concentrations but are nontoxic to either mammalian or plant cells (3-6). According to the morphogenic effects caused on treated fungal hyphae, plant defensins can be divided into two different subgroups namely "morphogenic" plant defensins, including Hs-AFP1¹ from Heuchera sanguinea and Rs-AFP2 from Raphanus sativus, which cause reduced hyphal elongation with a concomitant increase in hyphal branching, and "nonmorphogenic" plant defensins, including Dm-AMP1 from Dahlia merckii, Ah-AMP1 from Aesculus hippocastanum, and Ct-AMP1 from Clitoria ternatea, which slow down hyphal extension but do not induce marked morphological distortions (6).

Treatment of hyphae of Neurospora crassa with either of the antifungal plant defensins Rs-AFP2 or Dm-AMP1 induces a series of rapid membrane responses, including K⁺ efflux, Ca²⁺ uptake, alkalinization of the incubation medium, and membrane potential changes (7). In contrast with insect (8) and mammalian defensins (9), plant defensins do not form ion- or dye-permeable pores in artificial membranes (7, 10), nor do they exhibit substantial hyphal membrane permeabilization activity (7). Therefore, it seems likely that these membrane responses are initiated through interaction with a receptor that may either transduce a signal to endogenous ion channels in the membrane or, alternatively, facilitate insertion of the plant defensin into the membrane with subsequent ion channel formation (7). Mutational analysis of the plant defensin Rs-AFP2 from radish revealed two adjacent subsites that appear to be important for antifungal activity and possibly for interaction with a putative receptor (11). The identification of this receptor can be approached by a search for binding sites in the fungal membrane with a high specificity and affinity for plant defensins.

In the present study we have used a ³⁵S-labeled derivative of Hs-AFP1, an antifungal peptide belonging to the morphogenic plant defensins, to demonstrate the existence of specific, high affinity binding sites for plant defensins on N. crassa hyphae and microsomal membranes. Our results point to the presence of specific binding sites for plant defensins that have properties expected for plant defensin receptors. To our knowledge, this is the first example of an antimicrobial protein from plants shown to act through interaction with a receptor.

EXPERIMENTAL PROCEDURES

The antifungal peptides Hs-AFP1, Dm-AMP1, Ah-AMP1, and Ct-AMP1 were isolated as described previously by Osborn et al. (6). Rs-AFP2, Ace-AMP1, Ac-AMP1, Mj-AMP2, and Ib-AMP4 were isolated as described previously by Terras et al. (3), Cammue et al. (12), Broekaert et al. (13), Cammue et al. (14), and Tailor et al. (15), respectively. Rs-AFP2(Y38G) was purified from the culture medium of recombinant yeast (Saccharomyces cerevisiae) as described previously by De Samblanx et al. (11). t-Butoxycarbonyl-L-methionine N-hydroxysuccinimidyl ester (Boc-Met-NHS) was obtained from Bachem (Bubendorf, Switzerland), and Boc-[³⁵S]Met-NHS was from Amersham Int. (Slough, UK). Proteinase K and carbonylcyanide m-chlorophenylhydrazone (CCCP) were purchased from Sigma. All other reagents were of reagent grade and were obtained from commercial sources.

Antifungal activities of protein samples were assayed by microspectrophotometry as described previously (3, 16). Briefly, in a well of a 96-well microplate, 10 µl of the protein sample was mixed with 90 µl of potato dextrose broth (12 g/liter; Difco) containing fungal spores that had been pregerminated for 16 h. Growth was recorded after 48 h of further incubation at 22 °C. The absorption at 595 nm served as a measure for microbial growth (3). IC₅₀ values (i.e. the concentration of the antifungal protein that is required to inhibit 50% of the fungal growth) were calculated from dose-response curves with 2-fold dilution steps (3). N. crassa strain MUCL 19026, Botrytis cinerea strain MUCL 30158, Fusarium culmorum strain IMI 180420, and Alternaria brassicicola strain MUCL 20297 were used as test fungi.

Microsomal membranes were isolated from N. crassa essentially as described (17). Briefly, N. crassa mycelium was digested with -glucuronidase (type H1, Sigma) and subsequently homogenized in a Potter-Elvehjem-type glass tissue grinder with a Teflon pestle (Arthur H. Thomas Co., Philadelphia). The homogenate was subjected to a series of subsequent centrifugation steps as follows: 1000 × g for 10 min (with recovery of the supernatant); 15,000 × g for 30 min (with recovery of the supernatant); 12,000 × g for 30 min (with recovery of the supernatant) and 40,000 × g for 40 min (with recovery of the pellet). The final pellet was suspended in 1 mM Tris-EGTA (pH 7.5), flash-frozen in liquid nitrogen, and stored at 80 °C. This fraction contains microsomes that are mainly derived from the plasma membrane and the endoplasmic reticulum. The specific ATPase activity of this faction was approximately 40 µmol of phosphate released in 1 h at 30 °C per mg of protein in the absence of Na₃VO₄ (a plasma membrane ATPase inhibitor), and the ATPase activity was inhibited for 80% in the presence of 0.02 mM Na₃VO₄. ATPase activity was measured as described previously (17). Total protein in the membranes was determined after addition of SDS to a final concentration of 0.5% (w/v), using the bicinchoninic acid assay (18) with bovine serum albumin as standard.

Plant defensins were radiolabeled using Boc-[³⁵S]Met-NHS at 30 Tbq/mmol (Amersham Int.). For each labeling reaction 9.5 Mbq was aliquoted into small glass vials and the solvent removed by passing a stream of nitrogen over the surface. Ten µl of peptide solution (2.5 mg/ml in 0.1 M Tris borate buffer (pH 8.5)) was added to the glass vial on ice, and the reaction was allowed to proceed for 20 min. One ml of 20 mM ammonium acetate (pH 6) was added to the reaction mixture, which was subsequently applied to a 1 ml of HiTrap S-Sepharose column (Pharmacia, Uppsala, Sweden) previously equilibrated in 20 mM ammonium acetate (pH 6). The column was washed with 20 mM ammonium acetate (pH 6) until all the unincorporated label was removed from the column (about 5 ml of buffer). The labeled peptide was then eluted from the column with 500 mM ammonium acetate (pH 6), and 0.5-ml fractions were collected. Five µl of each fraction was added to liquid scintillant and counted in a liquid scintillator counter (Wallac 1410, Pharmacia, Uppsala, Sweden), and the fractions with the highest counts were pooled. Labeled peptides were stored frozen at 20 °C. In parallel with the radiolabeling reaction, equal aliquots of the peptide solutions were allowed to react with unlabeled Boc-Met-NHS (Bachem) and subjected to chromatography on a HiTrap S-Sepharose column. After the chromatography step, eluate fractions were freeze-dried and redissolved in distilled water, and the protein concentration was measured using the bicinchoninic acid method (18) with the unlabeled plant defensins as standard. This allowed us to estimate the protein concentration in the final labeled peptide solutions and hence to calculate their specific radioactivity. Specific activities of the labeled peptides were typically around 1 Tbq/mmol. To produce cold-labeled peptides for determination of the antifungal activity, 200 µl of peptide solution (2.5 mg/ml in 0.1 M Tris borate buffer (pH 8.5)) was added to 100 µg of Boc-Met-NHS on ice and incubated for 20 min. Trifluoroacetic acid was applied to the reaction mixture at a final concentration of 0.1%, and the mixture was separated by high pressure liquid chromatography on a reversed-phase column (C₂/C₁₈ silica, 25 × 0.93 cm, Pharmacia) using a linear gradient from 0 to 50% acetonitrile in 0.1% trifluoroacetic acid. Peaks containing the unmodified peptide were identified based on comparison of their retention time with those of the purified native peptides separated under identical conditions. Peak fractions containing labeled peptides were collected, dried in a rotating vacuum concentrator to remove the solvents, and redissolved in distilled water. The specific antifungal activity of labeled peptides was compared with that of native peptides in the antifungal activity assay described above.

For binding studies with N. crassa hyphae, N. crassa was grown at an inoculum density of 3 × 10⁵ spores/ml as described previously (7). The medium consisted of potato dextrose broth (12 g/liter, Difco). After 20 h of incubation at 22 °C, 80-µl aliquots of the N. crassa culture were incubated, unless stated otherwise, with 3 pmol of [³⁵S]Hs-AFP1 (~1.2 TBq/mmol) either with or without appropriate concentrations of competing antifungal peptides in a total volume of 100 µl in siliconized polypropylene microcentrifuge tubes (Sigma) at 22 °C. Total protein in the N. crassa cultures was measured after homogenization by high speed reciprocal shaking in the presence of 10% (w/v) 400-600-µm glass beads (Sigma) in a Fastprep apparatus (Bio 101/Savant, Farmingdale, NY). Protein measurement was done by the bicinchoninic acid method (18) using bovine serum albumin as standard. Total protein content in N. crassa cultures was approximately 1.8 mg of protein/ml. For binding studies with microsomal membranes, 2-µl aliquots of a membrane suspension containing 1.6 mg of protein/ml were incubated with 3 pmol of [³⁵S]Hs-AFP1 in a total volume of 100 µl of binding buffer (20 mM MES, 50 mM KCl, 1 mM MgCl₂ (pH 6.2)) in siliconized polypropylene microcentrifuge tubes at 4 °C. After incubation for the times indicated, the samples were transferred to wells of a MultiScreen Durapore 96-well filtration plate (0.65-µm pore size, Millipore, Bedford, MA). After filtration using a MultiScreen vacuum filtration manifold (Millipore), harvested hyphae or microsomal membranes were washed three times with 100 µl of ice-cold binding buffer. Filters with the hyphae or microsomal membranes were punched out manually with MultiScreen punch tips (Millipore), suspended in scintillation mixture (Ready Safe scintillation mixture, Analis, Gent, Belgium), and counted for ³⁵S in a Liquid Scintillation Counter. Nonspecific binding was determined as the amount of radioactivity bound in the presence of 10 µM unlabeled Hs-AFP1. Specific binding was calculated by subtracting nonspecific binding from total binding.

Reduction of Hs-AFP1 was performed by the addition of 100 mM Tris-HCl (pH 8) and 100 mM dithioerythritol, followed by incubation at 45 °C for 1 h. Reagents were removed on a disposable reversed phase C₈ silica column (Bond Elut, 500 mg solid phase, Varian). The C₈ silica column was subsequently rinsed with 6 ml of distilled water containing 0.1% (v/v) trifluoroacetic acid. The reduced peptide was eluted from the latter column with 5 ml of 50% (v/v) acetonitrile containing 0.1% (v/v) trifluoroacetic acid. Protein fractions (monitored by measurement of the absorbance at 280 nm) were vacuum dried and redissolved in distilled water.

RESULTS

As in previous studies on binding sites for glycopeptides (19) and chitin fragments (20), we have chosen to use sulfur-35 in the form of Boc-[³⁵S]Met-NHS as a radioactive isotope to label two different morphogenic plant defensins, namely Rs-AFP2 and Hs-AFP1 from R. sativus and H. sanguinea seed, respectively. To test the effect of the labeling reaction on the antifungal activity of these plant defensins, nonradioactive Boc-Met-NHS was linked to the peptides, and the specific antifungal activity of the modified peptides was assayed in an antifungal activity assay. The antifungal activity against the fungus N. crassa of modified Rs-AFP2 was reduced by more than 4-fold compared with that of unmodified native Rs-AFP2, whereas modified and unmodified Hs-AFP1 had essentially the same activity, irrespective of whether the activity was tested in a potato dextrose growth medium with or without addition of 1 mM MgCl₂ and 50 mM KCl (Table I). In addition, modified and unmodified Hs-AFP1 had essentially the same antifungal potency against three other fungal species with different sensitivities to Hs-AFP1. Therefore, Hs-AFP1 was chosen for labeling with Boc-[³⁵S]Met-NHS, and this radiolabeled derivative, hereafter called [³⁵S]Hs-AFP1, was used in subsequent binding studies.

Table I. Antifungal activity of Hs-AFP1 and Hs-AFP1 modified with Boc-Met-NHS (Boc-Met-IIsAFP1) on pregerminated fungal spores Fungus Growth mediuma IC50 (µg/ml)b Hs-AFP1 Boc-Met-HsAFP1 N. crassa A 4 5 N. crassa B 8 8 A. brassisicola A 7 7 A. brassisicola B 18 20 B. cinerea A 75 65 B. cinerea B >100 >100 F. culmorum A 12 10 F. culmorum B 25 28 a Growth medium A is potato dextrose broth (12 g/liter, Difco) and growth medium B is potato dextrose broth (12 g/liter) including 1 mM MgCl2 and 50 mM KCl. b Concentration required for 50% inhibition of fungal growth as defined under "Experimental Procedures."

[³⁵S]Hs-AFP1 was found to bind to both living N. crassa hyphae and microsomal membrane fractions derived from this fungus. To investigate whether this binding activity was specific, [³⁵S]Hs-AFP1 was added to either N. crassa hyphae or microsomal membranes together with increasing concentrations of native unlabeled Hs-AFP1. Binding of [³⁵S]Hs-AFP1 to either N. crassa hyphae or microsomal membrane preparations could be competed by about 95 and 98% in the presence of 1 and 10 µM of excess unlabeled peptide, respectively (Fig. 1).

[View Larger Version of this Image (19K GIF file)]

To assess further the specificity of binding of ³⁵S-labeled Hs-AFP1, the ability of different antifungal peptides to compete with [³⁵S]Hs-AFP1 was tested, both on N. crassa hyphae and microsomal membranes. The antifungal peptides used in this experiment were Ah-AMP1, Dm-AMP1, and Ct-AMP1, isolated from seed of horse chestnut (A. hippocastanum), dahlia (D. merckii), and C. ternatea, respectively, which belong to the nonmorphogenic plant defensins (6); Ace-AMP1, isolated from onion (Allium cepa) seed and structurally related to lipid transfer proteins (12); Ac-AMP1, isolated from amaranth (Amaranthus caudatus) seed and belonging to the hevein-type antimicrobial peptides (2, 13); Mj-AMP1, isolated from Mirabilis jalapa seed and belonging to the knottin-type antimicrobial peptides (2, 14); Ib-AMP4, a 20-amino acid antimicrobial peptide with two disulfide bridges isolated from Impatiens balsamina seed (15); Rs-AFP2, isolated from radish (R. sativus) seed and, like Hs-AFP1, a member of the morphogenic plant defensins (3, 6); and finally Rs-AFP2(Y₃₈G), an Rs-AFP2 analog carrying a single amino acid substitution from tyrosine to glycine at position 38 (11). As a result of this amino acid substitution, Rs-AFP2(Y38G) is only weakly active against N. crassa in a low ionic strength medium and is totally inactive in the presence of 1 mM MgCl₂ or 1 mM CaCl₂, whereas Rs-AFP2 is equally active in the absence or the presence of 1 mM divalent cations (11). Binding of [³⁵S]Hs-AFP1 to hyphae and microsomal membranes (Fig. 2, A and B) could be competed to some extent by the plant defensins Rs-AFP2, Ah-AMP1, Dm-AMP1, and Ct-AMP1, but in all these cases, competition was weaker compared with competition by Hs-AFP1 itself. Competition of [³⁵S]Hs-AFP1 binding by all plant defensins, except Hs-AFP1 itself, was stronger with hyphae than with microsomal membranes. Interestingly, no significant competition with [³⁵S]Hs-AFP1 binding was observed for Rs-AFP2(Y38G), the inactive single amino acid variant of the plant defensin Rs-AFP2. In addition, none of four different cysteine-rich antifungal peptides (Mj-AMP1, Ac-AMP1, Ace-AMP1, and Ib-AMP4), each belonging to a class that is structurally unrelated to plant defensins (2), were able to displace [³⁵S]Hs-AFP1 from its binding site. It was also observed that binding of [³⁵S]Hs-AFP1 to microsomal membranes, but not to hyphae, was stimulated by Mj-AMP1 and Ib-AMP4 at concentrations of 10 µM or above (Fig. 2B).

[View Larger Version of this Image (26K GIF file)]

It has previously been shown that plant defensins lose their antifungal activity upon reduction of their disulfide bridges (3) as well as upon addition of divalent cations to the growth medium (4, 6, 7). Reduced Hs-AFP1 at 40 µM was unable to compete with binding of [³⁵S]Hs-AFP1 to N. crassa hyphae. Addition of 20 mM MgCl₂ to the binding assay reduced binding of [³⁵S]Hs-AFP1 to N. crassa hyphae by over 95% (results not shown).

Association kinetics of binding of Hs-AFP1 to N. crassa hyphae and microsomal membranes was determined by incubation of [³⁵S]Hs-AFP1 for increasing periods. Dissociation was initiated after 30 min of incubation by addition of an excess of unlabeled Hs-AFP1. Incubation of [³⁵S]Hs-AFP1 with N. crassa hyphae and microsomal membranes resulted in rapid binding of the radioligand, and half-maximal binding occurred within 1 min (Fig. 3, A and B). Addition of 10 µM unlabeled Hs-AFP1 30 min after addition of the radioligand resulted in rapid (but partial) dissociation of radioligand binding, indicating reversibility of [³⁵S]Hs-AFP1 binding. Nonspecific binding, determined in presence of 10 µM unlabeled Hs-AFP1, remained at about 10% of total binding throughout the experiment. Rate constants for association and dissociation of [³⁵S]Hs-AFP1 to N. crassa hyphae on the one hand, and microsomal membranes on the other hand, were calculated to be 2.7 10⁸ M¹ min¹ and 8.5 min¹, respectively, and 2.8 10⁸ M¹ min¹ and 8.5 min¹, respectively. The K_d values given by the rate constant ratios were 31 nM for binding of [³⁵S]Hs-AFP1 to N. crassa hyphae and 30 nM for binding of [³⁵S]Hs-AFP1 to microsomal membranes.

[View Larger Version of this Image (21K GIF file)]

Saturability of the binding of Hs-AFP1 to N. crassa hyphae and microsomal membranes was examined by incubation with increasing concentrations of [³⁵S]Hs-AFP1 (2-80 nM). The saturation curve of [³⁵S]Hs-AFP1 binding to N. crassa hyphae was hyperbolic (Fig. 4A); half-maximal binding of [³⁵S]Hs-AFP1 occurred at 24 nM. Scatchard analysis of the binding data in the 10-100% saturation range revealed a K_d of 29 and a B_max of 1.4 pmol/mg protein (Fig. 4B). In the case of binding of [³⁵S]Hs-AFP1 to N. crassa microsomal membranes, half-maximal binding occurred at 22 nM (Fig. 4C). A K_d of 26 nM and a B_max of 102 pmol/mg protein was calculated from a Scatchard plot drawn with the data in the range of 10-100% saturation (Fig. 4D). The Hill coefficients determined from the specific binding data gave a value of 1.046 for binding of [³⁵S]Hs-AFP1 to N. crassa hyphae and a value of 0.996 for binding of [³⁵S]Hs-AFP1 to microsomal membranes, indicating a single class of binding sites for Hs-AFP1 on N. crassa hyphae and microsomal membranes that does not show cooperativity (results not shown). The K_d values calculated from saturation data (29 nM for binding of [³⁵S]Hs-AFP1 to N. crassa hyphae and 26 nM for binding of [³⁵S]Hs-AFP1 to microsomal membranes) are close to the K_d values calculated from the rate constant ratios (31 and 30 nM, respectively).

[View Larger Version of this Image (33K GIF file)]

To address the question whether proteins are involved in the binding of ³⁵S-labeled Hs-AFP1, microsomal membranes were preincubated with proteinase K at 50 µg/ml. After incubation of microsomal membranes with protease, heat-inactivated protease, or water for 1.5 h at 37 °C, binding of [³⁵S]Hs-AFP1 was assayed at 4 °C. Preincubation with native proteinase K reduced binding of [³⁵S]Hs-AFP1 by about 40% (50% reduction versus heat-inactivated control) (Fig. 5). Proteinase K did not affect the antifungal activity of Hs-AFP1 when assessed under conditions identical to those in the binding assays (results not shown).

[View Larger Version of this Image (37K GIF file)]

Mammalian and insect defensins, two families of defensins that are structurally related to plant defensins, have been shown to form voltage-dependent ion channels in membranes (8, 9). Furthermore, membrane-depolarizing agents such as CCCP or 3,3,4,5-tetrachlorosalicylanilide protect cells from the effects caused by either insect or mammalian defensins (9, 20). To investigate a possible antagonism between CCCP and binding of plant defensins to N. crassa cells, 10-50 µM CCCP was added to N. crassa cells prior to addition of radiolabeled Hs-AFP1. After 1.5 h of incubation, binding of [³⁵S]Hs-AFP1 was assayed. Binding of [³⁵S]Hs-AFP1 to N. crassa hyphae was not affected by CCCP (results not shown). CCCP inhibited growth of N. crassa hyphae over the concentration range tested (10-50 µM) (results not shown).

DISCUSSION

Our previous work has shown that plant defensins induce various physiological effects on the fungus N. crassa, including rapid Ca²⁺ uptake, K⁺ efflux, and medium alkalinization, and that such effects are most probably not mediated by direct peptide-phospholipid interactions (7). In the present study we have examined the binding of plant defensins to living N. crassa hyphae under conditions used to detect the above cited physiological effects. Two different plant defensins, namely Hs-AFP1 and Rs-AFP2, were subjected to a labeling procedure consisting of the derivatization of primary amino groups with Boc-Met-NHS. The resulting modified peptides exhibited the same antifungal activity as the native peptides in the case of Hs-AFP1 but not in that of Rs-AFP2. We have not verified the impact of the derivatized side groups on the conformation of Hs-AFP1, but the fact that the labeled peptide retained its full antifungal activity against four fungal species differing in sensitivity to native Hs-AFP1 is considered to be the best possible proof that the modifications had not affected side chains that are important for interaction with the fungal target. We have previously shown for Rs-AFP2 that substitution of particular amino acids by other amino acids can either drastically reduce or have little or no effect on the antifungal activity, depending on the position of that particular amino acid (11).

Binding of a Boc-[³⁵S]Met-NHS derivative of Hs-AFP1 to N. crassa hyphae showed all properties expected for a ligand-receptor interaction such as competitiveness, specificity, reversibility, and saturability. Data analysis (Scatchard and Hill plots) of the ligand saturation experiments with [³⁵S]Hs-AFP1 indicated that N. crassa hyphae possess a single class of high affinity binding sites for Hs-AFP1 with an apparent K_d of 29 nM. Binding of radiolabeled Hs-AFP1 could be competed at relatively low concentrations (0.1 µM) of unlabeled Hs-AFP1, but also, albeit at a somewhat lower efficiency, by the other plant defensins tested, including Rs-AFP2, Dm-AMP1, Ah-AMP1, and Ct-AMP1. These results suggest that all tested plant defensins bind to either the same receptor as Hs-AFP1 but with different affinities or on different subsites of this receptor. The specificity of the interaction was demonstrated by the inability to displace [³⁵S]Hs-AFP1 from its binding site by any of four different cysteine-rich, basic peptides that are structurally unrelated to plant defensins.

Binding of [³⁵S]Hs-AFP1 to a N. crassa microsome preparation containing mainly plasma membranes and endoplasmic reticulum showed the same properties as seen with N. crassa hyphae; binding was specific, competitive, reversible, and saturable with an apparent K_d of 27 nM and a B_max of 102 pmol/mg protein. This K_d value is very similar to that obtained with hyphae, which indicates that the Hs-AFP1 binding sites reside in the plasma membrane. Binding of radiolabeled Hs-AFP1 to microsomal membranes was weakly but consistently decreased upon proteinase K treatment, suggesting either that the receptor may be proteinaceous in nature or that certain proteins may be involved in binding of Hs-AFP1 to its receptor. Assuming a M_r of 100,000 for the putative proteinaceous receptor and given the B_max value of 102 pmol/mg protein, about 1% of all membrane proteins would correspond to the receptor, indicating that it is relatively abundant.

Direct interaction with lipid components of the plasma membrane and subsequent formation of voltage-dependent ion channels have previously been proposed to explain the antimicrobial effects of insect defensins and mammalian defensins (8, 9). Such interaction has previously been shown to be dependent on the membrane potential in the case of insect defensins (8) and mammalian defensins (9, 21). In contrast to insect and mammalian defensins, plant defensins are unable to influence the conduction of ion currents through artificial phospholipid membranes (7) or to cause release of fluorescent dyes entrapped in artificial phospholipid vesicles (10). Based on our binding data we propose that the membrane responses induced by plant defensins, including generation of ion fluxes, require the presence of specific receptors in the phospholipid membrane. Binding of [³⁵S]Hs-AFP1 to N. crassa hyphae was shown not to be affected by CCCP-induced membrane depolarization, whereas CCCP treatment abolished the increased Ca²⁺ influx induced by plant defensins (results not shown) indicating that binding and ion flux generation are separable events. Two different hypotheses can be forwarded to explain how binding of a plant defensin to a membrane receptor might influence ion fluxes through the membrane. A first possibility is that the receptor would act as an anchor point allowing plant defensins to insert in the membrane and to form ion-permeable pores. The observation that binding of [³⁵S]Hs-AFP1 can only be partially reversed by addition of excess unlabeled Hs-AFP1 would be consistent with this model. Examples of pore-forming proteins that are known to insert into phospholipid membranes after interaction with specific membrane receptors are the antibacterial colicins form Escherichia coli (22) and the antifungal killer toxin K1 from S. cerevisiae (23). On the other hand, it is also possible that binding of plant defensins to their receptor would entail activation of endogenous signal transduction components which in turn may affect the activity of endogenous ion channels or ion transporters. One of the ion fluxes generated upon treatment of fungi with plant defensins is the influx of Ca²⁺. As controlled Ca²⁺ influx is known to be essential for guided vesicle transport and sustained growth of tip-growing cells (24) including fungi (25), receptor-mediated Ca²⁺ influx caused by plant defensins is likely to contribute to growth inhibition.

Several lines of evidence suggest that binding of plant defensins to their receptor sites is linked to their antifungal effects. First, reduction of their disulfide bridges abolishes both antifungal activity and binding displacement capacity. Second, the presence of divalent cations reduces both binding and antifungal activity. Third, Rs-AFP2(Y38G), a single amino acid substitution variant of Rs-AFP2 that has severely reduced antifungal properties but a nearly identical circular dichroism spectrum (11), is unable to compete for the Hs-AFP1 binding sites, unlike Rs-AFP2 itself. Interestingly, Rs-AFP2(Y38G) was also found to be unable to trigger Ca²⁺ influx in N. crassa hyphae (11), whereas Rs-AFP2 is a potent mediator of enhanced Ca²⁺ influx (7, 11). Hence, the fact that a single amino acid substitution can affect both binding to a membrane receptor and Ca²⁺ flux across a membrane is consistent with the above proposed model whereby plant defensins interact with receptors in the fungal plasma membrane, thus affecting either directly or indirectly the generation of ion fluxes. However, as the available evidence is indirect, it can not be excluded as yet that the plant defensin receptor identified in this study is not directly involved in the mode of action of these antifungal peptides.