Maize EMBRYO SAC family peptides interact differentially with pollen tubes and fungal cells

Highlight EMBRYO SAC family peptides show multiple activities in mediating pollen tube burst and fungal growth inhibition. They interact differently with cell surface receptors of pollen tube and fungal cells.


Introduction
. Defensins in metazoan species, mainly from the arthropoda, include antibacterial defensins and scorpion α-neurotoxins (Smith et al., 2013), but also α-amylase inhibitors, antifungal proteins, and proteins annotated as proteinase inhibitors (Schultz et al., 1998;Letunic et al., 2015). Plant defensins have previously been shown to be required as part of the innate immune system to protect organs such as roots, flowers, fruits, and seeds from pathogen attack (e.g. Iwai et al., 2002;Dracatos et al., 2014;Ji et al., 2015).
Recent studies have demonstrated that plant defensins and defensin-like proteins (DEFL) also act as key signalling molecules during reproductive processes, especially along the pollen tube journey (Dresselhaus and Franklin-Tong, 2013;Qu et al., 2015). As polymorphic small secreted peptides, they are involved, for example, in signalling during pollen-pistil interactions (SCR/SP11: Schopfer et al., 1999), in pollentube guidance (LUREs: Okuda et al., 2009;Goto et al., 2011;Takeuchi and Higashiyama, 2012), and in sperm cell release (ES4: Amien et al., 2010). The ES gene family was previously shown to be highly expressed in the cells of the egg apparatus of maize. After fertilization, the genes were switched off and expression was not detectable during seed development or in vegetative tissues of maize (Cordts et al., 2001;Amien et al., 2010). ES4 was further studied and shown to mediate pollen tube burst via opening of the potassium channel KZM1 (Amien et al., 2010). In contrast, scorpion toxins related to ES peptides appear to act at multiple ion channels and were shown to block the pore of potassium channels (Quintero-Hernandez et al., 2013).
Considering that the various subclasses of defensins mainly act either as antimicrobial peptides or as toxins against higher organisms, the aim of this study was to elucidate the various functional activities and specificity of the whole ES1-4 peptide family, both in reproductive processes and defence reactions. Thus, the activity of various ES peptides and their mutant versions on pollen tube burst was investigated. The biotrophic fungus Ustilago maydis and the necrotrophic fungus Fusarium graminearum served as microbial targets to quantitatively study the effect of these peptides on fungal growth behaviour and response(s). Investigations were restricted to fungi because it is known that ES4 does not inhibit growth of bacterial pathogens (Amien et al., 2010). Moreover, important peptide fragments and individual residues within ES1-4 involved in the interactions with their pollen tube and fungal receptors were identified and characterized.

Materials and methods
Growth, culture, and harvest of plant and fungal material Maize plants of inbred line A188 were grown under standard greenhouse conditions at 26°C with 16 h light (22°C in the dark) and a relative air humidity of about 60%. Pollen was collected at anthesis (stage R1; stages after Abendroth et al., 2011) on paper sheets after tassels had been shaken softly the evening before to remove old pollen grains. Conidiation of F. graminearum strain ph1 was induced on synthetic nutrient-poor medium on plates incubated at 18°C under near-UV white light condition after Woriedh et al. (2011). Sporulation of U. maydis strain sg200 was induced in liquid complete medium and spores were incubated on a rotor at 28°C in the dark. F. graminearum conidia and U. maydis spores were collected and washed with sterile deionized water. Aliquots with 10% glycerol were kept at −80°C at OD 600 of 1 for further studies. F. graminearum and U. maydis were cultured in pollen germination medium [PGM: 0.0005% H 3 BO 3 , 10 mM CaCl 2 , 0.05 mM KH 2 PO 4 , 10% sucrose, and 6% PEG 4000 (Schreiber et al., 2004)].

Pollen tube burst assay
The assay was applied as described (Woriedh et al., 2013). Briefly, freshly collected pollen was immediately distributed onto Petri dishes containing droplets of 10 µL PGM at different pH values. Pollen tubes started to burst or to grow after about 10 min. Pollen tubes were subjected to further studies only if germination rates exceeded 80%. After optimization, PGM at pH 5 was applied to all pollen tube assays. A 10 µL solution of the above peptides or acetonitrile as a negative control was added to each droplet of germinated pollen using a Cell-TramH Air pump (Eppendorf) and a glass micropipette. During this assay, droplets were observed using an Eclipse TE2000 inverted microscope (Nikon) equipped with a 1.4 Megapixel digital AxioCam MRm camera (black and white) and AxioVision digital image processing software (Release 4.8).

Fungal germination inhibitory assay
Examination of the inhibitory effect of peptides on germination of F. graminearum conidia and U. maydis spores was performed in 96-well microtitre plates. A 3 µL solution of peptides or control PGM was added to 96 µL PGM supplemented with 1 µL (OD 600 = 1) of F. graminearum conidia or U. maydis spores each expressing 3xGFP marker protein. Microtitre plates were incubated at 28°C at 150 rpm in the dark and measured using spectrophotometry by relative absorbance at 595 nm using a Multiskan GO spectrophotometer (Thermo Scientific). Inhibition studies of F. graminearum-3xGFP and U. maydis-3xGFP in culture were controlled microscopically using an Eclipse TE2000 inverted microscope (Nikon) as described above. GFP was excited with UV light using a HC Alexa488/EGFP filter (F36-525, EX: 472/30, DM: 495, BA: 520/35 BP).

Peptide binding assays
For maize pollen tubes, maize pollen was germinated on a cellulose acetate membrane (Sigma: dialyser capacity 10-200 μL, MWCO 500 Da, diameter 11 mm) and flooded with a droplet of 10 µL PGM. Cellulose acetate membranes were washed three times with PGM before use. Around 10 min after germination of pollen tubes, 10 µL of TAMRA-ES-a, Dabcyl-ES-c, TAMRA-ES-d, or Rhodamine-ES4 was applied to the cellulose membrane, gently washed with PGM, and immediately visualized under bright field and fluorescence using an Aoptome.2 microscope (Zeiss) equipped with an Axio camera 503 mono and an Axio camera 105 colour. The ZEN 2012 SP2 digital image software (Blue edition) was used to capture and process images. TAMRA and rhodamine were excited with UV light using filter set 43 HE, 489047-9901 (EX: BP 550/25, Beamsplitter: FT 570, EM: BP 605/70). Dabcyl was excited with UV light using filter set 49, Beamsplitter: FT 395,EM: BP 445/50).
For F. graminearum conidia or U. maydis spores, a 6 µL solution of TAMRA-ES-a, Dabcyl-ES-c, TAMRA-ES-d, or Rhodamine-ES4 was added to 192 µL PGM supplemented with 2 µL (OD 600 = 1) of conidia or spores at 28°C in the dark. Cells were germinated for 3, 6, 12, and 24 h, and then washed with PGM before visualization on a cellulose acetate membrane using an Aoptome.2 microscope (Zeiss) as described above. Cellulose acetate membranes were washed three times with PGM before usage.
For F. graminearum wild-type mycelia, 2 µL (OD 600 = 1) of conidia were germinated in 192 µL PGM at 28°C in the dark for 24 h, and then incubated with 6 µL of labelled peptides for 3, 6, 12, and 24 h before visualization as described above.
Histochemical detection of reactive oxygen species Production of reactive oxygen species (ROS) was detected by staining of the superoxide (O 2 -) anion with nitrotetrazolium blue chloride (NBT; Sigma) according to Semighini and Harris (2008); 2.5 mM NBT (Sigma) was dissolved in 5 mM Mops. To stain wild-type F. graminearum conidia or U. maydis spores, a 6 µL solution of unlabelled peptides was added to 192 µL PGM supplemented with 2 µL (OD 600 = 1) of conidia or spores at 28°C in the dark, germinated for 24 h, and then incubated again under the same conditions with 200 µL NBT solution for 2 h. Samples were washed with PGM before visualization on a cellulose membrane under bright field using an Aoptome.2 microscope (Zeiss) as described above.
To stain F. graminearum wild-type mycelia, 2 µL (OD 600 = 1) of conidia were germinated in 192 µL PGM at 28°C in the dark for 24 h, then incubated with 6 µL of peptides for 3, 6, 12, and 24 h and finally incubated with 200 µL NBT solution for 2 h. After washing with PGM, mycelia were visualized as described above.

Statistical analysis
An independent-samples t-test and an ANOVA were used to examine the effect of peptides on pollen tubes and fungal cells (where data were normally distributed) and an independent-samples t-test was used for the effect of the peptides (where data were not normally distributed). The data are presented as the corrected mean ± SD (P < 0.001).

Structure of maize ES family peptides
The sequences of ES1-4 peptides are highly similar to each other (Cordts et al., 2001). ES4 shows a sequence identity of about 97% with ES2 and ES3, and of about 91% with ES1 ( Fig. 1A, B). Three C-terminal mismatches distinguish mature peptides of ES1 and ES4 and another mismatch occurs in the hypervariable region ES-d (see below). Thus, predicted mature ES1 and ES4 peptides consisting of 61 amino acids were selected for functional studies. Mature ES family peptides were first divided into five highly conserved fragments named ES-a, ES-b, ES-c, ES-d, and ES-e (13-16 amino acids in length; Fig. 1A, B). These peptides and variants of ES-d were used for interaction and mapping studies. Moreover, considering that ES1-4 genes encode peptides with structural homology to defensins and DEFLs, PyMOL (version 1.7.4; The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC) was used to generate homology models based on known 3D structures deduced from the protein data bank (Bernstein et al., 1977). The 3D structure is characterized by a welldefined 3-stranded anti-parallel beta-sheet and a short alpha helix. Four disulfide bridges are located in the hydrophobic core between helix and sheet, forming a cysteine-stabilized alpha-helical motif (Fig. 1C). This structure is homologous to plant defensins and DEFLs, and analogous to scorpion toxins and insect defensins. However, an extended loop occurs in the ES-d peptide between Cys 5 and Cys 6 . Therefore, the two variants mature ES-d1 (mES-d1) and mES-d2, as well as an altered mES4 peptide, were generated (Fig. 1A). For ES4, the electrostatic surface was calculated by means of PyMOL in order to visualize the distribution of charges and of hydrophobic and polar surfaces in an amino-acid-specific manner in 3D (Fig. 1D). The model shows that ES1-4 are hydrophobic peptides, contain a flexible C-terminus, and the domain containing ES-d is exposed from the surface of the molecule.

Mapping of ES domains involved in pollen tube burst in maize
Previous studies have shown that the female gametophyte of maize likely secretes large amounts of ES1-4 required to induce pollen tube burst prior to fertilization (Cordts et al., 2001;Amien et al., 2010). ES1, ES4, mES4 (each 61 amino acid long); smaller peptide fragments ES-a, ES-b, ES-c, ES-d, and ES-e; and mutated versions mES-d1 and mES-d2 (13-16 amino acid long) were applied to growing pollen tubes in order to study their effects on pollen tube growth behaviour. First, the pH optimum of maize PGM was determined. Fig. 2A shows the germination and growth of maize pollen tubes for 1 h at different pH values (pH 3.5 to pH 8). The maize pollen tubes germinated well, with between 92% and 99% germinating with no burst between pH 5 and 5.5. As a result, PGM at pH 5 was used for all further maize pollen tube assays. Different concentrations (100 nM, 250 nM, and 500 nM) of peptides were applied to germinated maize pollen and showed that ES1 and ES4 induces maize pollen tube burst of 60.8 ± 3.3% at 500 nM (P < 0.001). Furthermore, a small peptide of 15 amino acids (ES-d) was even more active than full-length ES peptides and showed induction of pollen tube burst of 82.5 ± 5.2% at 500 nM. Peptides covering other ES domains showed no burst at all or a maximum burst efficiency of 39.3 ± 2.4%. Mutants of ES-d (mES-d1 and mES-d2) and mutants of ES4 (mES4) showed significant reduction of pollen tube burst to about 29.2 ± 7.0 to 41.2 ± 4.4% (P < 0.001) (Fig. 2B). Moreover, after application of ES-d, maize pollen tubes started to burst within a few seconds, which is significantly faster than the response to mature peptides and other small peptides, in which the tubes started to burst after 1-2 min. A series of concentrations was applied to obtain the optimal concentrations for the burst assay. Additionally, the time until burst was correlated with the concentration of the applied peptides. As a result, quantification of pollen tube burst was measured in the following experiments after 5 min.
To identify amino acid residues involved in maize pollen tube burst and target interaction, ES-d (15 amino acids) and 15 mutated versions of ES-d were applied to germinated maize pollen at the above described concentrations. In each of these mutants, one amino acid was replaced by alanine (Ala). Analogously, ES-c (15 amino acids), which showed 29.3 ± 4.2% pollen tube burst, was applied as well as 15 mutants in which one amino acid was replaced by Ala, while Ala in ES-c was replaced with valine (Val). The results showed that a mutation of Leu 3 , Ile 4 , or Tyr 15 in ES-d suppressed maize pollen tube burst to <8 ± 1.5% (P < 0.001) at a concentration of 500 nM, suggesting that these amino acids are essential for target interaction. In contrast, a mutation of Gly 1 , Arg 3 , Ala 4 , or Glu 5 in ES-c strongly increased induction of maize pollen tube burst to ~45 ± 3.0 to 65 ± 5.7% (Fig. 2C, D), indicating that ES peptides can be further modified to increase their activity/stability.

ES-d binds at the surface of maize pollen tubes
It is difficult to investigate the binding of ES peptides at maize pollen tubes because application of ES-d (similar to ES1 and ES4) causes swelling of tube tips and growth arrest prior to explosive burst at the tube apex within a few seconds (Fig. 3A, B). To visualize the interaction and binding sites of ES-d on maize pollen tubes and to prolong swelling time to 2 min prior to burst, a low concentration of 50 nM of ES-d labelled with the fluorescent dye TAMRA was applied to germinated pollen tubes. The tubes showed fluorescence on the entire surface, which increased at the tips ( Fig. 3C-J), causing swelling prior to explosive burst at the apex (Fig. 3E, F). Fluorescence was not detected on pollen grains. The same binding behaviour was observed after application of ES4 labelled with the fluorescent dye rhodamine. Here, burst occurred at the same concentration after 5 min, indicating that longer peptides need more time to interact with the pollen tube surface (Fig. 3K-N). The same binding assay was repeated using ES-a and ES-c at the same concentration labelled with the fluorescent dyes TAMRA (ES-a) or Dabcyl (ES-c). For both assays, fluorescence was not observed in pollen grains and tubes ( Supplementary Fig. S1).

ES-c and ES-d inhibit fungal germination and growth and interact differently with targets compared with pollen tubes
Considering that maize ES1-4 are cysteine-rich peptide precursors belonging to the plant defensin/DEFL group, a fungal germination inhibitory assay was applied to study their antifungal activity and defence functions. Two maize fungal pathogens of different infection types were tested: the necrotrophic fungus F. graminearum and the biotrophic fungus U. maydis (CIMMYT, 2004;Dean et al., 2012). Thus, Fusarium conidia and Ustilago spores were inoculated in vitro with increasing concentrations of ES peptides. To quantify fungal biomass production and growth, F. graminearum-3xGFP and U. maydis-3xGFP strains that constitutively express triple versions of the gene for GFP were used. Fusarium conidia and Ustilago spores were inoculated in PGM, which was supplemented with varying concentrations of peptides. Peptide concentrations of 0.1-0.5 μM used for pollen tube burst assays did not show effects on fungal growth behaviour. However, strongly increased concentrations of 10-90 μM demonstrated a dose-dependent inhibition of germination after 24 h. Remarkably, 90 μM of ES1 inhibited 66.7 ± 1.1%, and ES4 inhibited 67.8 ± 1.5% of the germination of Fusarium conidia, while ES-c inhibited 77.1 ± 1.1% and ES-d inhibited 79.3 ± 1.6% of conidia (P < 0.001). In Ustilago spores, 90 μM of ES1 inhibited 55.9 ± 0.9% of germination and ES4 inhibited 56.4 ± 0.9%, while ES-c inhibited 80.9 ± 1.0% and ES-d inhibited 78.7 ± 1.0% (P < 0.001). A significant inhibition was not observed for other small peptides and their variants (Fig. 4A, B). Conidia and spores showed a severely malformed appearance at a concentration of 90 µM of ES-c or ES-d. Approximately 80% of conidia and spores did not germinate at all, while the rest stopped growing after forming a short germination mycelium. Complete suppression of conidia and spore germination was observed at 120 μM. A peptide concentration of 10 µM resulted only in minor fungal growth differences. In contrast, at concentrations of 30-60 µM, a reduction in germination and fungal growth occurred, including severe swelling and malformed appearance ( Fig. 4C-F). Moreover, Fusarium mycelium developed from germinated conidia failed to form hyphal colonies. Instead, they showed swelling and ballooning at mycelia tips, in comparison to untreated wild-type conidia, which showed development of normal long and thin filaments (Fig. 4C, D). Swelling and ballooning was not observed after peptides were applied to Ustilago spores, but germination of mycelium was strongly inhibited (Fig. 4E, F  To determine the most important amino acid residues involved in inhibiting fungal germination, ES-c and ES-d as well as 15 mutant versions of ES-c or ES-d (see above) were applied to Fusarium conidia and Ustilago spores as described. Mutations of Gly 8 , Tyr 9 , or Thr 10 in ES-c ( Supplementary  Fig. S2A) and mutations of Lys 13 , Cys 14 , or Tyr 15 in ES-d ( Supplementary Fig. S2B) significantly increased fungal germination to 66.7 ± 3.3% to 78 ± 2.2% in Fusarium. A similar result was obtained in Ustilago. Mutations of the same amino acids increased germination of spores to 67.2 ± 1.3% to 72.6 ± 1.4% (P < 0.001) compared with the controls at a concentration of 90 µM (ES-c and ES-d, respectively; Supplementary Fig. S3A, B). Germination above 60% was considered as a threshold for both fungi. ES-c and ES-d showed fungal germination of 19.1 ± 3.8% to 20.6 ± 2.1%, respectively, at the same conditions in both fungi (Supplementary Figs. S2 and S3). In conclusion, ES peptides showed similar effects on fungal cells compared with maize pollen tubes. However, significantly higher concentrations (about 100×) had to be applied to observe comparable effects. Moreover, in contrast to pollen tubes, both ES-c and ES-d showed activity on fungal cells.

ES-c and ES-d bind differently to fungal cells and induce ROS production
To detect and visualize the effect and interaction of ES-c and ES-d on germinated Fusarium conidia and developed mycelium, conidia of F. graminearum were inoculated in PGM following the addition of final concentrations of 30-90 μM of labelled peptides Dabcyl-ES-c, TAMRA-ES-d, and Rhodamine-ES4, as well as TAMRA-ES-a as a control at the four different stages according to Woriedh et al. (2011): (i) fresh conidia (0 hours post induction, hpi), (ii) activated conidia (3 hpi), (iii) germinated conidia (6 hpi), and (iv) developed hyphae (24 hpi), respectively. The analysis of localization and binding of Dabcyl-ES-c, TAMRA-ES-d, and Rhodamine-ES4 on germinated conidia and developed mycelium showed a dose-and time-dependent inhibition, as mentioned above. Germinated conidia took up labelled peptides and accumulated them 12 and 24 hpi in swollen and ballooned compacted cells of developed mycelia. In general, Fusarium mycelia appeared compacted, shorter, and thicker, and contained swollen tips (Fig. 5A-H) compared with germinated conidia treated with TAMRA-ES-a, where a weak accumulation of fluorescence was observed (Fig. 5I, J). Here, developed mycelia showed normal hyphal growth of thin and long branched filaments. Observation at earlier time points at 3 and 6 h was also conducted: TAMRA-ES-d was localized and bound only at the surface of activated conidia, germinated conidia, and developed mycelia (24 hpi), while Dabcyl-ES-c bound at the surface and accumulated inside activated conidia, germinated conidia, and developed mycelia (Fig. 6A-D and G-J). Furthermore, increasing the concentration to 90 µM resulted in swelling and enlargement of fresh conidia and formation of big vacuoles, which finally fused to a single large vacuole within the cells, forcing conidia to stop germination. Although TAMRA-ES-d bound only at surfaces of vacuolated conidia, Dabcyl-ES-c was detected both at cell surfaces and appeared to be inside the cell in vacuolated conidia (Fig. 6E, F, K, L). Increasing the concentration of ES-c or ES-d to 90 µM in developed mycelia (24 hpi) caused, after 6 h, swelling of cells, formation of vacuoles, and subsequent breaking/bursting of cell walls and release of vacuoles (Fig. 6M). In contrast, fresh conidia, germinated conidia, and developed conidia showed normal growth behaviour and lack of fluorescence 3 h and 6 h after application of the control peptide TAMRA-ES-a at concentrations of 30-90 µM ( Supplementary Fig. S4). Application of Rhodamine-ES4 showed inhibition behaviour similar to ES-c and ES-d. At the same time, fluorescence was observed at the cell surface and accumulated inside conidia and mycelia. Furthermore, conidia and mycelia treated with ES-a, ES-c, or ES4 were also stained with NBT to detect the production of ROS. NBT reacts with superoxide (O 2 -) anions to form a dark blue insoluble formazan compound. Conidia and mycelia treated with ES-c, ES-d, or ES4 strongly produced ROS. In addition, swollen and vacuolated conidia and mycelia were formed. Staining of conidia and mycelia treated with ES-a showed weak to no traces of ROS ( Fig. 6N-Q). All of the above described assays for localization and binding were also applied to U. maydis in order to quantify the effect of ES-c and ES-d on Ustilago spores. Similar to Fusarium, binding and activity of Dabcyl-ES-c, TAMRA-ES-d, and Rhodamine-ES4 on germinated Ustilago spores showed a dose-and time-dependent inhibition. Ustilago spores treated with 30 or 60 µM Dabcyl-ES-c and TAMRA-ES-d showed an accumulation of fluorescence. Spores became thicker and bigger than untreated wild-type spores and mycelia germination was inhibited or prevented completely (Fig. 7A, B, F,  G). At a concentration of 90 µM, treated spores were malformed within 3 h. Spores appeared swollen and enlarged and formed many vacuoles, which later assembled into few large vacuoles. TAMRA-ES-d was detected in whole spores while Dabcyl-ES-c was detected mainly in vacuoles (Fig. 7C, D, H, I). After 6 h, swollen and vacuolated cells started to burst, releasing vacuoles and cellular components (Fig. 7K). In contrast, germinated spores, which were treated with the control TAMRA-ES-a, showed normal spore growth behaviour and fluorescence could not be detected (Fig. 7L). Application of Rhodamine-ES4 resulted in the same inhibition behaviour as ES-c and ES-d. Fluorescence was observed in whole spores and in vacuoles. Furthermore, ROS production was also detected by NBT in spores treated with ES-c, ES-d, and Rhodamine-ES-4. Germinated Ustilago spores produced significant ROS amounts after application of 30 and 60 µM of the above-mentioned peptides (Fig. 7E, J). Traces of ROS were detected when ES-a was applied (Fig 7M).

ES-d loop domain is maize-specific
A protein sequence comparison of ES1-4 and entries in the RefSeq protein database showed that most regions of the ES family (including the peptide signal cleavage site) are well conserved in plant defensins and DEFLs. However, while ES-e (C-terminal region) is less conserved, the ES-d region has no similarity with other defensins/DEFLs and is specific for maize. Furthermore, amino acid residues of ES-c and ES-d targeting fungal growth inhibition are more conserved within plant defensins/DEFLs. In contrast, ES-c and ES-d residues affecting maize pollen tube burst are highly conserved within the ES family, but less in other plant defensins/ DEFLs ( Supplementary Fig. S5 and Table S1). In order to indicate the localization of important residues, Fig. 8 shows the 3D model of ES4 in two different orientations. Gly 1 , Arg 3 , Ala 4 , and Glu 5 in ES-c, and Leu 3 , Ile 4 , or Tyr 15 in ES-d affect pollen tube burst, while Gly 8 , Tyr 9 , and Thr 10 in ES-c, and Lys 13 , Cys 14 , and Tyr 15 in ES-d affect germination and inhibition in fungi. As shown, there is little overlap, indicating that ES peptides interact differentially or with different targets at the surface/cell wall of pollen tubes and fungal cells.

Discussion
A previous study (Amien et al., 2010) showed that extracellular cysteine-rich peptide precursors such as a Bowman-Birk trypsin inhibitor from soybean (Glycine max), pollen tube attractant LURE2 of wishbone flower (Torenia fournieri), or anti-fungal protein AFP2 from radish (Raphanus sativus) did not affect maize pollen tube growth behaviour, and thus pointed towards a very specific activity of ES peptides. Moreover, application of ES4 caused a significant delay of pollen tube rupture in the close maize relative eastern gamagrass (Tripsacum dactyloides) but did not have any effect on tubes of unrelated maize species such as tobacco (Nicotiana benthamiana and N. tabaccum) or lily (Hemerocallis fulva). These observations further indicate that ES peptides are species-specific. In the present study, these findings were extended to show that other ES members (ES1-3) are capable of inducing maize pollen tube burst in the same manner. Additionally, an ES peptide fragment that is specific for maize, named ES-d, was found to bind at the pollen tube surface, accumulate at tube tips, and cause swelling prior to burst. These effects were not observed for other less-specific ES peptide fragments. ES-d contains a putatively exposed loop region of the mature peptide as well as amino acids essential for the induction of pollen tube burst in maize. Thus, ES-d represents the most likely active region of the peptide that physically interacts with receptor(s) at the pollen tube surface. The specificity of this region, which is also different from corresponding regions of homologous peptides of other grasses, suggests that it may have played an important role during speciation and the formation of the genus Zea. Future experiments using pollen tubes from other grasses and the identification  of orthologous proteins in incompatible maize inbred lines and closely related species such as T. dactyloides will help to prove this hypothesis and thus may become important in overcoming hybridization barriers among grass species.
At significantly higher concentrations (10 µM and above), ES peptide fragments were found to possess broad antifungal activity. However, inhibition of cereal rust fungi, for example, is usually achieved by defensins NaD1 and NaD2 from flowers of tobacco (N. alata) at a concentration of 1-2 µM (Dracatos et al., 2014). Peach (Prunus persica) defensin DFN1 also displayed antifungal activity at 1 µM through specific interactions with membrane lipids of Botrytis cinerea, Penicillium expansum, and Monilinia laxa (Nannia et al., 2012). These findings indicate that ES peptides have lost most of their antifungal activity. Antimicrobial activity of ES4 has been studied previously, but direct antibacterial and antifungal activities against Pseudomonas syringae pv. tomato DC3000 (Pto DC3000) and Hyaloperonospora arabidopsidis (previously known as Peronospora parasitica), respectively, were not observed in in vitro experiments (Amien et al., 2010). However, when continuously expressed at high amounts under regulation of the 35S promoter, ES4-overexpressing seedlings infected with H. arabidopsidis recovered faster than controls. Additionally, fungal hyphae were no longer macroscopically visible a few weeks after infection. The mild effect on H. arabidopsidis, an obligate parasite and causal agent of the downy mildew of the plant model organism Arabidopsis thaliana (Schlaich and Slusarenko, 2009;Fry and Grünwald, 2010), might, however, also be due to the fact that this pathogen is unspecific for maize and other grasses. In contrast, ES1-4 showed a high and significant activity of two fungal pathogens of maize in the present study. Both fungi cause severe diseases during flowering time in maize through infection of reproductive tissues (Banuett, 1995;Klosterman et al., 2007;Shea Miller et al., 2007;Dean et al., 2012). Interestingly, both ES-c and ES-d peptides showed antifungal responses, but they bound differently to Fusarium and Ustilago and even differently among fungus cell types. These experiments indicate that ES peptides likely interact with multiple receptors/ targets in fungi.
In this study, applications of ES1-4 resulted in swelling and rupture of both pollen tube and fungal cells. Swelling and vacuole formation (in fungal cells) culminating in rapid cell rupture points towards water uptake. This may be caused by altered solute transport across the plasma membrane and/ or increased membrane permeability. Killing of cells by antimicrobial peptides has been proposed to occur in two main ways: (i) through disruption of the plasma membrane leading to leakage of cytoplasmic contents or (ii) through interaction with intracellular targets (Brogden, 2005). The results from this study suggest that cell death occurred through both ways because peptides localized first at the cell surface then later inside cells. The plant defensin NaD1 from tobacco (N. alata) flowers was previously shown to display a potent antifungal activity against a variety of agronomically important filamentous fungi including F. oxysporum f. sp. vasinfectum (Fov). NaD1 entering the cytoplasm of hyphae results in granulation of the cytoplasm and cell death through ROS production (Van Der Weerden et al., 2008) similar to the activity of the studied ES peptides. NaD1 and the antifungal defensin AFP2 from seed of Raphanus sativus were also shown to possess activity against the major human pathogen Candida albicans. The mechanism involves interaction between NaD1 or AFP2 and the fungal cell surface followed by membrane permeabilization, entry into the cytoplasm, production of ROS, and cell death induced by oxidative damage (Aerts et al., 2007;Hayes et al., 2013). It is still unknown whether permeabilization by ES-c or ES-d causes cell death by inducing leakage of cytoplasmic contents, by ion fluxes, or by facilitating entry of peptides into cells to access intracellular targets.
Plant defensins possess a variety of biological activities beside inhibition of ion channel function and have been reported, for example, to inhibit protein synthesis and enzyme activity (Vriens et al., 2014). Using the microelectrode impalement technique, Amien et al. (2010) showed that ES4 application leads to H + and/or K + flux and that ES4 target(s) likely represent ion-channel(s). Detailed studies expressing the maize potassium channel KZM1 in Xenopus oocytes showed that ES4 application leads to opening of the channel, which is almost closed at physiological conditions. The current model suggests that opening of KZM1 leads to flux of the strong osmoticum K + , which may cause membrane depolarization, but which may also influence the conductivity of a battery of voltage-dependent transporters located at the pollen tube plasma membrane and, ultimately, culminate in osmotic burst (Dresselhaus and Franklin-Tong, 2013). ES targets localize almost uniformly at the surface (likely plasma membrane) of the pollen tube, displaying some accumulation at the very tip of the pollen tube. This pattern further indicates that KZM1 may not represent the only interactor as its accumulation in the tip was not observed in pollen tubes transiently transformed with KZM1.
ES peptides also strongly inhibit the growth of F. graminearum and U. maydis in vitro, indicating that ES peptides may be able to regulate channels of the KZM1 type in fungal cells, but may also interact with other yet unknown channels. The maize defensin subclass of γ-zeathionins, for example, was shown to block Na + channels in animal cells (Kushmerick et al., 1998) and defensin MsDef1 from alfalfa (Medicago sativa) seeds was reported to block the mammalian L-type Ca 2+ channel. Notably, antifungal activity of MsDef1 is markedly reduced in the presence of Ca 2+ (Spelbrink et al., 2004). Treatment of hyphae of Neurospora crassa with antifungal plant defensins, for example with Rs-AFP2, Dm-AMP1, and α-hordothionin isolated from radish, dahlia (Dahlia merckii), or barley (Hordeum vulgare) seeds induced a rapid K + efflux, Ca 2+ uptake, and alkalinization of the incubation medium, suggesting that plant defensins act on multiple channels and/or receptors (Thevissen et al., 1996;Vriens et al., 2014). K + channels were recently also found to be inhibited by animal toxin-like human β-defensin 2, which interacts with the extracellular channel pore region . The K + channel-blocking reagent plectasin, a defensin blocker from the fungus Pseudoplectania nigrella, could block Kv1.3 channel currents in a dose-dependent manner. Besides Kv1.3, plectasin could also inhibit other channels like Kv1.1, Kv1.2, IKCa, SKCa3, hERG, and KCNQ with less activity . Scorpion defensins, analogues of ES1-4, were also reported to affect functions of multiple ion channels for Na + , K + , and Ca 2+ (Quintero-Hernandez et al., 2013). Together, these findings not only highlight the function of defensins as ion channel targets, but also imply that defensins from different organisms functionally evolved different ion channels as targets. How ES1-4 peptides interact with fungal hyphae or with maize pollen tubes to generate plasma membrane ion fluxes remains unclear and needs further investigation.
The picture may become even more complex, as defensins/ DEFLs also interact with receptor-like kinases. In A. thaliana, for example, TAPETUM DETERMINANT 1, EPIDERMAL PATTERNING FACTOR (EPF), and related EPF-Like, and S-LOCUS CYSTEINE RICH proteins were shown to interact with receptor-like kinases rather than ion channels (for review see Butenko and Aalen, 2012). To understand the complexity of ES activity and the specificity of its subdomains, a systematic, preferentially biochemical, approach is now necessary to elucidate all interaction partners of ES peptides both in pollen tubes and fungal cells.

Supplementary data
Supplementary data are available at JXB online. Fig. S1. ES-a and ES-c peptides do not bind to maize pollen tubes. Fig. S2. Identification of amino acid residues in ES-c and ES-d peptides affecting germination of Fusarium graminearum. Fig. S3. Identification of amino acid residues in ES-c and ES-d peptides affecting germination of Ustilago maydis. Fig. S4. ES-a peptide does not bind to F. graminearum. Fig. S5. ES family peptides lack homology to plant defensins/DEFLs in regions responsible for their pollen tube burst activity. Table S1. List of aligned plant defensin/DEFL proteins homologous to ES1-4.