Ca2+-binding proteins in the watery saliva of Megoura viciae counteract Ca2+-dependent occlusion of sieve plates in Vicia faba and so prevent the shut-down of food supply in response to stylet penetration. The question arises whether this interaction between aphid saliva and sieve-element proteins is a universal phenomenon as inferred by the coincidence between sieve-tube occlusion and salivation. For this purpose, leaf tips were burnt in a number of plant species from four different families to induce remote sieve-plate occlusion. Resultant sieve-plate occlusion in these plant species was counteracted by an abrupt switch of aphid behaviour. Each of the seven aphid species tested interrupted its feeding behaviour and started secreting watery saliva. The protein composition of watery saliva appeared strikingly different between aphid species with less than 50% overlap. Secretion of watery saliva seems to be a universal means to suppress sieve-plate occlusion, although the protein composition of watery saliva seems to diverge between species.
Sieve tubes – arrays of longitudinally arranged sieve elements (SEs)– are the transport conduits for an assortment of metabolites such as carbohydrates, amino acids, phytohormones and vitamins (reviewed by van Bel, 2003). SEs are elongated cells that lack a nucleus and vacuole and contain only an intact plasma membrane, phloem plastids, a SE endoplasmic reticulum (Sjölund and Shih, 1983), a few mitochondria, and an extensive set of phloem-specific proteins (reviewed by Hayashi et al., 2000). During SE development, the terminal walls transform into sieve plates, perforated by sieve pores (Evert, 1990) to connect adjacent SEs end to end.
Following production in the mesophyll and loading into sieve tubes, plant nutrients are translocated through the transport phloem towards the sinks, e.g. growing leaves, roots, stem or storage organs (van Bel, 1996). Mass flow through sieve tubes is driven by a turgor-pressure gradient between sources (high pressure) and sinks (low pressure) (Münch, 1930; Gould et al., 2005).
Aphids (Homoptera: Aphididae: Aphidinae) puncture sieve tubes with their stylets and passively ingest the nutrient-rich sieve-tube content owing to the high endogenous pressure. During stylet penetration, aphids are confronted with sieve-tube occlusion mechanisms (Sjölund, 1997; Knoblauch and van Bel, 1998; Ehlers et al., 2000). When damage occurs, occlusion of sieve pores prevents the loss of sieve-tube sap (Evert, 1982; Schulz, 1998) and the invasion of pathogens through the wound site (van Bel, 2003).
Sieve tubes are occluded by proteins and callose in response to injury. The nature of protein-occlusion mechanisms depends upon the plant family. Fabaceae possess parietal proteins, protein networks and reversibly dispersive protein bodies, the so-called forisomes (Knoblauch and van Bel, 1998; Ehlers et al., 2000; Knoblauch et al., 2001). Cucurbitaceae possess parietal proteins and solubilized phloem proteins (PP1) that form protein filaments after oxidation (Leineweber et al., 2000). In electron-microscopic images of Brassicaceae phloem, SEs show merely protein networks that span the SE lumen and are attached to the cell periphery (Sjölund, 1997). Sieve tubes of grasses appear virtually empty but may have an occlusion mechanism based on solidification of soluble proteins (Will and van Bel, 2006). To complement protein plugging, sieve pores are occluded by callose deposition (McNairn and Currier, 1968), a universal mechanism in all families studied so far [summarized by Behnke and Sjölund (Behnke and Sjölund, 1990)].
Induction of callose synthesis (Kauss et al., 1983), coagulation of soluble proteins (Will and van Bel, 2006) and dispersion of forisomes (Knoblauch et al., 2001) appear to be Ca2+ dependent. Occlusion is triggered by Ca2+ influx induced by damage (Knoblauch and van Bel, 1998). Aphids seem to have developed strategies to prevent sieve-tube occlusion during stylet penetration, one of which is the secretion of Ca2+-binding watery saliva (Tjallingii, 2006; Will and van Bel, 2006; Will et al., 2007).
As shown by the electrical penetration graph (EPG) technique [for an overview, see Walker (Walker, 2000)], watery saliva is massively secreted into the SE lumen at the beginning of sieve-tube penetration (Prado and Tjallingii, 1994). Some watery-saliva proteins of the aphid species Megoura viciae (Buckton) have Ca2+-binding properties (Will et al., 2007). This was demonstrated by confronting forisomes with salivary proteins in vitro. Forisome dispersal (which occludes sieve plates) can be triggered in vitro by supply of Ca2+, and dispersal can subsequently be reversed by adding Ca2+ chelators (Knoblauch et al., 2001) or concentrated aphid saliva (Will et al., 2007).
While feeding on SEs, aphids switch from phloem sap ingestion (EPG waveform E2) to secretion of watery saliva (EPG waveform E1) upon leaf-tip burning (Will et al., 2007). Leaf-tip burning was observed to induce an electropotential wave, accompanied by Ca2+ influx into the sieve-tube lumen resulting in sieve-tube occlusion (Fromm and Bauer, 1994; Furch et al., 2007). As demonstrated by Gould and colleagues (Gould et al., 2004), sieve-tube occlusion is accompanied by a decrease of sieve-tube pressure (Gould et al., 2004). This pressure decrease inside the sieve-tube lumen most likely triggers the switch in aphid behaviour (Will et al., 2008).
We investigated whether watery saliva acts as a universal tool in aphid–plant interactions through suppression of sieve-tube occlusion. We used the behavioural switch from phloem-sap ingestion to secretion of watery saliva (Will et al., 2007) as an indicator for suppression of sieve-tube occlusion by watery saliva in several aphid–plant combinations. Watery saliva from different aphid species was collected to compare their protein composition and investigate potential overlap with regard to Ca2+-binding proteins.
MATERIALS AND METHODS
Aphid and plant breeding
Aphid species were reared on their host plants (Table 1) in Perspex cages with large gauze-covered windows in a controlled room environment with a 17 h:7 h light (L):dark (D) regime at a temperature of 25°C. The same aphid/host plant combinations for rearing were also used for experiments.
Plants (Table 1) were grown in a greenhouse at 20°C with natural lighting plus additional lamp light (SONT Agro 400 W; Phillips, Eindhoven, The Netherlands) with a 14 h:10 h L:D period and were used in a vegetative phase.
Sieve-tube occlusion visualised by confocal laser scanning microscopy
In order to document sieve-tube occlusion in response to leaf-tip burning, Brassica napus and Hordeum vulgare plants were prepared as described previously (Knoblauch and van Bel, 1998). A few cortical cell layers were locally removed down to the phloem from the lower side of the main vein of a mature leaf that remained attached to the plant during the entire experiment. The layers were removed by manual paradermal slicing with a new razor blade, while avoiding damage to the phloem. Immediately after cutting a shallow window (∼10 mm long, 2 mm wide and 6 cm away from the leaf tip), we bathed the free-lying phloem tissue in a buffer solution (Hafke et al., 2005) and the plant was allowed to recover for 30–60 min. The leaf was fixed in an upside-down position with double-sided adhesive tape onto a small Perspex table. The state of the tissue was checked under the microscope. If the phloem appeared to be undamaged (indicated by thin sieve plates), a loading area (5 mm×5 mm) was made by gently rubbing the lower side of the leaf vein, located 2–3 cm upstream from the observation window, with fine sandpaper (Fig. 1A).
A stock solution of the phloem-mobile fluorochrome 5(6) carboxyfluorescein diacetate (CFDA)-mixed isomers (Invitrogen, Karlsruhe, Germany) was prepared by solubilisation of 1 mg CFDA in 1 ml DMSO. A working solution of 1μl stock solution in 2 ml buffer solution (Hafke et al., 2005) was applied at the loading site. CFDA permeates the plasma membrane in the non-fluorescent acetate form and is cleaved by cytosolic enzymes producing membrane-impermeant fluorescent carboxyfluorescein (CF) (handbook from Molecular Probes, Eugene, OR, USA). CF trapped inside SEs is transported by mass flow in the sieve tubes as observed by confocal laser scanning microscopy (CLSM; Leica TCS 4D; Leica Microsystems, Heidelberg, Germany) (Knoblauch and van Bel, 1998). After CF had reached the observation window, the leaf tip was burnt for 3 s (Fig. 1A,B; maximum area 1 cm2) to induce blockage of SEs. To visualize stoppage of mass flow in SEs, CF was photobleached at the observation window for approximately 3 min. If phloem transport was blocked, the photobleached SEs would remain non-fluorescent because of the lack of fluorochrome supply (Fig. 1C).
After burning and subsequent photobleaching, a time course of the events in the SEs was recorded. The experiment was repeated four times for each plant species with new individuals. Digital image processing was done with Corel PHOTO-PAINT®10 to optimize brightness, contrast and colouring.
Aphid behaviour monitored by EPGs
Aphid behaviour was monitored using the EPG technique modified by Tjallingii (Tjallingii, 1988). A gold wire electrode (1 cm length and 20μm diameter) was attached to the dorsum of an adult apterous aphid using electrically conductive silver glue (Electrolube, Swadlincote, Derbyshire, UK). A vacuum device was used to immobilize aphids during electrode attachment (van Helden and Tjallingii, 2000). The aphid electrode was connected to a DC EPG Giga-8 (Tjallingii, 1978; Tjallingii, 1988) and the EPG output was recorded with PROBE 3.5 (hardware and software from EPG-Systems, Wageningen, The Netherlands, www.epgsystems.eu). A second electrode (plant electrode) was inserted into the soil of potted plants. In this electrical circuit, the `living' components, represented by the aphid and plant, act as a variable resistor, and plant cell membrane potentials present additional voltage sources (reviewed by Walker, 2000). The whole setup was placed in a Faraday cage to shield it from electromagnetic influence.
Randomly selected adult and apterous aphids were placed on their particular host plant species (Table 1) on the midrib of the lower side of a mature leaf 6±0.5 cm away from the leaf tip. When aphids showed ingestion of phloem sap for longer than 30 min, recognized by an E2 waveform in the EPG (Tjallingii, 1988), the leaf tip (of maximum area 1 cm2) was carefully burned for 3 s to trigger a phloem-mediated electrical long-distance signal (Furch et al., 2007). In EPG recordings, the time point of burning was marked by generating three –50 mV pulses by pressing the EPG amplifier calibration button. EPG recordings of control insects (no leaf-tip burning) were made under identical experimental conditions, including the three –50 mV pulses. Burning experiments and controls were repeated 12 times in most cases (except for Rhopalosiphum padi, burning N=13; Macrosiphum euphorbiae, control N=7) and recording time was set to 1 h.
EPG waveforms were analyzed by waveform pattern and auto power spectra (APS) (example waveforms and APS are presented in Fig. 2) according to Prado and Tjallingii (Prado and Tjallingii, 1994), using the PROBE analysis module. APS present the waveform frequency (Hz) vs the relative magnitude with a maximum of 1, which provides a major characteristic of waveform identity.
Recorded EPGs were analyzed with regard to switches in behaviour expressed as transitions from one waveform to another, e.g. from E2 (ingestion of sieve-tube sap) to E1 (secretion of watery saliva into sieve-tube lumen) as observed for M. viciae (Will et al., 2007). EPGs from each aphid were scored as a `reaction' if they changed from E2 to E1 and as a `non-reaction' if they remained in E2 for 10 min after leaf-tip burning, or for controls, within 10 min of the three– 50 mV pulses.
Statistical analysis was done with Fisher's exact test using SigmaStat 3.0 software (SPSS, Chicago, IL, USA). Each treatment was compared against its control group. The level for significance was set to P=0.05.
Aphid saliva collection and saliva concentration
Watery saliva secreted by the aphid species M. viciae, Acyrtosiphon pisum (red and green variant), M. euphorbiae, Aphis fabae and Schizaphis graminum was collected and concentrated according to a previous method (Will et al., 2007). Saliva from 100 aphids in 1 μl of saliva/diet-concentrate was used as a benchmark for concentrating the diet/saliva mixture of each aphid species. The final sample was partitioned in aliquots of 10μl and kept frozen at –80°C.
1-D SDS-PAGE of the saliva concentrate (sample of 1000 aphids per lane) was carried out according to Laemmli (Laemmli, 1970) using a 4% stacking gel and a 10% separation gel in a MiniProtean 3 Electrophoresis System (Bio-Rad Laboratories, Hercules, CA, USA). Precision Plus Protein All Blue standards (Bio-Rad) was used as a protein size marker. Fourfold concentrated reducing sample buffer (Roti-Load 1; Carl-Roth, Karlsruhe, Germany) was added to each saliva sample in the proportion 1:3. Protein gels were silver stained (Switzer and Merril, 1979; Schoenle et al., 1984). Similarities of protein band patterns were calculated using Quantity One® 1-D Analysis Software (Bio-Rad) including all detected bands and setting the band match tolerance to 1%. Protein band intensity was not weighted because of the varying amounts of protein in the respective samples.
Time course of events in sieve tubes after leaf-tip burning
CFDA was applied to the loading site on B. napus leaves and transported downstream in the CF form through the sieve tubes. Following phloem transport of CF in B. napus leaves, the leaf tip was burnt (Fig. 1A,B) in order to induce sieve-plate occlusion. Subsequently, CF was photobleached at the observation window (Fig. 1C). The rationale for these treatments is that fluorescence will not recover at the site of observation until the sieve plates are unblocked (Fig. 1D). Fluorescence inside the observed SEs (marked with an asterisk in Fig. 1) was reduced by photobleaching (Fig. 1E,F). A continued bleached appearance (Fig. 1F–I) demonstrates the absence of fluorochrome supply and is evidence of stoppage of mass flow over a period of 30 min (Fig. 1C). Between 30 and 60 min after burning (Fig. 1J) fluorescence recovered, indicating that mass flow had resumed (Fig. 1D). Similar results were obtained for all replicates (N=4) and for both plant species (data for H. vulgare not shown).
EPG and leaf-tip burning
As with V. faba (Will et al., 2007), leaf-tip burning was adopted to observe whether food deprivation caused by sieve-tube occlusion triggers behavioural responses of other aphid species. Each aphid species (Table 1) switched its behaviour from ingestion (E2: Fig. 2B; Fig. 3) to secretion of watery saliva (E1: Fig. 2C; Fig. 3), as shown in detail for the aphid species M. euphorbiae (Fig. 2A–D) and in 1 h overviews for other tested aphid species (Fig. 3). The behavioural switch was observed for all tested aphid–plant combinations (Table 2) shortly after leaf-tip burning (Figs 2 and 3) (M. persicae: 6 s; M. euphorbiae: 10 s; A. fabae: 13 s; S. graminum: 17 s; R. padi: 18 s; A. pisum: 23 s). The response is significant with P-values that range from <0.001 (M. viciae, A. fabae, M. euphorbiae, R. padi and S. graminum) and 0.012 (M. persicae) to 0.036 (A. pisum, green variant) (Table 2). Macrosiphum euphorbiae resumes ingestion, preceded by mixed E1/E2 behaviour, approximately 25 min after leaf-tip burning (Fig. 2A,D).
Apart from the typical E2–E1 switch, individuals showed different behaviour profiles within a species. An example of analysis of the behaviour of 12 M. viciae individuals in reaction to SE occlusion demonstrates the variability of the responses (Fig. 4). Ten of 12 aphids reacted to leaf-tip burning by secretion of watery saliva followed by a mixed E1/E2 behaviour. The return from mixed E1/E2 behaviour to ingestion occurred most frequently (5/10; Fig. 4). Another frequent switch (4/10) was that from mixed behaviour back to secretion of watery saliva (E1), followed by a broad variety of behavioural profiles (Fig. 4). Other aphid species showed a similar diversity of behaviour.
Comparison of watery saliva from six aphid species by 1-D SDS-PAGE
Protein band patterns of the different aphid species/biotypes (Fig. 5) show only partial overlap (Fig. 5, red bands) when matched to M. viciae as the reference species (Fig. 5, green bands). Even if aphids are reared and fed on the same host plant [M. viciae, A. pisum (both variants) and A. fabae on V. faba], the similarity is low. A calculation for similarity shows values of 19.2–43.1% [43.1% S. graminum; 34.1% A. pisum (green variant); 32.9% A. pisum (red variant); 19.2% A. fabae; 30.1% M. euphorbiae] when compared with M. viciae. The protein pattern varies not only between species but also between variants of A. pisum (green and red). Watery saliva proteins of most species are distributed over the whole protein marker range (Fig. 5).
Protein concentration varies strongly between species (Fig. 5), although each sample was collected from 1000 aphids. Roughly, the amount of secreted saliva seems to be positively correlated with the size of the aphid species. The total protein amount in watery-saliva samples can affect the number of detectable protein bands (Will et al., 2007). Thus, a comparison of samples with differing protein concentrations is not conclusive for determining differences in protein bands (e.g. Fig. 5, red and green biotype of A. pisum). Nevertheless, samples with comparable protein concentrations (e.g. Fig. 5, M. viciae, A. fabae and M. euphorbiae) show a low similarity as well, which suggests that dissimilarity is not due to a difference in the amount of protein at least in the comparison among the samples of M. viciae, S. graminum, A. fabae and M. euphorbiae in Fig. 5.
Sieve-tube occlusion mechanisms are presumably involved in plant defence against phloem-feeding insects by blockage of nutrient supply (Caillaud and Niemeyer, 1996). Aphids appear to be able to counteract sieve-tube occlusion by secretion of watery saliva (Tjallingii, 2006; Will and van Bel, 2006; Will et al., 2007). This notion is based on the following observations: (i) aphids start secreting watery saliva (E1) at the beginning of each SE penetration (Prado and Tjallingii, 1994), (ii) M. viciae switches to secretion of watery saliva in response to sieve-tube occlusion triggered by distant leaf-tip burning (Will et al., 2007), (iii) watery saliva concentrate of M. viciae inhibits Ca2+-induced dispersion of forisomes (legume proteins involved in sieve-plate plugging) (Will et al., 2007), and (iv) several watery saliva proteins of M. viciae are able to bind Ca2+ in biochemical assays (Will et al., 2007). To answer the question of whether suppression of sieve-tube occlusion by aphid watery saliva is a general phenomenon, the leaf-burning test (Will et al., 2007) was extended to a variety of aphid–host plant combinations (Table 1). Aphid behavioural responses were monitored with EPG recordings (Figs 2 and 3; Table 2) and protein composition of watery saliva was compared among different aphid species/biotypes (Fig. 5).
Callose-mediated sieve-tube occlusion induced by remote leaf-tip burning has been visualized for broad bean, tomato (Furch et al., 2007) and cucurbits (A.C.U.F., unpublished). Prior to callose deposition, sieve tubes are plugged by proteins such as giant protein bodies (forisomes) in V. faba (Knoblauch et al., 2001) or probably structural phloem proteins in cucurbits (Leineweber et al., 2000). Sieve-plate plugging by forisomes was reversed by the time (15–20 min after burning) that callose deposition reached its maximum level (Furch et al., 2007). Thus, sieve-plate occlusion seems to be a two-step event in which protein plugging precedes callose deposition (Will and van Bel, 2006).
Both Rosopsida and Liliopsida employ callose deposition on sieve plates for occlusion (Eleftheriou, 1990; Evert, 1990). In addition, Rosopsida also employs structural proteins to occlude SEs (Evert, 1990), whilst we (Will and van Bel, 2006) speculated about the existence of soluble proteins in phloem sap of Liliopsida which coagulate in response to wounding. All in all, the rapidity of the observed sieve-tube occlusion in B. napus (Rosopsida) and H. vulgare (Liliopsida) (Fig. 1) and the speed of aphid reaction to leaf-tip burning (Figs 2 and 3) suggest sieve-plate plugging by proteins prior to callose deposition, because the latter is a slower response that requires several minutes for synthesis (Furch et al., 2007).
Activation of occlusion seems to depend on increasing Ca2+ concentration in SEs (Kauss et al., 1983; Knoblauch and van Bel, 1998; Knoblauch et al., 2001). Induction of sieve-plate occlusion by electropotential waves (Fromm and Bauer, 1994; Furch et al., 2007) and turgor shocks (Knoblauch et al., 2001; Hafke et al., 2007) triggers a release of Ca2+ into the SE lumen which then induces occlusion. Furch and colleagues (Furch et al., 2007) demonstrated that sieve-tube occlusion is reversible probably through active removal of Ca2+ ions.
The crucial role of Ca2+ in sieve-tube occlusion makes it an outstanding target for insects and pathogens in order to interfere with occlusion mechanisms. Therefore, the switch from ingestion to secretion of watery saliva, provoked by leaf-tip burning (Figs 2 and 3; Table 2), has been interpreted as an attempt to suppress sieve-plate occlusion (Will et al., 2007). In addition to the Ca2+-binding proteins that were described for M. viciae, one was recently identified in the saliva of A. pisum (Carolan et al., 2009). Aphids may use proteins in watery saliva as a tool for Ca2+ binding (Will and van Bel, 2006; Will et al., 2007). Thus, secretion of watery saliva into the SE lumen can be regarded as a method to counteract sieve-plate occlusion, although it does not seem to succeed in all cases (Figs 3 and 4).
This study indicates that the involvement of watery salivation in the suppression of plant defence is independent of aphid species [N=7; monophagous (1), oligophagous (3) and polyphagous (3)], host plant species (N=4) and host family (Fabaceae, Brassicaceae, Cucurbitaceae and Poaceae) (Figs 2 and 3; Table 2).
Detailed behavioural analysis of the aphid species M. viciae (Fig. 4) shows that 10 of 12 aphids reacted to leaf-tip burning and that seven of these returned to ingestion without interruption of sieve-tube penetration. The mixed E1/E2 behaviour appears to express testing of the penetrated sieve tube for its further suitability as a source of nutrition. A change of turgor pressure in sieve tubes caused by occlusion (Gould et al., 2004) represents a relevant signal for aphids to initiate the behaviour switch from ingestion to salivation, observed in vitro for M. persicae (Will et al., 2008).
Aphid behaviour that has been observed by EPG thus far [e.g. Brevicoryne brassicae, M. viciae and M. euphorbiae (Tjallingii, 1985); R. padi (Prado and Tjallingii, 1994); Aphis gossypii and M. persicae (Martin et al., 1997); A. pisum (Tjallingii and Gabrys, 1999)] suggests that aphids possess only a small and conserved repertoire of behaviour which hardly varies among aphid species and host plants. The key for adaptation to the respective host plant(s) may lie in the watery saliva, which is astonishingly variable in quantity and protein composition, even if aphid species feed on the same host plant (Fig. 5). The variable composition may reflect an evolutionary adaptation to a variety of defence mechanisms of host plant(s) (e.g. phenols, hydrogen peroxide) other than Ca2+-induced occlusion. Several watery-saliva proteins which are not involved in Ca2+ binding may function in the detoxification of such compounds (Miles and Oertli, 1993; Urbanska et al., 1998).
LIST OF ABBREVIATIONS
- auto power spectrum
- 5(6)carboxyfluorescein diacetate
- confocal laser scanning microscopy
- direct current
- electrical penetration graph
- sieve element
We thank Gregory P. Walker (Department of Entomology, University of California, Riverside, USA) for critical reading of the manuscript. We are grateful to Edgar Schliephake (Julius Kühn-Institut, Institut für Resistenzforschung und Stresstoleranz, Quedlinburg, Germany) for supplying a starter colony of M. persicae.