Arabidopsis ACTIN-DEPOLYMERIZING FACTOR3 Is Required for Controlling Aphid Feeding from the Phloem^,

Green peach aphid feeding from the sieve elements is shown to be restricted by ACTIN-DEPOLYMERIZING FACTOR3, thus implicating actin-dependent process in controlling insect feeding from the phloem.

Abstract

The actin cytoskeleton network has an important role in plant cell growth, division, and stress response. Actin-depolymerizing factors (ADFs) are a group of actin-binding proteins that contribute to reorganization of the actin network. Here, we show that the Arabidopsis (Arabidopsis thaliana) ADF3 is required in the phloem for controlling infestation by Myzus persicae Sülzer, commonly known as the green peach aphid (GPA), which is an important phloem sap-consuming pest of more than fifty plant families. In agreement with a role for the actin-depolymerizing function of ADF3 in defense against the GPA, we show that resistance in adf3 was restored by overexpression of the related ADF4 and the actin cytoskeleton destabilizers, cytochalasin D and latrunculin B. Electrical monitoring of the GPA feeding behavior indicates that the GPA stylets found sieve elements faster when feeding on the adf3 mutant compared to the wild-type plant. In addition, once they found the sieve elements, the GPA fed for a more prolonged period from sieve elements of adf3 compared to the wild-type plant. The longer feeding period correlated with an increase in fecundity and population size of the GPA and a parallel reduction in callose deposition in the adf3 mutant. The adf3-conferred susceptibility to GPA was overcome by expression of the ADF3 coding sequence from the phloem-specific SUC2 promoter, thus confirming the importance of ADF3 function in the phloem. We further demonstrate that the ADF3-dependent defense mechanism is linked to the transcriptional up-regulation of PHYTOALEXIN-DEFICIENT4, which is an important regulator of defenses against the GPA.

In eukaryotes, the actin cytoskeleton, which is composed of filamentous (F)-actin, has a central role in multiple cellular processes, including cell growth, division and differentiation, regulation of polarity, and facilitating cytoplasmic streaming, organelle movement and response to the environment (Staiger, 2000; Hussey et al., 2006; Staiger and Blanchoin, 2006; Pollard and Cooper, 2009; Szymanski and Cosgrove, 2009; Day et al., 2011; Henty-Ridilla et al., 2013). This microfilament network is dynamic and requires continuous reorganization. Remodeling of the actin network involves severing, depolymerization, and polymerization of F-actin, which needs to be spatially and temporally regulated. A variety of actin-binding proteins are involved in remodeling of the actin cytoskeleton (Hussey et al., 2006). These include the actin-nucleating and filament-stabilizing proteins like the formins and fimbrins, respectively, and the actin-depolymerizing factor (ADF) family of proteins (Vidali et al., 2009; Ye et al., 2009; Wu et al., 2010). As a result of their ability to sever and depolymerize F-actin and yield products with ends that can serve as sites for new filament initiation, the ADFs, which are small proteins (approximately 17 kD), increase the dynamics of the actin cytoskeleton and the balance of F-actin versus the free globular (G)-actin (Andrianantoandro and Pollard, 2006; Pavlov et al., 2007). ADFs are also involved in shuttling actin into the nucleus (Nebl et al., 1996; Jiang et al., 1997), where actin is a component of chromatin-remodeling activities that control gene expression (Farrants, 2008; Jockusch et al., 2006). In addition to the cytosol, some ADFs also localize to the nucleus (Ruzicka et al., 2007).

The ADF family in Arabidopsis consists of 11 expressed genes, which based on their relatedness to each other have been grouped into four subclasses (Feng et al., 2006). These ADF subclasses exhibit novel and differential tissue-specific and developmental expression patterns (Ruzicka et al., 2007). For example, the subclass I genes, which include ADF1, ADF2, ADF3, and ADF4, are constitutively expressed in a variety of vegetative and reproductive tissues, including roots, leaves, flowers, and young seedlings. Among the subclass I genes, ADF3 was the most strongly expressed. The subclass II genes can be subdivided into clade IIa (ADF7 and ADF10) and clade IIb (ADF8 and ADF11). ADF7 and ADF10 exhibit high levels of expression in the reproductive tissues, with strongest expression in mature pollen grain. In comparison, expression of the clade IIa genes is relatively poor in roots and leaves. By contrast, the clade IIb ADF8 and ADF11 genes show highest expression in epidermal cells that develop into root hairs. ADF5 and ADF9, which belong to subclass III, exhibit strongest expression in the fast-growing tissues, including the root meristem and emerging leaves. ADF6, which is the lone member of subclass IV, is expressed in all tissues including pollen. Biochemical characterization of Arabidopsis ADFs indicated that all class I, II, and IV ADFs possess F-actin-severing/depolymerizing activity, with class I ADFs possessing the strongest activities (Nan et al., 2017). In comparison, the class III ADFs lacked F-actin-severing/depolymerizing activity, and instead possessed F-actin bundling activity (Nan et al., 2017).

ADFs and actin cytoskeleton dynamics are involved in plant response to pathogens. In Arabidopsis, the actin polymerization inhibitor cytochalasin E attenuated nonhost resistance against Blumeria graminis f. sp. tritici (Yun et al., 2003). Cytochalasin E also interfered with the targeting of the resistance protein RPW8.2 to the extrahaustorial membrane in Arabidopsis inoculated with the powdery mildew Golovinomyces spp. fungi (Wang et al., 2009). Similarly, RPW8.2 localization was also affected in plants overexpressing ADF6, thus suggesting that the specific targeting of RPW8.2 to the extrahaustorial membrane is dependent on actin cytoskeleton dynamics (Wang et al., 2009). Arabidopsis ADF4 is required for resisting infection by the bacterial pathogen Pseudomonas syringae pv tomato DC3000 expressing the AvrPphB effector protein (Tian et al., 2009). ADF4 function in defense against this pathogen was linked to the transcriptional regulation of the Arabidopsis RPS5 gene, which encodes a resistance protein that facilitates the recognition of the AvrPphB effector protein and the activation of downstream defense signaling (Porter et al., 2012). Thus, it was suggested that ADF4 links actin cytoskeleton dynamics to pathogen perception and defense activation (Porter et al., 2012). ADF4 was also shown to regulate actin dynamics and callose deposition in response to elf26, a microbial elicitor of immunity (Henty-Ridilla et al., 2014). In contrast, knockdown of ADF4 resulted in enhanced resistance against an Arabidopsis adapted strain of the powdery mildew fungus (Inada et al., 2016), while knockdown of ADF2 resulted in enhanced resistance to the root-knot nematode Meloidogyne incognita (Clément et al., 2009). Hence, there is a wider involvement of ADFs in plant defense as well as susceptibility to pests.

Aphids (Hemiptera: Aphididae) encompass a large group of insects that consume phloem sap. Nearly 250 species of aphids are considered pests of plants that limit plant growth and productivity due to removal of nutrients from sieve elements and their ability to alter source-sink patterns and thus the flow of nutrients to growing parts of the infested plant (Pollard, 1973; Blackman and Eastop, 2000; Goggin, 2007). Furthermore, some aphids also vector viral diseases (Kennedy et al., 1962; Matthews, 1991; Guerrieri and Digilio, 2008). Aphids utilize their mouthparts, which are modified into slender stylets, to feed from the sieve elements. Prior to inserting stylets into the plant tissue, aphids utilize chemosensory hairs on their antennae to assess the plant surface, potentially for gustatory cues (Powell et al., 2006). Subsequently, the stylets briefly penetrate nonvascular cells to sample cell contents for additional gustatory cues. The puncturing of nonvascular cells along the path of the stylet penetration likely results in the activation of host defenses that could potentially interfere with the ability of the stylet to reach a sieve element. Aphids produce two distinct salivary secretions that facilitate infestation (Miles, 1999). The proteinaceous gelling saliva, which is released by the stylets on their way to the sieve elements, forms a sheath that facilitates stylet movement through the plant tissue and simultaneously minimizes wound responses in the plant by quickly sealing off wounds. In comparison, the watery saliva, which is released when the stylets penetrate the sieve elements, contains factors suggested to enable the insect to prevent and maybe reverse phloem occlusion and thus facilitate feeding from the sieve elements (Miles, 1999; Will et al., 2007, 2009).

The plant surface provides the first line of defense (e.g. trichomes and glandular secretions) that could deter aphid settling on a plant (Walling, 2008). In addition, during the different stages of stylet penetration into the plant tissue, the insect encounters defenses that deter feeding and adversely impact insect fecundity (Walling, 2008; Louis and Shah, 2013). These include factors that contribute to sieve element occlusion (e.g. callose deposition and phloem protein aggregation) as well as insecticidal factors present in the phloem sap (Pedigo, 1999; Smith, 2005; Powell et al., 2006; Goggin, 2007; Walling, 2008). The interaction between Arabidopsis and the green peach aphid (GPA; Myzus persicae Sülzer) has been utilized as a model system to characterize plant genes and mechanisms that contribute to defense (Louis et al., 2012; Louis and Shah, 2013). GPA is a polyphagous insect that is an important pest of more than 400 plant species belonging to over 50 plant families, and the vector of more than 100 viral diseases (Kennedy et al., 1962; Matthews, 1991; Blackman and Eastop, 2000). The PHYTOALEXIN-DEFICIENT4 (PAD4) gene is an important regulator of Arabidopsis defenses that deters GPA feeding and adversely impacts GPA fecundity (Pegadaraju et al., 2005, 2007; Louis et al., 2010a). Genetic studies have indicated that the up-regulation of PAD4 expression in response to GPA infestation correlated with PAD4’s involvement in deterring GPA feeding from sieve elements. This feeding deterrence was intensified in plants overexpressing PAD4 from the Cauliflower mosaic virus 35S gene promoter (Pegadaraju et al., 2007). In contrast, basal expression of PAD4 was sufficient for the accumulation of an antibiotic activity that adversely impacts insect fecundity (Louis et al., 2010a, 2012).

ADF proteins and profilin, an actin-binding protein that influences actin polymerization, have been identified in phloem exudates from a variety of plants (Schobert et al., 1998, 2000; Kulikova and Puryaseva, 2002; Lin et al., 2009; Rodriguez-Medina et al., 2011; Fröhlich et al., 2012). Furthermore, microfilament meshwork has been revealed in sieve elements of fava bean (Vicia faba) injected with the actin-binding fluorescent phalloidin (Hafke et al., 2013). Immunolabeling with anti-actin antibodies further confirmed the presence of actin cytoskeleton in the sieve elements (Hafke et al., 2013). The identification of an actin-binding protein in aphid saliva has led to the suggestion that an actin-dependent process contributes to plant defense in the phloem and aphids’ attempt to curtail these defenses by targeting actin dynamics (Nicholson et al., 2012). We therefore investigated the contribution of ADF genes in the interaction of Arabidopsis with the GPA. We show that an ADF3-dependent mechanism is required for controlling GPA feeding from the sieve elements. We further demonstrate that the PAD4 gene is a critical downstream component of this ADF3-dependent defense mechanism.

RESULTS

ADF3 Is Required for Limiting GPA Infestation on Arabidopsis

To determine if the ADF genes influence infestation of Arabidopsis by the GPA, Arabidopsis lines that were previously shown to lack or accumulate reduced levels of the ADF1 (At3g46010), ADF2 (At3g46000), ADF3 (At5g59880), ADF4 (At5g59890), ADF5 (At2g16700), and ADF9 (At4g34970) transcripts (Clément et al., 2009; Tian et al., 2009) were utilized in no-choice bioassays with the GPA. For the no-choice bioassay, twenty adult apterous asexually reproducing insects were released on each plant and the insect population size (adults + nymphs) determined 2 d post-infestation. As shown in Figure 1, A and B, only plants lacking ADF3 function repeatedly showed GPA numbers that were significantly higher than GPA numbers on the wild-type plant. Compared to the wild-type plant, the GPA population was significantly larger (P < 0.05) on the adf3-1 (Salk_139265) and adf3-2 (SAIL_501_F01) mutants. adf3-1 and adf3-2 contain T-DNA insertions within the first intron and last exon of ADF3, respectively (Supplemental Fig. S1A), which is associated with reduced accumulation of the ADF3 transcript in these mutant lines compared to the wild type (Supplemental Fig. S1B). The increase in GPA population on the adf3-1 mutant correlated with the significantly higher fecundity of GPA on the mutant compared to the wild type (Fig. 1C). Compared to an average of 1.2 nymphs/day produced by a GPA on the wild-type plant, 1.9 nymphs/day were produced on the adf3-1 mutant.

Figure 1.

Arabidopsis ADF3 gene is required for limiting GPA infestation. A, GPA population size on Arabidopsis wild-type (WT) accession Col-0 and adf1 (a1), adf2 (a2), adf3-1 (a3-1), adf4 (a4), adf5 (a5), and adf9 (a9) mutants (n = 10) 2 d post-infestation. B, GPA population size on WT accession Col-0 and adf3-1 (a3-1) and adf3-2 (a3-2) mutant plants (n = 10) 2 d post-infestation. C, One nymph was released on each WT and a3-1 plant, and the number of progeny nymphs produced by each insect over a 10-d period was determined (n = 6). D, GPA population size on WT accession Col-0, a3-1 mutant, and two independently derived gADF3 plants in which a genomic clone of ADF3 was transformed into the a3-1 mutant (n = 6). Top panel shows RT-PCR analysis of ADF3 and as a control ACT8 expression in uninfested plants of indicated genotypes. E, GPA population size on WT, adf3-1, and a transgenic 35Spro:ADF3 line #1 in which the ADF3 coding sequence was expressed from the heterologous 35S promoter (n = 6). See Supplemental Figure S2 for results from an independently derived 35Spro:ADF3 line. In A, B, D, and E, twenty adult apterous aphids were released on each plant and the population size (adults + nymphs) on each plant determined 2 d later. All values are mean ± se for each genotype. In A and B, asterisks above bars denote values that are significantly different from the WT (P < 0.05; Dunnett test). In C, an asterisk denotes significant difference from the WT (P < 0.05; t test). In D and E, different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

To confirm that loss of ADF3 function is indeed responsible for the better performance of the GPA on the adf3 plants compared to the wild-type plants, a genomic clone of ADF3 was transformed into the adf3-1 mutant. As shown in Figure 1D, ADF3 expression was restored in two independently derived gADF3 transgenic lines. In comparison to the adf3-1 mutant, the GPA population size was significantly lower on these gADF3 transgenic lines and comparable to that on the wild-type plants (Fig. 1D). Similarly, resistance was restored in adf3-1 plants expressing ADF3 from the heterologous Cauliflower mosaic virus 35S gene promoter. The GPA population size on two independently derived 35Spro:ADF3 (in adf3-1 background) lines was significantly smaller than that on the adf3-1 mutant and comparable to that on the wild-type plant (Fig. 1E; Supplemental Fig. S2). Taken together, these results confirm that ADF3 has a critical role in limiting GPA infestation on Arabidopsis.

Actin Cytoskeleton Destabilizers Restore Resistance to the adf3 Mutant

ADF3 is an actin-binding protein that was recently shown to possess F-actin-severing/depolymerizing activity (Nan et al., 2017). F-actin depolymerization assays conducted by us confirm the ability of ADF3 to depolymerize F-actin (Supplemental Fig. S3). To determine if ADF3-dependent actin reorganization is critical for controlling GPA infestation, the ability of the actin cytoskeleton destabilizers cytochalasin D and latrunculin B to compensate for ADF3 deficiency in the adf3-1 mutant was evaluated. Two adult aphids were caged onto cytochalasin D- and latrunculin B-treated, and as control on dimethyl sulfoxide (DMSO)-treated leaves of adf3-1 and wild-type plants, and the number of nymphs born per adult aphid monitored 2 d later. As expected, a significantly larger number of nymphs were born on the DMSO-treated adf3-1 compared to the wild-type plant (Fig. 2A). However, compared to DMSO treatment, significantly fewer nymphs were born on adf3-1 leaves that were treated with cytochalasin D and latrunculin B. The number of nymphs born on the cytochalasin D- and latrunculin B-treated adf3-1 leaves was comparable to that observed on the cytochalasin D- and latrunculin B-treated leaves of wild-type plants, thus indicating that these actin destabilizers compensate for the lack of ADF3 function in the adf3-1 mutant. Taken together, these results confirm the importance of ADF3’s actin-depolymerizing function in Arabidopsis defense against the GPA.

Figure 2.

Actin cytoskeleton destabilizers restore resistance to GPA in the adf3 mutant. A, Effect of the actin destabilizers cytochalasin D (CytD, left) and latrunculin B (LatB, right) on insect numbers. Number of nymphs on leaves of wild type (WT) and adf3-1 (a3-1) mutant treated with 2 μm CytD or LatB and as control with DMSO (0.5%), which was used as a solvent for CytD and LatB. The number of nymphs per adult was determined 2 d after release of two adult aphids on each leaf (n = 12). B, Complementation of adf3-1-conferred susceptibility to GPA by constitutive expression of ADF4 from the 35S promoter. Top, RT-PCR analysis of ADF4 expression from the 35S promoter in three independent a3-1 35SPpro:ADF4 lines and as control in WT and a3-1 plants. EF1α expression provided the control for RT-PCR. Bottom, no choice assay comparing aphid numbers 2 d post-release of 20 adult aphids per leaf on the WT, a3-1, and three independently derived a3-1 35Spro:ADF4 transgenic lines. Values are mean ± se of a minimum of six replicates for each genotype. Different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

To further test the involvement of ADF3’s actin-depolymerizing function in controlling GPA infestation, we tested if the adf3-1 defect could be compensated by overexpression of another actin depolymerizing factor, ADF4 (Henty et al., 2011; Henty-Ridilla et al., 2014; Nan et al., 2017), which exhibits 83% identity and 91% similarity to ADF3 (Supplemental Fig. S4). A previously described 35Spro:ADF4 construct (Henty-Ridilla et al., 2014), in which the ADF4 protein coding sequence is expressed from the 35S promoter, was transformed into the adf3-1 mutant. No choice assays were conducted on three independently derived adf3-1 35Spro:ADF4 lines and as a control on the wild type and adf3-1 mutant. As shown in Figure 2B, GPA population size was comparable on the wild type and the adf3-1 35Spro:ADF4 plants, and significantly lower than that on the adf3-1 mutant, thus confirming the ability of ADF4 overexpression to overcome the adf3-1 defect in limiting GPA infestation. Collectively, the above results in conjunction with the studies of Nan et al. (2017) lead us to conclude that the impact of ADF3 on controlling GPA population size is linked to ADF3’s function in actin cytoskeleton reorganization.

ADF3 Expression in the Phloem Is Required for Controlling GPA Infestation

ADF3 expression was reported to be the strongest among the ADF family of genes in Arabidopsis (Ruzicka et al., 2007). Gene expression data available on the Arabidopsis eFP Browser (http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi) revealed that ADF3 is expressed throughout the development of Arabidopsis in most organs, except pollen. Furthermore, biotic stress, including GPA infestation did not have a pronounced effect on ADF3 expression (http://bbc.botany.utoronto.ca/efp/cgi-bin/efpWeb.cgi). Real-time RT-PCR analysis confirmed that ADF3 expression was not significantly different between uninfested and GPA-infested plants (Fig. 3A). Histochemical analysis for GUS activity in leaves of transgenic ADF3pro:UidA plants in which the bacterial UidA-encoded GUS reporter was expressed from the ADF3 promoter further confirmed that ADF3 promoter activity was comparable between uninfested and GPA-infested leaves (Supplemental Fig. S5). ADF3 promoter activity was strong in the lateral and minor veins compared to the nonvascular tissues (Fig. 3B; Supplemental Fig. S5). Within the vasculature, GUS activity was found to be strong in the phloem (Fig. 3C).

Figure 3.

ADF3 function in the phloem is required for controlling GPA infestation. A, Real-time RT-PCR analysis of ADF3 expression relative to that of EF1α in uninfested (−GPA) and GPA-infested (+GPA) leaves of the Arabidopsis accession Col-0 at the indicated h post-infestation (hpi). Values are mean ± se (n = 3). B, Histochemical staining for GUS activity (blue color) in a GPA-infested leaf of a ADF3pro:UidA plant in which UidA expression is driven from the ADF3 promoter. Right, close-up showing strong GUS activity in the veins. C, Histochemical staining in a section through a vein showing strong GUS activity in the phloem (P). X, xylem. D, ADF3 expression in the phloem is sufficient for controlling GPA infestation. Top, RT-PCR analysis of ADF3 expression from the SUC2 promoter in three independent a3-1 SUC2Pro:ADF3 lines and as control in wild-type (WT) accession Col-0, a3-1, and a3-1 35Spro:ADF3 plants. Primers specific for the SUC2 promoter-driven ADF3 construct were used. EF1α expression provided the control for RT-PCR. Bottom, GPA population size on Arabidopsis plants of indicated genotypes. Twenty adult apterous aphids were released on each plant and the population size (adults + nymphs) on each plant determined 2 d later. Values are mean ± se of a minimum of six replicates for each genotype. Different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

To test if ADF3 expression in phloem is important for ADF3’s role in controlling GPA infestation, we tested if the ADF3 coding sequence expressed from the phloem-specific SUC2 promoter (Gottwald et al., 2000) was sufficient to restore resistance against the GPA in the adf3-1 mutant background. Transgenic adf3-1 plants expressing the ADF3 coding sequence from the 35S promoter provided the positive controls for this experiment. As shown in Figure 3D, the SUC2pro:ADF3 chimera complemented the adf3-1 defect. GPA population size was significantly lower on three independently derived adf3-1 SUC2pro:ADF3 lines compared to the adf3-1 mutant. The level of resistance observed in the adf3-1 SUC2pro:ADF3 lines was comparable to that observed in adf3-1 35Spro:ADF3 line. There, results confirm an important role for ADF3 in the phloem in controlling GPA infestation on Arabidopsis.

ADF3 Is Required for Limiting GPA Feeding from Sieve Elements

The adverse effect of ADF3 on insect fecundity could result from its effect on the accumulation of antibiosis activity or alternatively due to its impact on the insects feeding behavior. Previous studies have indicated the presence of an antibiotic activity, which is detrimental to GPA fecundity, in leaf petiole exudates that are enriched in phloem sap (Louis et al., 2010a; Nalam et al., 2012). To determine if ADF3 controls the accumulation of this antibiotic activity, petiole exudates collected from leaves of wild type, and the adf3-1 and adf3-2 mutants were added to a synthetic diet and GPAs were reared on the supplemented diet for 5 d. As shown in Figure 4, in comparison to the GPA population on the synthetic diet lacking petiole exudates, insect population size was significantly smaller (P < 0.05) on diet supplemented with petiole exudates from wild-type plants. The GPA population size was similarly lower on diet supplemented with petiole exudates collected from the adf3-1 and adf3-2 mutant plants, thus confirming that ADF3 does not significantly influence the accumulation of this antibiotic activity in the phloem sap-enriched petiole exudates.

Figure 4.

ADF3 is not required for accumulation of antibiosis activity in petiole exudates. Artificial diet assay showing a comparison of GPA numbers on diet containing petiole exudates (Pet-ex) collected from leaves of wild-type (WT), adf3-1 (a3-1), and adf3-2 (a3-2) plants. Control was an artificial diet supplemented with the buffer used to collect petiole exudates. Values are mean ± se of a minimum of seven replicates for each treatment. Different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

To test if ADF3 adversely influences GPA feeding, GPA feeding behavior was monitored on leaves of the wild-type and adf3-1 mutant plants with the electrical penetration graph (EPG) technique in which the plant and the insect, with a wire glued to its dorsum, are part of a low-voltage circuit (van Helden and Tjallingii, 2000). The different waveform patterns in an EPG are characteristic of the different feeding behavioral activities of the insect. EPG provides information on the time the insect takes to first probe (FP) the plant with its stylets, the time spent by the insect to find and tap into a sieve element for the first time (f-SEP; first sieve element phase), and the sum of time spent during the recording period by the insect in all the SEPs (s-SEP), in the xylem phase (XP) when the stylet is in the xylem and the insect is consuming xylem sap, the pathway phase (PP) when the stylet is inserted into the leaf tissue but is outside the sieve elements and likely sampling other cells, and the nonprobing (NP) phase when the stylet is not inserted into the plant tissue. As shown in Figure 5A, comparison of these behavioral activities of the GPA on the wild type and the adf3-1 mutant revealed that the insect spent significantly less time (P < 0.05) attaining the f-SEP on the adf3-1 mutant. In addition, the s-SEP was significantly longer (P < 0.05) when the insects were on the adf3-1 mutant compared to the wild type, thus indicating that the insects spent significantly more time feeding from the sieve elements of the adf3-1 mutant than the wild-type plant. A corresponding reduction in the time spent by GPA in PP was observed on the adf3-1 mutant compared to the wild-type plant, thus confirming that the insect encounters fewer obstacles that allow it to find sieve elements faster and feed for longer periods on the adf3-1 mutant. In contrast to f-SEP, s-SEP, and PP, the insect took comparable amount of time to reach the FP and spent comparable time in the NP and XP on the wild type and adf3-1 mutant. These results suggest that ADF3 and/or an ADF3-dependent mechanism(s) interferes with the ability of the GPA to find and feed from the sieve elements.

Figure 5.

ADF3 is required for controlling GPA feeding on Arabidopsis leaves. A, Electrical penetration graph (EPG) comparison of time spent by GPA in various activities on the wild-type (WT) and adf3-1 (a3-1) mutant plants in an 8-h period. FP, time taken for first probe; f-SEP, time taken to reach first sieve element phase; PP, time spent in pathway phase during the 8-h period; NP, total time spent nonprobing during the 8-h period; s-SEP, total time spent in sieve element phase during the 8-h period; XP, total time spent in the xylem phase during the 8-h period. B, EPG comparison of time spent by GPA in the E1 (salivation) and E2 (ingestion) phases during the first SEP. In A and B, values are the mean ± se (n = 10). Asterisks indicate significant differences (P < 0.05; Kruskal-Wallis test) in an individual feeding parameter between insect feeding on the a3-1 compared to the WT plant.

During the SEP, the watery saliva injected by the insect into the sieve elements may inhibit phloem-sealing mechanisms (Powell et al., 2006; Tjallingii, 2006; Will and van Bel, 2006). This watery saliva injection phase, the E1 phase, is followed by the E2 phloem sap-ingestion phase. A prolonged E1 and/or reduced E2 is suggestive of a lowered ability of the insect to suppress rapid phloem-sealing mechanisms (Garzo et al., 2002). To determine if insects feeding from sieve elements of the adf3 mutant encounter less resistance that results in a shorter duration of watery saliva release into the sieve elements, the length of the E1 waveform in the first SEP was compared between insects released on the wild type and the adf3-1 mutant. As shown in Figure 5B, the length of the E1 phase was comparable between insects placed on the wild type and adf3-1 mutant, thus indicating that ADF3 does not influence the length of the period of watery saliva release by GPA into Arabidopsis sieve elements. However, the E2 phase was significantly shorter (P < 0.05) on the wild type compared to the adf3-1, thus confirming that ADF3 and/or an ADF3-dependent mechanism in the wild-type plant limits ingestion of phloem sap by the GPA.

ADF3 Controls Callose Deposition in GPA-Infested Plants

Callose deposition is one of the processes involved in phloem occlusion (Will and van Bel, 2006; Kempema et al., 2007; Hao et al., 2008). Similarly, callose deposited outside cells that are in the path of the stylets trying to reach the sieve element, could interfere with the ability of the stylets to reach sieve elements. Considering that ADF3 is required for limiting GPA feeding, and callose synthesis is subject to the participation of the actin cytoskeleton (Cai et al., 2011), we tested the impact of ADF3 on callose deposition in GPA-infested plants. As shown in Figure 6, in response to GPA infestation, a significant increase in the number of callose spots was observed in wild-type plants. In comparison to the wild type, the number of callose deposits were lower in the GPA-infested leaves of adf3-1 mutant. The number of callose spots in the GPA-infested adf3-1 mutant was comparable to that observed in the GPA-infested pad4-1 mutant (Fig. 6), which, like adf3-1, is also defective in controlling GPA feeding from the sieve elements (Pegadaraju et al., 2007; Louis et al., 2012), thus further relating callose deposition with the ability of the plant to control GPA feeding.

Figure 6.

ADF3 is required for promoting callose deposition in response to GPA infestation. Top, callose deposits (light blue spots) in leaves of wild-type (WT) plants of the Arabidopsis accession Col-0 and the adf3-1 (a3-1) mutant, 24 hpi with the GPA (+GPA) or as control in uninfested (−GPA) plants. Bar = 200 μM. Bottom, mean number of callose spots ± se per mm² of leaf tissue in GPA-infested and uninfested leaves (n = 5). Different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

An ADF3-Dependent Mechanism Controls PAD4 Expression in GPA-Infested Plants

In the nucleus, actin is a component of the chromatin remodeling complexes that function to regulate gene expression (Jockusch et al., 2006; Farrants, 2008). Furthermore, some ADFs are involved in shuttling actin into the nucleus and thus in regulating gene expression (Nebl et al., 1996; Jiang et al., 1997). We noted that the up-regulation of PAD4 expression, which is observed in the GPA-infested leaves of the wild type, was attenuated in the adf3-1 mutant. As shown in Figure 7A, in comparison to the significant (P < 0.05) increase in PAD4 expression observed in GPA-infested compared to uninfested wild-type plants, in the adf3-1 mutant, PAD4 transcript levels did not exhibit a significant increase in response to GPA infestation. These results suggest that an ADF3-dependent mechanism regulates PAD4 transcript accumulation in GPA-infested wild-type plants.

Figure 7.

ADF3 is required for promoting PAD4 expression in response to GPA infestation. A, qRT-PCR analysis of PAD4 expression relative to EF1a expression at 24 hpi in GPA-infested and uninfested wild-type (WT) and adf3-1 (a3-1) mutant plants. Values are mean ± se (n = 3). Asterisks indicate values that are significantly different from the corresponding mock-treatment (P < 0.05; t test). B, GPA population size on WT, a3-1, a3-1 35Spro:ADF3 line #1, a3-1 35Spro:PAD4 line #1, pad4-1, and pad4-1 3Spro:ADF3 line #1. See Supplemental Figure S6 for data for additional independently derived transgenic lines. The no-choice assay was conducted by releasing twenty adult apterous aphids on each plant and the population size (adults + nymphs) on each plant determined 2 d later. Values are mean ± se (n = 6) for each genotype. Different letters above bars denote values that are significantly different from each other (P < 0.05; Tukey test).

To further define the relationship between ADF3 and PAD4 in Arabidopsis defense against the GPA, adf3-1 35Spro:PAD4 and pad4-1 35Spro:ADF3 plants in which PAD4 and ADF3, respectively, are expressed from the 35S promoter were generated. adf3-1 35Spro:ADF3 plants provided the controls. As expected, ADF3 expression from the 35S promoter restored basal resistance against GPA in the adf3-1 35Spro:ADF3 plants (Fig. 7B; Supplemental Fig. S6). By comparison, in two independently derived transgenic lines, ADF3 expression from the 35S promoter was unable to restore resistance in the absence of a functional PAD4 gene in the pad4-1 35Spro ADF3 plants (Fig. 7B; Supplemental Fig. S6), thus suggesting a critical role for PAD4 in ADF3-dependent defense against the GPA. Moreover, in no-choice assays conducted on plants of two independent adf3-1 35Spro PAD4 lines, GPA population size was comparable to that on the wild-type plants but smaller than on the adf3-1 mutant (Fig. 7B; Supplemental Fig. S6), thus indicating that overexpression of PAD4 from the heterologous 35S promoter is sufficient to restore significant levels of resistance in the absence of ADF3 function. Collectively, these results reaffirm the importance of ADF3-dependent regulation of PAD4 expression in Arabidopsis defense against the GPA.

DISCUSSION

The actin cytoskeleton and actin-binding proteins are present in the phloem (Schobert et al., 1998, 2000; Kulikova and Puryaseva, 2002; Lin et al., 2009; Rodriguez-Medina et al., 2011; Fröhlich et al., 2012; Hafke et al., 2013). However, their biological function in the phloem is poorly understood. We provide multiple lines of evidence indicating an important role for ADF3 and actin reorganization in Arabidopsis defense against the GPA, particularly in controlling GPA feeding from the sieve elements: (1) Compared to the wild-type plant, GPA population was larger on adf3 mutant plants. (2) The adf3 deficiency in controlling GPA infestation was compensated by the actin cytoskeleton destabilizers cytochalasin D and latrunculin B and by overexpression of actin depolymerizing factor ADF4. (3) ADF3 was expressed in the phloem, and expression of ADF3 from the phloem-specific SUC2 promoter was sufficient to restore resistance against the GPA in the adf3-1 mutant. (4) In agreement with it functioning in the phloem, ADF3 was required for limiting GPA feeding from the sieve elements. Our results further demonstrate that ADF3’s involvement in controlling GPA infestation is in part mediated via its influence on PAD4 expression and hence defense signaling.

EPG analysis indicated that besides limiting GPA feeding from sieve elements, an ADF3-dependent mechanism interferes with the ability of the insect stylets to reach the sieve elements. Thus, we propose that an ADF3-dependent process also impacts events occurring prior to the first penetration of a sieve element by the aphid stylets. On their way to finding a sieve element, the stylets penetrate and sample contents of nonvascular cells (Pollard, 1973; Powell et al., 2006). These punctures of plant cells by the stylets could stimulate plant defenses. Salivary components that are released into the plant tissue are also known to stimulate host defenses (DeVos and Jander, 2009; Bos et al., 2010). For example, callose deposition increases in response to GPA infestation (Elzinga et al., 2014) as well as application of GPA-derived elicitors (Prince et al., 2014). Callose has been implicated in the control of infestation by insects that feed from the phloem (Will and van Bel, 2006; Kempema et al., 2007; Hao et al., 2008). Callose deposited in the sieve elements interferes with the ability to the aphid to feed from the sieve elements. In addition, callose deposited outside the sieve elements could interfere with the ability of the stylet to reach sieve elements to further limit insect feeding. The GPA infestation-associated increase in callose deposition was attenuated the adf3-1 mutant, suggesting an important role for ADF3 in this process. The actin cytoskeleton is involved in the localization of callose synthases to the cell membrane, presumably due to the involvement of the actin cytoskeleton in vesicular trafficking (Cai et al., 2011). Whether ADF3 is similarly involved in the localization of callose synthase is not known. However, since the GPA infestation-associated increase in callose deposition was also attenuated in the pad4 mutant, we suggest that the impact of ADF3 on callose deposition in response to GPA infestation is likely exerted via ADF3’s engagement of the PAD4 defense-signaling pathway. The contribution of PAD4 in promoting callose deposition is further supported by a recent study, which showed that the endophytic Bacillus velezensis-induced resistance against the GPA was accompanied by an increase in callose deposition, which was dependent on PAD4 (Rashid et al., 2017).

Actin is a component of chromatin remodeling activities that control gene expression in the nucleus (Jockusch et al., 2006; Farrants, 2008), and ADF’s are involved in shuttling actin into the nucleus (Nebl et al., 1996; Jiang et al., 1997) and in regulating gene expression (Burgos-Rivera et al., 2008; Porter et al., 2012). Among the subclass I ADFs in Arabidopsis, ADF4’s involvement in effector-triggered immunity is linked to downstream activation of gene expression (Porter et al., 2012). Similarly, our results indicate that ADF3’s involvement in defense against the GPA is linked to the up-regulation of PAD4. However, although PAD4 is required for ADF3-conferrred resistance to the GPA, our results also indicate that ADF3 and PAD4 have additional functions in Arabidopsis defense against the GPA that are independent of each other. For example, unlike PAD4, an ADF3-dependent mechanism hinders with the ability of the insect to find sieve elements. In contrast, unlike ADF3, a PAD4-dependent mechanism is required for the accumulation of an antibiosis activity in the vascular sap (Pegadaraju et al., 2005, 2007; Louis et al., 2010a). Previous studies have shown that while basal expression of PAD4 contributes to the antibiosis activity, the up-regulation of PAD4 expression is associated with the feeding deterrence function of the PAD4-dependent pathway (Pegadaraju et al., 2007; Louis et al., 2010a, 2010b). These conclusions are in agreement with the observations described here that ADF3, which is required for the up-regulation of PAD4 expression in response to GPA infestation, but not for the basal expression of PAD4, is required for controlling GPA feeding, but not for the accumulation of the antibiotic activity. It is likely that different cell types are involved in exerting these effects. For example, the contribution of ADF3 and PAD4 to feeding deterrence is exerted at the level of the sieve elements, while ADF3’s involvement in hindering with the ability of the GPA to find sieve elements is exerted outside the sieve elements. Similarly, PAD4’s contribution to antibiosis is likely exerted in tissues where these antibiotic factors are produced.

ADF3 (At5g59880) and ADF4 (At5g59890), which are located in tandem on chromosome 5, are both subclass I ADFs that share a high level of sequence identity. Both possess actin-severing/depolymerizing activity (Henty-Ridilla et al., 2014; Nan et al., 2017). However, while ADF3, not ADF4, is essential for defense against the GPA, ADF4, but not ADF3, is required for AvrPphB -triggered immunity against the bacterial pathogen P. syringae (Tian et al., 2009). The ability of ADF4 expressed from the strong and ubiquitously expressed 35S promoter to limit GPA population in the adf3-1 mutant background indicates that ADF3 and ADF4 have overlapping biochemical functions. We therefore conclude that the unique biological functions of ADF3 in defense against the GPA is likely determined by differences in the spatial expression pattern and/or overall level of expression of ADF3, compared to ADF4, and likely other subclass I ADF genes.

CONCLUSION

Our results demonstrate that an ADF3-dependent mechanism hinders with the ability of the GPA to find and feed from sieve elements. Considering that ADF3 is required for up-regulating PAD4 expression and callose deposition in GPA-infested leaves, we postulate that an ADF3-dependent mechanism is involved in signaling associated with Arabidopsis defense against the GPA.

MATERIALS AND METHODS

Plant and Insect Materials

All plants were cultivated at 22°C under a 14-hr light (100 μE m² s⁻¹) and 10-hr dark regime on a peat-based soil mix (Fafard #2, fafard.com). The Arabidopsis (Arabidopsis thaliana) adf1 (Salk_144459; Tian et al., 2009), adf2 (RNAi line; Clément et al., 2009), adf3-1 (Salk_139265C; Tian et al., 2009), adf3-2 (SAIL_501_F01), adf4 (Garlic_823_A11.b.1b.Lb3Fa; Tian et al., 2009), adf5 (Salk_030145C), and adf9 (Salk_056064; Tian et al., 2009) lines were used in this study. Silencing of ADF2 in the ADF2 RNAi line was induced by treating plants with 0.5% ethanol solution (Clément et al., 2009).

A GPA (Kansas State University, Museum of Entomological and Prairie Arthropod Research, voucher specimen 194) colony was reared at 22°C under a 14-hr light (100 μE m² s⁻¹) and10-hr dark regime on a 1:1 mixture of commercially available radish (Raphanus sativus; Early Scarlet Globe) and mustard (Brassica juncea; FL Broadleaf).

No-Choice Tests

No-choice assays were conducted as previously described (Pegadaraju et al., 2005; Louis et al., 2010a). Unless stated otherwise, 20 adult asexually reproducing apterous (wingless) aphids were placed on each 4-week-old plant. Two days later, the number of insects on each plant was counted. To monitor aphid fecundity, one young nymph (3–4 d old) was placed on each plant, and the total number of progeny born over a 10-d period was recorded.

Monitoring Aphid Feeding Behavior

The EPG technique (van Helden and Tjallingii, 2000; Walker, 2000) was used to simultaneously monitor the feeding behavior of GPA on 4-week-old plants of two different genotypes, as previously described (Pegadaraju et al., 2007). An eight-channel GIGA-8 direct current amplifier (http://www.epgsystems.eu/systems.htm) was used for simultaneous recordings of eight individual aphids. The waveform recordings were analyzed using the EPG analysis software PROBE 3.4 (provided by W.F. Tjallingii, Wageningen University, the Netherlands). The analysis of the first sieve element phase was divided into the length of the first salivation E1 and the time of active ingestion phase E2.

Petiole Exudate Collection and Artificial Diet Feeding Assays

Phloem sap-enriched petiole exudates were collected from Arabidopsis leaves as previously described (Chaturvedi et al., 2008). After concentration, equal volumes of petiole exudates were added to a synthetic diet (Mittler and Dadd, 1965) contained in a feeding chamber (Louis et al., 2010b). Three adult aphids were released on each feeding chamber, and the total number of GPA in each feeding chamber was determined 5 d later.

Actin Cytoskeleton Destabilizer Treatment

Two leaves of each 4-week-old plant were infiltrated with 2 μm of cytochalasin D or latruculin B solubilized in 0.5% dimethyl sulfoxide (DMSO), followed by release of two adult GPA on each leaf. Each leaf was caged in a perforated 2-mL microfuge tube to prevent escape of insects. Total number of nymphs on each leaf were determined 2 d later. Plants treated with 0.5% DMSO provided the negative control. A minimum of twelve leaves of each genotype were used for each treatment.

PCR Analysis for Mutant Screening

All primers used in this study to confirm the knockdown of individual ADF genes in the adf mutants and the ADF2 RNAi knockdown lines are listed in Supplemental Table S1. PCR with gene-specific primers was performed under the following conditions: 94°C for 5 min, followed by 36 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min, with a final extension of 72°C for 5 min. PCR for genotyping the adf mutants utilized the T-DNA left border primer and a gene-specific primer. The PCR conditions included a 5 min denaturation at 94°C, followed by 36 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 1 min, with a final extension of 72°C for 5 min.

Gene Expression

RNA extraction from leaves was performed as previously described (Pegadaraju et al., 2005). Each sample included a minimum of two leaves. A minimum of three samples was analyzed for each treatment. For real-time RT-PCR, gene expression levels were normalized to that of EF1α, while for RT-PCR ACT8 was used as a control. Gene-specific primers used for real-time RT-PCR and RT-PCR are listed in Supplemental Table S1.

Expression and Purification of Recombinant ADF3 Proteins

The primers 5′-CGCCCCATATGGCTAATGCAGCATCAGGAATGGCAGTCC-3′ and 5′-AAGCTTCTCGAGTCAATTGGCTCGGCTTTTGAAAAC-3′ were used to amplify the ADF3 coding sequence, using cDNA prepared from mRNA harvested from Arabidopsis Col-0 as the template. The resultant product was digested with NdeI and XhoI, which cut within the two primer regions. The digested product was ligated between the NdeI and XhoI sites of pET28a vector. The resultant pET28a-ADF3 plasmid in which a 6X-His tag was incorporated at the N-terminal end of ADF3 was transformed into Escherichia coli strain BL21-DE3. Expression of the recombinant ADF3 protein was induced by the addition of Isopropyl-β-d-thiogalactopyranoside (IPTG) prior to bacterial cell harvest. The recombinant, 6X-His-tagged protein was purified over an affinity Ni-NTA column (QIAexpress, Qiagen, http://www.qiagen.com). The purified ADF3 protein was centrifuged at 20,000g for 30 min at 4°C before use in the F-actin cosedimentation assays. The GST-ADF3 and GST-ADF4 clones used in the actin depolymerization assays encode recombinant proteins that contain GST fused to the N-terminal end of ADF3 and ADF4. These were a gift of Brad Day. The recombinant proteins were purified over a GST-affinity column before use in actin depolymerization assays.

Transgenic ADF3- and ADF4-Expressing Plants

The ADF3 genomic fragment (gADF3) was amplified using the primers 5′-TAAATGAATTTTTTTTACGGGA-3′ and 5′-TCAATTGGCTCGGCTTTTGA-3′, and the resultant product cloned into the pCR8/GW/TOPO vector (Life Technologies; www.lifetechnologies.com). Gateway LR clonase (Life Technologies; www.lifetechnologies.com) was used to mobilize the cloned fragment into the binary vector pMDC107 (Curtis and Grossniklaus, 2003) to yield pMDC107-gADF3. To generate a construct for in planta expression of the ADF3 coding sequence from the Cauliflower mosaic virus 35S gene promoter, the pET28a-ADF3 plasmid (see above) was used as a template with the primer 5′-ATGGGCAGCAGCCATCATC-3′, which is derived from the pET28a vector and primer 5′-TCAATTGGCTCGGCTTTTGA-3′ to amplify the ADF3 coding region. The resultant PCR product was cloned into pCR8/GW/TOPO, from where it was mobilized into the destination vector pMDC32 (Curtis and Grossniklaus, 2003) to yield the pMDC32-35Spro:ADF3 plasmid.

The pMDC107-gADF3 and pMDC32-35Spro:ADF3 plasmids were individually electroporated into Agrobacterium tumefaciens strain GV3101, which was subsequently used for transforming the adf3-1 plants by the floral-dip method (Zhang et al., 2006) to generate transgenic plants in which the adf3-1 mutant phenotype was complemented by expression of the gADF3 and 35Spro:ADF3. Transgenic plants were selected on 1/2 strength Murashige and Skoog agar plates containing hygromycin (25 μg mL⁻¹). The 35Spro:PAD4 construct has been previously described (Xing and Chen, 2006).

The 35Spro:ADF4 construct (Henty-Ridilla et al., 2014), in which the ADF4 coding sequence containing an N-terminal T7 epitope tag is expressed from the 35S promoter was used for generating the adf3-1 35Spro:ADF4 plants by transforming adf3-1 plants with the recombinant construct. Transformed plants were screened based on their resistance to kanamycin (50 μg mL⁻¹), and the presence of the recombinant construct and its expression confirmed by PCR and RT-PCR, respectively. To verify the construct in the adf3-1 35S:ADF4 transgenic lines, a forward primer 5′-GGTGGTCAACAAATGGGT- 3′, which is specific to the region containing the T7 tag, and a reverse primer 5′-TTAGTTGACGCGGCTTTTC- 3′, which is specific to the ADF4 coding sequence, were utilized. The same primers were also utilized to monitor expression of the recombinant constructed by RT-PCR. The PCR conditions were as follows: 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s, followed by a final extension step of 72°C for 5 min.

To generate a SUC2pro:ADF3 construct, the ADF3 coding sequence with a 6X-His tag at the N terminus was amplified from plants containing the gADF3 construct, which contains a His tag on the 5′ end. The primers 5′-ATGGGCAGCAGCCATCATC-3′ and 5′-TCAATTG GCTCGGCTTTTGA-3′ were used to amplify the ADF3 coding sequence, and the resultant PCR product was cloned into pCR8/GW/TOPO. The pCR8 product was then mobilized into a pMDC85 destination vector (Curtis and Grossniklaus, 2003), which had been modified by the removal of the 2× 35S promoter and replaced by the SUC2 promoter (Elzinga et al., 2014). The resultant construct, pMDC85-SUC2pro:ADF3, was transformed into adf3-1 using A. tumefaciens via the floral-dip method (Zhang et al., 2006). Transformed plants were screened based on resistance to hygromycin (20 μg mL⁻¹) and presence of the recombinant construct and its expression confirmed by PCR and RT-PCR, respectively. A forward primer 5′-AGCCATCATCATCATCATCAC-3′ designed to the 6X-HIS tag at the 5′ end of the ADF3 coding sequence and a reverse primer 5′-TCAATTG GCTCGGCTTTTG-3′ that is specific to the ADF3 coding sequence were utilized in PCR reactions to confirm the presence of the SUC2pro:ADF3 construct. These same primers were also utilized to monitor expression of the recombinant construct by RT-PCR. The PCR conditions were as follows: 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 57°C for 30 s, and 72°C for 45 s, followed by a final extension step of 72°C for 5 min.

GUS Reporter Constructs and Histochemical Analysis

A 2,035-bp DNA fragment upstream of the transcriptional start site of ADF3 was amplified with Platinum Taq polymerase (Invitrogen) using Arabidopsis Col-0 genomic DNA as template and the primers 5′-TAAATGAATTTTTTTTACGGGA-3′ and 5′-GGTTGAATCAAAGCTAGTCTCA-3′. The PCR product was cloned into pCR8/GW/TOPO, from where it was mobilized into the pMDC132 vector (Curtis and Grossniklaus, 2003), which contains the coding sequence of the bacterial UidA gene, which encodes the GUS protein (Jefferson et al., 1987). Transgenic plants containing the ADF3pro:UidA construct were generated in the accession Col-0 background. Histochemical analysis of GUS activity was performed using X-gluc (5-bromo-4-chloro-3-indolyl-β-d-glucopyranosiduronic acid) as the substrate.

Callose Staining

Callose staining and quantification of callose deposits was conducted using a modification of a previously described protocol (Ton and Mauch-Mani, 2004). In brief, leaf samples were cleared overnight in 96% ethanol, followed by a 30-min incubation in sodium phosphate buffer (0.07 m, pH 9). Leaves were then incubated for 60 min in a solution containing 0.005% aniline blue in sodium phosphate buffer (0.07 m, pH 9), followed by the addition of calcofluor to a final concentration of 0.005%. The tissues were immediately observed under a Nikon e600 epifluorescence microscope equipped with a X-cite 120 Fluor System, UV (DAPI) filter and a digital SPOT color camera. Digital images were used to quantify callose spots, which were expressed as number of callose spots per mm² of leaf tissue.

Actin Depolymerization Assays

Rabbit skeletal muscle G-actin was prepared by the method of Spudich and Watt (1971) and flash frozen in liquid nitrogen in 0.5 mL aliquots for storage at −70°C until use. After rapid thawing, the G-actin was chromatographed on a superdex 75 gel filtration column immediately prior to measurements. Concentrations of G-actin were determined using an extinction coefficient of 0.63 mg/mL/cm at 290 nm. A portion of the actin was labeled with pyrene iodoacetamide by the method of Cooper et al. (1983) and stored by the same method as the unlabeled G-actin.

Measurements of pyrene-labeled actin were performed with an Aminco-Bowman II luminescence spectrometer using methods similar to those previously described (Xu and Root 2000). In brief, 15% pyrene-labeled actin was polymerized by the addition of 0.1 m KCl and 2 mm MgCl₂ yielding a 20-fold enhancement of pyrene fluorescence intensity. Actin concentrations were tested in the ranges of 3 to 7.1 μm for separate experiments. The polymerized actin was titrated at 25°C with a range of concentrations of recombinant GST-tagged ADF3 and ADF4 or buffer as a control, and changes in fluorescence intensity measured. Control experiments indicated a reproducibility of ±10% error or better.

Statistical Analysis

When comparing two treatments or genotypes, two-tail t test was used to determine if the mean values were significantly different from each other (P < 0.05). When simultaneously comparing multiple genotypes and/or treatments to each other, ANOVA performed following the General Linear model GLM was used followed by Tukey’s multiple comparison test to identify mean values that were significantly different from each other (P < 0.05; Minitab v15; www.minitab.com). When comparing multiple experimental groups to a single control group, Dunnett’s multiple comparison test was used to compare means to identify values that were significantly different (P < 0.05) from the control group. All data conform to the assumptions of ANOVA, and no transformations were necessary. For the EPG analysis, the nonparametric Kruskal-Wallis test (Minitab v15; www.minitab.com) was used to analyze the mean time spent by aphids on various activities.

Accession Numbers

ADF1 (At3g46010); ADF2 (At3g46000); ADF3 (At5g59880); ADF4 (At5g59890); ADF5 (At2g16700); ADF9 (At4g34970); ACT8 (At1g49240); EF1α (At5g60390); PAD4 (At3g52430).

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Arabidopsis lines carrying T-DNA insertions at the ADF3 locus.

• Supplemental Figure S2. ADF3 expression from the 35S promoter restores resistance against the GPA in the adf3-1 mutant.

• Supplemental Figure S3. ADF3 exhibits actin-depolymerizing activity.

• Supplemental Figure S4. Alignment of ADF3 and ADF4

• Supplemental Figure S5. Histochemical staining for GUS activity in uninfested (−GPA) and GPA-infested (+GPA) leaves of a ADF3pro:UidA plant.

• Supplemental Figure S6. Expression of PAD4 from the constitutively expressed 35S promoter restores resistance against GPA in plants lacking ADF3 function.

• Supplemental Table S1. Primers used for monitoring gene expression.