Effects of the cucumber mosaic virus 2a protein on aphid-plant interactions in Arabidopsis thaliana.

The cucumber mosaic virus (CMV) 2a RNA-dependent RNA polymerase protein has an additional function in Arabidopsis thaliana, which is to stimulate feeding deterrence (antixenosis) against aphids. Antixenosis is thought to increase the probability that aphids, after acquiring CMV particles from brief probes of an infected plant's epidermal cells, will be discouraged from settling and instead will spread inoculum to neighbouring plants. The amino acid sequences of 2a proteins encoded by a CMV strain that induces antixenosis in A. thaliana (Fny-CMV) and one that does not (LS-CMV) were compared to identify residues that might determine the triggering of antixenosis. These data were used to design reassortant viruses comprising Fny-CMV RNAs 1 and 3, and recombinant CMV RNA 2 molecules encoding chimeric 2a proteins containing sequences derived from LS-CMV and Fny-CMV. Antixenosis induction was detected by measuring the mean relative growth rate and fecundity of aphids (Myzus persicae) confined on infected and on mock-inoculated plants. An amino acid sequence determining antixenosis induction by CMV was found to reside between 2a protein residues 200 and 300. Subsequent mutant analysis delineated this to residue 237. We conjecture that the Fny-CMV 2a protein valine-237 plays some role in 2a protein-induced antixenosis.

Cucumber mosaic virus (CMV) is an insect‐transmitted virus that modifies interactions between its infected host plants and its aphid vectors in ways that increase the probability of transmission over various ranges and timescales (Donnelly et al., 2019; Carr et al., 2020). CMV does not infect its aphid vectors but influences their behaviour by altering the biochemistry of infected host plants. The paradigmatic example of this phenomenon is the interaction of the aphids Aphis gossypii and Myzus persicae with Cucurbita pepo plants infected with the Fny strain of CMV (Fny‐CMV) (Mauck et al., 2010). CMV infection causes infected cucurbits to emit a mix of volatile organic compounds that attract aphids, but infection also induces accumulation of antixenotic, that is, feeding‐deterrent, compounds in the leaves that ensure that aphids feed for only a brief time before moving to another plant (Mauck et al., 2010; Carmo‐Souza et al., 2014). Because CMV is a nonpersistently transmitted virus (virus particles acquired from an infected plant are attached loosely to an aphid's stylet mouthparts), this short feed is sufficient to render the aphids competent to transmit infection to neighbouring hosts (Krenz et al., 2015). A similar phenomenon has been observed in the interactions between M. persicae and plants of the model species Arabidopsis thaliana infected with Fny‐CMV (Westwood et al., 2013). Using this system, it was found that antixenosis was induced in some fashion by the CMV 2a protein (Westwood et al., 2013). Plants possess a variety of mechanisms that enable them to resist aphid infestation or discourage feeding by these insects (Nalam et al., 2019). In the case of CMV‐infected A. thaliana, Westwood et al. (2013) proposed that the 2a protein induces production of the antixenotic compound 4‐methoxy‐indol‐3‐yl‐methylglucosinolate (Kim and Jander, 2007; Mewis et al., 2012) by activation of the pathogen‐associated molecular pattern‐triggered immunity system.

The primary function of the CMV 2a protein is to act as the viral RNA‐dependent RNA polymerase, which catalyses synthesis of new genomic and subgenomic RNA molecules (Palukaitis and García‐Arenal, 2003; Seo et al., 2019). The 2a protein is one of five proteins encoded by the tripartite, positive‐sense RNA genome of CMV. RNA 1 is translated directly to yield the 1a methyltransferase/helicase protein, which associates with the 2a protein during replicase complex formation (Palukaitis and García‐Arenal, 2003; Seo et al., 2019). The 97 kDa 2a protein is translated from the 5′‐proximal open reading frame (ORF) of RNA 2. An overlapping ORF encodes the 2b counterdefence protein (Figure 1a), which is expressed from a viral subgenomic mRNA (RNA 4A) (Palukaitis and García‐Arenal, 2003). CMV RNA 3 acts as a translation template for the viral moment protein and also encodes the viral coat protein, which is expressed by translation of subgenomic RNA 4 (Palukaitis and García‐Arenal, 2003). The coat protein is the sole viral factor needed for attachment of CMV particles to acrostyle receptors in the aphid stylet (Webster et al., 2018; Fereres and Perry, 2019).

Figure 1

Design of recombinant CMV RNA 2 molecules to identify CMV 2a protein sequences that influence plant–aphid interactions. (a) Amino acid sequence alignment and similarity between the Fny‐CMV and LS‐CMV 2a proteins. The line graph illustrates the degree of similarity obtained using plotcon (EMBOSS Explorer: http://emboss.bioinformatics.nl/cgi‐bin/emboss/plotcon) with a window size of 10 residues. A diagram of the 2a protein ORF is displayed above the similarity plot showing the RNA‐dependent RNA polymerase (RdRp) domain and the overlapping 2b ORF. (b) Five recombinant CMV RNA 2 molecules were produced by ligating together segments of the LS‐CMV and Fny‐CMV RNA 2 molecules to produce in‐frame 2a ORFs encoding chimeric 2a proteins (Methods S1). Sequences derived from the Fny‐CMV 2a protein ORF are depicted with a red outline, and construct names are based on the sequence coordinates (in subscript text) for residues derived from each strain's 2a protein, for example the 2a ORF of F200‐300L encodes the Fny‐CMV 2a amino acid sequence between residues 200 and 300, with the remainder derived from the LS‐CMV 2a protein sequence. (c) The amino acid sequences for the LS‐CMV and Fny‐CMV 2a proteins between residues 201 and 300 are shown. Shading indicates amino acid sequence conservation between the two 2a protein sequences. Site‐directed mutagenesis was used to substitute LS‐CMV RNA 2 sequences with those from Fny‐CMV RNA 2 at 13 sites and the resulting changes in amino acid sequence are indicated in red

In a previous study of the effects of the CMV 2a protein on interactions of A. thaliana Col‐0 and M. persicae (isolate USL1: Devonshire and Sawicki, 1979), the mean relative growth rate (MRGR) of aphid nymphs confined on plants had been used as the sole proxy for aphid performance (Westwood et al., 2013). However, as shown with tobacco, virus‐induced changes in MRGR do not always correlate with decreased reproduction (Ziebell et al., 2011). Nevertheless, on 2a‐transgenic A. thaliana plants the MRGR and colony growth of M. persicae (mean number of offspring produced per aphid) were both impeded, showing that the Fny‐CMV 2a protein induces effects in this plant that decrease both measures of performance (Figure S1 and Spreadsheet S1). Therefore, in this study we used both assays to control for the possibility that different sequences within the CMV 2a protein influence different aspects of CMV‐induced antixenosis.

In this study we sought to identify amino acid residue(s) in the CMV 2a protein involved in induction of antixenosis in CMV‐infected A. thaliana plants. Our approach took advantage of our finding that although the Fny strain of CMV induces antixenosis in A. thaliana, the strain LS‐CMV does not (Westwood et al., 2013). Comparison of the amino acid sequences of the 2a proteins encoded by each strain enabled us to begin delineating which residue or residues might determine antixenosis induction (Figure 1a). The amino acid sequences of the RNA‐dependent RNA polymerase domains showed the highest degrees of conservation (Figure 1a). The greatest dissimilarity between the two 2a proteins occurs in the N‐proximal 300 residues, and in the C‐terminal regions of the 2a proteins, which correspond to the region of the 2a ORF that overlaps with part of the ORF encoding the 2b protein (Figure 1a). Although the region of the 2a ORF that overlaps with the 2b ORF has a number of effects on the pathology and movement of CMV (Du et al., 2008; Khaing et al., 2020), it is not required for antixenosis induction in A. thaliana (Westwood et al., 2013). Therefore, we hypothesized that the region spanning residues 1–300 of the 2a protein is the most likely to contain amino acid(s) that determine antixenosis induction. Five recombinant cDNA clones encoding chimeric RNA 2 molecules were constructed in which the regions encoding all or part of the N‐proximal 300 residues of the 2a protein comprised sequences exchanged between the RNA 2 sequences of Fny‐CMV and LS‐CMV (Figure 1b). Constructs were derived from plasmids pFny206 and pLS‐CMV2, the respective infectious cDNA clones for the Fny‐CMV and LS‐CMV RNA 2 molecules (Rizzo and Palukaitis, 1990; Zhang et al., 1994) (Table S1). Wild‐type or recombinant RNA 2 molecules were synthesized by in vitro transcription using T7 RNA polymerase, and infectious RNA mixtures produced by mixing these with in vitro‐synthesized Fny‐CMV RNAs 1 and 3. Infectious RNA mixtures for these reassortant and recombinant viruses were used to inoculate Nicotiana benthamiana plants for preparation of virions to use as inocula for experiments with A. thaliana (Palukaitis, 2019). We recently used a similar approach with Fny‐CMV/LS‐CMV reassortants to successfully identify the viral RNA conditioning antibiosis (strong resistance) against aphids in Nicotiana tabacum (Tungadi et al., 2020). RNA was isolated from systemically infected leaves, subjected to reverse transcription‐PCR (RT‐PCR) to amplify RNA 2‐specific sequences and these amplicons sequenced to confirm that all recombinant RNA 2 molecules were genetically stable and did not undergo further mutation during replication and movement through the plant. Plants of A. thaliana and N. benthamiana infected with viruses containing recombinant RNA 2 molecules exhibited easily discernable systemic disease symptoms that were less severe than those induced by reconstituted Fny‐CMV (F1‐F2‐F3: generated by mixing Fny‐CMV RNAs 1, 2, and 3), and similar in severity to symptoms induced by the reassortant virus F1‐L2‐F3 (constituted from Fny‐CMVs RNAs 1 and 3, and LS‐CMV RNA 2) (Figure S2 and Spreadsheet S2), and accumulated to similar levels in A. thaliana (Figure S3a).

A. thaliana plants were used for aphid performance experiments at 10 days following either inoculation with virions or mock inoculation with water. The induction of resistance to M. persicae confined on plants was assessed using two measures of aphid performance: MRGR of aphid nymphs over the first 6 days of life and subsequent reproduction. Both assays were carried out using the same batches of aphids and plants, and carried out on three separate occasions (Figure 2). MRGR was decreased for aphid nymphs placed on plants infected with F1‐F2‐F3, but not with the reassortant virus F1‐L2‐F3 (Figure 2). This is consistent with the conclusion of Westwood et al. (2013) that the 2a protein induces feeding deterrence, resulting in aphid growth inhibition. In this study it was found that aphid reproduction was also inhibited on plants infected with F1‐F2‐F3, but not on plants infected with the F1‐L2‐F3 reassortant virus (Figure 2).

Figure 2

Performance of Myzus persicae on Arabidopsis thaliana plants infected with Fny‐CMV/LS‐CMV reassortant viruses with Fny‐CMV/LS‐CMV recombinant RNA 2 molecules. One‐day‐old aphid nymphs (10–12 aphids per treatment per experiment) were weighed using a microbalance, placed on plants, reweighed 5 days later, and the mean relative growth rate (MRGR) calculated as previously described (Leather and Dixon, 1984; Stewart et al., 2009; Ziebell et al., 2011; Westwood et al., 2013) (histograms on left). Ten days later the offspring produced by each aphid were counted and the mean colony size calculated (Ziebell et al., 2011; Westwood et al., 2013; Tungadi et al., 2020) (histograms on right). Plants were mock‐inoculated (Mock) with sterile water or inoculated with purified virions of reconstituted Fny‐CMV (F1‐F2‐F3), the reassortant Fny‐CMV/LS‐CMV virus F1‐L2‐F3, or viruses possessing RNAs F1, F3, and one of the indicated Fny‐CMV(F)/LS‐CMV(L) recombinant RNA 2 molecules (described in Figure 1b). Plants were used for aphid performance experiments at 10 days postinoculation. Panels (a), (b), and (c) are the results of three independent experiments. Error bars represent standard error around the mean. Statistical analysis was performed using one‐way analysis of variance followed by Dunnett's post hoc multiple comparisons test (significant differences indicated by ****p < .0001; ***p < .001; **p < .01, and *p < .01) using R (Dalgaard, 2008)

The RNA 2 of F1‐L1‐300F‐F3 possesses a 2a ORF in which residues 1–300 are derived from LS‐CMV RNA 2 (Figure 1b). Neither aphid growth nor fecundity was affected on plants infected with the reassortant/recombinant virus F1‐L1‐300F‐F3 (Figure 2). The recombinant RNA 2 used to constitute this reassortant virus possesses the ORF for the Fny‐CMV 2b protein (Figure 1b). The Fny‐CMV 2b protein can induce a variety of effects on performance of aphids on tobacco plants infected with Fny‐CMV (Ziebell et al., 2011; Tungadi et al., 2020) and in 2b‐transgenic A. thaliana plants (Westwood et al., 2013). However, because infection with F1‐L1‐300F‐F3 did not induce resistance to aphids, it appears that the 2b protein is not conditioning aphid resistance induced in A. thaliana by CMV infection. This is consistent with the conclusion of Westwood et al. (2013) that the 2a protein conditions feeding deterrence during CMV infection. That F1‐L1‐300F‐F3 does not induce resistance to aphids is consistent with our starting hypothesis that the sequence determining aphid resistance induction lies within the region that shows the most dissimilarity between the 2a proteins of Fny‐CMV and LS‐CMV (Figure 1a). Neither growth nor reproduction of aphids was affected on A. thaliana plants infected with the reassortant/recombinant viruses F1‐F1‐200L‐F3 or F1‐F1‐100L‐F3. Conversely, aphid growth and reproduction were decreased on plants infected with F1‐F1‐300L‐F3 and F1‐F200‐300L‐F3 (Figure 2). These results were not only consistent with the hypothesis that the N‐proximal 300 residues of the Fny‐CMV 2a protein determine aphid resistance induction in A. thaliana, but also suggested that residue(s) important in CMV‐induced resistance to aphids lie between positions 200 and 300 in the Fny‐CMV 2a protein sequence.

Comparison of the 2a protein sequences encoded by LS‐CMV and Fny‐CMV revealed 27 differences in the sequence of amino acids lying between residues 200 and 300 (Figure 1c). Using site‐directed mutagenesis, 13 recombinant versions of LS‐CMV RNA 2 were generated in which between one and five codons from the Fny‐CMV 2a ORF sequence were substituted for corresponding codons of the LS‐CMV 2a ORF (Tables 1 and S2, and Figure 1c). RNA for each RNA 2 recombinant was synthesized by in vitro transcription and infectious RNA mixtures constituted by combining with in vitro‐synthesized RNAs 1 and 3 of Fny‐CMV. These mixtures were used to infect N. benthamiana plants and virions were purified from systemically infected leaves for use as inocula for experiments with A. thaliana. Viral RNA from systemically infected leaves was amplified by RT‐PCR with RNA 2‐specific primers and amplicons sequenced to confirm that the introduced mutations were stable in planta.

Table 1

Specific alterations in the primary amino acid sequence of the 2a protein produced by site‐specific mutagenesis of LS‐CMV RNA 2

Amino acid sequence replacement	Specific 2a protein residue(s) replaced
LIL202‐204MIM	202 and 204
RE216‐217KK	216 and 217
KTFSH222‐226RTFAD	222, 225, and 226
I237V	237
L251V	251
HCSYS254‐258KCQYE	254, 256, and 258
TED260‐262CPT	260, 261, and 262
DFV264‐266TIA	264, 265, and 266
VDPS267‐270IPPD	267, 268, and 270
ARP272‐274NRA	272 and 274
VHV280‐282INL	280, 281, and 282
AVCDT285‐289SICDS	285, 286, and 289
V295S	295

The amino acid sequence of the 2a protein of LS‐CMV was modified at 13 sites by replacement with corresponding sequences from the Fny‐CMV 2a protein (Table S2). The positions of these modifications within the 2a protein primary sequence are shown graphically in Figure 1c.

Symptoms induced by several of these mutant viruses differed from those induced by the F1‐L2‐F3 reassortant virus, suggesting that this region of the CMV 2a protein or its corresponding RNA sequence may influence symptomology, especially with respect to the effects of infection on leaf shape (Figure S4). There were no statistically significant differences in accumulation between the viruses in A. thaliana (Figure S3b). Aphid nymphs were placed on systemically infected plants at 10 days postinoculation, and aphid growth and reproduction were measured as already described, and the effects of each recombinant RNA 2 on aphid performance observed in three to five independent experiments (Spreadsheet S3). Only the I237V mutant induced statistically significant decreases in aphid performance (MRGR and colony growth) consistently in three out of three independent experiments (Figure 3 and Spreadsheet S2). The mutant VHV280‐282INL induced a statistically significant decrease in aphid performance in one out of three experiments, so two additional experiments were performed with this mutant to confirm that this was not a biologically relevant result (Spreadsheet S3). No other mutations induced statistically significant decreases in aphid performance.

Figure 3

Effects on performance of Myzus persicae on Arabidopsis thaliana plants of infection with Fny‐CMV/LS‐CMV reassortant viruses with site‐specific mutations in the 2a open reading frame (ORF). The results shown here are for one experiment using a subset of viral mutants. The full range of data, including this experiment (Experiments 2 and 3), for all mutants is available in Spreadsheet S3. One‐day‐old aphid nymphs (10–12 aphids per treatment per experiment) were weighed using a microbalance, placed on plants, reweighed 5 days later, and the mean relative growth rate (MRGR) calculated as previously described (Leather and Dixon, 1984; Stewart et al., 2009; Ziebell et al., 2011; Westwood et al., 2013) (histograms on left). Ten days later the offspring produced by each aphid were counted and the mean number of aphids in each colony calculated (Ziebell et al., 2011; Westwood et al., 2013; Tungadi et al., 2020) (histograms on right). Plants were mock‐inoculated (Mock) with sterile water or inoculated with purified virions of reconstituted Fny‐CMV (F1‐F2‐F3), or versions of the reassortant Fny‐CMV/LS‐CMV virus F1‐L2‐F3, with site‐specific mutations in RNA 2 affecting the 2a ORF (described in Table 1 and Figure 1c). Plants were used for aphid performance experiments at 10 days postinoculation. Error bars represent standard error around the mean. Statistical analysis was performed using one‐way analysis of variance followed by Dunnett's post hoc multiple comparisons test (significant differences indicated by ****p < .0001; ***p < .001; **p < .01, and *p < .01) using R (Dalgaard, 2008)

Thus, we think it probable that the valine at position 237 in the Fny‐CMV 2a protein sequence plays some role in induction of antixenosis against aphids in CMV‐infected A. thaliana. That replacement of isoleucine at this position in the LS‐CMV 2a protein sequence with valine had such a marked effect was initially surprising because both amino acids have hydrophobic side chains, making the I237V replacement conservative relative to some of the other sequence replacements. However, the literature provides several examples where replacement of isoleucine with valine, or vice versa, has profound effects on the biological activity of proteins. These examples include, among others, the effect of the brome mosaic virus movement protein on symptomology in N. benthamiana (Rao and Grantham, 1995), effects on herbicide resistance of the photosynthetic D1 protein (Mengistu et al., 2000; Dumont et al., 2016), Escherichia coli recA protein function (Knight et al., 1984), human 5‐hydroxytryptamine receptor and glutathione S‐transferase activity (Zimniak et al., 1994; Nakhai et al., 1995), and on β‐amyloid induced neuropathology (Yoshioka et al., 1991). It remains unknown how the Fny‐CMV 2a protein valine‐237 might be involved in inducing antixenosis against aphids in A. thaliana plants, and we cannot exclude additional roles for other residues of the 2a protein. Our working hypothesis is that this residue directly or indirectly aids an interaction between the Fny‐CMV 2a protein and a host factor or factors involved in either defensive signalling or the regulation of metabolism, leading to increased production of 4‐methoxy‐indol‐3‐yl‐methylglucosinolate and/or other antixenotic compounds.

CONFLICT OF INTEREST

The authors declare that they have no conflicts of interest.

FIGURE S1

Click here for additional data file.

FIGURE S2

Click here for additional data file.

FIGURE S3

Click here for additional data file.

FIGURE S4

Click here for additional data file.

TABLE S1

Click here for additional data file.

TABLE S2

Click here for additional data file.

METHOD S1

Click here for additional data file.

SPREADSHEET S1

Click here for additional data file.

SPREADSHEET S2

Click here for additional data file.

SPREADSHEET S3

Click here for additional data file.