Effects of the virus satellite gene βC1 on host plant defense signaling and volatile emission

Abstract

Tomato Yellow Leaf Curl China virus spreads together with its invasive vector, the silverleaf whitefly B biotype, which exhibits higher growth rates on infected plants. Previous studies indicate that the virus satellite gene βC1 accounts for the visible symptoms of infection and inhibits the constitutive expression of jasmonic acid (JA)—a phytohormone involved in plant defense against whiteflies—and of some JA-regulated genes. Here we present new details of the effects of on plant signaling and defense, obtained with (non-host) transgenic Arabidopsis thaliana and Nicotiana benthamiana plants. We found that JA induction in response to wounding was reduced in plants expressing βC1. This result implies that βC1 acts on conserved plant regulation mechanisms and might impair the entire JA defense pathway. Furthermore, transformed N. benthamiana plants exhibited elevated emissions of the volatile compound linalool, suggesting that βC1 also influences plant-derived olfactory cues available to vector and non-vector insects.

The fitness of plant viruses is critically influenced by interactions with the physiology of the host plant, which not only mediate successful propagation within the host but—for insect-vectored viruses—may also influence plant traits that affect vector attraction, pathogen acquisition and dispersal to new hosts.¹ Direct suppression by viruses of plant anti-pathogen defenses (e.g., through RNA silencing) is well documented,² and specific viral proteins involved in subverting defenses that would otherwise inhibit propagation have been identified in many systems.^3,4 But viruses have also been shown to induce effects on plant phenotypes that favor transmission through indirect effects on the nature of interactions between the host plant and insect vectors, for example by modifying host plant nutritional quality, anti-herbivore defenses and the olfactory cues emitted by infected plants.^5-8 The specific effectors and physiological mechanisms underlying these indirect effects are currently not well understood. Improved knowledge of these factors would enhance our understanding of plant disease ecology and also inform efforts to develop more effective control strategies for the many plant diseases that plague agriculture worldwide.

The well-documented association between Tomato yellow leaf curl China virus (TYLCCNV) and its whitefly vector provides an outstanding opportunity to explore the mechanisms underlying virus effects on plant phenotypes that significantly impact plant-vector interactions. TYLCCNV—a single-stranded DNA virus in the begomovirus genus—is transmitted exclusively by the whitefly Bemisia tabaci⁹ and has spread along with the invasive B. tabaci B Biotype in south China, causing major losses in tomato and tobacco crops.^10,11 Recent studies^10,12 demonstrated that B Biotype whiteflies have higher fecundity on TYLCCNV-infected tobacco plants than on healthy hosts. This suggests a mutualistic relationship between the pathogen and vector, in which the presence of the virus enhances vector reproduction and spread, via a potentially adaptive manipulation of the host plant phenotype. The specific mechanisms underlying the effects of TYLCCNV on host plants are not yet well understood, but the obligatory satellite DNA-β of the virus and its single gene, βC1 appear to play important roles. Previous studies have shown that this satellite is essential for pathogenicity, including the onset of visible symptoms and the accumulation of high levels of virus particles.^11,13 βC1 has also been shown to influence many plant regulation systems including the inhibition of RNA silencing,¹⁴ the disruption of leaf developmental regulation (causing the visual symptoms) and the inhibition of the jasmonic acid signaling pathway.^12,15

In the current study, we explored how βC1 impacts host plant phenotype likely involved in herbivores attraction and defense by using transgenic lines of two model plant species, Arabidopsis thaliana and Nicotiana benthamiana, transformed to express this virus gene.¹⁵ Although not usual hosts of the virus the insertion of the βC1 gene in these two biological systems produces morphological symptoms similar to those of TYLCCNV, notably leaf curliness and outgrowth tissues, with a level of phenotype severity correlated with the expression of the gene in the transformed line.^13,15 These transgenic lines are thus useful tools to investigate the action of this gene on conserved plant regulation mechanisms, in isolation from other physiological perturbations that an actual viral infection could create.

We specifically explored how phytohormone-mediated defense signaling pathways are affected by the presence of this gene. Two key phytohormones involved in plant responses to antagonists are jasmonic acid (JA) and salicylic acid (SA). The former is typically activated in response to physical damage and/or feeding by chewing herbivores and the latter in response to biotrophic pathogens or phloem-feeding insects, though these roles frequently vary.^16,17 There is furthermore considerable evidence for crosstalk—including mutual inhibition—between these and other defense signaling pathways,¹⁸ which presumably allows plants to fine-tune defense responses to specific antagonists, but also creates opportunities for manipulation by insect herbivores and pathogens.^19-21 Despite being phloem feeders, whiteflies are susceptible to JA-mediated defenses, and are furthermore known to inhibit this hormone in Arabidopsis and by so doing increase population growth rates.²² In tobacco hosts, Zhang et al.¹² showed that constitutive JA production was reduced when both the virus and its satellite are present and that JA-responsive genes were inhibited specifically by βC1, which results in a reduction of defenses against whiteflies. In Arabidopsis, Yang et al.¹⁵ also reported reduced constitutive expression of some JA-responsive genes in βC1-transformed Arabidopsis despite no changes in the expression of JA-synthesis genes. Both these studies, however, focused on constitutive states of the plants. Building on these observations, we directly investigated the induced levels of JA and SA as well as their precursors, after mechanical wounding in βC1-transformed Arabidopsis and N. benthamiana plants, in order to explore whether their ability to mount a JA pathway defense in response to external stress was also impaired in the presence of this gene—for example, through a “decoy strategy” exploiting the mutual inhibition of SA and JA²⁰—as a manipulation by TYLCCNV to favor the growth and dissemination of whitefly vectors.

We also explored the impacts of this gene on the olfactory phenotype of host plants. Plant volatile emissions are known to be influenced by the induction of phytohormone signaling pathways, and several aphid-transmitted viruses (from different genera) have been reported to induce qualitative and quantitative alterations in the volatile emissions of host plants that enhance vector attraction.^6,7,23 The manipulation of such cues has significant potential to influence rates of pathogen transmission through effects on the frequency of opportunities for virus particles to be acquired from infected plants and introduced to healthy ones.²⁴

Results

Phytohormones levels in A. thaliana and N. benthamiana βC1-transformed plants

The parental wild type of A. thaliana (Col-0) and four βC1-transformed lines representing two class levels of symptoms severity (Class I: mild phenotype, lines βC1-#2 and βC1-#42; class II: severe phenotype, lines βC1-#5 and βC1-#30) were compared for their production of JA (cis and trans isomers), SA, cinnamic acid (CA), Indole-3-acetic acid (Auxin, IAA), linoleic acid (LA) and linolenic acid (LN) after damage. As constitutive levels of the phytohormone JA are undetectable or very low in most undamaged Arabidopsis and peaks about one hour after wounding,²⁵ mechanical damage by leaf perforation was employed to elicit a JA response that could be compared between treatments. This method also ensured consistent amounts of damage to plants independently of their transformation status. After preliminary experiments at different sampling intervals, a time lapse of 45 min between damage and sampling was selected.

Overall, total jasmonic acid production (cis and trans isomers) after damage was significantly lower on the two severity classes of βC1 transformed lines than on the wild type, despite showing some variation among individual lines with lines βC1/Col-0#2 and βC1/Col-0#2 (Table 1; Fig. 1A). Similarly, linoleic and linolenic acid production was lower on βC1 lines, though these differences were not significant, with a P value of 0.07. On the other hand, SA production was higher on all βC1 lines, though again without reaching a significant difference (Table 1; Fig. 1B). The other hormones examined, cinnamic acid and auxin, were not significantly different between any lines.

Table 1. ANOVA results on phytohormones production in A. thaliana wild type and transgenic lines.

	Block		Phenotype class		Line (class)
Variable	F7,24	P value	F2,24	P value	F2,24	P value
Total JA	2.02	0.094	7.16	0.0036	3.53	0.045
SA	2.32	0.058	2.95	0.072	1.71	0.20
CA	1.89	0.12	1.31	0.29	1.66	0.21
IAA	1.59	0.19	0.55	0.59	1.33	0.28
LA	5.75	0.0006	2.85	0.078	0.096	0.91
LN	4.51	0.0027	2.97	0.071	0.97	0.39

For each variable tested, the F-ratios and associated P values of the following explanatory variables are indicated: block, phenotype class and A. thaliana line nested within phenotype class. Abbreviations: Total jasmonic acid, Total JA (i.e. sum of cis and trans isomers); salicylic acid, SA; cinnamic acid, CA; indole-3-acetic acid (i.e. auxin), IAA; linoleic acid, LA; linolenic acid, LN

Figure 1. Phytohormones responses to mechanical damage in Arabidopsis thaliana lines. (A) Total Jasmonic acid production (cis and trans isomers); (B) Salicylic acid production (Mean values +/− s.e., letters indicate significant differences). Left side graphs present the results for each tested line, and the transgenic lines are pooled by phenotype class in the right side graphs.

Compared with A. thaliana, N. benthamiana plants produced smaller amounts of JA for a given amount of damage. However, there was a difference between the two tested tobacco lines (Fig. 2A), with the βC1-transformed line having lower levels of JA compared with the wild type line (ANOVA on Log transformed values for total JA: F_1,18 = 11.18, P value = 0.0036). SA values, as well as linoleic and linolenic acids products were not statistically different between the βC1 and wild type lines (SA: F_1,18 = 0.27, P value = 0.61 ; LA: F_1,18 = 0.06, P value = 0.81 ; LN: F_1,18 = 0.27, P value = 0.61, Fig. 2B). Cinnamic acid and auxin levels were not detectable.

Figure 2. Phytohormones responses to mechanical damage in N. benthamiana lines. (A) Total Jasmonic acid production (cis and trans isomers); (B) Salicylic acid production (Mean values +/− s.e., letters indicate significant differences).

Volatile organic compounds emission in N. benthamiana βC1-transformed plants

The collection of volatile organic compounds was focused on the N. benthamiana lines, as the in-vitro cultured A. thaliana plants used in this study did not produce any quantifiable volatiles. The volatile organic compounds emitted by the two tobacco lines (wild type and βC1-transformed) were collected during separate day and night collections on HaysepQ filters and analyzed by gas chromatography. All plants consistently emitted linalool during the day collection, but exhibited little or no volatile emissions during the night; thus, we quantified only the linalool emitted during the 14h daytime collections. Relative to plant fresh mass, this emission was higher in the transformed βC1 line than in the wild type N. benthamiana (ANOVA on Log-transformed values of Linalool: F_1,24 = 5.59, P value = 0.0265) (Fig. 3).

Figure 3. Linalool emission during 14 h of day by Nicotiana benthamiana lines. Mean values +/− s.e., letters indicate significant differences.

Discussion

As noted above, the coincident spread of TYLCCNV and its whitefly vector in China appears to be favored by the synergistic effects of virus and vector on transmission.¹⁰ Yang et al.¹⁵ found that Arabidopsis transformed with the βC1 gene—carried by the virus’s β satellite—reduced the constitutive expression genes PDF1.2, PR4 and CORI3 acting in downstream pathways of the JA defense response, which is effective against whiteflies, but no effect on constitutive expression of the JA biosynthesis genes LOX2 and AOS, while Zhang et al.¹² found that the JA biosynthesis genes FAD3 and FAD7 were inhibited in N. tabacum transformed plants. Our results show that the βC1 gene also impairs the production of JA in response to wounding. It suggests that this gene can have an even larger impact on the JA resistance pathway by preventing the accumulation of this phytohormone in response to external stress. In the two transformed plant systems we examined, A. thaliana and N. benthamiana, the presence of the βC1 gene led to a lower accumulation of JA in damaged plants compared with their untransformed conspecifics. Since the gene was expressed in the transgenic lines in the absence of virus replication, we can rule out the possibility that the response observed is simply a reaction of the plant to infection. Furthermore, as independent gene insertions in the different Arabidopsis lines tested present the same pattern (with two lines having significantly lower JA and two being intermediate), it is unlikely that the observed effect was due to a disruption of plant gene expression by the insertion of the plasmid in a critical portion of the genome. The finding that two unrelated host species show the same deficiency in JA response in the presence of βC1 also suggests that this gene interferes with plant regulation in a non-specific way, especially since A. thaliana is not a host for TYLCCNV and N. benthamiana is not indigenous in infested areas—although its agro-infiltration is possible. Interestingly, there was no apparent correlation between the class of symptoms severity, linked to the expression level of βC1,¹⁵ and the intensity of JA reduction in A. thaliana. One explanation might be that the mechanisms involved in JA impairment can saturate at higher βC1 levels.

In contrast to JA, the other hormones and precursors we examined were not significantly affected. Specifically, we found little evidence of SA induction in βC1-transfomed plants—although SA levels were consistently, though not significantly higher in transformed Arabidopsis lines. This suggests that the observed suppression of JA induction may not be accomplished via a “decoy strategy” exploiting cross talk between the JA and SA signaling pathways,²⁰ but by another mechanism, potentially involving direct inhibition of JA synthesis genes with shared homologs between brassicaceae and solanaceae species.

While JA-mediated responses are not known to have any direct effect on TYLCCNV, this hormone appears to be key to plants’ defense against whiteflies.¹² Zarate et al.²² demonstrated that B biotype silverleaf whiteflies (the vector for TYLCCNV) exhibit a higher developmental rate on JA deficient hosts, and are themselves able to downregulate JA production to their advantage. The disruption of the entire JA pathway in plants expressing βC1 is thus very likely to further benefit the vector and might explain the increase in B. tabaci B biotype growth rate on virus infected plants observed in Jiu et al.¹⁰ For a persistent virus such as TYLCCNV, the successful acquisition of transmissible virions by their whiteflies vectors requires prolonged feeding (16–24 h) vector on an infected plant.²⁶ In contrast to non-persistent viruses that benefit from rapid vector dispersal, TYLCCNV should thus be more likely to benefit from effects on host plants that improve plant quality for vectors.^7,27

In addition to changes in phytohormonal response, our analyses of plant volatile emission revealed elevated constitutive emission of volatile linalool by N. benthamiana plants. Organic volatile compounds emitted by plants are a key cue for their location by phytophagous insects²⁸ and pathogen-induced elevation of plant volatile emissions has previously been implicated in vector attraction to infected plants.^6,7 Whitefly adults are believed to locate their host plants mostly through visual cues but olfactory cues also play a significant role in host plant choice.²⁹ Moreover, while tests of whitefly responses to specific olfactory cues are rare, owing to the significant challenges entailed in conducting behavioral assays with these insects, linalool was one of several individual plant-derived compounds reported to elicit positive behavioral responses from adult B biotype B. tabaci females, along with 1,8-cineole, eugenol, limonene and myrcene.³⁰ Thus, the elevated emission of linalool might be expected to increase the appearance, and potentially the attractiveness, of potential host plants for whitefly vectors.

Despite being located on a satellite isolated from the main virus genome particle (DNA-A) and incapable of independent replication, the βC1 gene appears to perform at least two critical functions for TYLCCNV: it directly provides protection against host-plant RNA silencing¹⁴ and it indirectly facilitates vector establishment through effects on host plant phenotype. While the DNA-A part of the virus can replicate to intermediate levels within the host in the absence of the DNA-β, the satellite is required for full pathogenicity and effective transmission.¹³ It is also responsible for most of the morphological deformations of the leaves on diseased plants—which may be a by-product of the disruption of regulatory pathways that play a role in both defense and development¹⁵ or could conceivably play an adaptive role in improving plant quality for whiteflies (e.g., by providing enhanced shelter). A similar dual function has been observed for the gene 2b of Cucumber mosaic virus, which codes for a protein that was initially identified as a RNA-silencing inhibitor but also disrupts host JA pathways, potentially favoring transmission of this virus by aphid vectors.³¹ The identification of such genes, and elucidation of their functions, is an important step toward an improved understanding of the mechanisms that govern plant virus effects on host plant phenotypes that mediate transmission by insect vectors, and might provide a lead for control strategies to limit the spread of these diseases by breaking the mutualistic part of the virus-vector association.

Materials and Methods

Transformed Arabidopsis thaliana lines

Arabidopsis lines were originally transformed from the accession Columbia-0 by inserting the βC1 gene expressed as a HA fusion protein and associated to promoters 35S and XVE, the later allowing β-estradiol induction of gene expression.¹⁵ Different insertions were produced independently and exhibited various levels of severity of morphological abnormalities, ranging from mild leaf curling (Class I) to severe curling associated with outgrowths tissues from abaxial leaves (Class II).¹⁵ The level of phenotype severity was correlated with expression of the gene in the transformed line.^13,15 Two lines presenting a class I phenotype (XVE:HA:βC1/Col-0#2 and XVE:HA:βC1/Col-0#42) and two lines with a severe class II phenotype (XVE:HA:βC1/Col-0#5 and XVE:HA:βC1/Col-0#30) were chosen for this study and compared with the Col-0 accession from which they were originated. Plants used in our experiments were grown in vitro on Murashige and Skoog (MS) medium supplemented with 20 µM β-estradiol.³²

Transformed Nicotiana benthamiana lines

N. benthamiana is a wild relative of cultivated tobacco. For our experiments, we used a line of N. benthamiana transformed with the construct 35S:HA:βC1 as well as the inbred wild type N. benthamiana line from which the transgenic line was produced. In almost all transformed lines, the phenotype was too severe to permit the production of viable fruits, but a few fruits with viable seeds were recovered and 2nd generation plants showed a phenotype similar to the parent. Plants grown from these seeds were employed in the experiments reported here. All plants were first germinated and grown on MS medium for three weeks, then transplanted on 10 cm × 10 cm pots of Pro-Mix potting soil supplemented with 3g of Osmocote fertilizer (NPK 14:14:14, The Scotts Company) and macronutrients.

Phytohormones production in A. thaliana βC1-transformed lines

Eight plants of each A. thaliana line were grown in vitro on MS medium supplemented with 20 µM β-estradiol in a growth chamber (16 h:8 h light:dark, 22°C) under a randomized design. The five treatments were distributed across eight PlantCon^™ boxes (MP Biomedicals), treated as blocks, with one plant of each treatment located randomly in each box. At one month, all leaves of all plants were perforated on average twice with a 0.5 mm diameter punch, the largest leaves received three holes and the smallest only one. After 45 min, the plants were cut, weighted and immediately frozen in liquids nitrogen. Some plants failed to germinate and two extracted samples were of too bad quality to be analyzed so only five replicates of line #30 and seven of Col-0 were used. The extraction and quantification of phytohormones followed the protocol described in Schmelz et al.^33,34 Briefly, we derivatized free carboxylic acids to methyl esters, which were isolated using vapor phase extraction and analyzed by GC-MS with isobutane chemical ionization using selected-ion monitoring. We quantified amounts of JA, using 100 ng of dihydrojasmonic acid, SA using 100 ng of [2H6]SA, IAA using 100 ng of [2H5]IAA, CA using 100 ng of [2H5]CA (CDN Isotopes) and LA using 100 ng of gamma-linolenic acid (Matreya LLC). These internal standards were added to the samples prior to processing. To confirm the identity of methyl linolenate (meLA), methyl linoeate (meLA), methyl jasmonate (meJA), methyl salicylate (meSA), methyl cinnamate (meCA) and methyl indole-3-acetate (meIAA) in our samples, we analyzed extracts by GC-MS with electron ionization, comparing retention times and spectra with that of pure compounds. The final amounts of methylated phytohormones were then corrected for the initial fresh mass of the sample. The values obtained for the cis and trans isomers of JA showed similar patterns of production and were pooled as a total JA value for the subsequent analyses.

Phytohormones production in N. benthamiana βC1-transformed line

Ten replicates of each HA:βC1 and wild type N. benthamiana were germinated and grown in vitro on MS medium in individual PlantCon^™ boxes in a growth chamber (16 h:8 h light:dark, 22°C:20°C). After three weeks, the young plants were delicately transplanted into compost soil pots as described above, and covered by a plastic dome until sampling. Wounding and Phytohormones sampling was performed three days later. As with A. thaliana, mechanical damage was applied to ensure the production of induced JA levels by perforating each of the leaves with a 2 mm punch and holes were distributed equally between small and larger leaves at 1 cm intervals (Fig. 4A and B). About 0.2 g of leaf material was sampled after one hour, weighted and immediately frozen in liquid nitrogen for later extraction (using the protocol described above).

Figure 4. Typical phenotypes of N. benthamiana wild and transformed lines. (A and B) 25-d-old plants, wild type and βC1 line respectively, that were used in the phytohormones extraction experiment. (C and D) 30-d-old plants, wild type and βC1 line respectively, that were used in the volatile collection. White bars correspond to 2 cm.

Volatile collections from N. benthamiana

Plants used in volatile collection were grown in vitro for three weeks then transplanted into compost soil 10 d prior to the collection, which was performed in two consecutive blocks (for a total of 12 replicates of βC1 and 15 of the wild type). The potted plants were placed under four liters glass domes with a Teflon^© base covering the pot and cotton around the plant stem to prevent air contamination. Total volatiles were collected consecutively for 8 h of night and 14 h of day in a climate-controlled growth chamber (16 h:8 h light:dark, 22°C:20°C) on two separate HaysepQ filters using a clean-air system that pushed 1.5 L/min of air into the domes and pulled 1L/min through the filters. The trapped volatiles were then eluded in 150 µL of dichloromethane with 5 μL of an internal standard added after eluting (80 ng/μL nonyl acetate, 40 ng/μL n-octane). Samples were injected in 1-μL aliquots into an Agilent model 6890 gas chromatograph fitted with a flame ionization detector (column: Agilent 19091Z-331, 0.25 mm internal diameter, 0.1 μm film thickness). The column was held at 35°C for 0.5 min then increased by 7°C per min to 150°C and further increased by 20°C per min to a maximum temperature of 220°C. Linalool was the only compound systematically emitted by all plants during the day and was produced in sufficient quantities to be identified and quantified. This compound was quantified using MSD Chemstation (Agilent Technologies 2003) by measuring volatile output in nanograms relative to the internal standard and corrected by fresh plant mass, rather than surface area as is usually done, in order to avoid bias due to the curliness of βC1-transformed leaves (Fig. 4C and D).

Statistical analyses were performed with JMP (SAS institute). Data were analyzed by ANOVA with treatment (line type) and block (when needed) as explanatory variables. In the A. thaliana phytohormones experiment, treatments were divided into three classes of severity (Wild type controls, mild transgene phenotype and severe transgene phenotype), with the individual lines nested within, since the different severity classes have different levels of βC1 expression. N. benthamiana JA values as well as volatile emissions of linalool were log-transformed in order to ensure equal variances and a normal distribution of the ANOVA residuals.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.