Cauliflower mosaic virus protein P6-TAV plays a major role in alteration of aphid vector feeding behaviour but not performance on infected Arabidopsis

Abstract

Emerging evidence suggests that viral infection modifies host plant traits that in turn alter behaviour and performance of vectors colonizing the plants in a way conducive for transmission of both nonpersistent and persistent viruses. Similar evidence for semipersistent viruses like cauliflower mosaic virus (CaMV) is scarce. Here we compared the effects of Arabidopsis infection with mild (CM) and severe (JI) CaMV isolates on the feeding behaviour (recorded by the electrical penetration graph technique) and fecundity of the aphid vector Myzus persicae. Compared to mock-inoculated plants, feeding behaviour was altered similarly on CM- and JI-infected plants, but only aphids on JI-infected plants had reduced fecundity. To evaluate the role of the multifunctional CaMV protein P6-TAV, aphid feeding behaviour and fecundity were tested on transgenic Arabidopsis plants expressing wild-type (wt) and mutant versions of P6-TAV. In contrast to viral infection, aphid fecundity was unchanged on all transgenic lines, suggesting that other viral factors compromise fecundity. Aphid feeding behaviour was modified on wt P6-CM-, but not on wt P6-JI-expressing plants. Analysis of plants expressing P6 mutants identified N-terminal P6 domains contributing to modification of feeding behaviour. Taken together, we show that CaMV infection can modify both aphid fecundity and feeding behaviour and that P6 is only involved in the latter.

Keywords: aphid vector; plant virus; vector behaviour; vector modification; vector transmission; viral factors.

FIGURE 1

Symptoms of CaMV infection and viral protein accumulation in Arabidopsis Col‐0 plants. (a) Mock‐inoculated, (b) CM‐infected, and (c) JI‐infected plant 21 days postinoculation with virus‐free or viruliferous aphids. Scale bar = 5 cm. (d) Western blot analysis of accumulation of CaMV proteins in CM‐ and JI‐infected Arabidopsis 21 days postinoculation. The membranes were stained for P2 (left panel), P4 (middle panel), and P6 (right panel). Ponceau red staining of the large RuBisCO subunit is shown as a loading control. Molecular weights are indicated on the right of the blots

FIGURE 2

Fecundity and feeding behaviour of Myzus persicae on mock‐inoculated or CaMV‐infected Arabidopsis. Two‐week‐old plants were inoculated with the indicated isolate and used 3 weeks later for the experiments. (a) Aphid fecundity after 5 days of infestation (n = 28–33). (b) Aphid feeding behaviour parameters recorded with the electrical penetration graph (EPG) technique (n = 22–27). Letters show significant differences between plant infection status as tested by GLM followed by pairwise comparisons using “emmeans” (p < .05; method: Tukey)

FIGURE 3

Phenotype of 5‐week‐old transgenic Arabidopsis plants expressing various P6 proteins. The images show (a) an untransformed Col‐0 plant and (b–h) transgenic Col‐0 plants expressing (b) P6 from the CaMV isolate CM; (c) P6 from isolate D4; (d) P6 from isolate JI; (e) the P6 dsR domain deletion mutant from isolate JI, referred to as JI‐ΔdsR; (f) the P6 Eki mutant from isolate JI, referred to as JI‐Eki; (g) HA‐tagged wild‐type P6 from CM (CM‐HA), and (h) the HA‐tagged P6 Vir/Avr domain deletion mutant from isolate CM, referred to as CM‐Δd23‐HA. Scale bar = 5 cm. (i) Western blot analysis of P6 protein accumulation in 5‐week‐old transgenic plants expressing the indicated P6 mutants. The figure shows the same blot revealed with a short (left) or a longer (right) exposure time to visualize weak bands. Extracts prepared from untransformed Col‐0 and from CM‐infected leaves were loaded as negative and positive controls, respectively. (j) Detection of P6‐JI‐Eki and P6‐CM‐Δd23‐HA by western blot. The two proteins that were not detected in the blot shown in (i) could be revealed in a different experiment using more concentrated extracts. Each lane presents extract from a different transgenic plant. Signals from wild‐type P6 loaded on the same blots are shown to the left of the panels and either the blots were exposed much shorter (P6‐JI‐Eki) or extracts were diluted (P6‐CM‐Δd23‐HA). Note that the mutant P6 concentration varied considerably. Ponceau red staining of the large RuBisCO subunit is shown as a loading control

FIGURE 4

Myzus persicae fecundity 5 days after deposit on untransformed or transgenic 5‐week‐old Arabidopsis plants expressing P6 proteins. Different letters show significant differences between plants as tested by GLM followed by pairwise comparisons using “emmeans” (p < .05; method: Tukey, n = 21–24). No statistically significant differences were recorded

FIGURE 5

Aphid feeding behaviour parameters recorded by EPG on untransformed Col‐0 and transgenic P6‐expressing 5‐week‐old Arabidopsis. Different letters indicate significant differences between plants as tested by GLM followed by pairwise comparisons using “emmeans” (p < .05; method: Tukey, n = 25–31)

FIGURE 6

Functional domains of P6 and P6 mutants used in this study. Only relevant domains and interacting proteins are shown. The N‐terminal half of P6 contains three major regions: The N‐terminus (amino acids 4–31) contains one of several P6/P6 self‐interaction domains required for P6’s function as the matrix protein of virus inclusions. This region is followed by the Vir/Avr domain delimited by the d23 region, which is involved in chlorosis induction and dwarfism. The mini‐TAV domain comprises amongst others an N‐terminal region binding to ribosomal proteins eL13 and eL18, the central double‐stranded RNA‐binding (dsR) region that interacts with TOR kinase, and a C‐terminal region that interacts with reinitiation supporting protein (RISP) and contains also a nuclear localization signal (NLS1). The mini‐TAV region is important for antiviral autophagy, innate immunity, and polycistronic translation. One of the functions of the C‐terminal half of P6 is probably suppression of RNA silencing, in which also the N‐terminal nuclear export signal (NES) and the two nuclear localization signals (NLS1 in the mini‐TAV domain and NLS2 in the C‐terminal region) play a role. The dsR region is deleted in P6‐JI‐ΔdsR. In P6‐JI‐Eki, the conserved Eki motif just preceding the NES is substituted by three alanines. CM‐HA contains an N‐terminal HA‐tag as modification and CM‐Δd23‐HA in addition a deletion of the d23 region. RBa, RBb, RNA‐binding regions. Figure adapted from Pooggin and Ryabova (2018)