Characterization of Grapevine Fanleaf Virus Isolates in 'Chardonnay' Vines Exhibiting Severe and Mild Symptoms in Two Vineyards

Abstract

Fanleaf degeneration is a complex viral disease of Vitis spp. that detrimentally impacts fruit yield and reduces the productive lifespan of most vineyards worldwide. In France, its main causal agent is grapevine fanleaf virus (GFLV). In the past, field experiments were conducted to explore cross-protection as a management strategy of fanleaf degeneration, but results were unsatisfactory because the mild virus strain negatively impacted fruit yield. In order to select new mild GFLV isolates, we examined two old 'Chardonnay' parcels harbouring vines with distinct phenotypes. Symptoms and agronomic performances were monitored over the four-year study on 21 individual vines that were classified into three categories: asymptomatic GFLV-free vines, GFLV-infected vines severely diseased and GFLV-infected vines displaying mild symptoms. The complete coding genomic sequences of GFLV isolates in infected vines was determined by high-throughput sequencing. Most grapevines were infected with multiple genetically divergent variants. While no specific molecular features were apparent for GFLV isolates from vines displaying mild symptoms, a genetic differentiation of GFLV populations depending on the vineyard parcel was observed. The mild symptomatic grapevines identified during this study were established in a greenhouse to recover GFLV variants of potential interest for cross-protection studies.

Keywords: cross-protection; fanleaf degeneration; genetic and phenotypic diversity; grapevine; grapevine fanleaf virus; high throughput sequencing; mild isolates; symptomatology; virome.

Figure 1

Comparative analyses of qualitative and quantitative traits for the three vine categories in 2016 (M−, mild symptomatic GFLV-free vines (in grey); M+, mild symptomatic vines infected with GFLV (in cyan) and S+, severe symptomatic vines infected with GFLV (in black)). Number of vines (n) are given in brackets. * For the vines in the M+ category, only three vines were considered in June 2016 for the evaluation of qualitative traits (leaf discolouration, leaf deformation, vine stunting, overall symptoms). Significance was tested with Student’s t, Welch’s t or Kruskal–Wallis H tests. Boxplots show the median (horizontal bold line) and the interquartile range with lonely dots representing extreme data. Different letters (a,b) indicate significant differences (p ≤ 0.05). Raw data and results of the statistical tests are available in Tables S1 and S3.

Figure 2

Principal components analysis (PCA) and hierarchical clustering of 21 vines according to their mean symptom scores and yield values over the four-year study (from 2016 to 2019). (A) Biplot representing the ‘M−’ (grey), ‘M+’ (cyan) and ‘S+’ (black) individuals. Projection of M−, M+ and S+ categories are represented by squares with their confidence ellipses (95%). (B) Dendrogram representing the hierarchical clustering based on the individuals coordinates on the 10 principal components. (C) Loading biplot, the vectors show the relative importance of each variable in discriminating amongst the observations. Longer vectors indicate greater contribution. Vectors point in the direction they ordinate the observations (DIS: Leaf discolouration, DEF: leaf deformation, STU: vine stunting, ALL: overall, CLU: cluster number, FY: fruit yield, FYpc: fruit yield per cluster, CAN: cane number, PWW: pruning wood weight per vine, PWWpc: pruning wood weight per cane). Results of the Fisher’s test are available in Table S4.

Figure 3

Evolution of yearly mean overall symptom scores and fruit yields of vines in the M−, M+ and S+ categories and their standard error. Results for vines in the M− (in grey), M+ (in cyan) and S+ (in black) categories are shown. Number of vines (n) is given in brackets. Means values are connected by bold lines and standard errors are materialized by vertical bars. * For the M+ category, only three vines were considered in June 2016 for the evaluation of the qualitative traits.

Figure 4

Comparison of overall symptom scores and fruit yields over four years for vines in M−, M+ and S+ categories. Results for vines in the M− (in grey), M+ (in cyan) and S+ (in black) categories are shown. Boxplots show the median (horizontal bold line) and the interquartile range with lonely dots representing extreme data. Different letters (a,b,c) indicate significant differences (p ≤ 0.05). Raw data and results of the statistical tests are available in Tables S1 and S3, respectively. Number of vines (n) is given in brackets. * For the M+ category, 15 replicates were considered for overall symptom scores.

Figure 5

Phylogenetic relationships between grapevine fanleaf virus (GFLV) molecular variants. Maximum likelihood trees were inferred from the complete nucleotide sequences of ORF1 (A), ORF2 (B) and ORF3 (C) recovered from GFLV-infected ‘Chardonnay’ vines. Each sequence name indicates the vineyard site (Pa, in orange and, Py, in deep blue), the sampled vine, the molecular variant and the vine affiliation to a phenotypic category (M+, in cyan and S+, in black) in parenthesis. Clades are named with the Roman numeral corresponding to the GFLV ORF and classified from the most [A] to the least [L] represented clade in sequence number. The names of the distinct clades are highlighted in bold. Sequence variants sharing at least 95% nucleotide identity were considered in the same clade; pairwise nucleotide identities values between GFLV sequences are available in Table S6. Scale bars below each tree show genetic distance. Only bootstrap values ≥ 0.95 are indicated.

Figure 6

Genetic diversity analyses of grapevine fanleaf virus (GFLV) ORF1 and 2 sequences. Graphics represent genetic diversity (π, substitution per site) along each ORF1 and ORF2 sequence and Tajima’s D (D_T) for evolution study (# p < 0.10, * p < 0.05, *** p < 0.001). Sequences were grouped according to either the phenotypic category or the vineyard site. A colour code was applied for each population: black for ‘S+’, cyan for ‘M+’, orange for ‘Pa’, deep blue for ‘Py’ and dark grey for all sequences. Number of sequences (n), overall genetic diversity π (± standard error, S.E.); the diversity of synonymous (dS) and nonsynonymous (dN) substitutions (dN − dS < 0: negative/purifying selection; dN − dS = 0: neutral/conservative selection; dN − dS > 0: positive/diversifying selection) and the overall Tajima’s D (D_T) (for all D_T values p > 0.10, non-significative) are given in the tables below the graphs. D_T = 0 corresponds to a mutation-drift equilibrium, D_T > 0 indicates balancing selection, sudden population contraction and D_T < 0 distinguishes a recent selective sweep, population expansion after a recent bottleneck. Genetic differentiation of GFLV populations is expressed as the fixation index (F_ST) with associated p-value (p). Significance (p ≤ 0.05) is indicated in bold.

Figure 7

Comparison of grapevine fanleaf virus (GFLV) RNA1 and RNA2 mean RPKM values (left) or mean number of total GFLV genomic RNA variants (right) detected in vines in the M+ versus S+ categories (A) or from the Pa versus Py vineyard sites (B). Boxplots show the median (horizontal bold line) and the interquartile range with lonely dots representing extreme data. Significance was tested with a Student’s t or Kruskal–Wallis H tests (for all tests p > 0.05, non-significant as indicated by letter ‘a’). Results obtained from vines of the M+ (in cyan) and S+ (in black) categories, and from the Pa (in orange) and Py (in deep blue) are shown. Number of vines (n) is given in brackets. Raw data and results of the statistical tests are available in Tables S5 and S8, respectively.