Abstract
Grape phylloxera (Daktulosphaira vitifoliae) feeds exclusively on Vitis species, preferentially on leaves of American Vitis species and roots of European Vitis vinifera. Over the last 15 years, extensive feeding and galling on V. vinifera leaves have been observed in Italy, Brazil, and Peru. In Uruguay, D. vitifoliae infestations on V. vinifera leaves were detected in very high densities. The cause of this unexpected insect behavior is unknown, but possible explanations include selection pressure for more aggressive native strains in the new context of vigorous plants replacing old vineyards, loss of resistance in plants due to improvement programs, or importation of exotic strains of the insect. The aims of this research were to evaluate genetic diversity of leaf-galling populations of Uruguayan phylloxera, to estimate genetic distances among them, and to compare Uruguayan and foreign phylloxera populations (Brazilian, Peruvian, and European). Genetic distances between root and leaf samples from the same plant were also estimated. Four polymorphic microsatellite primers were used in this study. In the analysis of leaf- and root-insect populations from the same plant, different insect genotypes were found on grafted vines, with one genotype on the rootstock and one on the V. vinifera (cultivar scion). For Uruguayan leaf-galling insect populations, the average number of alleles per locus was 4.25. Genetic variance found among individuals within populations was 88% (SE = 2.298, p < 0.001), and 12% between populations (SE = 0.319, p < 0.001). An FST of 0.211 (p < 0.001) suggests limited genetic flow among populations. Significant deviation from Hardy-Weinberg equilibrium detected for the loci analyzed and the negative FIS values together suggest that parthenogenesis could be the reproductive mode. Genetic diversity found in this work shows considerable potential for host adaptation to environmental variability.
Grape phylloxera, Daktulosphaira vitifoliae (Homoptera: Phylloxeridae; D. vitifoliae) is a tiny insect (1.5 mm long) first described by Asa Fitch (1856) on North American Vitis species. For over 100 years, it has been considered one of the world’s most important vineyard pests after it spread to Europe and began feeding on roots on the highly susceptible European grapevine (Vitis vinifera). The insect spread rapidly and soon destroyed the majority of European vineyards (Granett et al. 2001). By the end of the 19th century, grape phylloxera had arrived in Brazil, Argentina, and Peru (Botton and Walker 2009, Gironés de Sanchez 2007, Huertas 2004); in 1888, it was found in Uruguay on V. vinifera plants coming from Europe (Alvarez 1909). The current distribution of the pest includes almost all viticulture regions of the world, with the exception of Chile and part of Australia. In South America, grape phylloxera is found in Venezuela, Colombia, Bolivia, Peru, Argentina, Brazil, and Uruguay (CABI 2013).
When breeders realized that roots of American Vitis species were not severely damaged by phylloxera feeding, they began to use these resistant species to create rootstocks. Because these rootstocks have proven to be durably resistant, research support for studies on the biology, ecology, and control of this important pest has diminished (Granett et al. 2001).
Grape phylloxera feeding induces galls on roots and leaves. Leaf galls are pocket-like cavities surrounded by trichomes in which the insect feeds and lays eggs. Leaf galls are commonly seen on American Vitis species’ leaves but are less common on V. vinifera leaves. Root galls on the tips of feeder roots (nodosities) occur on European and American Vitis species. Galls can also form on mature storage roots (tuberosities), but these galls normally occur only on the roots of V. vinifera or hybrids with V. vinifera (Galet 1982). When tuberosity galls swell and crack, root-rotting pathogens enter and damage large portions of the roots, which eventually results in vine death (Granett et al. 2001).
The phylloxera life cycle is both complex and not well understood; it includes sexual and asexual, winged and wingless reproductive forms. Under favorable conditions in spring and summer, successive generations of wingless parthenogenetic individuals appear in root and leaves. The winged individuals appear under certain conditions, generally from midsummer to autumn, and they can asexually produce males and females that mate and originate overwintering eggs. The eggs hatch in spring, and when the first female asexual crawlers (instars) mature, they recommence the asexual parthenogenetic reproduction cycle. In regions where parthenogenetic reproduction predominates and sexual forms are rare or absent, phylloxera overwinter as first instars on the roots (Granett et al. 2001, Powell et al. 2013). In the southwestern United States, Downie and Granett (1998) documented a variation in the life cycle where only the leaf-galling portion of the life cycle exists on V. arizonica; winged forms were not found. In Europe, winged forms have been found in vineyards and sexual individuals have been observed in laboratory situations, but the completion of the sexual cycle in the vineyard has not been confirmed (Forneck et al. 2001).
Sexual and asexual grape phylloxera lifecycles are completed on roots and leaf galls from the same plant (Vorwerk and Forneck 2007). Feeding by the first instars initiates galls on the leaves, and phylloxera soon begin producing eggs by apomictic parthenogenesis. All offspring are genetically identical except for changes caused by mutations, chromosomal rearrangements, and rare mitotic recombination events (Hales et al. 2002). Worldwide, the predominant form of reproduction occurs through parthenogenesis (Corrie et al. 2002, Granett et al. 2001); however, population genetics data strongly suggest that rare sexual events do occur (Forneck et al. 2001, Islam et al. 2013, Vorwerk and Forneck 2007). In Uruguay, the winged sexual form has been detected only in laboratory conditions, and all reproduction seems to be parthenogenetic (Scatoni et al. 1981).
Leaf galling on V. vinifera is normally absent, but limited amounts of leaf galling have been observed in Italy, France, New York, Hungary (Granett et al. 2001, Molnar et al. 2009), Austria (Könnecke et al. 2011), and more recently in Panama (Quirós et al. 2009). D. vitifoliae infestations on V. vinifera leaves have been detected in Italy (Crovetti and Rossi 1989), Brazil (Botton and Walker 2009), Germany (Kopff 2000), and Peru (Walker personal communication).
Although gall formation on V. vinifera leaves in Uruguay has been rare, new plantations have recently suffered D. vitifoliae leaf infestations in very high densities and at many locations (Vidart et al. 2013). Uruguayan vineyards are primarily planted with certified, high-quality stock because of the Vineyard Recovery Plan (VRP), which made it possible to remove old vineyards and replant them with certified grapevines imported from European nurseries. This program has replanted more than 1,700 hectares (21% of the total area) in Uruguay (Macagno 2006).
The unexpected presence of high densities of leaf galls on V. vinifera in Uruguayan vineyards could be explained by three phenomena: (a) selection pressure for more-aggressive native strains in the new context of more vigorous and healthy plants replacing old vineyards (Kimberling et al. 1990); (b) loss of resistance in plants from plant improvement programs, which may have altered the plants’ polyphenol composition or other innate resistance mechanism; or (c) importation of exotic strains from other countries (Granett et al. 1996). A first step in determining which of these phenomena may be responsible for leaf galling would be a thorough evaluation of phylloxera’s genetic diversity. The aims of our research were to evaluate the genetic diversity of leaf-galling populations of Uruguayan phylloxera (gallicole form); to estimate genetic distances among these populations; and to compare Uruguayan and foreign phylloxera populations (Brazilian, Peruvian, and European) by estimating Nei’s genetic distances. We also performed a descriptive genetic analysis comparing populations of leaf- and root-galling forms from the same individual plant in Uruguay.
Materials and Methods
Sample collection.
Phylloxera samples were collected across Uruguay from all vineyards and nurseries in which leaf galling on V. vinifera was detected. The Uruguayan regions of Durazno (Du), Florida (Fl), Canelones (Ca), Colonia (Co), San José (Sj), and Montevideo (Mv) were sampled between February 2005 and March 2006 (growing season) (Figure 1, Table 1). Each vineyard or nursery was considered a sample site, and each site was managed in a homogenous manner. At each site, leaves were randomly selected. Individual galls on a leaf were chosen, making sure that they could be individually distinguished from other galls on the same leaf. Eggs coming from a unique gall were removed with a brush to ensure that all offspring from a single plant originated from a single female. Eggs were stored at −20°C until processed. Foreign leaf-gall samples were obtained from Brazil (BR; provided by Marcos Botton: EMBRAPA, CNPUV), Peru (PE; provided by Juan Carlos Brignardello, VITICOLA S.A.), and Europe (EU; provided by Walker’s laboratory).
Roots were sampled for phylloxera at all sites where leaf galls were collected, but it was difficult to obtain enough eggs from root galls to allow analysis. Root-gall samples from Ca (n = 4), Fl (n = 2), and Du (n = 2) were successfully collected from plants that were also sampled for leaf-gall phylloxera (Table 1). Because of the small number of root samples, only a descriptive analysis was performed to compare root and leaf phylloxera from the same plants. Leaf and root samples were processed following the same methodology.
Microsatellite analysis.
Eggs were ground in 1 μL of ultrapure water per egg. The mix was centrifuged at 6000 × g for 1 min. Primers DVIT1, DVIT2, DVSSR3, and DVSSR4 (Table 1) were tested in 20 μL polymerase chain reactions (PCR) containing 2 μL of the egg mixture, 1.5 mM MgCl2, 0.2 mM dNTP (Applied Biosystems, Foster City, CA), 0.2U Taq DNA polymerase (Applied Biosystems), 1X reaction buffer, and 0.5 μm of simple sequence repeat (SSR) primers. DVIT1 and DVIT2 primers were included because, at the beginning of the research, DVSSR7, DVSSR9, DVSSR16, and DVSSR17 resulted in monomorphic allelic profiles for the populations analyzed. PCR reactions were conducted in a Thermo Scientific Hybaid PX2 thermal cycler (Cole-Parmer, Vernon Hills, IL). The following temperature profiles were used: denaturation at 94°C for 5 min, followed by 40 cycles of 94°C for 45 sec, 58°C for 30 sec, 72°C for 1 min, and a final extension at 72°C for 10 min. Annealing temperatures were 58°C for DVSSR 3 and DVSSR4, 55°C for DVIT1, and 52°C for DVIT2. DVIT1 and DVIT2 were altered with Primer3 software (http://simgene.com/Primer3) to improve annealing temperatures (Table 2). PCR products were mixed with sample-loading dye (10 mM NaOH, 95% formamide, 0.05% bromophenol blue, and 0.05% xylene cyanol) at a 2:1 ratio. A 4 μL aliquot of this mixture was resolved on a 5% polyacrylamide, 7 M urea sequencing gel. Gels were run at a constant 55 W, 1,650 V for 35 min to 2 hr, depending on amplicon size. A 10 bp molecular-size marker (Invitrogen, Carlsbad, CA) was included in every run. Silver staining (Promega Biosciences, Madison, WI) was used to visualize gels and record polymorphisms among samples.
Genetic analysis.
Characterization of grape phylloxera populations was conducted with primers DVSSR3, DVSSR4 (Lin et al. 2006), DVIT1, and DVIT2 (Corrie et al. 2002). Allele sizes at each locus were assigned using the data obtained from the polyacrylamide gels, considering existing alleles for each primer amplification product (available at NCBI, http://www.ncbi.nlm.nih.gov/nucleotide/?term=daktulosphaira+vitifoliae). Analysis of molecular variance (AMOVA), principal coordinates analysis (PCoA), allele frequencies, number of alleles per locus, observed heterozygosity (Ho), expected heterozygosity (He), and deviations from Hardy-Weinberg equilibrium were estimated using GenAlex V6 (Peakall and Smouse 2006). A dendrogram was constructed with genetic distance analysis software (Lewis and Zaykin 2002) using Nei’s genetic distance and co-ancestry coefficients. Results were presented as graphical clusters using TreeView 1.6.6 (Page 1996). To test for significant isolation by distance, a Mantel test was carried out between genotypic-distance matrix (Nei’s distance) and geographic-distance (km) matrix. This test was also performed using the GenAlex V6.
Results
Polymorphism in Uruguayan populations.
The four primers used in our study were polymorphic. The average number of alleles per locus was 4.25, ranging from 3 to 7 alleles per locus (Table 3). For DVSSR3, the predominant allele size was 240 bp (allelic frequency, 0.65); for DVSSR4, it was 254 bp (allelic frequency, 0.53). For DVIT1, the predominant allele size was 180 bp (allelic frequency, 0.59), and for DVIT2, it was 150 bp (allelic frequency, 0.76). There was a significant deviation from the Hardy-Weinberg equilibrium (p < 0.05) for the gallicole Uruguayan populations. For DVSSR3, the allele of 278 bp was found only in two samples from Sj2.
Genetic diversity in Uruguayan populations.
A dendrogram using Nei’s genetic distances among populations shows that the samples clustered according to geographic location, except for Co3 which, although geographically distant, was genetically close to the Mv population (Table 4, Figure 2A). Eighty-eight percent of the genetic variance was found among individuals within populations (df = 56, SS = 128.69, SE = 2.298%, p < 0.001, FST = 0.211, p < 0.001; Table 5), while 12% was found among populations (df = 12, SS = 47.21, SE = 0.319%, p < 0.001). An FST value of 0.211 (p < 0.001) suggests that gene flow is limited between populations (Table 2). The Mantel test did not reveal a meaningful association between genetic and geographical distance (r = 0.096, p = 0.017). Some genotypes were restricted to certain vineyards, while others were more evenly distributed. Genetic relationships among samples evaluated through UPGMA (Unweighted Pair Group Method with Arithmetic Mean) cluster analysis revealed that clustering of samples was not associated with the geographic location of the samples (Figure 2B). Additionally, there were only two clusters that grouped samples by plant host (V. vinifera or rootstocks derived from American Vitis species).
Leaf and root populations.
There were differences between root and leaf populations from the same grafted plants. However, Nei’s genetic distances were always smaller between root and leaf samples from the same plant than between different sample locations (Supplemental Figure 1, Table 5).
Genetic distance between Uruguayan, European, Peruvian, and Brazilian populations.
Nei’s genetic distances among foreign and Uruguayan populations were calculated, and PCoA analysis was carried out. The Peruvian populations were the most genetically distant, whereas European and Brazilian populations were in the same quadrant as the Co and Sj populations (Figure 3).
Discussion
New plantations in Uruguay have recently suffered D. vitifoliae infestations on V. vinifera leaves in very high densities and at various locations (Vidart et al. 2013). The cause of this change in insect behavior is unknown. This infestation was also coincidental with the VRP, which resulted in vine removal and replanting of nearly 1,700 hectares (Macagno 2006) and required seven million grafted plants to be imported from European nurseries (INAVI, unpublished data). These vines could have resulted in the importation of an exotic phylloxera strain or increased selective pressure for a more aggressive native strain. Alternately, the imported vines may have simply not been resistant to the existing native strains of phylloxera.
Microsatellite or SSR markers have been intensively used for various applications in many different species (Wang et al. 2009). This molecular tool has been extensively used to study the genetic diversity of Aphidoidea insects (Islam et al. 2013, Sandrock et al. 2011), and has overturned many classical predictions about patterns of genotypic diversity and heterozygosity (Wilson et al. 2002).
Genetic differences and structure of phylloxera populations in California (Lin et al. 2006, Islam et al. 2013) and Australia (Corrie et al. 2004) have been studied with microsatellites. We performed microsatellite analysis in this study at four loci (DVSSR3, DVSSR4, DVIT1, and DVIT2), all of which detected polymorphisms among the Uruguayan grape phylloxera populations. Initially, we used eight loci, but discarded four because they produced monomorphic data for the populations we analyzed. Although these results should be analyzed with caution, other studies have achieved interesting results with this number of microsatellites (Corrie et al. 2002, Corrie and Hoffmann 2004).
An FST value of 0.211 suggests limited genetic flow among populations. The significant deviations from Hardy-Weinberg equilibrium detected for all the loci we analyzed, together with negative FIS values, suggest that the major reproductive mode is through parthenogenesis. However, the detection of high genotypic diversity and the occurrence of unique phylloxera genotypes (e.g., the Sj2 population) indicate that rare sexual recombination events occur. Similar findings were detected in Australian, Californian, and European grape phylloxera populations (Corrie et al. 2002, Islam et al. 2013, Lin et al. 2006, Vorwek and Forneck 2007).
Given phylloxera’s restricted mobility, the clustering we observed in the dendrogram within geographically close populations could be explained by geographic isolation. Samples from Ca grouped into two close clades, whereas samples from Sj and Co (except for Co3) grouped together. Our result is in accordance with geographic distances among sample sites, as Co and Sj are relatively close viticultural regions. The Co3 population came from an old planting of the rootstock Rupestris du Lot (more than 35 years old, not replanted during VRP) and was genetically closer to the Mv population, which came from an old American hybrid vineyard that had been isolated for more than twenty years. Given geographic distance, isolation, and different hosts, we expected Co3 and Mv to be more genetically distant. However, our cluster analysis among samples revealed high diversity within geographic locations, which was also detected with AMOVA (88% variation within populations). The high diversity we observed could be explained partly by multiple plant introductions (Granett et al. 1996) capable of shaping the genetic structure of populations. Future studies should conduct genetic analysis within geographically isolated vineyards (with isolated phylloxera populations) to evaluate the impact of introduced plant material on the evolution of populations and on phylloxera’s reproductive mode.
Our PCoA analysis indicated that the Peruvian populations were the most genetically distant, yet European and Brazilian populations appeared to be very similar and grouped in the same quadrant as the Co and Sj populations. The Peruvian samples were clearly distinct from the Uruguayan samples, and there was no evidence of movement of phylloxera from Peru to Uruguay. The geographic proximity of Uruguay to Brazil coincides with the genetic similarity of the phylloxera populations. Although the flow of plants between these countries is regulated, the genetic similarity between Uruguayan and Brazilian populations, given the limited natural movement of the insect, suggests that either both countries received infested plant material from the same source, or introduction and subsequent movement to the adjacent country occurred. The clustering of these phylloxera populations with the European populations supports introduction from a European source.
We also compared leaf- and root-phylloxera populations taken from the same host plant, and we detected differences in five of the eight pairs of samples; these differences were on V. vinifera grafted on American Vitis species rootstock (Ca and Fl populations). No such differences were found in un-grafted rootstock mother vines or French/American hybrids. We sampled the leaf- and root-gall phylloxera from a homogeneous single-host environment, whereas the paired root- and leaf-gall samples with genetic differences were from grapevines with two host tissues: a V. vinifera scion for the leaf galls and an American Vitis species rootstock for the root galls. Although these observations are consistent with those reported by Corrie et al. (2003), more samples need to be studied so that these host-based associations can be better detailed and the cause or source of these differences determined.
Using microsatellite DNA markers, Corrie et al. (2003) detected strong associations between D. vitifoliae asexual lineages and vine-host type within a vineyard. They also used excised root bioassays to show host-specific differences in life table parameters of reproductive rate and intrinsic rate of increase. A discontinuous plant host, such as a grafted vine, could favor population differentiation within the same plant. Since movement within a plant is not physically restricted, the presence of host-associated genotypes suggests that certain genotypes have an adaptive advantage and provides evidence of grape phylloxera host adaptation.
We found the leaf-galling form on all of the V. vinifera cultivars we sampled, with the exception of Tannat, a high-tannin grape variety commonly grown in Uruguay. Further work is needed to understand the lack of leaf galling on Tannat, and whether it is due to host preference or Uruguayan grape phylloxera’s restricted genetic diversity.
The primers employed in this study (DVSSR3, DVSSR4, DVIT1, and DVIT2) were able to differentiate the analyzed populations. We observed variability between individuals from the same population, which is in accordance with results obtained for grape phylloxera with random amplified polymorphic DNA (Downie 2000) and SSR (Islam et al. 2013) markers, both of which detected differences in samples that were adjacent to each other. Detection of multiple phylloxera genotypes suggests that the colonization of Uruguay’s vineyards was not a single unique event, but rather that phylloxera were imported on multiple occasions.
Conclusions
Microsatellite analysis of gallicole grape phylloxera populations detected high genetic diversity in Uruguayan vineyards. The leaf and root samples we analyzed from the same grafted plants in Ca, Fl, and Du were genetically different. When compared with international samples, the Peruvian populations were the most genetically distant from the Uruguayan samples. Several lines of evidence, including an FST value of 0.211, negative FIS values, deviation from Hardy-Weinberg equilibrium, and the presence of multicopy genotypes, suggest that parthenogenesis is the principal mode of reproduction. The facts that phylloxera have a high parthenogenetic reproductive capacity and the potential to move on planting materials, vineyard equipment, and through the air, have important implications for management strategies. These biological and cultural factors in conjunction with the recent VRP, which has introduced imported grafted grapevines with greater vigor into places where older plants were uprooted, could have promoted the appearance and selection of host-adapted populations in Uruguay, and may have increased the risk of grape phylloxera outbreaks.
Acknowledgments
Acknowledgments: The authors gratefully acknowledge INAVI and CSIC for financial support: CSIC SP “Manejo de plagas de la Viña,” and the INIA Las Brujas, vine growers and nurseries for permission to sample from their vines. We also thank Valeria Vidart and Valentina Mujica who kindly provided the samples from Canelones and Florida.
Footnotes
-
Supplemental data is freely available with the online version of this article at www.ajevonline.org.
- Received February 2014.
- Revision received August 2014.
- Accepted September 2014.
- Published online December 1969
- ©2015 by the American Society for Enology and Viticulture