Resistance to thrips (Enneothrips flavens) in wild and amphidiploid Arachis species

Thrips (Enneothrips flavens) is a pest that causes severe damage and yield losses to peanut crop if not properly controlled. The main control method currently used by farmers is bi-weekly application of insecticides during crop development, which, in addition to its toxicity, is very costly. Thus, new sources of resistance must be identified in order to reduce the use of insecticides and effectively manage the pest. This study aimed to evaluate the occurrence and symptoms of E. flavens infestations in 12 accessions of 10 wild species of Arachis and nine amphidiploids, as well as to compare their morphoagronomic characteristics to those of commercial cultivars. To this end, we conducted experiments during two summer seasons, using a randomized block design with four replications. We conducted evaluations of the severity of infestation, noting visual symptoms of E. flavens and morphological and reproductive characteristics of the Arachis plants. Results indicated that wild accessions V 7635 (A. vallsii), V 13250 (A. kempff-mercadoi), K 9484 (A. batizocoi), Wi 1118 (A. williamsii), V 14167 (A. duranensis) and V 13751 (A. magna) are the most promising for obtaining useful new amphidiploids. Among the amphidiploids, An 12 (A. batizocoi x A. kempff-mercadoi)4x, An 9 (A. gregoryi x A. stenosperma) 4x, and An 8 (A. magna x A. cardenasii)4x showed high level of resistance to E. flavens. The identified thrips resistant wild and amphidiploid Arachis species may be used in future breeding program to produce thrips resistant peanut cultivars.


Introduction
The thrips Enneothrips flavens Moulton (Thysanoptera: Thripidae) is considered a major pest of peanut due to its widespread occurrence and high population levels [1]. The main control method currently used by farmers is bi-weekly application of insecticides during crop development. During infestations, E. flavens lodges within buds and developing shoots, damaging the vegetative growth of plants. Because of its mode of attack, the chemical products required for its control have compositions and rates of application that make them costly. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 The use of cultivars with resistance to E. flavens could lead to gains in productivity or could promote a significant reduction in production costs by reducing or eliminating chemical control [2]. Thus, the search for resistant varieties is very important for the genetic improvement of the species. In Brazil, studies searching for resistance have been conducted with genotypes of the cultivated peanut, Arachis hypogaea L. [3][4][5][6][7], and the form of resistance known as "tolerance" was observed in some cultivars. However, the degree of resistance seen in many cultivars is considered too small in order to reduce or eliminate chemical control.
An alternative approach is the incorporation into breeding programs of wild Arachis species that present great genetic variability and are a potential source of genes for important agronomic traits such as resistance to diseases and pests [8][9][10][11][12][13]. However, the potential of this variability in wild species of Arachis has not been widely studied. The main obstacle to such studies is that the vast majority of wild species are diploid, while the cultivated species (A. hypogaea) is allotetraploid. The barrier of ploidy, as well as other factors such as genomic differences between species, precludes hybridization between wild species and A. hypogaea [14]. Obtaining fertile amphidiploid individuals by using colchicine to double chromosomes is a viable route for the introgression of wild Arachis spp. genes into A. hypogaea, as it is possible to produce amphidiploids with the same or similar genomic constitution as the cultivated species [15].
To effectively identify resistant plant material, assessments done under field conditions in locations near commercial plantations are required in order to select accessions and amphidiploids that could be crossed with A. hypogaea. Therefore, this study aimed to evaluate the occurrence and symptoms of E. flavens infestations in the wild and amphidiploid species of Arachis as well as to compare their morpho-agronomic characteristics to those of commercial cultivars.

Materials and methods
The experiments were performed during two summer seasons, 2011/2012 and 2012/2013, in an experimental area of the Agência Paulista de Tecnologia dos Agronegócios (APTA; São Paulo Agency for Agribusiness Technology), Polo Centro Norte, Pindorama, São Paulo, Brazil.
In the first season (2011/2012), the experiment consisted of 24 treatments (genotypes), comprising 12 accessions of 10 wild species of Arachis, nine amphidiploids, and three genotypes of A. hypogaea (including two commercial cultivars used as controls; Table 1). In the second season, two amphidiploids were not evaluated due to an insufficient number of seeds.
The seeds used in the experiments were pretreated with the commercial products Ethrel 1 (active ingredient: ethephon) at a dose of 2.0 mL/kg seed in order to break seed dormancy, and the fungicide Vitavax 1 -Thiram 200 SC (carboxin + thiram) at a dose of 2.5 mL/kg seed to protect against soil fungi. Seeds were germinated in 200-mL plastic cups containing soil substrate and manure (3:1) and placed in the greenhouse. When plants reached a height of approximately 15 cm, they were transferred to the field.
In both trials, we used a randomized block with four replications. Each plot consisted of a row containing five plants spaced one meter apart, with 1.80 meters between rows.
Plants were sprayed every 15 days with the fungicide chloratalonil (Bravonil 500 1 ) at a dose of 1.75 L ha -1 per application, alone or in an admixture with triazoles (Score 1 ) at 0.2 L ha -1 or with strobilurins (Opera 1 ) at 0.6 L ha - Weeds were controlled with pre-plant-incorporated herbicide application of trifluralin (Trifluralin Nortox 1 ) at a dose of 2.5 L ha -1 . Manual hoeing was performed whenever necessary during plant development.

Infestation and visual symptoms of E. flavens
During the experiments, we conducted seven (between 35 to 95 DAP) and eight (between 45 to 110 DAP) evaluations in the 2011/2012 and 2012/2013 seasons, respectively, to determine the number of insects (adults and nymphs) in young leaflets that were still closed, examining 10 leaves per plot.
We also evaluated the symptoms of thrips infestation by assigning visual symptom scores to 10 leaflets on a scale from 1 (leaf with no symptoms) to 5 (completely infested leaves), based on previous scales [16,17].
After graphical analysis, we selected evaluations that showed the highest number of insects per leaflet or the highest visual symptom scores and performed an analysis using both the selected evaluations and the mean of all evaluations conducted during the growing season for number of insects and symptom score. Scores were transformed into (x + 0.5) 1/2 , submitted to analysis of variance by F test, and means were compared by Tukey's test at 5% probability. In order to rank the genotypes by resistance, we created an attack index (AI) by multiplying the average number of E. flavens by the average symptom score. More resistant genotypes had lower AI values.
We conducted principal component analysis using SAS 1 [18] and Microsoft Excel. The variables analyzed included the number of thrips per 10 leaflets, the visual symptom score, and the attack index, using the mean values from both the seasons.
The values were submitted to analysis of variance by F test, and means were compared with the Scott-Knott test at 5% error probability.

Infestation and visual symptom scores of E. flavens
In both years, the insect population fluctuated unevenly over the course of the experiments, with some peaks of higher incidence (Figs 1 and 2). Visual symptom scores showed similar variation. In the 2011/2012 growing season, the level of thrips infestation was highest 85 days after planting (DAP), when a mean of approximately 8.0 thrips/10 leaflets was observed ( Fig  1). The highest visual symptom score (2.5) was recorded 45 DAP.
In the 2012/2013 growing season, the highest number of thrips was observed 62 DAP, which had a mean similar to that of the previous year (8.0 thrips/10 leaflets; Fig 2). The highest visual symptom score occurred 77 DAP, with a mean score that was also similar to that of the previous season (2.5). The observed differences between years may be attributable to the effect of weather conditions on the field experiments.
When averaged over the seven evaluations conducted during the 2011/2012 growing season, the highest levels of infestation (> 10 thrips/10 leaflets) were observed in the control  Table 2). The number of thrips per 10 leaflets at 85DAP had a high CV because the quantitative behavior of this characteristic.
The lowest levels of thrips infestation and lowest AI values observed during the 2011/2012 experiment were among the parental lines. Accession A. vallsii (V 7635) had the lowest values of the two measures (1.6 thrips/10 leaflets and AI 1.98). The highest level of infestation among the parents was found in Arachis gregoryi (V 6389), which had 7.6 thrips/10 leaflets and an AI of 24.02, surpassing the levels observed for the amphidiploids, which is in line with the great genetic variability found in wild species of the genus Arachis (Table 2, S1 Table). Relative to their parents, amphidiploids had AI values that were either intermediate, close to that of one parent, or higher than that of both parents. No amphidiploid showed resistance greater than both parents.
Generally, the number of thrips observed was positively correlated with the amount of damage to the plant. However, this was not true for the parents A. gregoryi (V 6389) and A. duranensis (V 14167), which had high visual symptom scores despite intermediate thrips values (Table 2).
In  (Table 3). As in the previous year, presence of the pest was generally well correlated with plant damage as measured by the visual symptom score, although again there were exceptions. The amphidiploid An 7 (A. vallsii x A. williamsii) 4x , for example, had one of the lowest pest levels but suffered relatively high damage. Similarly, the IAC Caiapó cultivar had low pest incidence but a high visual symptom score ( Table 3). The IAC Caiapó cultivar is resistant to thrips, as its productivity is less affected by thrips attack in comparison to other cultivars [4].
When averaged over all eight evaluations conducted during the 2012/2013 growing season, accession GKP 10017 (A. cardenasii) showed the least amount of thrips damage. Among the amphidiploids, we found that An 8 (A. magna x A. cardenasii) 4x , An 9 (A. gregoryi x A. stenosperma) 4x and An 2 (A. gregoryi x A. linearifolia) 4x were the least damaged ( Table 3).
The  Table 3). Once again, when compared to their parents, amphidiploids had AI values that were either intermediate, close to that of one parent, or higher than that of both parents. No amphidiploids showed greater resistance than both parents. Fig 3 shows the dispersion of the studied genotypes with regard to thrips resistance. The two-season results are based on the average number of thrips per 10 leaflets, the visual symptom score, and the attack index. Principal component analysis revealed that all three variables were correlated. The variable that best explained the dispersion was the attack index, followed by the number of thrips per 10 leaflets. The most susceptible genotypes were grouped (IAC 503, IAC Caiapó, and V 12549), while wild accessions and amphidiploids were dispersed on the left side of the graph. Furthermore, it can be seen that the amphidiploids are not necessarily near the parental accessions, and can be between the two parents or even far away from them on the graph. This shows that resistance to thrips is observed both in diploid species and in diploid and tetraploid synthetic hybrids. It should be noted that the less resistant wild and amphidiploid genotypes (A. gregoryi V 6389, An 4, and An 10) in the analysis of variance were also closer to the controls (A. hypogaea). It is noteworthy that the An 4 amphidiploid, which is closest to the grouping of A. hypogaea, was obtained from the cross between A. ipaënsis and A. duranensis, the leading candidates for ancestors of the cultivated peanut [19].

Morphoagronomic traits
The morphological and reproductive data from the 2011/2012 season are presented in Table 4.
Only the amphidiploid An 6 (A. batizocoi x A. cardenasii) 4x exhibited a main stem height similar to that of the A. hypogaea controls. For average branch length and average number of branch nodes, no amphidiploid resembled the controls, whereas for length between nodes, only amphidiploids An 9 (A. gregoryi x A. stenosperma) 4x and An 11 (A. vallsii x A. stenosperma) 4x had values similar to those of the controls. With regard to reproductive traits, it was found that the accessions of wild species and the amphidiploids were less productive than the A. hypogaea controls, with the exception of accessions V 13250 (A. kempff-mercadoi) and V 15076 (A. stenosperma) and amphidiploids An 6 (A. batizocoi x A. cardenasii) 4x and An 9 (A. gregoryi x A. stenosperma) 4x (Table 4).
In the 2012/2013 season, we chose to evaluate only the amphidiploids (Table 5). Amphidiploids An 8 (A. magna x A. cardenasii) 4x and An 6 (A. batizocoi x A. cardenasii) 4x had the shortest main stem heights. Number of primary branches and average number of branch nodes (data not shown) did not differ between genotypes. The amphidiploid with the shortest average branch length was An 7 (A. vallsii x A. williamsii) 4x with a value of 61.9 cm, which was nearly twice as long as that of the control genotypes. Reproductive traits followed the same trend observed in the previous season. The number of seeds per plant (NSP) had a high CV because the quantitative behavior of this characteristic.

Discussion
Thrips infestations of runner peanut cultivars were most severe between 15 and 54 DAP [7]. Lourenção et al. [5], who evaluated cultivars treated with chemical insecticides to control thrips, reported that regardless of cultivar, the most severe infestation was observed at the third evaluation, conducted at 56 days after seeding. Thereafter, there was a decreasing trend in the number of thrips per leaflet until the end of the plant cycle. In contrast, the most severe infestations of wild species in the current study occurred later, at 85 and 62 DAP in the 2011/ 2012 and 2012/2013 growing seasons, respectively. Rainfall is one of the factors that can reduce the occurrence of thrips by washing off and drowning individuals and by ensuring moisture levels that are favorable for microorganisms that cause morbidity and mortality in this insect pest [20].
Previous studies of E. flavens infestation of wild Arachis species and peanut cultivars have been conducted with the former also evaluating two synthetic amphidiploids that were included in the present study, An 2 -(A. gregoryi x A. linearifolia) 4x and An 4 -(A. ipaensis x A. duranensis) 4x [10,17]. Both amphidiploids were more resistant to thrips when compared to the evaluated A. hypogaea genotypes [10]. These authors emphasized the higher incidence of insects in commercial cultivars and suggested that due to their wide genetic variability, wild species are a promising source of resistance gene transfer to the cultivated peanut via synthetic amphidiploids of wild species. Runner IAC 886, IAC Caiapó and IAC 503 were the cultivars with the highest infestation rates and symptom scores. Accessions GKP 10017 (A. cardenasii) and V 7639 (A. kuhlmannii) exhibited similar levels of resistance to both E. flavens and the red-necked peanut worm, Stegasta bosquella (Chambers) [17].
In the current study, the amphidiploids An 12 (A. batizocoi x A. kempff-mercadoi) 4x , An 8 (A. magna x A. cardenasii) 4x and An 11 (A. vallsii x A. stenosperma) 4x had the lowest attack indexes. Studies have highlighted the importance of wild species as a source of pest resistance [21,17]; however, this is the first report of thrip resistance in an amphidiploid peanut species.
It is interesting to note that the amphidiploid An 8 stood out in terms of low insect incidence and visual symptom scores over both years of the study, reinforcing the idea that incorporating resistance to E. flavens (which is known to infest wild species of the genus Arachis) can be achieved using synthetic amphidiploids in future crosses with A. hypogaea breeding lines. It should be noted that there are great differences between the amphidiploids as well as between the amphidiploids and their parents. The evaluation of accessions in terms of resistance to thrips is a good indication of how they will behave when the amphidiploids are obtained. However, assessment of these synthetic hybrids is absolutely essential for the confirmation of their behavior after crossing two different genomes. Thus, in addition to obtaining a large number of amphidiploids, one must select those that are most resistant, which will allow gene pyramiding and more horizontal resistance. Agronomic characterization is of great importance in studies involving synthetic amphidiploids and accessions of wild species. Although amphidiploids are currently the most feasible route for the introgression of resistance genes from wild species to the cultivated peanut, these amphidiploids closely resemble the wild species in terms of architecture and seed production.
Currently, there are no commercial peanut cultivars that have been genetically modified for pest resistance, partly because the costs associated with regulation and release are extremely high for species considered to be "minor crops," and partly because, to date, the efficiency of genetic techniques for controlling peanut pests is not as high as that for other crops such as corn, soybean, and cotton [22]. Thus, the search for sources of resistance to pests and diseases in wild peanut species remains a useful alternative for reducing the use of pesticides and consequently production costs.
Our findings support the use of amphidiploids in peanut breeding programs. Backcrossing of the best amphidiploids with A. hypogaea elite lines followed by selection of thrips resistant progeny can potentially lead to the development of new thrips resistant peanut cultivars based on genes found in wild Arachis species.
To conclude, accessions of wild Arachis species with the lowest levels of E. flavens infestation over both years of the study were: V7635 (A.  V14167 (A. duranensis) and V13751 (A. magna). These six accessions are therefore the most promising for the production of new amphidiploids. The tested amphidiploids with the lowest levels of E. flavens infestation were An 12 (A. batizocoi x A. kempff-mercadoi) 4x , An 9 (A. gregoryi x A. stenosperma) 4x and An 8 (A. magna x A. cardenasii) 4x . The latter has high level of resistance to E. flavens and should therefore be used in future crosses.
Amphidiploids resemble wild species in terms of their architecture and reproductive characteristics.
Supporting information S1 Table. Raw data of nine amphidiploids and 15 accessions of Arachis evaluated during two summer seasons for resistance to thrips and by morphological and reproductive traits. (PDF)