Characterization of diploid Arachis interspecific hybrids for pest resistance

In Brazil, the thrips (Enneothrips flavens Moulton) and rednecked peanutworm (Stegasta bosquella Chambers) are considered key pests for the peanut crop. Wild species show resistance to both of these pests, and can be used in breeding programs. The production of the sterile interspecific hybrids is necessary, which could be colchicine treated to get a synthetic amphidiploid with the same or similar genomic configuration of cultivated peanut. In this context, this study proposed the hybridization of wild pest-resistant species in 18 distinct combinations, obtaining the interspecific hybrids of Arachis and completing their characterization by (i) the reproductive characterization through pollen stainability with 2% acetocarmine (AC) solution with glycerin and 0.25% tetrazolium PLANT PROTECTION Article Characterization of diploid Arachis interspecific hybrids for pest resistance Ailton Ferreira de Paula1*, Naiana Barbosa Dinato1, Bianca Baccili Zanotto Vigna2, Alessandra Pereira Fávero2 1.Universidade Federal de São Carlos Departamento de Morfologia e Patologia São Carlos (SP), Brazil. 2.Embrapa Pecuária Sudeste São Carlos (SP), Brazil. *Corresponding author: ailtonfp_biolog@yahoo.com.br Received: Apr. 16, 2018 – Accepted: Aug. 1, 2018 solution (TZ), (ii) the molecular certification of hybridization using microsatellite markers, and (iii) the morphological characterization using 61 morphological traits with Principal Component Analysis. Using reproductive, morphological and molecular characterizations, we identified six hybrid plants of the following crosses: three from A. magna V 13751 × A. kuhlmannii V 9243, two from A. magna V 13751 × A. kempff-mercadoi V 13250, and one from A. magna K 30097 × A. kuhlmannii V 7639. Amphidiploids can be reached based on these diploid plants and they can be used in breeding programs aiming pest resistance introgression.


INTRODUCTION
Peanut (Arachis hypogaea L.) is an oilseed crop and due to its flavor and high oil content, it is used in food industries and has potential for biofuel production (Santos et al. 2005).There are many pests that limit the production of peanut, however, thrips and rednecked peanutworm are currently considered the key pests limiting the production for the crop in the state of São Paulo in Brazil (Michelotto et al. 2013;2017).Thrips are small sucking insects of the order Thysanoptera, usually measuring between 0.5 and 5.0 mm in length.Most thrips feeds on plant material (phytophagous) and, when feeding, they destroy the leaflets of plants while they are still closed, and can also be vectors for other organisms (Borror and De Long 1969).The rednecked peanutworm (Stegasta bosquella) is an insect of the order Lepidoptera, whose adult usually has between 6.0 and 7.0 mm in length.The caterpillar of this insect is phytophagous attacking the leaflet of peanut while they are still closed (Borror and De Long 1969).
One way to control these pests is by obtaining resistant cultivars.However, the resistance located in A. hypogaea genotypes is partial.Several wild species of Arachis are highly resistant to pests, such as rednecked peanutworm and thrips (Janini 20113 ).Brazil is the center of origin and holder of the largest genetic diversity of the genus Arachis (Freitas et al. 2007), thus the appropriate conservation and characterization of germplasm are of great importance for the use in breeding and biotechnology of peanut.
The main barrier to breeding is that A. hypogaea is allotetraploid (4n = 40), while the vast majority of wild species of Section Arachis is diploid (2n = 20) (Fernández and Krapovickas 1994).The amphidiploid is an interspecific hybrid having a complete diploid chromosome set from each parent form and this is a viable route for introgression of genes located in wild species of Arachis.Thus, the strategy consists of crossing species of two distinct genomes (for example, A × B), obtaining sterile diploid hybrids AB (2n = 20).These hybrids should be treated with colchicine to double the chromosomes and to obtain fertile hybrids with two distinct full genomes, the AABB amphidiploid (4n = 40), which can be crossed with A. hypogaea AABB (4n = 40) (Simpson 1991).The same can be carried out using species of genome K in place of genome B. In this way, this study aimed to obtain interspecific hybrids of Arachis and characterize them by using morphological and reproductive descriptors and to certify the hybridization using molecular markers, with an objective of future introgression of resistance to two key pests in elite genotypes of peanut.

MATERIALS AND METHODS
In this study, we performed two crossings and characterization seasons besides the hybridization certification using molecular markers.
The crossings, reproductive and morphological characterization were performed in plants 90 days after germination.

Germination
Eight seeds from each parent and all progenies seeds were treated with Thiram ® and placed on soaked 0.65% Ethrel blotter paper.Germination conditions were 16 h at 20 °C in the dark and 08 h at 35 °C under fluorescent light.Seedlings were transplanted in pots with 180 mg of substrate.When the plants were vigorous they were transplanted to 25 × 40 × 40 cm pots.

Crosses
The crosses were manually performed in a greenhouse, at Embrapa Pecuária Sudeste (São Carlos, state of São Paulo, Brazil)    The success of hybridization percentage (SP) was calculated by the formula: SP = (number of hybrids/number of pollinations) × 100.
Although the combination A. magna (V 13751) × A. kempff-mercadoi (V 13250) was previously hybridized (Paula et al. 2017), the crossings were performed again aiming at the characterization and comparison to other diploid hybrids.

Reproductive characterization
Four flowers of each genotype were randomly collected.With the aid of tweezers, pollen was taken from the anthers, placed on slides and stained with 2% acetocarmine solution with glycerin (Stalker et al. 1991;Wondraceck-Lüdke et al. 2015) or 0.25% tetrazolium (Beyhan and Serdar 2008).The anthers that did not shed pollen were put on the slides with a drop of dying solution and macerated.Pollen was analyzed as stained or non-stained under the optical microscope.The criteria to consider pollen as stained and possibly viable were the good development and staining of the grains.Two hundred grains were counted from each flower, totaling 800 grains per individual.The stained pollen percentage was calculated for each sample from all evaluated plants.Analyses of variance and Tukey's test for means comparison were performed using the software Statistical Analysis System (SAS ® 9.3) (SAS Institute 2012).

Molecular certification of the hybrids
Total genomic deoxyribonucleic acid (DNA) was isolated from young leaves using the protocol based on Cetyl Trimethyl Ammonium Bromide (CTAB) described by Grattapaglia and Sederoff (1994), with the inclusion of an additional precipitation with 1.2 M NaCl, immediately after CTAB buffer.Quantification of total DNA was performed with a spectrophotometer (NanoDrop ND-1000) and DNA integrity was checked on 1% agarose gel stained with ethidium bromide.
Ten simple microsatellite markers (Table 2) were selected from the literature (Moretzsohn et al. 2013) according to the species included in the present study.The loci were amplified by polymerase chain reaction (PCR).The conditions of markers amplification were established with a temperature gradient PCR to determine the primers annealing temperature in the PCR.After setting specific temperatures for each locus, PCR reactions were performed in a thermocycler BioRad Markers that succeeded in amplification were subjected to electrophoresis in 6% polyacrylamide gels and stained with silver nitrate (Creste et al. 2001) for fragments visualization.The size of these fragments was estimated by using a 10 bp ladder (Invitrogen).From all SSR tested, we selected the markers that better amplified in the samples evaluated and that presented polymorphism for the parents used in crosses.The progenies that had an allele from the male parent not common to the female parent were considered hybrids.

Morphological characterization
Four leaves of the lateral branch, one leaf of the main axis and four flowers of all individuals were randomly collected.The first totally expanded leaves from the lateral branches and main axis were used to perform the characterization.Sixty-one morphological traits were analyzed (Fávero et al. 2015a;2015b).According to trait, the sample was measured by using a ruler or a caliper and was observed under a stereomicroscope.
The following traits were evaluated on the main axis and lateral branch: proximal leaflet length and width, distal leaflet length and width, petiole and petiolule length, lenght and width of stipule free part of stipule, lenght of stipule adnate part, trichomes on the abaxial and adaxial leaflet border, trichomes on the abaxial and adaxial leaflet center and midvein, bristles on the leaflet border, trichomes on the petiole and on the petiolule, Bristles on the petiole and on the petiolule, trichomes on the stipule (free and adnate part) center or border, bristles on the stipule (free and adnate part), anthocyanin in the stipule.The flower traits were: standard length and width, wing length and width, lower and upper lip length and hypanthium length.
Data were analyzed using Principal Component Analysis (PCA) in SAS® 9.3 software (SAS Institute 2012).The results of components 1 and 2 were multiplied by the mean values of each trait for each individual and the resulting values were used to build a biplot graph using the software Microsoft Excel®.

Hybridizations
During the hybridization season 3,140 pollinations were performed (Table 3), from which 166 pegs were generated and 49 F 1 seeds were produced (119 pegs produced no seeds).The combination A. magna K 30097 × A. kuhlmannii V 7639 received the highest number of pollinations ( 424) but produced only six pegs and four seeds.The accession A. vallsii V 7635 crossed with four distinct species showed the greatest difficulty in producing hybrids: 936 pollinations resulted in 11 pegs and five seeds.After all characterizations and certifications performed in this study, six plants were identified as a result of hybridization and the percentage of success was 0.19%.
The accession A. magna K 30097, in addition to the resistance to pests (Michelotto et al. 2017;Janini 2011), also presents disease resistance (Fávero et al. 2009), being a material of great importance for the peanut breeding program.Fávero et al. (2015a) performed crosses with four combinations involving A. magna K 30097 (A.simpsonii V 13710, A. kuhlmannii V 10506, A. kempff-mercadoi V 13250 and A. diogoi K 10602) as female parent and had difficulties to obtain seeds from crosses involving these accessions.Only one of these combinations was used in this study, A. magna K 30097 × A. kempff-mercadoi V 13250, from which we obtained nine seeds that germinated but died at seedling stage.
Wondraceck-Lüdke et al. ( 2015) obtained four hybrids from A. valida V 13514 × A. magna K 30097.But both accessions involved had genome B, unlike combinations with species genome A in this study, indicating that the genetic distance between the materials involved directly affects the production of the hybrid.
All combinations involving accessions A. krapovickasii Wi 1291 and A. vallsii V 7635 as female parent showed difficulties in obtaining seeds.Three combinations A. vallsii V 7635 × A. helodes V 6325 and A. gregory V 14760 × A. helodes Co 6862, A. kuhlmannii V 6413 produced no seeds.Several factors can interfere with the hybrids production in Arachis, as that the fertilization may not occur after pollination, or be late (pollen tube delayed development), or the proembryo doesn't grow after the peg reaches the ground or grow very slowly.Other factors that can interfere, as the bad flower manipulation technique in the emasculation and pollination process or hybrid seeds don't germinate or they germinate but without enough vigor to survive in the pots (Nigam et al. 1990;Tallury et al. 1995).
Moreover, three (A.magna V 13751 × A. kuhlmannii V 9243) and two (A.magna V 13751 × A. kempff-mercadoi V 13250) hybrid plants had no dormancy and germinated in pots before harvesting.These were transplanted into other pots to develop.
Out of the 49 seeds of the progenies produced from crosses, only nine germinated and survived in pots.Besides these nine seedlings, the five mentioned above were incorporated, which had spontaneous germination in the pots prior to harvesting.Thus, 14 plants were evaluated.
As for the difficulty in obtaining interspecific hybrids in Arachis, the number of abortions is a major factor that decreases the percentage of success.In this study, during the crosses, abortions were observed for 117 pegs, which have not developed seeds (Table 3), and during the germination of seeds, two other types of abortions were verified: abortion of well developed seeds that did not germinate, and death of seedlings without enough vigor to survive in the pots.Abortions observed in this study, both pre-and post-zygotic, may indicate that these individuals were hybrids that failed to survive.
The plants resulting from the crossing between A. magna V 13751 × A. kempff-mercadoi V 13250 were characterized as hybrids.This hybrid was artificially doubled, which generated the amphidiploid An13 (A. magna V 13751 × A. kempff-mercadoi V 13250) 4× .When An13 was crossed with the cultivar IAC OL4, 98 pollinations were undertaken, which resulted in 10 seeds.All these 10 seeds were germinated and four were characterized hybrids.The An13 (AABB) and IAC OL4 (AABB) have similar genomic formulas, so there were fewer abortions of pegs, seeds and seedlings (Paula et al. 2017).

Reproductive Characterization
The analysis totaled 22 plants, eight parents and 14 individuals of the progenies (F 1 ).For both pollen grain staining methods (Fig. 1), analysis of variance evidenced significant differences between individuals, and no differences between replications.
With respect to results of TZ staining (Table 4), most parents showed high viability (above 77% stained grains) except the plant 7 (V 7639), which continued to show viability lower than expected (33%), however, much higher than plants characterized as hybrid.For progenies plants, five of them showed high viability (above 90%) and six showed viability below 1.50%.Tukey's test at 5% probability indicated that the groups a to f encompassed all parents and individuals from selfing, but the group g included the hybrid individuals of progenies A. magna V 13751 × A. kempff-mercadoi V 13250 (pls.16 and 17), A. magna K 30097 × A. kuhlmannii V 7639 (pl.19), A. magna V 13751 × A. kuhlmannii V 9243 (pls.20, 21 and 22).Three plants (pl.11, 12 and 13) were not evaluated because they died.
In the case of reproductive and speciation strategies, Krapovickas and Gregory (1994) stated that selfing is considered the normal mode of reproduction in Arachis, but the flower can also receive sporadic visits of insects.In this sense, fertility, as estimated by pollen grain staining, can be used to determine the degree of genetic isolation between species.If more than 50% pollen grains are stained, the anther dehiscence is normal.If less than 50% are stained, anthers begin to have difficulty in dehiscence.Near 25% staining, anthers should not be dehiscent, but the grains may still be captured by a pollinator, such as bees.With less than 15% stained pollen, anthers must be dissected to obtain the pollen grains, and when the rate is below 10%, due to the small number of viable pollen grains, it is doubtful whether it is possible to achieve a successful pollination.
In individuals A. magna V 13751 × A. kempff-mercadoi V 13250 (pls.16 and 17), A. magna K 30097 × A. kuhlmannii V 7639 (pl.19), and A. magna V 13751 × A. kuhlmannii V 9243 (pls.20, 21 and 22), anthers were indehiscent and there was difficulty in taking the pollen; anthers had to be macerated on the slide.These F 1 plants obtained from interspecific crosses between accessions of genome A and B were classified as hybrids, and all of them showed a low pollen stainability, indicating high genetic isolation between the parents species.Species with A genome and species with B genome have approximately 20% genetic similarity (Moretzsohn et al. 2013).Stalker (1991) compared the pollen viability and meiosis in intra and interspecific crosses.All parents presented over 98% of stained grains, a low presence of univalents (below 0.05 univalent) and high rate of bivalents (approximately 10 bivalents) during meiosis, thus, they were considered normal.All progenies individuals resulting from intraspecific crosses showed similar results from the parents, indicating high genetic similarity between the accessions used for the hybrids production.Individuals of the progenies from interspecific crosses exhibited less than 11% stained pollen with a high number of univalents and low amount of bivalents, indicating low genetic similarity.Moretzsohn et al. (2013) showed the genetic similarity between distinct Arachis species.The studies of Moretzsohn et al. (2013) and Stalker (1991), combined, show that the genetic similarity between the parents can influence the percentage viability of pollen.

Molecular Identification
The SSR loci selected were Seq3D09, IPAHM406, RI2A06 and RM14B11, based on their amplification pattern in the    2014), the microsatellite markers were useful in identifying hybridization in natural or controlled crosses.

Morphological Characterization
Individuals were characterized morphologically and PCA eigenvalues revealed that the first three components explained 95% of the total variance in morphological traits.
Table 6 lists the descriptors that explained most of the variation observed according to the Principal Component 1.Among the 15 most important descriptors, seven were collected from the main axis of the plant, four from the lateral branch and four from flower.All other descriptors, not present in this table, were considered irrelevant in this study to determine the morphological variation between individuals.
From the components 1 and 2 multiplied by the mean values of each trait for each individual, it was possible to build a biplot graph (Fig. 3).Plants 11, 21 and 22 had the central axis broken during development.When analyzing the 15 most important descriptors for the characterization, seven of these belonged to the main axis of the plant, so they were very discriminative and plants that did not have it remained completely scattered in the graph.
The parents were well distributed on the plot.Plants A. magna K 30097 × A. kuhlmannii V 9243 (pls. 11, 12, 13) and A. magna K 30097 × A. helodes V 6325 (pl.18), identified as a result of selfing, were located near the female  (pls. 20, 21 and 22), identified as hybrid, were located a little away from their parents.The plant A. magna V 13751 × A. kuhlmannii V 9243 (pl.20), the only one that had the main axis, was located closer to the female parent A. magna V 13751 (pl. 3) than to the male parent A. kuhlmannii V 9243 (pl.6), the values that caused this distribution in the graph can be seen in Table 6.As well as in A. magna K 30097 × A. kuhlmannii V 9243 (pl.11), which had no main axis, hybrid plants A. magna V 13751 × A. kuhlmannii V 9243 (pls.21 and 22) were located far from the parents.
Hybrids A. magna V 13751 × A. kempff-mercadoi V 13250 (pls.16 and 17) were located closer to the male parent A. kempff-mercadoi V 13250 (pl.8) than to the female parent A. magna V 13751 (pl.3).Analyzing the most important descriptor for these two hybrids, the female and male parents had, respectively, 7.67 mm and 7.28 mm for the length of adnate part of stipule in the lateral branch (cpdRL).In this way, plants A. magna V 13751 × A. kempff-mercadoi V 13250 (pls.16 and 17), with 8.42 mm and 8.61 mm, respectively, were located closer to the male than to the female parent.
The hybrid A. magna K 30097 × A. kuhlmannii V 7639 (pl.19) was located closer to the male parent A. kuhlmannii V 7639 (pl.7) than the female parent A. magna K 30097 (pl.2).Considering the most important descriptor, it is observed that the female and male parents presented, respectively, 5.57 mm and 8.21 mm for the length of adnate part of stipule in the lateral branch (LapsLB), and thus the plant A. magna K 30097 × A. kuhlmannii V 7639 (pl.19), with 5.57 mm, was located closer to the male than to the female parent.This cross presented an important morphological trait, as the female parent A. magna K 30097 (pl.2) has orange flower and the male parent A. kuhlmannii V 7639 (pl.7) has yellow flower.As soon as the first flower appeared in A. magna K 30097 × A. kuhlmannii V 7639 (pl.19), this had yellow color, as the male parent.
According to Janini (2011), who evaluated 35 accessions of Arachis, including wild species, amphidiploid and cultivars, the accessions of A. kempff-mercadoi V13250, A. kuhlmannii V7639, V 9243, A. magna V 13751 and A. magna K 30097 are among the accessions that stood out for smaller infestations, greater tolerance and antibiosis to E. flavens and S. bosquella.Some accessions have multiple defense, protecting against both pests at the same time.In this study, six hybrid plants in three different combinations were produced, two hybrids from A. magna V 13751 × A. kempff-mercadoi V 13250 (pls.16 and 17), one from A. magna K 30097 × A. kuhlmannii V 7639 (pl.19), and three from A. magna V 13751 × A. kuhlmannii V 9243 (pls. 20, 21 and 22).
The two hybrid plants from A. magna V 13751 × A. kempff -mercadoi V 13250 (pls.16 and 17) had more phenotypical traits of the accession A. kempff -mercadoi V 13250 (pl.8), thus making the individual to locate closer to the male parent than to the female parent in the PCA.In agreement with Janini (2011), accession A. kempffmercadoi V 13250 was among those less attacked by E. fl avens (resistance) and among the least damaged by thrips and rednecked peanutworm (multiple tolerance).According to Michelotto et al. (2017), A. kempff -mercadoi V 13250 presented an average of 2.81 thrips per leaf while the controls had averages between 28.99 and 33.64 thrips per leaf.Th e accession A. magna V 13751 was among those with

Figure 1 .
Figure 1.(a) Pollen of the accession V 9243 that was stained with 2% acetocarmine with glycerin.(b) Pollen of the hybrid V 13751 × V 13250 that was stained with 0.25% tetrazolium.Arrows indicate non-stained pollen.100× magnification in optical microscope.
absence; P = allele presence; NE = not evaluated in this sample; NA = not applicable.
in 2012/2013 and 2013/2014 crossing seasons, from December to March.Flower buds of the female parents were emasculated in the late afternoon and pollinated in the morning of the next day.Eighteen combinations were performed between species of genome B (A. gregory C.E.

Table 1 .
Accessions used as parents resistant to pest, species, genome, individual, collectors, latitude, longitude and altitude of the collection area.
1 The missing data of the V 7635 genome means that its genome was not yet identified.

Table 5 .
Amplification profile of informative alleles from the four SSR loci used for hybridization confirmation.

Table 6 .
Order of descriptors that contributed most to the morphological variation observed in the principal component 1 (Prin 1) of the Principal Component Analysis of the parents and hybrids.
parent A. magna K 30097 (pl.2), in the top right of the graph except A. magna K 30097 × A. kuhlmannii V 9243 (pl.11), which did not have the main axis, and stood opposite to these plants, being located in the bottom left of the graph.Plants of A. magna V 13751 × A. kuhlmannii V 9243