Ecology, Behavior and Bionomics First Genotyping of Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae) Progeny from Crosses between Bt-Resistant and Bt-Susceptible Populations, and 65-Locus Discrimination of Isofamilies

Abstract

Concerns exist that Bt-resistant populations of Spodoptera frugiperda (J. E. Smith) (Lepidoptera: Noctuidae) found in Puerto Rico, could spread to continental United States, and about the lack of molecular tools to monitor potential crosses or distinguish populations. In this work, the feasibility of genotyping S. frugiperda crosses between Bacillus thuringiensis (Berliner) (Bacillales: Bacillaceae) (Bt) resistant and susceptible populations using simple sequence repeats (SSRs, or microsatellites) was assessed. Parents and their corresponding progeny (five resistant, five susceptible phenotype) were genotyped using 192 SSRs on three reciprocate crosses alternating male and female from Bt-susceptible and Bt-resistant populations. Oviposition, mortality and fecundity were evaluated for five pairs of each of the crosses. Cluster analysis showed that progeny of each cross was associated with the paternal and not the maternal genotype, probably due to higher heterozygosity in male parents compared to females. Seven SSR markers had one or more alleles correlated (p≤0.002) with the Bt-resistant phenotype, and 18 SSRs showed uniparental inheritance. Analysis of additional samples showed genetic differences among isofamilies of S. frugiperda at 65 loci. The best 35 markers that discriminated isofamilies are reported, three of them related to components of the communication system in moths: pheromone, olfactory receptor and antennal esterase. This is the first report of: 1) a multi-locus genotyping in crosses of Btresistant and susceptible individuals of S. frugiperda, 2) seven codominant markers associated to Bt-resistance in this species, 3) discrimination of isofamilies by genetic differences at three loci that could affect the behavior of S. frugiperda.

Key words

MSSR, Behavior, Microsatellite, Marker, Cry toxin, Pheromone

Introduction

S. frugiperda is a polyphagous insect that causes significant losses in Zea mays (L.) (maize), Arachis hypogaea (L.) (peanut), Gossypium spp. (cotton), Sorghum bicolor (L.) (sorghum), Oryza sativa (L.) (rice), and other species [1,2]. S. frugiperda is endemic to the American continent, and its control relies in certain areas on the use of transgenic crops expressing the Cry toxin of Bacillus thuringiensis (Berliner) [3]. In the year 2011, 75% of the acres planted with corn and 65% of the acres planted with cotton in the United States corresponded to Bt-toxin expressing crops according to the Economic Research Service (ERC) http://www.ers. usda.gov/Data/BiotechCropsD. In order to maintain the effectiveness of insect control by Bt toxins is important to implement management practices that minimize the development of Bt resistance in insect pests. Concerns that Bt-resistant populations found in Puerto Rico could spread to continental United States [4-6] made evident the lack of sufficient molecular tools to monitor their potential dispersal. Many studies were done to understand migration of S. frugiperda in the U.S.A., though, as pointed out by Nagoshi and Meagher [7] these studies sometimes had inconsistent results. Analysis of several loci were performed in order to characterize populations and find host associations in S. frugiperda [8,9]; but resolving issues of strain hybridization would require analysis of a large number of autosomal-codominant markers [10]. Microsatellites, also called SSRs (simple sequence repeats) or STRs (short tandem repeats), are codominant markers that could be suitable for that purpose. Few years ago, concerns were reported that transposable elements widespread throughout Lepidopteran genomes had rendered existing rnicrosatellites ineffective for their use in population studies of these insects [11]. In 2011, Arias et al. reported 192 microsatellite markers were for S. frugiperda, and shown their effectiveness in discriminating populations of this insect. Thus, the goal of the present work was to determine the feasibility of analyzing strain hybridization in S. frugiperda using a large number of microsatellites, as a further step towards their use in population genetics. For that purpose, crosses of Bt-resistant and Bt-susceptible populations of S. frugiperda were made: a) analyzed their biometrics and phenotyped the progeny for Bt resistance, b) genotyped parents and progeny using 192 microsatellites, c) explored the possible association between sex and heterozygocity at the studied loci, d) screened the microsatellite markers for their potential association to the Bt-resistant phenotype, and e) the potential discrimination of isofamilies using these markers. Some statistics on the abundance of transposon-related sequences in our microsatellite-enriched libraries generated from genomic DNA are also reported in this study.

Materials and Methods

S. frugiperda colonies

For reciprocal crosses, two strains of S. frugiperda were used: colony FAW-512 which has resistance to Bt toxins Cry1Ac and Cry1Fa, collected from a maize field in Santa Isabella, Puerto Rico in 2008 and was previously reported [12]; and colony FAW-Mon (Bt-susceptible control), obtained from Monsanto Company, Memphis, TN, USA. FAW-512 has been reared for several generations by bioassay selection exposing the neonates to Bt-Cry proteins, while FAW-Mon has been cultured separately for few generations using regular diet [13]. Additional males and females from three colonies FAW-512, FAW-W and FAW-STV were also genotyped to determine their heterozygosity levels. The colony FAW-W was collected from a maize field in Washington County, MS; and FAW-STV was collected from maize expressing Bt-Cry1Ab in a field in Stoneville, both in MS, U.S.A. All colonies have been cultured by groupmated according to a method previously described [13] and were maintained at the Southern Insect Management Research Unit, USDA-ARS, in Stoneville, MS.

Crosses, Biometrics and Challenge with Crylfa Toxin

FAW-512 and FAW-Mon colonies were used for these crosses. Five single pairs of virgin F10 FAW-512♀ × virgin F5 FAWMon ♂ and five single pairs of virgin F10 FAW-512♂ × virgin F5-FAWMon♀ were confined for mating and oviposition in 500 ml containers (Model 42505LY, Consolidated® Plastics Co., www.consolidatedplastics.com) with an open top cover with Batist cloth (15 x 15 cm) (Zweigart®, Piscataway, NJ) and inner walls covered by thin wax paper (Reynolds Cut-Rite) to receive the eggs. A plastic cup of 37 rnl (T-125, Solo®, Urban, IL) containing approximately 15 ml of a 10% solution of sucrose-water and benzoic acid 1000 mg kg- 1 was provided to each single pair as a food source. For each pair, data of pre-oviposition period, oviposition period, number of eggs laid per day per female, and mortality were collected. One hundred neonates from the first group of eggs collected from each pair were bio-assayed by overlaying the artificial diet with 10 ng cm-2 of Cry1Fa purified proteins produced in Pseudomonas fluorescens (Flǖggel) [13]. After seven days of exposure to Cry1Fa, surviving larvae were transferred to regular diet until 5111 instar (Bt-resistant). Bt-susceptible individuals were considered those larvae that died at 3rd instar, or remained at 3rd instar after transfer to normal diet. Two sets of parents with 10 progeny each (five Bt-susceptible and five Bt-resistant) were randomly selected for SSR-genotyping at 192 loci. The experiment was repeated at a later time using the next generation of each colony: five pairs of virgin F11 FAW-512♀ × virgin F6 FAW-Mon♂ and five pairs of virgin F11 FAW-512♂ × virgin F6 FAW-Mon♀. On the repeated experiment, the first crosses reproduced normally; however, no progeny were obtained from any of the five crosses of F11 FAW-512♂ × F6 FAW-Mon ♀. Thus, only 12 individuals (2 parents, 10 progeny) from one of these crosses were used for genotyping. Differences in pre-oviposition period, oviposition, fecundity and mortality between crosses were analyzed using PROC GLM on SAS program [14].

DNA Extractions, SSR Genotyping of S. frugiperda Parents and their Progeny

DNA was extracted from each individual using DNeasy Plant Mini Kit (Qiagen, Valencia, CA). Though the original microsatellites (SSRs) were developed using DNA from thorax tissue from adult individuals, the genotyping in the present work was done using DNA from entire individuals. Also is important to indicate that even from small-individual 3rd instar larvae, enough DNA was obtained (≥2 μg) to run all the markers. All 192 S. frugiperda SSR markers previously reported [15] and developed by the method of [16], were used to genotype each individual in the present work. Genotyping was done for 36 individuals from three crosses, that is, two parents and 10 progeny (five with Bt-resistant phenotype and five with Bt-susceptible phenotype) per cross. To simplify the notation, samples were named as follows: a single letter for parents (A, B, C, D, E, F). Crosses were named AB, CD and EF, where the first letter corresponds to the female and bold/underlined means Bt-resistant. Thus, A and E were F10 and F11 females from the Bt-resistant population FAW-512, C was an F5 female from the Bt-susceptible population FAW-Mon, D was an F10 male from FAW-512, and B and F were F5 and F6 males of the FAW-Mon. The progeny was labeled with the name of the cross, followed by another letter and a number. The letter N (normal larvae, resistant to Bt toxin) or D (dead or delayed development, susceptible to Bt toxin) were added to the name of each cross, followed by the numbers 1 to 5 to identify each sample. Thus, parents A and B were crossed, and the 10 progeny analyzed corresponded to five normal/resistant larvae (ABN1 to ABN5) and five progeny dead or with delayed development (ABD1 to ABD5). Genotyping of the SSRs on the samples was done with the software GeneMapper 3.7 (Applied Biosystems, Foster City, CA) by visual examination of allele sizes in base pairs (bp) at each locus. DNA sequences containing SSRs were screened using NCBI Nucleotide Databases (BLASTnr and BLASTx) [17].

Cluster Analysis, Heterozygosity and Possible Association of Markers to Bt-resistance

To determine if the SSR genotyping would group samples by Bt-resistance, a cluster analysis was performed by Neighbor Joining [18] using the software NTSYSpc v. 2.2 (Exeter Software, Setauket, NY). The confidence levels for the dendrograms were assessed by bootstrap resampling (1000 replicates) [19,20] using WINBOOT [21]. Heterozygosity (H0) was also analyzed, defined here as the presence of more than one allele (amplicon) size per locus per individual, and it was calculated for each sample from crosses based on 100 SSR markers using the program UPIC [22]. Markers and phenotypes were analyzed for putative association as indicated by the Fisher's Exact Test using a 2-way contingency table to analyze allele frequencies as computed using the Proc Freq procedure of SAS ver. 9.2 [23,24].

Given the higher H0 observed in males that participated in the crosses, additional samples were analyzed to determine if indeed males are in general more heterozygous than females in this species; and also to determine if the SSRs were discriminating individuals by sex. Twelve additional samples, six females and six males of S. frugiperda from three isofamilies, FAW-512, FAW-W and FAW-STV, were genotyped with all 192 SSRs to determine if a gender bias existed in the level of heterozygosity. Cluster analysis and H0 were performed as indicated for crosses, using in this case 98 SSRs that amplified all samples.

Results

Biometrics of FAW Crosses and Challenge with crylfa Toxin

Ovipositions started two days after the pairs were formed, and though for some pairs with Bt-susceptible females (F10 FAW-512♂ × F5FAW-Mon♀) ovipositions started a day later, the differences were not statistically significant. The total period of oviposition was 10 ± 1 days, slightly shorter for susceptible females, but again without significant differences between crosses, Figure 1. Fecundity was significantly higher (p ≤ 0.001) for Bt-resistant females (F10 FAW-512♀ × F5 FAW-Mon♂). The cumulative number of eggs laid by five Bt-susceptible females (FAW-Mon♀) was <2000, whereas for five Bt-resistant females (FAW-512♀) was >4000, data are plotted in Figure 1. The average mortality in crosses of Bt-resistant females with susceptible males was 13%, whereas in crosses of susceptible females with Bt-resistant males the mortality was 11%, these differences were not statistically significant. In the repeated experiment, crosses with Bt-resistant females (FAW-512♀) resulted in progeny, but crosses with susceptible females (F11 FAW-512♂ × F6 FAW-Mon♀) resulted in no progeny.

Figure 1: Cumulative number of eggs on daily oviposition of S. frugiperda crosses between a Bt- resistant (FAW-512) and a Bt-susceptible (FAWMon) population. Filled circles and solid lines correspond to F5 FAW-512♀ x F to FAW-Mon♂; empty circles and dash lines correspond to F10 FAW-Mon♀ x F5 FAW-512♂. Data points are the total cumulative number of eggs of five crosses with their corresponding standard errors of the means. Arrows indicate the last day of oviposition period.

SSR genotyping of S. frugiperda reciprocal crosses

Parents, five progeny with Bt-resistant phenotype and five progeny with Bt-susceptible phenotype were genotyped with the SSR markers. A total of 36 S. frugiperda individuals corresponding to three pairs of parents with 10 progeny each were screened with 192 SSR markers previously developed for this species [15]. A total of 100 out of 192 markers (52%) amplified all 36 samples, 59 of these markers were polymorphic and 41 monomorphic, Table 1. Examples of the feasibility of genotyping parents and progeny of the crosses are shown in Figure 2a and 2b. Among the markers that amplified some but not all 36 samples: 55 did not show amplification in one or more samples, 17 had amplicons difficult to score, and 20 failed to amplify all samples. The specific markers within each group are listed in Table 1. The 59 SSR markers that were polymorphic and amplified all 36 samples were used for cluster analysis. This analysis showed that the progeny of each cross was closely associated to the paternal and not to the maternal genotype, and these clades were supported by 82 to 99% confidence using bootstrap resampling, Figure 3a. The three females from the crosses grouped in a separate clade that was supported by a 99% probability in bootstrap resampling, Figure 3a.

Figure 2: Examples of SSR genotyping of parents and progeny in crosses between Bt-resistant and Bt-susceptible Spodoptera frugiperda populations. Panels a and b show examples of good quality markers. Panels c, d, e, f, correspond to marker Stv_Spf406 (in cross CD), Stv_ Spf858 (in crosses AB and CD), and Stv_Spf1783 (in cross AB), where A, B, C, D are parents, the first letter in the pair corresponds to the female, and bold-underlined font indicates Bt-resistant parent. Alleles observed for parents and their corresponding Bt-susceptible progeny (ABD1 to ABD5) and Bt-resistant progeny (ABN1 to ABN5).

Figure 3 (a): Cluster analysis of three Spodoptera frugiperda crosses between Bt-resistant (FAW-512) and Bt-susceptible (FAW-Mon) populations. Neighbor Joining calculated using data from 59 polymorphic-SSR markers that showed amplification in all 36 individuals - three pairs of parents with 10 progeny each. Crosses were AB, CD and EF, where A, B, C, D, E, F are parents, the first letter in the pair corresponds to the female, and bold-underlined font indicates Bt-resistant parent. The progeny were named with the two letters of the corresponding parents followed by N (for Bt-resistant) or D (for Bt-susceptible phenotype), and the numbers 1 to 5 to identify each sample. (b) Cluster analysis by Neighbor Joining of male and female individuals from three isofamilies (FAW-512, FAW-W, and FAW-STV) using 98 SSRs; markers clearly distinguish isofamilies and do not group samples by sex. Bootstrap coefficients of 1000 resampling are indicated at the nodes for values higher than 50%.

Table 1: SSR markers used in the studies, organized by their efficiency in amplifying 36 samples in crosses of Bt-resistant and Bt-susceptible populations of Spodoptera frugiperda. Marker names indicated by a bordered cell, effectively discriminated isofamilies of this insect, and those in bold-italic font correspond to the 18 loci that showed some degree of uniparental inheritance. All marker sequences were published before [15].

Heterozygosity in reciprocal crosses and additional samples.

The 100 SSR markers that amplified all the samples in the crosses were processed with UPIC software to calculate heterozygosity (H0) of each sample. The percentage of H0 of the parents in each cross was significantly lower for females (p ≤ 0.002) A, C, E than for males B, D, F, independently of the population from which they originated, Table 2. Heterozygosity of each individual progeny was calculated and the average values for the Bt-susceptible and Bt- resistant phenotypes of each cross (AB, CD and EF) are shown on Table 2. The H0 of the progeny, whether they were resistant or susceptible, was much closer to the paternal H0 (±7%) than to the maternal H0 (±19%), Table 2. No significant differences in Ho were observed between Bt-resistant and susceptible progeny phenotypes within any of the crosses.

Table 2: Heterozygosity observed in samples analyzed in crosses of Bt-resistant and Bt-susceptible Spodoptera frugiperda across 100 loci. The crosses were AB, CD and EF, where A, B, C, D, E, F are parents, the first letter in the pair corresponds to the female, and bold-underlined font indicates Bt-resistant parent. The progeny were named with the two letters of the corresponding parents followed by N (for Bt-resistant) or D (for Bt-susceptible phenotype), and the numbers 1 to 5 to identify each sample. S.E.: standard error of the mean. Additional samples are 6 males and 6 females from three isofamilies (FAW-512, FAW-W and FAW-STV).

To determine if H0 was indeed higher for males as compared to females of this species, two males and two females from three isofamiles (FAW-512, FAW-W and FAW-STV) were fingerprinted using the 192 SSRs. The analysis showed that only in isofamily FAW-512 the H0 of the females was slightly higher than in males, whereas on the other isofamilies no gender bias was observed in the level of H0. Overall, the H0 of isofamily FAW-STV was significantly higher (p≤ 0.05) than for FAW-512 and FAW-W. Cluster analysis of the 98 markers that amplified all these 12 samples, indicated that the markers not necessarily distinguished the gender of the samples, though they clearly distinguished isofamilies, as shown by high confidence values of bootstrap resampling, Figure 3b. Genetic differences among isofamilies were observed at 65 loci. Table 1 indicates by a square box the best 35 of these markers that discriminated one isofamily or more, and some examples are shown in Figure 4.

Figure 4: Examples of SSRs that discriminate isofamilies of Spodoptera frugiperda. Panels show alleles that characterized three isofamilies of this insect (FAW-512, FAW-W and FAW-STV), independently of the gender of the samples.

Potential association of SSR markers to Bt-resistant phenotypes

Out of the 59 polymorphic SSR markers that amplified all samples in the crosses, seven had one or more alleles possibly linked to the Bt-resistant phenotype (p≤ 0.002) in one or more of the crosses, as shown by association tests. These were Stv_ Spf858, Stv_Spfl125, Stv_Spf409, Stv_Spf1783, Stv_Spf1136, Stv_Spf1651, Stv_Spf1176. Examples of markers Stv_Spfl783, Stv_Spf858, and StvSpf406 for which between four and five progeny either Bt-resistant or susceptible carried the same allele/s, are shown in Figure 2c-2f

Loci with uniparental inheritance

The analysis of SSRs in crosses of fall armyworm revealed that at 18 loci either the maternal or the paternal alleles were absent in some of the progeny of the crosses. The maximum absence of maternal alleles was observed on Stv_Spf406, for which only paternal alleles were present in all 30 progeny analyzed from three crosses. Interestingly, BLAST results showed that the DNA sequence corresponding to Stv_Spf406 had certain homology to a protocadherin. The other extreme case was Stv_Spf1315, for which only maternal alleles were present in 29 out of 30 progeny on the three crosses. Sequences of three markers (Stv_Spf1502, Stv_Spf578 and Stv_Spf967) that showed absence of paternal alleles (in 4, 7, and 12 out of 30 samples, respectively) had short regions of homology to mitochondrial DNA. An interesting marker in this group was Stv_Spf764, because in all 30 individuals, progeny from three crosses, each individual received either the paternal or the maternal allele, but not both.

Discussion

Fecundity and fitness cost

Oviposition started earlier in crosses between Bt-resistant females and susceptible males than in reciprocate crosses, and the oviposition period was slightly longer when the resistant parent was a female, Figure 1, but these differences were not significant. However, the fecundity of Bt-resistant females in crosses with susceptible males was significantly higher, Figure 1. There are examples of Lepidoptera in which no fitness cost was associated to acquiring Bt resistance, i.e., acquiring resistance to Cry2Ab in Helicoverpa armigera (Hubner) (Lepidoptera: Noctuidae) [25], and in Pectinophora gossypiella (pink bollworm) [26]. However, a fitness cost such as reduced fecundity is expected in most insects acquiring Bt resistance [27]. This has been observed in other Lepidoptera, Helicoverpa armigera, when a different toxin was used (Cry1Ac) [28,29]. In fall armyworm, S. frugiperda, growth rate of F1 hybrids of Bt-resistant females by susceptible males was significantly higher from those obtained in F1 hybrids from resistant males by susceptible females (Portilla, unpublished). This change in fecundity could have contributed to the absence of eggs during our replicated experiment, in which crosses of resistant males (FAW-512) with susceptible females (FAW-Mon) resulted in no progeny.

Feasibility of genotyping parents and progeny in crosses between Bt-resistant and Bt-susceptible populations

In 2004, Prowell suggested that a large number of co-dominant markers would be necessary to study strain hybridization in S. frugiperda. Since then, ten microsatellites for Spodoptera exigua [30] and six for S. frugiperda [31] have been reported and successfully used for the genus Spodoptera, though this number of markers would still be 12 insufficient to study strain hybridization in this insect. From the 192 used microsatellites for genotyping parents and progeny of S. frugiperda crosses between Bt-resistant and Btsusceptible populations, 100 markers (59 polymorphic, 41 monomorphic) amplified all 36 samples, and the majority were easy to score, examples in Figure 2a and 2b. Transposon-related elements were believed to be present throughout Lepidoptera genomes, and were considered a limiting factor for the use of microsatellites in population studies of these insects [32]. In the present work with crosses of individuals from two isofamilies, the feasibility of genotyping parents and progeny using the previously reported microsatellites were tested [15]. From the prepared microsatellite-enriched libraries of S. frugiperda, no particular abundance of transposon-related sequences was observed. BLASTn and BLASTx analysis of 1904 contigs containing microsatellites detected only three sequences related to retrotransposons, these corresponded to markers Stv_Spf69, Stv_Spfl582 and Stv_Spf1713. Two of these markers (Stv_Spf69 and Stv_Spf1582) were monomorphic across all the samples analyzed (36 in the crosses and 12 additional males and females) and were easy to score. Marker Stv_Spf1713, though it failed to amplify some samples in the crosses, it was actually effective discriminating isofamilies.

Ten percent of the 192 SSR markers did not amplify any of the samples within crosses of S. frugiperda. In addition to possible null alleles, many factors can affect amplification, including DNA quality, genetic diversity of the samples and robotic liquid handling. Using the same method of SSR development for other species, the percentage of SSR markers that failed to amplify all samples was found with 3% in Chionanthus [33], 6% in Macrophomina [34], 12% in Rotylenchulus [35] and 29% in Cyperus [36]. Thus, the 10% of markers without amplification in S. frugiperda was not necessarily high compared to other species. Overall, the microsatellites had a good quality with not limitations for their use.

Dominant or recessive Cry resistance

In this study 7 (Stv_Spf858, Stv_Spfl125, Stv_Spfl409, Stv_Spfl783, Stv_Spfl136, Stv_Spfl651, Stv_Spf1176) out of 60 polymorphic-SSR markers tested had alleles significantly correlated with Bt-resistance (p 0.02), some examples are shown in Figure 2c-2f. In general, high levels of resistance to Bt toxins in Lepidoptera have been interpreted as the result of one or few autosomal recessive or incompletely recessive genes [4]. In contraposition, low levels of resistance have been interpreted as the result of dominant or co-dominant genes [37-39]. Since SSR markers are inherently co-dominant, the list of seven markers associated to Bt-resistance provided in the present work could be used as a starting point to study low levels of Bt- resistance in S. frugiperda.

Cluster analysis and heterozygosity

It was expected that cluster analysis of parents and progeny of reciprocal crosses using microsatellites would group Btresistant and Bt-susceptible individuals; instead, the analysis showed association of progeny to the male parent, and grouped the three females from the crosses in a separate clade, Figure 3a. Two hypotheses were explored: a) that males were more heterozygous than females; or b) that the progeny were not randomly expected 50% males and 50% females, but probably they were all males and the markers were grouping samples by sex. Since heterozygosity (H0) level is usually considered an indicator of environmental fitness in organisms (Livshits & Kobyliansky 1985), levels H0 of parents and progeny as possible explanation to the results of cluster analysis were explored. Females in the crosses had a H0 between 27-38%, whereas Hoof their male counterparts was significantly higher, between 35-53%. In S. frugiperda, the female is heterozygous carrying the chromosomes ZW, whereas the homozygous is the male, which carries the chromosomes ZZ [40]. Thus, the next question was, if H0 in S. frugiperda was gender specific, as this could have consequences in the biology of the insect. So, another 12 individuals were fingerprinted, 6 females and 6 males, from three isofamilies; the results showed no significant differences in H0 for males vs. females (Table 2), and that the microsatellites did not discriminate by sex, but showed a clear distinction by isofamilies, Figures 3b and 4.

Isofamilies

A total of 65 microsatellite markers showed some distinction of isofamilies. In this study are reported the best 35 microsatellites that discriminated one isofamily or more of S. frugiperda on the 12 samples of males and females from three isofamilies (FAW- 512, FAW-W and FAW-STV) tested. These markers are indicated by bordered cells in Table 1, and examples shown in Figure 4. Within this group of markers, one was Stv_Spf1221 related to a protein-folding calnexin, but the most interesting markers were: Stv_Spf173, Stv_Spf544 and Stv_Spf292, whose DNA sequences had homology to a pheromone, and olfactory receptor and an antennal esterase, respectively. All three of these components interact with each other and are essential in the social life of moths. Pheromones are released by female moths to indicate their readiness to mate [41], and are detected by olfactory receptors located in the male antennae where the odor guides them in flight toward the females [42]. In order to maintain high sensitivity to the signal during flight, pheromones are rapidly inactivated [43], what is achieved in microseconds by antennal esterases [44]. This means that the polymorphism observed in markers Stv_Spfl73, Stv_Spf544 and Stv_Spf292 is evidence of genetic variations in three important components of the communication system in S. frugiperda. Even more so, the pheromone/olfactory receptor is an extremely sensitive and selective system, "tuned to its key compound" and used for intra-specific communication [45], where the selectivity of the pheromone receptor neurons determine the chemical response specificity [46]. Given the high specificity of the system, would not be surprising that the simultaneous genetic variations on pheromone, olfactory receptor and antennal esterease was detected by the three microsatellites, may impact the "intra-strain" or "intra-colony" behavior in S. frugiperda. Thus, markers Stv_Spfl73, Stv_Spf544 and Stv_Spf292, may not only distinguish populations, but also provide valuable information regarding the social life of this insect, and deserve further investigation.

Uniparental inheritance, and gender effect on resistance

In addition to the issue of dominant or recessive inheritance of Bt-resistance, the gender of the resistant parent has also been reported as a determinant factor in the level of resistance of the progeny. In Helicoverpa armigera resistance to Cry1Ac was suggested as dominant and maternally enhanced inheritance [47]. In Heliothis virescens Fabricius and in S. frugiperda isofamily 456, the susceptibility to Bt-toxin was dependent on the gender of the susceptible parent, suggesting that resistance was linked to sex-genes [4,12]. In the present work, SSR analysis showed that at 18 loci either the maternal or the paternal alleles were absent in one or more samples, markers indicated in bold-italics in Table 1. Three of these loci (Stv_Spf967, Stv_Spf578, Stv_Spfl502) had some homology to mitochondrial DNA. That could explain absence of paternal alleles if the SSR marker amplified maternal mitochondria; or absence of maternal alleles if the mitochondria were paternally inherited. In insects, paternal mitochondria can be inherited [48]. Among the 18 loci with various levels of uniparental inheritance, an interesting marker was Stv_Spf406 for which maternal alleles were absent in all 10 progeny from each of the three crosses (30 individuals). The DNA sequence of marker Stv_Spf406 had some homology to a protocadherin. Cadherin-like proteins confer susceptibility to Bt- toxins in vitro by binding to the toxin, whereas disruption of a cadherin gene confers resistance to Bt-toxin [27,49,50]. Thus, if a protocadherin associated to marker Stv_Spf406 is only paternally inherited, it could explain the Bt-resistance linked to sex genes previously observed in this insect [4]. Another particularly interesting locus in the group of 18loci was marker Stv_Spf764. This marker was monoallelic, both in parents and progeny, showing either the maternal or the paternal allele but not both, in all 36 samples from three crosses. BLAST analysis did not link the DNA sequence of this marker to any known sequence.

Though absence of maternal or paternal alleles was observed at 18 loci, the total number of genotypes with absent maternal alleles (137) was similar to the total number of genotypes with absent paternal alleles (147). Therefore, uniparental inheritance per se cannot explain the close association of the progeny with their paternal genotype observed during cluster analysis Figure 3a. It was assumed that the progeny in the crosses AB, CD and EF had 50% probability of being males or females. Additional male and female samples are shown that the microsatellites did not group the individuals by sex, Figure 3b, and that the number of samples with uniparental maternal or paternal inheritance was not significantly different. Then, other than the higher heterozygosity of the males involved in the crosses, there was not an explanation for the association of progeny and male parent.

Conclusions

In this work was best demonstrated that it is feasible to genotype parents and progeny in crosses between Bt-resistant and Bt-susceptible S. frugiperda individuals, and 100 SSRs out of 192 were effective for this purpose. Though higher heterozygosity was observed in males during the crosses, this was not a general trend when additional samples from more isofamilies were analyzed. Seven microsatellites were found associated to the Bt-resistant phenotype of the progeny in one or more crosses; these could be the starting point to study low levels of Bt-resistance, which are generally associated to dominant or codominant genes. SSR analysis also showed alleles characteristic of individual isofamilies at 65 loci, the best 35 that showed those genetic differences are reported in this study, including three markers with DNA sequence homology to a pheromone, an olfactory receptor and an antennal esterase. This is the first report of such genetic differences at the isofamily level for this insect, what could have significant implications on the behavior and communication of S. frugiperda.

Acknowledgements

This work was supported by USDA-ARS projects 6402-21310-003-00; 6604-42000-008-00 and 6604-21000-003-00. Mention of trade names or commercial products in this article is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture

Data Accessibility

Data outputs from ABI 3730 are several hundred Mb, and will be provided upon request.

Author Contributions

Arias RS: conceived the ideas, wrote the entire manuscript, did the fingerprinting analysis, cluster analyses, boostrapping, graphics, tables, designed additional experiments.

Portilla M: made the insect crosses and evaluated their biometrics and phenotypes, did statistical analysis of biometrics

Ray JD: performed the statistical analysis of association of markers to Bt-resistance

Simpson SA: performed all the fingerprinting reactions and contributed to discussions on the experimental design.

Blanco CA: started and maintained the insect colonies, and provided the Bt-toxin for the experiments.

Schemer BE: contributed the equipment, reagents and discussion

References

1. Hall DG. Insects and mites associated with sugarcane in Florida. Florida Entomologist.1988;71:138-150.
2. Sparks AN. Review of the biology of the fall armyworm (Lepidoptera, Noctuidae). Florida Entomologist. 1979; 62:82-87.
3. Jenkins JN. Transgenic plants expressing toxins from Bacillus thuringiensis. Methods in Biotechnology. 1999: 5:211-232.
4. Blanco CA et al. Susceptibility ofisofamilies of Spodopterafrugiperda (Lepidoptera: Noctuidae) to Cry1Ac and Cry1Fa proteins of Bacillus thuringiensis. Southwestern Entomologist. 2010;35:409- 415.
5. Matten SR and Reynolds AH. Current resistance management requirements for Bt cotton in the United States. Journal of New Seeds. 2003;5:137-178.
6. StorerNPet al. Discovery and characterization of field resistance to Btmaize:Spodopterafrugiperda (Lepidoptera: Noctuidae) in Puerto Rico. J Econ Entomol. 2010;103:1031-1038.
7. Nagoshi RN and Meagher RL. Review of fall armyworm (Lepidoptera: Noctuidae) genetic complexity and migration. Florida Entomologist. 2008;91:546-554.
8. Nagoshi RN. The fall armyworm triose phosphate isomerase (Tpi) gene as a marker of strain identity and interstrain mating. Annals of the Entomological Society of America. 2010;103:283-292.
9. Prowell DP et al.Multilocus genetic analysis of host use, introgression, and speciation in host strains of fall armyworm (Lepidoptera :Noctuidae). Annals of the Entomological Society of America. 2004;97:1034-1044.
10. Tay WT et al. Generation of microsatellite repeat families by RTE retrotransposons in lepidopteran genomes. BMC Evol Biol. 2010;10:144.
11. Blanco CA et al. Soybean flower and wheat germ proportions in artificial diet and their effect on the growth rates of the tobacco budworm, Heliothisvirescens. J Insect Sci. 2009;9: 1-9
12. Blanco CAet al. Baseline susceptibility of tobacco budworm (Lepidoptera: Noctuidae) to Cry1F toxin from Bacillus thuringiensis. J Econ Entomol. 2008;101:168-173.
13. Blanco CA et al. Baseline susceptibility of tobacco budworm (Lepidoptera: Noctuidae) to Cry1F toxin from Bacillus thuringiensis. J Econ Entomol. 2008;101:168-173.
14. SAS (2001) Statistical analysis system, SAS computer program, release 9.1 (ed. Institute S), Cary, NC, USA.
15. Arias RS et al. First Microsatellites from Spodopterafrugiperda (Lepidoptera: Noctuidae) and their potential use for population genetics. Annals of the Entomological Society of America. 2011;104:576-587.
16. Techen N et al. Optimized construction of microsatellite-enriched libraries. MolEcolResour. 2010;10:508-515.
17. Altschul SF et al. Basic local alignment search tool. J MolBiol2009;215:403-410.
18. Saitou N and Nei M. The Neighbor-Joining method- a New method for reconstructing phylogenetic trees. MolBiolEvol1987;4:406-425.
19. Efron B et al. Bootstrap confidence levels for phylogenetic trees. ProcNatlAcadSci USA. 1996;93:13429-13434.
20. Felsenstein J. Confidence-limits on phylogenies - an approach using the bootstrap. Society for the Study of Evolution.1985;39:783-791.
21. Yap IV and Nelson RJ. Winboot: A program for performing bootstrap analysis of binary data to determine the confidence limits of UPGMA based dendrograms. IRRI Discussion Paper Series No. 14. International Rice Research Institute, Manila, May 2009.
22. Arias RS et al. UPIC: Perl scripts to determine the number of SSR markers to run. Bioinformation. 2009;3: 353-361.
23. SAS/Genetics (2009) SAS/Genetics TM, Statistical Analysis System, SAS computer program, 9.2. SAS Institute Inc.,Cary, NC, USA.
24. Arias RS et al. First Microsatellites from Spodopterafrugiperda (Lepidoptera: Noctuidae) and their potential use for population genetics. Annals of the Entomological Society of America. 2011;104:576-587.
25. Mahon RJ et al. Why do F1 screens estimate higher frequencies of Cry2Ab resistance in Helicoverpaarmigera (Lepidoptera: Noctuidae) than do F2 screens? J Econ Entomol. 2010;103:472-481.
26. Higginson DM et al. Evolutionary trade-offs of insect resistance to Bacillus thuringiensis crops: fitness cost affecting paternity. Evolution. 2005;59:915-920.
27. Gassmann AJ et al. Fitness costs of insect resistance to Bacillus thuringiensis. Annu Rev Entomol. 2009;54: 147-163.
28. Akhurst RJ et al. Resistance to the Cry1Ac delta-endotoxin of Bacillus thuringiensis in the cotton bollworm, Helicoverpaarmigera (Lepidoptera: Noctuidae). J Econ Entomol. 2003;96: 1290-1299.
29. Rahman M et al. Developmental penalties associated with inducible tolerance in He/icoverpaarmigera to insecticidal toxins from Bacillus thuringiensis. Appl Environ Microbiol. 2011;77:1443-1448.
30. Kim M et al. Isolation and characterization ofmicrosatellite loci from Spodopteraexigua (Lepidoptera: Noctuidae). Applied Entomology and Zoology. 2012;47:149-152.
31. Pavinato VAC et al. Microsatellite markers for genetic studies of the fall armyworm, Spodopterafrugiperda Genet Mol Res. 2013;12:370-380.
32. Tay WT et al. Generation of microsatellite repeat families by RTE retrotransposons in lepidopteran genomes BMC Evol Biol. 2010;10:144.
33. Arias RS et al. Isolation and characterization of the first microsatellite markers for Cyperusrotundus. Weed Research. 2011;51:451-460.
34. Arias RS et al. Development of Simple Sequence Repeat Markers for Chionanthusretusus (Oleaceae) and Effective Discrimination of Closely Related Taxa. HortScience. 2011;46:23-29.
35. Arias RS et al. Discriminating microsatellites from Macrophominaphaseolina and their potential association to biological functions. Plant Pathology. 2011;60:709-718.
36. Arias RS et al. Microsatellites reveal genetic diversity in Rotylenchulusreniformis populations. J Nematol. 2009;41:146-156.
37. Gould Fet al. Broad-spectrum resistance to Bacillus thuringiensis toxins in Heliothisvirescens. ProcNatlAcadSci USA. 1992;89:7986-7990.
38. Liu YB and Tabashnik BE Inheritance of resistance to the Bacillus thuringiensis toxin Cry1C in the diamondback moth. Appl Environ Microbiol. 1997;63:2218-2223.
39. Sayyed AH et al. Genetic and biochemical characterization of field-evolved resistance to Bacillus thuringiensis toxin Cry lAc in the diamondback moth, Plutellaxylostella. Appl Environ Microbiol. 2004;70:7010-7017.
40. Machado Vet al. Molecular characterization of host strains of Spodopterafrugiperda (Lepidoptera :Noctuidae) in Southern Brazil. Annals of the Entomological Society of America. 2008;101:619- 626.
41. Roelofs WLet al. Evolution of moth sex pheromones via ancestral genes. ProcNatlAcadSci USA. 2002;99:13621-13626.
42. O'Connell RJ. Olfactory receptor responses to sex pheromone components in redbandedleafroller moth. The J Gen Physiol. 1975;65:179-205.
43. Merlin Cet al. Antennal esterase cDNAs from two pest moths, Spodopteralittoralis and Sesamianonagrioides, potentially involved in odourant degradation. Insect Mol Biol. 2007;16: 73-81.
44. Ishida Y and Leal WSRapid inactivation of a moth pheromone. ProcNatlAcadSci USA. 2005;102:14075-14079.
45. Kaissling KE. Chemo-electrical transduction in insect olfactory cells. Annual Review of Neuroscience. 1986;9:121-145.
46. Sakurai Tet al. A single sex pheromone receptor determines chemical response specificity of sexual behavior in the silkmothBombyxmori. Plos Genetics. 2011;7: e1002115.
47. Nair R et al. Inheritance ofCry1Ac resistance and associated biological traits in the cotton bollworm, Helicoverpaarmigera (Lepidoptera: Noctuidae). J InvertebrPathol. 2010;104:31-38.
48. Sherengul W et al. Analysis of paternal transmission of mitochondrial DNA in Drosophila. Genes Genet Syst. 2006;81:399-404.
49. Gahan LJ et al. Identification of a gene associated with Bt resistance in Heliothisvirescens. Science. 2001;293:857-860.
50. Pigott CR et al. Role of receptors in Bacillus thuringiensis crystal toxin activity. MicrobiolMolBiol Rev. 2007; 71:255-281.