COMMODITY TREATMENT AND QUARANTINE ENTOMOLOGY Detection and Genetic Diversity of a Heliothine Invader (: ) From North and Northeast of Brazil

T. MASTRANGELO,1,2 D. F. PAULO,1 L. W. BERGAMO,1 E.G.F. MORAIS,3 M. SILVA,4 4 1 G. BEZERRA-SILVA, AND A.M.L. AZEREDO-ESPIN

J. Econ. Entomol. 107(3): 000Ð000 (2014); DOI: http://dx.doi.org/10.1603/EC13403 ABSTRACT The cotton bollworm, armigera (Hu¨ bner), was recently introduced in Brazil. During the 2012Ð2013 harvest, producers reported reduced yields up to 35% on major crops. The economic losses reached US$ 1 billion only in western Bahia, triggering a phytosanitary crisis. The deÞciencies in existing taxonomic keys to deal with the morphologically indistinct larvae of H. armigera and the native Helicoverpa zea (Boddie) constrained the detection of new incursions of this heliothine invader. This study explored the identity of heliothine larvae that were found infesting soybean- and corn-growing areas from Roraima state, northern Brazil, through sequences of the mitochondrial cytochrome c oxidase subunit I gene. The inter- and intraspecies sequence variations of DNA barcodes in H. armigera and H. zea were analyzed. The genetic diversity and population structure of the specimens from Roraima and two populations from Piauõ´ and Bahia states, northeastern Brazil, were assessed by adding the cytochrome c oxidase subunit II gene to the analysis. Owing to the lack of studies on genetic introgression for the two species, the suitability of using three different nuclear genes to distinguish the two species was also investigated. The results showed strong evidence that the heliothine larvae from north and northeast of Brazil are conspeciÞc with H. armigera, suggesting that this invasive has already crossed the Amazon basin. Surveys in the north of South America should start as soon as possible to monitor the entry or spread of this moth in the Caribbean, Central America, and the United States.

KEY WORDS Helicoverpa armigera, mtDNA, elongation factor-1␣, isocitrate dehydrogenase, ribo- somal protein S5

The cotton bollworm, Helicoverpa armigera (Hu¨ bner) Owing to its dispersal capability and omnivorous (Lepidoptera: Noctuidae), is a polyphagous species, behavior, H. armigera is the widest distributed species described as feeding on 181 plant species from Ͼ60 in the genus Helicoverpa (Venette et al. 2003) and families (Venette et al. 2003, Pogue 2004, Srivastava et poses a constant threat in international trade of many al. 2005). Each female moth can lay singly up to 1,500 horticultural crops (Lammers and MacLeod 2007). eggs on several parts of the host preferably at night Venette et al. (2003) reported 4,431 interceptions of (European and Mediterranean Plant Protection Or- Helicoverpa species in the United States since 1985 on ganization [EPPO] 1981). Larvae develop through six fruits, vegetables, and ornamentals. From 2000Ð2004, instars and feed on both vegetative and reproductive Helicoverpa spp. was intercepted Ͼ1,400 times at structures of the host (Wang and Li 1984). Large United States ports, and most samples were identiÞed larvae (longer than 2.4 cm) are the most damaging to two species only, H. armigera and Helicoverpa as- stage, as they consume Ϸ80% of their overall diet in sulta (Guenee) (Passoa 2004). the Þfth and sixth instars (Srivastava et al. 2005). Once Until recently, H. armigera was listed as an A1 quar- larvae are fully grown, they pupate in the soil and can antine pest (not present but of potential economic go into diapause to overwinter (Karim 2000). H. ar- importance) in Brazil (Czepak et al. 2013, Embrapa migera is a polyvoltine species and, depending on the 2013). However, disturbing reports started coming weather conditions, the entire life cycle can be com- from a two million hectares agricultural area in the pleted in 4Ð6 wk (Fitt 1989). cerrado biome that encompasses parts of the Brazilian states of Maranha˜o, Tocantins, Piauõ´, and Bahia (also known as the MATOPIBA region) in early 2013. Pro- 1 Centro de Biologia Molecular e Engenharia Gene´tica, Universi- dade Estadual de Campinas, Avenida Candido Rondon 400, 13083-875, ducers, especially from western Bahia, observed pop- Campinas, SP, Brazil. ulation levels never seen before of larvae morpholog- 2 Corresponding author, e-mail: [email protected]. ically similar to Helicoverpa zea (Boddie), and the 3 Embrapa Roraima, BR 174, km 8, Distrito Industrial, Caixa Postal 2012Ð2013 harvest was severely affected. Infestations 133, 69301-970, Boa Vista, RR, Brazil. 4 Universidade Estadual do Piauõ´ (UESPI), Rua Almir Benvindo s/n, of larvae reduced yields up to 35% on soybean, cotton, 64860-000, Uruc¸uõ´, PI, Brazil. corn, beans, sorghum, and millet. Most producers in-

0022-0493/14/0000Ð0000$04.00/0 ᭧ 2014 Entomological Society of America 2JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 107, no. 3 creased the number of pesticide applications by Brazil is still not known. In July 2013, heliothine larvae Ͼ15%, and the costs in cotton Þelds, for example, were found infesting soybean- and corn-growing areas jumped from US$ 400 to US$ 800 per hectare. The from Roraima state, northern Brazil. The major goal of extensive economic losses reached US$ 1 billion by this study was, therefore, to identify these larvae from July 2013, generating a phytosanitary crisis at the the north by using sequences of mitochondrial genes. MATOPIBA region (Embrapa 2013). Research insti- The interspeciÞc and intraspeciÞc sequence variation tutes identiÞed the new pest as H. armigera based of the COI gene fragment (“DNA barcode”) in the two mainly on the morphology of male genitalia (Hard- morphologically alike species, H. armigera and H. zea, wick 1965, Pogue 2004) and also conÞrmed its pres- was also analyzed. The genetic diversity and popula- ence in other two states from central Brazil (Goia´s and tion structure among the specimens from Roraima and Mato Grosso; Czepak et al. 2013, Tay et al. 2013). In two populations from Piauõ´ and Bahia states, north- 1992, H. armigera provoked a similar crisis in North eastern Brazil, were assessed by adding a second mi- China, with estimated losses of US$ 1.3 billion to ag- tochondrial gene, cytochrome c oxidase subunit II gene ricultural crops in the Yellow River cotton region (COII), to the analysis. (Sheng 1993). Many aspects of heliothine phylogeny remain un- Infestations in legumes and other horticultural clear by either morphology or some molecular mark- crops are being reported in other Brazilian states, but ers (Cho et al. 2008), and the genetic boundaries producers and entomologists are having trouble iden- between distinct lineages of H. armigera and H. zea tifying the pest. The based on morpholog- were not established yet owing to the lack of studies ical characters of Helicoverpa spp. is complicated on mitochondrial introgression for the two species. In (Pogue 2004) and presents three problems. First, as view of this, the suitability of using three different the available morphological keys are effective only for nuclear genes, which are used in phylogenomic stud- a particular life stage or gender, a large number of ies with lepidopterans owing to their stability and specimens collected from the Þeld are discarded be- considerable nucleotide substitution rates (Wahlberg cause they cannot be identiÞed. Second, the use of and Wheat 2008), was investigated for the populations these keys demands a high level of expertise, which analyzed. increases the frequency of incorrect identiÞcations. Finally, as H. armigera and H. zea are capable of in- Materials and Methods terbreeding to produce fertile offspring (Laster and Hardee 1995), the nonmapped phenotypic variability Specimens and DNA Extraction. In total, 65 he- in the characters of the hybrids could lead to misdi- liothine larvae suspected to be H. armigera were col- agnosis of the two species. lected from infested farms of eight geographic sites H. armigera and H. zea are difÞcult to distinguish from the states of Roraima (n ϭ 14), Piauõ´ (n ϭ 39), especially in the larval stage. Currently, only genitalia and Bahia (n ϭ 12). The specimens from Bahia came and wing characteristics of male adults are being used from populations from which adult had been to discriminate the two species (Pogue 2004, Czepak previously identiÞed as H. armigera based on morpho- et al. 2013). The lack in Brazil of sufÞcient number of logical characters (Czepak et al. 2013). Larvae were trained taxonomic experts in Lepidoptera and the de- collected directly from host plants and immediately Þciencies in existing morphology-based taxonomic preserved in 100% ethanol at Ϫ20ЊC until DNA ex- keys to deal with the morphologically indistinct larvae traction. of the two species prompted the need for a new iden- Total DNA was extracted from the last three ab- tiÞcation approach. dominal segments of the larvae (Behere et al. 2013) In the past decade, Hebert et al. (2003) proposed using the phenol: chloroform method, adapted for the Þrst half of the mitochondrial cytochrome c oxidase microcentrifuge tubes (Lyra et al. 2009), resuspended subunit I gene (COI) as the core (“DNA barcode”)of in 100 ␮l of TE buffer and stored at Ϫ20ЊC. The DNA a global identiÞcation system for metazoans and from six H. zea larvae reared at laboratory for four showed a model COI proÞle that was 100% successful generations (Callahan 1962) at the Piauõ´ State Uni- in correctly identifying 200 specimens from closely versity (UESPI) and two wild Spodoptera frugiperda allied species of lepidopterans. Sequences of COI and (Smith) (Lepidoptera: Noctuidae) larvae (collected other genes from mitochondrial DNA (mtDNA) have from corn in BonÞm, Roraima) was also extracted and been successfully used for population genetic studies used as control for some of the molecular analyses. of Lepidoptera groups (Caterino et al. 2000, Behere et Absence of cross-contamination during the extrac- al. 2007, Albernaz et al. 2012). Armstrong and Ball tions was conÞrmed by inclusion of a blank extraction (2005) also demonstrated the suitability of DNA bar- among each extraction batch. codes in providing a practical, cost-effective, and PCR Amplification and Sequencing. The partial se- ßexible framework for the accurate diagnostic of quences of the COI gene were ampliÞed by PCR using morphologically indistinct intercepted specimens. In- the primers COIF (5Ј-ATTCAACCAATCATAAA creasingly, molecular diagnostic tests are accepted as GATATTGG-3Ј) and COIR (5Ј-TAAACTTCTGGAT an essential component of detection and identiÞcation GTCCAAAAAATCA-3Ј; Li et al. 2011). The primers systems of exotic invasive species (Jenkins et al. 2012). A-tLEU (5Ј-ATGGCAGATTAGTGCAATGG-3Ј) and The exact distribution and extension of economic B-tLYS (5Ј-GTTTAAGAGACCAGTACTTG-3Ј; Liu damages of the recently introduced H. armigera in and Beckenbach 1992) were used to amplify part of June 2014 MASTRANGELO ET AL.: DETECTION OF H. armigera IN RORAIMA 3 the COII gene. Partial sequences of three protein-en- GQ892857, GQ892859), Þve H. zea (“H. zea 01Ð03” were coding nuclear genes, namely elongation factor-1␣ from UESPI laboratory, “H. zea 04” was a laboratory (EF-1␣), isocitrate dehydrogenase (IDH), and ribosomal strain from Mississippi (EU768942), and “H. zea 05” was protein S5 (RpS5), were ampliÞed by the pairs of primers a Brazilian voucher (HQ571107)), and the 16 H. armig- HibEfrcm4 (5Ј-ATTAACCCTCACTAAAGACAGCV era haplotypes (arm01Ðarm16; EU768935, EU768936, ACKGTYTGYCTCATRTC-3Ј) and HibAlf (5Ј-TAAT GQ892840, GQ892845, GQ892846, GQ892848, GQ892849, ACGACTCACTATAGGGGAGGAAATYAARAARG GQ892850, GQ892853, GQ892854, GQ995238Ð995244) AAG-3Ј), IDHdeg27f (5Ј-TAATACGACTCACTATA reported by Li et al. (2011). GGGGGWGAYGARATGACNAGRATHATHTGG-3Ј) Behere et al. (2007) investigated the genetic diver- and IDHdegR(5Ј-ATTAACCCTCACTAAAGTTY sity of H. armigera from Asia and Africa, but the COI Ј TTRCAIGCCCANACRAANCCNCC-3 ), HibRPS5FF sequences of this study could not be compared with (5Ј-TAATACGACTCACTATAGGGATGGCNGARGA theirs because they used only a 511 bp fragment lo- Ј Ј RAAYTGGAAYGA-3 ) and HibRPS5FR (5 -ATTAA cated in the region 980Ð1490 of the COI gene that was CCCTCACTAAAGCGGTTRGAYTTRGCAACACG-3Ј; not overlapped with the region 39Ð696 analyzed in Wahlberg and Wheat 2008), respectively. this study and by Li et al. (2011). The PCR reactions were conducted separately with Neighbor-joining (NJ) distance analysis (Saitou Ϸ 25 ng of total DNA, 2.5 mM MgCl2, 0.25 mg/ml of and Nei 1987) and sequence divergences were calcu- BSA, 100 ␮M dNTPs, 0.5 mM of each primer, 1.5 U of lated in MEGA 5.1 software (Tamura et al. 2011) with TaqDNA polymerase (Fermentas International Inc., Kimura two-parameters (K2P; Kimura 1980) and un- Burlington, Canada), and 10ϫ TaqBuffer for a Þnal corrected sequence divergences (p-distances) mod- reaction volume of 25 ␮l. AmpliÞcation was carried els. Node supports were measured by 5,000 replicates out on a GeneAmpPCR System 9700 thermal cycler of bootstrap (BS). (Applied Biosystems, Foster City, CA), with the fol- The Bayesian inference (BI) and maximum-likeli- lowing conditions: an initial denaturation at 94ЊC for hood (ML) methods were also used for the phyloge- 3 min, followed by 35 cycles of 94ЊC for 30 s, 55ЊC for netic analyses. The best Þtted substitution model se- 30 s (45ЊC for COI and 48ЊC for COII), 70ЊC for 1:30 lection for the dataset was carried out using MrAIC s (72ЊC for COII), and ended with a 7 min Þnal ex- 1.4.4 (Nylander 2004) software. The favored model for tension at 70ЊC (72ЊC for COII). After ampliÞcation, COI was the GTRϩI (General Time Reversible; I ϭ 2-␮l aliquots were analyzed by 1% agarose gel elec- invariable sites). BI analysis was conducted using mr- trophoresis in 1ϫ TAE (40 mM Tris-acetate, 1 mM Bayes v3.1.2 (Ronquist and Huelsenbeck 2003) with EDTA). the favored substitution model selected previously. Amplicons were puriÞed with the IllustraGFX kit Two independent analyses were run for 10,000,000 (GE Healthcare, Bucks, UK) and sequenced bidi- generations (sample frequency ϭ 1,000), with 25% of rectionally, to ensure correct basecalling, by the burn-in after checking for convergence. Node sup- ABI3730xl DNA Analyzer sequencer (Applied Bio- ports were analyzed by their posterior probabilities in systems, Foster City, CA), with the same primers the 70% majority rule consensus tree. The ML analysis used for the PCR reactions. was conducted using PhyML (Guindon and Gascuel Sequence Analyses. Sequences were assembled into 2003) with SPR tree topology search operation, 10 a contig for each specimen by using the Geneious 6.0.6 random starting trees as parameters and the MrAIC Software (Biomatters Ltd., Auckland, New Zealand), 1.4.4-favored substitution model. Node supports were considering Phred values (Ewing et al. 1998). All se- accessed with 1,000 replicates of BS. quences were aligned for each of the Þve genes sep- The DnaSP.V5 software (Librado and Rozas 2009) arately using the multiple sequences alignment algo- was used to investigate polymorphisms in COI se- rithm implemented in Clustal ⍀ (Sievers et al. 2011). quences and to identify possible diagnostic characters Protein coding sequences were checked for the pres- for species discrimination between H. armigera and H. ence of open reading frames in MEGA 5.1 Software zea. (Tamura et al. 2011). A haplotype network was constructed based on the Phylogenetic Analyses and Sequence Divergence. 96 COI sequences used in the phylogenetic analyses Only COI sequences were used for phylogenetic anal- by TCS version 1.21 (Clement et al. 2000), following yses owing to the lack of available sequences in public the guidelines proposed by Bandelt et al. (2000). The databases for the species and genetic regions investi- connection between haplotypes was limited by a gated in this study. The 65 COI fragments from the probability of parsimony for DNA pairwise differ- heliothine specimens from Roraima (RR01ÐRR14), ences lower than 95%. Piauõ´ (PI01ÐPI39), and Bahia (BA01ÐBA12) were used Structure and Genetic Diversity of the Populations to perform the phylogenetic analyses along with se- Sampled. The partial sequences of the COI and COII quences of three S. frugiperda (two from Roraima (“S. genes from the 65 heliothine specimens from Roraima, frugiperda (01)Ð(02)”) and a voucher under the Piauõ´, and Bahia were concatenated for the genetic GenBank GU090724 (“S. frugiperda (03)”)), one Heliothis diversity analyses. Individual sequences were col- virescens F. (Lepidoptera: Noctuidae) (GU087832), one lapsed in haplotypes and, in the absence of detailed Helicoverpa punctigera (Wallengren) (EU768941), one knowledge of gene ßow and for the purpose of this Helicoverpa gelotopoeon (EU768938), one Helicoverpa study, each one of the three Brazilian states sampled hawaiiensis (EU768939), three H. assulta (GQ892856, was regarded as a different “population” in the anal- 4JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 107, no. 3 yses. Haplotype frequencies, haplotype diversity (Hˆ ), The genetic distance (Table 1) between the inves- and nucleotide diversity (␲), as deÞned by Nei (1987), tigated heliothine specimens from Roraima, Piauõ´, were estimated using Arlequin v.3.5 (ExcofÞer and and Bahia and the 16 H. armigera COI sequences Lischer 2010). reported by Li et al. (2011) revealed very low se- Analysis of molecular variance (AMOVA) was per- quence divergence (0Ð0.8%), suggesting conspeci- formed to access the genetic structure among and Þcity among the specimens (Avise 2000). For in- within the three populations as implemented in Ar- stance, the sequence divergence between the 65 lequin v.3.5 (ExcofÞer and Lischer 2010). Genetic heliothine specimens and the H. zea group exceed differences among populations were determined that divergence, ranging approximately from 2 to through pairwise FST statistics (Reynolds et al. 1983, 3%, likely the range observed between the H. ar- Slatkin 1995). Statistical signiÞcance was accessed by migera group and H.zea (Table 1). For both distance 10,000 permutations, the computed distance matrix models used (K2P and p-distance), there was sub- used pairwise difference, and the gamma ␣ value was stantial sequence divergence between the 65 he- considered zero for the analysis. liothine specimens and the four other species of the Sequence Divergence Using Nuclear Genes. The genus Helicoverpa (Ϸ2.6Ð6.3%), He. virescens (Ϸ7Ð partial sequences of the nuclear genes EF-1␣, IDH, 8%), and S. frugiperda (Ϸ10Ð12%; Table 1). and RpS5 for 6 H. zea from laboratory rearing and 22 The sequences from the 65 heliothine specimens heliothine larvae from Roraima, Piauõ´, and Bahia (one (RR01ÐRR14; PI01ÐPI39; and BA01ÐBA12) and the 16 representative from each one of the 22 different hap- H. armigera (arm1Ðarm16; Li et al. 2011) were col- lotypes resulting from the concatenation of the COI lapsed into 21 different haplotypes, which were linked and COII fragments) were used for the estimation of in a unique parsimony network (Fig. 2). The general inter- and intraspeciÞc genetic distances between H. topology of the network showed one clade containing zea and H. armigera. Sequence divergence with K2P all H. armigera haplotypes, separated by Ͼ13 muta- (Kimura 1980) and uncorrected sequence divergence tional steps from H. zea and 11 steps from the other (p-distances) models were calculated by using MEGA heliothine species. The two most common H. armigera 5.1 software (Tamura et al. 2011). COI haplotypes, designated by Li et al. (2011) as “arm1” (found in Thailand and the Chinese province of Yunnan) and “arm2” (found in Yunnan and Henan provinces, China), are most likely to be ancestral Results (Castelloe and Templeton 1994). Five specimens from All specimens yielded high-quality DNA and were Roraima, 18 from Piauõ´, and 8 from Bahia were iden- successfully sequenced for the fragments of COI (658 tical to “arm1,” and 6, 11, and 2 specimens from Ro- bp), COII (554 bp), EF-1␣ (779 bp), IDH (560 bp), raima, Piauõ´, and Bahia, respectively, were identical to and RpS5 (593 bp) genes. The COI sequences ob- “arm2” (Fig. 2). Most of the other H. armigera haplo- tained for the specimens from Roraima, Piauõ´, and types were rare (single individuals) and their fre- Bahia were blasted against GenBank database and quency was comparatively high (15 out of 21). Hap- checked in BOLD IdentiÞcation System (BOLD lotypes of H. armigera differed from the center of the 2007), and all the 65 heliothine larvae matched 99Ð network by no more than three mutational steps. No 100% with H. armigera sequences. Chromatograms haplotype was shared among the six Helicoverpa spe- were reliable for all the gene fragments and, after cies, and He. virescens differed by 62 mutational steps terminal cutoff owing to sequences ambiguities, the from the H. armigera group (Fig. 2). absence of indels made the alignment straightforward. The DNA barcodes generated for H. armigera and The mtDNA and nuclear sequences generated in this H. zea (Table 2) based on combinations of diagnostic study have been deposited in GenBank (KF624811Ð nucleotides allowed the observation of 30 nucleotide KF625029). substitutions between the two species, with 14 unique Phylogenetic Analyses and Sequence Divergence. diagnostic characters that could be used to discrimi- Four different phylogenetic analyses (NJ under p-dis- nate the species. tance and K2P models, BI, and ML under MrAIC Structure and Genetic Diversity of the Populations 1.4.4-favored substitution model) conducted for the Sampled. The COI and COII fragments from the 65 COI sequences gave nearly identical tree topologies specimens from Roraima, Piauõ´, and Bahia were con- (Fig. 1). NJ and BI proved to be reliable measures of catenated, generating a Þnal 1212-bp sequence for species delimitation, as all identiÞable species with each individual. In total, 22 different haplotypes were more than one sequence (S. frugiperda, H. assulta, and deÞned (Fig. 3). Haplotypes were shared among sam- H. zea) had all individuals joined by Ͼ98% BS and 1.0 ples from the three Brazilian states, especially the posterior probability (PP). The 65 heliothine speci- haplotype one, the most frequent, suggesting a signif- mens from Roraima, Piauõ´, and Bahia joined by Ͼ95% icant gene ßow among these regions. In general, high BS and 0.9 PP values formed a monophyletic group haplotype diversity (Hˆ -mean Ͼ 0.85) and low nucle- with the 16 H. armigera sequences reported by Li et al. otide diversity (␲-mean Ͻ 0.003) were observed for (2011). The H. armigera group formed a robustly sup- the locations analyzed (Fig. 3). AMOVA detected no ϭ Ϫ5 ϭ ported monophyletic clade with H. zea (80%/77% NJ genetic structure (FST 7.10 ; P 0.44), with 99.9% BS, 0.91 BI PP, and 79% ML BS), as expected for these of variation accounted for at the within population species (Cho et al. 2008). level and with only 0.01% variation observed among June 2014 MASTRANGELO ET AL.: DETECTION OF H. armigera IN RORAIMA 5

Fig. 1. NJ strict consensus tree (topology under p-distance model) inferred using COI complete dataset for Heliothine species. The 65 specimens from Roraima (RR01ÐRR14), Piauõ´ (PI01ÐPI39), and Bahia (BA01ÐBA12) are in red. Sequences of S. frugiperda, He. virescens, and H. punctigera were used as outgroups. Numbers above branches refer to NJ BS proportions among 5,000 replicates, while numbers below branches refers to BI PP and ML BS proportions among 1,000 replicates. The four analyses gave nearly identical tree topologies, and node supports values below 70% (BS), 0.7 (PP), or both, were not recorded in the tree (Ϫ).

populations. Pairwise FST values (considering the in- clear genes were low (K2P and uncorrected p-dis- dividual three Brazilian states as separate populations) tance Ͻ3%; Table 4). The highest number of variable Ͻ Ͼ were low and not signiÞcant (FST 0.03; P 0.05) in sites (2.7%) was observed for the fragments of the the studied specimens (Table 3). RpS5 gene, but the highest sequence divergence Sequence Divergence Using Nuclear Genes. The (2.2%) was given by the EF-1␣ gene. As the intras- genetic distances between 22 heliothine specimens peciÞc range of differences from both H. armigera and from Roraima, Piauõ´, and Bahia and 6 H. zea from H. zea overlapped the interspeciÞc ranges (Table 4), laboratory rearing estimated by three different nu- the three nuclear gene fragments investigated failed to 6JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 107, no. 3

Table 1. Mitochondrial DNA COI pairwise genetic distances for the investigated heliothine specimens from Brazil and other Lepidoptera species

Kimura two-parameters (K2P) Uncorrected sequence Species compared divergences (p-distance) Range Range Investigated Heliothine specimensa vs S. frugiperda 0.114Ð0.120 0.099Ð0.103 He. virescens 0.076Ð0.079 0.068Ð0.071 H. punctigera 0.058Ð0.063 0.053Ð0.058 H. gelotopoeon 0.035Ð0.040 0.033Ð0.038 H. assulta 0.032Ð0.37 0.030Ð0.035 H. hawaiiensis 0.027Ð0.032 0.026Ð0.030 H. zea 0.023Ð0.027 0.024Ð0.029 H. armigera 0.000Ð0.008 0.000Ð0.008 H. armigerab vs H. zea 0.026Ð0.031 0.024Ð0.029

a The investigated heliothine specimens comprised the 65 larvae sampled from Roraima, Piauõ´, and Bahia states. b The H. armigera group encompassed only the 16 COI haplotypes described by Li et al.(2011). robustly discriminate the two morphologically alike cies. However, as rates of genetic change between taxa species. are dynamic processes (Rubinoff et al. 2006), thresh- olds used to deÞne species may vary widely among studies with lepidopterans. Hebert et al. (2004) used Discussion a threshold as low as 0.32% for butterßies, and The mtDNA COI phylogeny strongly supported Whinnett et al. (2005) observed divergences from 0.23 that the 65 heliothine larvae from Roraima, Piauõ´, and to 6.4% in Ithoninae butterßies, demonstrating that Bahia are conspeciÞc with H. armigera. The four phy- species thresholds can be extremely ßuid. logenetic analyses grouped these specimens with 16 H. The mtDNA COI haplotype network showed that armigera COI sequences described by Li et al. (2011) the 65 heliothine samples from Roraima, Piauõ´, and as a single monophyletic group. The other deÞned Bahia formed a homogeneous group of H. armigera species (S. frugiperda, He. virescens, and the other Þve haplotypes (Fig. 2). The connections between haplo- Helicoverpa species) were recovered in distinct posi- types was similar to that reported by Li et al. (2011), tions or clades on strongly supported branches (Fig. with H. armigera haplotypes differing from the center 1). The H. zea clade was monophyletic with the H. of the network by no more than three mutational steps armigera clade, and this same phylogenetic pattern and no shared haplotypes among the different species. had been observed in previous studies (Fang et al. Two Chinese haplotypes, “arm1” (EU768935) and 1997, Cho et al. 2008). Behere et al. (2007) suggested “arm2” (EU768936; Cho et al. 2008, Li et al. 2011), that the American H. zea populations were established encompassed 47.7 and 29.2%, respectively, of the hap- via founder event from H. armigera (or from their joint lotypes found in the samples from Roraima, Piauõ´, and common ancestor; Mallet et al. 1993) no Ͼ1.5 million Bahia, indicating that “arm1” is a potential ancestral years ago, which is further supported by the high haplotype of H. armigera. H. zea appeared linked to the similar morphology and the possibility of mating com- H. armigera group (by 15 mutational steps) through patibility between the two species (Laster et al. 1985, the haplotypes from samples PI-34, PI-3, PI-4, and Laster and Hardee 1995, Laster and Sheng 1995). RR-7 (Fig. 2). In the haplotype network reported by Despite the recent history of separation between Behere et al. (2007), the H. armigera and H. zea groups the two species (Behere et al. 2007) and the low were separated by Ϸ20 mutational steps. sequence divergences expected for the order Lepi- This study observed shared haplotypes between the doptera (Caterino et al. 2000, Hebert et al. 2003), the three Brazilian states, not signiÞcant FST values, and information content of the COI fragment selected in low nucleotide diversity between the sampled regions this study was sufÞcient to enable recognition of the from north and northeast of Brazil (Table 3; Fig. 3). In divergence among the lepidopterans analyzed. The general, high values of haplotype diversity and low estimated genetic distances (K2P and p-distance) in- nucleotide diversity, combined with a high number dicated divergence values between species Ͼ2%. Set- (15 out of 22) of low frequency haplotypes (Fig. 3), ting this value as a threshold for species diagnosis in are characteristic of species that have undergone a this study allowed the discrimination of all heliothine process of recent population expansion (ExcofÞer et species analyzed (Table 1). The low divergence (0Ð al. 2009). 0.8%) between the 65 heliothine specimens from Many authors, using different molecular methods, north and northeast of Brazil and the H. armigera have veriÞed similar patterns of genetic variation, sequences from Li et al. (2011) indicated conspeci- which seems to be common in pests capable of Þcity among these individuals, and both groups di- rapid spatial expansion and long-range dispersion such verged similarly with H. zea (Ϸ2Ð3% for K2P and as H. armigera. Daly and Gegg (1985) observed very ϭ p-distance;Table 1). Hebert et al. (2003) also desig- little genetic variation (FST 0.012) between Aus- nated a 2Ð3% mtDNA COI sequence divergence as a tralian H. armigera populations from a 3,000-km study “threshold” to discriminate insect and mammal spe- area using isozymes. Nibouche et al. (1998) veriÞed June 2014 MASTRANGELO ET AL.: DETECTION OF H. armigera IN RORAIMA 7

Fig. 2. Haplotype network based on partial mtDNA COI (658 bp) of heliothine species. Each haplotype is represented by a circle. The original names of the haplotypes reported by Li et al. (2011) (arm1Ðarm16) and the 65 specimens from Roraima (RR01ÐRR14), Piauõ´ (PI01ÐPI39), and Bahia (BA01ÐBA12) were maintained. Full lines between haplotypes represent one mutational step, while the dots are presumptive intermediate haplotypes that were not observed. that barriers such as the Sahara desert had not pre- spective of cropping seasons, were inferred from vented long-distance migration in H. armigera, as no EPIC-PCR DNA markers (Behere et al. 2013). signiÞcant isozyme allele frequency differences were Long-distance migration of this heliothine pest has found between populations from either side of the also been suggested by capture of adult moths on desert. Zhou et al. (2000) reported very low genetic Ascension Island (2,000 km from the African coast; distances among Israeli and Turkish H. armigera pop- Widmer and SchoÞeld 1983) and Willis Island in the ulations using RAPD-PCR analysis. Using microsatel- Coral Sea (450 km from the coast of Australia; Daly lite loci, Endersby et al. (2007) found no evidence of and Gregg 1985). Borne by wind, H. armigera can genetic structure among samples from Australia and reach the United Kingdom from southern Europe New Zealand collected at different times. Behere et al. (Ͼ1,000 km; Pedgley 1985). MarkÐrecapture experi- (2007) suggested the occurrence of long-distance ments have shown that H. armigera moths could ßy gene ßow in H. armigera, based on low FST values and 200Ð300 km in a single night (Armes and Cooter 1991). low among-group haplotype variance across Australia, Correct and prompt detection of this highly mobile China, India, Pakistan, Burkina Faso, and Uganda. No invasive pest is an essential step before initiating quar- signiÞcant population substructure across India, irre- antine actions, proper control measures, and quick 8JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 107, no. 3

Table 2. DNA barcodes for H. armigera and H. zea based on Table 3. Pairwise FST estimates among three investigated he- combinations of diagnostic nucleotides represented by the Inter- liothine populations from Brazil based on combined genes COI and national Union of Pure and Applied Chemistry symbols COII

N.Pa H. armigera H. zea N.P H. armigera H. zea Populations Piauõ´ Bahia Roraima 40 R A 340 C T Piauõ´ 0.0000 85 Y T 346 M A Bahia 0.0255 (Pϭ0.15) 0.0000 115 Tb C 368 R G Roraima Ϫ0.0056 (Pϭ0.48) Ϫ0.041 (Pϭ0.91) 0.0000 127 T C 385 Y T 130 R A 386 C T 181 K T 418 C T 217 A G 439 Y T strong and Ball 2005, Darling and Blum 2007). There 238 C T 477 G A 262 R A 478 C T are some shortcomings of mtDNA for “barcoding” 268 T C 496 R A identiÞcation, which includes recombination, incon- 274 Y T 536 T C sistent mutation rate, heteroplasmy, maternal inheri- 277 T C 539 Y T tance, and introgression (Rubinoff et al. 2006, Krish- 278 T C 548 T C 286 R A 595 Y C namurthy and Francis 2012). Although the possibility 313 Y T 616 Y T of mating incompatibility via mechanical isolation be- tween H. armigera and H. zea was reported (Hardwick a N.P. indicates the nucleotide position which is relative to the 1965), mating compatibility between laboratory H. zea beginning of the fragment investigated in this study. from the United States and wild H. armigera from Asia b The 14 pure diagnostic characters are shaded in bold italics. and Australia under controlled conditions (Laster et al. 1985, Laster and Hardee 1995, Laster and Sheng response to any incursion (Li et al. 2011). The current 1995) is also possible. These studies, allied to the in- study demonstrated the effectiveness of DNA bar- existence of hybridization trials between H. armigera codes for distinguishing H. armigera from H. zea (Ta- and H. zea populations from Central and South Amer- ble 2). The use of DNA barcodes has also succeeded ica, raise the question if the 65 heliothine larvae from in species-level identiÞcation for Ͼ50,000 species of north and northeast of Brazil are from a pure strain of Lepidoptera so far (Hebert et al. 2004, Hajibabaei et the recently introduced H. armigera or a hybrid orig- al. 2006, Burns et al. 2008, Silva-Branda˜o et al. 2009, inated from the cross between a H. zea male and a H. Jinbo et al. 2011). armigera female (Caterino et al. 2000, Hebert et al. Despite the high resolution to distinguish the six 2003). Helicoverpa species analyzed, DNA barcodes should New gene regions have been successfully se- not be used blindly in the context of biosecurity (Arm- quenced for Lepidoptera and primers for the genes

Fig. 3. Heliothine sampled sites, haplotype distribution, and genetic variability indices (ϮSE). June 2014 MASTRANGELO ET AL.: DETECTION OF H. armigera IN RORAIMA 9

Table 4. Interspecies and intraspecies divergences given by three nuclear genes for the H. armigera and H. zea investigated

Genetic distances Gene Amplicon No. of InterspeciÞc Intra. H.armigera Intra. H. zea No. of specimens sequenced length variable sites K2P p-distance K2P p-distance K2P p-distance EF-1␣ 28 (6 H. zea and 22 H. armigera 779bp 17 (2.1%) 0.004Ð0.022 0.004Ð0.022 0Ð0.008 0Ð0.006 0Ð0.022 0Ð0.022 IDH 560bp 7 (1.25%) 0.004Ð0.018 0.004Ð0.018 0Ð0.008 0Ð0.008 0 0 RpS5 593bp 16 (2.7%) 0Ð0.019 0Ð0.018 0Ð0.017 0Ð0.016 0 0

EF-1␣, IDH, and RpS5 were suggested to be universal as soon as possible to monitor its entry or spread in the in Lepidoptera, giving robustness for most species Caribbean, Central America, and the United States. (Wahlberg and Wheat 2008). Trying to overcome the The increasing rates of H. armigera infestations in unresolved hybridization issue, sequence divergence Brazil were favored by many conditions (tropical cli- between H. zea and H. armigera was also explored mate, extensive agriculture, wide host availability, in- using these three nuclear genes. Nonetheless, the vari- secticide resistance, and drought in northeastern ation of the fragments analyzed was not sufÞcient to states; Behere et al. 2013, Czepak et al. 2013). Finding robustly discriminate H. armigera from H. zea (Table similar conditions in other regions of America, this 4). The highest sequence divergence obtained was heliothine invader could trigger more phytosanitary 2.2%, but inter- and intraspeciÞc divergences were crisis. overlapped. In general, nuclear DNA undergoes rel- atively slow mutation compared with mtDNA, which may provide for the Þrst a smaller degree of taxonomic Acknowledgments resolution at the species level (Waugh 2007). A mono- We thank Francisco dos Santos Silva from UESPI and phyletic relationship between H. zea and H. armigera ␣ Suely Brito from the Ageˆncia de Defesa Agropecua´ria da was inferred previously based on the EF-1 gene (Cho Bahia (ADAB) for some heliothine samples from Piauõ´ and et al. 1995) and the Dopa Decarboxylase (DDC) gene Bahia states, respectively. We are thankful to Rosangela A. (Fang et al. 1997). The insufÞcient variation in the Rodrigues for technical and laboratory assistance, and to the EF-1␣, IDH, and RpS5 genes to accurately reconstruct Fundac¸a˜o de Amparo a` Pesquisa do Estado de Sa˜o Paulo the recent divergence between the species may be a (FAPESP) for a fellowship. consequence of the common ancestry (Mallet et al. 1993, Behere et al. 2007). Further studies to search for combining evidence References Cited from mtDNA and more rapidly evolving nuclear Albernaz, K. C., K. L. Silva-Brandao, P. Fresia, F. L. Consoli, DNA markers with pure strains and hybrids from H. and C. Omoto. 2012. Genetic variability and demo- armigera and H. zea will be necessary to enhance the graphic history of Heliothis virescens (Lepidoptera: Noc- robustness of any key and assignment of identiÞca- tuidae) populations from Brazil inferred by mtDNA se- tion for these two species. While comprehensive quences. Bull. Entomol. Res. 102: 333Ð343. protocols with new molecular markers giving higher tax- Armstrong, K. F., and S. L. Ball. 2005. DNA barcodes for biosecurity: invasive species identiÞcation. Phil. Trans. R. onomic resolution are not developed, mtDNA-based Soc. B 360: 1813Ð1823. identiÞcation systems will undoubtedly continue to pro- Armes, N. J., and R. J. Cooter. 1991. Effects of age and mated vide diagnostic and geographic origin information that status on ßight potential of Helicoverpa armigera (Lepi- exceeds that which can be achieved by using morpho- doptera: Noctuidae). Physiol. Entomol. 16: 131Ð144. logical studies alone (Armstrong and Ball 2005, Behere et Avise, J. C. 2000. Phylogeography: the history and formation al. 2007). The mtDNA-based systems can also help to of species. Harvard University Press, Cambridge, MA. speed the work of policy makers interested in the an- Bandelt, H. J., V. Macaulay, and M. Richards. 2000. Median ticipatory detection of heliothine invasive species. networks: speedy construction and greedy reduction, one The strong evidence that the heliothine larvae from simulation, and two case studies from human mtDNA. Mol. Phylogenet. Evol. 16: 8Ð28. Roraima are conspeciÞc with H. armigera under sig- Behere, G. T., W. T. Tay, D. A. Russell, D. G. Heckel, B. R. niÞcant gene ßow with northeastern populations sug- Appleton, K. R. Kranthi, and P. Batterham. 2007. Mito- gest that this invasive moth is not restricted to crop chondrial DNA analysis of Þeld populations of Helicov- Þelds from northeast and central Brazil and probably erpa armigera (Lepidoptera: Noctuidae) and of its rela- already has crossed the Amazon basin. The data of the tionship to H. zea. BMC Evol. Biol. 7: 117Ð127. current study, however, do not allow to precise the Behere, G. T., W. T. Tay, D. A. Russell, K. R. Kranthi, and P. likely source of the invasions. Tay et al. (2013) were Batterham. 2013. Population genetic structure of the also unable to deÞne the origin of the H. armigera cotton bollworm Helicoverpa armigera (Hubner) (Lepi- introduced in Brazil, but suggested that the samples doptera: Noctuidae) in India as Inferred from EPIC-PCR DNA Markers. PLoS ONE 8: e53448. from Mato Grosso might have come from either mul- (BOLD) The Barcode of Life Data System. 2007. Bold.v3. tiple recent incursions or a single incursion from Mol. Ecol. Notes 7: 355Ð364. (http://www.boldsystems. mixed populations of H. armigera. org/index.php/IDS_OpenIdEngine). A broader geographic sampling in northern Brazil Burns, J. M., D. H. Janzen, M. Hajibabaei, W. Hallwachs, and and surveys in Venezuela and Colombia should start P.D.N. Hebert. 2008. DNA barcodes and cryptic species 10 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 107, no. 3

of skipper butterßies in the genus Perichares in Area de Hajibabaei, M., D. H. Janzen, J. M. Burns, W. Hallwachs, and Conservacion Guanacaste, Costa Rica. Proc. Natl. Acad. P.D.N. Hebert. 2006. DNA barcodes distinguish species Sci. U.S.A. 105: 6350Ð6355. of tropical Lepidoptera. Proc. Natl. Acad. Sci. U.S.A. 103: Callahan, P. S. 1962. Techniques for rearing the corn ear- 968Ð971. worm, Heliothis zea. J. Econ. Entomol. 55: 453Ð457. Hardwick, D. F. 1965. The corn earworm complex. Mem. Castelloe, J., and A. R. Templeton. 1994. Root probabilities Entomol. Soc. Can. 40: 1Ð247. for intraspeciÞc gene trees under neutral coalescent the- Hebert, P.D.N., A. Cywinska, S. L. Ball, and J. R. deWaard. ory. Mol. Phylogenet. Evol. 3: 102Ð113. 2003. Biological identiÞcations through DNA barcodes. Caterino, M. S., S. Cho, and F. A. Sperling. 2000. The cur- Proc. R. Soc. B. 270: 313Ð322. rent state of insect molecular systematics: a thriving Hebert, P.D.N., E. H. Penton, J. M. Burns, D. H. Janzen, and Tower of Babel. Annu. Rev. Entomol. 45: 1Ð54. W. Hallwachs. 2004. Ten species in one: DNA barcoding Cho, S., A. Mitchell, J. C. Regier, C. Mitter, R. W. Poole, T. P. reveals cryptic species in the neotropical skipper butter- Friedlander, and S. Zhao. 1995. A highly conserved nu- ßy Astraptes fulgerator. Proc. Natl. Acad. Sci. U.S.A. 101: clear gene for low-level phylogenetics: Elongation factor- 14812Ð14817. 1␣ recovers morphology-based tree for Heliothine moths. Jenkins, C., T. A. Chapman, J. L. Micallef, and O. L. Reynolds. Mol. Biol. Evol. 12: 650Ð656. 2012. Molecular techniques for the detection and differ- Cho, S., A. Mitchell, C. Mitter, J. Regier, M. Matthews, and entiation of host and parasitoid species and the implica- R. Robertson. 2008. Molecular phylogenetics of helioth- tions for fruit ßy management. 3: 763Ð788. ine moths (Lepidoptera: Noctuidae: ), with Jinbo, U., T. Kato, and M. Ito. 2011. Current progress in comments on the evolution of host range and pest status. DNA barcoding and future implications for entomology. Syst. Entomol. 33: 581Ð594. Entomol. Sci. 14: 107Ð124. Clement, M., D. Posada, and K. A. Crandall. 2000. TCS: a Karim, S. 2000. Management of Helicoverpa armigera: a re- computer program to estimate gene genealogies. Mol. view and prospectus for Pakistan. Pakistan J. Biol. Sci. 3: Ecol. 9: 1657Ð1659. 1213Ð1222. Czepak, C., K. C. Albernaz, L. M. Vivan, H. O. Guimara˜es, Kimura, M. 1980. A simple method for estimating evolu- and T. Carvalhais. 2013. First reported occurrence of tionary rates of base substitutions through comparative Helicoverpa armigera (Hu¨ bner) (Lepidoptera: Noctu- studies of nucleotide sequences. J. Mol. Evol. 16: 111Ð120. idae) in Brazil. Pesqui. Agropec. Trop. 43: 110Ð113. Krishnamurthy, P. K., and R. A. Francis. 2012. A critical Daly, J. C., and P. Gregg. 1985. Genetic variation in Heliothis review on the utility of DNA barcoding in biodiversity in Australia: species identiÞcation and gene ßow in two conservation. Biodivers. Conserv. 21: 1901Ð1919. pest species H. armigera (Hubner) and H. punctigera ¨ Lammers, J. W., and A. Macleod. 2007. Report of a pest risk Wallengren (Lepidoptera: Noctuidae). Bull. Entomol. analysis: Helicoverpa armigera (Hu¨ bner, 1808). (http:// Res. 75: 169Ð184. www.fera.defra.gov.uk/plants/plantHealth/pestsDiseases/ Darling, J. A., and M. J. Blum. 2007. DNA-based methods documents/helicoverpa.pdf). for monitoring invasive species: a review and prospectus. Laster, M. L., and D. D. Hardee. 1995. Intermating com- Biol. Invasions 9: 751Ð765. patibility between North-American Helicoverpa zea and (Embrapa) Empresa Brasileira de Pesquisa Agropecua´ria. Heliothis armigera (Lepidoptera: Noctuidae) from Rus- 2013. Ac¸o˜es emergenciais propostas pela Embrapa para o sia. J. Econ. Entomol. 88: 77Ð80. manejo integrado de Helicoverpa spp. para a´reas agrõ´colas, 2013. (http://www.cpac.embrapa.br/publico/usuarios/ Laster, M. L., and C. F. Sheng. 1995. Search for hybrid uploads/destaques/manejo_helicoverpa2.pdf). sterility for Helicoverpa zea in crosses between the North Endersby, N. M., A. A. Hoffmann, S. W. McKechnie, and A. R. American Heliothis zea and Helicoverpa armigera (Lepi- Weeks. 2007. Is there genetic structure in populations of doptera: Noctuidae) from China. J. Econ. Entomol. 88: Helicoverpa armigera from Australia? Entomol. Exp. Appl. 1288Ð1291. 122: 253Ð263. Laster, M. L., C. E. Goodpasture, E. G. King, and P. Twine. (EPPO) European and Mediterranean Plant Protection Or- 1985. Results from crossing the bollworms Helicoverpa ganization. 1981. Data sheets on quarantine organisms armigera x H. zea in search of backcross sterility, pp. nЊ 110: Helicoverpa armigera. Paris: EPPO. (Bulletin, 11). 146Ð147. In Proceedings, 1985, Beltwide Cotton Produc- Excoffier, L., and H.E.L. Lischer. 2010. Arlequin suite ver tion and Research Conference National Cotton Council 3.5: A new series of programs to perform population of America, Memphis, TN. genetics analyses under Linux and Windows. Mol. Ecol. Li, Q. Q., D. Y. Li, H. Ye, X. F. Liu, W. Shi, N. Cao, and Y. Q. Res. 10: 564Ð567. Duan. 2011. Using COI gene sequence to barcode two Excoffier, L., M. Foll, and R. J. Petit. 2009. Genetic conse- morphologically alike species: the cotton bollworm and quences of range expansions. Annu. Rev. Ecol. Evol. Syst. the oriental tobacco budworm (Lepidoptera: Noctu- 40: 481Ð501. idae). Mol. Biol. Rep. 38: 5107Ð5113. Ewing, B., L. Hillier, M. Wendl, and P. Green. 1998. Base Librado, P., and J. Rozas. 2009. DnaSP v5: A software for calling of automated sequencer traces using phred. I. comprehensive analysis of DNA polymorphism data. Accuracy assessment. Genome Res. 8: 175Ð185. Bioinformatics 25: 1451Ð1452. (http://bioinformatics. Fang, Q. Q., S. Cho, J. C. Regier, C. Mitter, M. Matthews, oxfordjournals.org/content/25/11/1451). R. W. Poole, T. P. Friedlander, and S. Zhao. 1997. A new Liu, H., and A. T. Beckenbach. 1992. Evolution of the Mi- nuclear gene for insects phylogenetics: Dopa Decarbox- tochondrial Cytochrome Oxidase II Gene among 10 Or- ylase is informative of relationships within Heliothinae ders of Insects. Mol. Phylogenet. Evol. 1: 41Ð52. (Lepidoptera: Noctuidae). Syst. Biol. 46: 269Ð283. Lyra, M. L., L. B. Klaczko, and A.M.L. Azeredo-Espin. 2009. Fitt, G. P. 1989. The ecology of Heliothis species in relation Complex pattern of genetic distribution in populations of to agroecosystems. Annu. Rev. Entomol. 34: 17Ð52. the New World Screwworm ßy revealed by mitochon- Guindon, S., and O. Gascuel. 2003. A simple, fast and accu- drial DNA markers. Med. Vet. Entomol. 23: 32Ð42. rate algorithm to estimate large phylogenies by maximum Mallet, J., A. Korman, D. Heckel, and P. King. 1993. Bio- likelihood. Syst. Biol. 52: 696Ð704. chemical genetics of Heliothis and Helicoverpa (Lepidop- June 2014 MASTRANGELO ET AL.: DETECTION OF H. armigera IN RORAIMA 11

tera: Noctuidae) and evidence for a founder event in Srivastava, C. P., R. Ahmad, R. Ujagir, and S. B. Das. 2005. Helicoverpa zea. Ann. Entomol. Soc. Am. 86: 189Ð197. Helicoverpa armigera management in pulses-present sce- Nei, M. 1987. Molecular evolutionary genetics. Columbia nario and future strategies, pp: 265Ð286. In Recent Ad- University Press, New York, NY. vances in Helicoverpa armigera Management. Indian So- Nibouche, S., R. Bues, J. F. Toubon, and S. Poitout. 1998. ciety of Pulses Research and Development, Kanpur, Allozymepolymorphisms in the cotton bollworm Helicov- India. erpa armigera (Lepidoptera: Noctuidae): comparison of Tamura, K., D. Peterson, N. Peterson, G. Stecher, M. Nei, and African and European populations. Heredity 80: 438Ð445. S. Kumar. 2011. MEGA5: molecular evolutionary genet- Nylander, J.A.A. 2004. MrAIC.pl. Program distributed by ics analysis using maximum likelihood, evolutionary dis- the author. Evolutionary Biology Centre, Uppsala Uni- tance, and maximum parsimony methods. Mol. Biol. Evol. versity. (https://github.com/nylander/MrAIC). 28: 2731Ð2739. Passoa, S. 2004. IdentiÞcation guide to larval Heliothinae Tay, W. T., M. F. Soria, T. Walsh, D. Thomazoni, P. Silvie, (Lepidoptera: Noctuidae) of quarantine signiÞcance. G. T. Behere, C. Anderson, and S. Downes. 2013. A U.S. Dep. Agric. Report Columbus, OHIO 43212. brave new world for an old world pest: Helicoverpa ar- Pedgley, D. E. 1985. Windborne migration of Heliothis ar- migera (Lepidoptera: Noctuidae) in Brazil. PLoS ONE 8: migera (Hu¨ bner) (Lepidoptera: Noctuidae) to the Brit- e80134 2013 ish Isles. Entomol. Gazette. 36: 15Ð20. Venette, R. C., E. E. Davis, Z. Zaspel, H. Heisler, and M. Pogue, M. G. 2004. A new synonym of Helicoverpa zea (Bod- Larson. 2003. Mini risk assessment, Old World boll- die) and differentiation of adult males of H. zea and H. worm, Helicoverpa armigera (Hubner) (Lepidoptera: armigera (Hu¨ bner) (Lepidoptera: Noctuidae: Heliothi- Noctuidae). Cooperative Agricultural Pest Survey, Ani- nae). Ann. Entomol. Soc. Am. 97: 1222Ð1226. mal and Plant Health Inspection Service, US Department Reynolds, J., B. S. Weir, and C. C. Cockerham. 1983. Esti- _ mation for the coancestry coefÞcient: basis for a short- of Agriculture. (http://www.aphis.usda.gov/plant health/ _ _ _ term genetic distance. Genetics 105: 767Ð779. plant pest info/pest detection/downloads/pra/ Ronquist, F., and J. P. Huelsenbeck. 2003. MrBayes 3: harmigerapra.pdf). Bayesian phylogenetic inference under mixed models. Wahlberg, N., and C. W. Wheat. 2008. Genomic outposts Bioinformatics 19: 1572Ð1574. serve the phylogenomic pioneers: Designing novel nu- Rubinoff, D., S. Cameron, and K. Will. 2006. A genomic clear markers for genomic DNA extractions of lepidop- perspective on the shortcomings of mitochondrial DNA tera. Syst. Biol. 57: 231Ð242. for “barcoding” identiÞcation. J. Heredity 97: 581Ð594. Wang, N. C., and Z. H. Li. 1984. Studies on the biology of Saitou, N., and M. Nei. 1987. The neighbor-joining method: cotton bollworm (Heliothis armigera Hu¨ bner) and to- a new method for reconstructing phylogenetic trees. Mol. bacco budworm (Heliothis assulta Quenee). J. Shandong Biol. Evol. 4: 406Ð425. Agric. Univ. 1: 13Ð25. Sheng, C. F. 1993. Outbreak of Heliothis armigera in North Waugh, J. 2007. DNA barcoding in species: progress, China: possible causes and control strategies, pp. 841Ð844. potential and pitfalls. Bioessays 29: 188Ð197. In Proceedings, Beltwide Cotton Production Research Whinnett, A., M. Zimmermann, K. R. Willmott, N. Herrera, Conferences. National Cotton Council of America, Mem- R. Mallarino, F. Simpson, M. Joron, G. Lamas, and J. phis, TN. Mallet. 2005. Striking variable divergence times inferred Silva-Branda˜o, K. L., M. L. Lyra, and A.V.L. Freitas. 2009. across an Amazonian butterßy Ôsature zoneÕ. Proc. R. Soc. Barcoding Lepidoptera: current situation and perspec- B 272: 2525Ð2533. tives on the usefulness of a contentious technique. Widmer, M. W., and P. Schofield. 1983. Heliothis Dispersal Neotrop. Entomol. 38: 441Ð451. and Migration. Tropical Development and Research In- Slatkin, M. 1995. A measure of population subdivision based stitute, London, United Kingdom. on microsatellite allele frequencies. Genetics 139: 457Ð Zhou, X., O. Faktor, S. W. Applebaum, and M. Coll. 2000. 462. Population structure of the pestiferous moth Helicoverpa Sievers, F., A. Wilm, D. G. Dineen, T. J. Gibson, K. Karplus, armigera in the eastern Mediterranean using RAPD anal- W. Li, R. Lopez, H. McWilliam, M. Remmert, J. Soding, ysis. Heredity 3: 251Ð256. et al. 2011. Fast, scalable generation of high-quality pro- tein multiple sequence alignments using Clustal Omega. Mol. Syst. Biol. 7: 1Ð6. Received 17 September 2013; accepted 10 March 2014.