An Evaluation of the Status of invadens and the Systematics of the (Diptera: ) Complex Author(s): Michael San Jose , Luc Leblanc , Scott M. Geib , and Daniel Rubinoff Source: Annals of the Entomological Society of America, 106(6):684-694. 2013. Published By: Entomological Society of America URL: http://www.bioone.org/doi/full/10.1603/AN13017

BioOne (www.bioone.org) is a nonprofit, online aggregation of core research in the biological, ecological, and environmental sciences. BioOne provides a sustainable online platform for over 170 journals and books published by nonprofit societies, associations, museums, institutions, and presses. Your use of this PDF, the BioOne Web site, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/page/terms_of_use. Usage of BioOne content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research. SYSTEMATICS An Evaluation of the Species Status of and the Systematics of the Bactrocera dorsalis (Diptera: Tephritidae) Complex

1 1 2 1,3 MICHAEL SAN JOSE, LUC LEBLANC, SCOTT M. GEIB, AND DANIEL RUBINOFF

Ann. Entomol. Soc. Am. 106(6): 684Ð694 (2013); DOI: http://dx.doi.org/10.1603/AN13017 ABSTRACT The genus Bactrocera (Tephritidae) contains Ͼ500 species, including many severe pests of fruits and vegetables. Although native to tropical and subtropical areas of Africa, India, Southeast Asia, and Australasia, a number of the pest species, largely members of the Bactrocera dorsalis (Hendel) complex, have become widespread through accidental introduction associated with agri- cultural trade. The B. dorsalis complex includes several morphologically and ecologically similar pests, making species designations uncertain. One of these, Bactrocera invadens Drew, Tsuruta, and White, endemic to Sri Lanka, has spread across Africa in the last decade and become a major agricultural pest. We sequenced one mitochondrial and two nuclear genes from 73 specimens, belonging to 19 species to construct phylogenies and examine species relationships and limits within the genus Bactrocera and several species of the B. dorsalis complexÑspeciÞcally addressing the placement of B. invadens. Results indicate the B. dorsalis complex is polyphyletic. B. invadens and several other species within the B. dorsalis complex (B. dorsalis, Bactrocera papayae Drew & Hancock, and Bactrocera philippinensis Drew & Hancock) are also paraphyletic with respect to each other and probably represent a single genetically indistinguishable, phenotypically plastic, pest species that has spread throughout the world.

KEY WORDS Bactrocera dorsalis, Diptera, Tephritidae, B. invadens, invasive species

The genus Bactrocera (Tephritidae) is the largest in but the most invasive and economically important the tribe Dacini, with Ͼ500 described species (Drew species belong to the Bactrocera dorsalis complex 2004, Clarke et al. 2005). It is an Old World lineage, (Clarke et al. 2005). Hardy (1969) identiÞed 11 spe- with species native to Africa, tropical and subtropical cies closely related to B. dorsalis (Hendel) and Asia, Australasia, and the western PaciÞc Islands, with grouped them in the B. dorsalis complex. Drew and the highest species diversity within Southeast Asia and Hancock (1994) published the Þrst comprehensive Australasia (Drew 2004). Bactrocera includes at least revision of the complex, describing 40 new species, 68 economically important species whose larvae infest resulting in a total of 52 species in Asia. Currently, 75 a large variety of fruit and cucurbit crops (http:// described species are considered to be part of the B. www.herbarium..edu/fruitßy/), causing direct dorsalis complex, and many more await description damage even to unripe fruit. Many pest species have (Clarke et al. 2005). Four sibling species [B. dorsalis become widespread through accidental introductions (Hendel), Bactrocera papayae Drew & Hancock, Bac- associated with agricultural trade, and the genus now trocera philippinensis Drew & Hancock, and Bactro- occurs on every continent except Antarctica (Drew cera carambolae Drew & Hancock], herein collec- 2004). Bactrocera ßies also inßuence international tively referred to as Bactrocera dorsalis sensu lato trade; regions currently infested with Bactrocera pose (s.l.), are also among the most damaging and polypha- a signiÞcant risk to areas that are not infested. As a gous pests of the complex (Clarke et al. 2005). These result, noninfested countries severely limit or prohibit are almost identical morphologically, with minor color imports from infested areas to prevent further spread pattern differences (Drew and Hancock 1994) and of pest species. few measurable characters, the length of the female Some of the most widespread and damaging species aculeus and male aedeagus (Iwahashi 1999a,b), and of Bactrocera include Bactrocera () cucur- wing shape morphometrics (Schutze et al. 2012), to bitae (Coquillett), Bactrocera () oleae (Gme- separate them with limited reliability. Males of most of lin), and Bactrocera (Bactrocera) tryoni (Froggatt), the B. dorsalis complex species are attracted to either of two kairomone lures, methyl eugenol (ME) or cue- 1 Department of Plant and Environmental Protection Sciences, lure (CL), although B. dorsalis s.l. species are only University of Hawaii at Manoa, 3050 Maile Way, Gilmore 310, Ho- attracted to ME, and a few species do not respond to nolulu, HI 96822. male lures. This male response has been used both as 2 U.S. Department of Agriculture PaciÞc Basin Agricultural Re- search Center, 64 Nowelo St., Hilo, HI 96720. a diagnostic tool to separate morphologically similar 3 Corresponding author, e-mail: [email protected]. species when they are attracted to different lures and

0013-8746/13/0684Ð0694$04.00/0 ᭧ 2013 Entomological Society of America November 2013 SAN JOSE ET AL.: STATUS OF B. invadens AND SYSTEMATICS OF B. dorsalis 685 as a way to monitor and control ßies in agricultural of the B. dorsalis complex, using two mitochondrial settings. and four nuclear loci. They found that several de- Understanding the evolutionary relationships of scribed pest species are paraphyletic (B. dorsalis, B. pest groups is important for many aspects of quaran- papayae, and B. philippinensis) and that B. caram- tine and control. Assessing species limits among mor- bolae was genetically distinct from the other B. phologically similar species is essential because of spe- dorsalis s.l. species. cies-centric trade policies, which inform agricultural Although Drew et al. (2008) argued that B. dorsalis quarantine decisions. Species that are morphologically s.l. comprises multiple species, recent studies by Kro- similar may have different ecological tolerances or sch et al. (2013) and Boykin et al. (2013)found no host preferences, which dictate the amount of damage evidence for distinctly different lineages separating B. they cause. dorsalis s.str., B. philippinensis, and B. papayae when B. dorsalis s.l. has invaded many places around the using microsatellites, wing shape, aedeagus length, world, including the PaciÞc (B. philippinensis in Palau; and multiple nuclear genes. A broader phylogenetic B. dorsalis in Hawaii, French Polynesia, the Mariana analysis of B. dorsalis s.l. based on multiple genes and islands [eradicated], and Nauru [eradicated]; and B. including several nonpest species should provide philippinensis and B. papayae in Australia [both erad- more insight on species limits and . All pre- icated]) and South America (B. carambolae in Suri- vious studies with the exception of Krosch et al. (2012) name, Guyana, French Guiana, and northern Brazil). and Boykin et al. (2013) were based on relatively small They are regularly intercepted by quarantine agencies sample sizes and a limited number of mitochondrial in California and Florida, and risk of widespread es- genes. The use of only mitochondrial genes for phy- tablishment is of great concern (California Depart- logenetic reconstruction of species relationships has ment of Food and Agriculture [CDFA] 2012, Florida serious shortcomings (Funk and Omland 2003, Ballard Department of Agriculture and Consumer Services and Whitlock 2004, Cognato 2006, Rubinoff 2006). In [FDACS] 2012). In 2003, a new invasive Bactrocera our study, we used one mitochondrial gene and two species was detected attacking fruits in Kenya (Lux et nuclear genes to elucidate species relationships in the al. 2003, Drew et al. 2005), and it has rapidly spread genus Bactrocera under a phylogenetics framework. throughout tropical and subtropical Africa and is now We explored the relationship between B. invadens and encroaching on northern South Africa (De Meyer et species of the B. dorsalis complex, speciÞcally mem- al. 2010, International Plant Protection Convention bers of the B. dorsalis s.l. clade. By including several [IPPC] 2012). Described in 2005 as B. invadens Drew, noneconomic species of the B. dorsalis complex along Tsuruta, & White, this species was thought to be en- with species of economic importance, we further pro- demic to Sri Lanka, and while morphologically similar vide insight on whether the B. dorsalis complex is a to B. dorsalis, it has variable reddish coloration on the monophyletic evolutionary unit or merely a paraphyl- mesonotum, markedly different from other species etic assemblage of morphologically similar ßies. of the B. dorsalis complex that have a predominantly black mesoscutum (Drew and Hancock 1994). Drew et al. (2008) included B. invadens in an analysis of four Materials and Methods B. dorsalis s.lat. species and concluded it was signiÞ- cantly different from B. carambolae, Bactrocera dor- Taxon Sampling. A list of taxa and specimens used salis s.str., B. papayae, and B. philippinensis. However, in this study, including collecting locality, species B. invadens has been regarded as part of the B. names, and voucher codes, is given in Table 1. Spec- dorsalis complex in literature (Khamis et al. 2009, imens from Hawaii and Asia were collected either De Meyer et al. 2010, Liu et al. 2011) and shown to using male-lure-baited traps or the general tephritid be similar to B. dorsalis s.str. using morphometrics protein lure torula yeast by M.S.J. and D.R. (Asia) and and cytochrome oxidase I DNA sequences (Khamis L.L. and M.S.J. (Hawaii) between 2010 and 2011. et al. 2012). Specimens were also provided by colleagues as adult Despite their importance in agriculture and trade, ßies preserved in ethanol. Fifteen described and four relatively few studies have been carried out on Bac- undescribed species of Bactrocera in two subgenera trocera molecular systematics, and most of them were were included in this study (Table 1). We used Bac- limited to surveying mitochondrial genes. Smith et al. trocera (Zeugodacus) cucurbitae (Coquillett) as an (2003) used two mitochondrial genes (COII and 16s), outgroup. Specimens were identiÞed to species level and concluded that the B. dorsalis complex was mono- by L.L., using currently available keys (Drew 1989, phyletic. Using the genes COI and II, Nakahara and White and ElsonÐHarris 1992, Drew and Hancock Muraji (2008) found that the B. dorsalis complex was 1994, Drew et al. 2005), and determinations were ver- rendered paraphyletic, as Bactrocera musae (Tryon) iÞed by R.A.I. Drew. Four species are referred to by falls within the B. dorsalis complex species lineage in numbers (22, 25, 26, and 45), because they are new their analysis. In a more recent comprehensive phy- species with descriptions soon be published (Drew logenetic survey of Dacini using one nuclear and and Romig 2013). three mitochondrial genes, Krosch et al. (2012) DNA Extraction, Amplification, and Sequencing. found that the B. dorsalis complex was monophy- For each specimen, three legs were used for total letic. Most recently, Boykin et al. (2013) conducted genomic DNA extraction. The remainder of the spec- the most comprehensive survey of the pest species imen was deposited as a voucher in the University of 686 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 6

Table 1. Species, lure response, collection locality, and identification numbers for species in the genera Bactrocera, , and used in this study

No. Species (subgenera) Male lure Locality and sample numbers specimens Bactrocera (Bactrocera) dorsalis complex species B. (Bactrocera) carambolae Drew & Methyl eugenol 8 Malaysia: Kedah (ms1441, 1442); Penang (ms1439, 1440); Hancock Perak (ms1443Ð1446) B. (Bactrocera) dorsalis (Hendel) Methyl eugenol 12 Cambodia: Koh Kong (ms1116). China: Hong Kong (ms0932). French Polynesia: Tahiti (ms0889, 0890). Hawaii: Molokai (ms1128, 1130); Oahu (ms0853). IAEA lab colony (Vienna): ms1183, 1184, 1289). Thailand: Chiang Mai (ms1119) B. (Bactrocera) dorsalis (Hendel) Methyl eugenol 1 Thailand: Bangkok (ms1122) (with red thorax) B. (Bactrocera) dorsalis complex with Methyl eugenol 3 Cambodia: Koh Kong (ms1113). Thailand: Chiang Mai costal band not expanded (ms1120, 1121) B. (Bactrocera) fuscitibia Drew & Cue-lure 2 Cambodia: Koh Kong (ms1176, 1177) Hancock B. (Bactrocera) papayae Drew & Methyl eugenol 10 IAEA Lab colony (Vienna) (ms1186), Malaysia: Kedah Hancock (ms1432, 1433); Penang (ms1428Ð1430); Perak (ms1434, 1435, 1437, 1438) B. (Bactrocera) philippinensis Drew Methyl eugenol 2 IAEA Lab colony (Vienna) (ms1187, 1292) & Hancock B. (Bactrocera) species 26 Methyl eugenol 2 Laos: Luang Nam Tha (ms1110, 1294) (undescribed) Bactrocera (Bactrocera) species other than dorsalis complex B. (Bactrocera) correcta (Bezzi) Methyl eugenol 2 Cambodia: Koh Kong (ms1099). Thailand: Chiang Mai (ms1103) B. (Bactrocera) invadens Drew, Methyl eugenol 12 Burkina Faso (ms0944, 0947, 0949, 0952). Kenya (from Tsuruta, & White lab colony) (ms1188, 1293). Se´ne´gal (ms0898, 0899, 0900, 0904, 0905, 0909) B. (Bactrocera) kirki (Froggatt) Cue-lure 2 French Polynesia: Tahiti (ms0894, 0895) B. (Bactrocera) latifrons (Hendel) Latilure 2 Hawaii: Hawaii Island (ms0882); Oahu (ms0963) B. (Bactrocera) nigrotibialis (Perkins) Cue-lure 2 Cambodia: Koh Kong (ms1034, 1035) B. (Bactrocera) species 22 Cue-lure 2 Laos: Luang Nam Tha (ms1160, 1161) (undescribed)a B. (Bactrocera) species 25 Cue-lure 2 Cambodia: Koh Kong (ms1171). Laos: Luang Nam Tha (undescribed)a (ms1166) B. (Bactrocera) species 45 Cue-lure 2 Cambodia: Koh Kong (ms1170). Laos: Luang Nam Tha (undescribed)a (ms1167) B. (Bactrocera) tryoni (Froggatt) Cue-lure 2 French Polynesia: Tahiti (ms0892, 0893) B. (Bactrocera) tuberculata (Bezzi) Methyl eugenol 2 Thailand: Chiang Mai (ms1084, 1090) B. (Bactrocera) umbrosa (F.) Methyl eugenol 2 Cambodia: Koh Kong (ms1002, 1003) B. (Zeugodacus) cucurbitae Cue-lure 2 Cambodia: Koh Kong (ms0987). Hawaii: Molokai (Coquillett) (ms1126)

GenBank accession numbers available as Supp. Table 1 ͓online only͔. a Species 22, 25, and 45 are consistent with the B. dorsalis complex deÞnition of Drew and Hancock (1994), but have red markings on mesoscutum.

Hawaii Museum (UHIM) for morphological for 30 s, and 70ЊC for 1 min) with a Þnal 70ЊC extension studies (Table 1). Genomic DNA was extracted using for 10 min. EF-1␣ was ampliÞed as a single fragment the DNeasy blood and tissue extraction kit with primers M46Ð1 and M4rc under the same thermal following manufactureÕs protocol (Qiagen, Inc., Va- proÞle as COI. Period was ampliÞed as one fragment lencia, CA). with primers Per2612f and Per3105r, ampliÞed under Three different gene regions were ampliÞed: the the following thermal conditions: 2 min at 94ЊC, 40 mitochondrial gene cytochrome c oxidase I (COI, 780 cycles of (94ЊC for 30 s, 55ЊC for 30 s, and 70ЊC for 1 bp) and the nuclear genes, elongation factor-1␣ (EF- min) with a Þnal 70ЊC extension for 10 min. Primer 1␣, 759 bp) and period (PER, 450 bp). These three sequences are listed in Table 2. All polymerase chain genes were selected, as each has been demonstrated reaction (PCR) products were visualized on 1% aga- to be informative in distinguishing populations, spe- rose gel and puriÞed using QIAquick spin columns cies complexes, species, or genera in Diptera (Folmer (Qiagen, Inc.) according to the manufacturerÕs pro- et al. 1994, Simon et al. 1994, Cho et al. 1995, Bauzer tocol. Bidirectional DNA sequencing was performed et al. 2002, Moulton and Wiegmann 2004, Barr et al. at the Advanced Studies of Genomics, Proteomics and 2005, Foley et al. 2007, Virgilio et al. 2009, Gibson et al. Bioinformatics (ASGPB) sequencing facility of the 2011). COI was ampliÞed using the primers HCO- University of Hawaii at Manoa (http://asgpb.mhpcc. 2198rc and Pat-k508 under the following thermal con- hawaii.edu/). Sequences were deposited into GenBank ditions: 2 min at 94ЊC, 40 cycles of (94ЊC for 30 s, 53ЊC KF184032-KF184250 (Supp. Table 1 [online only]). November 2013 SAN JOSE ET AL.: STATUS OF B. invadens AND SYSTEMATICS OF B. dorsalis 687

Table 2. Primers names and sequences used in this study

Primer name Sequence 5Ј-3Ј Reference HCO-2198rc GCTCAACAAATCATAAAGATATTGG Folmer et al. 1994 Pat-k508 TCCAATGCACTAATCTGCCATATTA Simon et al. 1994 Per2612F ATTCATGGGAAGGAGATGCC Barr et al. 2005 Per3105R AABGACATGGGTTGGTACATC Barr et al. 2005 M46-1 CAGGAAACGCTATGACCGAGGAAATYAARAAGGAAG Cho et al. 1995 M4rc TGTAAAACGACGGCCAGTACAGCVACKGTYTGYCTCATRTC Cho et al. 1995

Sequence Alignment, Nucleotide Composition, and with the best likelihood score. To assess branch sup- Phylogenetic Analysis. Sequence alignments were port, one thousand Maximum Likelihood bootstrap performed with the software package Geneious replicates were conducted in GARLI; Maximum Like- (Drummond et al. 2011). Heterozygosity in the nu- lihood bootstrap trees were summarized in Sumtrees clear genes was present in most samples. Ambiguity ver. 3.1.0 (Sukumaran and Holder 2010) with a min- codes (i.e., notation according to International Union imum clade frequency of 50%, and branch support was of Pure and Applied Chemistry (IUPAC)) were used mapped onto the best scoring ML tree. For the con- to denote heterozygous base pairs, and these codes catenated dataset, we partitioned the data by gene and were used in the subsequent analysis. Sample ran MrBayes using the same settings as the individual heterozygosity was measured by calculating percent- gene analyses except the parameters statefreq, revmat, ages of sequences that contained one or more shape, and pinvar were unlinked between partitions. heterozygous sites and the number of heterozygous For the Maximum Likelihood analysis of the parti- sites within each gene. Sequence alignment for each tioned concatenated dataset, we ran GARLI using the gene was conducted in Geneious using the “Muscle” same settings and analyses for each partition as when option with default settings (Edgar 2004). Genetic genes were analyzed individually. Trees were visual- variation between genes and taxa, nucleotide compo- ized using FigTree v1.4.0 (Rambaut 2012) and rooted sition, pairwise distances, and average base frequen- with B. cucurbitae. cies for each gene, and level of signiÞcance were estimated by ␹2 tests and were calculated in PAUP* (Swofford 2003). All three gene regions were tested Results separately for an appropriate nucleotide substitution model with jModelTest v2.1.3 (Darriba et al. 2012) Nucleotide Composition and Sequence Diver- under the Akaike information criterion with correc- gence. Average base frequencies for each gene and ␹2 tion (AICc) (Darriba et al. 2012). The model HKYϩ level of signiÞcance were estimated by tests, and ϭ ϭ ϭ ⌫ was selected, and applied to all genes. Phylogenetic results were as follows: COI (A 0.31, C 0.18, G ϭ ␣ ϭ ϭ ϭ analyses were performed with both Maximum Like- 0.15, T 0.35), EF-1 (A 0.24, C 0.26, G 0.24, ϭ ϭ ϭ ϭ lihood and Bayesian Inference. MrBayes 3.2.1 (Ron- T 0.26), and period (A 0.25, C 0.32, G 0.22, ϭ quist et al. 2012) was used for Bayesian analyses and T 0.19). There was no signiÞcant difference in base ␹2 ϭ ϭ GARLI 2.0 (Zwickl 2006) was used for maximum like- frequencies for any of the genes (COI: 100.38, P ϭ ␣ ␹2 ϭ ϭ ϭ lihood (ML). We Þrst analyzed each gene separately 1.00, df 216; EF-1 : 9.22, P 1.00, df 216; ␹2 ϭ ϭ ϭ and subsequently concatenated them into a single period: 24.32, P 1.00, df 216). This demon- dataset partitioned, by gene, using Maximum Likeli- strates that there is no compositional bias in nucleo- hood and Bayesian inference. For each individual tides frequencies between the genes that might affect gene analysis (COI, period, and EF-1␣) we ran four phylogenetic resolution. Each gene contained various independent Bayesian runs in MrBayes 3.2.1 each with levels of genetic diversity (Table 3). Percentage of one hot chain and three cold chains. Each run started sequences that contained heterozygous base-pairs for ␣ from a random tree using default priors sampling every each nuclear gene were as follows: EF-1 (45.5%) and one thousand generation for 10 million generations period (81.0%). Percentage of heterozygous sites for ␣ with a relative burn-in of 25%. We used the program each nuclear gene were as follows: EF-1 (6.1%) and Tracer 1.5 (Rambaut and Drummond 2009) to assess period (20.7%). COI pair-wise difference between B. convergence of Bayesian analyses. For GARLI anal- dorsalis s.lat. species (B. carambolae, B. dorsalis, B. Յ ysis, we conducted 10 ML tree searches with default papayae, and B. philippinensis) was low ( 1.8%). settings using a random starting tree to Þnd the tree Phylogenetic Analysis. Although topological differ- ences exist among the three gene trees, differences were not supported by high bootstrap values. Conse- Table 3. Genetic variability between genes quently, we combined the genes into a single concat- enated dataset to compare with the individual gene Parsimony Variable Total Gene trees. A total evidence approach using the combined informative sites sites length (bp) data produced high branch support across much of the Period 133 (29.56%) 156 (34.67%) 450 phylogeny (Figs. 1Ð4). The position of B. dorsalis COI 277 (35.51%) 291 (37.31%) 780 ␣ complex species across both the concatenated and EF-1 131 (17.26%) 153 (20.16%) 759 individual gene trees suggests that the complex as a 688 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 6

Fig. 1. Maximum likelihood tree, concatenated, three gene data set (COI, period, and EF-1␣). Support values above branches are Maximum Likelihood Bootstrap values/Bayesian Posterior Probabilities. Scale bar indicates the number of substitutions per site. Black circle indicates B. dorsalis complex. ME and CL indicate attractants methyl eugenol and Cuelure, respectively. whole is polyphyletic. Two species, Þtting in the B. all our phylogenies contains four species previously dorsalis complex based on gross morphology, are sep- understood to be part of the complex (B. carambolae, arated in all trees by several non-B. dorsalis groups B. dorsalis, B. papayae, and B. philippinensis), but also from members of B. dorsalis s.l., which otherwise form B. invadens, which is morphologically distinct from a well-supported monophyletic clade (Fig. 1; 96 ML other dorsalis complex species by its light-colored BS/1.0 Bayesian PP). They are Bactrocera fuscitibia mesoscutum (Drew et al. 2005). B. invadens samples Drew and Hancock (attracted to Cuelure) and Bac- collected from the same localities are polyphyletic in trocera species 26, an undescribed species attracted to the analysis, being placed in four different locations methyl eugenol. However, internal resolution of the across the B. dorsalis s.l. clade. But the COI pair-wise species relationships within the B. dorsalis s.l. clade is genetic divergence among sequences within the B. low, with Ͻ50% posterior probabilities and bootstrap dorsalis s.l. clade is less than or equal to 1.8%, suggesting values for most branches. The B. dorsalis s.l. clade in unresolved and minor divergence between the taxa. November 2013 SAN JOSE ET AL.: STATUS OF B. invadens AND SYSTEMATICS OF B. dorsalis 689

Fig. 2. Maximum likelihood tree, COI dataset. Support values above branches are Maximum Likelihood Bootstrap values/Bayesian Posterior Probabilities. Scale bar indicates the number of substitutions per site. Black circle indicates B. dorsalis complex. ME and CL indicate attractants methyl eugenol and Cuelure, respectively.

B. carambolae, another major pest species in the B. of the world, but also trade restrictions between coun- dorsalis complex, is also paraphyletic within the B. tries are based on concepts of species that our results dorsalis s.l. clade in two out of the three individual suggest may not be valid. B. dorsalis s.str. is paraphyl- gene trees (COI [Fig. 2], period [Fig. 4]). However, etic with respect to B. carambolae, B. papayae, and B. in the EF-1␣ tree six B. carambolae specimens are philippinensis, indicating either that gene regions cho- placed basally in the subgenus Bactrocera clade and sen are not appropriate to detect genetic divergence two B. carambolae specimens are placed within the B. subsequent to speciation (␭ too low) or this assem- dorsalis s.l. clade. blage of ßies is a single species. Previous studies using mitochondrial DNA (Armstrong and Ball 2005), mic- rosatellites (Krosch et al. 2013), and morphological Discussion characters (Krosch et al. 2013), as well as mating The B. dorsalis Complex. The B. dorsalis complex is studies (Schutze et al. 2013), are in conßict with the polyphyletic in our analysis across multiple genes taxonomy proposed by Drew (Drew and Hancock (Figs. 2Ð4). The importance of robust species con- 1994, Drew et al. 2008). Our results using both mito- cepts and well-deÞned species for the B. dorsalis com- chondrial and nuclear genes also disagree with the plex cannot be overemphasized. Not only is B. dorsalis current morphology based taxonomy. The gene re- s.l. a major pest across tropical and subtropical areas gions chosen for this study were previously shown to 690 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 6

Fig. 3. Maximum likelihood tree, EF-1␣ dataset. Support values above branches are Maximum Likelihood Bootstrap values/Bayesian Posterior Probabilities. Scale bar indicates the number of substitutions per site. Black circle indicates B. dorsalis complex. ME and CL indicate attractants methyl eugenol and Cuelure, respectively. be able to differentiate species in a wide range of levels (e.g., COI pair-wise distance of the species ([COI, Simon et al. 1994]; [EF-1␣, Cho et al. within this clade is Ͻ1.8%). 1995]; [period, Barr et al. 2005]; Caterino et al. 2000). We included specimens of B. invadens from three As a result, current concepts of these important spe- invasive populations across Africa in this analysis. Al- cies, and consequently quarantine regulations, may though these samples have the typical morphology of need reevaluation. In the concatenated data set, the B. invadens and are being agriculturally regulated in above mentioned four species collectively form a well- these countries as B. invadens, the genetic analysis of supported clade (Fig. 1; 96 ML BS/ 1.0 PP) in which these samples demonstrates that these samples con- all taxa are attracted to methyl eugenol, but indepen- sistently appear as a component of the B. dorsalis s.l. dent species status for these taxa is not supported, clade across all analyses and genes. This indicates that based on a lack of monophyly and very low divergence either B. invadens is not only part of the B. dorsalis November 2013 SAN JOSE ET AL.: STATUS OF B. invadens AND SYSTEMATICS OF B. dorsalis 691

Fig. 4. Maximum likelihood tree, period dataset. Support values above branches are Maximum Likelihood Bootstrap values/Bayesian Posterior Probabilities. Scale bar indicates the number of substitutions per site. Black circle indicates B. dorsalis complex. ME and CL indicate attractants methyl eugenol and Cuelure, respectively. complex but also is genetically indistinguishable from tion the reliability of the morphological characters many of the pest species of this complex, or that the that have been previously considered important in recent introductions in Africa are being mis-identiÞed deÞning Bactrocera species and their complexes, with as B. invadens, demonstrating a failure in the morpho- obvious implications for quarantine and agriculture. In logical characters which deÞne these species. Either fact, in our phylogenetic analysis B. invadens is ren- option does not support the existence of B. invadens as dered polyphyletic with respect to other species in the an independent species outside of the clade based on B. dorsalis s.l. clade. This suggests that the criteria used its current morphological description. Support for our to deÞne species within the B. dorsalis complex need results also comes from a recent comparison of phe- to be reconsidered. Further, from an applied perspec- nylpropanoid volatiles in the adult male rectal gland tive, the genetic diversity of B. invadens samples from between four Bactrocera species, which found that B. Africa suggests either multiple introductions of B. in- invadens and B. dorsalis sequestered the same com- vadens to Africa or a genetically diverse founding bination of volatiles, while the other two species in the population. This indicates that quarantine efforts for study (Bactrocera zonata (Saunders) and Bactrocera Africa need signiÞcant attention to prevent future correcta (Bezzi)) sequestered different volatiles (Tan invasive species introductions. et al. 2011). The results from our study, as well as those The relationship between B. carambolae and the from Krosch et al. (2013) and Tan et al. (2011), ques- other B. dorsalis s.l. species is not as clear. In two of the 692 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 6 three gene trees (period and COI), B. carambolae is putative differences in host use. Reassessing the tax- paraphyletic within the B. dorsalis s.l. However, in the onomy of Bactrocera in general will be important, not EF-1␣ tree, two specimens are grouped within the B. only for understanding patterns of diversiÞcation in a dorsalis s.l. clade, but the other six are basal and sep- genus with hundreds of species across the Paleotro- arate within the subgenus Bactrocera. Morphologi- pics, but also for its implications in reassessing trade cally, B. carambolae is fairly distinct from the other B. barriers between countries, as they are often contin- dorsalis s.l. species (B. papayae, B. dorsalis, B. philip- gent on presence or absence of pest species. pinensis, and now B. invadens), and it is also sympatric with B. papayae and B. dorsalis. Owing to the para- phyly between the individual genes and the discor- Acknowledgments dance in the placement of B. carambolae between genes, it is possible that hybridization and introgres- We are indebted to Rudolph Putoa, Kemo Badji, and Sylvain Ouedraogo for providing specimens from French sion are occurring between B. carambolae and sym- Polynesia and Africa. We also thank Will Haines, Jesse Eiben, patric members of the B. dorsalis s.l. clade. This gene Tony Wong, Emmett Easton, and David Haymer for provid- ßow between morphologically distinct species com- ing additional material from Asia and Hawaii. Saba Young and plicates the diagnostic utility of morphological char- Dan Nitta helped with molecular work. We greatly appre- acters. Recent mating studies conducted by Schutze et ciate the help of Chi-Yeh Chien (Thai Royal Project Foun- al. (2013) indicated that B. carambolae will mate with dation), Thongsavanh Taipangnavong (Food and Agricul- the other B. dorsalis complex ßies tested in the study ture Organization (FAO)-integrated pest management (B. dorsalis, B. philippinensis, and B. papayae); how- (IPM) Laos), Vornthalom Chanthavong (FAO-IPM Laos), ever, they were relatively incompatible, as B. caram- Lira Chea (FAO-IPM Cambodia), Ajay Markanday (FAO- IPM Cambodia), Johannes Ketellar (FAO-Regional OfÞce bolae males had a lower mating success rate with the for Asia and the PaciÞc), Prabhat Kumar (Asian Institute of other species in the study than with B. carambolae Techonology) and Po-Yung Lai (University of Hawaii Manoa females. College of Tropical Agriculture and Human Resources), with Ultimately, our data, based on multiple genes, sug- permits and logistics in our international collecting trips. We gest that the taxa within the B. dorsalis s.l. clade (B. especially thank Norman Barr (U.S. Department of Agricul- dorsalis, B. papayae, B. philippinensis, and now B. in- tureÐAnimal and Plant Health Inspection Service [USDAÐ vadens) all represent a monophyletic and apparently APHIS]) and two anonymous reviewers for their helpful monotypic evolutionary lineage, despite some mor- comments. This work was made possible, in part, by a Co- phological variation in different parts of its geographic operative Agreement from the United States Department of Agriculture through Farm Bill funding (project 3.0251) ad- range with coloration of mesonotum varying from ministered by the College of Tropical Agriculture and Hu- black to reddish, not only in populations of B. invadens man Resources. It may not necessarily express the views of in Africa, but also to some extent in populations of the USDA. USDA is an equal opportunity provider and em- other B. dorsalis complex species in Bangladesh, ployer. Southeast Asia and Hawaii (L.L., unpublished data). Based on these results, we propose that these four major pest species represent a single, widespread, References Cited phenotypically plastic species from an evolutionary perspective. Our results corroborate recent studies Armstrong, K. F., and S. L. Ball. 2005. DNA barcodes for biosecurity: invasive species identiÞcation. Phil. Trans. R. from Schutze et al. (2012) and Krosch et al. (2013), in Soc. B 360: 1813Ð1823. which microsatellites, wing morphometric, and Ballard, J.W.O., and M. C. Whitlock. 2004. The incomplete mtDNA haplotype data were used in a population natural history of mitochondria. Mol. Ecol. 13: 729Ð744. genetic analysis of three of the species (B. dorsalis, B. Barr, N. B., L. Cui, and B. A. McPheron. 2005. Molecular papayae, and B. philippinensis). They found no dis- systematics of nuclear gene period in genus Anastrepha cernible genetic or morphometric differences be- (Tephritidae). Ann. Entomol. Soc. Am. 98: 173Ð180. tween the species. Based on the preponderance of Bauzer, L. G., J. S. Gesto, N. A. Souza, R. D. Ward, J. G. recent data in support of this group actually repre- Hamilton, C. P. Kyriacou, and A. A. Peixoto. 2002. Mo- senting a single species, we suggest that modiÞcations lecular divergence in the period gene between two pu- tative sympatric species of the Lutzomyia longipalpis to agricultural regulations are in order so that regu- complex. Mol. Biol. Evol. 19: 1624Ð1627. lations match the best available science. Regulations Caterino M. S., S. Cho, and F. A. Sperling. 2000. The current based on the assumptions of the current multiple spe- state of insect molecular systematics: a thriving tower of cies descriptions by Drew and Hancock (1994) could Babel. Annu. Rev. Entomol. 45: 1Ð54. lead to shortfalls in mitigating and controlling future Boykin, L. M., M. K. Schutze, M. N. Krosch, A. Chomicˇ,T.A. introductions, as a country or locality deemed infested Chapman, A. Englezou, K. F. Armstrong, A. R. Clarke, D. by one species may not be regulated for another spe- Hailstones, and S. L. Cameron. 2013. Multi-gene phylo- cies within the clade. genetic analysis of southÐeast Asian pest members of the Instead of a group of distinct species, our results Bactrocera dorsalis species complex (Diptera: Tephriti- dae) does not support current taxonomy. J. Appl. Ento- suggest that, across multiple genes, none of the main mol. (in press) (doi:10.1111/jen.12047). pest species of B. dorsalis s.l. clade (B. carambolae, B. (CDFA) California Department of Food and Agriculture. dorsalis, B. invadens, and B. papayae) are monophy- 2012. Exotic fruit ßy project press releases. (http:// letic. This suggests alternative reasons for the variation www.cdfa.ca.gov/plant/pdep/treatment/press_releases. in morphology and even questions the signiÞcance of html). November 2013 SAN JOSE ET AL.: STATUS OF B. invadens AND SYSTEMATICS OF B. dorsalis 693

Cho, S. W., A. Mitchell, J. C. Regier, C. Mitter, R. W. Poole, polymerase chain reaction ampliÞcation primers of use in T. P Friedlander, and S. W. Zhao. 1995. A highly con- molecular phylogenetic research. Ann. Entomol. Soc. served nuclear gene for low-level phylogenetics: elonga- Am. 104: 976Ð997. tion factor-1-␣ recovers morphology-based tree for he- Hardy, D. E. 1969. Taxonomy and distribution of the ori- liothine moths. Mol. Biol. Evol. 12: 650Ð656. ental fruit and related species (Tephritidae-Diptera). Clarke, A. R., K. F. Armstrong, A. E. Carmichael, J. R. Milne, Proc. Hawaii Entomol. Soc. 20: 395Ð428. S. Raghu, G. K. Roderick, and D. K. Yeates. 2005. Inva- (IPPC) International Plant Protection Convention. 2012. sive phytophagous pests arising through a recent tropical South Africa. (https://www.ippc.int/index.pxp?idϭ evolutionary radiation: the Bactrocera dorsalis complex of nppozaf). fruit ßies. Annu. Rev. Entomol. 50: 293Ð319. Iwahashi, O. 1999a. Distinguishing between the two sym- Cognato, A. I. 2006. Standard percent DNA sequence dif- patric species and B. papayae ference for insects does not predict species boundaries. J. (Diptera: Tephritidae) based on aedeagal length. Ann. Econ. Entomol. 99: 1037Ð1045. Entomol. Soc. Am. 92: 639Ð643. Darriba, D., G. L. Taboada, R. Doallo, and D. Posada. 2012. Iwahashi, O. 1999b. Distinguishing between two sympatric jModelTest 2: more models, new heuristics and parallel species, Bactrocera occipitalis and B. philippinensis (Dip- computing. Nat. Methods 9: 772Ð772. tera: Tephritidae), based on aedeagal length. Ann. En- De Meyer, M., M. P. Robertson, M. W. Mansell, S. Ekesi, K. tomol. Soc. Am. 92: 182Ð187. Tsuruta, W. Mwaiko, J. -F. Vayssieres, and A. T. Peterson. Khamis, F. M., N. Karam, S. Ekesi, M.D.E. Meyer, A. Bonomi, 2010. Ecological niche and potential geographic distri- L. M. Gomulski, F. Scolari, P. Gabrieli, P. Siciliano, D. bution of the invasive fruit ßy Bactrocera invadens (Dip- Masiga, et al. 2009. Uncovering the tracks of a recent and tera, Tephritidae). Bull. Entomol. Res. 10: 35Ð48. rapid invasion: the case of the fruit ßy pest Bactrocera Drew, R.A.I. 1989. The tropical fruit ßies (Diptera: Te- invadens (Diptera: Tephritidae) in Africa. Mol. Ecol. 18: phritidae: ) of the Australasian and Oceanic re- 4798Ð4810. gions. Mem. Qld. Mus. 26: 1Ð521. Khamis F. M., D. K. Masiga, S. A. Mohamed, D. Salifu, M. de Drew, R.A.I. 2004. Biogeography and speciation in the Da- Meyer, and S. Ekesi. 2012. Taxonomic identity of the cini (Diptera: Tephritidae: Dacinae), pp. 165Ð178. In N. L. invasive fruit ßy pest, Bactrocera invadens: concordance Evenhuis and K. Y Kaneshiro (eds.), D. Elmo Hardy in morphometry and DNA barcoding. PLoS ONE 7: memorial volume. Contributions to the systematics and e44862. (doi:10.1371/journal.pone.0044862). evolution of Diptera. Bishop Museum, Honolulu, HI. Krosch, M. N., M. K. Schutze, K. A. Armstrong, G. C. Graham, Drew, R.A.I., and D. L. Hancock. 1994. The Bactrocera dor- D. K. Yeates, and A. R. Clarke. 2012. A molecular phy- salis complex of fruit ßies (Diptera: Tephritidae: Daci- logeny for the Tribe Dacini (Diptera: Tephritidae): sys- nae) in Asia. Bull. Entol. Res. Suppl. 2: 1Ð68. tematic and biogeographic implications. Mol. Phylo- Drew, R.A.I., and M. Romig. 2013. Tropical fruit ßies of genet. Evol. 64: 513Ð523. SouthÐEast Asia (Tephritidae: Dacinae). CAB Interna- Krosch, M. N., M. K. Schutze, K. A. Armstrong, Y. Boontop, tional, Wallingford, United Kingdom. (in press). L. M. Boykin, T. A. Chapman, A. Englezou, S. L. Cam- Drew, R.A.I., K. Tsuruta, and I. M. White. 2005. A new eron, and A. R. Clarke. 2013. Piecing together an inte- species of pest fruit ßy (Diptera: Tephritidae: Dacinae) grative taxonomic puzzle: microsatellite, wing shape and from Sri Lanka and Africa. Afr. Entomol. 13: 149Ð154. Drew, R.A.I., S. Raghu, and P. Halcoop. 2008. Bridging the aedeagus length analyses of Bactrocera dorsalis sl (Dip- morphological and biological species concepts: studies on tera: Tephritidae) Þnd no evidence of multiple lineages the Bactrocera dorsalis (Hendel) complex (Diptera: Te- in a proposed contact zone along the Thai/Malay Pen- phritidae: Dacinae) in southÐeast Asia. Biol. J. Linn. Soc. insula. Syst. Entomol. 38: 2Ð13. Lond. 93: 217Ð226. Liu, L., J. Liu, Q. Wang, P. Ndayiragije, A. Ntahimpera, E. Drummond, A., B. Ashton, S. Buxton, M. Cheung, A. Cooper, Nkubaye, Q. Yang, and Z. Li. 2011. IdentiÞcation of Bac- C. Duran, M. Field, J. Heled, M. Kearse, and S. Markow- trocera invadens (Diptera: Tephritidae) from Burundi, itz. 2011. Geneious v5. 4. (http://www. geneious.com). based on morphological characteristics and DNA bar- Edgar, R. C. 2004. MUSCLE: a multiple sequence alignment code. Afr. J. Biotechnol. 10: 13623Ð13630. method with reduced time and space complexity. BMC Lux, S. A., R. S. Copeland, I. M. White, A. Manrakhan, and Bioinformatics 5: 1Ð19. M. K. Billah. 2003. A new invasive fruit ßy species from (FDACS) Florida Department of Agriculture and Con- the Bactrocera dorsalis (Hendel) group detected in East sumer Services. 2012. Oriental fruit ßy information. Africa. Insect Sci. Appl. 23: 355Ð360. (http://www.freshfromßorida.com/pi/enpp/ento/off. Moulton, J. K., and B. M. Wiegmann. 2004. Evolution and html). phylogenetic utility of CAD (rudimentary) among Me- Foley, D. H., R. C. Wilkerson, R. D. Cooper, M. E. Volovsek, sozoic-aged Eremoneuran Diptera (Insecta). Mol. Phy- and J. H. Bryan. 2007. A molecular phylogeny of Anoph- logenet. Evol. 31: 363Ð378. eles annulipes (Diptera: Culicidae) sensu lato: the most Nakahara, S., and M. Muraji. 2008. Phylogenetic analyses of species-rich anopheline complex. Mol. Phylogenet. Evol. Bactrocera fruit ßies (Diptera: tephritidae) based on nu- 43: 283Ð297. cleotide sequences of the mitochondrial COI and COII Folmer, O., M. Black, W. Hoeh, R. Lutz, and R. Vrijenhoek. genes. Res. Bull. Plant Prot. 44: 1Ð12. 1994. DNA primers for ampliÞcation of mitochondrial Rambaut, A. 2012. FigTree, version 1.4.0. (http://tree.bio. cytochrome c oxidase subunit I from diverse metazoan ed.ac.uk/software/Þgtree/). invertebrates. Mol. Mar. Biol. Biotechnol. 3: 294Ð299. Rambaut, A., and A. J. Drummond. 2009. Tracer, version Funk, D. J., and K. E. Omland. 2003. Species-level para- 1.5. (http://beast.bio.ed.ac.uk/Tracer). phyly and polyphyly: frequency, causes, and conse- Ronquist, F., M. Teslenko, P. van der Mark, D. Ayres, A. quences, with Insights from animal mitochondrial DNA. Darling, S. Hohna, B. Larget, L. Liu, M. A. Suchard, and Annu. Rev. Ecol. Evol. Syst. 34: 397Ð423. J. P. Huelsenbeck. 2012. MrBayes 3.2: Ecient Bayesian Gibson, J. F., S. Kelso, M. D. Jackson, J. H. Kits, G.F.G. phylogenetic inference and model choice across a large Miranda, and J. H. Skevington. 2011. Diptera-speciÞc model space. Syst. Biol. 61: 539Ð542. 694 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 6

Rubinoff, D. 2006. Utility of mitochondrial DNA barcodes Swofford, D. L. 2003. PAUP*. Phylogenetic Analysis Using in species conservation. Conserv. Biol. 20: 1026Ð1033. Parsimony (*and Other Methods), version 4. Sinauer, Schutze, M. K., M. N. Krosch, F. K. Armstrong, T. A. Chap- Sunderland, MA. man, A. Englezou, A. Chomic, S. L. Cameron, D. Hail- Tan, K. H., I. Tokushima, H. Ono, and R. Nishida. 2011. stones, and A. R. Clarke. 2012. Population structure of Comparison of phenylpropanoid volatiles in male rec- Bactrocera dorsalis ss, B. papayae and B. philippinensis tal pheromone gland after methyl eugenol consump- (Diptera: Tephritidae) in southeast Asia: evidence for a tion, and molecular phylogenetic relationship of four single species hypothesis using mitochondrial DNA and global pest fruit ßy species: Bactrocera invadens, B. wingshape data. BMC Evol. Biol. 12Ð130: 1Ð15. dorsalis, B. correcta and B. zonata.. Chemoecology 21: Schutze, M. K., A. Jessup, I. Ul-Haq, M.J.B. Vreysen, V. Wor- 25Ð33. noayporn, M. T. Vera, and A. R. Clarke. 2013. Mating Virgilio, M., M. De Meyer, I. M. White, and T. Backejau. compatibility among four pest members of the of Bactro- 2009. African Dacus (Diptera: Tephritidae): molecular cera dorsalis fruit ßy species complex (Diptera: Tephriti- data and host plant associations do not corroborate mor- dae). J. Econ. Entomol. 106: 695Ð707. phology based classiÞcations. Mol. Phylogenet. Evol. 51: Simon, C., F. Frati, A. Beckenbach, B. Crespi, H. Liu, and P. 531Ð539. Flook. 1994. Evolution, weighting, and phylogenetic White, I. M., and M. M. Elson–Harris. 1992. Fruit ßies of utility of mitochondrial gene-sequences and a compila- economic signiÞcance: their identiÞcation and bion- tion of conserved polymerase chain-reaction primers. omics. CAB International, Wallingford, United King- Ann. Entomol. Soc. Am. 87: 651Ð701. dom. Smith, P. T., S. Kambhampati, and K. A. Armstrong. 2003. Zwickl, D. J. 2006. Genetic algorithm approaches for the Phylogenetic relationships among Bactrocera species phylogenetic analysis of large biological sequence data- (Diptera: Tephritidae) inferred from mitochondrial sets under the maximum likelihood criterion. Ph.D. dis- DNA sequences. Mol. Phylogenet. Evol. 26: 8Ð17. sertation, The University of Texas, Austin, TX. Sukumaran, J., and M. R. Holder. 2010. DendroPy: a Python library for phylogenetic computing. Bioinformatics 26: 1569Ð1571. Received 15 February 2013; accepted 24 June 2013.