Phylogenetic Analysis of North AmericanRhagoletis(Diptera: Tephritidae) and Related Genera Using Mitochondrial DNA Sequence Data

MOLECULAR PHYLOGENETICS AND EVOLUTION

Vol. 7, No. 1, February, pp. 1–16, 1997 ARTICLE NO. FY960369

Phylogenetic Analysis of North American Rhagoletis (Diptera: Tephritidae) and Related Genera Using Mitochondrial DNA Sequence Data BRUCE A. MCPHERON

AND

HO-YEON HAN1

Department of Entomology and Institute of Molecular Evolutionary Genetics, Pennsylvania State University, University Park, Pennsylvania 16802-3508 Received November 6, 1995; revised June 10, 1996

comprehensive bibliographies on damage, control, and research related to pest management (see Boller and Prokopy, 1976; White and Elson Harris, 1992). Bush’s (1966, 1969) arguments on the nature of speciation within certain Rhagoletis species groups resulted in an extensive literature on the formation of host races and its potential role in sympatric speciation (Bush, 1975; Feder et al., 1988, 1993; Feder and Bush, 1989; McPheron et al., 1988; Frias, 1992). Bush (1966) first proposed that some members of the R. pomonella species group (hereafter, species groups within the genus will be referred to by the specific epithet for which they are named) might have diverged in sympatry following a shift to a novel host plant. Several of the species groups, including the pomonella, cingulata, tabellaria, and nova groups (Bush, 1966; Foote, 1981; Berlocher, 1984), contain complexes of sympatrically distributed sibling species that differ in host plant choice. Bush suggested that a mutation driving choice of host plant could lead to reproductive isolation because of the biology of the organism. Members of the genus generally use the host fruit as the mating site (e.g., Smith and Prokopy, 1980), so preference for a novel host could cause immediate isolation from the parent population. Debate has raged over both the plausibility and universality of this mode of speciation (e.g., Futuyma and Mayer, 1980; Diehl and Bush, 1989), but available evidence supports the fact that, at least in sympatric R. pomonella populations on different host species, gene flow is restricted (Feder et al., 1988, 1990, 1993; McPheron et al., 1988; McPheron, 1990). A well-resolved phylogenetic hypothesis is required in order to better understand patterns of speciation within Rhagoletis and further resolve the evolution of phenotypic traits important for both reproductive isolation and pest management. No such resolution is currently available. Bush (1966) revised the North American species and defined some of the species groups discussed above (Table 1). This provides a rough indication of evolutionary relationships, but predicts

Partial sequences (approximately 850 bp) of the mitochondrial 16S ribosomal RNA gene were determined for 21 members of the fruit fly genus Rhagoletis and 6 related tephritid taxa by sequencing DNA amplified by the polymerase chain reaction. Sequences were highly A 1 T rich, with an average G 1 C content of 19.2%. Sequence divergence ranged from 0 to 11.7% among the included taxa. Sequences were used to reconstruct the phylogeny of this group of flies with neighbor-joining and maximum parsimony methods. A group of 18 North American Rhagoletis species formed a monophyletic clade, and three morphologically based species groups were identified with strong statistical support. Relationships among the species groups were partially resolved. The tabellaria species group appears to be paraphyletic with respect to the pomonella species group, in contrast to previous allozyme analyses that united the cingulata and pomonella groups. The relations of the North American clade to other genera within the tribe Carpomyini were not clearly resolved. Rhagoletis striatella, a North American species that uses members of the tomato family (Solanaceae) as a larval host, may be more closely associated with South American Rhagoletis species and related genera than with the other North American Rhagoletis species. r 1997 Academic Press

INTRODUCTION The genus Rhagoletis Loew has been studied extensively for two reasons: it contains important agricultural pests, and members of the genus have been central in studies of speciation. The agricultural importance of such species as R. pomonella (Walsh), the apple maggot, R. cerasi (L.), the European cherry fruit fly, and R. completa Cresson, the walnut husk fly, has led to 1 Current address: Department of Life Science, Yonsei University, Maji-ri 234, Heungup-myon, Wonju City, Kangwon-do 220-710, Republic of Korea.

1

1055-7903/97 $25.00 Copyright r 1997 by Academic Press All rights of reproduction in any form reserved.

2

MCPHERON AND HAN

TABLE 1 Species Group Designation within the Genus Rhagoletis from the New World (data from Bush (1966), Foote (1981), Herna´ndez-Ortiz (1985), and Foote et al. (1993)); Species Marked with an Asterisk Were Sequenced for This Study Species group pomonella cingulata suavis tabellaria ribicola alternata striatella nova psalida ferruginea unplaced

Species pomonella,* zephyria,* mendax,* cornivora* cingulata,* indifferens,* osmanthi,* chionanthi,* turpinae suavis,* completa,* juglandis,* boycei,* zoqui,* ramosae tabellaria,* juniperina,* persimilis, ebbettsi, electromorpha* ribicola,* berberis alternata, basiola,* meigenii striatella,* macquartii, jamaicensis nova, lycopersella, tomatis, conversa,* willinki, penela psalida, rhytida, metallica ferruginea, adusta, blanchardi acuticornis, fausta*

little about relationships of species within these groups or between the species groups. The same is true for delimitation of species groups in the Neotropical fauna (Foote, 1981) (Table 1). The only attempts at resolution of broad phylogenetic relationships within the genus have been by Berlocher (1981, Berlocher and Bush, 1982). Berlocher’s (1981) analysis of 78 morphological characters and his subsequent allozyme analysis (Berlocher and Bush, 1982) contrasted with several features of Bush’s (1966) species group designations and left the monophyly of the genus in question. They acknowledged that reanalysis of morphological and karyotypic data sets was necessary, along with the addition of new characters. Rhagoletis is one of at least nine genera comprising the tribe Carpomyini (treated as subtribe Carpomyina of the tribe Trypetini by Norrbom (1989, 1994), but these taxa do not belong to the tribe Trypetini, s. str. (Han, 1992; Han and McPheron, 1994)). Previous studies of the genus have assumed its monophyly, but authors dealing with the North American Rhagoletis fauna have noted close relationships with certain other genera, particularly Zonosemata Benjamin (Bush, 1966; Berlocher and Bush, 1982; Foote et al., 1993). Thus, any analysis of the phylogenetic relationships within the genus should consider representatives of as many of these related genera as possible. Molecular data sets provide a wealth of new characters that have contributed to resolution of phylogenetic issues left unresolved by other means (e.g., see Avise (1994) and references therein). In particular, mitochondrial DNA (mtDNA) sequence data from the large subunit (16S) ribosomal RNA gene have proven useful for examination of insect relationships at the specific to

family level (DeSalle et al., 1987; Xiong and Kocher, 1991; Derr et al., 1992; DeSalle, 1992; Simon et al., 1994). We provide a demonstration of the utility of 16S ribosomal DNA sequence data for resolution of species group relationships within New World members of the genus Rhagoletis. Our data identify new relationships among several North American species groups, including delimitation of a large proportion of North American species as a monophyletic group. MATERIALS AND METHODS Isolation of DNA Taxa included in this study are listed in Table 2, along with collection and voucher information. Vouchers are deposited in the Frost Entomological Museum, Pennsylvania State University. Single flies were used for all extractions; most flies were collected fresh and preserved frozen, but some specimens were pinned or preserved in ethanol (indicated in Table 2). Total nucleic acids were isolated by methods previously described (Sheppard et al., 1992) for frozen flies. Pinned or alcohol-preserved specimens were extracted by the modification of this protocol described in Han and McPheron (1997). PCR Amplification and Sequencing Primers LR-N-13398 (58-CGCCTGTTTATCAAAAACAT-38) and LR-J-12883 (58-CTCCGGTTTGAACTCAGATC-38) (primers A and B, respectively, of Xiong and Kocher (1991)) and LR-N-13770 (58-AGAAATGAAATGTTATTCGT-38) and LR-J-13323 (58-ACTAATGATTATGCTACCTT-38) were used for symmetrical DNA amplifications by the polymerase chain reaction (PCR). PCR was performed in 50-µl reaction volumes using the following conditions: 13 ProMega reaction buffer, 250 µM each of dATP, dCTP, dGTP, and dTTP, 1.25 µM each primer, 2.5 units of ProMega Taq polymerase, and 1 µl of template DNA, overlaid with approximately 50 µl of mineral oil. The cycle program consisted of 35 cycles of 1 min at 93°C, 1 min at 45°C, and 2 min at 72°C with fastest possible transitions between temperatures. The final cycle had an additional 7 min at 72°C. Doublestranded amplification products were gel purified in 1.5% agarose gels (13 TAE) as described by Han and McPheron (1997). DNA thus released from the gel slice was used as template for asymmetrical amplifications in 60-µl volumes, with amplification conditions as above except for the concentration of the limiting primer (0.0125–0.05 µM). Single-stranded DNA was prepared for sequencing by washing (300 µl sterile water) and concentrating three times in Millipore Ultrafree-MC 30,000 MW filters. Sequencing of single-stranded DNA followed the Sequenase 2.0 (U.S. Biochemical, Cleveland, OH) protocol. All primers listed above for symmetrical amplification were used as limiting primers (at ratios between

Rhagoletis MOLECULAR PHYLOGENY

3

TABLE 2 Collection Data, Preservation Condition of Specimen, and Voucher Information (in Parentheses) for Taxa Sequenced in This Study; BAM and HYH Refer to Collections Made by the Authors Cryptodacus tau (Foote). Guatemala: Escuintala, Palin area, McPhail trap, 1992–1993, J. Lopez; in ethanol (wings glued to card point, additional specimens from same collecting lot in USNM) Euphranta canadensis (Loew). USA: Wyoming, 17–18.VII.1979, S. H. Berlocher and G. J. Steck; frozen (wings glued to card point) Haywardina cuculi (Hendel). Argentina: Tucuman, Burruyacu, Taruca Pampa, Finca San Augustine, ex Solanum trichoneuron fruit, col. 18.V.1991, A. L. Norrbom; pinned (wings glued to card point) Oedicarena latifrons (Wulp). Mexico: Mexico, Rt. 890, km. 9, ex Solanum brachycarpum fruit, col. 1991, em IX.1992, A. L. Norrbom; frozen (wings glued to card point) Rhagoletis basiola (Osten Sacken). USA: Pennsylvania: Centre Co., near Port Matilda, malaise trap, 9–14.VII.1993, BAM; frozen (left wing glued to card point) Rhagoletis boycei Cresson. USA: Arizona: Graham Co., Arcadia Campground in Pinalen˜o Mtns., ex Juglans major fruit, 1.IX.1993, D. Papaj; frozen (1?, 1/ from same series) Rhagoletis chionanthi Bush. USA: Georgia: Peach Co., near Fort Valley, ex Chionanthus virginiana fruit, 18.VII.1989, S. Berlocher; frozen (wings glued to card point) Rhagoletis cingulata (Loew). USA: Illinois: Champaign Co., Lake of the Woods, ex Prunus serotina fruit, 1986, BAM; frozen (1?, 1/ from same series) Rhagoletis completa Cresson. USA: Kansas: Riley Co., Fort Riley, ex Juglans nigra fruit, S. Berlocher; frozen (1?, 1/ from same series) Rhagoletis conversa (Bre`thes). Chile: D. Frias; pinned (1?, 1/ from same series) Rhagoletis cornivora Bush. USA: Pennsylvania: Centre Co., Houserville, ex Cornus amomum fruit, 1989, BAM; frozen (1?, 1/ from same series) Rhagoletis electromorpha Berlocher. USA: Pennsylvania: Centre Co., Penn State campus, ex Cornus racemosa fruit, 14.VIII.1991, HYH; frozen (1?, 1/ from same series) Rhagoletis fausta (Osten Sacken). USA: Wisconsin: Door Co., 1984, J. Feder; frozen (1?, 1/ from same series) Rhagoletis indifferens Curran. Canada: British Columbia, Peachland, ex Prunus emarginata fruit, VII.1974, S. Berlocher; frozen (1? from same series) Rhagoletis juglandis Cresson. USA: Arizona: Santa Cruz Co., Patagonia, ex Juglans major fruit, 29.VII.1993, D. Papaj; frozen (1?, 1/ from same series) Rhagoletis juniperina Marcovitch. USA: Illinois: Champaign Co., Univ. of Illinois campus, ex Juniperina virginiana fruit, X.1979, S. Berlocher; frozen (1? from same series) Rhagoletis mendax Curran. USA: Michigan: Berrien Co., Chickaming, ex Vaccinium corymbosum fruit; frozen (1?, 1/ from same series) Rhagoletis osmanthi Bush. USA: Florida: Osceola Co., Alligator Lake, ex Osmanthus americana fruit, 4.XI.1984, S. Berlocher; frozen (1? from same series) Rhagoletis pomonella (Walsh). USA: Pennsylvania: Centre Co., Penn State campus, ex Crataegus mollis fruit, col. IX.1988, em. VII.1989, BAM; frozen (1?, 1/ from same series) Rhagoletis pomonella. USA: Pennsylvania: Centre Co., Penn State campus, ex Crataegus mollis fruit, col. IX.1989, em. VIII.1990, BAM; frozen (1?, 1/ from same series) Rhagoletis ribicola Doane. USA: Wyoming, 17–18.VII.1979, S. H. Berlocher and G. J. Steck; frozen (1 frozen individual) Rhagoletis striatella Wulp. USA: Pennsylvania: Centre Co., Pleasant Gap, ex Physalis subglabrata fruit, col. 3.IX.1989, em 15–22.VI.1990, HYH; frozen (1?, 1/ from same series) Rhagoletis suavis (Loew). USA: Pennsylvania: Centre Co., Penn State campus, ex Juglans cinerea fruit, col. IX.1988, em. 1989, BAM; frozen (1?, 1/ from same series) Rhagoletis tabellaria (Fitch). USA: Pennsylvania: Centre Co., Houserville, ex Cornus stolonifera fruit, col. VII.1989, em. 1990, BAM; frozen (1?, 1/ from same series) Rhagoletis zephyria Snow. USA: Washington: Skamania Co., Stevenson, ex Symphoricarpos sp. fruit, col. 4.VIII.1986, em. 1987, BAM; frozen (1?, 1/ from same series) Rhagoletis zephyria. USA: Pennsylvania: Centre Co., Penn State campus, ex Symphoricarpos sp. fruit, 20.VIII.1989, HYH; frozen (1?, 1/ from same series) Rhagoletis zoqui Bush. Mexico: Veracruz, near Perote, ex Juglans sp. fruit, 1985, S. H. Berlocher; pinned (wings glued to card point) Rhagoletotrypeta pastranai Acze´l. Brazil: Santa Catarina, Cocador, ex. Celtis iguanae, col. 16.III.1987; pinned (wings glued to card point) Zonosemata electa (Say). USA: Texas: Nacogdoches Co., 7 mi. S of Nacogdoches, ex Solanum carolinense fruit, col. 30.VI.1984, S. H. Berlocher; frozen (wings glued to card point)

1:25 and 1:100, determined empirically) in asymmetrical amplifications and as sequencing primers. In addition, primers LR-J-13021 (58-ACGCTGTTATCCCTAAAGTA-38) and LR-N-13182 (58-TTAAAAGACGAGAAGACCCTA-38) were used as limiting and sequencing primers. Sufficient overlap in primer location generated approximately 40% overlap of base positions from the two strands. Regions for which sequence data from both strands were not available were free from

compressions and other irregularities that would introduce errors in reading autoradiographs. Phylogenetic Analysis Initial sequence alignment was conducted using the CLUSTAL software package (Higgins and Sharp, 1989), followed by manual refinement using ESEE software (Cabot and Beckenbach, 1989) (Appendix 1). Membership in the tribe Carpomyini, as defined by morphologi-

4

MCPHERON AND HAN

cal criteria (Norrbom, 1989), comprised our a priori ingroup designation. Monophyly of the ingroup was tested against two taxa, Euphranta canadensis (Loew) and Oedicarena latifrons (Wulp), selected as outgroups on the basis of previous morphological and molecular investigations (Berlocher and Bush, 1982; Norrbom, 1989; Han and McPheron, 1997). Relationships were analyzed using distance and maximum parsimony methods. A neighbor-joining (NJ) tree was generated from the Kimura two-parameter distance (Kimura, 1980) using MEGA (Kumar et al., 1993). We also examined NJ topologies using Jukes–Cantor, Tamura, and Tamura– Nei distances and gamma corrections (gamma 5 0.5) for the Jukes–Cantor and Tamura–Nei distances. All positions containing gaps or ambiguous characters were deleted. Support for the NJ topology was tested by bootstrapping (Felsenstein, 1985) (2000 replicates) and the standard error test of interior branch lengths (Rzhetsky and Nei, 1992; Kumar et al., 1993). The heuristic search procedure implementing the treebisection-reconnection algorithm and the MULPARS option was used to find minimum length trees using the maximum parsimony (MP) procedures in PAUP (Swofford, 1993). All characters were treated as unordered. Bias due to taxon input order was evaluated by stepwise random addition of sequences (20 replications). A strict consensus of the minimum length MP trees was calculated. Bootstrapping of the MP analysis (200 replicates) was also conducted under the heuristic search procedure, with a maxtree setting of 150 trees. Percentage divergence among species and transition/ transversion ratios were calculated using MEGA. The average frequency of nucleotide changes was calculated in MacClade 3.04 (Maddison and Maddison, 1992) over the NJ topology. Sequences from this study have been deposited in GenBank (Accession Nos. U39390 and U39413– U39440). The alignment is available as a Nexus file from the authors. RESULTS Sequence Evolution in 16S Ribosomal DNA Between 843 and 853 base pairs were sequenced, amounting to a total of 862 alignable positions, including gaps inserted to improve alignments. This sequence corresponds to positions 12912–13754 in the Drosophila yakuba mtDNA sequence (Clary and Wolstenholme, 1985). The uncorrected sequence divergence among taxa in this study ranged from 0% (R. mendax and R. zephyria East in the pomonella group and R. chionanthi and R. cingulata in the cingulata group) to 11.7% (between R. striatella and Zonosemata electa). The distribution of sequence divergence is trimodal. Several comparisons among members of the pomonella and cingulata groups have sequence divergences of less

than 0.5%. A second cluster of sequence divergences between 3 and 5% represent comparisons among most of the North American Rhagoletis species. A final group of divergences, from 7 to 11%, represents pairwise comparisons of these North American Rhagoletis with the other included taxa and comparisons of these taxa with each other. This portion of the sequenced region is highly A–T rich. Average nucleotide composition across all taxa was 36.7% A, 44.1% T, 6.7% C, and 12.5% G. There are relatively few transversions up to a genetic divergence of about 5%, but there is a tendency toward saturation at distances greater than this. The divergences of less than 5% correspond to comparisons among and within North American Rhagoletis species. Variable sites are not randomly distributed across the region of the 16S rRNA gene we sequenced. Substantial variation (56% of variable sites and 53% of sites informative for parsimony) lies in the 58 portion of our sequence data (roughly, positions 13400–13754 in the D. yakuba sequence, or the DNA region amplified by primers LR-N-13770 and LR-J-13323). The average number of substitutions inferred over the NJ topology (uncorrected for multiple hits) is asymmetrical (Table 3). The pattern of transitions shows the greatest bias, with A = G changes occurring about 4 times more often than G = A. Likewise, T = C transitions are 1.7 times more common than the reciprocal change. Phylogenetic Reconstruction After exclusion of positions containing gaps or ambiguous character states from the alignment, 823 base pairs were used for phylogenetic analysis. Of these, 231 sites were variable and 136 sites were informative for parsimony analysis. All distance measures except the Tamura–Nei distance yielded identical topologies. NJ analysis with either the Tamura–Nei distance alone or the Tamura–Nei distance with a gamma correction placed E. canadensis into a position near R. basiola. Our analyses with greater taxonomic representation of tephritids (Han and McPheron, 1997) suggests that E. canadensis belongs outside the Carpomyini. We chose to report an NJ topology (Fig. 1) based upon the Kimura two-parameter distance measure (following the guide-

TABLE 3 Average Number of Substitutions (Uncorrected for Multiple Hits) Inferred over the Neighbor-Joining Topology To: A From:

A C G T

5 33 129

C

G

T

2

136 1

105 25 17

2 42

26

FIG. 1. Phylogeny of Rhagoletis and related species based on neighbor joining reconstruction using the Kimura two-parameter distance. Numbers on branches are confidence estimates from the standard error test/bootstrap test (only standard error values greater than 70 are reported). Species group names for North American species groups are indicated at the right side of the figure.

Rhagoletis MOLECULAR PHYLOGENY

5

6

MCPHERON AND HAN

lines of Nei (1991) and Kumar et al. (1993), we chose a relatively simple distance measure that minimizes the variance of the distance estimate). The MP topology derived from equal weighting of characters generated 128 trees of length 522. These trees had a consistency index (calculated from informative characters only) of 0.48, a retention index of 0.65, and rescaled consistency index of 0.31. A strict consensus tree of all most parsimonious trees is shown in Fig. 2, with bootstrap support indicated. We attempted to calculate Bremer’s (1994) branch support index, but failed due to computational limits. The apparent substitution bias mentioned above (Table 3) might suggest that differential weighting is appropriate for MP analysis. We tried several options, but weight matrices generated by MacClade violated the triangle inequality (data not shown), and differentially weighted MP trees were not better resolved than the consensus tree presented. The topology generated by successive weighting (consensus of four most parsimonious trees) was identical to the consensus results from equal weighting, with the exception of the placement of a single member of the suavis species group (but this placement occurred in some of the most parsimonious trees from equal weighting).

Monophyly of the Tribe Carpomyini Our initial hypothesis of monophyly for this tribe was based upon morphological characters (Norrbom, 1994). While both NJ and MP analyses provide topological support for a single carpomyine lineage relative to E. canadensis and O. latifrons (Figs. 1 and 2), the level of confidence is not robust. The low confidence estimates of branches near the apparent root position of the Carpomyini (Fig. 1) suggest that the true root may not be recovered by this analysis. Relationships among Neotropical Genera of Carpomyini Representatives of the related genera Cryptodacus, Haywardina, Rhagoletotrypeta, and Zonosemata formed a single lineage along with South American R. conversa and North American R. striatella (Figs. 1 and 2). However, statistical support for this group and relationships of taxa within the group is very low. Monophyly of the North American Rhagoletis Twenty of 26 species of Rhagoletis native to North America (including Mexico) were examined. A group of 18 of these species (from R. pomonella to R. completa in Figs. 1 and 2) was recovered with high confidence as a monophyletic assemblage. This group includes members of four previously defined species groups, R. ribicola, whose species group assignment was tentative, and R. fausta, which has never been assigned to a species group (Bush, 1966). R. basiola branches from the other taxa near the root in both NJ and MP analyses (Figs. 1 and 2). This species is the only North American native of a Palaearctic species group (Table 1). R. striatella appears to be more closely related to a South American species, R. conversa, and, possibly, the New World genera Cryptodacus, Haywardina, Rhagoletotrypeta, and Zonosemata than it is to other North American Rhagoletis spp. Status of Species Groups

FIG. 2. Phylogeny of Rhagoletis and related species based on maximum parsimony reconstruction. This topology is a 50% majority rule consensus tree, with numbers on branches showing results of bootstrap analysis.

Phylogenetic analysis of the 16S rDNA data recovered monophyletic pomonella, cingulata, and suavis species groups (Figs. 1 and 2). Three members of the pomonella group are very similar, with R. mendax and one of the R. zephyria samples showing no differences over the entire sequence. This is the only species group for which multiple individuals were analyzed (two each of R. zephyria and R. pomonella). The two R. zephyria individuals represent the native range (western) and an area to which the species has been introduced and now coexists with R. pomonella (eastern). There were no diagnostic sequence differences among these three species. Association of R. cornivora, a sibling species to the other three described pomonella group species (Bush, 1966), was topologically supported, but statistical support was weak.

Rhagoletis MOLECULAR PHYLOGENY

The cingulata group was recovered with high support (Figs. 1 and 2). Genetic distances among these four species are very low, so no resolution of relationships among the constituent species is provided by this analysis. Inclusion of five walnut-infesting species in the suavis group was clearly supported by our analysis (Figs. 1 and 2). Sister relationships between R. zoqui and R. completa and between R. suavis and R. juglandis are indicated. The tabellaria species group was not monophyletic in this analysis. Rhagoletis juniperina clustered outside the pomonella group, while the sibling species pair R. tabellaria and R. electromorpha, sister taxa in our analysis, formed a sister taxon to this group (Fig. 1). The support for the nodes linking this assemblage is low. R. fausta has never been placed in a species group within the genus. Our analysis places it just outside the tabellaria–pomonella group assemblage (Figs. 1 and 2) discussed previously. Monophyly of this group is supported by our data. The phylogenetic position of R. ribicola is the least clear among these North American taxa. It definitely belongs to the North American species cluster and may be a sister taxon to the cingulata species group, although the statistical support for this relationship is not strong (Fig. 1). DISCUSSION Sequence Evolution in 16S Ribosomal DNA Evolution of fly mitochondrial ribosomal genes has been examined by several authors (e.g., DeSalle et al., 1987; Xiong and Kocher, 1991; DeSalle, 1992; Simon et al., 1994). Rhagoletis and related genera display many of the same features previously discussed, specifically, a clear A 1 T bias, a tendency to accumulate transversions as distance between taxa increases, and a substitutional bias. The last feature is certainly influenced by whatever molecular events have led to the A 1 T bias. These two sequence features are intimately tied to the transition/transversion ratio, because the sequence difference most commonly observed is A & T (Table 3), and, once a transversion occurs, evidence of transitions at that site is obliterated (Brown et al., 1982; DeSalle et al., 1987). The prevalence of A = G transitions (Table 3) reflects the number of relatively closely related comparisons among the North American Rhagoletis species. Most previous studies have focused on the 38 500 bp of the 16S gene (DeSalle et al., 1987; Xiong and Kocher, 1991; DeSalle, 1992). Our examination of an additional 350 bp of sequence 58 to this region provided us with a substantial increase in phylogenetically informative characters. This region should be considered for future studies using sequence from the 16S gene.

7

Phylogenetic Reconstruction NJ and MP methods recovered similar information from the 16S data set analyzed here. While, in theory, composition and substitution biases could compromise our ability to recover phylogenetic signal, it appears that the divergences among species in this study are sufficiently small to obviate this problem. The NJ tree was actually one step longer than the most parsimonious MP trees, a difference due to the placement of R. basiola relative to the root position. This is a trivial change, since bootstrap support for nodes in that portion of the topology is low. Species for which only pinned or alcohol-preserved specimens were available were recovered at expected positions within the topology, suggesting that contamination was not a problem. It is desirable to combine molecular, morphological, biogeographical, and ecological data sets into a single analysis, although the exact approach remains controversial. No comprehensive compilation of morphological characters is currently available for Rhagoletis that would make this possible. However, these data may soon be accessible (J. Jenkins, Michigan State University Ph.D. dissertation, personal communication) and, combined with data from Norrbom (1994), would permit a very informative investigation of evolutionary relationships in the group. Relationships among Neotropical Genera of Carpomyini Genetic distances between taxa in our study cluster in three groups. The highest genetic divergences are among non-Rhagoletis members of the Carpomyini and between those taxa and North American Rhagoletis species. There is little phylogenetic resolution of these taxa. The South American species R. conversa clusters with North American R. striatella (Figs. 1 and 2). Both species use members of the Solanaceae as larval hosts, as do many of the South American Rhagoletis species (Foote, 1981). Other included taxa, Zonosemata, Haywardina, and Oedicarena, also use the Solanaceae and reach their highest species diversity in Latin America (Norrbom et al., 1988; Norrbom, 1994). The allozyme analysis of Berlocher and Bush (1982) placed R. striatella in a cluster with Z. electa and O. latifrons, an affinity also suggested by morphology and host plant use (Bush, 1966), and Smith and Bush (1997) recovered a clade containing R. striatella and O. latifrons based on cytochrome oxidase II sequence data. However, the largest genetic divergence among any two species in our study is between R. striatella and Z. electa (11.7%). Relationships among genera of Carpomyini and their association with North American (or Palaearctic) Rhagoletis are not resolved by morphology (Norrbom, 1994) or molecular data (Han and McPheron, 1997). We do not feel that it is simply a matter of having reached a genetic distance where multiple hits have reduced the information content of the 16S gene to a useless level.

8

MCPHERON AND HAN

Our results (Han and McPheron, 1997) support the monophyly of many other tephritid tribes and subfamilies. We suspect that the genus Rhagoletis is not a monophyletic taxon relative to other carpomyine genera, but this conclusion must be verified by a comprehensive molecular and morphological analysis including more representatives of these taxa. Monophyly of the North American Rhagoletis The second level of genetic divergence, from 3 to 5%, involves comparisons among the members of the North American Rhagoletis species groups. The 18 North American species identified as a monophyletic group in this study were also recovered by allozyme analysis (Berlocher and Bush, 1982) and cytochrome oxidase II sequence analysis (Smith and Bush, 1997). Only two of the five species groups represented in this clade, the tabellaria and ribicola groups, are reported to have members outside the New World (Bush, 1966; Berlocher and Bush, 1982). Inclusion of some of those taxa to further test the evolutionary history of divergence in the genus would be appropriate. Significant support was demonstrated for two species groups, the cingulata and suavis groups, as defined by Bush (1966). The pomonella group sensu Bush was topologically supported, but inclusion of the sibling species R. cornivora was poorly supported (Figs. 1 and 2). So much of the allozyme change within the pomonella species group occurs on the branch leading to R. cornivora that UPGMA analyses of allozyme data placed other species between R. cornivora and the remaining members of the pomonella group (Berlocher and Bush, 1982; Berlocher et al., 1993). Smith and Bush (1997) also failed to recover R. cornivora as a sister to the remaining members of the pomonella group. As with allozymes, most of the pomonella group sequence changes in the 16S gene occur on the branch leading to R. cornivora. The tabellaria group was paraphyletic with respect to the pomonella group (Figs. 1 and 2), but statistical support for branching patterns was not strong. The sibling species R. tabellaria and R. electromorpha were sister taxa. R. juniperina is excluded from a clade with R. tabellaria and R. electromorpha, but, otherwise, its position is not clear within the group. The association of the tabellaria group with the pomonella group contrasts with results from electrophoretic analyses (Berlocher and Bush, 1982; Berlocher et al., 1993), which have consistently recovered the cingulata group with the pomonella group. However, using MacClade, we found that the shortest topology with a monophyletic cingulata group as a sister taxon to the pomonella group is 10 steps longer than the most parsimonious topologies. Smith and Bush (1997) included R. persimilis and R. tabellaria from western North America in their analysis and also supported an association of the

tabellaria group (minus R. juniperina) with the pomonella group. Bush (1966) was unable to place R. fausta within any species group, noting its similarity to various Rhagoletis species and some non-Rhagoletis tephritids. Allozyme analysis linked R. fausta with the suavis group (Berlocher and Bush, 1982). Our results (Fig. 1) unambiguously place this species with the tabellaria– pomonella clade. Analysis of the evolution of host use patterns would be of great interest in this lineage, since there is widespread use of the plant families Rosaceae, Cornaceae, and Ericaceae. Definition of the plesiomorphic condition and direction of evolutionary transitions would add to our understanding of the role of host plants in speciation patterns. R. ribicola has been synonymized with R. tabellaria in the past (Cresson, 1929), and these two species were sister taxa in the allozyme analysis of Berlocher and Bush (1982). However, Bush (1966) suggested that the morphological similarities were not convincing and predicted that R. ribicola was related to the cingulata group. Our results are not strong, but we do demonstrate topological support for Bush’s prediction. Relationships within the suavis species group recovered in this study are at odds with assessments based upon morphology (Bush, 1966) and allozymes (Swofford and Berlocher, 1987). We found a significant relationship between R. completa and R. zoqui, with R. boycei apparently related to this pair (Fig. 1), as also found by Smith and Bush (1997). The morphological and allozyme analyses united R. zoqui and R. boycei, but differed in their placement of R. completa. Five additional steps were required with our data set to recover the topology found by Swofford and Berlocher (1987). R. suavis, whose distribution in the eastern United States is most disjunct from other members of the species group, is also the most genetically distinct member of the group. The suavis species group, which is composed of morphologically distinct flies using a single host genus (Juglans), is different in its patterns of intragroup genetic divergence than the cingulata and pomonella groups. Whether this reinforces Bush’s (1966, 1969, 1975) arguments about modes of speciation within these groups remains to be rigorously tested. Status of the pomonella and cingulata Species Groups Members of the pomonella group (minus R. cornivora) and the cingulata group are genetically very similar, exhibiting divergences of less than 0.5%. This low level of divergence is accompanied by a relatively high transition/transversion ratio. This combination suggests a relatively recent divergence among members of these species groups. Members of both groups have been put forward as candidates for sympatric speciation mechanisms by host race formation by Bush (1966, 1969, 1975), who argued that such speciation could occur rapidly (in ecological time).

9

Rhagoletis MOLECULAR PHYLOGENY

Relationships within the pomonella group have been studied in detail. Berlocher et al. (1993) found a close relationship between R. zephyria and R. pomonella based on allozyme analysis (the undescribed taxon referred to as the flowering dogwood fly in that reference was not included in the present study). R. zephyria, R. pomonella, and R. mendax display the same broad patterns of shared alleles at enzyme loci that we observed in mitochondrial DNA sequences. Hybridization between R. pomonella and R. zephyria has been documented in areas where they occur sympatrically, as a result of either introduction (McPheron, 1990) or natural occurrence (Williams, Feder, McPheron, and Bush, unpublished data), while hybridization has not been observed among other members of the species group (Feder and Bush, 1989; Smith et al., 1993). The lack of diagnostic differences among the these pomonella group species could reflect shared ancestral polymorphisms, incomplete reproductive isolation at present, or both. The large genetic difference between R. cornivora and the other members of the pomonella species group (Fig. 1; Berlocher et al., 1993) places it in an ancestral position relative to the group. The cingulata group also displayed genetic divergences of less than 0.5%. This group has been less well studied than the pomonella group, but Bush (1966, 1969, 1975) discussed potential patterns of speciation among the constituent species. R. indifferens may exhibit the same sort of host race formation, involving native and domestic hosts in the western United States, that has occurred in R. pomonella in the eastern United States (Bush, 1969). Bush (1969) proposed three hypotheses for the evolution of R. osmanthi and R. chionanthi, the two species attacking Oleaceae. All of his hypotheses suggest that one of these species is derived from the other. Our data do not clearly resolve the relationships within the cingulata group, but also do not provide evidence for a sister relationship between R. osmanthi and R. chionanthi. Other gene regions with higher variability will be required to further resolve questions involving the evolution of species in the pomonella and cingulata groups.

Perspectives The mitochondrial 16S ribosomal RNA gene permits us to address certain key questions regarding the evolution of North American Rhagoletis species. First, members of the pomonella, tabellaria, cingulata, and suavis species groups (Table 1) plus R. fausta and R. ribicola form a clade that is well supported by both NJ and MP analysis (Figs. 1 and 2). Further analysis including putative Palaearctic members of the tabellaria species group is needed. Second, patterns of genetic divergence are markedly different among different species groups. The pomonella and cingulata species groups contain very closely related taxa, while the suavis species group consists of species that are more divergent. Speciation mechanisms within these groups still require study, but our results are not inconsistent with past arguments (e.g., Bush, 1969) that the suavis group has speciated quite differently from the pomonella and cingulata groups. A more rapidly evolving gene will be required for further study of phylogenetic relationships within these species groups. The relationship of the genus Rhagoletis to other members of the tribe Carpomyini remains unclear. Our results place R. conversa and R. striatella near exemplars of the genera Zonosemata, Haywardina, Rhagoletotrypeta, and Cryptodacus, but there is virtually no resolution of intergeneric ties. While we suspect that the Solanaceae-infesting Rhagoletis may be more closely related to some of these Neotropical taxa than to the Nearctic and Palaearctic representatives of Rhagoletis, our analysis does not provide sufficient resolution of this question. Additional species from these lineages should be sampled to investigate these relationships. APPENDIX 1 Alignment of 16S rDNA sequences. Gaps inserted to improve alignment are indicated by a hyphen, and missing or ambiguous characters are indicated by a question mark. Abbreviated taxon names can be interpreted from Table 1.

10

MCPHERON AND HAN

Rhagoletis MOLECULAR PHYLOGENY

11

12

MCPHERON AND HAN

Rhagoletis MOLECULAR PHYLOGENY

13

14

MCPHERON AND HAN

Rhagoletis MOLECULAR PHYLOGENY

ACKNOWLEDGMENTS We thank Stewart Berlocher, Jeff Feder, Daniel Frias, Allen Norrbom, and Dan Papaj for providing specimens for analysis. Stewart Berlocher, Gail Fitzhugh, Blair Hedges, Jim Smith, and anonymous reviewers commented on the manuscript. Support for this research came from Pennsylvania Agricultural Experiment Station Projects 3128 and 3463.

REFERENCES Avise, J. C. (1994). ‘‘Molecular Markers, Natural History and Evolution,’’ Chapman & Hall, New York. Berlocher, S. H. (1981). A comparison of molecular and morphological data, and phenetic and cladistic methods, in the estimation of phylogeny in Rhagoletis (Diptera: Tephritidae). In ‘‘Application of Genetics and Cytology in Insect Systematics and Evolution. Proc. Symp. of the Natl. Meeting of the Entomol. Soc. Amer., Atlanta, December 1–2, 1980’’ (M. W. Stock, Ed.), pp. 1–31, Forest, Wildlife and Range Experimental Station, Univ. of Idaho, Moscow. Berlocher, S. H. (1984). A new North American species of Rhagoletis (Diptera: Tephritidae), with records of host plants of Cornusinfesting Rhagoletis. J. Kans. Entomol. Soc. 57: 237–242. Berlocher, S. H., and Bush, G. L. (1982). An electrophoretic analysis of Rhagoletis (Diptera: Tephritidae) phylogeny. Syst. Zool. 31: 136–155. Berlocher, S. H., McPheron, B. A., Feder, J. L., and Bush, G. L. (1993). Genetic differentiation at allozyme loci in the Rhagoletis pomonella (Diptera: Tephritidae) species complex. Ann. Entomol. Soc. Am. 86: 716–727. Boller, E. F., and Prokopy, R. J. (1976). Bionomics and management of Rhagoletis. Annu. Rev. Entomol. 21: 223–246. Bremer, K. (1994). Branch support and tree stability. Cladistics 10: 295–304. Brown, W. M., Prager, E. M., Wang, A., and Wilson, A. C. (1982).

15

Mitochondrial DNA sequences of primates: Tempo and mode of evolution. J. Mol. Evol. 18: 225–239. Bush, G. L. (1966). The taxonomy, cytology, and evolution of the genus Rhagoletis in North America (Diptera, Tephritidae). Bull. Mus. Comp. Zool. 134: 431–562. Bush, G. L. (1969). Sympatric host race formation and speciation in frugivorous flies of the genus Rhagoletis (Diptera: Tephritidae). Evolution 23: 237–251. Bush, G. L. (1975). Sympatric speciation in phytophagous parasitic insects. In ‘‘Evolutionary Strategies of Parasitic Insects and Mites’’ (P. W. Price, Ed.), pp. 187–206, Plenum, New York. Cabot, E. L., and Beckenbach, A. T. (1989). Simultaneous editing of multiple nucleic acids and protein sequences with ESEE. Comput. Appl. Biosci. 5: 233–234. Clary, D. O., and Wolstenholme, D. R. (1985). The mitochondrial DNA molecule of Drosophila yakuba: Nucleotide sequence, gene organization, and genetic code. J. Mol. Evol. 22: 252–271. Cresson, E. T., Jr. (1929). A revision of the North American species of fruit-flies in the genus Rhagoletis (Diptera: Trypetidae). Trans. Am. Entomol. Soc. 55: 401–414. Derr, J. N., Davis, S. K., Woolley, J. B., and Wharton, R. A. (1992). Reassessment of the 16S rRNA nucleotide sequence from members of the parasitic Hymenoptera. Mol. Phylogenet. Evol. 1: 338–341. DeSalle, R. (1992). The phylogenetic relationships of flies in the family Drosophilidae deduced from mtDNA sequences. Mol. Phylogenet. Evol. 1: 31–40. DeSalle, R., Freedman, T., Prager, E. M., and Wilson, A. C. (1987). Tempo and mode of sequence evolution in mitochondrial DNA of Hawaiian Drosophila. J. Mol. Evol. 26: 157–164. Diehl, S., and Bush, G. L. (1989). The role of habitat preference in adaptation and speciation. In ‘‘Speciation and Its Consequences’’ (D. Otte and J. A. Endler, Eds.), pp. 345–365, Sinauer, Sunderland, MA. Feder, J. L., and Bush, G. L. (1989). A field test of differential host-plant usage between two sibling species of Rhagoletis po-

16

MCPHERON AND HAN

monella fruit flies (Diptera: Tephritidae) and its consequences for sympatric models of speciation. Evolution 43: 1813–1819. Feder, J. L., Chilcote, C. A., and Bush, G. L. (1988). Genetic differentiation between sympatric host races of the apple maggot fly Rhagoletis pomonella. Nature 336: 61–64. Feder, J. L., Chilcote, C. A., and Bush, G. L. (1990). The geographic pattern of genetic differentiation between host-associated populations of Rhagoletis pomonella (Diptera: Tephritidae) in the eastern United States. Evolution 44: 570–594. Feder, J. L., Hunt, T. A., and Bush, G. L. (1993). The effects of climate, host plant phenology and host fidelity on the genetics of apple and hawthorn infesting races of Rhagoletis pomonella. Entomol. Exp. Appl. 69: 117–135. Felsenstein, J. (1985). Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791. Foote, R. H. (1981). ‘‘The genus Rhagoletis Loew South of the United States (Diptera: Tephritidae),’’ U.S. Department of Agriculture Technical Bulletin 1607. Foote, R. H., Blanc, F. L., and Norrbom, A. L. (1993). ‘‘Handbook of the Fruit Flies (Diptera: Tephritidae) of America North of Mexico,’’ Comstock, Ithaca, NY. Frias, D. (1992). Genetica, ecologia y evolucion de las especies Chilenas del genero Rhagoletis (Diptera: Tephritidae). Acta Entomol. Chilena 17: 211–223. Futuyma, D. J., and Mayer, G. C. (1980). Non-allopatric speciation in animals. Syst. Zool. 29: 254–271. Han, H.-Y. (1992). ‘‘Classification of the Tribe Trypetini (Diptera: Tephritidae: Trypetinae),’’ Ph.D. thesis, Pennsylvania State University. Han, H.-Y., and McPheron, B. A. (1994). Phylogenetic study of selected tephritid flies (Insecta: Diptera: Tephritidae) using partial sequences of the nuclear 18S ribosomal DNA. Biochem. Syst. Ecol. 22: 447–457. Han, H.-Y., and McPheron, B. A. (1997). Molecular phylogenetic study of Tephritidae (Insecta: Diptera) using partial sequences of the mitochondrial 16S ribosomal DNA. Mol. Phylogenet. Evol. 7: 17–32. Herna´ndez-Ortiz, V. (1985). Descripcion de una nueva especie Mexicana del genero Rhagoletis Loew (Diptera: Tephritidae). Folia Entomol. Mex. 64: 73–79. Higgins, D. G., and Sharp, P. M. (1989). Fast and sensitive multiple sequence alignments on a microcomputer. Comput. Appl. Biosci. 5: 151–153. Kimura, M. (1980). A simple method for estimating evolutionary rate of base substitutions through comparative studies of nucleotide sequences. J. Mol. Evol. 16: 111–120. Kumar, S., Tamura, K., and Nei, M. (1993). ‘‘MEGA: Molecular Evolutionary Genetics Analysis, Version 1.0,’’ Institute for Molecular Evolutionary Genetics, Penn State University, University Park. Maddison, W. P., and Maddison, D. R. (1992). ‘‘MacClade, Version 3.04,’’ Distributed by Sinauer Associates, Sunderland, MA.

McPheron, B. A. (1990). Genetic structure of apple maggot fly (Diptera: Tephritidae) populations. Ann. Entomol. Soc. Am. 83: 568–577. McPheron, B. A., Smith, D. C., and Berlocher, S. H. (1988). Genetic differences between host races of Rhagoletis pomonella. Nature 336: 64–66. Nei, M. (1991). Relative efficiencies of different tree-making methods for molecular data. In ‘‘Phylogenetic Analysis of DNA Sequences’’ (M. M. Miyamoto, and J. Cracraft, Eds.), pp. 90–128, Oxford Univ. Press, New York. Norrbom, A. L. (1989). The status of Urophora acuticornis and U. sabroskyi (Diptera: Tephritidae). Entomol. News 100: 59–66. Norrbom, A. L. (1994). New species and phylogenetic analysis of Cryptodacus, Haywardina, and Rhagoletotrypeta (Diptera: Tephritidae). Insecta Mundi 8: 37–65. Norrbom, A. L., Ming, Y., and Hernandez-Ortiz, V. (1988). A revision of the genus Oedicarena Loew (Diptera: Tephritidae). Folia Entomol. Mex. 75: 93–117. Rzhetsky, A., and Nei, M. (1992). A simple method for estimating and testing minimum-evolution trees. Mol. Biol. Evol. 9: 945–967. Sheppard, W. S., Steck, G. J., and McPheron, B. A. (1992). Geographic populations of the medfly may be differentiated by mitochondrial DNA variation. Experientia 49: 1010–1013. Simon, C., Frati, F., Beckenbach, A., Crespi, B., Liu, H., and Flook, P. (1994). Evolution, weighting, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann. Entomol. Soc. Am. 87: 651–701. Smith, D. C., Lyons, S. A., and Berlocher, S. H. (1993). Production and electrophoretic verification of F1 hybrids between the sibling species Rhagoletis pomonella and R. cornivora. Entomol. Exp. Appl. 69: 209–213. Smith, D. C., and Prokopy, R. J. (1980). Mating behavior of Rhagoletis pomonella (Diptera: Tephritidae). VI. Site of early-season encounters. Can. Entomol. 112: 585–590. Smith, J. J., and Bush, G. L. (1997). Phylogeny of the genus Rhagoletis (Diptera: Tephritidae) inferred from DNA sequences of mitochondrial cytochrome oxidase II. Mol. Phylogenet. Evol. 7, in press. Swofford, D. L. (1993). ‘‘PAUP: Phylogenetic Analysis Using Parsimony, Version 3.1,’’ Smithsonian Institution, Washington, DC. Swofford, D. L., and Berlocher, S. H. (1987). Inferring evolutionary trees from gene frequency data under the principle of maximum parsimony. Syst. Zool. 36: 293–325. White, I. M., and Elson-Harris, M. M. (1992). ‘‘Fruit Flies of Economic Significance: Their Identification and Bionomics,’’ CAB International, Wallingford, UK. Xiong, B., and Kocher, T. D. (1991). Comparison of mitochondrial DNA sequences of seven morphospecies of black flies (Diptera: Simuliidae). Genome 34: 306–311.