• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Oct 11, 2005; 102(41): 14706–14711.
Published online Oct 3, 2005. doi:  10.1073/pnas.0502054102
PMCID: PMC1253542
Evolution

Consequences of reproductive barriers for genealogical discordance in the European corn borer

Abstract

Speciation involves the origin of trait differences that limit or prevent gene exchange and ultimately results in daughter populations that form monophyletic or exclusive genetic groups. However, for recently diverged populations or species between which reproductive isolation is often incomplete, gene genealogies will be discordant, and most regions of the genome will display nonexclusive genealogical patterns. In these situations, genome regions for which one or both species are exclusive groups may mark the footprint of recent selective sweeps. Alternatively, such regions may include or be closely linked to “speciation genes,” genes involved in reproductive isolation. Therefore, comparisons of gene genealogies allow inferences about the genetic architectures of both reproductive isolation and adaptation. Contrasting genealogical relationships in sexually isolated pheromone strains of the European corn borer moth (Ostrinia nubilalis) demonstrate the relevance of this approach. Genealogies for five gene regions are discordant, and only one molecular marker, the sex-linked gene Tpi, has evidence for pheromone strain exclusivity. Tpi maps to a position on the sex chromosome that is indistinguishable from a major factor (Pdd) affecting differences in postdiapause development time. The major factor (Resp) determining male behavioral response to pheromone is also sex-linked, but maps 20-30 cM away. Exclusivity at Tpi may be a consequence of these linkage relationships because evidence from phenotypic variation in natural populations implicates both Pdd and Resp as candidates for genes involved in recent sweeps and/or reproductive isolation between strains.

Keywords: genealogy, genetic linkage map, introgression, selective sweep, speciation

Describing and interpreting historical patterns of descent and diversification are the principal goals of evolutionary biology. These patterns are commonly inferred from DNA sequence data by using genealogies or phylogenies based on single gene regions. However, the reconstructed relationships between recently diverged lineages often fail to reveal species as monophyletic or exclusive groups (groups in which all members are more closely related to each other than to individuals outside the group). For an extended period after speciation, patterns of descent inferred from DNA sequence data will vary across the genome by chance, and shared ancestral polymorphism will be common (1-3). Consequently, multiple gene genealogies for the same set of closely related species often produce discordant trees (4-9). However, genetic variation linked to advantageous alleles within daughter species can “hitchhike” to fixation along with the selected allele, resulting in the purging of ancestral polymorphism and a more rapid approach to exclusivity or monophyly (10). The size of the chromosomal region affected by these “selective sweeps” is determined by the strength of selection and the rate of recombination.

Once a pattern of exclusivity becomes established in geographically isolated populations it will persist, but hybridization and gene flow may cause the pattern to erode when populations come into secondary contact. At “speciation genes,” loci for which trait differences result in hybrid unfitness or positive assortative mating, gene exchange is impeded or prevented between races or species, effectively maintaining exclusive patterns (11-13). Elsewhere in the genome introgression may be extensive, leading to shared alleles and an erosion of exclusive relationships. The result is a semipermeable species boundary, with permeability depending on the genetic marker (11-13). Thus, the genetic architectures of both reproductive isolation and selective sweeps have important consequences for genealogical patterns. Specifically, markers closely linked to speciation genes or to loci that have experienced a recent selective sweep are more likely to exhibit an exclusive relationship between pairs of recently diverged species (or subspecies, races, and strains).

In this study, we focus on reproductively isolated strains of the European corn borer moth (ECB), Ostrinia nubilalis. ECBs are native to Europe, North Africa, and parts of western Asia but were introduced into North America on broom corn from Italy and Hungary early in the 20th century (14). In both Europe and North America, the ECB consists of two behaviorally isolated strains that differ in the sex pheromone produced by females and the response elicited in males. In the Z strain, females produce and males respond to a 3:97 mixture of (E)- and (Z)-11-tetradecenyl acetate, whereas in the E strain, females produce and males respond to a 99:1 (E)/(Z) blend (15). Sexual isolation stems from a stereotypic male response to these alternative pheromone blends (16, 17). Although there are clearly two distinct sexual communication systems, with cross attraction rarely occurring in the laboratory, hybrid genotypes can be found in nature at North American sites (18, 19). The genetic factors responsible for major differences in female pheromone blend production, Pher, and male behavioral response to those blends, Resp, exhibit simple Mendelian inheritance (16, 17, 20, 21). Pher is autosomal, whereas Resp is sex-linked.

ECB populations are also characterized by variation in numbers of generations per year (voltinism). In New York state, populations are either bivoltine, with an early first generation and a late second generation, or univoltine, with a single generation in the middle of the season (22). Adults from the univoltine population are temporally isolated from those of the bivoltine population (23). Voltinism patterns reflect differences in postdiapause development (PDD) time, the time to pupation for over-wintering larvae under temperature and photoperiod conditions conducive to breaking diapause. PDD is controlled by a major factor on the sex chromosome (24). Based on voltinism and sex pheromone blend, New York ECB populations consist of three distinct races: univoltine Z (UZ), bivoltine E (BE), and bivoltine Z (BZ) (18, 19).

Although diverged with respect to sexual communication and life history, strains or races of ECB are otherwise difficult to distinguish (25, 26). Genetic surveys of allozymes (27, 28), mitochondrial DNA (29), random amplified polymorphic DNAs (RAPDs) (30), and a nuclear gene (31) have revealed extensive shared genetic variation, but little differentiation. However, significant allele-frequency differences occur at a sex-linked gene that encodes the enzyme triose-phosphate isomerase (TPI). At this locus, BE populations in New York are fixed for the Tpi-1 allele, and UZ and BZ populations are segregating for both Tpi-1 and Tpi-2, with Tpi-2 being the more common allele (19). Allele-frequency differences between pheromone strains for the sex-linked Tpi locus might be maintained by tight linkage to the gene for male response, but, by using interstrain crosses, Dopman et al. (21) mapped Resp >20 cM away from Tpi on the sex chromosome and argued that the origin of patterns of variation at Tpi cannot be explained by linkage to this factor.

Here, we use the ECB genetic linkage map (21) as a frame-work for a comparative genealogical analysis of sequence data that includes Tpi and four additional molecular markers. We infer genealogical relationships for these five markers (three sex-linked, one autosomal, and mtDNA) for each strain and an outgroup species. We then show that genealogies are discordant and that only Tpi reveals evidence for exclusivity between Z and E moths. Finally, we address the relationship between genealogical patterns and the genetic architecture of reproductive barriers by mapping our markers relative to the positions of Resp and Pdd.

Materials and Methods

Insect Populations. Insect mapping families. Two cultures of ECB, maintained by Wendell Roelofs and colleagues at the New York State Agricultural Experiment Station (NYSAES) in Geneva, NY, were used as sources for initiating mapping families. The first culture consisted of UZ-strain insects that were derived from field-collected larvae, pupae, and adults found in corn stubble from Bouckville, NY, in April 1994. The second culture consisted of BE-strain insects derived from corn stubble near Geneva, NY, in May 1996. Details of mass rearing of insects can be found in ref. 22.

Field-collected insects. Both pheromone strains occur in New York, and sympatric populations have been documented at a number of localities (19, 22). The Geneva population has been monitored for ECBs and consists of BE and UZ moths (e.g., ref. 22). Thus, female larvae and pupae collected from corn stubble in May of 2000 and 2004 that were reared under diapause-breaking conditions to assess PDD time could be scored as BE or UZ. Fourteen BE and six UZ females were used for genetic analysis from this locality. Eight UZ females were collected for analysis from a UZ population in Madison, NY, in May 2000, and four BZ females were collected from a BZ population in Eden, NY, in October 2000.

Both pheromone strains also occur in North Carolina, with Z borers found in the western part of the state and E borers more common in the east. A zone of overlap exists where males are found in pheromone traps baited with either the Z or E blends (32). For genealogical analysis, we used three female ECBs collected in May 2002 from potato plants in Weeksville, NC, (E strain) and two females collected from corn in Fletcher, NC, (Z strain). An Asian corn borer (ACB) female (Ostrinia furnicalis) was used as an outgroup species.

Marker Development. Amplified fragment length polymorphism (AFLP) and microsatellite markers. Development of AFLP and microsatellite markers is described in ref. 21.

Triose-phosphate isomerase (Tpi). A Tpi fragment was initially amplified with degenerate primers designed by using Tpi sequences from Heliothis virescens (GenBank accession no. U23080), Spodoptera littoralis (GenBank accession no. L39011), Drosophila melanogaster (GenBank accession no. AE003772), and Anopheles merus (GenBank accession no. U82707). ECB-specific primers were then designed and used for PCR amplification of most of the Tpi sequence, with the exception of the furthest 3′ exon. The 3′ end of the gene was obtained by PCR amplification from cDNA. mRNA was isolated by using Oligotex extraction (Qiagen, Valencia, CA), reverse transcribed, and PCR amplified with ECB and poly(T)-linker primers. Thermal cycling used the profile: 94°C for 45 sec, 50°C for 45 sec, and 72°C for 1.3 min. The ≈1.6-kb genomic fragment that includes the entire coding region plus introns was amplified by using primers ECBtpi_for1A (5′-AGATGTCAAAATTCAACTCAG) and ECBtpi_rev5 (5′-AGCACCCTTCGGCACTT), and sequenced by using this primer pair and two internal primers, ECBtpi_for5A (5′-AGGCAGACCAAGGCACTCTTGCC) and ECBtpi_rev3A (5′-TTCGGTACCGATGGCCCATACAG). Tpi was placed on the Z linkage map as described in Dopman et al. (21). We assessed electrophoretic mobility for TPI in a subset of field-collected insects to compare mobility class with Tpi sequence in an effort to identify the amino acid substitution responsible for electrophoretic mobility variation. Because female Lepidoptera are heterogametic, female ECBs possess a single copy of either the Tpi-1 or the Tpi-2 allele.

Kettin (Ket). Degenerate primers provided by P. Andolfatto (University of California at San Diego) were used for initial Ket amplification, and ECB-specific primers amplified an ≈1.3-kb fragment (ECBketF, 5′-TGAAATCCCGGAACCAGTAACA; ECBketR, 5′-TTGAGGTGAGTAGTGAAAATAGGAG) under the amplification profile of 94°C for 45 sec, 53.5°C for 45 sec, and 72°C for 1.3 min. Sequencing primers were ECBket308F: 5′-CTAGGTGAAGCAGTAACGACAGC and ECBket348R: 5′-ATCCAAAGTAACGAATCCGAAATC. Ket was mapped either by a diagnostic restriction site (Rsa I) or by amplification using a primer that extends over a polymorphic insertion/deletion (indel) (ECBket_RSA3F: 5′-TATGAATCAGTTACCTACATAACTAGGTAC). Restriction enzyme digestion followed supplier's protocols (NEB, Beverly, MA), and PCR followed previous Ket cycling conditions with a change in annealing temperature to 50.7°C.

Lactate dehydrogenase (Ldh). A blastn search of the Bombyx mori EST database (www.ab.a.u-tokyo.ac.jp/silkbase/) with a Papilio glaucus Ldh sequence (P. Andolfatto, personal communication) yielded a significant alignment (e-52; EST sequence number wdS30992). Primers were designed from a B. mori sequence that was conserved with respect to the P. glaucus sequence (Ldhbm_65F, 5′-ATCGCCAGTAACCCCGTGG; Ldhbm_376R, 5′-CGATAGCCCAGGA AGTGTATCCCTTC), and ECB-specific internal primers were developed (ECB_ldhF, 5′-GGCTCCGGCACCA ACCTGGACTC; ECB_ldhR, 5′-CGTAGGCGCTCTTCACCACCATCTCA). Other than a 55°C annealing temperature for the B. mori-derived primers and a 58°C annealing temperature for the ECB-derived primers, cycling conditions for Ldh were identical to those used for Ket. PCR products were ≈600 and ≈500 bp, respectively, for the two primer pairs. Like Ket, Ldh was mapped by using either a diagnostic restriction site (Alu I) or by amplification using a primer that extends over a polymorphic indel (ECBldh_E84R, 5′-GAATATCAGAACAAACAAAGGTC). Digestion followed the supplier's protocols (NEB), and PCR followed previous cycling conditions with a change in annealing temperature to 64.2°C.

Pheromone binding protein (Pbp). A ≈1.6 kb fragment of DNA that included Pbp was amplified by using primers ECEP5 and ECPA as described by Willett and Harrison (31). Amplified fragments were cloned by using the TOPO TA cloning kit (Invitrogen), and sequence data were obtained for the 5′ end of the gene (31). ACB sequence was AF133630 from GenBank.

Cytochrome oxidase I (COI). An ≈1.2-kb fragment of COI was amplified by using primers Ron (5′-GGATCACCTGATATAGCAT TCCC) and Pat (5′-TCCA ATGCACTA ATCTGCCATATTA) at 50°C annealing, but otherwise, thermal cycling was the same as above.

PCR and sequencing. All PCRs and sequencing with a PRISM 377 (Applied Biosystems) followed protocols described in ref. 21. Sequence data were manipulated with the dnastar programs (DNASTAR, Madison, WI) with default parameters.

Genealogical Analysis. Tests of exclusivity between E and Z moths were conducted by comparing maximum likelihood (ML) trees with those obtained from heuristic searches with the constraint trees: ACB, ((E), (Z)) or ((E), (Z)). When ACB was not specified in the constraint, it could be attached on any branch and the resulting tree would still be compatible with the constraint. Exclusivity could therefore be defined in terms of the monophyly of E and paraphyly of Z, monophyly of Z and paraphyly of E, or reciprocal monophyly. Shimodaira-Hasegawa tests (33) in paup* (version 4.0b10) (34) were used to compare the ML tree with the ML tree compatible with the constraint by using 10,000 resampling estimated log-likelihood (RELL) bootstrap replicates. Optimal substitution models under ML for each data set were identified by using modeltest (version 3.6) (35), but a full-parameterized general time reversible (GTR) + Γ + I model was used for Shimodaira-Hasegawa tests to maximize the ML scores and to minimize biases due to differences in the number of free parameters between trees.

Other genealogical analyses were conducted in paup* under the maximum parsimony (MP) optimality criterion. All MP analyses used heuristic searches that assumed unordered, equal-weight characters with gaps treated as a “fifth base.” Multiple-base indels were down-weighted by reducing to single-base indels under the assumption that such characters represent single evolutionary changes. Starting trees were obtained by using stepwise addition; the addition sequence was random by using 10 replicates; the branch-swapping algorithm was tree-bisection-reconnection (TBR); and multiple trees (multrees) were in effect. Bootstrap values were obtained by using a heuristic search with stepwise addition, random-addition sequence, TBR branch-swapping, and 1,000 replicates. Finally, the congruence of characters from each marker was assessed by performing a partition-homogeneity test with 1,000 replicates.

Z Chromosome Genetic Map. Details of crosses, linkage map construction, and phenotype assessment of male behavioral response can be found in ref. 21 and references therein. The reported Z linkage map includes data from 116 BC1 progeny (78 males and 38 females), all of which were genotyped for Ket (this study), Ldh (this study), and Tpi (21). Microsatellite marker ma169 was genotyped for all male progeny, which were also phenotyped for male behavioral response. Finally, 79 progeny (41 males and 38 females) were genotyped for 8 AFLP markers.

PDD has been shown through its association with TPI electrophoretic mobility to be linked to the Tpi gene and therefore must map to the Z chromosome (24). A BC1 family was generated to map PDD differences on the Z chromosome by using a three-point test cross with Tpi and Ldh as markers. After hatching BC1 progeny, a 12:12 light:dark photoperiod was used to induce maximum diapause (24). After 35 days, borers in diapause were reared under diapause-breaking conditions of 16:8 light:dark and the time to pupation was noted every 2 days. Under these conditions, the average PDD for parental stocks is 14.50 (SE ± 0.55) days for bivoltine females and 43.94 (SE ± 1.64) days for univoltine females (24).

Results

Genealogical Analysis. Aligned sequence data for COI, Ket, Ldh, and Tpi consisted of 17 E- and 20 Z-strain ECBs, plus the ACB outgroup (Table 1). Only 12 of the 17 E-strain insects were sequenced for Pbp. After reduction of multiple-base indels, the aligned data sets ranged in size from 1,449 characters for Tpi to 291 characters for Ldh (Ldh had 167 intronic characters removed that were difficult to align) (Table 1). Although the length of the Ldh data set was smaller than the lengths for the other markers, it contained the second largest proportion of variable to total characters (Tpi, 55/1,449; Ket, 37/690; Ldh, 31/291; Pbp, 59/477; and COI, 35/1,195). Loci showed a wide range of ingroup nucleotide diversity and average sequence divergence to the outgroup (Table 1). The pairwise-sequence divergence within ECB ranged from 0.1% for COI to 2.6% for Ldh. Average sequence divergence to the outgroup ranged from 1% to >3.5%.

Table 1.
Summary information for sampled markers after reducing multiple-base indels

Characters from the five markers were significantly incongruent, as indicated by the partition-homogeneity test (P = 0.001). Nonparametric bootstrap support for clades varied by locus, with COI showing little overall clade support and the other loci each containing at least two nodes with ≥70% support (Fig. 1). Although strict reciprocal monophyly between E and Z moths was not present in any gene tree, the Tpi genealogy came close. Only two clades had bootstrap support ≥70% in Tpi; one contained 16 of the 17 E-strain borers (99% support) and the other contained 19 of the 20 Z-strain borers (74% support). Additionally, there were no shared haplotypes between E and Z moths at Tpi (Table 1). Each of the other markers had at least one group that contained multiple Z and E moths with identical haplotypes.

Fig. 1.
Genealogies and mapping positions for molecular markers and traits divergent between ECB pheromone strains. A genetic linkage map of the Z (sex) chromosome shows the mapping positions for Ket, Tpi, Ldh, AFLPs (A1-A8), microsatellite ma169, male behavioral ...

For each locus, we tested the hypothesis that Z and E strains of ECB form exclusive genealogical groups (Fig. 1). We compared the likelihood of the ML unconstrained tree with the likelihood of the tree that enforced a pattern of exclusivity for the Z and E strains. The exclusivity hypothesis, ((E), (Z)), was rejected for COI, Ket, Ldh, and Pbp, but not for Tpi (Shimodaira-Hasegawa test, COI, P = 0.019; Ket, P < 0.001; Ldh, P < 0.001; Pbp, P < 0.001; and Tpi, P = 0.399) (Fig. 1). The ML tree for Tpi also did not differ significantly from one in which a reciprocal monophyletic relationship exists between Z and E pheromone strains [i.e., (ACB, (E), (Z)); P = 0.399].

The choice of phylogenetic analysis had no significant effect on tree topology. For each locus, the ML tree did not significantly differ from the MP tree(s) (gaps included, Wilcoxon sign-ranks test, COI, P = 1.0; Ket, P ≥ 0.18; Ldh, P = 1.0; Pbp, P ≥ 0.21; and Tpi, P ≥ 0.26). Gene genealogies shown in Fig. 1 represent the consensus of reconstructions for the MP analysis.

Character-state optimization on an MP tree for moths characterized electrophoretically for TPI located a single character-state change (a nonsynonymous substitution) on the branch separating females that are hemizygous for the Tpi-1 allele from females that have the Tpi-2 allele. Moths that carry the Tpi-1 allele have an asparagine at amino acid residue 189, whereas moths that carry the Tpi-2 allele have a lysine in this position. The substitution of a lysine for an asparagine, which accounts for the slower anodal migration of Tpi-2, is the only character-state change that maps to the branch separating females with different TPI electromorphs (Fig. 1). In both ACB and the moth H. virescens, Tpi sequences have an asparagine at residue 189, suggesting that the Tpi-2 allele may be recently derived within the Z-strain ECBs.

Z Chromosome Genetic Map. The expanded Z chromosome map has a total length of 76 cM, and is bounded by Ket on one end and by an AFLP marker on the other end (Fig. 1). Tpi and four AFLP markers map 29 cM from Ket; Ldh is 6 cM further along the sex chromosome; and Resp maps 20 cM beyond Ldh. In the three-point test cross used to map the major factor for PDD, hemizygous females from the BC1 mapping family showed a clear bimodal distribution for female PDD time (Fig. 2). A 12-day period starting on day 25 and ending on day 36 divides females into bivoltine (fast PDD) and univoltine (slow PDD) groups that correspond with parental PDD. Of 41 fast-developing bivoltine females, 37 expressed the Tpi-1/Ldh-1 genotype and 4 expressed the Tpi-1/Ldh-2 genotype. Of 30 slow-developing univoltine females, 27 expressed the Tpi-2/Ldh-2 genotype and 3 expressed the Tpi-2/Ldh-1 genotype. Thus, we observed no recombinants between PDD time and Tpi, and 7 recombinants between Tpi and Ldh (which did not differ from ref. 21; χ2 = 0.823; P = 0.364). The factor for PDD (Pdd) therefore maps to the same position as Tpi (Fig. 1).

Fig. 2.
Histogram showing bimodal PDD time for female BC1 offspring genotyped for Tpi and Ldh. The 12-day period starting on day 25 and ending on day 36 divides females into bivoltine (fast PDD, n = 41) and univoltine (slow PDD, n = 30) groups that correspond ...

Discussion

Although speciation ultimately results in divergence across the entire genome, random lineage sorting and introgression will cause shared polymorphism to persist among recently diverged populations and closely related species. Indeed, multilocus studies of closely related species frequently report discordant genealogical patterns that fail to support species boundaries based on morphological, behavioral, and ecological characters (7-9). In the ECB, gene genealogies based on COI, Ket, Ldh, Pbp, and Tpi exhibit significant discordance (Fig. 1), and exclusivity of Z and E pheromone strains is rejected for four of five markers. Only Tpi supports the existence of two distinct haplotype groups composed of Z- and E-strain moths.

Exclusivity at Tpi and extensive haplotype sharing at the other four markers results from some combination of random lineage extinction, selection, or differential introgression. Because mtDNA has an effective copy number one-fourth that of autosomes, monophyly because of random lineage extinction at mitochondrial genes would be expected with high probability well before monophyly at even a small number of diploid loci (3). That mtDNA does not show any evidence of exclusivity between strains argues against random lineage extinction as the only explanation for pattern differences between Tpi and the other markers.

Patterns of discordance in the genealogical data are best interpreted within the context of the ECB linkage map. Combining genealogical and mapping data reveals (i) that disequilibrium persists between Tpi and Resp in sympatric localities (19) despite a >20-cM distance separating these two loci on the Z chromosome (Fig. 1); (ii) that Ldh maps closer to Resp than Tpi and yet, unlike Tpi, Ldh shows a high degree of haplotype sharing between pheromone strains; and (iii) that the mapping positions of Tpi and Pdd are indistinguishable. That is, the only locus that shows evidence for exclusivity maps close to a factor that affects differences in development time and contributes to reproductive isolation.

Given our knowledge of genomic location, the observed pattern of locus-specific exclusivity in the ECB can be explained by two alternative scenarios. One explanation is that limited introgression (and persistent disequilibrium) characterizes much of the sex chromosome, but disequilibrium and monophyly have become discernible only in those regions that have experienced a recent selective sweep (e.g., the Pdd/Tpi region). For the ECB, sharing of haplotypes (e.g., at Ldh) and absence of discernible disequilibrium where Z and E moths are sympatric would then reflect persistent ancestral polymorphism and not contemporary gene exchange. In this scenario, it remains possible that Resp does play an important role in slowing the decay of linkage disequilibrium between Tpi and Resp by limiting or preventing introgression between the Z and E strains.

A second explanation for locus-specific exclusivity is that contemporary gene exchange (rather than shared ancestral polymorphism) between Z and E strains is responsible for shared haplotypes at Ket, Ldh, Pbp, and COI. A process of differential introgression would be expected to generate a positive association between the genomic location of genes contributing to isolation and markers showing genetic divergence, a pattern that has been found among closely related Drosophila species (9, 36). Under this scenario, Tpi marks the location of a speciation gene that renders the surrounding chromosomal region immune to the homogenizing effects of introgression. F1 hybrids between Z and E moths do exist in nature (18, 19), but the reproductive fate of these moths remains unknown. If F1 hybrids backcross, the factor linked to Tpi would have to be working in concert with Resp through epistatic selection for disequilibrium between Tpi and Resp to persist across 26 cM of map distance.

The observation that the only marker that reveals strains of the ECB to be monophyletic (Tpi) also maps to the same position as a major factor for a divergent phenotype (Pdd) strongly suggests a role for natural selection. This observation lends support to a model of speciation in which species become differentiated at some gene regions through natural selection, whereas, at other loci, shared variation persists because of recent or historical gene flow. The Drosophila pseudoobscura species group has also been cited as an example of the importance of selection in speciation because reciprocal monophyly occurs only at markers linked to regions containing hybrid sterility genes (9, 36). In the ECB, if natural selection has had a role in shaping patterns of genetic variation at Tpi, either through a recent selective sweep or by eliminating incompatibilities in hybrid offspring, then Pdd, which shows tight linkage with Tpi, provides a candidate target of selection. Confirming that both selection and demography have shaped patterns of genetic variation among markers represents the next step in understanding the genetics of speciation in the ECB.

The closely related species, subspecies, races, or strains that form the foundation for speciation research often exhibit discordant phylogenies or gene genealogies for different molecular markers. The origin of monophyly, often associated with selective sweeps in these groups, may provide the diagnostic markers that allow us to identify the genomic location of reproductive barriers or adaptations. Genealogies from these and other gene regions help to clarify patterns of descent and diversification and, when combined with a genetic map containing the locations of barriers and sweeps, reveal a more complete picture of the speciation process.

Acknowledgments

We thank Wendell Roelofs, Paul Robbins, Charles Linn, Kathy Poole, Jim Walgenbach, and George Kennedy for assistance and advice with ECB collecting and rearing. We thank Peter Andolfatto for providing information on Ket and Ldh, and Shankar Iyer for sequencing Pbp. Two anonymous reviewers provided constructive criticism of this paper. This work was supported by U.S. Environmental Protection Agency “Science to Achieve Results” fellowship for graduate environmental study U-94589501-0 (to E.B.D.), National Research Initiative of the Cooperative State Research Education and Extension Service Grant 2001-35302-11123 (to R.G.H.), and National Science Foundation Grant DEB-0415343 (to R.G.H.).

Notes

This paper was submitted directly (Track II) to the PNAS office.

Abbreviations: ECB, European corn borer moth; ACB, Asian corn borer moth; AFLP, amplified fragment length polymorphism; PDD, postdiapause development; TPI, triose-phosphate isomerase; Pbp, pheromone binding protein; ML, maximum likelihood; MP, maximum parsimony; indels, insertions/deletions.

Data deposition: The sequences reported in this paper have been deposited in the GenBank database (accession nos. DQ204878-DQ205062).

References

1. Tajima, F. (1983) Genetics 105, 437-460. [PMC free article] [PubMed]
2. Neigel, J. E. & Avise, J. C. (1986) in Evolutionary Processes and Theory, eds. Karlin, S. & Nevo, E. (Academic, New York), pp. 515-534.
3. Hudson, R. R. & Coyne, J. A. (2002) Evolution 56, 1557-1565. [PubMed]
4. Wang, R. L., Wakeley, J. & Hey, J. (1997) Genetics 147, 1091-1106. [PMC free article] [PubMed]
5. Ting, C. T., Tsaur, S. C. & Wu, C. I. (2000) Proc. Natl. Acad. Sci. USA 97, 5313-5316. [PMC free article] [PubMed]
6. Kliman, R. M., Andolfatto, P., Coyne, J. A., Depaulis, F., Kreitman, M., Berry, A. J., McCarter, J., Wakeley, J. & Hey, J. (2000) Genetics 156, 1913-1931. [PMC free article] [PubMed]
7. Beltran, M., Jiggins, C. D., Bull, V., Linares, M., Mallet, J., McMillan, W. O. & Bermingham, E. (2002) Mol. Biol. Evol. 19, 2176-2190. [PubMed]
8. Broughton, R. E. & Harrison, R. G. (2003) Genetics 163, 1389-1401. [PMC free article] [PubMed]
9. Machado, C. A. & Hey, J. (2003) Proc. R. Soc. London Ser. B 270, 1193-1202. [PMC free article] [PubMed]
10. Maynard Smith, J. & Haigh, J. (1974) Genet. Res. 23, 23-35. [PubMed]
11. Barton, N. H. & Hewitt, G. M. (1981) in Evolution and Speciation, eds. Atchley, W. R. & Woodruff, D. S. (Cambridge Univ. Press, Cambridge, U.K.), pp. 109-145.
12. Harrison, R. G. (1990) in Oxford Surveys in Evolutionary Biology, eds. Futuyma, D. & Antonovics, J. (Oxford Univ. Press, Oxford), Vol. 7, pp. 69-128.
13. Wu, C. (2001) J. Evol. Biol. 14, 851-865.
14. Caffrey, D. J. & Worthley, L. H. (1927) A Progress Report on the Investigations of the European Corn Borer (U.S. Department of Agriculture, Washington, DC), USDA Bull. 1476.
15. Klun, J. A., Chapman, O. L., Mattes, J. C., Wojtowski, P. W., Beroza, M. & Sonnet, P. E. (1973) Science 181, 661-663. [PubMed]
16. Roelofs, W., Glover, T., Tang, X. H., Sreng, I., Robbins, P., Eckenrode, C., Lofstedt, C., Hansson, B. S. & Bengtsson, B. O. (1987) Proc. Natl. Acad. Sci. USA 84, 7585-7589. [PMC free article] [PubMed]
17. Glover, T., Campbell, M., Robbins, P. & Roelofs, W. (1990) Arch. Insect Biochem. Physiol. 15, 67-77.
18. Klun, J. A. & Huettel, M. D. (1988) J. Chem. Ecol. 14, 2047-2061. [PubMed]
19. Glover, T. J., Knodel, J. J., Robbins, P. S., Eckenrode, C. J. & Roelofs, W. L. (1991) Environ. Entomol. 20, 1356-1362.
20. Klun, J. A. & Maini, S. (1979) Environ. Entomol. 8, 423-426.
21. Dopman, E. B., Bogdanowicz, S. M. & Harrison, R. G. (2004) Genetics 167, 301-309. [PMC free article] [PubMed]
22. Roelofs, W. L., Du, J.-W., Tang, X.-H., Robbins, P. & Eckenrode, C. J. (1985) J. Chem. Ecol. 11, 829-936. [PubMed]
23. Eckenrode, C. J., Robbins, P. S. & Andaloro, J. T. (1983) Environ. Entomol. 12, 393-396.
24. Glover, T. J., Robbins, P., Eckenrode, C. J. & Roelofs, W. L. (1992) Arch. Insect Biochem. Physiol. 20, 107-117.
25. Liebherr, J. (1974) M.S. thesis (Cornell Univ., Ithaca, NY).
26. Liebherr, J. & Roelofs, W. L. (1975) Ann. Entomol. Soc. Am. 68, 305-309.
27. Harrison, R. G. & Vawter, A. T. (1977) Ann. Entomol. Soc. Am. 70, 717-720.
28. Cianchi, R., Maini, S. & Bullini, L. (1980) Heredity 45, 383-388.
29. Marcon, P., Taylor, D. B., Mason, C. E., Hellmich, R. L. & Siegfried, B. D. (1999) Insect Mol. Biol. 8, 213-221. [PubMed]
30. Pornkulwat, S., Skoda, S. R., Thomas, G. D. & Foster, J. E. (1998) Ann. Entomol. Soc. Am. 91, 719-725.
31. Willett, C. S. & Harrison, R. G. (1999) Genetics 153, 1743-1751. [PMC free article] [PubMed]
32. Sorenson, C. E., Kennedy, G. C., Van Duyn, W., Bradley, J. R., Jr., & Walgenbach, J. F. (1992) Entomol. Exp. Appl. 64, 177-185.
33. Shimodaira, H. & Hasegawa, M. (1999) Mol. Biol. Evol. 16, 1114-1116.
34. Swofford, D. L. (2003) paup*, Phylogenetic Analysis Using Parsimony (* and Other Methods) (Sinauer, Sunderland, MA), Version 4.
35. Posada, D. & Crandall, D. A. (1998) Bioinformatics 14, 817-818. [PubMed]
36. Noor, M. A., Grams, K. L., Bertucci, L. A., Almendarez, Y., Reiland, J. & Smith, K. R. (2001) Evolution 55, 512-521. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...