Radiation and divergence in the Rhagoletis Pomonella species complex: inferences from DNA sequence data

Radiation and divergence in the Rhagoletis Pomonella speciescomplex: inferences from DNA sequence data

X. XIE,* A. P. MICHEL,* D. SCHWARZ,� J. RULL,� S. VELEZ,* A. A. FORBES,* M. ALUJA�& J. L. FEDER*

*Department of Biological Sciences, University of Notre Dame, Notre Dame, IN, USA

�Department of Entomology, Pennsylvania State University, University Park, PA, USA

�Instituto de Ecologıa, Asociacion Civil, Xalapa, Veracruz, Mexico

Introduction

Understanding what a species is and the processes

responsible for their formation are two central issues in

evolutionary biology. The standard answer to the first

question has been the biological species concept (BSC),

which defines species on the basis of intrinsic barriers to

gene flow (Mayr, 1963). Geographic isolation (allopatry)

has been argued to be the common denominator with

respect to answering the second question of speciation

mode (Mayr, 1963).

In recent years, a number of alternative, genetically

based species concepts have been proposed to the BSC.

These concepts include defining species based on the

existence of genotypic clusters in sympatry (Mallet, 1995;

Feder, 1998), on the presence of diagnostic autapomor-

phies distinguishing populations (Cracraft, 1989) and on

reciprocal monophyly between evolutionary lineages (de

Queiroz, 1998). Although debate continues on the

merits, practicality and relationship of these genetic

concepts relative to the BSC (Coyne & Orr, 2004), the

discussion has served to highlight that the genomes of

diverging populations, particularly taxa undergoing

divergence-with-gene-flow speciation, can be mosaic.

Regions containing loci under selection contributing to

gene flow barriers (islands of speciation; Turner et al.,

2005) can display greater differentiation than regions not

involved in reproductive isolation. A single value of

reproductive isolation between taxa may therefore not

apply uniformly across the genome (Mallet, 1995; Feder,

1998; Wu, 2001), as whole genomes need not be isolated

(Mayr, 2001) or monophyletic (Hudson & Coyne, 2002)

to have good species.

Correspondence: Jeffrey L. Feder, Department of Biological Sciences, PO

Box 369, Galvin Life Science Center, University of Notre Dame, Notre

Dame, IN 46556-0369, USA.

Tel.: +1 574 631 4159; fax: +1 574 631 7413;

e-mail: [email protected]

Present addresses: A. P. Michel, Department of Entomology, Ohio

Agricultural Research and Development Center, Ohio State University,

210 Thorne Hall, 1680 Madison Ave., Wooster, OH 44691, USA.

D. Schwarz, Department of Entomology, University of Illinois at Urbana-

Champaign, Morrill Hall, 505 S. Goodwin Ave., Urbana, IL 61801, USA.

S. Velez, Museum of Comparative Zoology, Harvard University, 26

Oxford St, Cambridge, MA 02138, USA.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

900 J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

Keywords:

biogeography;

differential introgression;

genetic divergence;

host races;

inversions;

sibling species;

speciation mode plurality;

sympatric speciation.

Abstract

Here, we investigate the evolutionary history and pattern of genetic

divergence in the Rhagoletis pomonella (Diptera: Tephritidae) sibling species

complex, a model for sympatric speciation via host plant shifting, using 11

anonymous nuclear genes and mtDNA. We report that DNA sequence results

largely coincide with those of previous allozyme studies. Rhagoletis cornivora

was basal in the complex, distinguished by fixed substitutions at all loci. Gene

trees did not provide reciprocally monophyletic relationships among US

populations of R. pomonella, R. mendax, R. zephyria and the undescribed

flowering dogwood fly. However, private alleles were found for these taxa

for certain loci. We discuss the implications of the results with respect to

identifiable genetic signposts (stages) of speciation, the mosaic nature of

genomic differentiation distinguishing formative species and a concept of

speciation mode plurality involving a biogeographic contribution to sympatric

speciation in the R. pomonella complex.

doi: 10.1111/j.1420-9101.2008.01507.x

The issue of speciation process has also been conten-

tious. Recently, however, a growing consensus has

emerged that allopatry is not the exclusive mode of

speciation, even for animals; divergence-with-gene-flow

speciation can and does occur, although under more

restrictive circumstances than speciation in allopatry

(Coyne & Orr, 2004). An important accompanying

development in this discussion has been an increasing

awareness of the possibility of speciation mode ‘plural-

ity’; that categorizing speciation as strictly allopatric vs.

nonallopatric may not adequately describe the geo-

graphic context of divergence for many taxa. If, for

example, some genetic changes leading to reproductive

isolation occur in allopatry and others in sympatry, then

it would seem appropriate for the geographic mode of

speciation to be mixed (Mallet, 2005). In this regard,

Coyne & Orr (2004) have proposed several new terms

describing situations of mixed geographic mode.

The Rhagoletis pomonella (Diptera: Tephritidae) complex

provides an opportunity to investigate questions con-

cerning speciation pattern and process. The complex

contains a number of host races (e.g. apple and haw-

thorn-infesting populations of R. pomonella) and sibling

species (e.g. R. mendax, R. zephyria, R. cornivora and the

undescribed flowering dogwood fly) at varying stages of

divergence. The close morphological similarity, distinct

host plant affiliations and broadly overlapping geographic

ranges of R. pomonella complex flies led Bush (1966,

1969) to propose that its members all speciated sympat-

rically via host plant shifting in North America. The

central premise of the sympatric hypothesis is that

differential adaptation to alternative host plants gener-

ates ecological barriers to gene flow that initiate specia-

tion. For R. pomonella, each fly in the complex infests a

unique, nonoverlapping set of host plants (see supple-

mentary Table S1) which tend to fruit at different times

of the field season (Smith, 1988; Feder et al., 1993;

Berlocher, 2000; Dambroski & Feder, 2007). Because

Rhagoletis is univoltine, differences in the depth of the

over-wintering diapause differentially adapts flies to

variation in host fruiting phenology, generating allo-

chronic mating isolation between R. pomonella taxa

(Smith, 1988; Feder et al., 1993, 1994; Dambroski &

Feder, 2007). Rhagoletis flies also mate almost exclusively

on or near the fruit of their respective host plants

(Prokopy et al. 1971, 1972). Traits related to host plant

discrimination, including differences in fruit volatile

preference, therefore also directly affect mate choice

and represent another important ecological barrier to

gene flow among R. pomonella flies (Feder et al., 1994;

Linn et al., 2003, 2004).

Here, we investigate patterns of DNA sequence diver-

gence for 11 anonymous nuclear genes and mtDNA in

the R. pomonella complex. The objectives of the study are

threefold. Our first goal is to test whether any ‘genetic

signposts’ of divergence exist for R. pomonella flies. By

genetic signpost, we are referring to identifiable patterns

of differentiation, such as degree of genotypic clustering

or monophyly, in gene trees indicative of stages (the

extent of ecological and intrinsic isolation) in the speci-

ation continuum, rather than specific conditions that

have to be met to satisfy a particular species concept.

With the exception of R. cornivora, previous allozyme and

mtDNA work has failed to uncover diagnostically fixed

allele differences distinguishing any of the other taxa

(McPheron et al., 1988; Berlocher et al., 1993; Feder,

1998; Feder et al., 1999; Berlocher, 2000). Some taxa,

such as R. mendax and R. zephyria, do possess high-

frequency allozyme variants that are rare in other

populations (Berlocher et al., 1993; Feder, 1998; Feder

et al., 1999; Berlocher, 2000), but they are not diagnostic.

This has lead to speculation that certain R. pomonella taxa

may represent ‘quantitative genetic’ species; species that

display marked phenotypic differences in host-related

ecological adaptations causing substantial reproductive

isolation because of the cumulative effects of significant

allele frequency, but not fixed differences, across con-

tributing loci (Berlocher & Feder, 2002). If more sensitive

sequence analysis reveals the same pattern, then we

must begin to ask why fixed differences are absent; are

most of the sibling species too young for the lineage

sorting of neutral variation to be complete or is persis-

tent, low-level gene flow among taxa sufficient to

prevent alleles from becoming fixed?

Our second goal was to determine whether gene trees

for the nuclear loci are consistent with a hypothesis of

speciation mode plurality for R. pomonella. Earlier work

has suggested that introgression from Mexico helped

facilitate the sympatric radiation of the R. pomonella

complex in the USA (Feder et al., 2003a, 2005; Xie et al.,

2007). Sometime during a period of past geographic

separation �1.57 Ma, inversions appeared to have arisen

and fixed in an isolated hawthorn fly population located

in the Eje Volcanico Trans Mexicano (EVTM; Fig. 1) for

three different genomic regions on chromosomes 1, 2

and 3 (haploid n = 6 for R. pomonella; Bush, 1966).

Following secondary contact, gene flow from the EVTM

into the USA probably through the conduit of the Sierra

Madre Oriental Mountains (SMO in Fig. 1) created

adaptive inversion clines for diapause life-history traits

in the USA (Xie et al., 2007). This latitudinal diapause

variation in conjunction with additional changes in host

discrimination aided the US hawthorn fly population in

sympatrically shifting and adapting to a variety of new

plants with differing fruiting times and other physical

and chemical characteristics, generating R. mendax,

R. zephyria and the flowering dogwood fly in the process.

In historical time (< 150 years), R. pomonella also shifted

from hawthorn to introduced, domesticated apple (Malus

pumila) resulting in the formation of the apple fly race in

the north-eastern and mid-western USA (Bush, 1966).

Therefore, a portion of the diapause variation contribut-

ing to the radiation of the R. pomonella complex in the

USA originated at an earlier time and in a different

Radiation of Rhagoletis pomonella complex 901

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

location than the sympatric host shifts triggering speci-

ation. In this study, we increase the sampling of nuclear

loci and taxa to confirm the chronology of the isolation of

Mexican EVTM flies relative to the radiation of the

R. pomonella complex in the USA.

Our third goal is to determine whether patterns of

genetic divergence vary across the R. pomonella genome.

Previous work implied that the inverted regions of

chromosomes 1–3 display greater levels of divergence

between Mexican and US hawthorn flies than loci

mapping elsewhere on chromosomes 4 and 5 (Feder

et al., 2005; Xie et al., 2007). From these data, we inferred

that after the initial period of introgression from Mexico,

the rearranged regions of chromosomes 1–3 may have

evolved to become more impervious to gene flow than

loci residing on other chromosomes. Through increased

sampling of sibling taxa, we investigate whether a similar

pattern is evident for the other US members of the

R. pomonella complex as well.

Methods

Flies and collection sites

Rhagoletis pomonella complex flies were genetically anal-

ysed from 14 different sites ⁄ populations, 10 from the

USA and four from Mexico (Fig. 1; see supplementary

Table S1 for complete details of the sites, including host

species and collecting dates). Two of the four Mexican

hawthorn sites [Coajomulco, Morelos (Evtm CJ) and

Tancitaro, Michoacan (Evtm MC)] reside in the Eje

Volcanico Trans Mexicano plateau (green outlined area

in Fig. 1). The other two Mexican hawthorn sites [San

Joaquin, Queretaro (Smo SJ) and Piletas, Veracruz

(Smo PL)] reside in the Sierra Madre Oriental Moun-

tains, and are located along the hypothesized conduit

for introgression from Mexico into the USA (orange

area in Fig. 1). We also sequenced Rhagoletis electromor-

pha from Dowagiac, MI, USA. to serve as an outgroup

(R. electromorpha belongs to the tabellaria species com-

plex, the sister group to R. pomonella; Bush, 1966). Flies

were sampled as larvae in infested fruit at all sites and

either immediately dissected from the fruit and frozen

for later genetic analysis or reared to adulthood in the

laboratory.

Genes sequenced

Sequence data were generated for 11 anonymous

nuclear loci and a 626-bp fragment of the mitochondrial

cytochrome oxidase II gene (see Table 1 and supple-

mentary information for details; accession numbers in

GenBank for DNA sequences are AY152477–AY152526,

AY930466–AY931013, DQ812553–DQ812885 and

EU108879–EU109174). Six of the 11 loci (P181,

P3072, P2956, P667, P7 and P22) map to chromosomes

1–3 and are subsumed by inversions (Roethele et al.,

2001; Feder et al., 2003b). Four of the remaining five

loci analysed in the study (P661, P2963, P1700 and

P309) map to chromosomes 4 and 5 (Table 1). The

exact map position of P3060 is not known, but it is not

located on chromosomes 1–3. The 11 loci analysed in

the current study are therefore broadly representative of

R. pomonella genome with respect to map position (all

five major autosomes were sequenced for at least two

genes each), location within rearranged vs. collinear

regions and the potential for host-related selection and

introgression. Consequently, if fixed genetic differences

do exist among R. pomonella complex flies, there are

reasonable expectations that they should be revealed by

the current survey.

Fig. 1 Collection sites for fly populations analysed in the study. See

supplementary Table S1 for full site information. The Rhagoletis

pomonella hawthorn fly sites Evtm CJ and Evtm MC (Coajomulco,

Morelos and Tancitaro, Michoacan, Mexico respectively) are located

in the Eje Volcanico Trans Mexicano (area outlined in green). Sites

Smo SJ and Smo PL (San Joaquin, Queretaro and Piletas, Veracruz)

are in the Sierra Madre Oriental Mountains of Mexico (area outlined

in orange). Sites Pom NY (Geneva, NY), Pom MI (Grant, MI) and

Pom TX (Brazos Bend, TX) are R. pomonella populations in the US

Sites Dog IN (Granger, IN) and Dog GA (Byron, GA) are flowering

dogwood populations. Sites Zeph WA (Dixie, Washington) and Zeph

PA (Munson, PA) are R. zephyria populations. Sites Mend MI

(Sawyer, MI) and Mend PA (Middleburg, PA) are R. mendax sites.

Site R. corn (Urbana, IL) is an R. cornivora population, whereas site

R. elect. (Dowagiac, MI) is a population of the outgroup Rhagoletis

electromorpha. Ranges for R. pomonella (red line), R. mendax (purple

line), R. zephyria (light blue line), R. cornivora (black line) and the

flowering dogwood fly (pink line) in the USA are also shown. The

majority of R. pomonella taxa are broadly sympatric in the eastern

USA. The exception is R. zephyria, which is parapatric and exten-

sively distributed in the western USA (solid light blue line), as well as

the East (stippled light blue line). Given their pattern of host plant

usage, eastern populations of R. zephyria appear to be endemic and

not introduced (Gavrilovic et al., 2007), but this hypothesis requires

further phylogeographic analysis.

902 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

DNA cloning and sequencing

Genomic DNA were isolated from individual flies and

PCR amplified for 35 cycles (94 �C, 30 s; 52 �C, 1 min;

72 �C, 1.5 min) using locus-specific primers for the 11

nuclear and mtDNA fragments as described in Roethele

et al. (2001). Products were TA cloned into pCR II vectors

(Invitrogen, Calsbad, CA, USA). PCR amplification prod-

ucts were initially cloned separately for a minimum of

two flies from each study site, with an attempt made to

sequence four to six clones per locus per fly in both the

5¢- and 3¢-directions on an ABI 3700 sequencer using the

ABI Prism� BigDyeTM Terminator v3.0 system (Applied

Biosystems, Foster City, CA, USA). To try and increase

sample sizes for certain sites, we also separately amplified

genomic DNA for four to eight flies from the site, and TA

cloned the pooled amplification products for sequencing.

To avoid analysis of identical alleles from the same

individual, sequences generated from the pooled library

were not included unless they differed from each other.

We were unable to generate reliable sequence data for

the Brazos Bend, Texas (Pom TX) and Geneva, New York

(Pom NY) hawthorn-infesting R. pomonella population

for the locus P1700. In addition, we only sequenced

R. mendax and R. zephyria flies from the Pennsylvania

(PA) sites for P1700 and mtDNA.

Gene tree construction and analysis

Parsimony and maximum-likelihood gene trees were

constructed using PAUP*b10 (Swofford, 2002). For the

parsimony analysis, gaps were treated as a fifth base pair,

with indels of identical length and sequence position

recoded to count as single mutational steps. Rhagoletis

electromorpha was used as an out-group taxon to root

trees. Parsimony and maximum-likelihood trees were

very similar and so we report the results for only the

parsimony trees here. Intragenic recombination was

statistically tested using the methods of Hudson & Kaplan

(1985). The molecular clock was tested for each locus for

R. pomonella and R. electromorpha sequences by comparing

log-likelihood scores enforcing vs. relaxing the clock

hypothesis for the best supported DNA substitution model

identified using MODELTEST (Posada & Crandall, 1998).

Neighbour-joining trees (Saitou & Nei, 1987) summa-

rizing the overall genetic relatedness of populations were

constructed using PHYLIP v 3.66 (Felsenstein, 1989).

Trees were constructed separately for the six loci map-

ping to chromosomes 1–3 plus P1700 and for the

remaining four loci residing elsewhere in the genome.

To construct the neighbour-joining trees, mean pairwise

uncorrected genetic distances were first computed sepa-

rately for each locus between each pair of populations

(except the PA sites), as well as between these popula-

tions and the outgroup R. electromorpha using Mega v3.1

(Kumar et al., 2004). The pairwise distance between two

populations for a locus was then divided by the average

distance of all R. pomonella populations to R. electromorpha

to standardize for sequence length and substitution rate

differences among loci. For the two SMO populations,

two diverged haplotypes could generally be identified at

loci, one more closely related to Mexican EVTM alleles

and the other to SN alleles from the USA (see Figs 2–4

and supplementary Figs S1–S4). We considered these

haplotype classes as separate SMO populations in the

calculations of genetic distance. Similarly, for the chro-

mosome 1–3 loci and P1700, taxa possessing North (N)

and South–North (SN) haplotypes at a locus were

considered different populations in genetic distance

calculations. Standardized distances were then averaged

across P181, P3072, P2956, P667, P7, P22 and P1700 and

across P661, P2963, P309 and P3060 to give the overall

pairwise distances used for tree construction. Because of

the variable presence of N haplotypes in taxa other than

R. pomonella, we only included the N haplotype popula-

tion from NY as a general representative of the N clade in

the network.

Unfortunately, other analytical approaches, such as the

isolation model with migration (IM or IMa; Nielsen &

Wakeley, 2001; Hey & Nielsen, 2007), are not readily

applicable to the R. pomonella data set because of viola-

tions of underlying assumptions of the model, including

that there be no unsampled populations exchanging

genes with the sampled populations or their ancestor, no

directional or balancing selection acting on sites (selec-

tive neutrality), no recombination within loci and free

recombination between loci.

Analysis of population differentiation and structure

A hierarchical analysis of molecular variance (AMOVAAMOVA)

was performed using Arlequin 2.0 (Schneider et al.,

Table 1 Loci sequenced in the study.

Locus Chr. Seq. Ln P Model Rec.

P181 1 288 1.000 TrN + I 2

P3072 1 465 0.377 TIM 5

P2956 2 510 0.244 TrN + I 5

P667 2 583 0.866 TVM + G 5

P22 3 593 0.757 HKY 2

P7 3 736 0.576 TrN + I 2

P2963 4 377 0.756 GTR + G 3

P661 4 436 0.088 TVM + G 6

P1700 5 484 0.997 HKY + G 6

P309 5 482 1.000 K81uf + G 3

P3060 ? 562 0.999 HKY 0

mtDNA – 626 0.990 TrN + I 0

Given are chromosome map positions (Chr.), the length of the

sequence in base pairs (Seq. Ln), the probability level (P) for

conformation to a molecular clock, the ML substitution model

(model) determined by MODELTEST using the AIC criterion (Posada

& Crandall, 1998), and the minimum number of recombination

events (Rec.) estimated by the method of Hudson & Kaplan (1985)

for each locus.

Radiation of Rhagoletis pomonella complex 903

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

2000) to test for genetic structuring among populations.

For the AMOVAAMOVAs, populations were divided into six

groups: (1) EVTM R. pomonella sites (MC and CJ); (2)

SMO R. pomonella sites (SJ and PL); (3) US R. pomonella

sites (Pom MI, NY and TX); (4) flowering dogwood fly

sites (Dog IN and GA); (5) R. mendax (Mend MI site); and

(a) (b)

(c) (d)

Fig. 2 Most parsimonious gene trees for (a) P3072, (b) P22, (c) P2956 and (d) mtDNA. Species and site designations are given in Fig. 1

legend and supplementary Table S1. Trees are scaled so that the longest distance from an allele to the outgroup Rhagoletis electromorpha

(R. elect.) are relatively the same across loci. Chromosome position for loci, sequence lengths (no. of bp), branch lengths (no. of steps), and

bootstrap support for nodes (10 000 reps.) are given. The three numbers in the arrow brackets (< >) following site designation for an allele

indicate the number of identical alleles sequenced for the population, the number of alleles sequenced that differ by a single substitution

from the haplotype and the number of alleles that differ by two substitutions respectively. N (blue coloured) designates North US alleles,

SN (red) = South ⁄ North alleles, SMO (orange) = Mexican Sierra Madre Oriental alleles and EVTM (green) = Mexican Eje Volcanico Trans

Mexicano. Yellow-coloured circles demark the relative depth of the node connecting N alleles for nuclear loci and SN alleles for mtDNA

with EVTM haplotypes. The congruence of relative N vs. SN ⁄ SMO ⁄ EVTM node depths among chromosomes 1–3 loci and mtDNA (compare

yellow-coloured nodes) is hypothesized to be a consequence of the past geographic separation of EVTM and US hawthorn fly populations

�1.57 Ma (Feder et al., 2003a). Black-coloured nodes represent the shallowest relative node depth connecting SN and EVTM haplotypes.

Comparisons of these nodes to those in Fig. 4a–d highlight the deeper coalescence times of SN and EVTM haplotypes between chromosome

1–3 loci, hypothesized to be because of the greater permeability and more recent gene flow of collinear loci outside of the rearrangements on

chromosomes 1–3.

904 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

(6) R. zephyria (Zeph WA site). Separate AMOVAAMOVAs were

performed for the six chromosome 1–3 loci and P1700

with N haplotypes both excluded ()N) and included

(+N). Neither R. cornivora nor R. electromorpha were

included in the AMOVAAMOVAs, as gene trees showed these

taxa to be basal and their inclusion would have inflated

the per cent variation explained by among-group differ-

ences for the other ‘in-group’ taxa in the analysis.

Tamura & Nei (1993) genetic distances between in-group

taxa and R. electromorpha for loci were calculated using

Mega v3.1 (Kumar et al., 2004).

Results

None of the 11 sequenced nuclear genes or mtDNA

deviated significantly from a molecular clock (Table 1).

All nuclear loci except P3060 displayed evidence for

possible recombination, as implied by the method of

Hudson and Kaplan (Table 1). Inferred recombination

was generally limited, however, to alleles within the

same haplotype class for loci on chromosomes 1–3 or

same geographic population ⁄ taxa. There was no evidence

for recombination among mtDNA sequences (Table 1).

Confirmation of clock-like evolution and intra-haplo-

type-limited recombination allowed for inferences to be

drawn concerning the three goals of the study based on

comparisons of branching topologies and node depths

among gene trees. We organize the presentation of these

results below into three subsections highlighting the

similarities and differences in patterns of genetic differ-

entiation for: (1) the six loci on chromosomes 1–3 (genes

residing in inverted regions of the genome); (2) the five

loci mapping to chromosomes 4 and 5 (genes in putative

co-linear regions of the genome); and (3) mtDNA.

Genes on chromosomes 1–3

Several features characterized the gene trees for the six

loci mapping to chromosomes 1–3 (Fig. 2a–c and sup-

plementary Figs S1–S3). First, R. cornivora (designated by

R. corn. in the figures) was clearly basal to the other

R. pomonella populations, displaying fixed, autapomor-

phic substitutions for P181, P3072, P2956, P667, P7 and

P22.

Second, with the exception of P7, gene trees for

chromosome 1–3 loci resolved US and Mexican R. pomo-

nella, R. zephyria, R. mendax and the flowering dogwood

fly as a distinct monophyletic ‘in-group’ clade from

R. cornivora (Fig. 2a–c and supplementary Figs S1–S3).

Third, within the ‘in-group’ taxa of US and Mexican

R. pomonella populations, R. zephyria, R. mendax and the

flowering dogwood fly, two distinct clades ⁄ classes of

haplotypes could be identified for each locus. One clade

consisted of North haplotypes (designated by the blue-

coloured letter ‘N’ in Fig. 2a–c and supplementary

Figs S1–S3) found in northern populations of R. pomo-

nella (NY and MI), and to varying degrees, depending on

the locus, in R. mendax, R. zephyria and the flowering

dogwood fly as well. The other clade was comprised of a

trinity of haplotypes present: (1) in northern and

southern populations of US R. pomonella, R. mendax,

R. zephyria and the flowering dogwood fly (designated

by the red-coloured letters ‘SN’ in the figures); (2) in the

Sierra Madre Oriental population of Mexican hawthorn

flies (designated by the orange-coloured letters ‘SMO’);

and (3) in the Eje Volcanico Trans Mexicano region

(designated by the green-coloured letters ‘EVTM’).

Fourth, the depths of the nodes uniting the N clade and

the triad of SN ⁄ SMO ⁄ EVTM haplotypes relative to the

(a)

(b)

Fig. 3 Neighbour-joining trees based on overall genetic distances

among Rhagoletis pomonella taxa relative to the outgroup taxon

Rhagoletis electromorpha for (a) chromosome 1–3 loci and P1700 and

(b) loci not residing on chromosomes 1–3. Trees are scaled so that the

lengths of the branch from the R. electromorpha ⁄ R. pomonella node to

terminal populations are the same in (a) and (b). For chromosome

1–3 loci, the N and SN clades of haplotypes present were considered as

separate populations. Only the NY N haplotypes were included in the

network to serve as a general indicated of the divergence of this clade

of alleles in the USA. For the SMO, alleles displaying affinity to SN

haplotypes and EVTM haplotypes were also treated as separate

populations. The N vs. SN ⁄ SMO ⁄ EVTM haplotype node is designated

by the compare yellow-coloured circle in (a) and the SN vs. EVTM

haplotype nodes by the black-coloured circle in (a) and (b).

Radiation of Rhagoletis pomonella complex 905

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

outgroup R. electromorpha were similar among loci and

concurred with a deep division seen for mtDNA between

EVTM flies in Mexico and the remainder of the in-group

taxa (compare yellow-coloured nodes in Fig. 2a–c and

supplementary Figs S1–S3 among loci). The congruence

of these nodes among gene trees is hypothesized to

reflect the initial isolation of the Mexican EVTM popu-

lation that predated the secondary contact and introgres-

sion of nuclear SN haplotypes into the USA.

Fifth, aside from P7, the Mexican population of

hawthorn-infesting R. pomonella in the EVTM designated

by the green-coloured haplotypes was monophyletic and

displayed fixed, autapomorphic differences from the

other taxa for the five other nuclear loci on chromosomes

1–3 (Fig. 2a–c and supplementary Figs S1–S3).

Sixth, the hawthorn-infesting populations of R. pomo-

nella in the SMO generally contained a mixture of alleles

that showed affinity to both the EVTM and the USA. For

P3072 and P667, SMO flies possessed EVTM haplotypes,

forming a monophyletic clade (Fig. 2a and supplemen-

tary Fig. S2). For P22 and P181, SMO flies had alleles

embedded within the SN class of haplotypes present in

US R. pomonella, R. mendax, R. zephyria and flowering

dogwood fly populations (Fig. 2b and supplementary

Fig. S1). The locus 2956 had haplotypes related to both

EVTM and the SN clades of alleles (Fig. 2b). Finally, P7

(supplementary Fig. S1) contrasted with the other chro-

mosome 1–3 genes in that SMO alleles formed a distinct

clade genealogically unrelated to either EVTM or SN

haplotypes.

Seventh, no fixed diagnostic substitution distinguished

R. pomonella, R. mendax, R. zephyria and the flowering

dogwood fly in the USA from each other for P3072, P181,

P2956, P667 or P7, although an autapomorphy possibly

(a) (b)

(c) (d)

Fig. 4 Most parsimonious gene trees

for (a) P309, (b) P3060, (c) P1700 and

(d) 2963. See Fig. 2 legend for additional

information.

906 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

exists for R. zephyria at P22 (Fig. 2a–c and supplementary

Figs S1–S4). However, private alleles were observed at

several genes for R. mendax and R. zephyria, and much

less so for the flowering dogwood fly.

The neighbour-joining network based on overall rel-

ative genetic distance measures among populations for

chromosome 1–3 loci and P1700 (Fig. 3a) highlighted the

general patterns discerned from the individual gene trees,

including: (1) the basal position for R. cornivora in the

pomonella complex; (2) the presence of a deep northern

clade of alleles (designated by Pom-NY) within the

in-group taxa in the USA; (3) the existence of a distinct,

monophyletic clade of Mexican hawthorn fly haplotypes

in the EVTM; and (4) the composite nature of the

Mexican hawthorn fly population in the SMO, contain-

ing genetic elements with affinity to both EVTM and US

haplotypes. The network also inferred that R. pomonella

flies in the USA are overall genetically most similar to

flowering dogwood flies, followed by R. zephyria, and

then R. mendax. In addition, the class of alleles within the

SMO showing affinity to SN haplotypes in the USA is

most closely related to R. pomonella, and, in particular,

the TX population.

Genes mapping outside chromosomes 1–3

The five loci mapping outside chromosomes 1–3 were

similar to the six chromosome 1–3 loci in several

respects. Gene trees for P661, P2963, P1700, P309 and

P3060 clearly positioned R. cornivora basal to the other

members of the R. pomonella complex (Fig. 4a–d and

supplementary Fig. S4). The in-group taxa comprised of

US and Mexican R. pomonella, R. zephyria, R. mendax and

the flowering dogwood fly were also monophyletic for

P661, P2963, P1700, P309 and P3060, sharing several

synapomorphic substitutions in common. In addition,

P661, P2963, P1700, P309 and P3060 showed substantial

genetic differentiation among R. pomonella populations,

as indicated by significant FCT values in a hierarchical

AMOVAAMOVA, similar to the six chromosome 1–3 loci (Table 2).

Indeed, P1700 displayed two potentially fixed autapo-

morphic substitutions diagnostically distinguishing

R. zephyria from all other R. pomonella complex flies

(Fig. 4c). The flowering dogwood fly also possessed a

clade of private haplotypes for P1700, as did R. mendax for

the loci P309 and P1700 (Fig. 4a) and R. zephyria for P661

(supplementary Fig. S4).

There were several pronounced differences, however,

between loci residing on chromosomes 1–3 and most

genes mapping elsewhere in the genome. First, except for

P1700, no locus mapping outside of chromosomes 1–3

possessed a highly diverged N haplotype class of alleles

(Fig. 4b–d and supplementary Fig. S4). Second, no

monophyletic clade of Mexican EVTM haplotypes was

present for P309, P3060 or P661 (Fig. 4a,b and supple-

mentary Fig. S4). Third, Mexican SMO alleles for P309,

Table 2 Hierarchical AMOVAAMOVA among EVTM, SMO and US populations of Rhagoletis pomonella, R. mendax, R. zephyria and the flowering dogwood

fly.

Locus Chr. N

Percent variation explained Fixation indices

Among

pops.

Among sites

in pops.

Within

sites

FCT

(pop. ⁄ total)

FSC

(site ⁄ pop.)

FST

(site ⁄ total)

P181 1 )N 50.2 )0.3 50.1 0.502*** )0.005 0.499***

+N 37.4 6.4 56.2 0.374* 0.102** 0.438***

P3072 1 )N 61.7 10.3 28.0 0.617*** 0.269** 0.720***

+N 51.9 16.7 31.4 0.519*** 0.347*** 0.686***

P2956 2 )N 80.3 )4.2 23.9 0.803*** )0.021 0.760***

+N 51.7 7.0 41.3 0.517*** 0.144* 0.586***

P667 2 )N 29.9 34.3 35.8 0.299* 0.490*** 0.642***

+N 31.1 26.5 42.4 0.311* 0.385*** 0.572***

P7 3 )N 40.7 2.7 56.6 0.407* 0.046 0.434***

+N 45.4 16.1 38.5 0.385* 0.262** 0.546***

P22 3 )N 30.6 15.7 53.7 0.306* 0.226* 0.462***

+N 28.7 18.3 53.0 0.283** 0.257*** 0.470***

P661 4 )N 13.8 18.5 67.7 0.138** 0.215*** 0.323***

P2963 4 )N 41.6 10.3 48.1 0.416** 0.176*** 0.519***

P1700 5 )N 71.8 1.3 26.9 0.717** 0.047 0.731***

+N 65.6 1.1 33.3 0.656*** 0.030 0.667***

P309 5 )N 38.7 2.1 59.2 0.387*** 0.033 0.408***

P3060 ? )N 10.4 8.7 80.9 0.104 0.097** 0.191***

See the Methods section and supplementary Table S1 for designation of sites within populations. Chr. = chromosome that locus resides on. For

the N column: )N = analysis performed without N haplotypes, +N = analysis performed including N haplotypes for chromosome 1–3 loci and

P1700. The loci P2963, P661, P309 and P3060 do not possess N haplotypes, and so are designated )N. *P < 0.05, **P < 0.01, ***P < 0.001, as

determined by permutation tests with 10 000 replicates.

Radiation of Rhagoletis pomonella complex 907

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

P3060 and P661 were not distinct and were embedded in

the gene tree with SN alleles from the USA; these three

loci provided no evidence for the SMO having genetic

affinity to the EVTM. Finally, although a significant

proportion of the genetic variation present for P309,

P3060, P2963 and P661 was partitioned among taxa, the

relative degree of genetic divergence separating the

alleles displaying variation was less for these loci than

for the six loci residing on chromosomes 1–3 (compare

the depths of the black-coloured circles signifying the

nodes uniting EVTM and SN between gene tree in Fig. 4

and supplementary Fig. S4 vs. Fig. 2 and supplementary

Figs S1–S3, as well as between the neighbour-joining

networks in Fig. 3a,b; note: P1700 was included with the

six chromosome 1–3 loci for network construction

because of the observation of distinct N and SN haplo-

types at the locus). The shallower divergence among

alleles for nonchromosomes 1–3 loci was not because

of a slower rate of evolution for these genes (mean

Tamura–Nei distance for in-group R. pomonella taxa to

R. electromorpha for loci not on chromosomes

1–3 = 0.053 ± 0.0152, n = 4, range 0.032–0.066; mean

for chromosome 1–3 loci = 0.046 ± 0.0222, n = 6, range

0.020–0.085; distance for P1700 = 0.056 ± 0.0104).

Moreover, the gene trees in Figs 2 and 4 are all scaled

relative to R. electromorpha and the networks in Fig. 3 are

based on relative, not raw, genetic distances to R. electro-

morpha. Rather, the difference reflects the greater genetic

similarity of P661, P2963, P309 and P3060 alleles among

the in-group R. pomonella taxa than P181, P3072, P2956,

P667, P7, P22 and P1700 haplotypes. As a consequence,

the neighbour-joining network for P661, P2963, P309

and P3060 (Fig. 3b) was less well resolved than that for

the chromosome 1–3 loci and P1700 (Fig. 3a), with US

R. pomonella and flowering dogwood fly populations

appearing paraphyletic and R. zephyria being embedded

within the Mexican SMO populations.

MtDNA

MtDNA displayed both similarities and differences from

the patterns of genetic variation observed for the nuclear

loci. Rhagoletis cornivora was again basal in the R. pomo-

nella complex. Like P661, P2963, P309 and P3060,

mtDNA did not display a highly diverged N haplotype

confined to northern US fly populations (Fig. 2d). How-

ever, the EVTM and the in-group R. pomonella taxa

showed a deep genetic divergence for mtDNA that was

congruent with the relative node depths between N and

SN ⁄ SMO ⁄ EVTM clades of haplotypes for chromosomes

1–3 loci (Fig. 2). However, there was no signature for

subsequent introgression in the mtDNA between the

EVTM and US or SMO fly populations. Moreover, unlike

the chromosome 1–3 loci, but like P661, P309 and P3600,

mtDNA, haplotypes for SMO flies were not a mixture of

EVTM and US types. Instead, they were all embedded

within and very similar to SN haplotypes found in the

USA (Fig. 2d). There was no diagnostic mtDNA autapo-

morphy distinguishing R. pomonella, R. mendax, R. zephy-

ria, the flowering dogwood fly or SMO flies for COII

(Fig. 2d). However, R. mendax and R. zephyria appeared

to possess potential private substitutions and, on this

basis, were the most genetically differentiated of the

in-group taxa for mtDNA.

Discussion

Genetic signposts for speciation

The first goal of the current sequence study was to

determine whether genetically distinguishable stages

of population divergence could be identified in the

R. pomonella complex. Our results suggest that on a

coarse, qualitative scale such stages may be recognized

with respect to: (1) R. cornivora (fixed, reciprocal mono-

phyly for all loci); (2) EVTM hawthorn flies (fixed

substitutions and monophyly for several, but not all

loci); (3) SMO hawthorn flies (fixed differences for a

couple of loci and some private alleles for a handful of

genes); (4) the in-group sibling taxa in the USA (possible

autapomorphies and private alleles at a few loci); and (5)

the apple and hawthorn host races (no fixed substitution

or private allele difference). The sequence data did not

reveal any finer genetic resolution than previous allo-

zymes studies, however (Feder & Bush, 1989; Berlocher

et al., 1993; Feder, 1998; Feder et al., 1999; Berlocher,

2000). Moreover, neither nuclear or mtDNA loci clearly

resolved the phylogenetic relationships (i.e. a bifurcating

gene tree) among the in-group populations of R. pomo-

nella, R. mendax, R. zephyria and flowering dogwood fly in

the USA. Reciprocal monophyly was not observed for

any locus among these four in-group taxa and it is

conceivable that with increased sampling the potential

private and autapomorphic variants seen for R. mendax

and R. zephyria for a couple of genes in the current study

will prove to be shared among taxa, albeit at low

frequencies, as is the case for allozymes. Such an

outcome is suggested by RFLP data for P1700 for

R. mendax and R. zephyria (Schwarz et al., 2005). The

sequence analysis for R. pomonella flies therefore did not

codify a clear demarcation of taxonomic status for US

populations based on phylogenetic or lineage concepts of

species, although the neighbour-joining network based

on overall relative genetic distances measures for chro-

mosome 1–3 loci and P1700 did imply that R. pomonella

and the flowering dogwood fly are sister taxa, followed

by R. zephyria and R. mendax (Fig. 3a).

There are two interpretations for the general lack of

phylogenetic resolution of the DNA data for the USA in-

group taxa. The first explanation is that diagnostic

differences exist, but that the sequencing of 11 nuclear

loci and mtDNA was not an extensive enough survey of

the genome to detect these differences. The current data

for the in-group taxa therefore mainly reflect the

908 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

consequences of incomplete lineage sorting among the

relatively recently diverged in-group fly populations in

the USA. The second interpretation is that the over 30

allozymes and 11 nuclear loci analysed to date actually

do represent an accurate snapshot of the state of genetic

differentiation among the in-group R. pomonella taxa. If

true, then these taxa are probably best characterized as a

continuum of quantitative genetic sibling species and

host races distinguishable as sympatric genotypic clusters

(Mallet, 1995) on the basis of near private mutations

and ⁄ or allele frequency differences (Feder, 1998). Mem-

bers of the group would therefore reflect a dynamic

equilibrium between low-level gene flow and the

strength of differential host selection, forming a complex

analogous to the syngameons of plants. The recent

findings of Schwarz et al. (2005) of an historical hybrid-

ization event between R. mendax and R. zephyria poten-

tially giving rise to a new race ⁄ species of fly in the eastern

USA infesting introduced honeysuckle from the Lonicera

tatarica complex are consistent with this dynamic equi-

librium view for the R. pomonella group.

Although the current data failed to reveal reciprocal

monophyly among the R. pomonella in-group taxa in the

USA, there were several a priori reasons to believe that

the 11 anonymous nuclear loci sequenced in the study

should have been sufficient to uncover fixed, diagnostic

differences (autapomorphies) provided they exist. First,

the in-group taxa R. pomonella, R. mendax, R. zephyria and

the flowering dogwood fly possess differential host-

associated adaptations related to diapause and host

discrimination that serve as significant ecological barriers

to gene flow (Smith, 1986; Bierbaum & Bush, 1990;

Feder et al., 1993, 1994; Linn et al., 2003; Schwarz et al.,

2007). Second, R. pomonella · R. mendax and R. men-

dax · R. zephyria display a degree of nonhost-related

prezygotic isolation. In particular, male R. mendax have

difficulty mating with female R. pomonella because of a

difference in adult body size between the species (Smith,

1986; J. L. Feder, pers. obs.) and even the similar-sized

R. mendax and R. zephyria show significant sexual isola-

tion (Schwarz & McPheron, 2007). Third, R. pomonella

and R. mendax also exhibit post-zygotic isolation (over

50%) that could be because of intrinsic genomic incom-

patibility and ⁄ or host fruit-related larval survivorship

differences (Smith, 1986; J. L. Feder, pers. obs.). No

prezygotic isolation appears to exist between R. pomonella

and the flowering dogwood fly. However, there is subtle

evidence for possible low-level post-zygotic isolation.

Although there is no reduction in fecundity in apple

race · dogwood fly matings, backcrosses of F1 hybrids to

parental taxa suggest that egg hatch may be reduced by

10% for second generation flies of mixed apple ⁄ dogwood

ancestry (Smith, 1986). Fourth, although there is also no

evidence for reduced fecundity or fertility in hybrid

crosses between the apple and hawthorn host races of

R. pomonella (Reissig & Smith, 1978; Smith, 1986), F1

apple · hawthorn flies fail to orient to host fruit volatiles

in flight tunnel assays (Linn et al., 2004; Dambroski et al.,

2005) indicative of an impaired olfactory system that

could have significant fitness consequences in the field.

Fifth, the 11 nuclear loci we sequenced in the study

representatively cover five of the six chromosomes

constituting the R. pomonella genome (we had no marker

on the small dot sixth chromosome). Moreover, six of

these loci reside in inverted regions on chromosomes 1–3

that are in linkage disequilibrium with allozymes show-

ing host-related differences among R. pomonella taxa

(Roethele et al., 2001; Feder et al., 2003b). These rear-

ranged regions contain genes affecting diapause life-

history variation that are responsible for generating

reproductive isolation between the sibling species and

host races (Feder et al., 1997a, b; Filchak et al., 2000). As

such, these six sequenced loci are located in what could

be considered ‘islands of speciation’ for R. pomonella and

are prime targets for reflecting population divergence.

Nevertheless, moving from the temporally flat, infinite-

allele framework of the allozymes to the genealogically

deeper, infinite-site perspective provided by the current

DNA sequence data did not appreciably change the

resulting view of the R. pomonella in-group taxa. It is

conceivable that advances in molecular techniques will

soon make it possible to cost effectively sequence a large

portion of the genome of R. pomonella for variation and

reveal an essential core of fixed, host-related substitu-

tions defining the different species. However, we must

also entertain the possibility that even more exhaustive

searching of the in-group taxa will not greatly change the

implications of this study; R. pomonella host races and

sibling species represent a seamless transition that form

without the complete closure of the genome to gene flow

and in the absence of the evolution of fixed genetic

differences. Only when we move to the level of R. corn-

ivora at the base of the R. pomonella complex where

nearly complete pre- and post-zygotic isolations are

present (Berlocher, 2000) do we observe reciprocally

discreet genetic differences throughout the genome.

Speciation mode plurality

With respect to the second issue of speciation mode

plurality, the sequence data are consistent with the

hypothesis that past geographic isolation and subsequent

introgression of inversion polymorphism from an isolated

hawthorn-infesting fly population of R. pomonella in the

EVTM of Mexico played a role in the adaptive radiation

of the USA in-group taxa (Feder et al., 2003a). From this

study, we could infer that the initial isolation of the

EVTM population in Mexico (estimated at �1.57 Ma

based on a mtDNA insect molecular clock; Feder et al.,

2003a) occurred after the divergence of R. cornivora from

the rest of the complex. Moreover, the introgression of

inversions from the EVTM into the USA appears to have

preceded the genesis of R. mendax, R. zephyria and the

flowering dogwood fly, as well as the apple race.

Radiation of Rhagoletis pomonella complex 909

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

Additional information is needed, however, to assess the

current taxonomic status of EVTM flies. Hawthorn-

infesting populations of R. pomonella in the EVTM of

Mexico clearly formed a monophyletic clade from the

rest of the group for several loci. The EVTM population

also does not presently appear to be in contact with

hawthorn-infesting populations in the SMO or USA (Rull

et al., 2006). The genetics and biogeography of EVTM

population are therefore consistent with these flies

currently being completely reproductively isolated from

the other R. pomonella taxa and potentially having been

formed via an allopatric mode of speciation not involving

host shifting. However, further studies involving crosses

of EVTM to SMO and US flies are needed to test for

nonhost-related reproductive isolation to confirm this

hypothesis.

The status of hawthorn flies in the SMO of Mexico is

also not definitive. The SMO population contains genetic

elements from both the EVTM and the USA, consistent

with it having served as a conduit for past gene flow

connecting the EVTM with an ancestral US population of

hawthorn-infesting flies. Incomplete lineage sorting in

combination with low levels of recombination in inverted

regions of chromosomes 1–3 could contribute to the

mosaic nature of the SMO gene pool, but this explanation

alone is insufficient to account for the totality of the

pattern. In the neighbour-joining networks (Fig. 3), the

alleles in the SMO showing affinity to SN haplotypes in

the USA are more closely related to each other than are

those genes in the SMO and EVTM showing affinity. In

addition, mtDNA indicates a close genetic relationship

between SMO and US R. pomonella populations (Fig. 2d).

These data imply a temporal dynamic of differential

introgression, with more extensive and recent gene flow

between the USA and the SMO than between the SMO

and EVTM. Indeed, it is not obvious whether the

differences seen between SMO and US hawthorn fly

populations are indicative of interspecific divergence or

conspecific geographic variation. Tentative support for

the latter hypothesis comes from a recent survey of

microsatellites in SMO and US hawthorn flies revealing

clinal variation for several loci between the regions

(Michel et al., 2007). More extensive genetic surveys

and tests for nonhost-related reproductive isolation are

needed to clarify the taxonomic status of SMO flies.

Patterns of genetic differentiation across the genome

Our third goal was to assess patterns of genetic differ-

entiation across the R. pomonella genome. In this regard,

with the exception of P1700, there was a general trend

for loci residing on chromosomes 1–3 to display more

pronounced levels of allelic divergence within and

among populations than genes mapping to chromo-

somes 4 and 5. Previous studies have documented this

pattern among EVTM, SMO and US hawthorn-infesting

populations of R. pomonella (Feder et al., 2005; Xie et al.,

2007). Here, we observe that the same pattern holds for

the in-group taxa R. mendax, R. zephyria and the flow-

ering dogwood fly. The result implies that the rear-

ranged regions of chromosomes 1–3 have become less

permeable to gene flow among fly taxa than loci in

putative co-linear regions of the genome, consistent

with recent inversion models for speciation (Noor et al.,

2001a; Rieseberg, 2001; Navarro & Barton, 2003;

Kirkpatrick & Barton, 2006). Against this backdrop,

the lack of a signature of mtDNA introgression between

the EVTM and SMO ⁄ US populations since the initial

isolation of Mexican highland flies �1.57 Ma is puz-

zling. Possible explanations for the disjunct nature of

mtDNA variation between the EVTM and SMO ⁄ USA

include male-driven gene flow, cytonuclear gene

incompatibilities, and ⁄ or a factor directly under differ-

ential selection in the mtDNA.

A delta of life?

In conclusion, our studies of the R. pomonella complex

suggest that many of the dichotomies we impose

concerning geographic modes ⁄ mechanisms of diver-

gence, the cladistic splitting of taxa and systematic

categories of organisms may blur during speciation.

Rather than the analogy of a branching ‘tree of life’, a

‘delta of life’ comprised of many inter-tangled channels

may be more appropriate for describing the formative

stages of R. pomonella speciation. Thus, a plurality of

mechanisms and processes including host-associated

selection, sympatric host shifts, biogeography, secondary

contact, past differential introgression, inversions, clines

and ongoing and low-level hybridization probably inter-

act to generate what we call species in the R. pomonella

complex. These species do not necessarily display

reciprocal monophyly at their inception. Rather, they

exist as genotypic clusters that can be statistically iden-

tified on the basis of near private alleles and frequency

differences at a subset of loci resulting from differential

selection and gene flow barriers across the genome. Thus,

we see no obvious qualitative genetic difference between

the apple and hawthorn host races of R. pomonella,

R. mendax, R. zephyria or the flowering dogwood fly in

the current sequence-based analysis or past allozyme

studies. Eventually, some of these species may continue

to be channelled by divergent ecological selection alone

or in combination with processes acting during periods of

geographic isolation to consolidate into the types of

discreet operational taxonomic units (OTUs) favoured by

systematists for classification. Such a scenario may help

explain the monophyly of R. cornivora and the EVTM

population of hawthorn flies. However, care must be

taken in interpreting phylogenies of such OTUs, as during

the formative period of population divergence, there may

be no one single bifurcating population history encapsu-

lating the genetic and biogeographic ⁄ ecological processes

that acted across the genome to generate these ‘species’.

910 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

By contrast, other taxa may persist for extended periods

of time in a dynamic equilibrium of semi-permanent

divergence dictated by the balance between disruptive

selection and introgression, whereas still others may fuse

if environmental ⁄ ecological circumstances were to

change. This view of life and the genesis of biodiversity

is by no means unique to Rhagoletis (e.g. Wang et al.,

1997; Rieseberg et al., 1999; Noor et al., 2001b; Beltran

et al., 2002; Machado & Hey, 2003; Stump et al., 2005;

Turner et al., 2005; Mallet et al., 2007). It will be

interesting to see whether and how this view changes

as the entire genome sequences of more and more of

life’s diversity become known.

Acknowledgments

The authors thank the following individuals for their

assistance: S. Berlocher, G. Bush, B. Matta, B. McPheron,

J.J Smith, F. Wang, F. Wang Jr, J. Wise and the Trevor

Nichols Research Station of Michigan State University at

Fennville, MI. This research was supported, in part, by

grants from the National Science Foundation, a National

Research Initiative grant from the United States Depart-

ment of Agriculture, and the 21st Century Fund of the

state of Indiana to JLF.

References

Beltran, M., Jiggins, C.D., Bull, V., Linares, M., Mallet, J.,

McMillan, W.O. & Bermingham, E. 2002. Phylogenetic

discordance at the species boundary: comparative gene gene-

alogies among rapidly radiating Heliconius butterflies. Mol. Biol.

Evol. 19: 2176–2190.

Berlocher, S.H. 2000. Radiation and divergence in the Rhagoletis

pomonella species group: inferences from allozymes. Evolution

54: 543–557.

Berlocher, S.H. & Feder, J.L. 2002. Sympatric speciation in

phytophagous insects: moving beyond controversy? Annu.

Rev. Entomol. 47: 773–815.

Berlocher, S.H., McPheron, B.A., Feder, J.L. & Bush, G.L. 1993. A

revised phylogeny of the Rhagoletis pomonella (Diptera: Tephri-

tidae) sibling species group. Ann. Entomol. Soc. Am. 86: 716–727.

Bierbaum, T.J. & Bush, G.L. 1990. Genetic differentiation in the

viability of sibling species of Rhagoletis fruit flies on host plants,

and the influence of reduced hybrid viability on reproductive

isolation. Entomol. Exp. Appl., 55: 105–118.

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: Tephriti-

dae). Evolution 23: 237–251.

Coyne, J.A. & Orr, H.A. 2004. Speciation. Sinauer Associates,

Sunderland, MA, 545 pp.

Cracraft, J. 1989. Speciation and its ontology: the empirical

consequences of alternative species concepts for understand-

ing patterns and processes of differentiation. In: Speciation and

its Consequences (D. Otte & J. A. Endler, eds), pp. 28–59.

Sinauer Associates Inc., Sunderland, MA.

Dambroski, H.R. & Feder, J.L. 2007. Host plant and latitude-

related diapause variation in Rhagoletis pomonella: a test for

multifaceted life history adaptation on different stages of

diapause development. J. Evol. Biol. 20: 2101–2112.

Dambroski, H.R., Linn, C. Jr, Berlocher, S.H., Forbes, A.A.,

Roelofs, W. & Feder, J.L.. 2005. The genetic basis for fruit odor

discrimination in Rhagoletis flies and its significance for

sympatric host shifts. Evolution 59: 1953–1964.

Feder, J.L. 1998. The apple maggot fly, Rhagoletis pomonella: flies

in the face of conventional wisdom about speciation? In:

Endless Forms: Species and Speciation (D. Howard & S. H.

Berlocher, eds), pp. 130–144. Oxford University Press, Oxford.

Feder, J.L. & Bush, G.L. 1989. A field test of differential host

plant usage between two sibling species Rhagoletis pomonella

fruit flies (Diptera: Tephritidae) and its consequences for

sympatric models of speciation. Evolution 43: 1813–1819.

Feder, J.L., Hunt, T.A. & 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.

Feder, J.L., Opp, S., Wlazlo, B., Reynolds, K., Go, W. & Spisak, S.

1994. Host fidelity is an effective pre-mating barrier between

sympatric races of the Apple Maggot Fly. Proc. Natl Acad. Sci.

USA 91: 7990–7994.

Feder, J.L., Roethele, J.B., Wlazlo, B. & Berlocher, S.H. 1997a.

Selective maintenance of allozyme differences between sym-

patric host races of the apple maggot fly. Proc. Natl Acad. Sci.

USA, 94: 11417–11421.

Feder, J.L., Stolz, U., Lewis, K.M., Perry, W., Roethele, J.B. &

Rogers, A. 1997b. The effects of winter length on the genetics

of apple and hawthorn races of Rhagoletis pomonella (Diptera:

Tephritidae). Evolution 51: 1862–1876.

Feder, J.L., Williams, S.M., Berlocher, S.H., McPheron, B.A. &

Bush, G.L. 1999. The population genetics of the apple maggot

fly and the snowberry maggot: implications for models of

sympatric speciation. Entomol. Exp. Appl. 90: 9–24.

Feder, J.L., Berlocher, S.H., Roethele, J.B., Dambroski, H.R.,

Smith, J.J., Perry, W.L., Gavrilovic, V., Filchak, K.E., Rull, J. &

Aluja, M.. 2003a. Allopatric genetic origins for sympatric host-

plant shifts and race formation in Rhagoletis. Proc. Natl Acad. Sci.

USA 100: 10314–10319.

Feder, J.L., Roethele, J.B., Filchak, K.E., Niedbalski, J. &

Romero-Severson, J. 2003b. Evidence for inversion polymor-

phism related to sympatric host race formation in the apple

maggot fly, Rhagoletis pomonella. Genetics 163: 939–953.

Feder, J.L., Xie, X., Rull, J., Velez, S., Forbes, A., Dambroski, H.,

Filchak, K. & Aluja, M. 2005. Mayr, Dobzhansky, Bush and

the complexities of sympatric speciation in Rhagoletis. Proc.

Natl Acad. Sci. 102: 6573–6580.

Felsenstein, J. 1989. PHYLIP – phylogeny inference package.

Cladistics 5: 164–166.

Filchak, K.E., Roethele, J.B. & Feder, J.L. 2000. Natural selection

and sympatric divergence in the apple maggot Rhagoletis

pomonella. Nature 407: 739–742.

Gavrilovic, V., Bush, G.L., Schwarz, D., Crossno, J.E. & Smith,

J.J. 2007. Rhagoletis zephyria (Diptera: Tephritidae) in the Great

Lakes basin: a native insect on a native host? Ann. Entomol. Soc.

Amer. 100: 474–482.

Hey, J. & Nielsen, R. 2007. Integration within the Felsenstein

equation for improved Markov chain Monte Carlo methods in

population genetics. PNAS 104: 2785–2790.

Radiation of Rhagoletis pomonella complex 911

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

Hudson, R.D. & Coyne, J.A. 2002. Mathematical conse-

quences of the genealogical species concept. Evolution 56:

1557–1565.

Hudson, R.R. & Kaplan, N.L. 1985. Statistical properties of the

number of recombination events in the history of a sample of

DNA sequences. Genetics 111: 147–164.

Kirkpatrick, M. & Barton, N. 2006. Chromosomal inversions,

local adaptation, and speciation. Genetics 173: 419–434.

Kumar, S., Tamura, K. & Nei, M. 2004. MEGA3: integrated

software for molecular evolutionary genetics analysis and

sequence alignment. Brief. Bioinform. 5: 150–163.

Linn, C., Feder, J.L., Nojima, S., Dambroski, H.R., Berlocher,

S.H. & Roelofs, W. 2003. Fruit odor discrimination and

sympatric host race formation in Rhagoletis. Proc. Natl Acad.

Sci. USA 100: 11490–11493.

Linn, C.E., Dambroski, H.R., Feder, J.L., Berlocher, S.H., Nojima,

S. & Roelofs, W.L. 2004. Postzygotic isolating factor in

sympatric speciation in Rhagoletis flies: reduced response of

hybrids to parental host-fruit odors. Proc. Natl Acad. Sci. USA

101: 17753–17758.

Machado, C.A. & Hey, J. 2003. The causes of phylogenetic

conflict in a classic Drosophila species group. Proc. R. Soc. Lond.

B 270: 1193–1202.

Mallet, J. 1995. A species definition for the modern synthesis.

Trends Ecol. Evol. 10: 294–299.

Mallet, J. 2005. Speciation in the 21st Century. Review of

‘‘Speciation’’, by Jerry A. Coyne and H. Allen Orr. Heredity 95:

105–109.

Mallet, J., Beltran, M., Neukirchen, W. & Linares, M. 2007.

Natural hybridization in heliconiine butterflies: the species

boundary as a continuum. BMC Evol. Biol. 7: 28.

Mayr, E. 1963. Animal Species and Evolution. Harvard University

Press, Cambridge, MA, 797 pp.

Mayr, E. 2001. Wu’s genic view of species. J. Evol. Biol. 14: 866–

867.

McPheron, B.A., Smith, D.C. & Berlocher, S.H. 1988. Genetic

differences between host races of Rhagoletis pomonella. Nature

336: 64–66.

Michel, A.P., Rull, J., Aluja, M. & Feder, J.L. 2007. The genetic

structure of hawthorn-infesting Rhagoletis pomonella popula-

tions in Mexico: implications for sympatric host race forma-

tion. Mol. Ecol. 16: 2867–2878.

Navarro, A. & Barton, N.H. 2003. Chromosomal speciation and

molecular divergence-accelerated evolution in rearranged

chromosomes. Science 300: 321–324.

Nielsen, R. & Wakeley, J. 2001. Distinguishing migration from

isolation. A Markov chain Monte Carlo approach. Genetics 158:

885–896.

Noor, M.A.F., Grams, K.L., Bertucci, L.A. & Reiland, J. 2001a.

Chromosomal inversions and the reproductive isolation of

species. Proc. Natl Acad. Sci. USA 98: 12084–12088.

Noor, M.A.F., Grams, K.L., Bertucci, L.A., Almendarez, Y.,

Reiland, J. & Smith, K.R. 2001b. The genetics of reproductive

isolation and the potential for gene exchange between

Drosophila pseudoobscura and D. persimilis via backcross hybrid

males. Evolution 55: 512–521.

Posada, D. & Crandall, K.A. 1998. MODELTEST: testing the

model of DNA substitution. Bioinformatics 14: 817–818.

Prokopy, R.J., Bennett, E.W. & Bush, G.L. 1971. Mating

behavior in Rhagoletis pomonella (Diptera: Tephritidae). I. Site

of assembly. Canadian Entomologist 103: 1405–1409.

Prokopy, R.J., Bennett, E.W. & Bush, G.L. 1972. Mating

behavior in Rhagoletis pomonella. II. Temporal organization.

Canadian Entomologist 104: 97–104.

de Queiroz, K. 1998. The general lineage concept of species.

Species criteria, and the process of speciation: a conceptual

unification and terminological recommendations. In: Endless

Forms: Species and Speciation (D. Howard & S. H. Berlocher,

eds), pp. 57–78. Oxford University Press, Oxford.

Reissig, W.H. & Smith, D.C. 1978. Bionomics of Rhagoletis

pornonella in Crataegus. Ann. Entomol. Soc. Am. 71: 155–159.

Rieseberg, L.H. 2001. Chromosomal rearrangements and speci-

ation. Trends Ecol. Evol. 16: 351–358.

Rieseberg, L.H., Whitton, J. & Gardner, K. 1999. Hybrid zones

and the genetic architecture of a barrier to gene flow between

two sunflower species. Genetics 152: 713–727.

Roethele, J.B., Romero-Severson, J. & Feder, J.L. 2001. Evi-

dence for broad-scale conservation of linkage map relation-

ships between Rhagoletis pomonella and Drosophila melanogaster.

Ann. Entomol. Soc. Am. 94: 936–947.

Rull, J., Aluja, M., Feder, J.L. & Berlocher, S.H. 2006. The

distribution and host range of hawthorn-infesting Rhagoletis

pomonella (Diptera: Tephritidae) in Mexico. Ann. Entomol. Soc.

Am. 99: 662–672.

Saitou, N. & Nei, M. 1987. The neighbor-joining method: a new

method for reconstructing phylogenetic trees. Mol. Biol. Evol.

4: 406–425.

Schneider, S., Roessli, D. & Excoffier, L. 2000. Arlequin ver. 2.000:

A Software for Population Genetics Data Analysis.

Schwarz, D. & McPheron, B.A. 2007. When ecological isolation

breaks down: sexual isolation is an incomplete barrier to

hybridization between Rhagoletis species. Evol. Ecol. Res. 9:

829–841.

Schwarz, D., Matta, B.M., Shakir-Botteri, N.L. & McPheron, B.A.

2005. Host shift to an invasive plant triggers rapid animal

hybrid speciation. Nature 436: 546–549.

Schwarz, D., Shoemaker, K.D., Shakir, N.L. & McPheron, B.A.

2007. A novel preference for an invasive plant as a mechanism

for animal hybrid speciation. Evolution 61: 245–256.

Smith, D.C. 1986. Genetics and Reproductive Isolation of Rhagoletis

flies. PhD dissertation. University of Illinois at Urbana, 97 pp.

Smith, D.C. 1988. Heritable divergence of Rhagoletis pomonella

host races by seasonal asynchrony. Nature 336: 66–68.

Stump, A.D., Fitzpatrick, M.C., Lobo, N.F., Traore, S., Sagnon,

N.F., Costantini, C., Collins, F.H. & Besansky, N.J. 2005.

Centromere-proximal differentiation and speciation in Anoph-

eles gambiae. Proc. Natl Acad. Sci. 102: 15930–15935.

Swofford, D.L. 2002. PAUP*: Phylogenetic Analysis Using Parsimony

(*and Other Methods), Version 4.0b10. Sinauer Associates,

Sunderland, MA.

Tamura, K. & Nei, M. 1993. Model selection in the estimation of

the number of nucleotide substitutions. Mol. Biol. Evol. 10:

512–526.

Turner, T.L., Hahn, M.W. & Nuzhdin, S.V. 2005. Genomic islands

of speciation in Anopheles gambiae. PLoS Biol. 3: 1572–1578.

Wang, R.L., Wakeley, J. & Hey, J. 1997. Gene flow and natural

selection in the origin of Drosophila pseudoobscura and close

relatives. Genetics 147: 1091–1106.

Wu, C.-I. 2001. The genic view of the process of speciation.

J. Evol. Biol. 14: 851–865.

Xie, X., Rull, J., Michel, A.P., Velez, S., Forbes, A.A., Lobo, N.F.,

Aluja, M. & Feder, J.L. 2007. Hawthorn-infesting populations

912 X. XIE ET AL.

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y

of Rhagoletis pomonella in Mexico and speciation mode plural-

ity. Evolution 61: 1091–1105.

Supplementary material

The following supplementary material is available for this

article:

Table S1 Collecting sites for flies genetically analysed in

the study.

Figure S1 Parsimony gene tree for locus P181.

Figure S2 Parsimony gene tree for locus P667.

Figure S3 Parsimony gene tree for locus P7.

Figure S4 Parsimony gene tree for locus P661.

This material is available as part of the online article

from: http://www.blackwell-synergy.com/doi/abs/10.1111/

j.1420-9101.2008.01507.x.

Please note: Blackwell Publishing are not responsible

for the content or functionality of any supplementary

materials supplied by the authors. Any queries (other

than missing material) should be directed to the corres-

ponding author for the article.

Received 1 November 2007; revised 19 December 2007; accepted 19

December 2007

Radiation of Rhagoletis pomonella complex 913

ª 2 0 0 8 T H E A U T H O R S . J . E V O L . B I O L . 2 1 ( 2 0 0 8 ) 9 0 0 – 9 1 3

J O U R N A L C O M P I L A T I O N ª 2 0 0 8 E U R O P E A N S O C I E T Y F O R E V O L U T I O N A R Y B I O L O G Y