© 2006 Society of Systematic Biologists
Patterns of Speciation in the Yucca Moths: Parallel Species Radiations within the Tegeticula yuccasella Species Complex
Edited by Adrian Paterson: Associate Editor
1 Department of Biological Sciences, University of Idaho Moscow Idaho 83844–3051 USA
2 Department of Biology & Life Sciences Consortium, Pennsylvania State University University Park Pennsylvania 16802 USA
 |
Abstract |
|---|
 TopNotes Abstract MethodsResultsDiscussionConclusionsReferences |
|---|
The interaction between yuccas and yucca moths has been central to understanding the origin and loss of obligate mutualism and mutualism reversal. Previous systematic research using mtDNA sequence data and characters associated with genitalic morphology revealed that a widespread pollinator species in the genus Tegeticula was in fact a complex of pollinator species that differed in host use and the placement of eggs into yucca flowers. Within this mutualistic clade two nonpollinating "cheater" species evolved. Cheaters feed on yucca seeds but lack the tentacular mouthparts necessary for yucca pollination. Previous work suggested that the species complex formed via a rapid radiation within the last several million years. In this study, we use an expanded mtDNA sequence data set and AFLP markers to examine the phylogenetic relationships among this rapidly diverging clade of moths and compare these relationships to patterns in genitalic morphology. Topologies obtained from analyses of the mtDNA and AFLP data differed significantly. Both data sets, however, corroborated the hypothesis of a rapid species radiation and suggested that there were likely two independent species radiations. Morphological analyses based on oviposition habit produced species groupings more similar to the AFLP topology than the mtDNA topology and suggested the two radiations coincided with differences in oviposition habit. The evolution of cheating was reaffirmed to have evolved twice and the closest pollinating relative for one cheater species was identified by both mtDNA and AFLP markers. For the other cheater species, however, the closest pollinating relative remains ambiguous, and mtDNA, AFLP, and morphological data suggest this cheater species may be diverged based on host use. Much of the divergence in the species complex can be explained by geographic isolation associated with the evolution of two oviposition habits.
Keywords: AFLP; coevolution; host use; mtDNA; parallel species radiation; yucca moths
Received June 22, 2005; Revised September 21, 2005; Accepted November 23, 2005
Mutualistic interactions are part of the biotic foundation upon which many communities and ecosystems are based. These interactions can range from extremely specific relationships, such as those between endosymbionts and their hosts, to very generalized ones, such as between some plants and their pollinators (Stanton, 2003). Most of the conceptual advances in mutualism, however, have been based on a few model systems. Obligate pollination mutualisms, in particular, have served as focal systems for study. For example, obligate mutualisms between seed-eating pollinators and plants, such as yuccas and yucca moths and figs and fig wasps, have provided the empirical and theoretical basis for examining the benefits and costs, stability, and the exploitation of mutualism (Pellmyr and Huth, 1994; Anstett et al., 1996; Addicott, 1996, 1998; West et al., 1996; Herre, 1999; Herre et al., 1999; Huth and Pellmyr, 2000; Marr et al., 2001; Bronstein et al., 2003; Cook and Rasplus, 2003; Morris et al., 2003; Peng et al., 2005). Research on the ecology and evolution of these systems has affirmed that mutualism is a balanced antagonistic interaction.
Understanding how mutualisms originate and subsequently diversify requires a phylogenetic approach. The incorporation of molecular systematics into the study of obligate pollination mutualism has provided this approach and has had two important consequences. First, the use of a phylogenetic framework allows testing of hypotheses dealing with coevolution, correlated evolution, cospeciation, modes of speciation, and the evolution of cheating from within mutualistic lineages (Herre et al., 1996; Pellmyr and Leebens-Mack, 2000; Machado et al., 2001, 2005; West et al., 1996; Weiblen, 2004). Second, molecular data have shown that some of these supposedly species-poor systems are in reality complexes of cryptic species that differ little in morphology (Weiblen et al., 2001; Weiblen and Bush, 2002; Kato et al., 2003; Molbo et al., 2003). Moreover, these species complexes are sometimes the result of rapid radiations, which suggests that interactions among species may impact the rate at which these lineages diversify.
In this study, we examined the phylogenetic history of a rapid radiation within the pollinating yucca moths of the genus Tegeticula. Previous phylogenetic work based on mtDNA sequence data demonstrated that the geographically widespread pollinator T. yuccasella was, in fact, a species complex of pollinator and cheater moths (Pellmyr et al., 1996; Pellmyr and Leebens-Mack, 1999, 2000) that differed in host use and genitalic morphology. Although each of the species within the T. yuccasella complex as a whole were monophyletic based on the mtDNA sequence data, there was little statistical support for the species relationships within a clade containing seven pollinator species and the two cheater species. This clade originated in a burst of speciation that occurred 3.2 ± 1.8 Mya (Pellmyr and Leebens-Mack, 1999) and was concomitant with the colonization of capsular-fruited Yucca, the evolution of a new oviposition habit, and the evolution of cheater moth species that do not pollinate yucca flowers but still feed on yucca seeds. Most of the ecological diversity in Tegeticula arose during this radiation. A robust phylogeny of this group is necessary for testing hypotheses about coevolution, mutualism reversal, and cospeciation between yuccas and yucca moths. Here, we build on previous phylogenetic studies by expanding the mtDNA data set to include sequence data from another mtDNA gene, cytochrome b, and adding genome-wide nuclear markers via AFLPs. AFLPs have been useful in resolving the phylogenetic relationships in other very closely related groups of organisms (Albertson et al., 1999; Parsons and Shaw, 2001; Ogden and Thorpe, 2002; Deprés et al., 2003; Mendelson and Shaw, 2005). We explore the congruence and conflict between these two data sets to examine the evolution of oviposition habit, mutualism reversal, and the diversification of the yucca moths.
The Study System
The well-known obligate pollination mutualism between yuccas and yucca moths consists of the interactions of many yucca and pollinator moth species. The yuccas (Yucca and Hesperoyucca) contain about 40 species, with Yucca divided into three sections based on fruit type: the spongy-fruited section Clistocarpa, the fleshy-fruited section Sarcocarpa, and the capsular-fruited section Chaenocarpa. Vegetative form also differs among the host plants, ranging from the tall branching tree-like form of the Joshua tree, Y. brevifolia, to small rosettes such as in Y. harrimaniae. Irrespective of fruit type or vegetative form, all yucca species are actively pollinated by at least one species of yucca moth (Pellmyr, 2003).
The pollinator yucca moths consist of two genera, Parategeticula and Tegeticula. Much of the research examining the obligate mutualism has focused on Tegeticula. This North American genus contains a total of fourteen described pollinator species and two derived nonpollinating cheater species (Pellmyr, 1999, 2003). The interaction between the Tegeticula pollinators and yuccas is highly specific, with 11 of the 14 pollinators utilizing only one yucca species (Pellmyr, 1999, 2003). All of the pollinator species actively pollinate yucca flowers using specialized mouthparts, but they differ in oviposition habit. Female yucca moths deposit eggs in yucca flowers and the resulting larvae feed on a subset of the developing yucca seeds. There are two major oviposition habits among Tegeticula species. Locule-ovipositing species use a long and narrow ovipositor to cut through the flower ovary wall into the locule and lay eggs next to the ovules. This is the ancestral oviposition habit for the genus (Pellmyr and Leebens-Mack, 1999). In contrast, superficially ovipositing species use a short, thick ovipositor to place eggs within or just below the tissue surface. The larva then chews into the locule to feed on the developing ovules.
Previous phylogenetic studies based on mtDNA sequence data suggested that much of the species and life history diversity within Tegeticula occurred in a burst of speciation (Pellmyr and Leebens-Mack, 1999, 2000). The colonization of capsular-fruited Yucca, the shift in oviposition habit from locule oviposition to superficial oviposition, and the evolution of cheater moths all occurred during this radiation (Pellmyr and Leebens-Mack, 1999, 2000). Within this explosion, several well-supported clades were identified, but the relationships among them were uncertain because of short internal branches. Based on the well-supported clades, Pellmyr and Leebens-Mack (2000) concluded that oviposition habit was labile, and that the evolution of cheating occurred twice within the radiation. Here, we demonstrate that the species relationships as portrayed by mtDNA contrasts with those inferred from the nuclear loci. The nuclear markers showed support for two clades of moths that differ in oviposition habit, and the two cheater species evolved from within the "superficially ovipositing" clade. Differences in moth morphology also supported the two clades identified by the nuclear loci. Furthermore, the results suggest that shifts between oviposition habits may be more phylogenetically conservative than previously indicated.
|
Methods |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
Five individuals from each of the studied Tegeticula species were included in the analyses (Table 1). A new species, which we have given the working name ‘T. california’ was identified based on this study. Individuals were chosen to maximize coverage of the host range and geographic range of each moth species (Fig. 1). We used T. treculeanella as the outgroup based on previous work by Pellmyr and Leebens-Mack (1999, 2000). We also included T. mojavella in the analyses to further polarize the phylogeny. This pollinator is more closely related to the ingroup than T. treculeanella, and we wanted to be certain that the ingroup remained monophyletic with additional mtDNA sequence and AFLP data. We removed the head, wings, and genitalia from each adult moth to keep as a voucher. Total genomic DNA from the remaining thorax and abdomen was extracted using a modified protocol of Harrison et al. (1987) or the IsoQuick DNA Extraction kit (Orca Research Inc., Bothell, WA). For larvae, the entire specimen was extracted.
|
|
Sequencing Protocol
We sequenced two regions of the mtDNA for each of the taxa. The first region was 2104 contiguous bp that included the 3' end of cytochrome oxidase I, the intervening tRNA lysine, and the 5' end of cytochrome oxidase II. We also sequenced 919 contiguous bp of the cytochrome b gene. The cytochrome oxidase region was amplified with four pairs of PCR primers that produced overlapping regions of sequence. The primer pairs were 1461F-2302R, 2231F-3020R, 2638F-3306R, and 3252F-3771R and the numbers refer to the nucleotide positions in the Drosophila yakuba mtDNA genome (Clary and Wolstenholme, 1985). The cytochrome b region was amplified with two pairs of overlapping PCR primers: cytB1-cytB2 and cytB3-cytB4. (Primer sequences are available upon request from the authors.) The thermal cycler profile for amplification of both regions was 95°C for 2 min, 35 cycles of 95°C for 1 min, 52°C for 45 s, 72°C for 1 min 30 s, and 72°C for 10 min. Amplicons were cleaned with the Qiagen PCR purification kit (Qiagen, Valencia, CA) before being used in cycle sequencing reactions. Sequencing reactions consisted of 4 µ l of DNA product, 2 µ l of ABI Big Dye Terminator Sequencing Mix, 2 µ l of 2 µ M primer, and 2 µ l of sequencing buffer. Sequencing products were cleaned using Centri-Sep sephadex columns (Princeton Separations, Inc.), lyophilized, and resuspended in 2.5 µ l of formamide and loading dye from Applied Biosystems. Products were sequenced on an ABI 377 automated DNA sequencer and analyzed using Sequencher 3.1 (Gene Codes Corp., Ann Arbor, MI).
AFLP Protocol
We also generated AFLP markers on an additional set of samples representing each of the species. Initial AFLP runs with the individuals used for mtDNA sequencing produced small numbers of fragments and had low repeatability. This was an artefact of low-quality nuclear DNA, even though mtDNA amplification and sequencing was successful. Because many of the samples were extracted 5+ years ago, the nuclear DNA had degraded, even though the mtDNA had remained stable. We extracted DNA using the IsoQuick DNA Extraction kit (Orca Research Inc.) from an additional set of samples representing each of the species to obtain high quality DNA for the AFLP procedure. As for sequencing, individuals were chosen to maximize coverage of the host range and geographic range of each moth species. None of the individuals sequenced were used in the AFLP analyses. We used a modified protocol of the Applied Biosystems Plant Genome kit developed by M. Gitzendanner (personal communication) to conduct the AFLP analysis. Restriction and ligation reactions were carried out in a single step. Genomic DNA was digested for three hours at 37°C with 5 units of EcoRI (Promega, Madison, WI), 0.5 units of MseI (New England Biolabs, Beverly, MA) in 11 µ l reaction volumes containing 0.15 units of T4 DNA ligase (Promega), 1 x T4 Ligase buffer (Promega), 0.05 M NaCl, 0.55 µ g BSA, 4.5 µ M MseI adapter (5'-GACGATGAGTCCTGAG-3' and 5'-TACTCAGGACTCAT-3'), and 0.45 µ M EcoRI adapter (5'-CTCGTAGACTGCGTACC-3' and 5'-AATTGGTACGCAGTCTAC-3'). Reactions were diluted by a factor of 20 in 1 x TE0.1 (20 mM Tris-HCl, 0.1 mM EDTA, pH 8.0), and followed by two rounds of selective amplification. The first selective amplification was conducted in 20-µl reaction volumes containing 4 µ l of the diluted restriction-ligation reaction, 1 unit Taq DNA polymerase (Sigma Chemical Co., St. Louis, MO), 10 x PCR buffer (10 mM Tris-HCl, 50 mM KCl, pH 8.3), 3.5 mM MgCl2, 0.8 mM dNTPs, 0.3 µ M EcoRI +1 selective primer (5'-GACTGCGTACCAATTCA-3'), and 0.3 µ M MseI +1 selective primer (5'-GACGATGAGTCCTGAGTAAC-3'). Reactions were heated to 72°C for 2 min, then cycled 20 times at 94°C for 30 s, 56°C for 30 s, 72°C for 120 s, and then held at 60°C for 30 min. These reactions were diluted by a factor of 14 in 1 x TE0.1 and used in the final selective amplification step.
The final amplification was performed in 10-µ l reactions containing 2.5 µ l dilute +1 PCR product, 0.5 units Amplitaq Gold DNA polymerase (Applied Biosystems), 1 x Amplitaq PCR Buffer (Applied Biosystems), 3 mM MgCl2, 0.8 mM dNTPs, 0.05 µ M of each EcoRI +3 primer (5'-[5-TET] GACTGCGTACCAATTCAAC-3'; 5'-[6-FAM] GACTGCGTACCAATTCACA-3'; [5-HEX]GACTGCGTACCAATTCAAG), and 0.125 µ M MseI +3 primer (5'-GACGATGAGTCCTGAGTAACAG-3'). The EcoRI +3 primers were fluorescently labeled for visualization on an ABI 377 sequencer. Because each primer was labeled with a different wavelength of dye, single reactions contained three primer combinations (i.e., each EcoRI +3 primer with the MseI +3 primer). The reactions were held at 94°C for 2 min, then cycled 10 times starting at 94°C for 30 s, 65°C for 30 s, 72°C for 2 min, with a reduction in the annealing temperature by 1°C per cycle. Reactions were then cycled 36 times at 94°C for 30 s, 56°C for 30 s, 72°C for 120 s, followed by a 30-min 60°C hold. A total of 1.2 µ l of loading dye (63% deionized formamide, 27% ABI blue loading dye, and 10% Tamra 500 size standard) from Applied Biosystems plus 1.5 µ l of the +3 PCR product was loaded onto a 5% acrylamide gel. Electrophoresis was conducted on an ABI 377 sequencer using the GeneScan2400 run parameters. Fragment sizes were assigned using the Tamra 500 size standard and the GeneScan software version 3.1.2 (Applied Biosystems). We used Genotyper 2.5 (Applied Biosystems) to develop a set of AFLP markers to score for all individuals. Scoring of AFLP markers was conducted in two steps. First, we screened every individual for markers whose peak fluorescence was 300 fluorescent units or above. Each marker greater than 300 fluorescent units was added to a marker database. We then rescreened with this composite marker database to determine which subset of markers was present in each individual. A cutoff threshold of 25 fluorescent units or above was used for scoring the presence of a marker in this second step. The default detection setting on the GeneScan data collection software (Applied Biosystems) is arbitrarily set at 50 fluorescent units. Because there is always variation in overall fluorescent intensities across runs, we chose a lower threshold to increase the likelihood that a band would be scored as present. That is, we wanted to reduce the chance of missing the presence of a band because of variation in fluorescence intensities that occur due to electrophoresis conditions. This procedure would not bias the pattern of relatedness because all samples would be scored with the same threshold value.
Phylogenetic Analyses
The mtDNA sequence data were analyzed using maximum likelihood following the procedures and recommendations in Sullivan (2005). We first determined the model of sequence evolution using the DT-ModSel program (Minin et al., 2003). This procedure is based on the Bayesian information criterion and incorporates relative branch-length error when choosing a model of sequence evolution. The model was used in a heuristic search with random addition of taxa, 10 replicate searches, and TBR branch swapping in PAUP 4.0b10 (Swofford, 2002). We performed an absolute goodness of fit test (Goldman, 1993) to determine whether the model chosen by DT-ModSel was appropriate. To conduct this test, the difference in the unconstrained likelihood and the likelihood from the heuristic search was recorded. The resulting tree and parameter estimates of the model were used in Seq-Gen 1.2.7 (Rambaut and Grassly, 1997) to generate 100 replicate data sets that were used in a subsequent parametric bootstrap test. We then performed a maximum likelihood search on each replicate data set to compare the unconstrained likelihood score with the score from the heuristic search. These searches were performed on a 108-node Beowulf cluster at the University of Idaho. The difference between the unconstrained and constrained likelihoods for each replicate was used to generate a null distribution. We then compared the difference between the unconstrained likelihood and the constrained likelihood for the original data set against this distribution. Once the model was confirmed to be appropriate for the sequence data, we ran another heuristic search via maximum likelihood and used the nonparametric bootstrap procedure (Felsenstein, 1985) to assess support for the nodes in the resulting tree topology.
The presence/absence of the AFLP markers were converted to the Nei-Li distance metric (Nei and Li, 1979) in PAUP 4.0b10. We chose this measure as a means to provide a composite index of the signal from all AFLP markers rather than assessing the influence of individual markers. A heuristic search using the minimum evolution criterion was then performed, and support for nodes in the resulting topology was assessed via 100 nonparametric bootstrap replicates. Because the mtDNA and AFLP analyses yielded different topologies, we used a parametric bootstrap approach to test if the topologies were statistically different (Huelsenbeck et al., 1996). We constrained the mtDNA data set to be congruent with the AFLP topology and performed another maximum likelihood analysis. The resulting tree and parameter estimates were used in Seq-Gen 1.2.7 to generate 100 replicate data sets. The likelihood score for the unconstrained search was compared to the likelihood for the AFLP constrained search for each of the data sets to generate the null distribution. We did not conduct a combined analysis of the mtDNA sequence data and the AFLP markers because we had to use different subsets of individuals to collect each data set and the geographic localities sampled within species were not entirely consistent between the two sets of taxa. Also, there was more than nine times the number of characters for the sequence data compared to the AFLP data which would have swamped out the phylogenetic signal from the AFLP markers.
Morphometric Measurements
Pellmyr (1999) measured 12 female and 11 male continuous morphological traits that were used in designating species within the T. yuccasella complex. We made the same measurements for 3 females and 2 males of ‘T. california’ and used data from Pellmyr (1999) for the other species. Forewing length and width were measured for both sexes. For females we also measured ovipositor traits (length of apophyses anteriores and posteriores, keel height, keel length, ovipositor tip to keel, and ovipositor height) and other reproductive traits (signum diameter, corpus bursae length, corpus bursae width, corpus bursae plus duct length). For males we measured aedeagus length and width, vinculum-saccus length, the number of spines in both pectinifers, and quantified valva shape by measuring distance from valva base to pectinifer, distance from pectinifer to valva apex, distance from valva base to apex, the valval crescent width, and valva crescent depth. We analyzed differences for both sexes among the ingroup using principal components analysis in JMP 5.0.1.2
[EC]
(SAS Institute). Sample sizes for each species were 5 individuals per sex except for ‘T. california,’ T. elatella (2 females), T. rostratella (3 males), T. altiplanella (4 males), and T. superficiella (4 males). We used ANOVA to test if the first two principal components were significantly different between the species groupings identified by the AFLP analyses.
|
Results |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
mtDNA Sequence Data
For each individual, we generated 3021 bp of mtDNA sequence data—2102 bp of cytochrome oxidase I, the intervening tRNA leucine and the 5' end of cytochrome oxidase II, and 919 bp of cytochrome b. Sequences were deposited in GenBank under the accession numbers in Table 1 and the data matrix was deposited in TreeBase. There was a single nucleotide insertion for T. altiplanella in the tRNA leucine and an indel in the tRNA for T. treculeanella and T. mojavella. The nucleotide insertion and the indel position were not used in the phylogenetic analyses because the former was an autapomorphy for T. altiplanella and the latter only occurred in the outgroup taxa. The greatest sequence divergence between species in the ingroup was 2.35% (uncorrected p). DT-ModSel returned the HKY+G+I model to describe sequence evolution for the taxa. The absolute goodness-of-fit test did not reject the null hypothesis of a perfect fit between model and data (Fig. 2).
|
The maximum likelihood analysis of the mtDNA sequence data produced a single tree with a score of –lnL = 7341.001 (Fig. 3). This tree supported monophyly for each of the moth species with the exception of the pollinator-cheater sister pair T. cassandra and T. intermedia (bootstrap values for species monophyly were 85 or above). There was strong bootstrap support for two major clades within the ingroup. One clade contained T. cassandra, T. intermedia, and T. yuccasella, and the second clade contained the other seven species. Branch lengths at the base of these two clades were much shorter than the branch lengths leading to each species.
|
AFLP/nDNA Data
The AFLP procedure generated 352 markers. The AFLP data matrix was deposited in TreeBase. We tested repeatability of the markers by running replicate samples on four individuals. Repeatability was 99.4%. The resulting minimum evolution tree identified two major clades of moths, but the composition of these clades was different than those inferred from the mtDNA data (Fig. 3). One clade comprised the three locule-ovipositing species, T. altiplanella, T. baccatella, and T. rostratella, which was found in the mtDNA analysis, but this clade also included the locule-ovipositing T. yuccasella. The phylogenetic relationships within this clade were well supported.
The second clade contained the three superficially ovipositing pollinators, the locule ovipositing pollinator ‘T. california,’ and the two cheater species. Species relationships within this clade were unresolved except for the pollinator-cheater sister pair T. cassandra and T. intermedia. The AFLP data also suggested that the locule-ovipositing cheater T. corruptrix may be genetically differentiated into two entities based on feeding on capsular or fleshy-fruited yuccas. A similar pattern was also seen in the mtDNA data, although T. corruptrix was monophyletic.
Morphological Data
The PCA of both female and male morphology showed patterns among the moth species that were very similar to the AFLP topology. The four locule-ovipositing species, T. altiplanella, T. baccatella, T. rostratella, and T. yuccasella, were very similar to one another and distinct from the "superficial and cheater" species (Fig. 5). Interestingly, ‘T. california’ was more similar to the superficially ovipositing taxa even though it is a locule-ovipositing species. The first two principal components were significantly different between the locular and superficial and cheater clades for females (PC 1, F1,49 = 7.003, P = 0.011; PC 2, F1,49 = 78.816, P < 0.0001) and males (PC 1, F1,49 = 87.681, P < 0.0001; PC 2, F1,49 = 6.871, P = 0.011). In terms of differences in female morphology, PC1 and PC2 together explained 81.5% of the variation. The loadings on these two components suggested that all of the measured traits are important in explaining the variation between the two groups (Table 2). No single trait had a disproportionately large eigenvector for either principal component. For PC1, ovipositor length (apophyses posteriores and anteriores) and the other components of reproductive morphology (signum diameter, corpus bursae length, corpus bursae width, corpus bursae plus duct length) were contrasting. For PC2, body size and ovipositor traits were most important. Morphological differences among males were influenced by body size, aedeagus length, and valval size for PC1, and vinculum-saccus length and aedeagus length for PC2 (Table 2). The late cheater, T. corruptrix, exhibited two distinct morphological clusters that reflected individuals that use capsular and fleshy-fruited yuccas, respectively.
|
|
|
Conflict between Molecular Data Sets
Because the AFLPs and morphology presented a similar picture of moth relationships, we tested whether the mtDNA topology was in conflict with these data. We used a parametric bootstrap to test if the mtDNA topology was statistically different from the AFLP topology, by constraining the mtDNA sequence data to the AFLP topology. The likelihood of the unconstrained mtDNA tree was –lnL = 7341.001 and the likelihood of the tree constrained to the AFLP topology was –lnL = 7355.108. The difference of 14.107 likelihood units was well outside of the probability of finding this difference by chance (Fig. 6). Thus, the organellar DNA yielded a significantly different topology than the nuclear loci.
|
|
Discussion |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
Species Radiation and Conflict between Data Sets
Some of the most striking cases of ecological diversity in evolutionary lineages are the result of species radiations (Schluter, 2000). Adaptation to different environments produces combinations of ecologically important traits that generate reproductive isolation among populations (McKinnon et al., 2004). In many radiations, adaptation is rapid and causes speciation events to happen in quick succession. As a consequence, resolving the phylogenetic relationships among rapidly radiating taxa is difficult, and studies that have incorporated both organellar and nuclear genes have frequently documented phylogenetic conflict (Shaw, 2002; Seehausen et al., 2003; Sullivan et al., 2004). In some cases, the combination of multiple genes still does not resolve species relationships and the resulting phylogenetic tree may contain a hard polytomy (sensu Madison, 1989).
For the yucca moths, previous phylogenetic research based on mtDNA sequence data indicated that this group of mutualists and cheaters had undergone recent and rapid diversification (Pellmyr and Leebens-Mack, 1999, 2000). In the present study, the mtDNA sequence data and AFLP nuclear markers gave somewhat differing pictures of the evolutionary relationships, but both supported the idea that speciation in this group has happened in quick succession. One phylogenetic characteristic of a species radiation is that internal branches are relatively short compared to terminal branches (e.g., Jackman et al., 1999; Fishbein and Soltis, 2004). This pattern was observed in both of the molecular data sets. For the mtDNA topology, the internal branches (i.e., branches uniting two or more species) were significantly shorter than the branches leading to terminal taxa (F1,19 = 5.58, P = 0.0296). The same was true for the AFLP topology (F1,21 = 396.39, P < 0.0001), although this result may be slightly biased by the large number of AFLP markers that were autapomorphies.
A second characteristic of a rapid species radiation is that phylogenetic relationships within major clades may be poorly supported or unresolved even though species are monophyletic. Again, this was true for both data sets. Of the 10 species in the ingroup, 6 were monophyletic for both mtDNA and nuclear markers. The pollinator T. superficiella and the cheater T. corruptrix were both monophyletic in at least one data set. Tegeticula superficiella was monophyletic based on the AFLP data but not the mtDNA data, whereas the converse was true for T. corruptrix. Moreover, both molecular data sets suggested that there was intraspecific divergence in T. corruptrix based on host use of capsular and fleshy-fruited yucca species. The pollinator-cheater sister pair T. cassandra–T. intermedia was the only species pair that was not monophyletic in at least one data set. Previous research has demonstrated that these two species were closest relatives and have had limited hybridization between them (Segraves and Pellmyr, 2004); thus, a combination of incomplete lineage sorting and hybridization may explain the lack of monophyly for this clade. Taken together, the pattern of short internal branches, poor resolution at the base of clades, and species monophyly are indicative of speciation events that happened in rapid succession during the early stages of this radiation.
Although both the mtDNA and nuclear data sets supported a rapid radiation of moths, there were differences in the phylogenetic relationships within the radiation. The mitochondrial and nuclear data sets produced significantly different topologies. Both data sets supported two major clades of moths, but the species composition of these clades differed. Which molecular marker data set is presenting a more accurate estimate of the species tree is unclear. There are two methodological reasons that could result in disparate estimates of the moth phylogenies—sampling and algorithm differences. Because we had to use a different set of individuals for the mtDNA and AFLP analyses, sampling could have impacted the phylogenies. That is, data from different individuals would contain different phylogenetic signal, especially in terms of species relationships. This problem has shown to be important by Funk and Omland (2003) when species are highly polyphyletic. The monophyly of many species for both data sets, however, suggests that sampling different individuals within species would not produce different topologies. Each individual within a species would likely contain the same pattern of genetic divergence with respect to other species surveyed. Due to the lack of models for AFLP data, we analyzed the AFLP data with minimum evolution rather than maximum likelihood as for the sequence data. The difference in the search algorithm could result in different topologies. We investigated this possibility by converting the mtDNA sequence data to the HKY85 distance measure and conducting a search using minimum evolution. The minimum evolution search yielded a topology identical to the maximum likelihood search (data not shown), which suggests that there are differences in phylogenetic signal between the two data sets. Neither sampling bias nor search algorithm appears to be a likely cause of the topological differences between the marker sets.
Studies that have surveyed both organellar and nuclear markers in other rapidly radiating lineages have also demonstrated conflict in the phylogenies produced by each data set (e.g., Shaw, 2002; Sullivan et al., 2004). There are three reasons tentatively to assume that the AFLP topology may be more accurate. First, there may not have been enough time during the early stages of the radiation for sibling taxa to obtain reciprocal monophyly for mtDNA prior to the next speciation event (i.e., incomplete lineage sorting). As a result, the distribution of the ancestral mtDNA haplotypes across the newly formed species was likely somewhat random rather than consistent with the true sequence of speciation. Once the species were formed, however, each species remained distinct and became monophyletic for the mtDNA inherited at the early stage of the radiation. Thus, species are currently monophyletic, but for ancestral mtDNA haplotypes that do not reflect the pattern of speciation. Second, hybridization with pollinators has been identified at least two times in the cheater T. intermedia (Segraves and Pellmyr, 2004; Segraves et al., 2005). Hybridization events, either early in the radiation or more recently, would likely significantly affect the mtDNA phylogeny (Ballard and Whitlock, 2004). For example, hybridization might result in capture of mtDNA from a distantly related moth lineage. Thus the organellar phylogeny would not be indicative of the species phylogeny. For AFLP markers, incomplete lineage sorting is also possible; however, the likelihood that the same pattern of sorting would occur across the many different markers is low. Instances of hybridization would result in a merging of the genome rather than wholesale transfer from one species to another as is possible for mtDNA. Finally, differences in genitalic morphology among moth species are more consistent with the AFLP topology than the mtDNA topology (see below).
Morphology and Life History Evolution
Differences in oviposition habit within Tegeticula have important consequences for our understanding of diversification within this genus. Morphology of both sexes varies with oviposition habit (Fig. 5). Females of locule-ovipositing species have a long, narrow ovipositor for depositing eggs into the locule next to developing yucca ovules or seeds (Pellmyr, 1999). In contrast, the ovipositors of superficially ovipositing females are short and wide. Analyses of morphology produced species groupings consistent with the major clades depicted by the AFLP topology. For males and females, the locule-ovipositing species T. rostratella, T. altiplanella, T. baccatella, and T. yuccasella were distinct from the superficially ovipositing species. The two largest locule-ovipositing species, T. corruptrix and ‘T. california,’ were more morphologically similar to the superficially ovipositing species as was also suggested genetically by the AFLP data. Within the cheater species T. corruptrix there was a strong difference in morphology associated with utilizing fleshy-fruited or capsular-fruited Yucca species. This distinctiveness is also mirrored in both molecular data sets; however, more data are required to test whether there is genetic differentiation partitioned by host use in this cheater species.
The number of evolutionary transitions between the two broad categories of oviposition habits differs between the phylogenies produced by the mitochondrial and nuclear data sets. For the mtDNA phylogeny, superficial oviposition evolved at least twice—once for the pollinators T. superficiella and T. elatella, and once for the pollinator-cheater sister pair, T. cassandra and T. intermedia. The AFLP topology, however, suggests a single origin for all of the superficially ovipositing species. Locule-oviposition is the ancestral condition within Tegeticula (Pellmyr and Leebens-Mack, 1999, 2000), which suggests that ‘T. california’ and T. corruptrix may be basal in this superficial clade, and that superficial oviposition evolved just once. Additional phylogenetic resolution is required to fully evaluate this hypothesis.
Pellmyr and Leebens-Mack (2000) used mtDNA sequence data to show that the transition from pollinating to cheating has occurred twice within the radiation, and that both cheaters evolved from superficially ovipositing pollinators. The expanded data set used in the present study corroborates two origins of cheating. For the cheater T. intermedia, both mtDNA and AFLP markers identified the superficially ovipositing pollinator T. cassandra as the closest relative. The sister species to the locule-ovipositing cheater, T. corruptrix, however, remains unclear. Pellmyr and Leebens-Mack (2000) suggested that the superficially ovipositing pollinators T. superficiella and T. elatella were sibling species to T. corruptrix. Based on the current data, we were unable to identify the closest pollinator species for this cheater.
Diversification and Coexistence of Moth Species
The differences in oviposition habit are also important in understanding species distributions and the pattern of speciation within Tegeticula. The AFLP and morphological data suggest that there have been two radiations of pollinator species—one locule-ovipositing and one mainly superficially ovipositing clade. Where pollinator species co-occur, the moths differ in oviposition habit. One possible explanation for this pattern is that differences in oviposition habit translate into differences in reproductive morphology. These differences may prevent interspecific hybridization and allows species differing in oviposition habit to coexist, whereas species with similar reproductive morphology would hybridize when in sympatry. In cases where species with similar oviposition habits do come in contact, hybridization has occurred. Segraves and Pellmyr (2004) and Segraves et al. (2005) have documented that the superficially ovipositing cheater T. intermedia has hybridized with the superficially ovipositing pollinators T. cassandra and T. elatella.
The current geographic distributions of the superficially ovipositing pollinator species are allopatric, and this pattern is true for the locule-ovipositing pollinator species with the exception of T. baccatella (Fig. 1). Tegeticula baccatella, however, uses a fleshy-fruited yucca species in contrast to the sympatric pollinators which use capsular-fruited yuccas from an entirely different Yucca section. Thus, host use may also serve as another means for isolating moth lineages (Leebens-Mack et al., 1998). The phylogenetic results and the current species distributions suggest that geographic separation has been very important in facilitating speciation rather than switches between oviposition habits within moth lineages.
|
Conclusions |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
Members of the T. yuccasella species complex comprise a recent species radiation that includes the evolution of a new oviposition habit and two origins of cheating species. Phylogenetic analyses of mtDNA and AFLP markers provided different evolutionary histories of this radiation. Most species were monophyletic for both data sets, but the composition of major clades varied between the data sets. Morphological data corroborated the AFLP data set and together these data suggest that there have been two parallel radiations within the species complex—one involving locule-ovipositing species and the other involving superficially ovipositing species. Discrepancies between the mtDNA and AFLP data sets are potentially the result of the lack of lineage sorting of the mtDNA markers during the early phase of the radiation and/or hybridization. Differences in oviposition habit may be an important mechanism isolating species lineages within this radiation.
|
Acknowledgements |
|---|
We thank Ron Leuschner and Dan Udovic for providing specimens used in the analyses, and Manuel Balcázar-Lara for invaluable help during the fieldwork in Mexico. Big Bend National Park, Black Gap Wildlife Management Area, Mojave National Preserve, Jornada Long-Term Ecological Research site, Laguna Atascosa National Wildlife Refuge, Torrey Pines State Park, Florida Park Service, Ocala National Forest, Archbold Biological Station, and the town of Cuatro Ciénegas kindly provided permits and access to field sites. The project was funded by NSF grants DEB 0321293 (DMA, OP), DEB 0242783, and DEB 0242778 and REU supplements (OP); NSF dissertation improvement grant DEB 0242776 to KAS; and NSF grant DEB 0234417 (JLM); and two travel grants to OP from National Geographic Society for fieldwork in the USA and Mexico, respectively. Jack Sullivan provided expert advice on the analyses and very helpful comments on the manuscript.
|
Notes |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
3 Current Address: Department of Biology, Syracuse University, Syracuse, New York 13244, USA; E-mail: dmalthof{at}syr.edu (D.M.A.)
|
References |
|---|
|
TopNotesAbstractMethodsResultsDiscussionConclusionsReferences |
|---|
-
Addicott J. F. Cheaters in yucca/moth mutualism. Nature (1996) 380:114–115.[CrossRef][Web of Science]
Addicott J. F. Regulation of the mutualism between yuccas and yucca moths: Population level processes. Oikos (1998) 81:119–129.[CrossRef][Web of Science]
Albertson R. C., Markert J. A., Danley P. D., Kocher T. D. Phylogeny of a rapidly evolving clade: The cichlid fishes of Lake Malawi, East Africa. Proc. Natl. Acad. Sci. USA (1999) 96:5107–5110.
Anstett M. C., Bronstein J. L., Hossaert-McKey M. Resource allocation: A conflict in the fig/fig wasp mutualism? J. Evol. Biol. (1996) 9:417–428.[CrossRef][Web of Science]
Ballard J. W., Whitlock M. C. The incomplete natural history of mitochondria. Mol. Ecol. (2004) 13:729–744.[CrossRef][Medline]
Bronstein J. L., Wilson W. G., Morris W. F. Ecological dynamics of mutualist/antagonist communities. Am. Nat. (2003) 162:S24–S39.[CrossRef][Web of Science][Medline]
Clary D. O., Wolstenholme D. R. The mitochondrial DNA molecule of Drosophila yakuba: Nucleotide sequence, gene organization, and genetic code. J. Mol. Evol. (1985) 22:252–271.[CrossRef][Web of Science][Medline]
Cook J. M., Rasplus J. Mutualists with attitude: Coevolving fig wasps and figs. Trends Ecol. Evol. (2003) 18:241–248.[CrossRef]
Deprés L., Gielly L., Redoutet W., Taberlet P. Using AFLP to resolve phylogenetic relationships in a morphologically diversified plant species complex when nuclear and chloroplast sequences fail to reveal variability. Mol. Phylo. Evol. (2003) 27:185–196.[CrossRef][Web of Science][Medline]
Felsenstein J. Confidence limits on phylogenies: An approach using the bootstrap. Evolution (1985) 39:783–791.[CrossRef][Web of Science]
Fishbein M., Soltis D. E. Further resolution of the rapid radiation of Saxfrigales (Angiosperms, Eudicots) supported by mixed-model bayesian analysis. Syst. Bot. (2004) 29:883–891.[CrossRef]
Funk D. J., Omland K. E. Species-level paraphyly and polyphyly: Frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annu. Rev. Ecol. Syst. (2003) 34:397–423.[CrossRef][Web of Science]
Goldman N. Statistical tests of models of DNA substitution. J. Mol. Evol. (1993) 36:182–198.[CrossRef][Web of Science][Medline]
Herre E. A. Laws governing species interactions? Encouragement and caution from figs and their associates. In: Levels of selection in evolution—Keller L., ed. (1999) Princeton, New Jersey: Princeton University Press. Pages 209–237.
Herre E. A., Knowlton N., Mueller U. G., Rehner S. A. The evolution of mutualisms: Exploring the paths between conflict and cooperation. Trends Ecol. Evol. (1999) 14:49–53.[CrossRef][Medline]
Herre E. A., Machado C. A., Bermingham E., Nason J. D., Windsor D. M., McCafferty S. S., Houten W. V., Bachmann K. Molecular phylogenies of figs and their pollinator wasps. J. Biogeogr. (1996) 23:521–530.[CrossRef]
Huelsenbeck J. P., Hillis D. M., Jones R. Parametric bootstrapping in molecular phylogenetics: Applications and performance. In: Molecular zoology: Advances, strategies and protocols—Ferraris J. D., Palumbi S. R., eds. (1996) New York: John Wiley & Sons. Pages 19–45.
Huth C. J., Pellmyr O. Pollen-mediated selective abortion in yuccas and its consequences for the plant-pollinator mutualism. Ecology (2000) 81:1100–1107.[Web of Science]
Jackman T. R., Larson A., De Queiroz K., Losos J. B. Phylogenetic relationships and tempo of early diversification in Anolis lizards. Syst. Biol. (1999) 48:254–288.[Abstract]
Kato M., Takimura A., Kawakita A. An obligate pollination mutualism and reciprocal diversification in the tree genus Glochidion (Euphorbiaceae). Proc. Nat. Acad. Sci. USA (2003) 100:5264–5267.
Leebens-Mack J., Pellmyr O., Brock M. Host specificity and genetic structure of two yucca moth species in a yucca hybrid zone. Evolution (1998) 52:1376–1382.[CrossRef][Web of Science]
Machado C. A., Jousselin E., Kjellberg F., Compton S. G., Herre E. A. Phylogenetic relationships, historical biogeography and character evolution of fig-pollinating wasps. Proc. R. Soc. Lond. Series B Biol. Sci. (2001) 268:1–10.[Medline]
Machado C. A., Robbins N., Gilbert M. T. P., Herre E. A. A critical review of host specificity and its coevolutionary implications in the fig/fig wasp mutualism. Proc. Natl. Acad. Sci. USA (2005) 102:6558–6565.
Maddison W. P. Reconstructing character evolution on polytomous cladograms. Cladistics (1989) 5:365–377.[CrossRef][Web of Science]
Marr D. L., Brock M. T., Pellmyr O. Coexistence of mutualists and antagonists: Exploring the impact of cheaters on the yucca-yucca moth mutualism. Oecologia (2001) 128:454–463.[CrossRef][Web of Science]
McKinnon J. S., Mori S., Blackman B. K., Kingsley L. D. D., Jamieson L., Chou J., Schluter D. Evidence for ecology's role in speciation. Nature (2004) 429:294–298.[CrossRef][Medline]
Mendelson T. C., Shaw K. L. Sexual behaviour: Rapid speciation in an arthropod. Nature (2005) 433:375–376.[CrossRef][Medline]
Minin V., Abdo Z., Joyce P., Sullivan J. Performance-based selection of likelihood models for phylogeny estimation. Syst. Biol. (2003) 52:674–683.
Molbo D., Machado C. A., Sevenster J. G., Keller L., Herre E. A. Cryptic species of fig-pollinating wasps: Implications for the evolution of the fig-wasp mutualism, sex allocation, and precision of adaptation. Proc. Natl. Acad. Sci. USA (2003) 100:5867–5872.
Morris W. F., Bronstein J. L., Wilson W. G. Three-way coexistence in obligate mutualist-exploiter interactions. Am. Nat. (2003) 161:860–875.[CrossRef][Web of Science][Medline]
Nei M., Li W. H. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA (1979) 76:5269–5273.
Ogden R., Thorpe R. S. The usefulness of amplified fragment length polymorphism markers for taxon discrimination across graduated fine evolutionary levels in Caribbean Anolis lizards. Mol. Ecol. (2002) 14:437–445.
Parsons Y. M., Shaw K. L. Species boundaries and genetic diversity among Hawaiian crickets of the genus Laupala identified using amplified fragment length polymorphism. Mol. Ecol. (2001) 10:1765–1772.[CrossRef][Medline]
Pellmyr O. Systematic revision of the yucca moths in the Tegeticula yuccasella complex (Lepidoptera: Prodoxidae) north of Mexico. Syst. Entomol. (1999) 24:243–271.[CrossRef]
Pellmyr O. Yuccas, yucca moths and coevolution: A review. Ann. Mo. Bot. Gard. (2003) 90:35–55.[CrossRef]
Pellmyr O., Huth C. J. Evolutionary stability of mutualism between yuccas and yucca moths. Nature (1994) 372:257–260.[CrossRef][Web of Science]
Pellmyr O., Leebens-Mack J. Forty million years of mutualism: Evidence for Eocene origin of the yucca-yucca moth association. Proc. Natl. Acad. Sci. USA (1999) 96:9178–9183.
Pellmyr O., Leebens-Mack J. Reversal of mutualism as a mechanism for adaptive radiation in yucca moths. Am. Nat. (2000) 156:S62–S76.[CrossRef][Web of Science]
Pellmyr O., Leebens-Mack J., Huth C. J. Non-mutualistic yucca moths and their evolutionary consequences. Nature (1996) 380:155–156.[CrossRef][Medline]
Peng Y. Q., Yang D. A., Wang Q. Y. Quantitative tests of interactions between pollinating and non-pollinating fig wasps of dioecious Ficus hispida. Ecol. Entomol. (2005) 30:70–77.[CrossRef][Web of Science]
Rambaut A. E., Grassly N. C. SEQ-GEN: An application for the Monte Carlo simulation of DNA sequence evolution along phylogenetic trees. Comp. Appl. Biosci. (1997) 13:235–238.[Medline]
Schluter D. The ecology of adaptive radiation (2000) Oxford, UK: Oxford University Press.
Seehausen O., Koetsier E., Schneider W. V. Nuclear markers reveal unexpected genetic variation and a Congolese-Nilotic origin of the Lake Victoria cichlid species flock. Proc. R. Soc. Lond. Series B Biol. Sci. (2003) 270:129–137.[Medline]
Segraves K. A., Althoff D. M., Pellmyr O. Limiting cheaters in mutualism: Evidence from hybridization between mutualist and cheater yucca moths. Proc. R. Soc. Lond. Series B Biol. Sci. (2005) 272:2195–2201.[CrossRef][Medline]
Segraves K. A., Pellmyr O. Testing the "Out of Florida" hypothesis on the origin of cheating in the yucca-yucca moth mutualism. Evolution (2004) 58:2266–2279.[Web of Science][Medline]
Shaw K. L. Conflict between nuclear and mitochondrial DNA phylogenies of a recent species radiation: What mtDNA reveals and conceals about modes of speciation in Hawaiian crickets. Proc. Natl. Acad. Sci. USA (2002) 99:16122–16127.
Stanton M. L. Interacting guilds: Moving beyond the pair-wise perspective on mutualisms. Am. Nat. (2003) 162:S10–S23.[CrossRef][Web of Science][Medline]
Sullivan J. Maximum-likelihood estimation of phylogeny from DNA sequence data. Methods Enzymol. (2005) 395:757–778.[CrossRef][Web of Science][Medline]
Sullivan J. P., Lavoue S., Arnegard M. E., Hopkins C. D. AFLPs resolve phylogeny and reveals mitochondrial introgression within a species flock of African electric fish (Mormyroidea: Teleostei). Evolution (2004) 58:825–841.[CrossRef][Web of Science][Medline]
Swofford D. L. PAUP*. Phylogenetic analysis using parsimony (*and other methods). Version 4 (2002) Sunderland, Massachusetts: Sinauer Associates.
Weiblen G. D. Correlated evolution in fig pollination. Syst. Biol. (2004) 53:128–139.
Weiblen G. D., Bush G. L. Speciation in fig pollinators and parasites. Mol. Ecol. (2002) 11:1573–1578.[CrossRef][Medline]
Weiblen G. D., Yu D. W., West S. A. Pollination and parasitism in functionally dioecious figs. Proc. R. Soc. Lond. Series B Biol. Sci. (2001) 268:651–659.[Medline]
West S. A., Herre E. A., Windsor D. M., Green P. R. S. The ecology and evolution of the New World non-pollinating fig wasp communities. J. Biogeogr. (1996) 23:447–358.[CrossRef]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  V. Merckx and M. I Bidartondo Breakdown and delayed cospeciation in the arbuscular mycorrhizal mutualism Proc R Soc B, May 7, 2008; 275(1638): 1029 - 1035. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. I. Smith, O. Pellmyr, D. M Althoff, M. Balcazar-Lara, J. Leebens-Mack, and K. A Segraves Pattern and timing of diversification in Yucca (Agavaceae): specialized pollination does not escalate rates of diversification Proc R Soc B, February 7, 2008; 275(1632): 249 - 258. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ripplinger and J. Sullivan Does Choice in Model Selection Affect Maximum Likelihood Analysis? Syst Biol, February 1, 2008; 57(1): 76 - 85. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Althoff, M. A. Gitzendanner, and K. A. Segraves The Utility of Amplified Fragment Length Polymorphisms in Phylogenetics: A Comparison of Homology within and between Genomes Syst Biol, June 1, 2007; 56(3): 477 - 484. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||