© 2006 Society of Systematic Biologists
Utility of Nuclear Allele Networks for the Analysis of Closely Related Species in the Genus Carabus, Subgenus Ohomopterus
Edited by Marshal Hedin: Associate Editor
Department of Zoology, Graduate School of Science, Kyoto University Sakyo, Kyoto 606–8502, Japan E-mail:sota{at}terra.zool.kyoto-u.ac.jp
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionConclusionsReferences |
|---|
Nuclear DNA sequence data for diploid organisms are potentially a rich source of phylogenetic information for disentangling the evolutionary relationships of closely related organisms, but present special phylogenetic problems owing to difficulties arising from heterozygosity and recombination. We analyzed allelic relationships for two nuclear gene regions (phosphoenolpyruvate carboxykinase and elongation factor-1
), along with a mitochondrial gene region (NADH dehydrogenase subunit 5), for an assemblage of closely related species of carabid beetles (Carabus subgenus Ohomopterus). We used a network approach to examine whether the nuclear gene sequences provide substantial phylogenetic information on species relationships and evolutionary history. The mitochondrial gene genealogy strongly contradicted the morphological species boundary as a result of introgression of heterospecific mitochondria. Two nuclear gene regions showed high allelic diversity within species, and this diversity was partially attributable to recombination between various alleles and high variability in the intron region. Shared nuclear alleles among species were rare and were considered to represent shared ancestral polymorphism. Despite the presence of recombination, nuclear allelic networks recovered species monophyly more often and presented genetic differentiation patterns (low to high) among species more clearly. Overall, nuclear gene networks provide clear evidence for separate biological species and information on the phylogenetic relationships among closely related carabid beetles.
Keywords: Gene genealogy; introgression; mitochondrial DNA; nuclear DNA; phylogeny; recombination; species relationship
Received June 18, 2005; Revised August 12, 2005; Accepted October 25, 2005
In the study of evolutionary relationships among closely related taxa, diploid nuclear gene genealogies have been used to overcome the shortcomings of mitochondrial gene genealogies. Although mitochondrial DNA has enabled studies of the evolution of various organisms, the resulting gene genealogies frequently indicate unresolved species relationships (e.g., polyphyly and paraphyly) or phylogenies that contradict other evidence (e.g., morphological) for various reasons (Brower et al., 1996; Funk and Omland, 2003). The merits of mitochondrial DNA (haploidy, small genome size, short coalescent time, maternal inheritance) also limit its usefulness for studying species evolution (Ballard and Whitlock, 2004). The vast information stored in nuclear DNA has the potential to compensate for or overcome these limitations. There are, however, difficulties in using diploid nuclear DNA data. In nonmodel organisms, few gene regions have been explored for use in interspecific phylogenetic studies, and the available genes generally exhibit low levels of divergence among species. In addition, heterozygosity and the occurrence of recombination present major difficulties (e.g., Doyle, 1997; Schierup and Hein, 2000; Posada and Crandall, 2002; Sota and Vogler, 2003).
Heterozygosity complicates ordinary phylogenetic analysis. The two alleles from a heterozygous individual are not necessarily most closely related to each other. Such a polymorphism may allow for gene genealogies for a given gene region, but the integration of gene genealogies from multiple loci is difficult because there is no unique sequence for an individual over two or more loci and the genetic relationship among individuals cannot be determined in any straightforward manner (Sota and Vogler, 2003). In practice, for many phylogenies based on diploid gene sequences, allele sequences are not separated and heterozygous sites are treated as missing (or polymorphic). This procedure can make phylogenetic inference ambiguous, especially for closely related taxa (Sota and Vogler, 2003). However, the pairing of different alleles provides information on interbreeding and may be useful in delimitating species, such as with the field for recombination (FFR) concept of Doyle (1995). Although the FFR species concept is sensitive to the presence of heterospecific introgressants, the occurrence of introgressants in turn provides evidence for historical or on-going gene flow.
Intragenic recombination poses another difficult issue in the study of species phylogeny because recombination with crossing-over or gene conversion generally breaks the treelike structure of the descent sequence. Recent methodological developments allow for the detection of recombination in different ways (see Posada and Crandall, 2001a). The prior assessment of recombination in sequence data may improve data analysis or the interpretation of results. Therefore, some studies of closely related species have considered the incidence of recombination in phylogenetic analyses (e.g., Rozas et al., 2001; Broughton and Harrison, 2003; Machado and Hey, 2003). Also, in phylogeographical analysis using nuclear allele networks (e.g., Antunes et al., 2002; Zhang et al., 2005), testing for recombination is a prerequisite for the nested clade analysis (e.g., Templeton, 1998). However, assessment for recombination on nuclear gene genealogy can play a role in detecting hybridization between closely related species (e.g., Okuyama et al., 2005).
It is important to test empirically whether allelic sequences of diploid nuclear genes from closely related species can provide substantial information on species relationships and history, which cannot be obtained from mitochondrial haplotype trees and normal direct sequences that do not account for allelic differences. For this purpose, we focused on closely related species of the genus Carabus, subgenus Ohomopterus. This subgenus is endemic to Japan and has radiated into 15 species that are well defined by morphology and geography. In an attempt to reconstruct the phylogeny of this subgenus using mitochondrial evidence, a marked contradiction between the gene genealogy and morphological species boundaries was found (Su et al., 1996). Subsequent molecular phylogenetic analyses have examined the incongruence between mitochondrial and nuclear gene genealogies (Sota and Vogler, 2001), the utility of nuclear gene sequences in phylogenetic reconstruction (Sota and Vogler, 2003), and phylogeographic patterns of mitochondrial introgression (Sota et al., 2001; Sota, 2002). Yet the relationships among the species in this subgenus are complicated and provide further challenges in terms of genetic and molecular phylogenetic analyses of populations, especially analyses based on nuclear gene genealogies.
Here, we studied two nuclear genes (phosphoenolpyruvate carboxykinase [PepCK] and elongation factor-1 [EF-1a]) and a mitochondrial gene (NADH dehydrogenase subunit 5 [ND5]) for five Ohomopterus species, making inference about intragenic recombination and introgression owing to interspecific hybridization. Along with analyses of species relationships using traditional phylogenetic analyses and population genetic summary statistics, we construct allelic networks for the gene sequences, as network approaches provide many opportunities to address the detailed evolutionary relationships among closely related sequences (Templeton, 1994, 2001a; Posada and Crandall, 2001b; Cassens et al., 2005).
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Organisms
The subgenus Ohomopterus (Coleoptera: Carabidae: genus Carabus) is endemic to Japan and has been classified into 15 species (Ishikawa, 1985, 1991; Takami, 2000). These species are diverse in genital morphology and body size, and two to three species with different body sizes occur sympatrically at many localities (Sota et al., 2000). The "species" of Ohomopterus do not simply follow the biological species concept (BSC; Mayr, 1963) due to problems of determining the species status of geographically isolated populations (or isolation by distance) and the presence of hybrid swarms. Thus, application of the cohesion concept of species (Templeton, 2001b) may be needed to define species of Ohomopterus. However, identifying species based on morphology and mating behavior is unambiguous at least for the species in the present analysis.
We studied five species occurring in the Mt. Kongo area, central Honshu, Japan (34°25'N, 135°40'E): Carabus (Ohomopterus) yamato (Nakane), C. (O.) iwawakianus (Nakane), C. (O.) uenoi (Ishikawa), C. (O.) yaconinus Bates, and C. (O.) dehaanii Chaudoir (Fig. 1). This species assemblage represents the greatest number of Ohomopterus species occurring in a single local area. The five species have different geographical ranges, from the widest in O. dehaanii to the narrowest in O. uenoi, which is confined to the Mt. Kongo area (Fig. 1). In the Mt. Kongo area, the altitudinal ranges vary among the species, although there is no case of complete allopatry among species (Fig. 1; see also Sota, 1985). For these species, identification can be made unambiguously based on external and genital morphologies. There is no substantial evidence of natural hybridization among these species (e.g., obvious hybrid individuals collected in the field), but trans-species polymorphisms in mitochondrial haplotypes are known for these species (Sota, 2002).
|
Beetles for DNA extraction were collected using pitfall traps at four sites within a 6-km distance at altitudes of 200, 400, 700, and 1000 m in the northwest area of Mt. Kongo. For each species, 2 to 14 beetles were sampled from each of two or three sites so as to avoid sampling bias by site. A total of 79 beetles (at least 10 specimens per species) were used in the present analysis (see Table 1). Specimens were identified to species by the authors and kept in the Kyoto University Museum as voucher samples.
|
DNA Extraction and Sequencing
General experimental procedures were as described in Sota and Vogler (2003). The thoracic muscles or testes of beetles preserved in 99% ethanol were digested with proteinase K in hexadecyltrimethylammonium bromide (CTAB) buffer, and total DNA was extracted using the standard phenol-chloroform method. The primers used for PCR and direct sequencing of mitochondrial ND5 and nuclear PepCK and EF-1a are described in Sota and Vogler (2003). For the nuclear genes, direct sequencing of the PCR product (from both the 5'-and 3'-end) and additional sequencing of two cloned PCR products were used to determine two allele sequences for each specimen. All sequencing procedures used the ABI Big Dye Terminator Cycle Sequencing FS Ready Reaction kit and an ABI377 sequencer (Applied Biosystems). In direct sequencing, potential heterozygous sites were detected as those giving equally strong signals of two bases. A heterozygous site was identified only when the same pair of bases was found in the sequence as determined from both directions. To separate allelic sequences, the original PCR product was cloned with TA-cloning plasmids using a TOPO TA cloning kit (Invitrogen). At least two clones were sequenced, and the allele sequences were determined. For the clone sequences, signals that were not found in direct sequencing were regarded as Taq errors. Therefore, we excluded errors attributable to cloning and subsequent PCR where possible.
Gene Genealogy with Traditional Parsimony Analysis
The studied sequences were 1020 bp of mitochondrial ND5 (to the 3'-end), a 627-bp partial fragment of the PepCK exon, and a partial sequence of EF-1a containing parts of two exons (initial 139 bp and last 152 bp) and an intervening 177-to 181-bp intron. The sequence data have been deposited to GenBank (accession numbers: ND5, AF219434
[GenBank]
, AF219439
[GenBank]
, AF219455
[GenBank]
, AF219462
[GenBank]
; AY945502
[GenBank]
to AY945515
[GenBank]
; PepCK, AY945516
[GenBank]
to AY945587
[GenBank]
; EF-1a, AY945588
[GenBank]
to AY945708. The sequences with AF-numbers were reported in Sota and Vogler, 2001; others are first reported in the present study). Although there are two copies of EF-1a in some insects (Danforth and Ji, 1998), the EF-1a sequence in our study matches only one of these copies judging from the intron position (see Sota and Vogler, 2003).
For each of the gene regions, we performed a standard parsimony analysis using PAUP* version 4.0b10 (Swofford, 2002). To accommodate ambiguously aligned regions of the EF-1a intron, we used the recoding method in the program INAASE (Integration of Ambiguously Aligned Sequences) (Lutzoni et al., 2000). A heuristic search was performed using 100 random addition analyses with TBR branch swapping. The support for each node on a strict consensus trees was evaluated by nonparametric bootstrap (1000 replications; heuristic analysis of 20 random addition analyses with TBR branch swapping) and Bremer support indices using TreeRot version 2 (Sorenson, 1999).
Network Analysis
Networks constructed for gene genealogies are advantageous over strictly bifurcating trees as they provide a way of representing more of the phylogenetic information present in a data set and can account effectively for processes acting at the species level (reviewed in Posada and Crandall, 2001b). The network approach can also be effective for analyzing closely related species (Templeton, 1994, 2001a). Here we constructed statistical parsimony networks using the method of Templeton et al. (1992) (TCS method). Recently Cassens et al. (2005) proposed a method to construct a network by uniting equally parsimonious trees, which can provide better estimation of gene genealogies compared to other commonly used methods. However, we used the TCS method because it was more easily integrated with other methods for testing cladogram structure and detecting recombination.
A network for each gene was constructed with a 95% connection limit of parsimony using the program TCS version 1.13 (Clement et al., 2000). For the EF-1a intron region, we used data with the alignment-ambiguous regions delimited as the input for the Integration of Ambiguously Aligned Sequences (INAASE) program, and we adjusted the alignment manually. However, because characters recoded with INAASE could not be used in TCS directly, each gap on the aligned data was treated as a fifth base.
Detecting Recombination
Although the nuclear DNA sequences studied were relatively short, intragenic recombination may have affected divergence in the sequences. We used a simulation-based analysis, the informative-sites test (Worobey, 2001) implemented in the program PIST (available at http://evolve.zoo.ox.ac.uk/software), to detect recombination in the PepCK and EF-1a gene regions. Following Worobey (2001), only third-position data for the exon regions were used. To obtain the trees used in the test, we first determined the best-fit substitution model for each of the two data sets using the hierarchical likelihood ratio test in ModelTest version 3.06 (Posada and Crandall, 1998). The GTR+I+G model was selected for PepCK data, and the HKY+I+G model was selected for EF-1a data. With these models, we performed Bayesian Markov chain Monte Carlo (MCMC) analysis using MrBayes version 3.1.1 (Huelsenbeck and Ronquist, 2001), with one million generations of run with four chains and a sample frequency of 100 generations. The burn-in period was determined graphically. After discarding nonstationary runs as burn-in, we obtained the means of the base frequencies and parameters in the substitution model. In addition, trees of the highest five posterior probabilities were sampled as the test trees, for use in the informative-site test. Each informative-site test was performed using 10,000 simulated runs. To calculate the informative-site index (ISI), which takes on values between 0 (recombination absent) and 1 (recombination pervasive), we used the mean tree length of 20 simulated data sets in which characters at each site of the original data were randomly reassigned among the taxa to remove linkage between sites (see Worobey, 2001).
We also attempted to estimate recombination using the approach proposed by Crandall and Templeton (1999), which begins by constructing statistical parsimony networks. Although their search for recombinants uses disconnected network parts as a result of tree construction with statistical parsimony, we searched for the possibility of recombinants for alleles located at isolated tips or on long loops in single networks of the PepCK and EF-1a exons. We searched for a clumped distribution of substitutions on either side (5' or 3') of the sequences. Simultaneously, we divided the sequence data into three segments for PepCK (200sites each) or into two segments for EF1a (before and after the intron) and checked for matching between different segments of the target alleles and those of other alleles to search for candidate parental alleles. Then, we constructed a parsimony tree for possible recombinants and three or more candidate parental alleles, and mapped substitution sites on the branches. Equation 5.3 in Crandall and Templeton (1999) was used to determine the probability of nonrandom clustering of substitutions, which is interpreted as the result of recombination (at P < 0.05). In this interpretation, caution is needed concerning the possibility of PCR-mediated recombination (Kanagawa, 2003). If a recombinant sequence has been created from two alleles during PCR and subsequently cloned and sequenced, the recombinant sequence must coincide partially (but not completely) with the other allele sequences within the same individual. Therefore, we paid attention to possible recombinants that had the parental sequence from the same individual. After determining recombinant and parental alleles as above, we reviewed the direct sequencing results and sequences of clones for the possibility of PCR-mediated recombination.
Nonrandom Association between Allelic Clades and Species Category
Among closely related species, different taxa with little current gene flow may not necessarily be recovered as monophyletic species. However, they may cluster in different parts (clades) of a statistical parsimony network (Templeton, 1994). To test statistically whether alleles are distributed randomly among network clades with respect to the morphological species category, we used the permutational contingency test (Templeton and Sing, 1993; Templeton, 1994; see Zhang et al., 2005, for an example of application). We considered the nonrandom distribution of species across different clades at each nesting level (one-step and higher steps) as evidence against panmixia and hence evidence for different reproductive entities.
Prior to the above analysis, cladogram nesting was performed following Templeton et al. (1987) and Templeton and Sing (1993). Due to highly complex loop structures of allele networks, we relied on an option for cladogram nesting implemented in TCS version 1.13 (this option has been removed in later versions of the TCS program). This program performed cladogram nesting without untying loops in a network. We used the program GeoDis (Posada et al., 2000) to perform the permutational test. At nesting level of n-step, we examined if alleles from different species were distributed randomly over n-step clades without considering distribution of alleles within each n-step clade. In addition, within each of n-step clade consisting of two or more species, we tested the nonrandomness of species distribution among n–1-step clades and counted the frequency of nonrandom n-step clades.
Genetic Divergence and Species Relationship
We determined the fixation index FST based on pairwise distance (Reynolds et al., 1983) using Arlequin version 2.000 (Schneider et al., 2000). The null hypothesis of FST = 0 (no differentiation between species) was tested using the permutational procedure implemented in Arlequin. In addition, we constructed neighbor-joining (NJ) trees (Saitou and Nei, 1987) using the observed pairwise sequence difference between species. For each sequence data set, mean pairwise sequence differences were obtained using Arlequin, and an NJ tree was obtained using the neighbor program in PHYLIP version 3.6 (Felsenstein, 2004). To assess tree stability in relationship to the sampling of different alleles within species, we conducted bootstrap resampling. One hundred bootstrap data sets were created for each gene region using an Excel macro program, and an NJ tree was obtained for each data set, as described above. Bootstrap frequencies of different nodes were obtained using the "CONSENSE" program in PHYLIP.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Sequence Divergence
The ND5 and PepCK sequences were exon regions only, whereas the EF-1a sequences consisted of separate exon regions with a variable intervening intron region. For the EF-1a data, we present analyses of both the exon sequences only and of the complete sequences. The mitochondrial ND5 sequences represented 18 haplotypes, whereas the allele sequences of the nuclear genes showed greater diversity (58 to 86 alleles) with high frequencies of heterozygous individuals (Table 1). This situation prevented the assessment of the association between alleles across loci. The level of nucleotide diversity across all of the samples was similar for different gene regions, but, within each gene region, the nucleotide diversity was much higher in O. iwawakianus and O. yaconinus than in the other three species (Table 1).
Species Monophyly in Traditional Parsimony Analysis
The phylogenetic relationships of alleles using traditional parsimony analysis revealed only a few cases of species monophyly (Fig. 2). The mitochondrial ND5 revealed only the monophyly of O. yamato haplotypes, and PepCK revealed monophyly for only O. dehaanii alleles. Although the EF-1a exon revealed the monophyly of O. dehaanii alleles only, the inclusion of the intron sequence resulted in monophyly of both O. dehaanii and O. yamato.
|
The sharing of identical sequences across species was most prevalent in the ND5 data, in which four haplotypes were shared by two or three species, except O. yamato. Interspecific sharing of alleles was limited to two sequences in PepCK and one in the EF-1a exon. EF-1a exhibited length variation in the intron, with 83 total allele sequences, none of which were shared between species.
|
Statistical Parsimony Network
Statistical parsimony network construction within the 95% connection limit resulted in three subnetworks for ND5 (Fig. 3), whereas single networks were constructed for the PepCK (Fig. 4) and EF-1a exons (Fig. 5). The inclusion of the EF-1a intron region resulted in a disconnected network with four components (Fig. 6).
|
|
|
|
A loop structure was found only in a small part of the ND5 network (Fig. 3), whereas loops were common and extensive for the nuclear gene regions. In particular, the PepCK alleles were highly divergent for O. iwawakianus and O. yaconinus, representing a highly complex loop structure, although these alleles were clustered for each of the other three species (Fig. 4). O. uenoi had only two sister alleles, both of which were shared by O. iwawakianus and O. yaconinus. O. iwawakianus and O. yaconinus exhibited similar degrees of sequence divergence, and the lineages were not separated according to species.
In the EF-1a exon data (Fig. 5), alleles of each species were generally clustered, and a clearer pattern of species differentiation was revealed. All of the O. uenoi alleles were close to those of O. iwawakianus. Complete EF-1a sequences, including the intron, resulted in disconnected networks due to the highly variable intron sequence (Fig. 6). Clustering of alleles largely followed the pattern of highly variable parts within the intron sequence, depicted as alignment-ambiguous regions 1 to 5. In these ambiguously aligned regions, O. yamato and O. dehaanii had species-specific patterns, resulting in species-specific clades, whereas each of the remaining three species had more or less variable patterns and shared intron types. For example, the very distinct type 1 sequence of region 3 was shared by six O. iwawakianus alleles, six O. yaconinus alleles, and one O. uenoi allele (see clades 4–5 and 4–6, Fig. 6).
Detecting Recombination
When the informative-sites test was performed on all of the sequence data irrespective of species, PepCK sequences revealed a significant sign of recombination (Table 2). When the test was applied to each species for PepCK, O. iwawakianus revealed a highly significant sign of recombination whereas O. yaconinus yielded marginally significant values around P = 0.05. No evidence for recombination was revealed in the EF-1a data.
|
We also tried to detect recombination by mapping substitutions on branches of the networks. In the PepCK network, nine potential recombinants of O. iwawakianus and O. yaconinus alleles were detected with the possible parental alleles (Fig. 7; see also Fig. 4). Three of these had heterospecific parents, and each of O. iwawakianus and O. yaconinus had O. dehaanii as one of the parents. Further, two alleles of O. uenoi that were shared with O. iwawakianus and O. yaconinus might be recombinants between O. iwawakianus and O. yaconinus alleles.
In the EF-1a exon, one allele of O. iwawakianus that was seven mutational steps from the other O. iwawakianus alleles and was rather close (two steps) to O. yaconinus alleles had exon sequences before and after the intron that matched those of two distant O. iwawakianus alleles (Fig. 5). This O. iwawakianus alleles could be a recombinant of other O. iwawakianus alleles with a probability of P = 0.029 (the inclusion of alignment-unambiguous intron sites increased this significance: P = 0.002). For the entire sequenced region of EF-1a, we obtained three disconnected networks and a single allele separated from the others. However, we could not detect recombination in these networks, except for the one O. iwawakianus allele mentioned above.
Nonrandom Association between Allelic Clade and Species Category
Using the nested clade designs, we tested for random association between species and clades as expected for panmictic populations. At all nesting levels, we found significant, nonrandom associations of species with clades (Table 3). As there were few shared sequences or closely related sequences with a one-site difference between species, clades with single species were common at lower nesting levels.
|
Within a clade with two or more species, the association between species and clades one-step lower could be random, especially at lower nesting levels (Table 3), but this can be attributed to the small sample sizes, as each allele was largely represented by a single specimen.
Species Relationship Based on Mean Genetic Distance
Genetically distinct entities were identified using the non-zero pairwise FST for all pairs of species for all genes, but the mean pairwise sequence difference between species showed variable interspecific relationships, depending on gene regions (Table 4). Therefore, the species tree constructed from genetic distance (mean pairwise sequence difference) using the neighbor-joining method revealed different sister relationships, depending upon the gene regions used (Fig. 8). Species relationship based on the mitochondrial ND5 gene, representing the grouping of O. dehaanii, O. yaconinus, and O. iwawakianus, was very different from those based on nuclear genes. Nuclear genes recovered a grouping of O. yaconinus, O. iwawakianus, and O. uenoi, but the sister relationship within this group differed between PepCK and EF-1a. For PepCK, O. uenoi had only two allele sequences with one-step differences from each other, which were shared with both O. iwawakianus and O. yaconinus. Therefore, the sister relationship between O. uenoi and O. yaconinus is attributable to lower sequence divergence within O. yaconinus than O. iwawakianus.
|
|
Bootstrap resampling of the data with respect to alleles within species revealed that the topologies could be affected by sampling (Fig. 8). Especially in the EF-1a exon, the relationships among O. yaconinus, O. yamato, and O. dehaanii are highly variable, whereas the inclusion of the intron data enhances the stability of the monophyly of O. dehaanii and O. yamato.
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
Although the five sympatric Ohomopterus species can be discriminated unambiguously based on morphology (Fig. 1), mitochondrial and nuclear gene genealogies revealed differences in the recovery of monophyletic species and the pattern of trans-species polymorphisms. The recovery of species as a monophyletic clade on a phylogenetic tree can be the basis of recognizing species (see Sites and Marshall, 2004). For example, a practical definition in line with the phylogenetic species concept (PSC) is expressed by Brower (1999): "each population is a phylogenetic species if the alleles of all its members are joined in a contiguous section of an unrooted network, separated from each other population by a single branch that represents parsimoniously inferred character state transformations." However, empirical studies like ours readily encounter situations in which a strict reliance upon recovery of monophyletic lineages does not work as the basis of determining species' boundaries. In general, a slower substitution rate may cause difficulty in recovering monophyletic species for nuclear genes. In addition, the presence of rare or extensive trans-species polymorphisms, that are not artifacts of the analytical procedure (see Brower et al., 1996), may complicate interpretation of species relationship. Introgressive hybridization and incomplete lineage sorting are historical processes leading to these phenomena (reviewed in Brower et al., 1996; Funk and Omland, 2003). Further analyses of the pattern of trans-species polymorphisms may provide clues to discerning morphological species, as supported by the genetic evidence.
Introgressive Hybridization
Introgressive hybridization is of primary interest as a cause of trans-species polymorphisms, as hybridization is particularly important in the study of species and speciation (e.g., Harrison, 1990; Arnold, 1997). The phylogenetic effects of introgression are considered greater for mitochondrial DNA than for nuclear DNA because of the lack of recombination and less linkage to selected loci (Funk and Omland, 2003). However, portions of the nuclear genome with weak linkage to loci important in speciation process can be introgressed as well (Wu, 2001). In Ohomopterus, the existence of hybrid zones has been well documented for several pairs of parapatric species (Kubota, 1988; Kubota and Sota, 1998; Sota and Kubota, 1998; Sota et al., 2001; Takami and Suzuki, 2005). These natural hybridizations appear to have resulted in introgression of mitochondria, which can be detected by the sharing pattern of the identical haplotypes between species (Sota et al., 2001; Sota, 2002; Takami and Suzuki, 2005). However, for sympatric species, current natural hybridization is virtually absent or has yet to be demonstrated, but introgressive hybridization may have occurred in the past for some sympatric species.
In the present study, we obtained no direct evidence for the introgression of nuclear DNA. In EF-1a, we found no cases of interspecific sharing of the same sequence. Although O. iwawakianus and O. uenoi shared one exon sequence, the intron sequence between the exons was not identical, indicating that the exon sequence had been inherited from a common ancestor. In PepCK, O. iwawakianus, O. yaconinus, and O. uenoi shared two alleles with one-step mutations. Of these species, O. uenoi possessed only these two alleles, and it is unlikely that these alleles are introgressants from the other two species that have replaced native alleles. It seems more likely that these alleles were inherited from the common ancestor O. uenoi shares with O. iwawakianus or O. yaconinus (i.e., retention of ancestral polymorphism).
In contrast, the mitochondrial ND5 gene data showed sharing of identical sequences by different pairs of species, suggesting the introgression of mitochondria due to interspecific hybridization. The sharing pattern of mitochondrial haplotypes may be explained by the feasibility of interspecific mating, depending on differences in body size and genital morphology (see Sota, 2002). As previously mentioned, no putative hybrid individual has been reported for these Ohomopterus species in the Mt. Kongo area; therefore, introgressive hybridization should have been rare or occurred in the remote past.
The contrasting results for introgression between nuclear and mitochondrial genes may simply reflect different susceptibility of these genes to recombination and selection during hybridization and subsequent backcrossing, as mentioned earlier. Further research with many nuclear loci using microsatellite or amplified fragment length polymorphism (AFLP) analyses may demonstrate differential introgression of nuclear DNA portions correlated with the pattern of mitochondrial introgression.
Recombination
Despite the short sequences considered in this study, the nuclear DNA networks appeared highly reticulated, and a statistical test revealed the possibility of recombination for at least PepCK. Because PCR-mediated recombination might have affected our results, we examined this possibility (see Materials and Methods). All of the PepCK recombinants had candidate parental sequences from other individuals and were not questionable. The sole recombinant type of EF-1a had one parental sequence from the same individual, but the reexamination of the direct sequencing results and sequences of cloned PCR products for the EF-1a data revealed no evidence of PCR-mediated recombination. Therefore, a PCR artifact in our results is probably negligible.
Sequences resulting from recombination will appear on phylogenetic trees in various ways (Schierup and Hein, 2000; Posada and Crandall, 2002). A recombinant between closely related sequences might be located close to the parental sequences, whereas a recombinant between distantly related sequences should appear far from the parental sequences. As indicated by the possible recombinant between the O. iwawakianus alleles in the EF-1a gene, even intraspecific intergenic recombination can create a distinct allele that may be closer to heterospecific alleles. Four of the other possible cases of recombination had heterospecific parents (Fig. 7). If these inferences are correct, the recombination, which should have followed the introgression of heterospecific genes, would provide additional evidence for introgressive hybridization that have not been detected from the sharing pattern of alleles, as discussed in the previous section. The combination of parental species for recombinants (Fig. 7) suggests introgressive hybridization between O. dehaanii x O. yaconinus, O. dehaanii x O. iwawakianus, and O. yaconinus x O. iwawakianus, which could also be inferred from the mitochondrial data. However, it may be possible that the sequence portion, which coincided with the heterospecific allele, is a rare allele sequence of conspecifics retained as an ancestral polymorphism. Also, given the high variability of some third codon positions in PepCK, these coincident sequence portions might have evolved in parallel by chance alone.
Overall, recombination was not considered to be a major process in sequence evolution. However, the confounding effect of recombinants on phylogenetic reconstruction warrants the assessment of recombination in various cases of seemingly complicated phylogenetic data for closely related taxa.
Species Relationships
Here, we discuss the utility of allelic data in recognizing species as genetically different entities and in inferring species phylogeny. Species relationships represented by gene genealogies are, a priori, vague in many areas, reflecting occasional gene flow after speciation (introgressive hybridization), as well as the sharing of ancestral polymorphisms (e.g., Maddison, 1997). Therefore, inferences about species relationships may sometimes rely on the frequency distribution of alleles or the distribution of genetic distances between species. Here, we have attempted to identify species as statistically different genetic populations. The cladogram association depicted a significant association between nested clades and the morphological species category for all sequence data and nesting levels. Similarly, the fixation index FST was greater than zero for all species pairs and showed varying degrees of relatedness. These analyses, applied to sympatric species, provided a statistical evidence for restricted gene flow due to reproductive isolation and not to geographical isolation. Compared with monophyletic criteria of delimiting species (e.g., Brower, 1999; Sites and Marshall, 2004), our procedures can recognize the existence of different species for a wide range of gene genealogies, which do not necessarily reveal strict species monophyly.
For inference of species phylogeny from the allelic data, we attempted to construct neighbor-joining trees using average genetic distances between species. Different genes were found to support different sister relationships, although a bootstrap experiment showed that stochastic sampling could have partially affected the results. As already mentioned, the mitochondrial gene genealogy was likely extensively affected by introgressive hybridization and could be excluded from this consideration. The two nuclear genes, PepCK and EF-1a, commonly indicated grouping of O. dehaanii and O. yamato, whereas these genes showed different sister relationships among the remaining three species; i.e., O. yaconinus with O. uenoi for PepCK, and O. iwawakianus with O. uenoi for EF-1a.
A previous morphological cladistic analysis (Takami, 2000) hypothesized the relationships of the five species as ((dehaanii, yamato), yaconinus), (iwawakianus, uenoi)), and a combined molecular phylogenetic analysis with five nuclear gene sequences (Sota and Vogler, 2003) resulted in topologies consistent with the morphological result. Thus, the grouping of O. dehaanii and O. yamato seems to be plausible, as is the grouping of O. yaconinus, O. iwawakianus, and O. uenoi with the most likely sister relationship of O. iwawakianus and O. uenoi. However, the allele networks presented here (Figs. 4 to 6) did not reveal a direct relationship between O. yamato and O. dehaanii alleles, and the EF-1a alleles of these species were intervened by those of O. yaconinus. A closer relationship between O. iwawakianus and O. uenoi was evident in the EF-1a network, and the PepCK network did not provide contradictory evidence, although it equally favored the opposite relationship; i.e., the two O. uenoi alleles are descents from either of O. iwawakianus or O. yaconinus. A simple interpretation of species phylogeny from each allelic network may be depicted as networks of species shown on the right side of Figure 8, indicating an inconsistency with the bifurcation pattern in the left. Thus, the bifurcation pattern obtained from the mean genetic distances between species may not represent actual process of species division due to some statistical artifact. For allelic networks to become more useful in the study of species phylogeny, we need to develop additional methods to reconstruct species trees based on allele networks, and methods allowing the integration of multiple, sometimes conflicting, allelic networks into one species tree hypothesis.
|
Conclusions |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
The present study has demonstrated that the assessment of networks of nuclear alleles, which exhibit detailed relationships among alleles belonging to different closely related taxa, can provide much insight into the process of divergence of lineages leading to speciation. In addition, nuclear allele networks can reveal confounding processes involved in gene genealogies, such as recombination and introgressive hybridization. Thus, network analysis of sequences from variable nuclear DNA regions should be used more often in the analysis of rapidly diverging groups of organisms. Of course, it is risky to infer the relationships among closely related species based on just one or two gene genealogies, because different portions of nuclear DNA can show different patterns of divergence among closely related species (e.g., Wu, 2001). To resolve the history of diversification among Ohomopterus species, assessment of allele networks for additional unlinked loci would be useful.
|
Acknowledgements |
|---|
We are grateful to Alfried Vogler for inspiring the present study, Atushi Kawakita for laboratory assistance, and to Yudai Okuyama for comment. Thanks are also due to M. Hedin, R. Page, G. Morse, and an anonymous referee for their helpful comments on our manuscript. Partially supported by grants-in-aid from JSPS (nos. 11304056; 12440217; 15207004) and from MEXT, Japan (21st Century COE Research, A14).
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionsReferences |
|---|
-
Antunes A., Templeton A. R., Guyomard R., Alexandrino P. The role of nuclear genes in intraspecific evolutionary inference: Genealogy of the transferrin gene in the brown trout. Mol. Biol. Evol. (2002) 19:1272–1287.
Arnold M. L. Natural hybridization and evolution (1997) New York: Oxford University Press.
Ballard J. W. O., Whitlock M. C. The incomplete natural history of mitochondria. Mol. Ecol. (2004) 13:729–744.[CrossRef][Medline]
Broughton R. E., Harrison R. G. Nuclear gene genealogies reveal historical, demographic and selective factors associated with speciation in field crickets. Genetics (2003) 163:1389–1401.
Brower A. V. Z. Delimitation of phylogenetic species with DNA sequences: A critique of Davis and Nixon's population aggregation analyis. Syst. Biol. (1999) 48:199–213.
Brower A. V. Z., DeSalle R., Vogler A. Gene trees, species trees, and systematics: A cladistic perspective. Annu. Rev. Ecol. Syst. (1996) 27:423–450.[CrossRef][Web of Science]
Cassens I., Mardulyn P., Milinkovitch M. C. Evaluating intraspecific "network" construction methods using simulated sequence data: Do existing algorithms outperform the global maximum parsimony approach? Syst. Biol. (2005) 54:363–372.
Clement M., Posada D., Crandall K. A. TCS: A computer program to estimate gene genealogies. Mol. Ecol. (2000) 9:1657–1660.[CrossRef][Medline]
Crandall K. A., Templeton A. R. Statistical approaches to detecting recombination. In: The evolution of HIV—Crandall K. A., ed. (1999) Baltimore: The Johns Hopkins University Press. Pages 153–176.
Danforth B. N., Ji S. Elongation factor-1 occurs at least two copies in bees: Implications for phylogenetic analysis of EF-1 sequences in insects. Mol. Biol. Evol. (1998) 15:222–235.
Doyle J. J. The irrelevance of allele tree topologies for species delimitation, and a non-topological alternative. Syst. Bot. (1995) 20:574–588.[CrossRef]
Doyle J. J. Trees within trees: Genes and species, molecules and morphology. Syst. Biol. (1997) 46:537–553.
Felsenstein J. PHYLIP (phylogeny inference package), version 3.6. (2004) Seattle: Department of Genome Sciences, University of Washington.
Funk D. J., Omland K. E. Species-level paraphyly and polyphyly: Frequency, causes, and consequences, with insights from animal mitochondrial DNA. Annu. Rev. Ecol. Evol. Syst. (2003) 34:397–423.[CrossRef]
Harrison R. G. Hybrid zones: Windows on evolutionary process. Oxford Surv. Evol. Biol. (1990) 7:69–128.
Huelsenbeck J. P., Ronquist F. MrBayes: Bayesian inference of phylogeny. Bioinformatics (2001) 17:754–755.
Ishikawa R. The subfamily Carabinae. In: The Coleoptera of Japan in color—Ueno S., Kurosawa Y., Sato M., eds. (1985) volume II. Osaka: Hoikusha. Pages 14–54.
Ishikawa R. The evolution of Carabus. In: Divergence and isolating mechanisms. (In Japanese.) (1991) Tokyo: Yasaka-shobo.
Kanagawa T. Bias and artifacts in multitemplate polymerase chain reactions (PCR). J. Biosci. Bioeng. (2003) 96:317–323.[Web of Science][Medline]
Kubota K. Natural hybridization between Carabus (Ohomopterus) maiyasanus and C. (O.) iwawakianus (Coleoptera, Carabidae) (1988) 56. Tokyo: Kontyu. 370–378.
Kubota K., Sota T. Hybridization and speciation in the carabid beetles of the subgenus Ohomopterus (Coleoptera, Carabidae, genus Carabus). Res. Popul. Ecol. (1998) 40:213–222.[CrossRef]
Lutzoni F., Wagner P., Reeb V., Zoller S. Integrating ambiguously aligned regions of DNA sequences in phylogenetic analysis without violating positional homology. Syst. Biol. (2000) 49:628–651.
Machado C. A., Hey J. The causes of phylgenetic conflict in a classic Drosophila species group. Proc. Roy. Soc. Lond. B (2003) 270:1193–1202.[Medline]
Maddison W. P. Gene trees in species trees. Syst. Biol. (1997) 46:523–536.
Mayr E. Animal species and evolution (1963) Cambridge, Massachusetts: Harvard University Press.
Okuyama Y., Fujii N., Wakabayashi M., Kawakita A., Ito M., Watanabe M., Murakami N., Kato M. Nonuniform concerted evolution and chloroplast capture: Heterogeneity of observed introgression patterns in three molecular data partition phylogenies of Asian Mitella (Saxifragaceae). Mol. Biol. Evol. (2005) 22:285–296.
Posada D., Crandall K. A. Modeltest: Testing the model of DNA substitution. Bioinformatics (1998) 14:817–818.
Posada D., Crandall K. A. Evaluation of methods for detecting recombination from DNA sequences: Computer simulations. Proc. Nat. Acad. Sci. USA (2001a) 98:13757–13762.
Posada D., Crandall K. A. Intraspecific gene genealogies: Trees grafting into networks. Trends Ecol. Evol. (2001b) 16:37–45.[CrossRef][Medline]
Posada D., Crandall K. A. The effect of recombination on the accuracy of phylogeny estimation. J. Mol. Evol. (2002) 54:396–402.[Web of Science][Medline]
Posada D., Crandall K. A., Templeton A. R. GeoDis: A program for the cladistic nested analysis of the geographical distribution of genetic haplotypes. Mol. Ecol. (2000) 9:487–488.[CrossRef][Medline]
Reynolds J., Weir B. S., Cockerham C. C. Estimation for the coancestry coefficient: Basis for a short-term genetic distance. Genetics (1983) 105:767–779.
Rozas J., Gullaud M., Blandin G., Aguadé M. DNA variation at the rp49 gene region of Drosophila simulans: Evolutionary inferences from an unusual haplotype structure. Genetics (2001) 158:1147–1155.
Saitou N., Nei M. The neighbor-joining method: A new method for reconstructing phylogenetic trees. Mol. Biol. Evol. (1987) 4:406–425.[Abstract]
Schierup M. H., Hein J. Consequences of recombination on traditional phylogenetic analysis. Genetics (2000) 156:879–891.
Schneider S., Roesseli D., Excoffier L. Arlequin version 2.000. In: A software for population genetics data analysis (2000) Geneva: University of Geneva.
Sites J. W. Jr., Marshall J. C. Operational criteria for delimiting species. Annu. Rev. Ecol. Evol. Syst. (2004) 35:199–227.[CrossRef]
Sorenson M. D. TreeRot, version 2 (1999) Boston: Boston University.
Sota T. Life history patterns of carabid beetles belonging to the subtribe Carabina (Coleoptera, Carabidae) in the Kinki District, western Japan. Kontyu, Tokyo (1985) 53:370–378.
Sota T. Radiation and reticulation: Extensive introgressive hybridization in the carabid beetles Ohomopterus inferred from mitochondrial gene genealogy. Popul. Ecol. (2002) 44:145–156.[CrossRef]
Sota T., Ishikawa R., Ujiie M., Kusumoto F., Vogler A. P. Extensive trans-species mitochondrial polymorphisms in the carabid beetles Carabus subgenus Ohomopterus caused by repeated introgressive hybridization. Mol. Ecol. (2001) 10:2833–2847.[Medline]
Sota T., Kubota K. Genital lock-and-key as a selective agent against hybridization. Evolution (1998) 52:1507–1513.[CrossRef][Web of Science]
Sota T., Vogler A. P. Incongruence between mitochondrial and nuclear gene trees in the carabid beetles Ohomopterus. Syst. Biol. (2001) 50:39–59.
Sota T., Vogler A. P. Reconstructing species phylogeny of the carabid beetles Ohomopterus using multiple nuclear DNA sequences: Heterogeneous information content and the performance of simultaneous analysis. Mol. Phylogenet. Evol. (2003) 26:139–154.[CrossRef][Web of Science][Medline]
Sota T., Takami Y., Kubota K., Ujiie M., Ishikawa R. Interspecific body size differentiation in species assemblages of the carabid subgenus Ohomopterus in Japan. Popul. Ecol. (2000) 42:279–291.[CrossRef]
Su Z. H., Tominaga O., Ohama T., Kajiwara E., Ishikawa R., Okada T. S., Nakamura K., Osawa S. Parallel evolution in radiation of Ohomopterus ground beetles inferred from mitochondrial ND5 gene sequences. J. Mol. Evol. (1996) 43:662–671.[CrossRef][Web of Science][Medline]
Swofford D. L. PAUP*, Phylogenetic analysis using parisimony (*and other methods), version 4. (2002) Sunderland, Massachusetts: Sinauer Associates.
Takami Y. Phylogeny of the subgenus Ohomopterus (Coleoptera, Carabidae, genus Carabus): A morphological aspect. Tokyo Metropolitan Univ. Bull. Nat. Hist. (2000) 4:1–32.
Takami Y., Suzuki H. Morphological, genetic and behavioural analyses of a hybrid zone between the ground beetles Carabus lewisianus and C. albrechti (Coleoptera, Carabidae): Asymmetrical introgression caused by movement of the zone? Biol. J. Linn. Soc. (2005) 86:79–94.[CrossRef][Web of Science]
Templeton A. R. The role of molecular genetics in speciation studies. In: Molecular ecology and evolution: Approaches and applications—Schierwater B., Streit B., Wagner G. P., DeSalle R., eds. (1994) Basel: Birkhäuser-Verlag. Pages 455–457.
Templeton A. R. Nested clade analyses of phylogeographic data: Testing hypotheses about gene flow and population history. Mol. Ecol. (1998) 7:381–397.[CrossRef][Medline]
Templeton A. R. Using phylogeographic analysis of gene trees to test species status and process. Mol. Ecol. (2001a) 10:779–791.[CrossRef][Medline]
Templeton A. R. The meaning of species and speciation: A genetic perspective. In: Speciation and its consequences—Otte D., Endler J. A., eds. (2001b) Sunderland, Massachusetts: Sinauer Associates. Pages 3–27.
Templeton A. R., Boerwinkle E., Sing C. F. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping. I. Basic theory and an analysis of alchol dehydrogenase activity in Drosophila. Genetics (1987) 117:343–351.
Templeton A. R., Crandall K. A., Sing C. F. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics (1992) 132:619–633.
Templeton A. R., Sing C. F. A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping. IV. Nested analysis with cladogram uncertainty and recombination. Genetics (1993) 134:659–669.
Worobey M. A novel approach to detecting and measuring recombination: New insights into evolution in viruses, bacteria, and mitochondria. Mol. Biol. Evol. (2001) 18:1425–1434.
Wu C. I. The genic view of the process of speciation. J. Evol. Biol. (2001) 14:851–865.[CrossRef][Web of Science]
Zhang A. B., Kubota K., Takami Y., Kim J. L., Kim J. K., Sota T. Species status and phylogeography of two closely related Coptolabrus species (Coleoptera: Carabidae) in South Korea inferred from mitochondrial and nuclear gene sequences. Mol. Ecol. (2005) 14:3823–3841.[CrossRef][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  T. Sota and N. Nagata Diversification in a fluctuating island setting: rapid radiation of Ohomopterus ground beetles in the Japanese Islands Phil Trans R Soc B, October 27, 2008; 363(1508): 3377 - 3390. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. A. Morrison Increasing the Efficiency of Searches for the Maximum Likelihood Tree in a Phylogenetic Analysis of up to 150 Nucleotide Sequences Syst Biol, December 1, 2007; 56(6): 988 - 1010. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||