| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
Department of Botany, University of Toronto, 25 Willcocks St., Toronto, Ontario, Canada M5S 3B2
Received for publication June 13, 2003. Accepted for publication October 16, 2003.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: allozymes • cpDNA • Melanthiaceae • phylogeography • pollen: seed flow • post-glacial migration • Reid's Paradox • Trillium
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Van der Pijl, 1969
; Jaquard et al., 1984
; Sauer, 1988
; Dingle, 1996
). In temperate regions, dispersal ability has directly influenced response to glacial cycles. During the Pleistocene (2.4 x 106 yr(myr)–10 000 yr BP), at least six glacial advances affected the physical and biological environments of the Northern Hemisphere (Cox and Moore, 2000
). In North America, the Wisconsin glaciation began 120 000 yr BP and ended approximately 8000 yr BP (Davis, 1983
). At its maximum, the ice sheet extended as far south as 40° N in eastern North America. As the Wisconsin ice sheet began to retreat approximately 18 000 yr BP, species that survived in ice-free refugia migrated north to previously glaciated habitats.
In plants, post-glacial colonization is especially paradoxical because observed average seed dispersal distances cannot account for the rapid northward migration that occurred in many species (Reid's Paradox; Clark et al., 1998
). For example, Cain et al. (1998)
modeled migration of ant-dispersed Asarum canadense L. and concluded that long-distance seed dispersal, by unknown mechanisms, was necessary to explain its present-day distribution in eastern North America. Similarly, analyses of preserved pollen samples suggest that Acer spp. migrated nearly 2000 km in approximately 10 000 yr, a rate of 200 m/yr (Davis, 1983
). Recent modeling studies also indicate that modes of colonization, particularly rare long-distance dispersal events, influence genetic diversity in glaciated regions (Ibrahim et al., 1995
). Indeed, reduced genetic diversity has been documented in glaciated ranges of several European and North American species (for a review see Hewitt [2000]
), although in most cases the occurrence and mechanisms of long-distance dispersal are not known.
Phylogeographic methods can be used to investigate post-glacial migration and dispersal and are especially useful for the study of herbaceous species that do not preserve well in the palynological record (Cruzan and Templeton, 2000
; Hewitt, 2000
). Early studies used intuitive approaches to decipher phylogeographic data (Avise, 2000
); subsequently, Templeton and colleagues (Templeton et al., 1995
; Templeton, 1998
; Posada et al., 2000
) developed nested clade analysis (NCA), a procedure that statistically assesses the process of migration and dispersal resulting in observed phylogeographic patterns. Recent concerns that NCA may lead to erroneous conclusions (Knowles and Maddison, 2002
) suggests that both intuitive and NCA approaches should be used in phylogeographic studies. Since in most angiosperms chloroplast DNA (cpDNA) is maternally inherited through seed alone, phylogeographic analysis of cpDNA haplotypes can provide an unambiguous marker for the study of seed dispersal. In addition, surveys of allozymes can provide insights into the genetic consequences of post-glacial migration (e.g., Broyles, 1998
). Furthermore, when used in combination, comparisons of the genetic structures of cpDNA and allozymes can allow estimates of rates of dispersal via pollen and seed (Ennos, 1994
).
Despite a wealth of palynological studies of the post-glacial history of eastern North America (Davis, 1983
), there is much less phylogeographical data from plant species from this region (Sewell et al., 1996
; Echt et al., 1998
; Tremblay and Schoen, 1999
; Maskas and Cruzan, 2000
; Walter and Epperson, 2001
). Furthermore, most phylogeographic studies of plant migration in North America and elsewhere have dealt with tree species (reviewed in Taberlet et al., 1998
). To our knowledge, no phylogeographic study to date has used molecular markers to investigate angiosperm species whose ranges span both glaciated and ice-free portions of eastern North America. Therefore, the principal goal of this study was to examine the phylogeography of a wide-ranging woodland herb to evaluate the role of long-distance dispersal in the migratory history of this species.
Trillium grandiflorum (Michx.) Salisb. (Melanthiaceae), a long-lived perennial that is widespread as an understory woodland herb in eastern North America (Fig. 1 inset; Case and Case, 1997
), is a useful candidate for phylogeographic studies of post-glacial migration. The species is predominately bumble bee pollinated and self-incompatible. Marker gene studies of mating patterns indicate that most populations are predominantly outcrossing (Broyles et al., 1997
; Kalisz et al., 1999
; Sage et al., 2001
). It is not known to hybridize with other members of the genus (Case and Case, 1997
). All Trillium species have ant-dispersed seeds (i.e., myremecochory: Gates, 1941
; Case and Case, 1997
) and ant-mediated dispersal of up to 
2 m in T. grandiflorum has been documented (Kalisz et al., 1999
). Seed dispersal by wasps of up to several hundred meters has also been reported in the genus (Jules, 1996
). These reports suggest limited dispersal range and raise the question of how T. grandiflorum could have achieved its current widespread distribution following glacial range contraction. Vellend et al. (2003)
recently demonstrated that dispersal of T. grandiflorum seeds can occur when white-tailed deer ingest the fruit and excrete them in their stool. They suggest that deer could have provided an important mechanism for long-distance dispersal in T. grandiflorum. In principle, phylogeographic approaches should enable long-distance dispersal to be distinguished from other modes of colonization.
|
). Moreover, the pollen record indicates that refugial populations of tree species (Acer spp., Fagus spp.) associated with Trillium and the guild of woodland herbs with which it occurs existed in Alabama and Arkansas and migrated northwards from this southern refugium following the retreat of the Wisconsin glacier (Davis, 1983
). Trillium species may have also survived in multiple refugia during the Wisconsin glaciation. For example, separate refugia could have existed in the Appalachians, along the Atlantic coast and elsewhere. To evaluate these ideas we used nested clade analysis in combination with allozyme data in an effort to discriminate between the single vs. multiple refugia hypotheses. This investigation examines the migratory history and mechanisms of gene flow in Trillium grandiflorum through a geographical survey of contemporary patterns of genetic diversity at cpDNA and allozyme loci. In our study, we addressed the following specific questions: (1) Using both NCA and intuitive approaches, is there a phylogeographic signal in the molecular data from which we can reconstruct the post-glacial history of this species? (2) What are the genetic consequences of glacial range contraction and subsequent recolonization in T. grandiflorum? In particular, is there evidence of reduced genetic diversity in northern population of the species? (3) What is the relative importance of gene flow via pollen vs. seeds in T. grandiflorum and has long-distance dispersal contributed towards northern migration?
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). In April–May 2000, we collected leaf samples from 35 T. grandiflorum populations throughout most parts of this range (Table 1). The mean distance between the 35 populations was 660 km (range 33–1460 km). We sampled flowering individuals only and these were collected at 2–3 m intervals within populations. Trillium grandiflorum does not propagate clonally and therefore each individual is a separate genet. We sampled a single leaf from up to 30 individuals from each population and leaf samples were subsequently stored on ice for up to 10 d.
|
. We used polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis to detect intraspecific cpDNA variation. First, we screened 20 universal cpDNA primer pairs (Taberlet et al., 1991
; Demesure et al., 1995
; Dumolin-Lapegue et al., 1997
). Ten primer pairs yielded reliable PCR fragments (TFab, TFcd, TFef, TC, VL, HK, K1K2, CS, ST, ML). One primer pair, CS, could not be optimized to eliminate nonspecific amplification. For the remaining nine primer pairs, we screened each PCR fragment with 17 restriction enzymes (four-, five-, and six-base pair cutters). All PCR reactions were carried out on a Hybaid PCRexpress (Thermo Electron Corp., Mississauga, Ontario, Canada) under the following conditions: 1.5 mmol/L MgCl2, 200 µmol/L of each dNTP, 0.00625 units/µL Taq, and 1 x PCR buffer; 1 min at 94°C, 1 min at the optimized annealing temperature (Tfab –57°C, TFcd –50°C, Tfef –61°C, TC –59°C, VL –57°C, HK –60°C, K1K2 –60°C, CS –60°C, ST –61°C, ML –61°C) and 2 min at 72°C for 35 cycles. The PCR fragments were assayed on 1.6% agarose gels stained with ethidium bromide. Through preliminary screening, we were able to resolve three sequence polymorphisms (gain/loss of a restriction site) in T. grandiflorum. Fragment VL exhibited a sequence polymorphism when cut by MspI. Fragment HK showed two sequence polymorphisms when cut by MspI or AluI. Thus, all cpDNA analyses are based on these three characters. To assay the extent of polymorphism, we assayed 8–10 individuals of T. grandiflorum from 35 populations for each polymorphic fragment/restriction enzyme combination. In total, 345 individuals were assayed for cpDNA polymorphisms.
Statistical analysis of cpDNA variation
We used HAPLODIV (Pons and Petit, 1995
) to calculate the within-population diversity (hS), which is the probability that two randomly chosen haplotypes in a population are different, and total gene diversity (hT), which is the probability that any two haplotypes are different. We used ARLEQUIN (Schneider et al., 2000
) to calculate the percentage of among- and within-population differentiation of the cpDNA genome with analysis of molecular variance (AMOVA) and to calculate the haploid equivalent of FST.
We used maximum parsimony to construct an intraspecific phylogeny of haplotypes as described by Templeton et al. (1995)
. To justify the use of parsimony, we followed the procedure outlined by Templeton et al. (1992)
to calculate H, the probability that a restriction site change has been caused by more than one mutation. So long as H < 0.05, then the use of maximum parsimony is justified. First, we estimated 
, which is the probability that two alleles chosen at random differ at a single nucleotide (Ewens, 1983
). In this case, = k/{2 rm [ln(n)]}, where k is the number of polymorphic sites, r is the length of the restriction enzyme recognition sequence, m is the total number of cut sites and n is the sample size. For T. grandiflorum, after the initial screen, k = 3, r = 5 (on average), m = 105 and n = 12. Substituting into Equation one of Templeton et al. (1992)
yields H = 0.011, well below the 0.05 criterion. Therefore, the use of parsimony was justified for this analysis.
To analyze the historical signal in T. grandiflorum cpDNA we used GeoDis (Posada et al., 2000
), which implements the nested-clade analysis (NCA) designed by Templeton et al. (1995)
. We assessed the significance of dispersion (DC) and displacement (DN) values by permutation tests of the data, based on 1000 permutations. To assess historical processes within each level of the nested cladogram we used the dichotomous key developed by Templeton (1998)
.
Allozyme analysis
In the laboratory, we ground small pieces of leaf tissue in an extraction buffer consisting of 0.7 mmol/L Borax, 4 mmol/L sodium metabisulfite, 40 mmol/L sodium diethyldithiocarbonate, 50 mmol/L sodium ascorbate, 0.2 mol/L Tris-HCl, 11 mmol/L DTT, and 3 mmol/L PVP-40. The extract was adsorbed onto Whitman #3 wicks and frozen at –80°C for later analysis. Following Broyles et al. (1997)
, we were able to assay eight loci for T. grandiflorum: a morpholine-citrate buffer system (pH 6.1; 50 mA) was used for isocitric dehydrogenase (Idh), malate dehydrogenase (Mdh-1 and Mdh-2), phosphoglucomutase (Pgm), 6-phosphogluconate dehydrogenase (Pgd), and shikimic dehydrogenase (Skdh). A discontinuous system consisting of lithium-borate electrode buffer (pH 8.3; 60 mA) and tris-citrate gel buffer was used to resolve glutamate dehydrogenase (Gdh) and phosphoglucose isomerase (Pgi). On average, we assayed 28.5 individuals (range 17–30) of T. grandiflorum per population.
Statistical analysis of electrophoretic data
To analyze genetic diversity, we used POPGENE (Yeh et al., 1997
) to calculate observed heterozygosity (Ho), expected heterozygosity (He) and the effective number of alleles per locus (A). To test the expectation that there is reduced genetic diverisity in glaciated vs. nonglaciated parts of the range, we used Pearson correlations in JMP (SAS, 2000
) to investigate the relations between A, Ho, and HE with latitude.
We used FSTAT (Goudet, 1995
, 2000
) to calculate f, , and F, which are analogous to Wright's (Wright, 1951
) measures FIS, FST, and FIT (Weir and Cockerham, 1984
). FSTAT calculates mean values for f, , and F by jackknifing over loci and calculates 95% confidence intervals (CI) for these measures by bootstrapping over loci. Standard errors (SE) and CIs were calculated by 1000 permutations of the data.
To assess genetic relatedness between populations, we used POPGENE to construct UPGMA dendrograms (Sneath and Sokal, 1973
) based on Nei's genetic distance calculated from population allele frequencies (Nei, 1972
). To detect if there was a significant isolation-by-distance effect, we used the ISOLDE program in GENEPOP (Raymond and Rousset, 1995
) to perform a Mantel test of [FST/(1 – FST)] against the natural logarithm of geographic distance. To determine if the variance in genetic distance increased with geographic distance, we performed a Mantel test on the residual of [FST/(1 – FST)] against the natural logarithm of geographic distance (Hutchinson and Templeton, 1999
). Distances between populations were calculated with ArcView (ESRI, 1998
).
As the results of cpDNA analysis for T. grandiflorum (see Results) revealed a striking pattern of haplotype distribution, we used population haplotype information to further analyze allozyme structure that could be related to cpDNA structure. First, we coded each population by cpDNA haplotype or clade (see Results); three polymorphic populations were coded by the predominant haplotype or clade while one population (NY-5) consisted of equal numbers of 1-1 and 1-2 clade haplotypes and was excluded from these analyses. We then used POPGEN (Yeh et al., 1997
) to perform hierarchical F-statistic analyses at two levels for consistency with nested clade analysis (NCA): populations grouped by haplotypes within one-step clades and populations grouped by one-step clade (1-1 or 1-2) within the two-step clade. Thus, FPT represents the differentiation between haplotypes within a one-step clade (at the 1-1 and 1-2 levels) or the differentiation between one-step clades at the two-step clade level.
Relative rates of gene flow via pollen and seed
We used equations developed by Ennos (1994)
to estimate the relative rates, r, of gene flow via pollen and seed. This method assumes that cpDNA is maternally inherited in Trillium, as it is in most angiosperms (Palmer, 1987
). The relative rate of pollen to seed migration, r, is r = pollen/seed = (A – 2C)/C, where A = 1/FST(b) – 1 and C = 1/FST(m) – 1 subscript b denotes a biparentally inherited marker and m a maternally inherited marker. However, this should be considered a special case if inbreeding is zero. As there is some biparental inbreeding in Trillium (see Results), we used the modified form of the equation that takes into account the inbreeding coefficient FIS as follows (see Ennos, 1994
for details), where pollen/seed = (A[1 + FIS] – 2C)/C.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
No clear spatial patterns were evident in the UPGMA dendrogram of allele frequencies at allozyme loci (Fig. 2). Furthermore, there was a significant although weak effect of isolation-by-distance (Mantel test; slope of regression = 0.028, P = 0.01) and the variance in genetic distance increased significantly with geographic distance (slope of regression = 0.17, P = 0.012).
|
). With five haplotypes and a closed loop, there are four possible networks. There is no phylogenetic rationale to choose among the four equally parsimonious trees. Caution in the following analysis is clearly warranted. While NCA of the four alternate trees were mostly consistent, in that long-distance range expansion and fragmentation were detected for all, two trees did yield inferences of contiguous range expansion. However, coalescent theory can be used to choose the most likely tree topology (Crandall and Templeton, 1993
). Therefore, the intraspecific phylogeny and interpretation we present below are the most likely based on coalescent predictions (and see Discussion).
|
|
and references therein). Following Crandall and Templeton (1993)
, tip haplotypes are defined as those connected to only one other haplotype, whereas interior haplotypes are connected to multiple haplotypes. The D haplotype is the most common (Table 3), suggesting that it is an interior haplotype and that C and E are tip haplotypes (Fig. 3, other networks not shown).
Further explanation is required for the designation of tip and interior clades at the two-step level. Inferences from NCA usually only apply to haplotypes within a species or species group (see Templeton et al., 1995
; Golden and Bain, 2000
). A recent phylogeny of Trillium by Osaloo and Kawano (1999)
placed the morphologically similar T. ovatum as the sister species of T. grandiflorum with very little molecular divergence between the two. Therefore, we argue that information from this closely related taxon could be used to infer the age of the T. grandiflorum one-step clades. Clade 1-1 is closer to the T. ovatum cpDNA haplotype than clade 1-2 (S. R. Griffin and S. C. H. Barrett, unpublished data). Presumably, older clades have had more time to accumulate mutations and the 1-1 clade comprises three haplotypes whereas the 1-2 clade comprises two haplotypes. Thus, we argue that clade 1-1 warrants interior status. Nevertheless, it is important to note that the qualitative interpretation using Templeton's (1998)
key for the 2-1 nesting level is the same (long-distance colonization; Table 3) regardless of whether clade 1-1 or 1-2 is designated as the interior clade.
Long-distance colonization and fragmentation can explain the distribution of haplotypes and clades in T. grandiflorum (Table 3). Within the 1-1 clade, past fragmentation has resulted in the separation of haplotype A from haplotypes B and C. Range expansion through long-distance colonization can be invoked to explain the distribution of haplotypes E and D (though see Discussion). Long-distance colonization is also inferred at the 2-1 clade level. Hence, NCA indicates that long-distance dispersal events characterize the post-glacial migration of T. grandiflorum in eastern North America.
Hierarchical F statistics of populations based on haplotypes/clades (Table 4) indicated little differentiation between haplotypes or clades. When analyzed at the one-step clade level (1-1 or 1-2), the differentiation between haplotypes (i.e., FPT) was small compared to the differentiation among all populations. Similarly, at the two-step clade level (2-1), there was little differentiation between the 1-1 and 1-2 clades (FPT) compared to differentiation among all populations (FST).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), Polygonella spp. (Lewis and Crawford, 1995
), and Asclepias exaltata L. (Broyles, 1998
), typically report reduced genetic variation in glaciated regions compared to regions that remained ice-free during the Wisconsin glaciation. In contrast, we found no significant correlation of allozyme diversity against latitude in T. grandiflorum, and indeed all three estimates of diversity exhibited weak positive trends (i.e., increased diversity) in higher latitudes. Several factors may account for the lack of correlation between nuclear genetic diversity and latitude in T. grandiflorum.
First, the southern populations of T. grandiflorum that we sampled may not have been growing in the refugial areas where the species survived during the Wisconsin glaciation. Based on the analysis of pollen cores (Davis, 1983
), refugial deciduous forests existed to the southwest (i.e., Alabama and Arkansas) of the current southern margins of the range of T. grandiflorum. Therefore, it is possible that all contemporary populations of this species are the result of post-glacial expansion and thus show reductions in genetic diversity relative to now-extinct refugial populations. Nonetheless, even if all present-day populations are indeed derived, both long-distance and stepping-stone models of migration predict a loss of genetic diversity in the most recently founded (i.e., northern) populations (Hewitt, 2000
). Second, if the current range does include some refugial populations (as our cpDNA haplotype analysis suggests; see below), contemporary processes may have obscured historical signals at allozyme loci. For example, while the center of origin for Trillium may be in the southeastern United States, the current center of abundance for T. grandiflorum lies much further north (e.g., Michigan, Ontario, New York, Quebec; Case and Case, 1997
). Population sizes in the north are typically much larger than in the south (S. R. Griffin, unpublished data), thus buffering populations against the effects of genetic drift. Moreover, contemporary gene flow between populations can also obscure historical signals, especially when populations are abundant and of large effective size, as they are in many northern locales.
Isolation-by-distance analysis indicates that gene flow occurs among populations of T. grandiflorum. Hutchinson and Templeton (1999)
developed a graphical model to estimate the relative influence of drift and gene flow on population genetic structure. Local gene flow between populations leads to a positive relation between genetic and geographic distance. On the other hand, drift will tend to increase the variance in allele frequencies and therefore increase the variance in genetic distance. Thus, if there is little gene flow at large distances, drift will lead to increasing variance in genetic distance as geographic distance increases. Our isolation-by-distance analysis of allozyme data indicates that T. grandiflorum displays this pattern: both genetic distance and the variance in genetic distance increased with the geographic distance between populations (i.e., Case I sensu Hutchinson and Templeton, 1999
). This result raises the question of what mechanisms might account for gene flow among populations of T. grandiflorum.
One possible explanation is that bumble bees, the predominant pollinators of T. grandiflorum, mediate pollination between populations. Although pollen flow between the particular populations we sampled seems unlikely because of the large geographical scale of our study, pollinator-mediated gene flow among unsampled populations may account, in part, for the pattern of isolation-by-distance we observed. An alternative mechanism of gene flow between populations has recently been suggested by Vellend et al. (2003)
who demonstrated that T. grandiflorum seeds could be dispersed via ingestion and defecation by deer. Indeed, our analysis of cpDNA diversity using NCA provided evidence that long-distance seed dispersal contributed to post-glacial recolonization in T. grandiflorum (Table 3 and see below). Nevertheless, our estimates of rates of gene flow via pollen vs. seed using the Ennos method (1994) implicates pollinators as the primary agents of between-population gene flow in T. grandiflorum. A third possibility is that the findings from our analysis of isolation-by-distance are unrelated to current evolutionary processes involving gene flow. The genetic structure at allozyme loci revealed in our geographical survey may have largely arisen during post-glacial range expansion and contemporary rates of pollen and seed dispersal have not had a strong impact in altering these historical patterns.
In contrast to the patterns observed at allozyme loci, cpDNA haplotype diversity in T. grandiflorum displays much greater population structure (Table 2). The high FST-cpDNA value (0.95) agrees with theoretical predictions and most empirical data from other plant species (Ennos, 1994
). Indeed, the differences in population structure between the allozyme and cpDNA markers suggest that gene flow via pollen is 45x higher than that via seed.
Phylogeographic patterns in Trillium grandiflorum
Despite considerable interest in the effects of past climate change on plant migration (reviewed in Comes and Kadereit, 1998
), relatively little is known about the historical processes that have shaped present-day genetic diversity in eastern North American plants, especially in glaciated regions. The geographical structuring of cpDNA haplotypes provided insights into the post-glacial history of T. grandiflorum. Given the recent concerns about NCA (Knowles and Maddison, 2002
), we evaluated our results from T. grandiflorum using both NCA-based and intuitive approaches. For the most part, NCA concurs with visual inspection of our data set, and our comparison strengthens our phylogeographic inferences for T. grandiflorum.
The haplotypes within the 1-1 clade (i.e., haplotypes A, B, and C; Fig. 3) are primarily limited to the ice-free portion of the range of T. grandiflorum, though some are found in populations from glaciated regions (Fig. 1). Furthermore, haplotype A is further restricted from the rest of the clade occurring largely along the eastern side of the Appalachian Mountains. This suggests a past fragmentation into two refugia on either side of the Appalachians during the Wisconsin glaciation. Indeed, NCA inferred a fragmentation event in the 1-1 clade (Table 3). The restriction of haplotype A to the eastern side of the Appalachians suggests that these mountains may have restricted this haplotype from dispersing from its refugium. Thus, most of the recolonization of glaciated regions (through long-distance dispersal; see below) may have occurred from a refugium west of the Appalachian Mountains, containing haplotypes B and C. In Europe, the Alps and Pyrenees are also thought to have blocked the dispersal of some animal and tree species from Italian and Iberian refugia, respectively (reviewed in Taberlet et al., 1998
; Hewitt, 2000
). However, haplotype C is found in two populations from North Carolina east of the Appalachians (Fig. 1) suggesting that there has been some secondary contact between the putative refugia. If this occurred, it implies that the barrier imposed by the Appalachians may not have been an absolute one. In contrast, Sewell et al. (1996)
found one common cpDNA haplotype in Liriodendron tulipifera L. spanning the area covered by the T. grandiflorum clade 1-1 (i.e., both east and west of the Appalachians) while another less common haplotype was restricted to populations in Florida. Thus, in this case the Appalachians did not restrict the migration of L. tulipifera in this region. Liriodendron tulipifera has wind-dispersed seeds that can presumably travel greater average distances than Trillium seeds and that may have facilitated L. tulipifera's dispersal across the Appalachians.
Haplotypes B and C occur in two separate clusters in Ontario and New York that are geographically separated by regions occupied by other haplotypes (Fig. 1). This pattern is consistent with models of range expansion where occasional long-distance dispersal results in a patchy distribution of genotypes in glaciated regions (Hewitt, 2000
) and has been observed in European oaks (Petit et al., 1997). The results of our NCA indicate that long-distance colonization has likely shaped the distribution of clades at the 2-1 level (Table 3).
Long-distance dispersal can also be inferred from NCA within the 1-2 clade (Table 3). However, this conclusion is at odds with an intuitive interpretation of the geographical distribution of haplotypes D and E (Fig. 1). The limited distribution of these haplotypes suggests a history of fragmentation, with the D and E haplotypes restricted to separate refugia during the last glaciation. These two haplotypes are restricted to the northeast and northwest corners of the range, respectively, with minimal geographical overlap between them. Alternatively, NCA inferred long-distance dispersal within this clade. It is it difficult to reconcile the NCA inference with our intuitive assessment of the geographic distribution of these haplotypes. Thus, caution in the use of Templeton's key (1998) seems warranted, as it may lead to suspect conclusions, especially for limited data sets as are commonly found in plant phylogeographic studies.
To attempt to discriminate among the different phylogeographic inferences of long-distance dispersal and fragmentation, and especially for clade 1-2, we examined hierarchical allozyme structure in our data set (Table 4). If T. grandiflorum was restricted to geographically separate refugia during glaciation and if different haplotypes/clades occurred in these refugia, then populations grouped by cpDNA haplotype should exhibit genetic differentiation between haplotype groups. On the other hand, if all haplotypes/clades within a nesting level originated from a single refugium, then relatively little genetic differentiation between haplotype/clades would be predicted. Our results revealed little differentiation in allozymes between haplotype/clade groups at all three nesting levels (i.e., FPT, Table 4). Hence, this analysis was equivocal with respect to the multiple refugia hypothesis. Elsewhere, Soltis et al. (1997)
reported similar findings for several plant species from the Pacific Northwest. Despite cpDNA structure that indicated distinct northern and southern refugia, allozymes and ribosomal DNA were not associated with the two refugia in this region. Discordance among nuclear and cytoplasmic loci largely reflects differences in the ability of these markers to detect historical events. In addition, such differences may also reflect contrasting patterns of gene dispersal via pollen or seed.
A closed haplotype loop presented an additional dilemma for the use of NCA, as different interpretations of the loop may affect the interpretations of NCA. As discussed above (see Results), we used coalescent predictions to choose the most likely haplotype network. Analyses of the other three networks yielded similar, though not identical, results. One consistent result was that one-step clades containing the B haplotype all showed a pattern of past fragmentation. This finding strengthens the argument that the Appalachians may have served as a barrier to recolonization. Long-distance colonization was also inferred in one-step clades from two of three alternate networks. A significant discrepancy, however, was that NCA inferred contiguous range expansion at the two-step clade level in two of the alternate trees. A conclusion of contiguous range expansion is difficult to reconcile with either the NCA analysis discussed above or an intuitive assessment of the distribution of haplotypes. If the phylogenetic relationships among haplotypes are ignored, the restricted distributions (Fig. 1) of haplotypes B, D, and E suggest fragmentation events and the disjunct populations containing B and C suggest long-distance seed dispersal. Resolution of additional haplotypes would likely refine our phylogeographic analysis of T. grandiflorum but, unfortunately, this species appears to have limited cpDNA variability.
In conclusion, our study initiates research on the post-glacial migration of woodland herbs in eastern North America. Notwithstanding the caveats of our phylogeographic analyses, fragmentation and range expansion, likely by long-distance dispersal, have been important factors in the post-glacial history of T. grandiflorum and the present-day distribution of genetic variation. Future work investigating phylogeographic patterns in the guild of woodland herbs commonly associated with T. grandiflorum (e.g., Asarum canadense L., Clintonia borealis (Ait.) Raf., Erythronium americanum Ker-Gawl., Sanguinaria canadensis L.) would be valuable to determine if these species also display similar migratory histories. Studies of this type will undoubtedly lead to a better understanding of the biogeographic history of this wide-ranging plant community that characterizes the forested regions of many parts of eastern North America.
|
FOOTNOTES |
|---|
2 barrett{at}botany.utoronto.ca
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Broyles S. B. 1998 Postglacial migration and the loss of allozyme variation in northern populations of Asclepias exaltata (Asclepiadaceae). American Journal of Botany 85: 1091-1097[Abstract]
Broyles S. B. S. L. Sherman-Broyles P. Rogati 1997 Evidence of outcrossing in Trillium erectum and Trillium grandiflorum (Liliaceae). Journal of Heredity 88: 325-329
Cain M. L. H. Damman A. Muir 1998 Seed dispersal and the Holocene migration of woodland herbs. Ecological Monographs 68: 325-347
Case F. W. R. B. Case 1997 Trilliums. Timber Falls Press, Portland, Oregon, USA
Clark J. S. et al 1998 Reid's paradox of rapid plant migration. BioScience 48: 13-24[CrossRef][ISI]
Comes H. P. J. W. Kadereit 1998 The effect of quaternary climatic changes on plant distribution and evolution. Trends in Plant Science 3: 432-438[CrossRef][ISI]
Cox C. B. P. D. Moore 2000 Biogeography: an ecological and evolutionary approach. Blackwell Science, London, UK
Crandall K. A. A. R. Templeton 1993 Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction. Genetics 134: 959-969[Abstract]
Cruzan M. B. A. R. Templeton 2000 Paleoecology and coalescence: phylogeographic analysis of hypotheses from the fossil record. Trends in Ecology and Evolution 15: 491-496
Davis M. B. 1983 Quaternary history of deciduous forests of eastern North America and Europe. Annals of the Missouri Botanical Garden 70: 550-563
Demesure B. N. Sodzi R. J. Petit 1995 A set of universal primers for amplification of polymorphic noncoding regions of mitochondrial and chloroplast DNA in plants. Molecular Ecology 4: 129-131[Medline]
Dingle H. 1996 Migration: the biology of life on the move. Oxford University Press, New York, New York, USA
Doyle J. J. J. L. Doyle 1987 A rapid isolation procedure for small quantities of fresh tissue. Phytochemical Bulletin 19: 11-15
Dumolin-Lapegue S. M. H. Pemonge R. J. Petit 1997 An enlarged set of consensus primers for the study of organelle DNA in plants. Molecular Ecology 6: 393-397[CrossRef][Medline]
Echt C. S. L. L. DeVerno M. Anzidei G. G. Vendramin 1998 Chloroplast microsatellites reveal population genetic diversity in red pine, Pinus resinosa Ait. Molecular Ecology 7: 307-316
Ennos A. R. 1994 Estimating the relative rates of pollen and seed migration among plant populations. Heredity 72: 250-259[ISI]
ESRI. 1998 ArcView GIS. Environmental Systems Research Institute, Redlands, California, USA
Ewens W. J. 1983 The role of models in the analysis of molecular genetic data, with particular reference to restriction fragment data. In B. S. Weir [ed.], Statistical analysis of DNA sequence data, 45–73. Marcel Dekker, New York, New York, USA
Gates B. N. 1941 Observation in 1940 on the dissemination by ants of the seeds of Trillium grandiflorum. Rhodora 43: 206-207
Golden J. L. J. F. Bain 2000 Phylogeographic patterns and high levels of chloroplast DNA diversity in four Packera (Asteraceae) species in southwestern Alberta. Evolution 54: 1566-1579[ISI][Medline]
Goudet J. 1995 FSAT version 1.2: a computer program to calculate F-statistics. Journal of Heredity 86: 485-486
Goudet J. 2000 FSTAT, a program to estimate and test gene diversities and fixation indices (version 2.9.3). Computer program and documentation available from author; website: http://www.unil.ch/izea/softwares/fstat.html [accessed 3 June 2003]
Hewitt G. M. 2000 The genetic legacy of the Quaternary ice ages. Nature 405: 907-913
Hutchinson D. W. A. R. Templeton 1999 Correlation of pairwise genetic and geographic distance measures: inferring the relative influences of gene flow and drift on the distribution of genetic variability. Evolution 53: 1898-1914[CrossRef][ISI]
Ibrahim K. M. R. A. Nichols G. M. Hewitt 1995 Spatial patterns of genetic variation generated by different forms of dispersal during range expansion. Heredity 77: 282-291[CrossRef][ISI]
Jaquard P. G. Heim J. Antonovics 1984 Genetic differentiation and dispersal in plants. Springer-Verlag, Berlin, Germany
Jules E. S. 1996 Yellow jackets (Vespula vulgaris) as a second seed disperser for the myrmecochorous plant, Trillium ovatum. American Midland Naturalist 135: 367-369[CrossRef][ISI]
Kalisz S. F. M. Hanzawa S. J. Tonsor D. A. Thiede S. Voigt 1999 Ant-mediated seed dispersal alters pattern of relatedness in a population of Trillium grandiflorum. Ecology 80: 2620-2634[CrossRef][ISI]
Knowles L. L. W. P. Maddison 2002 Statistical phylogeography. Molecular Ecology 11: 2623-2635[CrossRef][Medline]
Lewis P. O. D. J. Crawford 1995 Pleistocene refugium endemics exhibit greater allozymic diversity than widespread congeners in the genus Polygonella (Polygonaceae). American Journal of Botany 82: 141-149
Maskas S. D. M. B. Cruzan 2000 Patterns of intraspecific diversification in the Piriqueta caroliniana complex in southeastern North America and the Bahamas. Evolution 54: 815-827[CrossRef][ISI][Medline]
Nei M. 1972 Genetic distance between populations. American Naturalist 106: 283-292[CrossRef][ISI]
Osaloo S. K. S. Kawano 1999 Molecular systematics of Trilliaceae II. Phylogenetic analyses of Trillium and its allies using sequences of rbcL and matK genes of cpDNA and internal transcribed spaces of 18S–26S nrDNA. Plant Species Biology 14: 75-94[CrossRef]
Palmer J. D. 1987 Chloroplast DNA evolution and biosystematic uses of chloroplast DNA variation. American Naturalist 130: S6-S29[CrossRef]
Petit R. J. E. Pineau B. Demesure R. Bacilier A. Ducousso A. Kremer 1997 Chloroplast DNA footprints of postglacial recolonization by oaks. Proceedings of the National Academy of Sciences, USA 94: 9996-10001
Pons O. R. J. Petit 1995 Estimation, variance and optimal sampling of gene diversity. 1. Haploid locus. Theoretical and Applied Genetics 90: 462-470[ISI]
Posada D. K. A. Crandall A. R. Templeton 2000 GeoDis: a program for the cladistic nested analysis of geographic distribution of genetic haplotypes. Molecular Ecology 9: 487-488[CrossRef][Medline]
Raymond M. F. Rousset 1995 GENEPOP, version 1.2: population genetics software for exact tests and ecumenicism. Journal of Heredity 86: 248-249
Ridley H. N. 1930 The dispersal of plants throughout the world. L. Reeve, Ashford, Kent, UK
Sage T. L. S. R. Griffin V. Pontieri P. Drobac W. W. Cole S. C. H. Barrett 2001 Stigmatic self-incompatibility and mating patterns in Trillium grandiflorum and Trillium erectum (Melanthiaceae). Annals of Botany 88: 829-841
SAS. 2000 JMP, version 4.04. SAS Institute, Cary, North Carolina, USA
Sauer J. D. 1988 Plant migration: the dynamics of geographic patterning in seed plant species. University of California Press, Berkeley, California, USA
Schneider S. D. Roessli L. Excoffier 2000 Arlequin, version 2.000: a software for population genetics data analysis. Genetics and Biometry Laboratory, University of Geneva, Switzerland
Schwaegerle K. E. B. A. Schaal 1979 Genetic variability and founder effect in the pitcher plant Sarracenia purpurea L. Evolution 33: 1210-1218[CrossRef][ISI]
Sewell M. M. C. R. Parks M. W. Chase 1996 Intraspecific chloroplast DNA variation and biogeography of North American Liriodendron L (Magnoliaceae). Evolution 50: 1147-1154[CrossRef][ISI]
Sneath P. H. A. R. R. Sokal 1973 Numerical taxonomy. Freeman, San Francisco, California, USA
Soltis D. E. M. A. Gitzendanner D. D. Strenge P. S. Soltis 1997 Chloroplast DNA intraspecific phylogeography of plants from the Pacific Northwest of North America. Plant Systematics and Evolution 206: 353-373[CrossRef][ISI]
Taberlet P. L. Gielly G. Pautou J. Bouvet 1991 Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105-1109[CrossRef][ISI][Medline]
Taberlet P. L. Fumagalli A. G. Wust-Saucy J. F. Cosson 1998 Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 7: 453-464[CrossRef][Medline]
Templeton A. R. 1998 Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Molecular Ecology 7: 381-397[CrossRef][Medline]
Templeton A. R. K. A. Crandall C. F. Sing 1992 A cladistic analysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132: 619-633[Abstract]
Templeton A. R. E. Routman C. A. Phillips 1995 Separating population structure from population history: a cladistic analysis of the geographical distribution of mitochondrial DNA haplotypes in the tiger salamander, Ambystoma tigrinum. Genetics 140: 767-782[Abstract]
Tremblay N. O. D. J. Schoen 1999 Molecular phylogeography of Dryas integrifolia: glacial refugia and postglacial recolonization. Molecular Ecology 8: 1187-1198[CrossRef][Medline]
Van der Pijl L. 1969 Principles of dispersal in higher plants. Springer-Verlag, Berlin, Germany
Vellend M. J. A. Myers S. Gardescu P. L. Marks 2003 Dispersal of Trillium seeds by deer: implications for long-distance migration of forest herbs. Ecology 84: 1067-1072[CrossRef][ISI]
Walter R. B. K. Epperson 2001 Geographic pattern of genetic variation in Pinus resinosa: area of greatest diversity is not the origin of postglacial populations. Molecular Ecology 10: 103-111[CrossRef][Medline]
Weir B. S. 1990 Genetic data analysis. Sinauer, Sunderland, Massachusetts, USA
Weir B. S. C. C. Cockerham 1984 Estimating F-statistics for the analysis of population structure. Evolution 38: 1358-1370[CrossRef][ISI]
Wright S. 1951 The genetical structure of populations. Annals of Eugenics 15: 323-354[ISI]
Yeh F. C. R.-C. Yang T. J. B. Boyle Z.-H. Ye J. X. Mao 1997 POPGENE, the user-friendly shareware for population genetic analysis. Molecular Biology and Biotechnology Centre, University of Alberta, Calgary, Alberta, Canada
This article has been cited by other articles:
|
|
 |
|
  K. Aoki, T. Matsumura, T. Hattori, and N. Murakami Chloroplast DNA phylogeography of Photinia glabra (Rosaceae) in Japan Am. J. Botany, December 1, 2006; 93(12): 1852 - 1858. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Shaw and R. L. Small Chloroplast DNA phylogeny and phylogeography of the North American plums (Prunus subgenus Prunus section Prunocerasus, Rosaceae) Am. J. Botany, December 1, 2005; 92(12): 2011 - 2030. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||
