| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
2Department of Biology, University of New Mexico, Albuquerque, New Mexico 87131 USA 3School of Biological Science, University of New South Wales, Sydney, New South Wales 2052, Australia 4Jepson Herbarium and Department of Integrative Biology, University of California, Berkeley, California 94720 USA 5Department of Evolution, Ecology & Organismal Biology, Ohio State University, 1735 Neil Avenue, Columbus, Ohio 43210-1220 USA
Received for publication June 1, 2000. Accepted for publication December 5, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Asteraceae • Camptacra • ITS sequence • Minuria • Olearia • Peripleura • phylogeny • Tetramolopium • Vittadinia
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), which are mostly Australian, with one species each in New Zealand and New Caledonia. Tetramolopium comprises 38 perennial, woody species, of which all but one are insular. The genus has a remarkable distribution extending from the Hawaiian Islands (11 species) and Cook Islands (1 species) to New Guinea (25 species) and Australia (1 species) (Pedley, 1993
; Lowrey, 1995).
Richard (1832)
first described Vittadinia to include one species, V. australis, from New Zealand. Nees von Esenbeck (1832)
described Tetramolopium including one species, T. tenerrimum, from the Hawaiian Islands. De Candolle (1836)
described several new species of Vittadinia and transferred two species from Brachycome Cass. to Vittadinia. Additionally, he described the monotypic genus Eurybiopsis to include E. macrorhiza from Australia.
Generic delimitation within this group has been problematic historically. The convoluted history of the two genera since De Candolle (1836)
is summarized in Table 1. Vittadinia alone has undergone six generic recircumscriptions, and relationships with a number of other Australian genera have been stated or implied (Zhang and Bremer, 1993
; Nesom, 1994b; Table 1). Zhang and Bremer (1993)
stated, "There is a need for a more complete analysis of the Astereae with new hypotheses of generic interrelationships among groups from different continents." Following this, Nesom (1994b) presented his new hypotheses of generic relationships in the tribe within a new subtribal classification. His Vittadinia group (Table 1) differs from previous classifications by the segregation of species previously included in Vittadinia to form Peripleura (Nesom, 1994a
) and by the addition of two very small genera of annuals (Iotasperma [two spp.] and Dichromochlamys [monotypic]).
|
), suggesting that the species was derived from a recent secondary dispersal event from the Hawaiian Archipelago. However, evidence from nuclear restriction fragment length polymorphism (RFLP) marker analyses indicates that the Cook Island species is a distinct entity forming a sister group to the Hawaiian species suggesting a possible dispersal event from a New Guinea ancestor (Okada, Whitkus, and Lowrey, 1997
). Further data are needed to clarify the conflict between the morphology and ecology on the one hand and the molecular data on the other.
Chan et al. (1999)
recognized the need to evaluate phylogenetic relationships in this group using molecular sequence data in their preliminary investigation of the relationship of the Hawaiian and Cook Island species of Tetramolopium. They used the internal transcribed spacer (ITS) region of nuclear ribosomal DNA and included several species of the Australian Vittadinia group as outgroup taxa. In agreement with Baldwin et al. (1995)
, ITS data were shown to be useful for examining phylogenetic relationships among congeneric species and closely related genera. The purpose of this study is to reassess phylogenetic relationships within the Pacific taxa of Tetramolopium and among the Vittadinia-Tetramolopium group of genera using ITS sequences and morphological data. We feel it is important to include morphological data to assess the phylogenetic usefulness of such data in light of its various historical interpretations. These varying interpretations have produced the complex taxonomic history of the group. Additionally, we desired to: (1) evaluate the hypotheses of Okada, Whitkus, and Lowrey (1997)
about the origin of the Cook Island species of Tetramolopium, (2) test the monophyly of the genera within the group, (3) use the molecular estimate of phylogeny to evaluate the morphological characters that have been cited as indicators of affinity, and (4) assess phylogenetic relationships among the genera.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
Total DNA was extracted from fresh leaves, leaves dried in silica gel crystals, or leaves taken from existing herbarium specimens using either the CTAB method (Doyle and Doyle, 1990
) or the DNeasy Plant Minikit (QIAGEN, Clifton Hill, Victoria, Australia). The Hawaiian taxa of Tetramolopium were sequenced manually; all other taxa were sequenced using an automated sequencer. For manual sequencing, both double-stranded and single-stranded DNA were amplified. Double-stranded DNA was amplified with Taq DNA polymerase from Boehringer Mannheim (Roche, Indianapolis, Indiana) following manufacturer's recommendations. Following an initial melt at 95°C for 2 min, cycling began with 1 min at 95°C, 1 min at 48°C, 45 s at 72°C. The extension time was increased by 4 s for each of 30 subsequent reaction cycles. This was followed by a final extension of 72°C for 7 min. A fragment of 650 base pairs was then isolated from a low-melting agarose gel. Single-stranded DNA was amplified using asymmetric polymerase chain reaction (PCR) and a 1/10 or 1/100 dilution of the double-stranded product. Polymerase chain reaction conditions were identical to those used for double-stranded amplification except that asymmetric primer concentrations were used (500 nmol/L and 20 nmol/L), and the annealing temperature was 54°C. The single-stranded PCR product was purified by differential filtration in a Millipore Ultrafree-MC tube (Millipore UFC3 THK 00–30 000 nmol; Millipore, Bedford, Massachusetts, USA).
Manual sequencing was performed using the SequiTherm cycle sequencing kit (Epicentre Technologies, Madison, Wisconsin, USA) following the procedure for 32P-labeled primers but substituting primers labeled using biotinylated phosphoramidite (Clontech Laboratories Biotin-ON, Palo Alto, California, USA). Cycle sequencing parameters were 95°C for 5 min, 35 cycles at 95°C for 30 s, 54°C for 30 s, and 70°C for 1 min. Upon completion, 3 µL of stop dye solution was added. The sequencing reactions were separated using a 5% Long Ranger (FMC Bioproducts, Rockland, Maine, USA) and visualized using the procedure specified in the Sequenase Images non-isotopic DNA sequence detection kit (US Biochemical, Cleveland, Ohio, USA) with modifications to minimize background. These included a longer blocking step (1 h), decreasing the streptavidin-alkaline phosphatase conjugate (SAAP) concentration to 
1/2850 the volume of the blocking buffer, and performing a 20–30 s distilled water wash after each 1x post-SAAP/SDS wash. Optimal results were obtained by exposing the membrane to Kodak BioMax MR film for 2 h 15 min.
For automated sequencing, double stranded templates were amplified using the universal primers developed from fungal rDNA (White et al., 1990
). Samples for automated sequencing were amplified using PCR Supermix (GibcoBRL, Rockville, Maryland, USA) or AmpliTaq Gold (Perkin Elmer, Foster City, California, USA), following manufacturers' recommendations. Polymerase chain reaction conditions included an initial melt at 95°C for 2 min (PCR Supermix) or 10 min (AmpliTaq Gold), 35 cycles of 1 min at 95°C, 1 min at 48°C, and 1 min at 72°C, followed by a final 5-min extension at 72°C. Polymerase chain reaction products were cleaned using QIAquick PCR clean-up kits (Qiagen, Chatsworth, California, USA). Sequence reactions were performed using the Thermosequenase dye-terminator pre-mix kit (Amersham, Cleveland, Ohio, USA) following manufacturer recommendations, except that final reaction volumes were adjusted to 5 µL. Sequence reactions were analyzed using the ABI Model 377 DNA sequencer (Applied Biosystems, Foster City, California, USA) as described in Nelson et al. (1997)
.
Sequences were assembled and checked using ABI Prism software (Factura and Autoassembler) (Applied Biosystems). The sequences were aligned by determining the boundaries separating the rRNA genes from the ITS1 and ITS2 regions by comparison with the sequences of Daucus carota and Vicia faba (Yokota et al., 1989
), where the limits of mature 18S, 5.8S, and 26S rRNAs have been defined by S1 nuclease mapping. Internal transcribed spacer sequences were either manually aligned, using as a guide the above boundaries, and stored in DAPSA (DNA and protein sequence alignment program; E. Harley, University of Cape Town, Cape Town, South Africa) or were aligned using ClustalW (Thompson, Higgins, and Gibson, 1994
) and optimized by eye. Segments affected by insertion/deletion events (indels) were treated as missing data in the analyses, and potentially informative indels were scored as additional characters (present/absent) that were added to the sequence database. All sequences are lodged in GenBank (Table 2).
Heuristic searches were conducted in Phylogenetic Analysis Using Parsimony* (PAUP*) version 4.0b2a (Swofford, 1999
) for each data set individually and combined using tree bisection reconnection (TBR) branch-swapping and the MULPARS (save all most parsimonious trees) option. Random taxon addition (100 replicates) was employed to search for multiple islands of trees in all analyses of the molecular and combined data; random taxon addition (20 replicates) was used on the morphological data. Branch lengths for trees were calculated using the accelerated transformation optimization (ACCTRAN) option in PAUP. Relative support for the clades identified by parsimony analysis of the molecular data was estimated by the jackknife option in PAUP using 30% character deletion, fast stepwise addition, and 100 000 replicates, and also by decay analyses (Donoghue and Sanderson, 1992
) using PAUP and AutoDecay version 4.0.1 (Eriksson, 1998
) with a simple heuristic search on each constraint tree. Output trees were imported into MacClade to explore evolution of nonmolecular characters and to construct constraint trees in order to test alternative hypotheses against the data. Analyses were then performed in PAUP using 100 random addition replicates and the option "topological constraint enforced." The significance of differences between the strict consensus trees obtained in the constraint analysis and that of the most parsimonious trees was determined using the "compare two trees" option in PAUP. All analyses were polarized by the outgroup method, using two species of arborescent Olearia. A more inclusive analysis of Astereae, polarized on sequences from Calenduleae and Anthemideae, revealed these taxa to belong to an early diverging lineage within the tribe (Noyes and Reiseberg, 1999
; E. Cross and C. J. Quinn, unpublished data).
Combinability of the morphological and molecular data sets was assessed using the permutation tail probability (PTP) test (Faith and Cranston, 1991) to test for the presence or absence of phylogenetic signal and the partition homogeneity test (the incongruence length difference test of Farris et al., 1994
) to test for incongruence between the data sets. These two tests were applied using PAUP* version 4.0b2a. The partition homogeneity test was implemented with invariant characters excluded (Cunningham, 1997
) using TBR branch-swapping with 1000 replicates. The maximum of trees held in memory (MAXTREES) option was set to 25 to allow the test to proceed to completion. Analyses performed using an unrestricted MAXTREES setting exceeded the computer memory after only 3 repetitions.
Maximum likelihood (ML) analysis was performed using the stochastic search algorithm of Salter and Pearl (2001)
under the molecular clock assumption and the F84 model of substitution, estimating the transition/tranversion parameter simultaneously with the tree (ti/tv = 1.47).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
. Sequences included the complete ITS region except for the Western Australian accession of Minuria macrorhiza, where the first 30 bases of ITS-1 were missing, and Brachycome, Minuria cunninghamii, Vittadinia cuneata, and V. gracilis, for which the 5.8S region was excluded. The aligned data matrix comprised 44 sequences and 634 characters: a single sequence for each species, except Minuria macrorhiza, where there was a distinct divergence between those obtained from the Western Australian and Northern Territory accessions. Alignment required 20 indels, 4 affecting two positions and the rest only one. Twelve of these were potentially parsimony informative (Table 3); when these were scored, the matrix comprised 646 characters, 132 of which (20.5%) were parsimony informative, and 109 (16.9%) within the ingroup.
|
The heuristic search yielded a single island of 2520 trees of 490 steps having a rescaled consistency (RC) index of 0.44. The strict consensus is shown in Fig. 1 and a phylogram of one of the 2520 most parsimonious trees is shown in Fig. 2. There is no support for the monophyly of any ingroup genus. The two species of Camptacra form a very robust clade with the Queensland accession of Olearia arguta (98% jackknife support, +3 decay), but Camptacra gracilis is much closer to the latter than it is to C. barbata (93%, +3). The Hawaiian and Cook Island species of Tetramolopium form a well-supported clade (H in Fig. 1; 82%, +2). Four of the five species of the Australian genus Peripleura are associated with this clade, although there is low support for this link; P. bicolor groups with P. sericea and P. obovata (clade P; 56%, +1) within a polytomy with P. hispidula and clade H. Peripleura diffusa, however, is placed at a distance from this grade in M1, being the well-supported sister of one of the clades of Minuria species (83%, +5). The two accessions of M. macrorhiza are sister taxa within this clade, although support for their association is only weak (72%, +2). Minuria cunninghamii and M. sp. nov. are strongly nested within (100%, +13) clade M2 along with Kippistia as their well-supported (83%, +4) sister group. Minuria integerrima is the first lineage to diverge. Three clades of Vittadinia species were retrieved (V1, V2, and V3), but other species of this genus are ungrouped within the main polytomy. All species of Peripleura, Tetramolopium, and Vittadinia, with the exception of Peripleura diffusa and Tetramolopium vagans, constitute a well-supported major clade (94%, +4; clade V in Fig. 1). The phylogram (Fig. 2) shows that branch lengths are generally very short in the V clade indicative of few changes among the sequences of the taxa in this group.
|
|
The best tree produced in the maximum likelihood analysis (log likelihood –3566.32) contained all the labeled clades in Fig. 1. It differed only in some of the unsupported topology, particularly in the placement of Brachycome as the first diverging lineage and Tetramolopium vagans in a sister relationship with the C clade. Within the H clade, T. "mitiaroense" and T. humile were sequentially the first two taxa to diverge. Again, T. rockii grouped with T. sylvae and T. filiforme grouped with T. lepidotum. An examination of the nine next best ML trees showed they also only differed from Fig. 1 in aspects of the unsupported topology.
Morphological analysis
A database of 39 characters was assembled (Appendix). A systematic examination of the receptacle under a scanning electron microscope revealed hitherto unrecognized variation in the surface morphology that permitted the recognition of four states within the study group. After dispersal of the cypselas, the receptacle may appear smooth under the dissecting microscope except for the scar where the floret pedicel was attached (Fig. 3C). In other species the receptacle bears a pattern of single ridges that separate adjoining floret insertions (Fig. 3D). In some cases these ridges extend on the outer side of the insertion to form a pointed scale of variable size (Fig. 3A, B). In Brachycome ciliaris each floret is surrounded by the expanded base of a scale, so that there is a double ridge between adjacent insertions. These states are scored under character 36.
|
|
We believe that the more data exploration the better, so we have included both separate and combined analyses in order to get the most information possible to elucidate phylogenetic relationships while keeping in mind the caveat of the partition homogeneity test results.
Analysis of the combined morphological and molecular database gave 32 trees of 756 steps (CI = 0.49; RC = 0.31) in a single island. The strict consensus of these trees is shown in Fig. 5. The topology shows increased resolution of relationships when compared to the tree derived from molecular data only. All the labeled clades in Fig. 1 are present except for V3, which no longer contains Vittadinia pustulata, and P, which now also includes Peripleura hispidula.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The distributions of the indels are quite informative of relationships within the group. Indels 2, 3, and 12 confirm the distinction between the ingroup and the root on arborescent Olearia, 5 and 10 support the M1 and V clades, respectively, and indel 4 supports the sister relationship between the two accessions of Minuria macrorhiza.
Tetramolopium vagans is placed at a distance from all other members of the genus in Fig. 1. It shows no close affinity with any included taxon. An undescribed taxon with putative affinity with this species is recorded farther north in Queensland. It seems likely that a new genus may be required for these taxa, but further investigation of their relationships is needed before any final assessment is made. In particular, their relationships to other poorly known members of the tribe in New Guinea need to be considered.
Although all other species of Tetramolopium were not grouped in the analysis, when they were constrained in the one clade, tree length was increased by only a single step. Hence the molecular evidence against monophyly of this group is very weak. The inclusion of T. vagans within the constrained clade, however, gave a tree length 14 steps longer than the most parsimonious solution (P < 0.03 on the compare-two-trees test). This indicates that the molecular data provide very strong evidence against the inclusion of this species within the genus. Again, while the three New Guinea species of Tetramolopium were not grouped, and there was no support in the jackknife analysis for any association between them, only a single step was needed to make all the remaining species of the genus monophyletic (P > 0.9). Indels 7 and 8 characterize the Hawaiian clade of Tetramolopium and also occur in the New Guinea species: indel 8 is present in all three, whereas 7 occurs in only T. macrum and T. pumilum. These two species also share indel 1. This is direct evidence in favor of the monophyly of the non-Australian members of this genus, despite their not being grouped in the analysis. It further suggests that T. alinae was the first lineage to diverge after indel 8 had arisen but before indel 7, and that the other two New Guinea species share a unique common ancestor, as indicated by indel 1. This hypothesis conflicts with the topology obtained in the analysis, but all the conflicting aspects decay at +1 step. Further data are needed to robustly resolve these relationships.
Although four of the five species of Peripleura are grouped with the Hawaiian clade and are strongly supported members of the V clade (94%, +4), P. diffusa is the well-supported sister (83%, +5) to the Minuria macrorhiza complex (clade M1). This last association is also supported by the distribution of indels 5 and 6 (Fig. 1). Constraint analyses revealed that an extra 19 steps are required to render Peripleura monophyletic (P < 0.03). Hence the molecular data provide strong evidence against the monophyly of this genus also. The remaining species of Peripleura constitute a paraphyletic assemblage, but constraint analyses revealed that only one extra step is required to render the group monophyletic (P > 0.9). It is therefore clear from the molecular data that Peripleura diffusa has no close affinity with other members of that genus.
Vittadinia s.s. (sensu stricto) is rendered paraphyletic in the molecular analysis: there are several supported groups of species labeled V1, V2, and V3 in Fig. 1, and the relationships of V. australis, which is the type species, V. eremea, and V. pterochaeta are unresolved. When the genus was constrained to a single clade, tree length was increased by four steps. However, the compare-two-trees test showed no significant difference from Fig. 1 (P > 0.7).
Species of Minuria constitute three separate lineages: the two accessions of M. macrorhiza are strongly associated with P. diffusa in clade M1, which is the unsupported sister group of the V clade; M. cunninghamii is grouped with M. sp. nov. in a strongly supported clade (100%, +13), and these are in turn grouped with Kippistia; and M. integerrima is placed as the first diverging lineage among ingroup taxa. The analysis provides no evidence of any close relationship between these three groups, although all the branches separating them receive little or no support. When all species assigned to the genus were constrained into one clade, tree length increased by 17 steps. The inclusion of Peripleura diffusa in the constrained group reduced tree length by four steps; inclusion of Kippistia in the group decreased tree length by seven steps (to +10); the inclusion of both these taxa reduced it a further five steps (to +5). Comparisons of the strict consensus trees from these searches exceeded available memory before completion, but it seems highly likely that an extra ten or more steps represents a significant difference from the most parsimonious topology. We conclude that the molecular evidence does not support the current concept of Minuria and that there is strong support for the association of Peripleura diffusa with the Minuria macrorhiza clade and of Kippistia with the M2 clade.
Minuria also shows much larger sequence divergences than any other genus in this group, having a maximum pairwise divergence of 13.2% for M. cunninghamii to M. integerrima (Fig. 2). Neither Kippistia nor Peripleura diffusa are more divergent from any species of Minuria. Divergence within the M. macrorhiza complex is 4% and within M2 is 3%. These values can be contrasted with the respective maximum divergences observed in Peripleura (excluding P. diffusa), Tetramolopium (excluding T. vagans), and Vittadinia of 3.6, 5.4, and 6.1% (Fig. 2). Minuria therefore contains taxa that are more divergent in their ITS sequences than any of the other genera. Both Kippistia and Peripleura diffusa fit well within the group from both a cladistic and sequence divergence point of view. However, further data are needed to establish the monophyly of such an expanded concept.
Constraint analysis of the relationship between M. integerrima and the M1 clade revealed that only one extra step is required to render this group monophyletic (P > 0.9). Further sampling of taxa, including those outside the ingroup in this study, are needed to see whether this is a fair reflection of the distinction between the two lineages.
Indels 9 and 11 appear to have arisen twice, given the robustness of clades C, V, and M2. The remaining ten indels appear to be reflecting aspects of the phylogeny, although in four cases this is obscured by the lack of resolution and weakness in the topology.
The divergence between the sequences obtained from the two accessions of Minuria macrorhiza is of the same order as between the Northern Territory accession and Peripleura diffusa (4.4%) and between Minuria cunninghamii and M. sp. nov. (3%); it also compares favorably with the pairwise divergences observed within Vittadinia (1.6–6.1%). The molecular data therefore provide strong support for the morphological distinctions observed between these two populations and for the recognition of a new species.
Analysis of morphological characters
There are several morphological characters that unite the different groups delineated in the molecular analysis. All taxa of Vittadinia s.s. possess an apical position of the embryo in the mature cypsela and do not fill the cypsela completely. The Hawaiian Tetramolopium clade H and the Camptacra clade C are characterized by having a smooth receptacle surface. In addition, clade C is united by the absence of biseriate two-celled hairs with bifurcate tips on the cypsela.
However, when the morphological states are mapped on the strict consensus as well as several representative most parsimonious trees for the combined analysis there remains a high level of homoplasy in most characters. It is instructive to examine some of these, especially characters that have been used in distinguishing taxa. Female sterile disc florets (character 11), a character which has been widely used to distinguish genera in the family (Bremer, 1994
), appears to have been gained and lost more than once. Monoecy has arisen in Tetramolopium as well as Minuria, and there has been a subsequent loss within the Hawaiian clade of the former. Constraint analyses revealed that only two extra steps (P = 0.92) are required to group Tetramolopium humile with the other species of the Hawaiian group with female fertile disc florets, so that only a single loss is required within this clade. The evidence against monophyly of the ingroup taxa possessing female sterile disc florets, with or without a subsequent reversal, is very strong (+22 and +30 steps, P = 0.02 and 0.01, respectively). Hence the combined data provide strong support for more than one character state change in each direction. Recent genetic analysis of Hawaiian Tetramolopium indicates that female disc floret fertility/sterility is largely controlled by just two loci with modifiers (Whitkus, Doan, and Lowrey, 2000
). This accords well with the apparent lability of sex expression in the evolution of the group. If the relatively simple genetic control of female fertility/sterility in Tetramolopium were found to be a general pattern in the Asteraceae, systematists would do well to be very cautious in using variation in monoclinous sex expression as a taxonomic character for generic delimitation, particularly in the Astereae (Nesom, 2000
).
Similar tests of the following characters revealed significant support (P < 0.01) for homoplasy in each case: gain/loss of woodiness or a woody rootstock (character 1); ratio of ray to disc florets (9); length of ligule on ray florets (10); corolla form in disc florets (24); filament insertion (25); compression of the cypsela (28); marginal ribs on the cypsela (30); presence and number of facial ribs on the cypsela (31); and loss of scales and/or pattern on the receptacle (36). Characters 1, 24, 28, 30, 31, and 36 have been used rather extensively for generic delimitation as well as for higher ranks in the Asteraceae (Zhang and Bremer, 1993
; Bremer, 1994
; Nesom, 1994b
). We believe that studies similar to those of Whitkus, Doan, and Lowrey (2000)
and Bachmann (1991)
are needed to elucidate the genetic control of morphological characters in the Asteraceae and provide insight into their lability and potential usefulness as taxonomic and phylogenetic markers.
The morphological data show a very high degree of homoplasy. The CI value of 0.31 in our study is among the lowest (indicative of a high level of homoplasy) of the values reported in recent analyses of homoplasy in published phylogenetic studies (Sanderson and Donaghue, 1996
; Givnish and Sytsma, 1997
). In addition, the present study has fewer taxa than other studies that have reported high levels of homoplasy correlated with relatively large numbers of taxa being considered (Givnish and Sytsma, 1997
). The high level of morphological homoplasy in the Vittadinia group explains the poor resolution obtained in the morphological analysis and the lack of support for some of the generic concepts within this assemblage. The probable, main sources of homoplasy are convergence/parallelism and hybridization. Homoplasy in gain/loss of woodiness and length of ray ligules could result from selection for particular states in similar arid environments. There has been no documented case of natural hybridization in this group. However, artificial interspecific hybridization is easily accomplished in Hawaiian Tetramolopium and Vittadinia (Lowrey, 1986
; T. K. Lowrey, C. J. Quinn, and J. Avritt, unpublished data) and artificial intergeneric hybrids of the following crosses Tetramolopium x Vittadinia, Vittadinia x Camptacra, and Tetramolopium x Camptacra have recently been obtained (T. K. Lowrey, C. J. Quinn, and J. Avritt, unpublished data). Therefore it is possible that hybridization has played a role in the evolution of the group. Furthermore past intergeneric hybridization events could explain, in part, the considerable degree of homoplasy in the morphological data. We believe that if it is possible to have reticulate evolution via intergeneric hybridization in this group of genera in the Astereae it may well be an important process elsewhere in the tribe and the family. Future systematic studies of generic groups in the Asteraceae should consider the possibility of reticulate evolution above the species level when high levels of homoplasy are encountered.
Origin of the Cook Island Tetramolopium species
Intensive biosystematic (Lowrey, 1986
) and molecular studies (Lowrey and Crawford, 1985
; Okada, Whitkus, and Lowrey, 1997
) provide strong support for the monophyletic origin of and genetic cohesion among the Hawaiian taxa. The New Guinean species are montane shrublets that occur in submesic or mesic tropic-alpine habitats. The Cook Island species is a dwarf shrub occurring in uplifted, fossilized coral outcrops in coastal scrub vegetation on Mitiaro only. The Hawaiian species exhibit the greatest morphological and ecological diversity, constituting a classic example of adaptive radiation by insular plants (Funk and Wagner, 1995
; Lowrey, 1995
; Okada, Whitkus, and Lowrey, 1997
). They vary from prostrate, rosette perennials to tall, upright shrubs and occur in a diversity of submesic to xeric habitats from sea level to 3000 m on the five main islands (Lowrey, 1995
).
Based on morphological similarity and geologic evidence it has been assumed that the Hawaiian taxa developed from a single colonization event from a New Guinean progenitor. Smith (1977)
suggested that Tetramolopium evolved in the New Guinean highlands as a neo-endemic in the early Pleistocene and was subsequently dispersed to the Hawaiian Islands. The discovery of a new species of Tetramolopium in the Cook Islands (Lowrey, 1995
) has complicated the previously simple picture of dispersal from New Guinea to the Hawaiian Islands. The Cook Islands taxon is morphologically and ecologically similar to species in one of the Hawaiian sections (section Tetramolopium) of Tetramolopium (Lowrey, 1995
), suggesting that the taxon was derived from a recent secondary dispersal event. However, evidence from nuclear RFLP marker analyses indicates that the Cook Island taxon is a distinct entity forming a sister group to the Hawaiian species (Okada, Whitkus, and Lowrey, 1997
). In fact, nearly all analyses of the RFLP data indicate that sect. Tetramolopium is paraphyletic with the inclusion of the Cook Island species. Furthermore, the RFLP trees consistently fail to recognize a close affinity to sect. Tetramolopium despite the strong morphological and ecological similarities. This result does not support the hypothesis that the taxon was derived from a relatively recent secondary dispersal event from a Hawaiian founder. However, the RFLP data set used by Okada, Whitkus, and Lowrey (1997)
did not use any outgroup taxa nor did it include any New Guinean species of Tetramolopium.
Okada, Whitkus, and Lowrey (1997)
offered three equally parsimonious hypotheses to account for the two dispersal events necessary to explain the bihemispheric distribution on the Hawaiian and Cook Islands. These hypotheses are: (1) the Cook Islands species is the result of a secondary long-distance founding event from Hawaiian Tetramolopium, (2) the original dispersal event was to the Cook Islands with the Hawaiian group resulting from a secondary dispersal event, and (3) the Cook Island species and the Hawaiian group represent independent dispersal events. Okada, Whitkus, and Lowrey (1997)
argued that the biogeographic, morphological, and reproductive biology in totality provide support for hypothesis 1 despite the conflicting RFLP data.
Our ITS data alone do not resolve the relationships of T. "mitiaroense" to the Hawaiian species of Tetramolopium, but the combined analysis does place them in a sister relationship. This favors the second hypothesis advanced by Okada, Whitkus, and Lowrey (1997)
and is congruent with their RFLP data. However, there is no character support for this pattern of relationship. Clearly further data are required to provide a definitive answer to this question. Nonetheless, both the indel data and the combined analysis support the hypothesis of Smith (1977)
that the Pacific Island species of Tetramolopium are descended from a New Guinea ancestor.
General conclusions
Vittadinia is the only genus to be resolved in both the morphological and combined analysis. None of the analyses provide support for the monophyly of Camptacra, Minuria, Olearia, Peripleura, or Tetramolopium. The considerable amount of homoplasy in the morphological data set may reflect the influence of past intergeneric hybridization events. Tetramolopium vagans is placed at a distance from all other species of that genus. In the molecular and combined analyses Minuria is resolved into three separate clades. Olearia arguta is strongly grouped with Camptacra, Kippistia is grouped with the M1 clade of Minuria, and Peripleura diffusa is strongly grouped with the Minuria macrorhiza complex. Both Peripleura diffusa and Kippistia appear to fit within an expanded definition of Minuria. Indel data suggest this broader concept may be monophyletic, but an analysis with broader taxon sampling is required to demonstrate this unequivocally. In both the molecular and combined analyses, there is strong support for a close association between all taxa belonging to the V clade, i.e., Vittadinia, all Peripleura except P. diffusa, and all Tetramolopium except T. vagans. The combined analysis places T. "mitiaroense" as sister to the Hawaiian species of Tetramolopium. Although this is congruent with the RFLP data of Okada, Whitkus, and Lowrey (1997)
, there is not strong character support for this pattern of relationship. Clearly, further data are required to provide a definitive answer to this question. Both the indel data and the combined analysis support the hypothesis of Smith (1977)
that the Pacific Island species of Tetramolopium are descended from a New Guinea ancestor. Olearia arguta is closely related to Camptacra and shows no affinity with the arborescent species of Olearia included in this analysis. This taxon, which has been collected from both eastern and western parts of northern Australia and reputedly is a complex including several species (N. Lander, personal communication), requires further investigation before any firm conclusions can be drawn.
): 0, absent; 1, present.
): 0, absent; 1, present as a single row; 2, present as two or more rows.
).
|
FOOTNOTES |
|---|
6 Author for correspondence (tlowrey{at}unm.edu
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Baldwin B. G. M. J. Sanderson J. M. Porter M. F. Wojciechowski C. S. Campbell M. J. Donoghue 1995 The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: 247-277[CrossRef][ISI]
Bentham G. 1867 Flora Australiensis,. vol. 3, 447–680. Lovell Reeve, London, UK
Bremer K. 1994 Asteraceae: cladistics and classification. Timber Press, Portland, Oregon, USA
Burbidge N. T. 1982 A revision of Vittadinia A. Rich. together with reinstatement of Eurybiopsis DC. and description of a new genus, Camptacra. Brunonia 5: 1-72
Chan R. T. Lowrey D. Natvig R. Kimball R. Whitkus C. Quinn 1999 Molecular phylogeny of Tetramolopium (Asteraceae). XVI International Botanical Congress Abstracts, 3057. IBC, St. Louis, Missouri, USA
Cunningham C. W. 1997 Can three incongruence tests predict when data should be combined?. Molecular Biology and Evolution 14: 733-740[Abstract]
De Candolle A. P. 1836 . Prodromus 5: 280-281
Donoghue M. J. M. J. Sanderson 1992 Suitability of molecular and morphological evidence in reconstructing plant phylogeny, 340–368. In P. S. Soltis, D. E. Soltis, and J. J. Doyle [eds.], Molecular systematics of plants. Chapman and Hall, New York, New York, USA
Doyle J. J. J. L. Doyle 1990 Isolation of plant DNA from fresh tissue. Focus 12: 13-15
Eriksson T. 1998 AutoDecay version 4.0.1. (Program distributed by author). Department of Botany, Stockholm University, Stockholm, Sweden
Faith D. P. P. S. Cranston 1991 Could a cladogram this short have arisen by chance alone: on permutation tests for cladistic structure. Cladistics 7: 1-28
Farris J. S. M. Kallersjo A. G. Kluge C. Bult 1994 Testing significance of incongruence. Cladistics 10: 315-319[CrossRef][ISI]
Funk V. A. W. L. Wagner 1995 Biogeographic patterns in the Hawaiian Islands. In W. L. Wagner and V. A. Funk [eds.], Hawaiian biogeography, 379–419. Smithsonian Institution Press, Washington, D.C., USA
Givnish T. J. K. J. Sytsma 1997 Homoplasy in molecular vs. morphological data: the likelihood of correct phylogenetic inference. In T. J. Givnish and K. J. Sytsma [eds.], Molecular evolution and adaptive radiation, 56–97. Cambridge University Press, New York, New York, USA
Gray A. 1861 Characters of some Compositae in the United States Exploring Expedition under Captain Wilkes, with observations by Asa Gray. Proceedings of the American Academy of the Arts 5: 114-146
Hillebrand W. 1888 Flora of the Hawaiian Islands: a description of their phanerogams and vascular cryptogams. Carl Winter, Heidelberg, Germany. Facsimile ed., 1965. Hafner Publishing, New York, New York, USA
Lander N. S. 1987 New combinations in Minuria DC. Nuytsia 6: 63-66
Lowrey T. K. 1986 A biosystematic revision of Hawaiian Tetramolopium (Compositae: Astereae). Allertonia 4: 204-264
———. 1995 Phylogeny, adaptive radiation and biogeography of Hawaiian Tetramolopium (Asteraceae: Astereae). In W. L. Wagner and V. A. Funk [eds.], Hawaiian biogeography: evolution on a hot spot archipelago, 195–220. Smithsonian Institution Press, Washington, D.C., USA
———, and D. Crawford 1985 Allozyme divergence and evolution in Tetramolopium (Compositae; Astereae) on the Hawaiian Islands. Systematic Botany 10: 64-72[CrossRef][ISI]
Mattfeld J. 1929 Compositae. In L. Diels [ed.], Beitrage zur Flora des Saruwaged-Gebirges. Botanische Jahrbucher für Systematik, Pflanzengeschichte und Pflanzengeographie 62: 452-501
Nelson M. A. et al 1997 Expressed sequences from conidial, mycelial and sexual stages of Neurospora crassa. Fungal Genetics and Biology 21: 348-363[CrossRef][ISI][Medline]
Nees Von Esenbeck C. G. 1832 Genera et species Asterarum. 202-204
Nesom G. 1994a Peripleura (Asteraceae: Astereae), a new Australian genus segregated from Vittadinia. Phytologia 76: 125-135
———. 1994b Subtribal classification of the Astereae (Asteraceae). Phytologia 76: 193-274
———. 2000 Generic conspectus of the tribe Astereae (Asteraceae) in North America, Central America, the Antilles, and Hawaii. Sida, Botanical Miscellany, No. 20. Botanical Research Institute of Texas, Fort Worth, Texas, USA
Noyes R. D. L. H. Rieseberg 1999 ITS sequence data support a single origin for North American Astereae (Asteraceae) and reflect deep geographic divisions in Aster s.l. American Journal of Botany 86: 398-412
Okada M. R. Whitkus T. K. Lowrey 1997 Genetics of adaptive radiation in Hawaiian and Cook Island species of Tetramolopium (Asteraceae; Astereae). I. Nuclear RFLP marker diversity. American Journal of Botany 84: 1236-1246[Abstract]
Pedley L. 1993 Sigesbeckia fugax, new species, and Tetramolopium vagans, new species, new Asteraceae from Queensland. Austrobaileya 4: 87-92
Richard A. 1832 Essai d'une Flore de la Nouvelle Zelande,. 250
Salter L. A. D. K. Pearl 2001 A stochastic search strategy for estimation of maximum likelihood phylogenetic trees. Systematic Biology 50: 7–17
Sanderson M. J. M. Donoghue 1996 The relationship between homoplasy and confidence in a phylogenetic tree. In M. J. Sanderson and L. Hufford [eds.], Homoplasy: the recurrence of similarity in evolution, 67–89. Academic Press, San Diego, California, USA
Smith J. M. B. 1977 Origins and ecology of the tropicalpine flora of Mt. Wilhelm, New Guinea. Biological Journal of the Linnean Society 9: 87-131[CrossRef]
Swofford D. 1999 PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer, Sunderland, Massachusetts, USA
Thompson J. D. D. G. Higgins T. J. Gibson 1994 CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673-4680
White T. J. T. Bruns S. Lee J. Taylor 1990 Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, D. H. Gelfind, J. J. Sninsky, and T. J. White [eds.], PCR protocols: a guide to methods and applications, 315–322. Academic Press, New York, New York, USA
Whitkus R. H. Doan T. K. Lowrey 2000 Genetics of sex expression in Hawaiian Tetramolopium (Asteraceae). Heredity 85: 37-42
Yokota Y. T. Kayata Y. Iida A. Kato S. Tanifuji 1989 Nucleotide sequences of the 5.8s rRNA gene and internal transcribed spacer regions in carrot and broadbean ribosomal DNA. Journal of Molecular Evolution 29: 294-301[CrossRef][ISI][Medline]
Zhang X. K. Bremer 1993 A cladistic analysis of the tribe Astereae (Asteraceae) with notes on their evolution and subtribal classification. Plant Systematics and Evolution 184: 259-283[CrossRef][ISI]
This article has been cited by other articles:
|
|
 |
|
  F. Eggens, M. Popp, M. Nepokroeff, W. L. Wagner, and B. Oxelman The origin and number of introductions of the Hawaiian endemic Silene species (Caryophyllaceae) Am. J. Botany, February 1, 2007; 94(2): 210 - 218. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Lindqvist and V. A. Albert Origin of the Hawaiian endemic mints within North American Stachys (Lamiaceae) Am. J. Botany, October 1, 2002; 89(10): 1709 - 1724. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||
