| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Department of Botany, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan; 3Departamento de Ecología Vegetal, Instituto de Ecología, Xalapa, Veracruz 91000, México; and 4Department of Cell and Molecular Biology, Tulane University, New Orleans, Louisiana 70118 USA
Received for publication March 20, 2001. Accepted for publication July 3, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: boreotropical flora • disjunction • Dugandiodendron • Magnolia • Magnoliaceae • matK • Splendentes • Talauma
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Li, 1952
; Graham, 1972
; Boufford and Spongberg, 1983
; Wu, 1983
). The "Arcto-Tertiary Geoflora" theory was initially proposed to explain the origin and evolution of these patterns (Engler, 1879
; Chaney, 1947
; Axelrod, 1960
). This theory suggests a homogenous early Tertiary flora evolved in arctic regions and subsequently migrated during the mid-Tertiary to middle latitudes in response to climatic changes. Additional fossil discoveries and new geological discoveries, however, suggested a more complex chain of events was needed to fully understand the evolution of floras in the Northern Hemisphere in the Tertiary (Tralau, 1963
; Wolfe, 1975, 1978, 1980, 1997
; Tiffney, 1985a, b
; Taylor, 1990
; Tanai, 1992
; Mai, 1995
; Manchester, 1999
). Tiffney (1985a)
suggested five major periods of migration existed between eastern Asia and eastern North America: pre-Tertiary, Early Eocene, Late Eocene-Oligocene, Miocene, and Late Tertiary-Quaternary. He also noted that most evergreen disjunct taxa (e.g., Magnoliaceae, Lauraceae, and Theaceae) migrated during the Early Eocene through both the Bering and North Atlantic routes and that many deciduous lineages migrated during the Miocene.
To account for these new discoveries, Wolfe (1975)
proposed the "Boreotropical Flora" theory to explain the circumboreal early Eocene mixed heterogeneous floristic units (deciduous and evergreen vegetation). This theory suggests that tropical (megathermal) floristic units appeared at high latitudes in the Northern Hemisphere in the early Tertiary and that taxa were exchanged with other areas via the Bering and/or North Atlantic land bridge, resulting in a circumboreal heterogenous flora. A marked temperature deterioration occurred at the end of the Eocene (the terminal Eocene event), which caused the tropical elements to move south and disappear at high latitudes (Wolfe, 1978, 1980, 1997
; Collinson, Fowler, and Boulter, 1981
; Graham, 1999
). It should be noted that the movement of the boreotropical flora to lower latitudes resulted in the origin of disjunct tropical plants, as the climate following the mid- to late Eocene did not warm to levels prevalent in the early Eocene (Miller, Fairbanks, and Mountain, 1987
), i.e., most tropical plants could not reach high enough latitudes again and therefore could not migrate to other continents (Tiffney, 1985a
; Wolfe, 1997
).
Only a few studies have used molecular phylogenetic techniques to study evolutionary patterns or estimate divergence times of disjuncts, e.g., Liriodendron (Parks and Wendel, 1990
); Magnolia section Rytidospermum (Qiu, Chase, and Parks, 1995
; Qiu, Parks, and Chase, 1995
); Leguminosae subfamily Papilionoideae tribe Robinieae and allies (Lavin and Luckow, 1993
; Lavin, 1995
); Caulophyllum, Menispermum, Penthorum, and Phryma (Lee et al., 1996
); Symplocarpus (Wen, Jansen, and Kilgore, 1996
); Aesculus (Xiang et al., 1998
); Cornus, Boykinia, Tiarella, Trautvetteria, and Calycanthus (Xiang, Soltis, and Soltis, 1998
); Hamamelis (Wen and Shi, 1999
), and Aralia section Dimorphanthus (Wen, 1999, 2000
). These molecular analyses mainly focused on temperate disjunctions (deciduous and herbaceous taxa). According to Tiffney (1985a)
and Wolfe (1997)
, most temperate disjunct taxa probably migrated between continents during the Miocene or later via the Bering land bridge. Indeed, divergence times estimated by molecular analyses roughly correlate with this scenario, although the divergence times (from 2 to 25 million years ago [mya]) are scattered throughout the Miocene to the Pliocene (see Wen, 1999
). The wide range of divergence times suggests multiple migrations occurred between the continents.
Animal migrations also indicate multiple migrations. Woodburne and Swisher (1995)
indicate that ten major migrations of land mammals between North America and Eurasia took place during the Cenozoic Era. In addition, >30 minor dispersals occurred, involving a few land mammal taxa.
Although only a relatively small number of disjunct taxa have been analyzed using molecular techniques (Wen, 1999
), they are concordant with the scenario in Tiffney (1985a)
and Wolfe (1997)
regarding temperate disjunct taxa. This scenario, however, also predicted another important event that has not been tested, i.e., the origin of tropical disjunctions. As described above, if tropical taxa are not recently derived from regionally adjacent temperate taxa, then tropical disjunctions are relicts of the boreotropical flora and would have been formed during the middle to late Eocene (
50–34 mya).
Extant species of Magnoliaceae are distributed in temperate and tropical Asia (two-thirds of the species) from the Himalayas to Japan and southeastward through the Malay Archipelago to the New Guinea area. The remainder of the family is found in the Americas with tropical elements extending to Brazil and the West Indies (Takhtajan, 1969
; Heywood, 1978
; Law, 1984
; Nooteboom, 1993
; Frodin and Govaerts, 1996
). In the late Cretaceous and Tertiary, the family occurred throughout the Northern Hemisphere (Greenland, Alaska, and Europe; Mai, 1995
; Graham, 1999
). The disjunct distribution of the family in temperate and tropical areas provides an opportunity to test the boreotropical floral concept using molecular techniques. We would expect that tropical disjunct pairs of Magnoliaceae diverged during the middle to late Eocene and that the temperate disjunct pairs diverged after the Eocene, that is, in the Oligocene, Miocene, or more recently.
Phylogenetic and phytogeographic studies of temperate disjunct pairs of Magnoliaceae (Liriodendron and Magnolia section Rytidospermum) indicate the late Miocene to early Pliocene to be the formation time for these temperate disjunctions (Parks et al., 1983, 1994
; Parks and Wendel, 1990
; Qiu and Parks, 1994
; Qiu, Chase, and Parks, 1995
; Qiu, Parks, and Chase, 1995
; Sewell, Parks, and Chase, 1996
; Azuma, Thien, and Kawano, 1999
; Kim et al., 2001
). In this study, we conduct a molecular phylogenetic analysis of Magnoliaceae using chloroplast DNA sequence data and estimate divergence times (calibrated with fossil seeds) of tropical and temperate disjunctions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, Magnoliaceae is composed of 223 species of trees or shrubs classified into the following seven genera; Elmerrillia Dandy (4 spp.), Kmeria Dandy (2 spp.), Liriodendron L. (2 spp.), Magnolia L. (128 spp.), Manglietia Blume (29 spp.), Michelia L. (47 spp.), and Pachylarnax Dandy (2 spp.). Within the family, two well-defined monophyletic subfamilies are recognized: Liriodendroideae (only Liriodendron) and Magnolioideae. Delimitation of genera, subgenera, and sections in the subfamily Magnolioideae have been extremely controversial, hence different taxonomic systems have been proposed for the Magnolioideae (Dandy, 1927, 1950, 1978
; Lozano-Contreras, 1975, 1983
; Law, 1984
; Nooteboom, 1985, 1987, 2000
; Chen and Nooteboom, 1993
; Vázquez-Garcia, 1994
; Frodin and Govaerts, 1996
). In fact, recent molecular phylogenetic studies do not support the monophyly of Magnolia (Qiu, Chase, and Parks, 1995
; Azuma, Thien, and Kawano, 1999
; Kim et al., 2001
).
For example, Law (1984)
recognized two tribes, four subtribes, and 14 genera in the subfamily Magnolioideae. On the other hand, Nooteboom (1985)
merged most genera recognized by Law (1984)
into Magnolia or Michelia, creating two tribes and six genera in the Magnolioideae and three subgenera and 16 sections in the genus Magnolia. In addition, Nooteboom (1985)
placed section Manglietiastrum Noot. in the genus Magnolia subgenus Talauma (Juss.) Pierre, but later it was placed in the genus Manglietia (Chen and Nooteboom, 1993
).
In New World Magnoliaceae, Lozano-Contreras (1975, 1983)
described the genus Dugandiodendron Lozano (14 spp.), distributed in Colombia and adjacent areas. Lozano-Contreras (1975, 1983)
suggested that Dugandiodendron differs from other members of Magnoliaceae in the position of the flower and in the prefoliation (arrangement or form of leaves) in the vegetative buds, but Nooteboom (1985)
did not adopt this idea and retained it in the genus Magnolia. In turn, Vázquez-Garcia (1994)
separated section Splendentes Dandy ex A. Vázquez, which is composed entirely of Caribbean species from sect. Theorhodon Spach, creating a new section of genus Magnolia. Section Splendentes differs from members of section Theorhodon in stamen morphology in that the connective apex is extended into a long setiform appendage (Howard, 1948
; Vázquez-Garcia, 1994
). Interestingly, the same Splendentes-type stamen character is found in Dugandiodendron (Lozano-Contreras, 1975, 1983
) and section Aromadendron (Bl.) Noot. (Nooteboom, 1987
).
In this paper, we follow the taxonomic classification of Magnoliaceae by Frodin and Govaerts (1996)
for generic circumscription and scientific names. We use the subgeneric and sectional treatments of Nooteboom (1985, 1987)
except for the following genera and section that we use for our convenience in discussing taxa in this paper: Dugandiodendron, Manglietiastrum Y. W. Law., and section Splendentes in Magnolia. This and other voucher information is archived at the Botanical Society of America website (http://ajbsupp.botany.org/).
Disjuncts
There are three temperate disjunct taxonomic groups in Magnoliaceae, i.e., Liriodendron [L. tulipifera L. in North America and L. chinense (Hemsl.) Sarg. in China], Magnolia subgenus Magnolia section Rytidospermum Spach [M. dealbata Zucc., M. fraseri Walt., M. macrophylla Michx., and M. tripetala (L.) L. in North America to Mexico and the West Indies, and M. obovata Thunb., M. officinalis Rehder et E. H. Wilson, and M. rostrata W. W. Sm. in eastern Asia], and Magnolia subgenus Yulania (Spach) Reichenbach section Tulipastrum (Spach) Dandy [M. acuminata (L.) L. in North America and M. liliiflora Desr. in China]. Tropical disjunction of Magnoliaceae is found in the genus Magnolia subgenus Talauma in which there are four sections (Nooteboom, 1985
), i.e., sections Blumiana Blume (7 spp.), Aromadendron (5 spp.), and Manglietiastrum Noot. (1 sp., which is treated as a genus in this study) in eastern to southeastern Asia, and section Talauma Baill. (31 spp.) in the tropical to subtropical area in the New World (numbers of species are cited from Nooteboom, 2000)
. Southeastern Asian section Blumiana and American section Talauma are morphologically very similar to one another (Nooteboom, 1985, 1987
).
Plant materials, DNA extraction, amplification, and sequencing
Leaf materials of a total of 60 accessions from all genera (except Pachylarnax), including Manglietiastrum and Dugandiodendron, in Magnoliaceae and all sections (except Aromadendron) in genus Magnolia were used in this study (http://ajbsupp.botany.org/). Approximately one-half of the samples were previously utilized in an earlier phylogenetic analysis of temperate Magnolia species (Azuma, Thien, and Kawano, 1999
). Extracted DNA of Manglietiastrum sinica Y. W. Law. and Kmeria septentrionalis Dandy were provided from W.-B. Sun and Y.-L. Qiu (University of Zurich). Recent molecular phylogenetic analyses indicates that Degeneriaceae and Himantandraceae are more closely related to Magnoliaceae than to other members of the Magnoliales, i.e., Annonaceae, Eupomatiaceae, and Myristicaceae (Qiu et al., 1999
; Soltis et al., 2000
). In this study, Degeneriaceae (Degeneria roseiflora J. M. Miller and D. vitiensis Bailey et Smith) and Himantandraceae [Galbulimima belgraveana (F. Muell.) Sprague] were used as outgroups for Magnoliaceae.
The coding region of matK located between the two exons of the trnK gene and two intergenic spacer regions of chloroplast DNA (psbA-trnH and atpB-rbcL) were sequenced as described by Azuma, Thien, and Kawano (1999)
. It was not possible, however, to sequence the degraded DNA of material obtained from some herbarium specimens. In these taxa, only the matK region was sequenced using five newly designed overlapping primers (Table 1). Two intergenic spacer regions of Degeneriaceae and Himantandraceae also could not be sequenced because of degraded DNA or primer mismatching. Details of DNA amplification and sequencing strategy of fresh leaves were described by Azuma, Thien, and Kawano (1999)
. The amplification and sequencing of matK region from degraded DNAs is described below.
|
. Double-stranded DNA was amplified via 43 cycles in the polymerase chain reaction (PCR), with primer sets corresponding to each region as follows: HMF1-HMR1, HMF2-HMR2, HMF3-HMR3, HMF4-HMR4, and HMF5-HMR5 (Table 1). The PCR mixture (25 µL) contained 1 µL of template DNA, 200 µmol/L each deoxynucleotide, 1 µmol/L each primer, 2.5 mmol/L MgCl2, GeneAmp PCR Gold Buffer (Perkin-Elmer Applied Biosystems, Foster City, California, USA), and 0.6 units of AmpliTaq Gold (Perkin-Elmer Applied Biosystems). The initial cycle of the PCR reaction consisted of 0.5 min at 94°C for denaturation, 1 min at 50°C for primer annealing, and 1.5 min at 72°C for primer extension, with a pre-PCR heat step (9 min) to activate the AmpliTaq Gold in the first cycle. The last PCR cycle extended for 10 min. The PCR was performed by the GeneAmp PCR System 9700 (Perkin-Elmer Applied Biosystems). The PCR products were separated on 1% agarose gels and purified using the QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany). If a sharp band of amplified DNA did not appear on the gel stained with ethidium bromide (EtBr), additional PCR was conducted using 1 µL of the post-PCR mixture under the same condition. For sequencing, purified DNAs were reacted with BigDye Terminator Cycle Sequencing Ready Reaction (Perkin-Elmer Applied Biosystems). Sequencing was performed by an ABI PRISM 377 DNA Sequencer (Perkin-Elmer Applied Biosystems).
Phylogenetic analysis
Sequence data were manually aligned. Indels (insertion and/or deletion) were scored as binary characters (0 or 1) except length mutations of poly A, T, or G tracks. These polynucleotide tracks were ignored in all analyses. In addition, three short sequence inversions of 6 base pairs (bp), 3 bp, and 2 bp within the matK coding region and one inversion of 6 bp within the psbA-trnH spacer region were excluded from the data matrices (see Azuma, Thien, and Kawano, 1999
).
Maximum parsimony (MP) analysis was conducted with PAUP*, 4.0 beta version (Swofford, 2000)
. In addition, the neighbor-joining (NJ) and the maximum likelihood (ML) analyses were also performed for a combined data set with the same program. For MP analysis, a heuristic search was employed, using equal weighting of the TBR (tree bisection reconnection) branch-swapping option with MULPARS. One thousand bootstrap replications (Felsenstein, 1985
) and decay analyses (Bremer, 1988
) were conducted to measure confidence levels of each branch. We used AutoDecay, version 4.0 (Eriksson, 1998
) for the decay analysis. For NJ analysis, distance matrices were calculated using the Kimura two-parameter (Kimura, 1980
) with the transition/transversion ratio set at 2.0. In the ML analysis, ten replications of random-addition sequence searches with TBR branch-swapping were carried out. The option to collapse branches at zero length was activated. One hundred bootstrap replications were conducted under the same condition. In the analysis of a matK data set, Degeneria species were used as outgroups, whereas in the analyses of other data sets, Liriodendron species were used as outgroups.
Estimation of divergence times
The molecular clock or rate-constancy hypothesis is controversial (Kimura, 1983
), and in plants, nucleotide substitution rates are considered to be influenced by life history characters (including generation time; Bousquet et al., 1992
). To estimate divergence time in Magnoliaceae, a substitution rate for matK gene sequences was calculated. The rate-constancy of nucleotide substitutions of the matK gene in Magnolia was assessed using the relative rate test of Wu and Li (1985)
. The number of synonymous and nonsynonymous substitutions were calculated using method 2 of Ina (1995)
and computer programs (Ina, 1995
). In this test, Liriodendron tulipifera was used as a reference taxon.
We surveyed the literature for fossil seed records of Magnolioideae to correlate divergence times of some lineages within the Magnolioideae with molecular data. Although fossil records of Magnolia-like leaves are abundant throughout the Northern Hemisphere in the Tertiary (e.g., Tralau, 1963
; Dorofeev et al., 1974
; Hably, 1985
; Uemura, 1988
; Rember, 1991
; Mai, 1995
; Liu, Guo, and Ferguson, 1996
; Walther, 1999
), we only used seed characters because foliar characters of the Magnolioideae are shared with several other families (Peigler, 1989
; Collinson, Boulter, and Holmes, 1993
; Mai, 1995
). Fossil seeds of Magnolia are common throughout the Tertiary in Europe. However, fossils of modern Magnolia species become numerous in the late Miocene (Nooteboom, 1993
; Mai, 1995
).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; GenBank accession number GBAN-Z00044). There are 113 (7.7%) variable sites among 1460 total characters within Magnoliaceae, of which 73 (5.0%) are informative. The most parsimonious (MP) analysis of the matK data produced 174 equally most parsimonious trees with 218 steps, a consistency index (CI) (excluding uninformative characters) of 0.83, a retention index (RI) of 0.92, and a rescaled consistency index (RC) of 0.82. The 50% majority-rule consensus tree is shown in Fig. 1-I. The MP tree indicates tropical American section Talauma is monophyletic and ancestral within Magnolioideae (Fig. 1-I). The clade, however, is supported by relatively small bootstrap and decay values. There are several clades within the remaining clade, but phylogenetic affinities among them are not conclusive (Fig. 1-I).
|
2500 bp in which matK gene is located) and psbA-trnH and atpB-rbcL intergenic regions of 47 accessions to obtain more informative characters (http://ajbsupp.botany.org/). The data matrix of the trnK intron, including the matK region, consists of 2496 characters with four indels. Among the 2496 characters, 181 (7.3%) are variable within Magnoliaceae, of which 112 (4.5%) are informative. The MP analysis of trnK intron generated 99 equally most parsimonious trees with 210 steps, CI excluding uninformative characters = 0.85, RI = 0.92, and RC = 0.83. A strict consensus tree showed almost the same topology as generated by the matK data (tree not shown). The psbA-trnH data matrix consists of 455 characters including four indels. There are 62 (13.6%) variable characters within Magnoliaceae, of which 32 (7.0%) characters are informative. The MP analysis of the psbA-trnH data matrix generated four equally most parsimonious trees (70 steps, CI excluding uninformative characters = 0.87, RI = 0.96, and RC = 0.89, tree not shown). The data matrix of atpB-rbcL consists of 826 characters including five indels. There are 42 (5.1%) variable characters, of which 20 (2.4%) characters are informative. The MP analysis produced a single most parsimonious tree (44 steps, CI excluding uninformative characters = 0.95, RI = 0.97, and RC = 0.95, tree not shown). Phylogenetic resolutions for both trees were low. To obtain a better-resolved phylogenetic tree, the sequences from these different gene and noncoding regions were directly combined because they are all from the chloroplast genome and thus there is no recombination. The combined data matrix is composed of 3781 characters including 13 indels. There are 284 (7.5%) variable characters, of which 163 (4.3%) characters are informative. The MP analysis generated 26 equally most parsimonious trees (326 steps, CI excluding uninformative characters = 0.85, RI = 0.93, and RC = 0.84, tree not shown). The neighbor-joining (NJ) analysis generated a cladogram basically concordant with the MP tree (Fig. 1-II). The most likelihood analysis (ML) produced the tree with the best log-likelihood (–7610.11) obtained from examination of 19 577 different trees, largely concordant with both MP and NJ trees (Fig. 1-III). In the ML tree, tropical American section Talauma (clade A) branches first within Magnolioideae, only weakly supported by bootstrap analysis. The next clade (B) comprises three subclades, the tropical Asian group [clade C with M. liliifera (L.) Baill. (section Blumiana), M. coco (Lour.) DC., and M. delavayi Franch. (section Gwillimia DC.)], the West Indies section Splendentes (clade D), and the temperate-subtropical clade E, including two temperate disjuncts (clades K and M). Phylogenetic relationships among clades C, D, and E are not clearly resolved (Fig. 1-III).
Estimation of divergence times
We conducted the relative rate test (Wu and Li, 1985
) to check the rate-constancy of nucleotide substitution of matK region among 57 accessions of Magnoliaceae. Absolute numbers of synonymous sites in matK region were very low (2.95 sites on average within Magnolioideae), and there were no large differences between rates of synonymous and nonsynonymous substitutions (0.0085 vs. 0.0067) within Magnolioideae. Therefore, we used the Kimura two-parameter distance across the whole matK region for the test and detected no accelerated or decelerated lineages.
Fossil seed records (Tertiary) of subfamily Magnolioideae are abundant throughout the Northern Hemisphere (Fig. 2). The morphological characteristics of these fossils as well as those of extant taxa are well documented (e.g., Mai, 1971, 1975
; Dorofeev et al., 1974
; Tiffney, 1977
). Most extant taxa of Magnolioideae produce ovate to asymmetrical seeds with smooth (most Magnolia taxa) or rough coats (Manglietia and Michelia and most taxa of section Talauma; Tiffney, 1977
; nos. 36–38, 41, 42 in Fig. 2). Other taxa, i.e., sections Theorhodon, Magnolia, Tulipastrum, Buergeria (Siebold et Zucc.) Baillon, and Yulania (Spach) Dandy, produce seeds with flat, symmetrical, and cordiform to bean-like shapes (nos. 33–35, 39, 40), which are easily distinguishable from the ovate and asymmetrical seeds (Tiffney, 1977
; Fig. 2). Based on these characters, we classified the fossil seeds into two categories in North America and Eurasia (Fig. 2). Among extant taxa producing flat and symmetrical seeds, those of sects. Theorhodon (e.g., M. grandiflora L., no. 39) and Magnolia taxa are longer than wide, while those of section Buergeria [e.g., M. kobus DC. and M. stellata (Siebold et Zucc.) Maxim., nos. 33, 34] are usually wider than long, and those of sections Tulipastrum (e.g., M. acuminata, no. 40) and Yulania (e.g., M. denudata Desr., no. 35) are slightly longer than wide or vice versa (Fig. 2; Tiffney, 1977
).
|
). The seed fossils found in Eurasia (nos. 9b, 16–26) bear a greater resemblance to those of subgenus Yulania (sects. Buergeria, Tulipastrum, and Yulania) than those in sects. Theorhodon and Magnolia. In addition, Tiffney (1977)
described a fossil seed taxon, Magnolia waltonii (no. 32), from the Brandon Lignite of North America (Fig. 2), considered to be early Miocene (Tiffney, 1994
). Magnolia waltonii is not directly comparable to the seeds of any living taxa of Magnolia, but in some respects there are similarities between M. waltonii and sections Theorhodon and Tulipastrum (Tiffney, 1977
). Figlar (1993)
examined a leaf fossil (the slope Magnolia) from the Miocene Clarkia flora in North America (Idaho) that is comparable to the leaves of extant Magnolia acuminata (section Tulipastrum).
Based on these fossil records, we consider the ancestral lineage of subgenus Yulania (especially sect. Tulipastrum) to have diverged in Eurasia and North America during the middle Oligocene to early Miocene (30–20 mya). Molecular phylogenetic analysis indicates subgenus Yulania is monophyletic (clade K in Fig. 1-III) and that M. acuminata (sect. Tulipastrum) is the basal lineage. Therefore, we assume the disjunction of subgenus Yulania taxa (i.e., M. acuminata vs. Asian relatives) occurred at 25 mya (late Oligocene; Fig. 1-III).
Average nucleotide substitution rate of matK gene in subgenus Yulania is estimated to be 9.95 x 10–11 ± 1.85 x 10–11 (M. acuminata vs. other taxa). Based on this rate, the divergence time between tropical American section Talauma (clade A) and other taxa (clade B) is estimated to be 42.0 ± 6.6 mya on average (Fig. 1-III). Divergence time between the other tropical groups (tropical Asian clade C and the West Indies clade D) and temperate-subtropical clade E is estimated to be 35.6 ± 5.6 mya (Fig. 1-III). The divergence time between another temperate disjunct (M. tripetala vs. other taxa in clade M) is estimated to be 27.9 ± 4.4 mya (Fig. 1-III).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Although section Blumiana is placed in the subgenus Talauma and separated from the two other sections in subgenus Magnolia (sections Gwillimia and Lirianthe), it is morphologically indistinguishable from the other taxa except in the dehiscence of fruits (Nooteboom, 1985
). Section Gwillimia and monotypic section Lirianthe are also morphologically closely related and can also be distinguished only by fruiting characters (short beaks vs. long beaks; Nooteboom, 1985
). Thus, the close affinity among these Asian sections is supported by morphological characters as well as by molecular phylogenetic analysis.
Section Aromadendron, not included in this study, is taxonomically placed in the subgenus Talauma and is morphologically closely related to sect. Blumiana (Nooteboom, 1985, 1987
). A recent molecular study, however, indicates that sect. Aromadendron does not show a close affinity to other sections of subg. Talauma. Instead, it is the sister group to sect. Alcimandra (Dandy) Noot. [M. cathcartii (Hook. f. et Thomson) Noot.] of subg. Magnolia, hence this Aromadendron/Alcimandra clade is sister to the genus Michelia-section Maingola clade (Kim et al., 2001)
.
Although Dugandiodendron and section Splendentes are separately recognized and assigned to different sections or subgenera, the close affinity between them is also suggested by their stamen morphology, with the connective apex extended into a long setiform appendage (see Howard, 1948
; Lozano-Contreras, 1975, 1983
; Vázquez-Garcia, 1994
). The tip embeds in the gynoecium, and then the stamens abscise at the base in section Splendentes (Howard, 1948
; Vázquez-Garcia, 1994
). The long setiform appendage, however, is not unique, as most Aromadendron taxa have a similar setiform appendage atop the stamens (Nooteboom, 1987
).
Morphological studies support a sister relationship between tropical American section Talauma and tropical Asian section Blumiana (Nooteboom, 1985, 1987
). The molecular phylogenetic analysis, however, does not support this conclusion but rather indicates the tropical Asian clade (including Blumiana, clade C in Fig 1-III) might form a sister relationship with another tropical American lineage, the West Indies group (clade D, and perhaps with Dugandiodendron; see Fig. 1-I). Although a direct sister relationship between the tropical Asian group and the West Indies is not shown, we can conclude that a tropical disjunction of Magnoliaceae occurs between the tropical Asian and the West Indies taxa (and Dugandiodendron).
Circumscription of temperate disjunct groups
Present and previous molecular phylogenetic studies of Magnoliaceae indicate section Rytidospermum represents a para- or polyphyletic lineage (Qiu, Chase, and Parks, 1995
; Azuma, Thien, and Kawano, 1999
; Kim et al., 2001
; see Fig. 1). The disjunction is found in clade M (in Fig. 1-III), which is composed of Magnolia tripetala (North American taxon), Asian section Oyama Nakai taxa (clade O), and Asian Rytidospermum-Manglietia clade (clade P). The genus Manglietia (29 spp., all Asian taxa) is taxonomically separated from Magnolia by number of ovules (four or more) per carpel and leaf anatomical characters (Nooteboom, 1985, 1993
). However, other taxonomic features are comparable to those found in genus Magnolia, especially in the way the leaves are often arranged in false whorls on the ends of shoots. The feature is found only in section Rytidospermum and genus Manglietia. The close affinity between Asian Rytidospermum and Manglietia may be reliable, although the affinity is weakly supported by bootstrap analysis (Fig. 1).
The relationship between section Oyama clade and M. tripetala or Asian Rytidospermum-Manglietia clade is not supported by morphological characters. Morphological characters need to be reevaluated to support this relationship. Close affinity between M. tripetala and Asian members of section Rytidospermum (but not with other American members of that section) is found in seed and fruit morphologies (Qiu, Chase, and Parks, 1995
). However, this study does not support a direct sister relationship between the taxa (Fig. 1).
Section Tulipastrum (subgenus Yulania) with only two species (M. acuminata in North America and M. liliiflora in China) has been traditionally recognized as a disjunct. Molecular analysis, however, does not support this monophyly, but indicates M. acuminata is sister to all Asian members (including M. liliiflora) of subgenus Yulania (clade K in Fig. 1-III). The section Tulipastrum is taxonomically separated by the sepaloid tepals, the flowering after leaf emergence, and the color of tepals (Nooteboom, 1985
). However, these characters are not important because they appear in other members of subgenus Yulania [e.g., petaloid sepals in M. salicifolia (Siebold et Zucc.) Maxim., and flowering at the time of leaf emergence in M. kobus]. Although floral characters of M. acuminata are relatively distinct among subgenus Yulania taxa, i.e., relatively small flower and yellow-greenish and narrow tepals, Asian members of subgenus Yulania possess no commonly recognized morphological character.
Biogeographical implications
The origin of the family Magnoliaceae and divergence of the subfamilies Liriodendroideae and Magnolioideae extends back 100 mya (Late Cretaceous). For example, multifollicular fruits (Archaeanthus linnenbergeri) and associated leaves, tepals, and stipular bud scales are known from the Late Cretaceous (95 mya) in the Dakota Formation of central Kansas, USA (Dilcher and Crane, 1984
). The floral structure of the Archaeanthus is comparable to that of extant Magnoliaceae and is considered to be an ancestor or a lineage closely related to Magnoliaceae (Peigler, 1989
; Friis, Crane, and Pedersen, 1997
; Crane, 1998
).
Despite the long history of Magnoliaceae, diversification of extant taxa within the subfamily Magnolioideae seems to have occurred recently, as indicated by the long branch between Liriodendroideae (only Liriodendron) and Magnolioideae (Fig. 1). The phylogenetic tree of the combined sequence data indicate that divergence within Magnolioideae occurred at 42 mya (middle Eocene), in which the tropical American section Talauma branched off first, and then both tropical Asian and the West Indies groups diverged (Fig. 1-II and III).
According to paleobotanical and geological evidence, the early Eocene (50–54 mya) was the warmest period in the Tertiary, and the boreotropical flora (including an ancestral lineage of Magnolioideae) circumboreally spread at high latitudes in the Northern Hemisphere (Reid and Chandler, 1933
; Chandler, 1964
; Wolfe, 1978, 1997
; Collinson, Fowler, and Boulter, 1981
; Miller, Fairbanks, and Mountain, 1987
; Graham, 1999
; see Fig. 2). The paleobotanical evidence indicates that in the late early and early middle Eocene (50–48 mya) a significant cooling occurred, followed by two warm intervals (46–43 and 37–34 mya) separated by a cool interval (42–38 mya) in the middle to late Eocene (Wolfe, 1978, 1997
). Oxygen isotope records also suggest very warm temperatures in the early Eocene, but cooling temperatures during the middle to late Eocene proceeded with small fluctuations (Miller, Fairbanks, and Mountain, 1987
). The cooling events might have caused movement of the boreotropical floral elements to the lower latitudes, leading to disjunction of ancestral lineages of modern tropical plants between North America and Eurasia.
The molecular phylogenetic analysis indicates tropical American section Talauma (clade A in Fig. 1-III) diverged at 42 mya (middle Eocene). In addition, the tropical Asian group (clade C) may be the most ancestral lineage within the remaining clade B because the sequence distance between tropical Asian clade C and clade E (0.0076) is greater than between the West Indies clade D and clade E (0.0063; see Fig. 1-II). Therefore, the cool interval (42–38 mya) at the middle to late Eocene seems to have led to the diversification or disjunction between North America (tropical American section Talauma) and Eurasia (or clade B). In that period, southeastern Alaska's vegetation supported a temperate (mesothermal) flora (Wolfe, 1972
), suggesting difficulty of interchanges for tropical elements between the continents. However, temperature fluctuation during the middle to late Eocene (46–34 mya) may have enabled northern parts of the divided tropical floras to interchange between the continents via the North Atlantic and/or the Bering land bridge during the warm intervals. Indeed, fossil records of land mammals indicate that intercontinental migrations occurred twice during the Eocene (at 43 and 36 mya; Woodburne and Swisher, 1995
), which are concordant in time with the two warm intervals inferred from the paleobotanical records (Wolfe, 1978
).
The molecular phylogenetic analysis indicates that divergence and disjunction of tropical Asian and the West Indies groups (clades C and D) occurred at 36 mya (late Eocene). The warm interval (37–34 mya) in the late Eocene therefore might have prompted intercontinental migration of the divided lineage of Magnolioideae (clade B), and subsequent cooling by 33 mya may be responsible for the disjunction between clades C and D.
Paleobotanical evidence suggests major climate deterioration occurred at 33 mya (the Terminal Eocene Events) leading to a high rate of extinction in lineages and elimination of tropical elements from most areas of North America and Europe (Wolfe, 1978, 1997
). The molecular phylogenetic analysis could not resolve phylogenetic affinities among most temperate-subtropical clades (within clade E in Fig. 1-III). Diversification of temperate-subtropical clades (clade E) seems to have occurred near the time of major climate deterioration at 33 mya (Fig. 1-III), and perhaps the extinction of tropical taxa obscured relationships among temperate lineages and/or rapid speciation of temperate taxa took place. After the 33-mya climatic deterioration, temperature began to fluctuate and five warm-cool intervals took place between early Oligocene and middle Miocene (34–15 mya). The temperature, however, did not reach the level of the early Eocene (Miller, Fairbanks, and Mountain, 1987
; Wolfe, 1997
). The fluctuation of temperature may have allowed temperate elements to migrate between continents during warm intervals. Indeed, estimated times of disjunctions of temperate lineages of Magnolioideae (at 28 and 25 mya) correlate with this scenario.
Comparison with previous studies of temperate disjunction in Magnoliaceae
Parks and Wendel (1990)
show the divergence time of Liriodendron tulipifera and L. chinense to be 10–16 mya (middle to late Miocene) based on molecular analysis (allozyme and RFLP [restriction fragment length polymorphism] analysis of cpDNA) and paleobotanical evidence. However, if a substitution rate calculated in this study is employed to estimate the time of disjunction (relative rate tests for these taxa were not conducted because they were used as reference taxa in the test), L. tulipifera and L. chinense diverged at 27.9 ± 4.4 mya. Similarly, Qiu, Parks, and Chase (1995)
suggested the divergence time of M. tripetala and M. hypoleuca Siebold et Zucc. (M. obovata in this study; both deciduous taxa in section Rytidospermum) to be 4.1–5.5 or 1.9 mya using allozyme data and 1.7 ± 0.8 mya using RFLP data of cpDNA. This suggests the disjunction occurred in the late Miocene to early Pliocene (6–5 mya). The divergence time between M. tripetala and Asian relatives in our study (clade M in Fig. 1-III) was estimated to be 27.9 ± 4.4 mya. However, if only the two taxa (M. tripetala and M. obovata) are compared as in Qiu, Parks, and Chase (1995)
, the divergence time is estimated to be 20.9 ± 3.3 mya.
The estimation of divergence times conducted in this study strictly relies on fossil records adopted as an independent estimate. Therefore, the accuracy of assignment and incidental occurrence of the fossil records as well as variation of molecular evolutionary rate and pattern of extinction in a clade will affect the estimation and may be responsible for different estimations between previous and present studies.
The molecular phylogenetic tree constructed from cpDNA sequence data shows mostly maternal phylogeny and would miss tracing the tangled phylogeny that would occur in a clade composed of polyploid taxa. Polyploidy is known in several lineages of Magnoliaceae (Chen et al., 2000)
, especially in subgenus Yulania and section Theorhodon in genus Magnolia (Fig. 1-I). If reticulate evolution took place in the polyploid clade, the cpDNA phylogeny may obscure events that occurred in lineages and would then affect the estimated divergence time.
Recently, Kim et al. (2001)
conducted a molecular phylogenetic analysis of Magnoliaceae covering all genera and sections (total 99 taxa of 223 species) using ndhF sequence data. Generally, the topology of the ndhF tree is concordant with that of the tree constructed in this study. However, the ndhF tree indicated the clade of M. macrophylla and M. dealbata (North American taxa of section Rytidospermum) was basal in the subfamily Magnolioideae, but supported by a very low bootstrap value (39%). Qiu, Parks, and Chase (1995)
also showed the same relationship. On the other hand, our study shows that the tropical American group (section Talauma) is at the base, and then the tropical Asian group and the West Indies group diverged in the subfamily but with relatively low bootstrap values (Fig. 1). To clarify this discordance and phylogenetic relationship in unresolved clades, a sequence matrix combined with multiple genes and noncoding regions should be analyzed.
Other examples of tropical disjunction of Laurasian taxa
The Illiciaceae, with 42 species, are distributed in southeastern Asia, southeastern United States plus eastern Mexico, and the West Indies (Smith, 1947
). Molecular phylogenetic analysis of ITS sequence data of 15 species (not including Caribbean species) indicated the North American species (2 spp.) and the remaining east Asian species separated first, suggesting the age of disjunction is relatively old (Hao, Saunders, and Chye, 2000)
. Fossil records of seeds and fruits of Illiciaceae are found in Europe and North America from the Paleocene to the Miocene (Mai, 1970
; Tiffney and Barghoorn, 1979
; Friis, Crane, and Pedersen, 1997
). Therefore, the disjunction of the Illiciaceae suggests it also represents a relict of the boreotropical flora like Magnoliaceae.
Both Illiciaceae and Magnoliaceae are ancient Laurasian taxa extending to the early Cretaceous. However, members of Illiciaceae are not present in South America unlike taxa of Magnoliaceae (Dugandiodendron and Talauma). For most of the Tertiary and much of the Cretaceous, South America was essentially an island continent (Gentry, 1982
). Magnoliaceae is one of the few Laurasian elements that invaded South America via the Panamanian Isthmus after Pliocene or via the proto-Antillean chain (GAARlandia) during the Tertiary (Gentry, 1982
; Iturralde-Vinent and MacPhee, 1999
), which probably resulted in rapid speciation in American Talauma and Dugandiodendron.
In regard to the southward distribution of Magnoliaceae (New Guinea and Brazil), we could not eliminate a possibility of migration via the Southern Hemisphere (Cox, 1990
). However, the present and past (fossil) distribution patterns of Magnoliaceae do not support this hypothesis.
Conclusion
The boreotropical flora concept suggests that tropical disjuncts have occurred during the middle to late Eocene prior to the disjunctions of temperate taxa (Tiffney, 1985a
; Wolfe, 1997
). Molecular phylogenetic analysis and fossil evidence of Magnoliaceae clearly show the tropical disjunction occurred during the middle to late Eocene and that the temperate disjunction occurred in the Oligocene, which is concordant with the boreotropical flora concept. There are only two examples of molecular analyses on the intrageneric level of tropical disjunction (Magnoliaceae and Illiciaceae) and both indicate the tropical disjunctions are ancient.
|
FOOTNOTES |
|---|
5 Author for reprint requests (j52116{at}sakura.kudpc.kyoto-u.ac.jp
).
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Azuma H. L. B. Thien S. Kawano 1999 Molecular phylogeny of Magnolia (Magnoliaceae) inferred from cpDNA sequences and evolutionary divergence of the floral scents. Journal of Plant Research 112: 291-306[CrossRef][ISI]
Boufford D. E. S. A. Spongberg 1983 Eastern Asian–eastern North American phytogeographical relationships: a history from the time of Linnaeus to the twentieth century. Annals of the Missouri Botanical Garden 70: 423-439
Bousquet J. S. H. Strauss A. H. Doerksen R. A. Price 1992 Extensive variation in evolutionary rate of rbcL gene sequences among seed plants. Proceedings of the National Academy of Sciences, USA 89: 7844-7848
Bremer K. 1988 The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795-803[CrossRef][ISI]
Chandler M. E. J. 1964 The lower Tertiary floras of southern England. IV. A summary and survey of findings in the light of recent botanical observations. British Museum (Natural History), London, UK
Chaney R. W. 1947 Tertiary centers and migration routes. Ecological Monographs 17: 139-148
Chen B. L. H. P. Nooteboom 1993 Notes on Magnoliaceae III: the Magnoliaceae of China. Annals of the Missouri Botanical Garden 80: 999-1104[CrossRef][ISI]
Chen Z. X. Huang R. Wang S. Chen 2000 Chromosome data of Magnoliaceae. In Y. Liu et al. [eds.], Proceedings of the International Symposium on the Family Magnoliaceae, 192–201. Science Press, Beijing, China
Collinson M. E. M. C. Boulter P. L. Holmes 1993 Magnoliophyta (Angiospermae). In M. J. Benton [ed.], The fossil record 2, 809–840. Chapman and Hall, London, UK
Collinson M. E. K. Fowler M. C. Boulter 1981 Floristic changes indicate a cooling climate in the Eocene of southern England. Nature 291: 315-317
Cox C. B. 1990 New geological theories and old biogeographycal problems. Journal of Biogeography 17: 117-130[CrossRef][ISI]
Crane P. R. 1998 The phylogenetic position and fossil history of the Magnoliaceae. In D. Hunt [ed.], Magnolias and their allies, 21–36. David Hunt, Milborne Port, UK
Dandy J. E. 1927 The genera of Magnolieae. Bulletin of Miscellaneous Information (Royal Botanic Gardens, Kew) 7: 257-264
Dandy J. E. 1950 A survey of the genus Magnolia together with Manglietia and Michelia. In P. M. Synge [ed.], Camellias and Magnolias: report of the conference held by the Royal Horticultural Society, 64–81. Royal Horticultural Society, London, UK
Dandy J. E. 1978 A revised survey of the genus Magnolia together with Manglietia and Michelia. In N. G. Treseder [ed.], Magnolias, 29–37. Faber and Faber, London, UK
Dilcher D. L. P. R. Crane 1984 Archaeanthus: an early angiosperm from the Cenomanian of the western interior of North America. Annals of the Missouri Botanical Garden 71: 351-383[CrossRef][ISI]
Dorofeev P. I. 1959 Ob oligotsenovoi flore s. kozyulino v ustye r. tomi. Doklady Akademii nauk SSSR 127: 1103-1105 (in Russian)
Dorofeev P. I. 1960a Ob oligotsenovoi flore dunaevskogo yara na r. tym v zapadnoi sibiri. Doklady Akademii nauk SSSR 132: 659-661 (in Russian)
Dorofeev P. I. 1960b Novye dannye o tretichnykh florakh kireevskogo yara na obi. Doklady Akademii nauk SSSR 133: 211-213 (in Russian)
Dorofeev P. I. S. Iljinskaja N. Imchanitzkaja T. Kolesnikova E. Kutuzkina I. Shilkina N. Snigirevskaya I. Sveshnikova S. Zhilin 1974 Magnoliaceae-Eucommiaceae. In A. Takhtajan [ed.], Fossil flowering plants of the USSR, vol. 1. Nauka, Leningrad, Russia (in Russian)
Doyle J. J. J. L. Doyle 1987 A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemistry Bulletin 19: 11-15
Engler A. 1879 Versuch einer Entwicklungsgeschichte der Pflanzenwelt, inbesondere der Florengebiete seit der Tertiärperiode. 1. Die extratropischen Gebiete der Nördlichen Hemisphäre. Verlag von Wilhelm Engelmann, Leipzig, Germany
Eriksson T. 1998 AutoDecay, version 4.0 (program distributed by the author). Bergius Foundation, Royal Swedish Academy of Sciences, Stockholm, Sweden
Felsenstein J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Figlar R. B. 1993 Stone Magnolias. Arnoldia 53: 2-9
Friis E. M. P. R. Crane K. R. Pedersen 1997 Fossil history of magnoliid angiosperms. In K. Iwatsuki and P. H. Raven [eds.], Evolution and diversification of land plants, 121–156. Springer-Verlag Tokyo, Tokyo, Japan
Frodin D. G. R. Govaerts 1996 World checklist and bibliography of Magnoliaceae. Royal Botanic Gardens, Kew, UK
Gentry A. H. 1982 Neotropical floristic diversity: phytogeographical connections between Central and South America, Pleistocene climatic fluctuations, or an accident of the Andean orogeny?. Annals of the Missouri Botanical Garden 69: 557-593[CrossRef][ISI]
Graham A. 1972 Outline of the origin and historical recognition of floristic affinities between Asia and eastern North America. In A. Graham [ed.], Floristics and paleofloristics of Asia and eastern North America, 1–18. Elsevier, Amsterdam, The Netherlands
Graham A. 1999 Late Cretaceous and Cenozoic history of North American vegetation, north of Mexico. Oxford University Press, New York, New York, USA
Gray A. 1846 Analogy between the Flora of Japan and that of the United States. American Journal of Science and Arts Second Series 2: 135-136
Hably L. 1985 Catalogue of the Hungarian Cenozoic leaf-flora. Studia Botanica Hungarica 18: 5-58
Hao G. R. M. K. Saunders M.-L. Chye 2000 A phylogenetic analysis of the Illiciaceae based on sequences of internal transcribed spacers (ITS) of nuclear ribosomal DNA. Plant Systematics and Evolution 223: 81-90
Heywood V. H. 1978 Flowering plants of the world. Oxford University Press, London, UK
Howard R. A. 1948 The morphology and systematics of the West Indian Magnoliaceae. Bulletin of the Torrey Botanical Club 75: 335-357[CrossRef]
Ina Y. 1995 New methods for estimating the numbers of synonymous and nonsynonymous substitutions. Journal of Molecular Evolution 40: 190-226[CrossRef][ISI][Medline]
Iturralde-Vinent M. A. R. D. E. MacPhee 1999 Paleogeography of the Caribbean region: implications for Cenozoic biogeography. Bulletin of the American Museum of Natural History Number 238. American Museum of Natural History, New York, New York, USA
Kim S. C.-W. Park Y.-D. Kim Y. Suh 2001 Phylogenetic relationships in family Magnoliaceae inferred from ndhF sequences. American Journal of Botany 88: 717-728
Kimura M. 1980 A simple method for estimating evolutionary rates of base substitutions through comparable studies of nucleotide sequences. Journal of Molecular Evolution 16: 111-120[CrossRef][ISI][Medline]
Kimura M. 1983 The natural theory of molecular evolution. Cambridge University Press, London, UK
Knobloch E. D. H. Mai 1986 Monographie der Früchte und Samen in der Kreide von Mitteleuropa. Rozpravy ústredního ústavu geologického 47: 1-219
Lavin M. 1995 Tribe Robinieae and allies; model groups for assessing early Tertiary northern latitude diversification of tropical legumes. In M. Crisp and J. J. Doyle [eds.], Advances in legume systematics 7. Phylogeny, 141–160. Royal Botanical Gardens, Kew, UK
Lavin M. M. Luckow 1993 Origins and relationships of tropical North America in the context of boreotropical hypothesis. American Journal of Botany 80: 1-14
Law Y.-W. ( Liu Y.-H.). 1984 A preliminary study on the taxonomy of the family Magnoliaceae. Acta Phytotaxonomica Sinica 22: 89-109 (in Chinese)
Lee N. S. T. Sang D. J. Crawford S. H. Yeau S.-C. Kim 1996 Molecular divergence between disjunct taxa in eastern Asia and eastern North America. American Journal of Botany 83: 1373-1378
