| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
2Departamento de Botânica, Instituto de Biologia, Universidade Estadual de Campinas, Caixa Postal 6109, Cep 13083-970, Campinas, São Paulo, Brazil; 3Institute of Systematic Botany, University of Zürich, Zollikerstrasse 107, 8008 Zürich, Switzerland
Received for publication November 28, 2003. Accepted for publication April 29, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Apocynaceae • Apocynoideae • Mesechiteae • morphology • phylogenetic systematics • rpl16 • rps16 • trnS-G
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
) and includes nine genera and about 150 species. It is restricted to the Neotropics, where it has a broad distribution, ranging from the southwestern United States throughout Mesoamerica and the Caribbean to southern South America in rainforests, montane forest, savanna, and desert habitats. The tribe is extremely variable in habit and includes vines, erect subshrubs, and small trees. Floral structure is also remarkably diverse, especially the corolla, which ranges from small, inconspicuous, whitish, and tubular to large, variously colored, and infundibuliform.
Although several recent phylogenetic studies have addressed the circumscription of the Apocynaceae and their relationships with the Asclepiadaceae (Judd et al., 1994
; Sennblad and Bremer, 1996
, 2002
; Potgieter and Albert, 2001
), resulting in the amalgamation of these two families into Apocynaceae sensu lato (Endress and Bruyns, 2000
), many other aspects of classification within the family remain controversial. One main controversy has been the delimitation and composition of tribes (Table 1). This is exemplified by Mesechiteae, for which relationships with other tribes are unknown and the generic circumscription is confusing and shows little consistency (Zarucchi, 1991
; Henrickson, 1996
; Williams, 1999
).
|
recognized four tribes in Apocynoideae: Parsonsieae, Nerieae, Ecdysanthereae, and Ichnocarpeae. His tribal delimitations were based mainly on the form of the "retinacle," a term he coined for the specialized region of the anther that unites it with the style head (Pichon, 1948a
). Except for Galactophora and Secondatia, most of the genera included in the ingroup in our study were placed in his tribe Ichnocarpeae, which was defined by the presence of a glabrous, concave retinacle. Within Ichnocarpeae, five subtribes were recognized, only two of which are pertinent to our study: Forsteroniinae and Mandevillinae.
A new classification for the Apocynaceae sensu stricto was published by Leeuwenberg (1994)
. Although his work was clearly influenced by Pichon's ideas, only three tribes were recognized in Apocynoideae: Echiteae, Wrightieae, and Apocyneae. This was the first classification to take into account the priority rule, necessitating changing the name of Pichon's (1950)
tribe Parsonsieae to Echiteae and Nerieae to Wrightieae. Leeuwenberg's third tribe, Apocyneae, was more or less equivalent to Pichon's tribes Ecdysanthereae and Ichnocarpeae combined. Leeuwenberg's tribes, however, are difficult to distinguish, and Sennblad et al. (1998)
showed that they are artificial. Leeuwenberg (1994)
arranged the genera he included in his Echiteae into two subtribes, Echitinae and Parsonsiinae, without any explanation of the criteria he used to delimit them. In his classification, most of the genera that comprise the ingroup in this study were placed in Echitinae. The only exception is Tintinnabularia, which was placed near Beaumontia in Wrightieae. Although Leeuwenberg (1994)
gave no reasons for this placement of Tintinnabularia, presumably he was influenced by Woodson's (1936)
comment that Tintinnabularia and Beaumontia share three character states: (1) domatia on the lower surface of the leaves, (2) long anther filaments, and (3) foliaceous sepals. The first two are relatively rare in the family.
The most recent classification of the family is that by Endress and Bruyns (2000)
. Theirs was a unified classification taking into account both Apocynaceae and Asclepiadaceae, and the first one to incorporate phylogenetic considerations. Two of the five subfamilies they recognized, Rauvolfioideae and Apocynoideae, correspond to the traditional Apocynaceae sensu stricto (s.s.), the other three to the Asclepiadaceae as traditionally circumscribed. Their classification differed most dramatically from those of previous authors with regard to the circumscription of tribes in Rauvolfioideae and Apocynoideae. Although the subfamilies and tribes proposed by Endress and Bruyns (2000)
were based mainly on morphological characters, this was the first classification to incorporate evidence from molecular studies as well (e.g., Endress et al., 1996
; Sennblad and Bremer, 1996
; Sennblad et al., 1998
; K. Potgieter and V. Albert, University of Illinois, personal communication).
Endress and Bruyns (2000)
recognized five tribes in Apocynoideae: Wrightieae, Malouetieae, Echiteae, Mesechiteae, and Apocyneae. Two of the most important morphological characters they used to define these tribes are the structure of the style head and the retinacle, both of which are also important characters in Pichon's (1948a
, 1950)
classification, but were not considered reliable by A. Leeuwenberg (Wageningen Agricultural University, personal communication). Endress and Bruyns (2000)
significantly changed tribal circumscription and composition from those in the classification systems of Pichon (1950)
and Leeuwenberg (1994)
. Most taxa included in a more narrowly circumscribed Mesechiteae by Endress and Bruyns (2000)
were part of the heterogeneous Ichnocarpeae of Pichon (1950)
and the Echiteae of Leeuwenberg (1994)
.
Another controversy concerns whether or not smaller satellite genera should be recognized as distinct from their closely related, larger genus. Within Mesechiteae, one such case is whether Allomarkgrafia should be included in the synonymy of Mesechites (as proposed by Pichon, 1950
and followed by Leeuwenberg, 1994
) or recognized as a distinct genus (as in the classification of Endress and Bruyns, 2000
). Another problematic taxonomic question concerns the large genus Mandevilla and its smaller satellite genera Macrosiphonia, Telosiphonia, and, more recently, Quiotania. In this group of taxa, one to four genera have been recognized by specialists in the family (Zarucchi, 1991
; Leeuwenberg, 1994
; Henrickson, 1996
; Williams, 1999
).
The classification of Apocynaceae sensu lato (s.l.) proposed by Endress and Bruyns (2000)
represents a considerable advance in the systematics of the family and is a logical starting point for studies that test the phylogenetic relationships of the groupings proposed therein. This type of analysis was done by Potgieter and Albert (2001)
, based on the trnL-F intergenic spacer and morphology in a broad phylogenetic study of the family. They found well-defined clades in Rauvolfioideae, several of which are reflected in the current classification. For Apocynoideae, however, no clearly defined groups were retrieved using that plastid region.
The aims of the present article are to test the monophyly of Mesechiteae sensu Endress and Bruyns (2000)
and the relationships among its constituent genera using both morphology and molecular sequence data from the plastid rpl16 intron, rps16 intron, and trnS-G intergenic spacer. The resulting hypothesis of phylogenetic relationships within Mesechiteae is compared with the current classification, morphological features that characterize clades are discussed, and a modified circumscription of Mesechiteae is proposed.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
, were defined as the ingroup and included in this study (Allomarkgrafia, Galactophora, Macrosiphonia, Mandevilla, Mesechites, Secondatia, Telosiphonia, and Tintinnabularia). Mandevilla tenuifolia exhibits a high degree of polymorphism in its habit and vegetative parts, one morphotype (Mandevilla tenuifolia2) of which is very similar to M. myriophyllum. Therefore, in order to determine the relationship of these two species, we included two samples of the most extreme forms of M. tenuifolia and one from M. myriophyllum. Quiotania could not be included, as none of the type material could be located, suggesting that it was never distributed. Outgroup taxa were chosen from all but the basalmost tribe of the subfamily (Wrightieae), based largely on previous studies, which suggest that the closest relative of Mesechiteae is either Apocyneae or Echiteae (Sennblad et al., 1998
; Sennblad and Bremer, 2002
). Two genera from Echiteae (Prestonia and Rhodocalyx) and five genera from Apocyneae (Beaumontia, Chonemorpha, Forsteronia, Odontadenia, and Trachelospermum) were included. Three representatives of Malouetieae (two species of Pachypodium and one species of Mascarenhasia) were used to root the cladograms. Taxon names, voucher information, and GenBank accession numbers are given in Appendix 1 (see Supplemental Data accompanying the online version of this article).
DNA extraction, amplification, and sequencing
Total genomic DNA was extracted from silica-dried leaf material or from herbarium specimens using DNeasy Plant mini kits (Qiagen, Valencia, California, USA) following the manufacturer's protocol. Three noncoding plastid regions, the rpl16 intron, rps16 intron, and trnS-G intergenic spacer, were amplified for all taxa. Double-stranded DNA was amplified by polymerase chain reaction (PCR) on a Biometra Tgradient machine (Biometra, Göttingen, Germany), applying a thermal cycling program consisting of 34 cycles of denaturation at 95°C for 30 s, annealing at 52°C for 1 min, and extension at 72°C for 90 s. Reactions were terminated with a final extension of 4 min at 72°C. All PCR reactions were performed in a total volume of 25 µL, using 2.5 mmol/L MgCl, 10% PCR* Buffer (Amersham Biosciences, Otelfingen, Switzerland), 0.25 mmol/ L dNTP, 0.5 units Taq DNA polymerase (Amersham Biosciences), 1–4 µL bovine serum albumin (BSA, Sigma, Steinheim, Germany) and 0.1 mmol/L of each primer. Primer information is presented in Appendix 2 (see Supplemental Data accompanying the online version of this article). For a few taxa, internal primers were also used to amplify the rpl16 intron and trnS-G intergenic spacer, with the following changes in the thermal cycling program: 40 instead of 34 cycles and extension time shortened to 1 min. Successfully amplified PCR products were then purified using GFX PCR DNA and a gel band purification kit (Amersham Biosciences).
The same primers used for PCR amplification were also used for the cycle-sequencing reactions, carried out with the ABI Prism Big Dye Terminator Cycle Sequencing Ready Extraction Kit (Perkin Elmer, Applied Biosystems, Applera Europe BV, Rotkreuz, Switzerland). Sequence products were purified on MicroSpin G-50 columns (Amersham Pharmacia Biotech Europe, Dübendorf, Switzerland) and loaded on an ABI Prism 377 DNA sequencer (Perkin Elmer). Complementary strands were edited and assembled with Sequencher 3.1.1 (Gene Codes, Ann Arbor, Michigan, USA).
Data matrix composition and parsimony analysis
Nucleotide sequences of the rpl16 intron, rps16 intron, and trnS-G intergenic spacer were aligned using Clustal W, version 1.8 (Thompson et al., 1994
) and adjusted by eye. Regions of ambiguous alignment were excluded from the analysis. Individual gap positions were treated as missing data, unequivocally aligned gaps being coded as presence/absence of characters with the software Gapcoder (Young and Healy, 2003
) and then added to the sequence matrix.
Twenty-nine morphological characters were scored using herbarium and fresh specimens, pickled flowers, and when available, flower sections provided by the second author. For some taxa, the literature was also consulted (e.g., Woodson, 1933
; Pichon, 1950
; Leeuwenberg, 1997
; Morales, 1998
, 2002
). The morphological matrix, a list of the characters, character states, and explanatory notes on characters are given in Appendices 3 and 4 (see Supplemental Data). Exemplars were used as terminal taxa, because morphological variation is considerable in the two larger genera, Mandevilla and Forsteronia, which would lead to difficulties in coding character states.
The following data sets were subjected to phylogenetic analysis: (1) rpl16 intron, (2) rps16 intron, (3) trnS-G intergenic spacer, and (4) morphology. As the results of individual analyses did not show any major topological conflict, data partitions were combined in the following ways: all molecular data sets combined together (molecular combined) and all molecular and morphological data sets combined (total evidence).
Maximum parsimony analyses were performed using PAUP* 4.0b (Swofford, 2000
). All characters were unordered and equally weighted. Polymorphisms in the data matrix were treated as such, rather than as uncertainties. A heuristic search for most parsimonious trees (MPT) included an initial round of tree searches with 1000 random addition sequence replicates (RASR), holding 10 trees at each step, tree bisection-reconnection (TBR) branch swapping with MULTREES and steepest descent in effect, saving a maximum of 100 trees at each replicate. All shortest trees retained in memory were then included in a second round of searches involving exhaustive TBR branch swapping. Relative support for each node was estimated using the bootstrap resampling procedure (Felsenstein, 1985
) as implemented in PAUP employing a full heuristic search with 1000 replicates, 250 RASR, three trees held at each step, TBR branch swapping with steepest descent and MULTREES in effect, saving 10 trees at each RASR.
Morphological characters were optimized onto the strict consensus tree of the total evidence analysis using Winclada, version 1.00.08 (Nixon, 2002
), in order to identify the synapomorphies that were congruent with each of the major clades of the ingroup retrieved in our analyses. The proportion of nodes in the individual molecular data partitions that were congruent with the topology of the total evidence tree was assessed in WinClada by optimizing the unambiguous character changes of each data set separately onto the topology of the strict consensus tree obtained from the total evidence matrix. Subsequently, the ratio of nodes supported by such character changes divided by the total number of nodes was calculated. This method is analogous to the Partition Bremer Support (Baker and DeSalle, 1997
), although it has the disadvantage that, unlike Partition Bremer Support, it cannot identify whether or not a data partition contradicts a particular node.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
(see Table 1), Galactophora was also included in the initial round of taxon sampling. Results from a heuristic search showed that Galactophora did not group with the Mesechiteae, but rather came out as sister to the rest of the ingroup species, as illustrated in Fig. 1. Based on its position in the unrooted tree and its long branch compared to other taxa in our sampling, Galactophora might actually be less closely related to the ingroup than the taxa used for rooting (Pachypodium geayi, P. lamerei, and Mascarenhasia lisianthiflora). Due to this uncertainty, Galactophora was excluded from further analysis.
|
; Nixon and Carpenter, 1996
), and because none of the nodes involved in topological discrepancies in our analysis were supported by high bootstrap values (Fig. 2), the individual molecular data sets were combined. The best resolved cladograms were provided by the rpl16 and rps16 intron data sets, with most of the internal nodes receiving bootstrap support (BS) higher than 50%. Of the cladograms generated by the separate molecular data sets, only the rps16 intron tree defined a clade, corresponding to what will later be defined as the Mesechiteae clade, with BS higher than 50%, identical to that of the molecular combined and total evidence trees. The least resolved cladogram was that based on the trnS-G intergenic spacer data set.
|
|
|
The taxa belonging to tribe Apocyneae sensu Endress and Bruyns (2000)
are polyphyletic, with the six representatives included in this study dispersed among three different parts of the cladogram. The Neotropical genus Forsteronia is nested in the Mesechiteae clade, whereas Odontadenia, a large genus widely distributed in the Neotropics, forms a strongly supported clade (BS = 98%) with Secondatia, a Neotropical genus included in tribe Mesechiteae by Endress and Bruyns (2000)
. Beaumontia, Chonemorpha, and Trachelospermum, all genera of tropical and subtropical regions of Asia, form a strongly supported clade (BS = 96%) somewhat closer to the base of the tree. A clade composed of Prestonia and Rhodocalyx, both Neotropical members of Echiteae sensu Endress and Bruyns (2000)
, is strongly supported (BS = 100%).
Analysis of the percentage of nodes supported by unambiguously optimized characters for each individual partition onto the strict consensus tree of the total evidence analysis showed that the individual data sets had varying degrees of resolving power (Table 3). The highest power of resolution consistent with the total evidence tree was found for nucleotides in the rpl16 intron, with more than 80% of the nodes in the strict consensus tree of the combined analysis supported by at least one unambiguously optimized nucleotide substitution of the rpl16 intron. Slightly lower percentages were found for the rps16 intron and trnS-G intergenic spacer, with the lowest percentage provided by the gaps in the rps16 intron. A relatively high percentage (45%) of the nodes in the strict consensus of the total evidence tree was supported by at least one unambiguously optimized morphological character.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
for noncoding plastid regions, such as the occurrence of strings of mononucleotide repeats and small tandem repeat units. Considerable length variation among the individual sequences required the insertion of a large number of gaps in the alignments (Table 2), especially for the trnS-G intergenic spacer, which is consistent with the results reported by Perret et al. (2003)
for the same region. Gaps provided a considerable amount of information for our analysis, representing one-quarter of the total number of informative characters. The nucleotide substitutions were not uniform across sequences of the same DNA region, but rather alternated between more conserved and more variable regions. This heterogeneity was highest in the trnS-G intergenic spacer region. These more variable regions, which could provide potential sources of phylogenetic information at lower taxonomic levels, were excluded from our analysis because of ambiguous alignment.
Phylogenetic utility of individual data partitions
All three plastid regions were characterized by similar percentages of parsimony informative characters, ranging from 8.6% in the trnS-G intergenic spacer to 10.6% in the rps16 intron, consistent with the results reported by Schönenberger and Conti (2003)
. In contrast to this uniformity in the percentage of parsimony informative characters provided by each data set, marked differences were observed in the resolving power of the separate data partitions. The highest resolving power was found in the rpl16 intron, which yielded the highest percentage of nodes supported by at least one unambiguously optimized character onto the strict consensus of the total evidence tree (Table 3). All individual data partitions, however, including morphology, contributed partially to the resolution in the total evidence tree.
Although the rpl16 intron, rps16 intron, and trnS-G intergenic spacer have been used in several recent studies, either individually or combined with other plastid and/or nuclear regions (e.g., Oxelman et al., 1997
; Downie et al., 2000
; Asmussen and Chase, 2001
; Perret et al., 2003
), the combination of all three regions in the same data matrix has been applied only recently (Schönenberger and Conti, 2003
). This study shows their phylogenetic utility to resolve relationships at the tribal and intergeneric levels, especially when combined with morphology.
Comparison between phylogenetic hypothesis and current classification
The total evidence tree does not support the monophyly of Mesechiteae as circumscribed by Endress and Bruyns (2000)
with its nine constituent genera: Allomarkgrafia, Galactophora, Quiotania, Macrosiphonia, Mandevilla, Mesechites, Secondatia, Telosiphonia, and Tintinnabularia. Of the eight genera included in this study, six (Allomarkgrafia, Macrosiphonia, Mandevilla, Mesechites, Telosiphonia, and Tintinnabularia) form a strongly supported clade together with Forsteronia, a genus placed in Apocyneae by Endress and Bruyns (2000)
. Secondatia, on the other hand, groups with Odontadenia, the Neotropical representative of Apocyneae included among the outgroup taxa in our study. Galactophora clearly does not belong in Mesechiteae, but its relationships are uncertain at present (see Results and Fig. 1). The status of Quiotania could not be evaluated, because collection of leaf tissue was not possible (see Materials and Methods).
Addition of morphological characters to the combined molecular data set increased bootstrap support for the Mesechiteae clade from 81 to 97%. Four morphological synapomorphies were identified, which are congruent with the Mesechiteae clade, as defined in the total evidence tree: (1) leaf blade with one to many colleters on the adaxial surface, (2) anthers with a blunt-cordate to truncate base, (3) retinacle strongly united with the style head by cellular fusion, and (4) style head with five strongly protruding longitudinal ribs (Fig. 4). The structure of the style head and the manner in which it is united with the anthers (retinacle type) are both key characters in the specialized flowers of the Apocynaceae. The style head is a product of postgenital fusion of the two carpel apices, which develop into an enlarged, secretory structure. The retinacle is the region of the anther that becomes postgenitally united with the style head, thus forming a gynostegium (Fallen, 1986
; Sennblad et al., 1998
). All members of the Mesechiteae clade have a style head that is often referred to in the earlier literature as "umbraculiform" (umbrella shaped). This type of style head is characterized by having a star-like shape in cross section, from the five strongly projecting vertical ribs, which may extend along its entire length or be more or less restricted to the base. The anthers are postgenitally united with these ribs by the retinacle, which in Mesechiteae consists of a strip of specialized cells that are unusually short and strongly attached to the style head by cellular fusion. This type of retinacle sets the Mesechiteae clade apart from other Neotropical Apocynoideae, in which the retinacle is composed of longer hairs, characterized by a weaker attachment to the style head, without cellular fusion.
In contrast to these "good" characters, many of the morphological characters (e.g., habit, presence of domatia, absence vs. presence or arrangement of calycine colleters, corolla shape, length of staminal filaments, and nectary height), which have previously been used to define taxonomic groups (Woodson, 1933
; Zarucchi, 1991
; Morales, 1997
, 1998
; Williams, 1999
, 2002
), appear to be phylogenetically unreliable. For example, Woodson (1936)
and Leeuwenberg (1994)
considered Tintinnabularia to be related to Beaumontia (a genus of the outgroup tribe Apocyneae in this study), based on their shared possession of domatia and long staminal filaments.
Relationships between genera
Next we briefly discuss relationships among the three subclades of Mesechiteae, focusing on the morphological synapomorphies that are consistent within these clades.
Mesechites subclade
The Mesechites subclade is defined by one morphological synapomorphy: having the ribs of the style head restricted to the base (Fig. 4). This distinguishes the Mesechites subclade from the other two subclades of Mesechiteae, both of which are characterized by having ribs along the entire length of the style head. In the literature (e.g., Woodson, 1933
; Williams, 1999
), the inflorescences of Allomarkgrafia, Mesechites, and Tintinnabularia are often said to be cymose (to distinguish these genera from Mandevilla, which is considered to have racemose inflorescences). However, inflorescence type is a character that is difficult to interpret, especially as the inflorescence is often reduced (e.g., in Tintinnabularia mortonii, Mesechites minima, and M. rosea).
In the Mesechites subclade, Tintinnabularia is sister to a strongly supported clade composed of Allomarkgrafia and Mesechites. Tintinnabularia, described by Woodson (1936)
and comprising three species, is restricted to Mexico, Guatemala, and Honduras. It is one of the most rarely collected and thus poorly known genera in the Apocynoideae. It has been considered to be allied to other members of the Mesechiteae by most Neotropical Apocynaceae specialists (e.g., Zarucchi, 1991
; Henrickson, 1996
; Williams, 1999
), but except for Pichon (1950)
, who considered Tintinnabularia to be most closely related Forsteronia, no other specialist in the family has made an attempt to elucidate its closest relatives within the tribe.
Mesechites currently comprises 10 species, divided between the two subgenera, Eumesechites and Didymadenia, described by Woodson (1933)
. Of the three species of Mesechites included in our study, the two Cuban species, M. minima and M. rosea, representing subgenus Didymadenia, are strongly supported as sister taxa (BS = 94%). The South American representative from subgenus Eumesechites, M. mansoana, however, groups with Allomarkgrafia rather than with the other Mesechites species. This suggests that Mesechites may not be a monophyletic genus as currently circumscribed. The two genera have long been considered to be closely related. Pichon (1948b
, 1950
) considered Allomarkgrafia to be a synonym of Mesechites, arguing that the style head form was the same for the two genera and that no diagnostic character states supported the generic distinction of Allomarkgrafia. Our current results do not contradict this taxonomic interpretation; no morphological synapomorphy that could distinguish the two genera was found. More species need to be analyzed, however, before firm conclusions as to generic circumscription can be reached.
Forsteronia
The inclusion of Forsteronia in the Mesechiteae clade is somewhat unexpected; this relationship has never been proposed. Forsteronia is a relatively large Neotropical genus with some 46 recognized species, characterized by having small, subrotate flowers in thyrsiform, often dense inflorescences. Although easily distinguished from other Neotropical genera, the placement of Forsteronia within the Apocynoideae has proved difficult. Pichon (1950)
placed Forsteronia in his tribe Ichnocarpeae based on its glabrous, concave retinacle and created the subtribe Forsteroniinae to accommodate this genus together with Tintinnabularia. This subtribe was mainly defined by the presence of domatia on the abaxial surface of the leaves. Leeuwenberg (1994)
agreed for the most part with Pichon and placed Forsteronia in his tribe Apocyneae, together with many of the genera originally placed in Pichon's Ichnocarpeae. He did not recognize subtribe Forsteroniinae, however, and he placed Tintinnabularia in a different tribe, Wrightieae. Endress and Bruyns (2000)
included Forsteronia in Apocyneae, assuming that it shares the diagnostic characters states of the tribe: style head fusiform and retinacle formed by a horseshoe-shaped rim and a narrow longitudinal strip of hairs.
One of the main difficulties concerning the placement of Forsteronia is due to the morphological variation within the genus. In order to estimate this variation and the consistency of our phylogenetic results, the flower structure of an additional five species of Forsteronia, F. australis Müll.-Arg., F. portoricensis Woodson, F. rufa Müll.-Arg., F. spicata (Jacq.) G. Mey., and F. velloziana (A. DC.) Woodson, was examined, based on herbarium vouchers, pickled flowers and, when available, sections of flowers provided by the second author. All species were found to share two of the key characters states that define the Mesechiteae clade: the presence of colleters at the leaf base and anthers with a bluntly cordate to truncate base. Other characteristics, however, are more variable and are more or less intermediate between Mesechiteae and Apocyneae. One such character is the structure of the retinacle. As in other Mesechiteae, the retinacle of Forsteronia has a glabrous, concave region; but it also has a small to well-developed row of hair-like cells beneath this. These hair-like cells form a weak union between the anthers and the base of the style head, generally only by agglutination but sometimes with accompanying cellular fusion (e.g., F. spicata), as in other Mesechiteae. Furthermore, the longitudinal ribs of the style head are not always well developed. In some species they are quite conspicuous, with the characteristic "Mesechiteae" star-like shape in cross section. In other species, however, the ribs are scarcely developed (e.g., F. acouci), so that the style head is more or less pentagonal in cross section, which is more characteristic of the Apocyneae. Despite this variation in some of its morphological features, the results of our analyses of morphological variation indicate that the inclusion of Forsteronia in the Mesechiteae clade is warranted. In order to test the monophyly of Forsteronia, however, a broader taxon sampling of species chosen to adequately represent the range of variation within the genus is required.
The Mandevilla subclade
The three genera of this subclade (Macrosiphonia, Mandevilla, and Telosiphonia) have always been considered to be closely related and sometimes even synonymous. They all have racemose inflorescences, sometimes reduced to a single flower, and strongly protruding ribs extending along the entire length of the style head.
Mandevilla is the largest Neotropical genus in the Apocynoideae. It is extremely variable, with about 120 species distributed throughout the Neotropics, from Mexico to Argentina, and includes vines, erect shrubs, and even epiphytes. The flower size and structure also spans a broad range, from inconspicuous whitish, tubular flowers less than 1 cm long to brightly colored, showy infundibuliform flowers up to 5 cm long. Macrosiphonia and Telosiphonia, in contrast, contain only five and six species, respectively, which share a number of morphological characteristics. Both are erect shrubs or subshrubs with leaves covered by a dense, wooly indument on the abaxial side and occur in savannas or arid habitats. The flowers are white with a long slender tube and are presumably adapted to hawkmoth pollination. Telosiphonia was originally described as a subgenus of Macrosiphonia by Woodson (1933)
. Henrickson (1996)
, however, elevated subgenus Telosiphonia to generic rank based on characters such as inflorescence type, style head structure and pollen size. He suggested that the many similarities between Macrosiphonia and Telosiphonia are the result of adaptation to a similar habitat and pollination syndrome. The distribution of the two genera roughly coincides with the extreme northern and southern distribution of Mandevilla. Telosiphonia is restricted to the arid zones of Mexico and the southwestern United States, whereas Macrosiphonia is found in the savannas of central Brazil to Argentina. The separation of Macrosiphonia and Telosiphonia into two distinct clades strongly supports Henrickson's (1996)
ideas that the two taxa are not congeneric. Nevertheless, their recognition as distinct genera probably cannot be upheld in light of the present data because both clades are nested in Mandevilla. The morphological characteristics that have been used to distinguish Mandevilla from Macrosiphonia and Telosiphonia are rather minor, being based only on leaf indument and superficial aspects of flower structure. Woodson (1933
, p. 778) maintained them as distinct genera, but stated that "The existing distinctions between Macrosiphonia and Mandevilla are extremely tenuous." Pichon (1948b)
proposed the inclusion of Macrosiphonia in the synonymy of Mandevilla, arguing that the distinguishing characters used by Woodson (1933)
to differentiate between the two genera were inconsistent and arbitrary, making impossible an unambiguous distinction between them. Our morphological analyses identified no apomorphies exclusive to either genus, reinforcing the difficulty of upholding their current generic rank.
In his taxonomic revision of Mandevilla, Woodson (1933)
recognized two subgenera: Eumandevilla and Exothostemon. Within subgenus Eumandevilla, he recognized five sections: Laxae, Montanae, Tenuifoliae, Torosae, and Tubiflorae. The monophyly of subgenus Exothostemon, represented in our study by M. rugosa, M. rugellosa, and M. scabra, is strongly supported (BS = 100%). However, these results must be interpreted as preliminary due to the small number of Mandevilla species sampled for a genus of this size. Similarly, no conclusions can be made at this time about relationships among the sections of Mandevilla sensu Woodson (1933)
, also due to the insufficient taxon sampling. Our finding of paraphylly in M. tenuifolia with regard to M. myriophyllum suggests that the latter could be merely an extreme morphotype of the former. However, a significantly broader taxon sampling in M. tenuifolia would need to be undertaken in order to determine this with more certainty. Further studies based on more intensive taxon sampling in Mandevilla, needed to address these questions, are underway.
Galactophora and Secondatia
The exclusion of Secondatia and Galactophora from the Mesechiteae, suggested by our phylogenetic analyses, is congruent with morphology. In both genera, leaf colleters are absent, the anther base is strongly sagittate, and the retinacle is protuberant with no detectable concave region—all character states that are at odds with the synapomorphies that support the Mesechiteae clade. In Secondatia, the style head is almost circular in cross section, with no longitudinal ribs, similar to that found in Odontadenia. The retinacle structure is also of the same type as that in Odontadenia. Secondatia and Odontadenia comprised Pichon's (1950)
subtribe Secondatinae of Ecdysanthereae, and our results also support a close relationship between these two genera. In Galactophora, in contrast, the style head has five well-developed projecting ribs at the base. The presence of these ribs was the main reason for the inclusion of Galactophora in Mesechiteae by Endress and Bruyns (2000)
. Our results, however, suggest that the ribs on the style head are independently derived in Galactophora and the Mesechiteae clade. Preliminary analysis of the style-head structure of G. calycina reinforces this hypotheses. In this species, the ribs are not continuous with the main body of the style head, as in taxa of the Mesechiteae clade, but rather are formed by soft tissue that is distinct from that of the rest of the style head. The union between the style head and the anthers in Galactophora is also quite weak; the anthers are easily detached from the style head. This is in sharp contrast to the situation in taxa of the Mesechiteae clade in which the anthers are so strongly united with the style head that they can usually only be removed by ripping off an adjacent piece of the style head.
Apocyneae
Our results, especially in light of the position of Forsteronia in Mesechiteae, strongly suggest that Apocyneae sensu Endress and Bruyns (2000)
is not monophyletic. This is not unexpected, confirming their prediction (Endress and Bruyns, 2000
, p. 8) that "The Apocyneae, especially, will probably need to be divided in some way, and some rearrangement of taxa will no doubt be necessary as more data accumulate." No further conclusions can be drawn about relationships within the Apocyneae, however, due to the small number of taxa from this tribe sampled in our analysis.
Conclusions
The phylogenetic analysis presented here provides the first broad study of the Mesechiteae including representatives of all but one of its constituent genera, using both morphological and molecular characters. This represents the first step towards resolving long-standing disputes over generic delimitation and intergeneric relationships within the tribe. The newly defined Mesechiteae comprise taxa previously ascribed to eight genera: Allomarkgrafia, Forsteronia, Macrosiphonia, Mandevilla, Mesechites, Quiotania, Telosiphonia, and Tintinnabularia.
Topics to be addressed in a future study include testing the monophyly and determining the systematic position of Forsteronia in the Mesechiteae; defining the generic circumscription of Allomarkgrafia, Macrosiphonia, and Telosiphonia; testing the monophyly of the currently recognized sections within Mandevilla; and elucidating character evolution and the biogeographic history of the group.
|
FOOTNOTES |
|---|
4 mendress{at}systbot.unizh.ch
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Baker R. H. R. DeSalle 1997 Multiple sources of character information and the phylogeny of Hawaiian Drosophilids. Systematic Biology 46: 654-673[CrossRef][ISI][Medline]
Brower A. 1996 Combining data in phylogenetic analysis. Trends in Ecology and Evolution 11: 334-335
Downie S. R. D. S. Katz-Downie M. E. Watson 2000 A phylogeny of the flowering plant family Apiaceae based on chloroplast DNA rpl16 and rpoc1 intron sequences: towards a suprageneric classification of subfamily Apioideae. American Journal of Botany 87: 273-292
Endress M. E. P. V. Bruyns 2000 A revised classification of the Apocynaceae s.l. Botanical Review (Lancaster) 66: 1-56
Endress M. E. B. Sennblad S. Nilsson L. Civeyrel M. W. Chase S. Huysmans E. Grafström B. Bremer 1996 A phylogenetic analysis of Apocynaceae s. str. and some related taxa in Gentianales: a multidisciplinary approach. Opera botanica belgica 7: 59-102
Fallen M. E. 1986 Floral structure in the Apocynaceae: morphological, functional, and evolutionary aspects. Botanischer Jahrbücher für Systematik, Pflanzengeschichte und Pflanzengeographie 106: 245-286
Felsenstein J. 1985 Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Henrickson J. 1996 Studies in Macrosiphonia (Apocynaceae): generic recognition of Telosiphonia. Aliso 14: 179-195
Judd W. S. R. W. Sanders M. J. Donoghue 1994 Angiosperm family pairs: preliminary phylogenetic analysis. Harvard Papers in Botany 5: 1-51
Kelchner S. A. 2000 The evolution of noncoding chloroplast DNA and its application in plant systematics. Annals of the Missouri Botanical Garden 87: 482-498[CrossRef][ISI]
Leeuwenberg A. J. M. 1994 Series of revisions of Apocynaceae. XXXVIII. Taxa of the Apocynaceae above the genus level. Wageningen Agricultural University Papers 94: 45-60
Leeuwenberg A. J. M. 1997 Series of revisions of Apocynaceae XLIV. Craspidospermum Boj. ex A. DC., Gonioma E. Mey., Mascarenhasia A. Dc., Petchia Livera, Plectaneia Thou., and Stephanostegia Baill. Wageningen Agricultural University Papers 97: 1-124
Morales J. F. 1997 A synopsis of the genus Allomarkgrafia (Apocynaceae). Brittonia 49: 337-345[CrossRef][ISI]
Morales J. F. 1998 A synopsis of the genus Mandevilla (Apocynaceae) in Mexico and Central America. Brittonia 50: 214-232[CrossRef][ISI]
Morales J. F. 2002 Studies in neotropical Apocynaceae II: a review of the genus Fernaldia. Rhodora 104: 186-200[ISI]
Nixon K. C. 2002 WinClada, version 1.00.08. Computer program distributed by the author, at http://www.cladistics.com. Cornell University, Ithaca, New York, USA
Nixon K. C. J. M. Carpenter 1996 On simultaneous analysis. Cladistics 12: 221-241[CrossRef][ISI]
Oxelman B. M. Liden D. Berglund 1997 Chloroplast rps16 intron phylogeny of the tribe Sileneae (Carophyllaceae). Plant Systematics and Evolution 206: 393-410[CrossRef][ISI]
Perret M. A. Chautems R. Spichiger G. Kite V. Savolainen 2003 Systematics and evolution of tribe Sinningieae (Gesneriaceae): evidence from phylogenetic analyses of six plastid regions and nuclear ncpGS. American Journal of Botany 90: 445-460
Pichon M. 1948a Classification des Apocynacées: XIX. Le rétinacle des Echitoïdées. Bulletin de la société botanique de France 95: 211-216
Pichon M. 1948b Classification des Apocynacées: X. Genre Mandevilla. Bulletin Muséum National d'Histoire Naturelle, série 2 20: 211-216
Pichon M. 1950 Classification des Apocynacées: XXV, Échitöidées. Mé moires du Muséum National d'Histoire Naturelle, série B, Botanique 1: 1-143
Potgieter K. V. A. Albert 2001 Phylogenetic relationships within Apocynaceae s.l. based on trnL intron and trnL-F spacer sequences and propagule characters. Annals of the Missouri Botanical Garden 88: 523-549[CrossRef][ISI]
Schönenberger J. E. Conti 2003 Molecular phylogeny and floral evolution of Peneaceae, Oliniaceae, Rhynchocalycaceae and Alzateaceae (Myrtales). American Journal of Botany 90: 293-309
Sennblad B. B. Bremer 1996 The familial and subfamilial relationships of Apocynaceae and Asclepiadaceae evaluated with rbcL data. Plant Systematics and Evolution 202: 153-175[CrossRef][ISI]
Sennblad B. B. Bremer 2002 Classification of Apocynaceae s.l. according to a new approach combining Linnaean and phylogenetic taxonomy. Systematic Biology 51: 389-409[CrossRef][ISI][Medline]
Sennblad B. M. E. Endress B. Bremer 1998 Morphology and molecular data in phylogenetic fraternity: the tribe Wrightieae (Apocynaceae) revisited. American Journal of Botany 85: 1143-1158[Abstract]
Swofford D. L. 2000 PAUP*: phylogenetic analysis using parsimony, version 4.0b. Sinauer Associates, Sunderland, Massachusetts, USA
Thompson J. D. G. D. 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
Williams J. K. 1999 A new species of Tintinnabularia (Apocynaceae, Apocynoideae) from Honduras, with taxonomic notes. Lundellia 2: 136-141
Williams J. K. 2002 Thoreauea (Apocynaceae: Apocynoideae), a new genus from Oaxaca, Mexico. Lundellia 5: 47-58
Woodson R. E., Jr. 1933 Studies in the Apocynaceae IV. The American genera of Echitoideae. Annals of the Missouri Botanical Garden 22: 153-306[CrossRef]
Woodson R. E., Jr. 1936 Studies in the Apocynaceae IV. The American genera of Echitoideae. Annals of the Missouri Botanical Garden 23: 169-438[CrossRef]
Young N. D. J. Healy 2003 Gapcoder automates the use of indel characters in phylogenetic analysis. BMC Bioinformatics 4: 6[CrossRef][Medline]
Zarucchi J. L. 1991 Quiotania: a new genus of Apocynaceae—Apocynoideae from Northern Colombia. Novon 1: 33-36
This article has been cited by other articles:
|
|
 |
|
  A. G. Ghebretinsae, M. Thulin, and J. C. Barber Relationships of cucumbers and melons unraveled: molecular phylogenetics of Cucumis and related genera (Benincaseae, Cucurbitaceae) Am. J. Botany, July 1, 2007; 94(7): 1256 - 1266. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Marazzi, P. K Endress, L. P. d. Queiroz, and E. Conti Phylogenetic relationships within Senna (Leguminosae, Cassiinae) based on three chloroplast DNA regions: patterns in the evolution of floral symmetry and extrafloral nectaries Am. J. Botany, February 1, 2006; 93(2): 288 - 303. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||
