| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
(American Journal of Botany. 2008;95:1136-1152.) doi: 10.3732/ajb.0800096 © 2008 Botanical Society of America, Inc. |
What's this? |
Systematics and Phytogeography |
2 Rancho Santa Ana Botanic Garden, 1500 North College Avenue, Claremont, California 91711 USA 3 Department of Botany, California Academy of Sciences, 55 Concourse Drive, Golden Gate Park, San Francisco, California 94118 USA
Received for publication 17 March 2008. Accepted for publication 9 June 2008.
ABSTRACT
Acanthaceae (Asteridae; Lamiales) include 
4000 species and encompass a range of morphological diversity, habitats, and biogeographic patterns. Although they are important components of tropical and subtropical habitats worldwide, inadequate knowledge of the family’s phylogenetic framework has impeded comparative research. In this study, we sampled all known lineages of Acanthaceae including Andrographideae. Also included were eight of 13 genera whose relationships remain enigmatic. We used sequence data from nrITS and four chloroplast noncoding regions, and parsimony and Bayesian methods of analysis. Results strongly support most aspects of relationships including inclusion of Avicennia in Acanthaceae. Excepting Neuracanthus, newly sampled taxa are placed with strong support; Kudoacanthus is in Justicieae, Tetramerium lineage, and the remaining enigmatic genera are in Whitfieldieae or Barlerieae, and Andrographideae are sister to Barlerieae. This last result is unanticipated, but placement of Andrographideae based on structural characters has been elusive. Neuracanthus is monophyletic but placement relative to (Whitfieldieae (Andrographideae + Barlerieae)) is weakly supported. Many clades have clear morphological synapomorphies, but nonmolecular evidence for some remains elusive. Results suggest an Old World origin with multiple dispersal events to the New World. This study informs future work by clarifying sampling strategy and identifying aspects of relationships that require further study.
Key Words: Acanthaceae • Andrographideae • Barlerieae • biogeography • Lamiales • Lankesteria • Neuracanthus • phylogeny • Whitfieldieae
Over the last 10 years or so, students of Acanthaceae have made considerable progress toward delimiting the family (Schönenberger and Endress, 1998
; Schönenberger, 1999; Schwarzbach and McDade, 2002; Wortley et al., 2007), identifying the family’s major clades (Hedrén et al., 1995; Scotland et al., 1995; McDade and Moody, 1999; McDade et al., 2000b), and unraveling relationships within many of these clades (e.g., Acantheae: McDade et al., 2005; Justicieae: McDade et al., 2000a; Justicieae: Isoglossinae: Kiel et al., 2006; Tetramerium lineage: Daniel et al., 2008; Thunbergioideae: Borg et al., in press). In parallel, Scotland and colleagues have undertaken important family-wide studies of characters such as corolla aestivation (Scotland et al., 1994) and pollen (Scotland, 1993; Scotland and Vollesen, 2000) that shed light on relationships in Acanthaceae and provide morphological synapomorphies for many lineages.
On the basis of this body of work and the important historical taxonomic and comparative research on the group (e.g., Nees von Esenbeck, 1847; Lindau, 1895; Bremekamp, 1953, 1965; Raj, 1961; Balkwill and Getliffe-Norris, 1988), Scotland and Vollesen (2000) placed 201 of 221 genera accepted by them into seven infrafamilial taxa of Acanthaceae, leaving only 20 unplaced. At essentially the same time, Manktelow et al. (2001) presented results showing that three of these 20, Whitfieldia, Chlamydacanthus, and Lankesteria, compose a clade referred to as Whitfieldieae by these authors; with somewhat limited taxon sampling, Whitfieldieae was placed sister to Barlerieae. Of the 17 remaining unplaced genera, four more were resolved in preparing for the current study as set forth in Table 1: Aphelandrella, "Idiacanthus," Morsacanthus, and Perenideboles.
|
), based in part on unusual pollen morphology (Fig. 1; see also Figs. 25c,d and 26 in Scotland and Vollesen, 2000). However, monophyly of Andrographideae remains to be tested, and its relationships to other Acanthaceae need to be assessed. Knowledge of the phylogenetic framework of the family as a whole and understanding of evolutionary and biogeographic patterns in the group will be advanced by clarifying relationships of Andrographideae and by determining the placement of the remaining enigmatic genera.
|
MATERIALS AND METHODS
Taxon sampling
We were able to obtain material suitable for DNA extraction for species representing eight of the 13 accepted genera that were not assigned to lineages by Scotland and Vollesen (2000) or Manktelow et al. (2001). Among these genera, only Neuracanthus is circumscribed to include more than a few species (see Table 1); six of 30 species of Neuracanthus, including both African and Malagasy taxa, were sampled to test monophyly of the genus. We also included one of five species of Forcipella (placed by Scotland and Vollesen [2000] in Justicieae) because Kiel et al. (2006) excluded this taxon from Isoglossinae (Justicieae) and suggested that it might, in fact, be unrelated to Justicieae. Because the placement of Lankesteria with Whitfieldieae was not strongly supported in the analysis of Manktelow et al. (2001), we added two species (for a total of three), including both a Malagasy and an additional African species. Representatives of five of eight genera assigned to Andrographidinae by Scotland and Vollesen (2000) were included to test monophyly of the taxon and to place it relative to other lineages of Acanthaceae s.l. Sixty-six additional species were included to represent the other known lineages of Acanthaceae s.l.: Ruellieae (4 species), Justicieae (21 species), Barlerieae (11 species), Whitfieldieae (sensu Manktelow et al. [2001]; 8 species), Acantheae (7 species), Avicennia (4 species), Thunbergioideae (7 species), and Nelsonioideae (3 species). As knowledge of relationships within these lineages and the availability of material permitted, taxa were chosen to represent known sublineages, to span the basal node within each lineage to break long branches, and to encompass geographic ranges. Martynia annua and Sesamum indicum were used as outgroups; numerous studies of Lamiales have shown Acanthaceae s.l. to be monophyletic and Martyniaceae and Pedaliaceae to be among the lineages of Lamiales most closely related to the family (e.g., Schwarzbach and McDade, 2002; Oxelman et al., 2005; Wortley et al., 2005). The total data set thus included 85 taxa (Appendix 1).
Molecular methods
Fresh leaf material, leaf material dried in silica gel, or recently collected herbarium specimens were used as sources of DNA. Total genomic DNA was extracted using the modified CTAB method of Doyle and Doyle (1987) except that Qiagen DNeasy (Qiagen, Valencia, California, USA) kits for plant tissue were usually used for herbarium samples. Procedures for amplifying all of the genic regions used here have been described elsewhere as follows: trnL-F (McDade and Moody, 1999), rps16 and trnS-G (McDade et al., 2005), trnT-L (Kiel et al., 2006), and nrITS (McDade et al., 2000b; additional primers reported by Daniel et al., 2008).
To optimize sequencing results, we ran PCR products on a 1% agarose gel for several hours, the dominant band was excised, and the template was purified using Qiagen QIAQuick (Qiagen) gel extraction kits. Sequences were generated on Beckman (Beckman Coulter, Fullerton, California, USA) or ABI (Applied Biosystems, Foster City, California, USA) automated sequencers; both strands were sequenced for verification and to complete the sequences. Electropherograms of all sequences were proofread manually in the program 4Peaks version 1.7.2 (Griekspoor and Groothuis, 2006). Overlapping portions were reconciled by reverse-complementing one, aligning the two, and double-checking any inconsistencies against the electropherograms; mismatches that could not be resolved were coded as uncertain.
Sequencing strategy
We sought data for all five genic regions for all taxa except that we did not attempt to sequence the three most slowly evolving chloroplast (cp) regions (i.e., trnL-F, rps16, trnT-L) for all species of Neuracanthus or Avicennia (the most densely sampled lineages) because these vary little within genera. In practice, sequences for a few target taxa were not obtainable for each genic region (Table 2) despite attempts with multiple primers and reaction conditions. Notably, perhaps owing to mutations in priming sites, rps16 and trnT-L sequences were not obtainable for a number of Whitfieldieae (see Appendix 1). Thunbergioideae (notably species of Mendoncia) were also often difficult to sequence, especially for trnS-G and nrITS. Finally, nrITS presented sequencing challenges for taxa of Andrographideae, perhaps owing to DNA quality from herbarium specimens. We studied the impact of these taxonomically concentrated missing data by examining results from analyses of data sets that included all taxa but sequence data only for those regions for which representatives of the taxon in question (i.e., Whitfieldieae, Thunbergioideae, Andrographideae) had complete sequence data. For example, in the case of Whitfieldieae, results from analysis of the data set of sequences for trnL-F + trnS-G, with almost no missing data for Whitfieldieae, were compared to the analysis including all genic regions. In no case was there evidence of an impact of missing sequence data on our results.
|
) following methods discussed by Morrison (2006) for manual alignment of sequence data for phylogenetic analysis and Graham et al. (2000). Sequences from three of five genic regions were easily aligned across sampled taxa. In contrast, two regions of trnS-G and one portion of the ITS1 region of nrITS could not be aligned across all taxa; these were omitted from analyses. As noted by a number of authors (e.g., McDade and Moody, 1999; Shaw et al., 2007), the noncoding cp regions have a relatively high frequency of parsimony informative indels (i.e., length mutations). Indels were scored for phylogenetic analysis following Graham et al. (2000: S86) except that overlapping gaps were placed to minimize the number of mutations required as described by Kelchner and Clark (1997). Thirty-two, 24, 22, and 39 indels were added to the trnL-F, rps16, trnT-L, and trnS-G data matrices, respectively, as presence/absence characters. No indels were added to the nrITS data set because no length mutations met the criteria employed. Table 2 compares the five DNA regions in terms of length, variability, number of taxa sequenced, and missing data.
For analysis, data matrices for the five DNA regions were prepared as NEXUS files in the program MacClade version 4.06 (Maddison and Maddison, 2000). We tested for congruence among genic regions sequenced using the incongruence length difference [ILD] test (Farris et al., 1994; implemented in the program PAUP* version 4.0 [Swofford 2000] as the partition homogeneity test with 200 replicates, 25 random addition sequences, maxtrees = 5000). The cp regions were either not significantly incongruent by the ILD test or marginally so (i.e., P < 0.05 but some of the randomized data partitions yielded summed lengths 
the sum of tree lengths from the original partitions, and summed tree lengths from all random partitions < 1% longer than the original). As a result, data for all four cp regions were combined. The ILD test indicated significant incongruity between the combined cp and nrITS data (P < 0.05). We analyzed the cp and nuclear data separately, and then compared the results to identify taxa or clades that were placed differently by data from the two genomes (Appendices S1 and S2, in Supplemental Data with online version of this article, are bootstrap consensus trees from separate analyses of the cp and nrITS data, respectively). The nrITS bootstrap consensus tree was considerably less resolved than the cp tree. The only resolved difference involved the placement of Avicennia sister to Acanthoideae by the nrITS data albeit with weak support (BS = 59%), whereas the cp data place Avicennia sister to Thunbergioideae with stronger support (BS = 100%). In the absence of strongly supported differences in the hypotheses of phylogenetic relationships supported by the cp vs. nuclear data, we combined the data.
The combined data set (i.e., all four cp genic regions + nrITS; archived in TreeBASE as S2076 and M3890, http://TreeBASE.org) was analyzed using maximum parsimony (MP) as the optimality criterion in PAUP* and using Bayesian Markov chain Monte Carlo (MCMC) inference (Yang and Rannala, 1997) in the program MrBayes version 3.0 beta 4 (Ronquist and Huelsenbeck, 2003). All parsimony analyses used rigorous heuristic search strategies designed to find all islands (sensu Maddison, 1991) of equally parsimonious trees (i.e., >50 random addition sequences, tree-bisection-reconnection [TBR] branch swapping); results were summarized as strict consensus trees. Results of separate analyses in the program MrModeltest version 2 (Nylander, 2004) indicated that the best-fit model of evolution for data from both genomes was general time reversible (GTR; Tavaré, 1986) plus parameters for proportion of invariant sites (I; Reeves, 1992) and gamma-distributed rate variation (G; Yang, 1993). The Bayesian analysis was run with these settings and with data from the cp and nuclear data treated as "unlinked" data partitions; taxa missing data for nrITS were excluded from the Bayesian analysis. Two replicates of three heated and one "cold" chain were run for 5000000 generations, with trees saved every 1000 generations. Both runs reached log-likelihood stationarity after about 115000 generations. Bayesian posterior probability values for branches were determined by opening the tree file produced by MrBayes in PAUP*, filtering to remove the first 120 trees (corresponding to those sampled during the burn-in phase), and then constructing the majority rule consensus tree. The two replicates produced postburn-in consensus trees of identical topology and nearly identical posterior probability support for branches; when posterior probabilities differed between the two, the lower value is presented here.
Support for individual branches in the parsimony trees was evaluated using nonparametric bootstrap values (BS; Felsenstein, 1985) and decay indices (DI; Bremer, 1988; Donoghue et al., 1992). Bootstrap values are from at least 100 replicates with 10 random addition sequences and TBR branch swapping. The DIs for each branch were determined by using MacClade to prepare a batch file that directed PAUP* to find the shortest trees inconsistent with each of the branches resolved in the strict consensus of MP; these searches used 100 random addition sequences to find the shortest trees, saving only 500 trees per replicate as topology was not of interest. The difference between the length of these trees and the globally shortest trees is the DI for the branch in question.
Hypotheses regarding the phylogenetic placement of the taxa newly studied here have largely not been advanced. In the few cases where placements other than those found here have been suggested, the source of confusion was usually clear in the context of our results, as discussed later. We explicitly tested the placement of two taxa as follows. (1) Lankesteria was placed by Manktelow et al. (2001) as the basal member of Whitfieldieae, but this relationship was not strongly supported. The genus was placed by Lindau (1895) in Contortae, a group that largely paralleled our current concept of Ruellieae. In contrast, Bremekamp (1944) suggested a relationship with Pseuderanthemum, a genus that is part of the basal lineage of Justicieae (sensu McDade et al., 2000a). Given that Lankesteria has continued to be difficult to place with confidence (described later), we here explicitly tested these alternative hypotheses regarding relationships. (2) Schwarzbach and McDade (2002) placed Avicennia with Acanthaceae s.l. with strong support and as sister to Thunbergioideae with only moderate support from DNA data. This relationship is not, however, supported by any known morphological synapomorphies and, as a genus of mangrove shrubs and trees, Avicennia does not seem at home in a lineage of vines. We thus tested the hypothesis that Avicennia is instead part of Acanthoideae. In addition, the Old World genus Chlamydacanthus was not monophyletic in our results Because this unexpected result impacts interpretation of biogeographic patterns, we also tested whether our data can reject monophyly of the sampled members of Chlamydacanthus.
Tests of alternative hypotheses used trees with only the branch in question constrained. These trees were loaded into PAUP*, and the program was asked to find the shortest trees consistent with the constraint using the same search strategy as described for calculating decay indices. One randomly chosen MP tree consistent with the constraint was compared to one randomly selected MP tree using Templeton’s (1983) test (reported as z statistics). The same strategy was used to compare likelihood scores of trees reflecting the alternative phylogenetic hypothesis with all likelihood parameters (except base frequencies for which empirical values were used) estimated using one randomly selected MP tree. These parameters were then used as the model to compare likelihood scores of the most likely tree to that of the tree consistent with the alternative phylogenetic hypothesis using the Kishino–Hasegawa RELL test (Kishino and Hasegawa, 1989; K-H RELL) as implemented in PAUP*. Tests were one-tailed because an optimal tree was one of the trees being compared (Felsenstein, 2004:369).
Pollen morphology
Pollen of a number of the previously unplaced genera and of Andrographideae was studied for comparison to that of related plants. Unacetolysed pollen was prepared and examined as described by Daniel (1998) on a Zeiss/LEO 1450VP (LEO, Cambridge, UK; Carl Zeiss SMT, Peabody, Massachusetts, USA) scanning electron microscope. Specimens from which pollen was examined are listed in the figure legends.
RESULTS
Molecular evolution
Assessed as proportion of parsimony informative sites, nrITS was most variable and trnL-F least variable (Table 2). However, the nrITS sequences used here are considerably shorter than all of the cp regions, in significant part owing to omission of an unalignable portion of ITS1; nrITS also had no length mutations that met our criteria for scoring as indels. As a result, all the cp regions contributed more parsimony informative substitutions, and each also contributed 22–39 parsimony informative indels.
Phylogenetic relationships
Results from the parsimony and Bayesian analyses were entirely congruent except for taxa omitted from the Bayesian owing to missing nrITS data. As detailed in Materials and Methods, taxa omitted from the Bayesian analysis were placed by parsimony as expected based on existing taxonomic concepts. We thus present strict consensus MP trees with support values from both parsimony and Bayesian analyses; branches missing posterior probabilities reflect taxa omitted from the Bayesian analysis. There is strong support (i.e., BS > 90%; Bayesian posterior probability, BPP = 100%) for almost all aspects of relationships (see branch support values in Figs. 2–5); in describing relationships below, we highlight support only when it is weak (i.e., BS < 70%, BPP < 100%).
|
|
|
|
) of Justicieae (Fig. 4).
The remaining taxa that are studied phylogenetically for the first time here are part of a weakly supported clade (BS < 50%, DI = 1; BPP = 72%) that is sister to Ruellieae + Justicieae (Fig. 2). Within this weakly supported clade, there is strong support for monophyly of Barlerieae (including Acanthostelma, Golaea, Lasiocladus and Acanthura, all placed here for the first time), and of Andrographidieae, and for a sister relationship between these two: Barlerieae + Andrographidieae (Fig. 5). Core Whitfieldieae are strongly supported as monophyletic, including Camarotea and Leandriella, both placed here for the first time, and Forcipella, which was placed by Scotland and Vollesen (2000) in Justicieae. Lankesteria is monophyletic with strong support and sister to core Whitfieldieae but with only weak support for this last aspect of relationships (BS < 50, DI = 1; BPP = 91). The Bayesian analysis supports monophyly of all of these taxa together: ((Barlerieae + Andrographideae) Lankesteria, core Whitfieldieae), but this clade has little support from parsimony (i.e., BPP = 99%, but BS < 50%, DI = 1). Lastly, monophyly of the genus Neuracanthus is strongly supported, but placement of this genus sister to ((Barlerieae + Andrographideae) (Whitfieldieae s.l.)) is weakly supported by both Bayesian and parsimony analyses (Figs. 2, 5). For brevity, we refer to this entire lineage by the acronym BAWN. Relationships within each of the component lineages of BAWN, especially as regards taxa that are here placed for the first time are discussed, along with evidence other than DNA data that supports (or contradicts) the pattern of relationships presented in Figs. 2–5
.
Our data could not reject the hypothesis that Avicennia is sister to Acanthoideae rather than to Thunbergioideae; likewise, monophyly of Chlamydacanthus could not be rejected by our data (Table 3). In contrast, our data reject placement of Lankesteria with either Justicieae or Ruellieae (Table 3).
|
Relationships among major lineages
Our results confirm relationships among Nelsonioideae, Thunbergioideae, and other Acanthaceae s.l. that have been reported in other studies (e.g., Hedrén et al., 1995; Scotland et al., 1995; McDade and Moody, 1999; McDade et al., 2000b; Schwarzbach and McDade, 2002). Notably, Nelsonioideae and Thunbergioideae have been considered "transitional" taxa between other Lamiales and Acanthaceae (e.g., Cronquist, 1981) based on the shared presence of explosively dehiscent capsules, as well as on the absence of synapomorphies characteristic of all or most Acanthoideae (e.g., retinacula, cystoliths). Scotland and Vollesen (2000) indicate that plants belonging to Nelsonioideae and Thunbergioideae lack explosively dehiscent capsules, as indeed is the case for Mendoncia (with drupes) and Avicennia (with odd fleshy capsules). However, we have grown plants of multiple species of Nelsonioideae and of Thunbergia and have observed explosive dehiscence of the capsules of these plants. Unfortunately, assessment of this trait as a synapomorphy for Acanthaceae s.l. is made difficult by the fact that relationships among Lamiales remain unresolved (Wagstaff and Olmstead, 1997; Oxelman et al., 1999; Schwarzbach and McDade, 2002; Wortley et al., 2005, 2007; see also Tank et al., 2006) such that the closest relatives of Acanthaceae s.l. remain uncertain. In any event, the fruit types of Mendoncia and Avicennia are autapomorphic regardless of whether explosively dehiscent capsules are synapomorphic for Acanthaceae s.l. or for a larger group.
Nelsonioideae
Descending-cochlear aestivation of the corolla appears to be synapomorphic for Nelsonioideae (Scotland et al., 1994; Scotland and Vollesen, 2000). Reduced ovule number may be synapomorphic for Acanthaceae exclusive of nelsonioids; whereas Nelsonioideae are characterized by six to numerous ovules, reduction to four seems to be synapomorphic for Thunbergioideae + Acanthoideae, with secondary increases in Andrographideae and at least two lineages of Ruellieae. It has been suggested that absence of endosperm is synapomorphic for Thunbergioideae + Acanthoideae (seeds of Nelsonioideae have endosperm) but, as noted by Scotland and Vollesen (2000), this trait remains poorly studied across Acanthaceae s.l. and merits further study.
Thunbergioideae and Avicennia
The work of Schönenberger (1999; Schönenberger and Endress, 1998) has pointed to a number of synapomorphies that strongly support monophyly of Thunbergioideae. As reported previously by Schwarzbach and McDade (2002) and Borg et al. (2008), our results place Avicennia sister to Thunbergioideae (Figs. 2, 3). Schwarzbach and McDade (2002) reviewed nonmolecular data in detail and found considerable evidence consonant with the placement of Avicennia with Acanthaceae s.l., but not with Thunbergioideae. These authors suggested that placement of Avicennia sister to Acanthoideae better accommodates the nonmolecular evidence. In fact, our results cannot reject this placement (Table 3). Although the current study brings more than twice as many parsimony informative characters to bear on the problem of the phylogenetic placement of Avicennia compared to Schwarzbach and McDade (2002), the data set continues to be dominated by evidence from the cp genome, as is the data set used by Borg et al. (2008). Cleary, chloroplast data consistently support placement of Avicennia with Thunbergioideae. We are currently testing the placement of Avicennia by obtaining sequence data from mitochondrial loci and increasing our nuclear data set beyond nrITS.
Acanthoideae: Acantheae
The present results confirm monophyly of Acanthoideae (i.e., the retinaculate clade), and the sister relationship between Acantheae and the cystolith clade (Fig. 2) as has been shown in earlier studies (Hedrén et al., 1995; Scotland et al., 1995; McDade and Moody, 1999; McDade et al., 2000b). Acanthoideae are a very large clade (>4000 species with many New World Justicieae yet to be described) that is marked by the presence of retinacula. This structure is derived from the funiculus that persists and becomes woody as fruits mature, forming a hook that subtends each seed. As elegantly described by Witztum and Schulgasser (1995), the retinacula propel seeds away from the parent plant when fruits dehisce. Fruit type is remarkably conserved in acanthoids: fruits vary in size from a few mm to more than 10 cm in length but, so far as is known, all are explosively dehiscent capsules with seeds borne on retinacula.
McDade et al. (2005) proposed monothecate anthers as synapomorphic for Acantheae and discussed character evolution and biogeographic patterns within the clade. Acantheae also have colpate pollen grains that lack the pore-like endoapertures that are present in many other Acanthaceae and Lamiales. This trait has been interpreted as a synapomorphy for Acantheae by Scotland and Vollesen (2000), but this hypothesis needs to be evaluated with both additional palynological data and improved resolution among Lamiales. Acantheae also lack cystoliths; these structures apparently evolved in the common ancestor of their sister group, a lineage that includes the majority of Acanthaceae (3500 species). Although much remains to be learned about cystoliths (e.g., their chemical composition, range of form and distribution across plant structures, function), the presence of these plant crystals is, so far as is known, universal in this large clade.
Acanthoideae: Justicieae and Ruellieae
Justicieae and Ruellieae are each strongly supported as monophyletic and are sister taxa (Figs. 2, 4). Remarkably, no nonmolecular synapomorphies have been posited for Ruellieae + Justicieae. Manktelow (2000) described a lamellar structure that partitions the corolla tube obliquely lengthwise in almost all putative members of Ruellieae. This "filament curtain" is derived from the synstapetal portion of the corolla tube + filaments. Manktelow (2000) hypothesized that the filament curtain represents a structural synapomorphy for Ruellieae and that putative Ruellieae that lack the structure would be shown to have relationships elsewhere in Acanthaceae. Manktelow et al. (2001) indeed showed that a number of putative Ruellieae that lack the filament curtain belong elsewhere in the retinaculate clade, lending support to the idea that this structure is synapomorphic for Ruellieae. The work of Schmidt-LeBuhn (2003; Schmidt-LeBuhn et al., 2005), Tripp (2007; Tripp et al., in press) and of Scotland and colleagues (e.g., Carine and Scotland, 1998, 2000; Scotland, 1998; Bennett and Scotland, 2003; Moylan et al., 2004a, 2004b) is shedding further light on the evolution of this large (800 species) and morphologically diverse clade.
Within Justicieae, the analyses presented here returned relationships and lineages essentially as in the results of McDade et al. (2000a) (Fig. 4). McDade et al. (2000a) proposed tricolporate hexapseudocolpate pollen as synapomorphic for Justicieae, albeit with many subsequent shifts to other pollen types. Our results show that the heretofore unplaced genus Kudoacanthus is part of the Tetramerium lineage (sensu Daniel et al., 2008). In Flora of China (Hu, 2002), Kudoacanthus, a rarely collected, unispecific genus endemic to Taiwan was treated in Justicieae: Asystasinae. However, this last taxon is not a valid lineage in that it included genera placed in at least three lineages of Justicieae in our results (i.e., Asystasia, Pseuderanthemum: Pseuderanthemum lineage sensu McDade et al., 2000a; Isoglossa: Isoglossinae sensu Kiel et al., 2006; Kudoacanthus: Tetramerium lineage sensu Daniel et al., 2008). Like other members of the Tetramerium lineage, plants of this genus have two stamens; no staminodes; anthers with thecae subequal in size, parallel and lacking appendages; and four ovules per ovary. None of these characteristics is synapomorphic for the Tetramerium lineage, and in fact, Daniel et al. (2008) were not able to identify clear nonmolecular synapomorphies for the clade. Neither pollen nor karyotype of Kudoacanthus has been studied. It is important to note that the Tetramerium lineage is sparsely sampled in the current study such that the placement of this genus relative to other members of the lineage should not be over-interpreted. Notably, Kudoacanthus is only the second known Asian member of this clade (the other is Clinacanthus), and relationships between these and other members of the clade need to be assessed with a richer sample of members of the Tetramerium lineage. Kudoacanthus will be included in a forthcoming study revisiting the Tetramerium lineage with additional taxon sampling to test a number of hypotheses that stem from our initial study of this lineage and to establish biogeographic patterns.
Acanthoideae: BAWN (Barlerieae, Andrographideae, Whitfieldieae and Neuracanthus)
The remaining genera that are here placed phylogenetically for the first time are part of the weakly supported lineage that is sister to Justicieae + Ruellieae. We refer to this lineage as BAWN, an acronym derived from the names of the constituent lineages: Barlerieae, Andrographideae, Whitfieldieae, and Neuracanthus (Fig. 5). In discussing the BAWN clade, we begin from distal portions and proceed toward the base because that will establish strongly supported relationships and provide character context within which to examine the basal, weakly supported portions of this clade.
Barlerieae
Barlerieae are strongly supported as monophyletic and a number of the heretofore unplaced genera are in this lineage: Old World Acanthostelma, Golaea, Lasiocladus, and New World Acanthura. Quincuncial aestivation has been proposed as a synapomorphy for Barlerieae (Scotland et al., 1994, Scotland and Vollesen, 2000). Of the newly placed genera, Lasiocladus has this aestivation type (Benoist, 1967), but the trait has apparently not yet been studied for the other three. As noted by Manktelow et al. (2001), nearly all Barlerieae, as then understood, have seeds with hygroscopic trichomes. The presence of trichomes on the seeds of Acanthostelma and of Golaea was noted in the original description of each (respectively, Bidgood and Brummitt, 1998; Chiovenda, 1929). Mandy Cadman (Pt. Elizabeth, South Africa, unpublished data) observed trichomes also on seeds of Lasiocladus, but fruits and seeds are apparently unknown for Acanthura. As discussed in more detail later, among the BAWN clades, hygroscopic trichomes also occur on seeds of Lankesteria and Neuracanthus. Among other lineages of Acanthaceae, seeds with hygroscopic trichomes are known from at least Ruellieae and Blepharis (Acantheae), but these are apparently not homologous to those in Barlerieae (Grubert, 1974; Scotland et al., 1995; Manktelow, 1996). Many Barlerieae have pollen with coarsely reticulate interapertural sculpturing of the exine (Fig. 6C–I), but this trait is not consistent across the clade (Raj, 1961). Notably, Lasiocladus has pollen with the interapertural exine gemmate to rugulate (Muller et al., 1989; Fig. 6B). Pollen has not been studied for a number of other Barlerieae (e.g., Acanthostelma, Golaea). The phylogenetic status of pollen traits remains to be assessed. In sum, we propose quincuncial aestivation as synapomorphic for Barlerieae as here expanded. Additional synapomorphies may be found in pollen and other characters once additional, lineage-wide comparative studies can be undertaken. We return to hygroscopic trichomes on seeds later, in discussing the phylogenetic status of the BAWN clade.
|
discovered two new species that share characters of Golaea, Acanthostelma and Crabbea; notably, all these plants have inflorescences of densely bracteate spherical heads and fan-shaped stigmas. Thulin (2006) described the two new species in Crabbea and transferred Golaea and Acanthostelma to Crabbea. In the generic description, Thulin (2006) indicates that species of Crabbea have quincuncial aestivation but does not explicitly indicate that he has seen this feature in any of the Somali species. Our results are consistent with treating Golaea and Acanthostelma in Crabbea, provided that Lasiocladus is also treated in this last genus (Fig. 5).
Malagasy Lasiocladus was placed by Benoist (1967) in "Tribu des Crabbéés," a group that he distinguished from other Acanthoideae by the combination of cystoliths, quincuncial aestivation (shared with his Tribu Barléreéés), and a regular calyx (vs. zygomorphic in Tribu Barléreéés). As in Crabbea s.l., inflorescences of plants of Lasiocladus are subtended by an involucre formed by 6–8 free bracteoles. However, these structures are lanceolate so that inflorescences do not take the form of spherical heads. Stigmas of these plants are also fan-shaped, but this trait requires additional study. A number of additional species of Crabbea and Lasiocladus should be studied phylogenetically before reassessing the taxonomy of these plants.
The two sampled members of Barleria are sister taxa and are sister to the Crabbea clade just described. Barleria is a large genus of about 300 species (Balkwill and Balkwill, 1998) such that the present sample suggests monophyly but is inadequate for further interpretation. This genus has long been distinguished from other plants with quincuncial aestivation based on the calyx of four unequal segments; the anterior and posterior are largest and the lateral lobes are narrower and often shorter.
The Barleria clade (i.e., Acanthostelma through Barleria, Fig. 5) is sister to the Lepidagathis clade, which includes all sampled species of Lepidagathis, Lophostachys, and Acanthura. Notably, Lepidagathis is not monophyletic, with some species placed sister to the sampled members of Lophostachys and others more closely related to Acanthura. However, the generic status of both Acanthura and Lophstachys has been questioned. In androecial (including pollen) characters, Acanthura is consistent with some species of Lepidagathis (e.g., compare Fig. 6A, D–F) and D. C. Wasshausen (U. S. National Museum of Natural History, unpublished data) indicates that the name is a junior synonym of Lepidagathis (i.e., A. mattogrossensis conforms to a previously described species of Lepidagathis). Likewise, Benoist (1911) noted that species of Lophostachys, a New World group of 15 species, could as readily be accommodated in Lepidagathis. Cintia Kameyama (Institute of Botany, São Paolo, Brazil, unpublished data) has noted affinities between Lophostachys and Lepidagathis; plants of these genera also have similar pollen morphology (cf. Fig. 6D–I). Following Kameyama’s work, Wasshausen and Wood (2004) made new combinations for Bolivian species of Lophostachys in Lepidagathis. As traditionally treated (i.e., excluding Acanthura and Lophostachys), Lepidagathis is a large genus of 100 species, of which 10 occur in the New World. Clearly, additional sampling of species traditionally treated in both Lophostachys and Lepidagathis is necessary. Diagnostic morphological characteristics that might distinguish these genera from one another have not been described. If the patterns of relationships recovered here survive further sampling, then treating all of these in Lepidagathis would be phylogenetically valid. On the other hand, if the two strongly supported subclades recovered here are borne out with further sampling, two genera would be equally appropriate, especially if unambiguous morphological characters can be identified for these subclades.
Barlerieae are notable for their biogeographic patterns. Both Barleria and Lepidagathis are extremely wide-ranging in the Old World and are also present (but with few species) in the New World; geographic patterns among subgenera and sections of Barleria were documented in detail by Balkwill and Balkwill (1998). Although our sample of these genera is inadequate to determine patterns among Old World areas, both Asian and African taxa are present in two of seven sections of Barleria recognized by Balkwill and Balkwill (1997) and in both subclades of Lepidagathis recovered here, suggesting a complex pattern of vicariance, dispersal and/or extinction. Of special interest in the Old World, Lasiocladus is restricted to Madagascar, there are numerous Malagasy species of Barleria, and a few species of Lepidagathis occur on Madagascar as well. Our results regarding relationships among Barlerieae thus indicate at least three dispersal events to Madagascar, more if Malagasy Barleria or Lepidagathis is polyphyletic. The sister relationship between Lasiocladus and Golaea + Acanthostelma is not strongly supported by parsimony, and it is thus not clear whether the Somali taxa reached northeast Africa from farther south in Africa or from Madagascar. So far as known, only one member of the Barleria clade occurs in the New World, Barleria oenotherioides, a species that also occurs in West Africa (Daniel, 1995; Balkwill and Balkwill, 1998). This disjunct distribution would seem to be explainable only by a fairly recent dispersal event to the New World. Relationships in the Lepidagathis clade indicate at least two dispersal events from the Old World to the New: once to account for New World Acanthura and Lepidagathis alopecuroidea (this last species also occurs in the Old World) and once for the common ancestor of New World Lophostachys. In sum, Barlerieae are likely ancestrally an Old World clade with evidence of (1) a complex biogeographic pattern among paleotropical areas, (2) at least three dispersal events to Madagascar, and (3) at least three dispersal events to the New World.
Our results point to a number of directions for future research regarding Barlerieae. Clearly, it is necessary to add many more species of Barleria and Lepidagathis s.l. to test monophyly, to test the infrageneric classification of Barleria proposed by Balkwill and Balkwill (1997) and to identify sublineages of Lepidagathis. Better sampling of Crabbea and Lasiocladus is also necessary to test monophyly of these genera and clarify the taxonomic disposition of the Somali taxa. The present analysis is also missing a number of genera that were placed in Barlerieae by Scotland and Vollesen (2000) and that likely belong here, including a number that are of biogeographic interest: Barleriola (6 species, West Indies), Borneacanthus (6, Borneo), Boutonia (1, Madagascar), Chroesthes (3–4, China to SE Asia), and Hulemacanthus (1, New Guinea). Others that may belong here include three small, rarely collected and poorly known Malagasy genera with four stamens and bithecous anthers: Pericalypta (treated as unplaced by Scotland and Vollesen [2000] but with "prefloraison quinconciale" according to Benoist [1967, p. 125] and pollen similar to Lasiocladus according to Muller et al. [1989]), Podorungia (treated as Justicieae by Scotland and Vollesen [2000], but with "prefloraison quinconciale" according to Benoist [1967, p. 126] and pollen similar to Lasiocladus according to Muller et al. [1989]), and Pseudodicliptera (also treated as Justicieae by Scotland and Vollesen [2000] but placed by Benoist [1967] in "Tribu des Crabbéés" albeit without mention of aestivation and with at least two distinct pollen types, one suggestive of Ruellia and the other of Lasiocladus [Muller et al., 1989]).
Andrographideae
The sampled members of Andrographideae are monophyletic with strong support; a taxon corresponding to this lineage has long been recognized based on remarkable pollen (Lindau, 1895; Scotland, 1992; Fig. 1A, B) and earlier on other evidence (Bentham and Hooker, 1876). All plants that have been placed in this lineage have pollen with three compound apertures (pororate Diotacanthus, colporate in all others) associated with areas of highly ornamented and thickened exine either at the margin of the compound aperture (most genera; see Fig. 1A, B) or uniformly covering the compound aperture (Diotacanthus) (Scotland, 1992). Unsampled genera that have been treated as Andrographideae are included in Table 1; we predict that all will be part of this lineage based upon pollen morphology. Andrographideae usually have six or more ovules per ovary, which appears to be a secondary increase in ovule number and synapomorphic for the lineage. Among other Acanthoideae, at least two lineages of Ruellieae have likewise undergone secondary increase in number of ovules. Additionally distinguishing Andrographideae from its sister group, Barlerieae, is ascending cochlear aestivation, no doubt symplesiomorphic compared to quincuncial in Barlerieae; this aestivation type also occurs in Justicieae, most Acantheae and a few Thunbergioideae. Also, so far as is known, seeds of Andrographideae lack hygroscopic trichomes.
This lineage has not been included in previous broadly sampled molecular phylogenetic studies of Acanthaceae, nor have relationships among its constituent genera been studied. Species of Indoneesiella have been treated in Andrographis (Nees von Esenbeck, 1847; Cramer, 1996) and as a distinct genus (Sreemadhavan, 1967, 1968). Our results certainly indicate a close relationship between these, but additional sampling is necessary to support one or the other taxonomic treatment. Ridley (1924) pointed to confusion regarding generic assignment of species between Gymnostachyum and Phlogacanthus, but our results do not support a close relationship between these genera (i.e., synonymizing these two genera would likely require treating the entire lineage in a single genus). In describing Gymnostachymn glomeratum (Bl.) Brem., Bremekamp (1948, p. 23) noted the traits that argued against referring the species to other genera of Andrographideae and concluded that Gymnostachyum is "...negatively characterized..." Scotland’s (1992) comprehensive survey of pollen of Andrographideae did not reveal palynological characteristics useful for hypothesizing generic relationships within the tribe. Clearly, additional work is needed on Andrographideae, including denser sampling and assessment of morphological characters that may distinguish genera.
Our results indicate that Barlerieae and Andrographideae are sister taxa. We know of no nonmolecular synapomorphies that support (or refute) this relationship but anticipate that our results will stimulate more research in this phylogenetic neighborhood of Acanthaceae. Notably, the biogeographic complexity of Barlerieae and the fidelity of Andrographideae to Asia (a region that is likely less rich than Africa and the Americas in Acanthaceae) is paralleled by a fivefold difference in species richness (450 species of Barlerieae vs. <80 Andrographideae).
Whitfieldieae
As proposed by Manktelow et al. (2001), our results indicate a close relationship between Whitfieldia and Chlamydacanthus and also place Malagasy Leandriella, Camarotea, and Forcipella in this lineage (Fig. 5). Manktelow et al. (2001) proposed pollen and seed synapomorphies for the lineage, and so far as known, these newly placed Malagasy taxa share these traits. All three have lenticular, biporate pollen with a granular region surrounding the apertures, as do species of Whitfieldia and Chlamydacanthus (Fig. 7C–H). Seeds of Forcipella share the concentric ovals of coarse scales posited as synapomorphic for Whitfieldieae by Manktelow et al. (2001); seeds are apparently unknown for plants of the other two genera. Based presumably on pollen that is superficially similar to the "gürtelpollen" of Isoglossinae (Justicieae), Scotland and Vollesen (2000) placed Forcipella in Justicieae. Indeed, pollen of plants belonging to Whitfieldieae can be remarkably like gürtelpollen, but our results indicate that this similarity is homoplastic.
|
Lankesteria is weakly supported as the basal member of Whitfieldieae, as also posited by Manktelow et al. (2001); notably, our results provide no stronger support for this relationship even though the present analysis included more than twice as many characters. Manktelow et al. (2001) pointed to possible corolla (vascular traces to the lobes trifurcating), pollen (pores surrounded by a circular area with sculpting that is markedly different—usually densely granular—from the remainder of the surface of the grain), and seed (covered with concentric rings of ridges albeit obscured by hygroscopic trichomes in Lankesteria) synapomorphies for Lankesteria + core Whitfieldeae, and we know of no evidence to refute these hypotheses. Still, because of the weak support for placement of Lankesteria, we tested its placement with Ruellieae (as proposed by Lindau, 1895) and with Justicieae (as posited by Bremekamp, 1944), both of which alternative placements were rejected (Table 3). Future research should further test the placement of Lankesteria with additional characters; especially desirable would be additional nuclear data as cp data seem to support the sister relationship to Whitfieldieae consistently but weakly.
Relationships among Whitfieldieae s.l. (i.e., including Lankesteria) indicate dispersal between continental Africa and Madagascar. Lankesteria, with three of ca. eight species sampled, including two African and the only known Malagasy species, is monophyletic. Our data do not resolve relationships among sampled species of Lankesteria with confidence so we can say only that there has been one dispersal event between Africa and Madagascar, direction unknown. Until the ancestral area for Lankesteria can be inferred, we cannot posit an ancestral area for the entire lineage (i.e., Whitfieldieae s.l.) nor for core Whitfieldieae. Thus, among core Whitfieldieae, the MP and Bayesian trees support a single dispersal event between Africa and Madagascar, again direction unknown. However, our data cannot reject monophyly of Chlamydacanthus (i.e., including Malagasy C. euphorbioides), in which case two or more dispersal events might have been involved.
Whitfieldieae s.l. are sister to (Barlerieae + Andrographideae) with support from Bayesian analysis (BPP = 99%) but little support from parsimony. Manktelow et al. (2001) study did not sample Andrographideae: as a result, Whitfieldieae s.l. were sister to Barlerieae in their results. These authors suggested that hygroscopic trichomes (present in Barlerieae and in Lankesteria, lacking in core Whitfieldieae, but the concentric rings of ridges were proposed to constitute evidence of their ancestral presence) and trifurcating traces to the corolla lobes (present in Whitfieldieae s.l. and in at least some Barlerieae) might support a relationship between Barlerieae and Whitfieldieae s.l. The results of the current study intercalate Andrographideae between Barlerieae and Whitfieldieae as sister to the former and thus mandate reexamination of these and other traits. As noted, Andrographideae lack hygroscopic trichomes on the seeds (Bremekamp, 1953), and so far as we are aware, the pattern of vasculature to corolla lobes has not been studied. Corolla-aestivation pattern is homoplasious among these plants because the basal members of Whitfieldieae s.l., Lankesteria and Whitfieldia, have contort aestivation, whereas Chlamydacanthus, Forcipella, and Leandriella have ascending cochlear (aestivation pattern is apparently not known in Camarotea). Thus, as for the strongly supported sister relationship between Barlerieae and Andrographideae, we know of no nonmolecular evidence supporting the relationship of these two to Whitfieldieae s.l. We return to assess the phylogenetic status of hygroscopic trichomes on seeds in the context of support for the entire BAWN lineage.
Neuracanthus
Monophyly of Neuracanthus is strongly supported by our data (Fig. 5). These plants are unique among Acanthaceae in having sepals united in a 3 + 2 pattern, and there has been little doubt about the validity of the group (Bidgood and Brummitt, 1998). We propose this calyx morphology as a synapomorphy for the genus and predict that the species not sampled here are part of this clade based on shared possession of the calyx trait. Corolla aestivation has apparently not been studied in Neuracanthus. Pollen of numerous African and Asian species representing the three sections of Neuracanthus was noted by Furness (1998) to be uniform and characterized as tricolporate with a finely perforate tectum (i.e., the interapertural regions of the exine are foveolate; Fig. 1C, D). Furness (1998) further noted that pollen of Neuracanthus does not resemble that of any of its presumed relatives (all of which are here treated as Barlerieae, Fig. 6). Pollen of Neuracanthus (Fig. 1C, D, which illustrates a Malagasy endemic not studied by Furness, 1998) also differs from that of Andrographideae and Whitfieldieae (Figs. 1 and 7, respectively). We concur with Furness (1998) that Neuracanthus has distinctively relatively featureless pollen that does not provide evidence of relationships of the genus to other Acanthaceae.
The geographic range of Neuracanthus includes continental Africa (18 species with the greatest species richness in eastern and northeastern Africa), the Arabian Peninsula (4 species), Madagascar (6 species), and tropical Asia (India to Vietnam, 4 species). Unfortunately, we were not able to sample any Asian taxa, and whether these four are an Asian clade or represent multiple extensions of the genus into Asia remains to be assessed. The two sampled Malagasy taxa, N. umbraticus and N. brachystachyus, are not sister taxa, but their closer relationships to other taxa are not strongly supported. Thus, although we have not sampled enough species to address biogeographic patterns in detail, it is clear that this relatively small genus (30 species) has had a history of dispersal among Old World regions.
Relationships of Neuracanthus to other Acanthaceae have been enigmatic. Bidgood and Brummitt (1998) discussed earlier attempts to classify the genus but were unable to point to evidence to permit resolution with confidence. In fact, they noted that "it seems idle to speculate on the position of Neuracanthus within the family" (p. 4). Placement of the genus sister to the clade composed of ((Barlerieae + Andrographideae) (Whitfielideae s.l.)) is the most weakly supported aspect of the backbone of acanth relationships, indicating that relationships of these plants are nearly as enigmatic based on molecular data as on structural data. Among traits that have been discussed comparatively for the other lineages of BAWN, hygroscopic trichomes on seeds certainly merit further study. These are present in Neuracanthus, Lankesteria, and Barlerieae. It is possible that this trait is synapomorphic for BAWN, with losses in the common ancestor of core Whitfieldieae and Andrographideae. The first step to assess this hypothesis is to examine trichomes across all three clades for evidence to support or refute homology. Evidence to address the hypothesis that the concentric rings of ridges on the seeds of core "Whitfieldieae" provide evidence of the ancestral presence of trichomes should also be sought: if this is the case, then similar structures should be present on the surface of seeds with trichomes. Once studied in Neuracanthus, aestivation pattern may be informative although given the diversity already known to exist among the other three BAWN clades, this trait is unlikely to provide unambiguous phylogenetic signal. In sum, the results of this study clarify sampling strategies for comparative work on plants in the BAWN clade and may stimulate comparative work on these plants.
In this paper, we have established the phylogenetic placement of Kudoacanthus (Tetramerium lineage, Justicieae); Acanthostelma, Golaea, and Acanthura (Barlerieae); and Camarotea, Forcipella, and Leandriella (core Whitfielidieae). We have confirmed monophyly of Neuracanthus and placed it among the earliest diverging lineages of the cystolith clade, but relationships among these lineages are not strongly supported. In terms of relationships among major lineages, we contribute the unexpected and strongly supported result that Andrographideae are sister to Barlerieae. Other aspects of the newly proposed BAWN clade require further testing, specifically the relationships of Lankesteria to Whitfieldieae, and of Neuracanthus to the other clades, as these aspects are not strongly supported in our results.
These results, especially as informed by nonmolecular characters, permit predictions about four of the five remaining genera that were not classified by Scotland and Vollesen (2000). As already noted, based on aestivation and pollen characters, we infer placement of Pericalypta proximate to Lasiocladus in Barlerieae. Second, based on pollen morphology as reported in Muller et al. (1989) and an androecium of four stamens, we predict that Malagasy Vindasia will be placed among Whitfieldieae. Third, Malagasy Sphacanthus has pollen consistent with either Whitfieldieae or Isoglossinae of Justicieae (see Kiel et al., 2006); its androecium of two stamens suggests placement with the latter. Fourth, the description of the sole species of Vavara, V. breviflora (Benoist, 1962, p. 131), indicates that the two stamens have unequally inserted, muticous anthers and pollen grains that are "9-sulcata," traits strongly suggesting placement in Justicieae: "justicioids" (sensu McDade et al., 2000a). The other unplaced genus, Dolichostachys (one species, Madagascar), remains poorly known and will benefit from additional collections, study of structural characters, and inclusion in studies based on molecular characters.
This study also provides the phylogenetic framework to address the evolution and diversification of Acanthaceae. For example, Acanthaceae are important components of tropical and subtropical communities worldwide and have adapted to a wide range of habitat types, from extremely xeric to per humid. In terms of biotic interactions, so far as is known, all Acanthaceae are animal-pollinated. and many have other ecological interactions with animals (e.g., extrafloral nectaries visited by ants; McDade, 1984; Champluvier, 1994; Gracie, 1991; McDade and Turner, 1997). Schmidt-LeBuhn et al. (2007) suggested a relationship between hummingbird pollination and increased speciation rate, but phylogenetic knowledge of Acanthaceae has been inadequate to test this hypothesis. Our phylogenetic results provide the context in which to assess the role of biotic interactions and habitat specialization in the diversification of Acanthaceae, a process that has rather prolifically yielded at least 4000 species.
Comparative biogeographic studies are also in order. As discussed by McDade et al. (2005), the available evidence and analyses to date indicate that Acanthaceae evolved well after the breakup of Gondwanaland such that dispersal must be posited as the mechanism to explain intercontinental disjunctions. Among lineages of Acanthaceae, the BAWN clade has had a comparatively dynamic history of biogeographic events especially among OW areas. All the component clades (i.e., Barlerieae, Andrographideae, Whitfielidieae, Neuracanthus) are primarily Old World and likely originated there. Only among Barlerieae have there been dispersal events to the New World, and here, our results indicate at least three, one of which is likely recent (i.e., Barleria oenotherioides). This contrasts sharply with biogeographic patterns in other clades of Acanthaceae that have been studied phylogenetically. Most of these clades had an Old World origin followed by a single dispersal event to the NW with subsequent diversification there. This is true of Thunbergioideae (Borg et al., 2008), Acantheae (McDade et al., 2005), Ruellia (Tripp, 2007), and among the Justicieae, Isoglossinae (Kiel et al., 2006), and Tetramerium lineages (Daniel et al., 2008). Among "justicioids" (sensu McDade et al., 2000a), a clade with more than 1000 species, just two dispersal events to the NW (i.e., the common ancestor of NW Dicliptera and of NW "justicioids") have yielded extensive radiations totaling 500 species (L. A. McDade and C. A. Kiel, unpublished data). It is notable that Barlerieae have dispersed to the NW more often but, once there, have diversified less than other acanth lineages. The phylogenetic origins of the floras of specific biogeographic areas (e.g., Australia, Madagascar, New Guinea) also merit attention. Especially when floras of such areas are notably species-rich, the comparative contributions of dispersal vs. in situ speciation must be documented to understand the processes that yield diverse floras.
Finally, the BAWN clade of Acanthaceae requires additional work. It is intriguing that, just as affinities of Lankesteria and Neuracanthus have been enigmatic based on structural characters, placement of these taxa by our molecular data are the most weakly supported aspects of our results. Further, although our data place Andrographideae sister to Barlerieae with strong support, structural synapomorphies for this relationship are not apparent, reflecting the fact that relationships of this lineage have been enigmatic until the current study. Comparative study of structural characters such as those usefully compiled by Scotland et al. (1994) and Scotland and Vollesen (2000), for example, as well as additional molecular data, will be needed to strengthen or refute the relationships proposed here.
Appendix 1. Taxa, GenBank accessions (trnL-F, rps16, trnT-L, trnS-G, nrITS), sources of plant materials from which DNA was extracted for sequencing, and lineage and clade where placed by the analyses presented here (i.e., clade names as depicted in Figs. 1–4 and in the text). Taxa are listed in alphabetical order by genus and species. When plants in cultivation were used, we provide native range in parentheses. Abbreviations for herbaria follow Holmgren and Holmgren (1998). A dash (—) means that the sequence was not obtained.
|
FOOTNOTES
1 The authors thank F. Almeda, K. Balkwill, A. J. Borg, A. M. Boyd, M. Butterwick, M.-J. Cadman, D. J. Hearn, P. Jenkins, J. MacDougal, M. Manktelow, S. Manktelow, M. McMahon, R. Olmstead, H. Ranarivelo, T. Van Devender, K. Vollesen, J. Wood, and the staff of the Jodrell Laboratory (K) for help in obtaining material for DNA or in the field. The curators of the following herbaria graciously granted permission to sample herbarium specimens: ARIZ, BR, CAS, DAV, DS, DUKE, IEB, J, K, NY, P, PH, RSA, UPS, and US. The authors thank the Royal Botanic Gardens, Sydney and the Strybing Arboretum for permission to collect from plants cultivated at these institutions and the San Francisco Conservatory of Flowers for greenhouse space. They are grateful to D. C. Wasshausen for discussion and access to unpublished data on South American Acanthaceae. S. Serata assisted with scanning electron microscopy for palynological studies. The manuscript benefited from the input of M. Simmons and two anonymous reviewers, and the editorial assistance of L. Worlow. This work was supported by the U. S. National Science Foundation (DEB 0108589 to L.A.M. and T.F.D.), the Academy of Natural Sciences, Rancho Santa Ana Botanic Garden, California Academy of Sciences, and the Christensen Research Institute. 
4 Author for correspondence (e-mail: lucinda.mcdade{at}cgu.edu)
LITERATURE CITED
Balkwill, M.-J., AND K. Balkwill. 1997. Delimitation and infra-generic classification of Barleria (Acanthaceae). Kew Bulletin 52: 535–573.[CrossRef]
Balkwill, M.-J., AND K. Balkwill. 1998. A preliminary analysis of distribution patterns in a large, pantropical genus, Barleria L. (Acanthaceae). Journal of Biogeography 25: 95–110.[CrossRef][Web of Science]
Balkwill, K., AND F. Getliffe Norris. 1988. Classification of the Acanthaceae: A southern African perspective. Monographs in Systematic Botany from the Missouri Botanical Garden 25: 503–516.
Bennett, J. R., AND R. W. Scotland. 2003. A revision of Strobilanthes (Acanthaceae) in Java. Kew Bulletin 58: 1–82.[CrossRef]
Benoist, R. 1911. Les genres Lepidagathis et Lophostachys sont-ils distincts? Notulae Systematicae 2: 139–144.
Benoist, R. 1962. Nouvelles Acanthacées de Madagascar. Bulletin de la Société Botanique de France 109: 129–135.
Benoist, R. 1967. Contribution à la connaissance des Acanthacées malgaches. Bulletin de la Société Botanique de France 113: 531–536.
Bentham, G., AND J. D. Hooker. 1876. Genera plantarum, vol. 2. L. Reeve, London, UK.
Bidgood, S., AND R. K. Brummitt. 1998. A revision of the genus Neuracanthus (Acanthaceae). Kew Bulletin 53: 1–76.[CrossRef]
Borg, A. J., L. A. McDade, AND J. Schönenberger. 2008. Molecular phylogenetics and morphological evolution of Thunbergioideae (Acanthaceae). Taxon.
Bremekamp, C. E. B. 1944. Materials for a monograph of the Strobilanthinae (Acanthaceae). Verhandelingen der Nederlandsche Akademie van Wetenschappen, Afdeeling Natuurkunde. Tweede Sectie 41: 1–306.
Bremekamp, C. E. B. 1948. Notes on the Acanthaceae of Java. Verhandelingen der Koninklijke Nederlandsche Akademie van Wetenschappen, Afdeeling Natuurkunde. Tweede Sectie 45: 1–78.
Bremekamp, C. E. B. 1953. The delimitation of the Acanthaceae. Proceedings, Koninklijke Nederlandse Akademie van Wetenschappen, series C, Biological and Medical Sciences, Amsterdam 56: 533–546.
Bremekamp, C. E. B. 1965. Delimitation and subdivision of the Acanthaceae. Bulletin of the Botanical Survey of India 7: 21–30.
Bremer, K. 1988. The limits of amino acid sequence data in angiosperm phylogenetic reconstructions. Evolution 42: 795–803.[CrossRef][Web of Science]
Carine, M. A., AND R. W. Scotland. 1998. Pollen morphology of Strobilanthes Blume (Acanthaceae) from southern India and Sri Lanka. Review of Palaeobotany and Palynology 103: 143–165.[CrossRef][Web of Science]
Carine, M. A., AND R. W. Scotland. 2000. Taxonomy and biology of Stenosiphonium Nees (Acanthaceae). Botanical Journal of the Linnean Society 133: 101–128.[CrossRef][Web of Science]
Champluvier, D. 1994. Brachystephanus myrmecophilus (Acanthaceae), espèce nouvelle du Zaïre oriental: Un cas intéressant de myrmécophile. Belgian Journal of Botany 127: 45–60.
Chiovenda, E. 1929. Flora Somalia. Sindacato Italiano Arti Grafiche, Rome, Italy.
Cramer, L. H. 1996. Notes on Sri Lankan Acanthaceae. Kew Bulletin 51: 553–556.[CrossRef]
Cronquist, A. 1981. An integrated system of classification of flowering plants. Columbia University Press, New York, New York, USA.
Daniel, T. F. 1995. New and reconsidered Mexican Acanthaceae. VI. Chiapas. Proceedings of the California Academy of Sciences 48: 253–282.
Daniel, T. F. 1998. Pollen morphology of Mexican Acanthaceae: Diversity and systematic significance. Proceedings of the California Academy of Sciences 50: 217–256.
Daniel, T. F., L. A. McDade, M. Manktelow, AND C. A. Kiel. 2008. The "Tetramerium lineage" (Acanthaceae: Acanthoideae: Justicieae): Delimitation and intra-lineage relationships based on cp and nrITS sequence data. Systematic Botany 33: 416–436.[CrossRef][Web of Science]
Donoghue, M. J., R. G. Olmstead, J. F. Smith, AND J. D. Palmer. 1992. Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanical Garden 79: 333–345.[CrossRef][Web of Science]
Doyle, J. J., AND J. L. Doyle. 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Durkee, L. H. 2001. Acanthaceae. In W. D. Stevens, C. U. Ulloa, A. Pool, and O. M. Montiel [eds.], Flora de Nicaragua, vol. 1. Monographs in Systematic Botany from the Missouri Botanical Garden 85: 8–36.
Farris, J. S., M. Källersjö, A. G. Kluge, AND C. Bult. 1994. Testing significance of incongruence. Cladistics 10: 315–319.[CrossRef][Web of Science]
Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791.[CrossRef][Web of Science]
Felsenstein, J. 2004. Inferring phylogenies. Sinauer, Sunderland, Massachusetts, USA.
Furness, C. A. 1998. Pollen morphology of Neuracanthus (Acanthaceae). Kew Bulletin 53: 77–81.[CrossRef]
Gracie, C. 1991. Observation of dual function of nectaries in Ruellia radicans (Nees) Lindau (Acanthaceae). Bulletin of the Torrey Botanical Club 118: 188–190.[CrossRef][Web of Science]
Graham, S. W., P. A. Reeves, A. C. E. Burns, AND R. G. Olmstead. 2000. Microstructural changes in noncoding chloroplast DNA: Interpretation, evolution, and utility of indels and inversions in basal Angiosperm phylogenetic inference. International Journal of Plant Sciences 161 (6 Supplement):S83–S96.[CrossRef][Web of Science]
Griekspoor, A., AND T. Groothuis. 2006.4Peaks, version 1.7.2 (1.7.1). Website http://mekentosj.com/4peaks/[accessed May 2008]
Grubert, M. 1974. Studies on the distribution of myxospermy among seeds and fruits of Angiospermae and its ecological importance. Acta Biológica Venezuelica 8: 315–551.
Hedrén, M., M. W. Chase, AND R. G. Olmstead. 1995. Relationships in the Acanthaceae and related families as suggested by cladistic analysis of rbcL nucleotide sequences. Plant Systematics and Evolution 194: 93–109.[CrossRef][Web of Science]
Holmgren, P. K., AND N. H. Holmgren. 1998[continuously updated]. Index Herbariorum: A global directory of public herbaria and associated staff. New York Botanical Garden’s Virtual Herbarium. Website http://sweetgum.nybg.org/ih/[accessed May 2008].
Hu, J.-Q. 2002. Acanthaceae. In Z. Wu, P. H. Raven, and D. Y. Hong [eds.], Flora of China, vol. 19. Science Press, Beijing, China, and Missouri Botanical Garden Press, St. Louis, Missouri, USA.
Kelchner, S. A., AND L. G. Clark. 1997. Molecular evolution and phylogenetic utility of the chloroplast rpl16 intron in Chusquea and the Bambusoideae (Poaceae). Molecular Phylogenetics and Evolution 8: 385–397.[CrossRef][Web of Science][Medline]
Kiel, C. A., L. A. McDade, T. F. Daniel, AND D. Champluvier. 2006. Phylogenetic delimitation of Isoglossinae (Acanthaceae: Justicieae) and relationships among constituent genera. Taxon 55: 683–694.[Web of Science]
Kishino, H., AND M. Hasegawa. 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in Hominoidea. Journal of Molecular Evolution 29: 170–179.[CrossRef][Web of Science][Medline]
Lindau, G. 1895. Acanthaceae. In A. Engler, and K. Prantl [eds.], Die natürlichen Pflanzenfamilien 4(3b), 274–353. W. Engelmann, Leipzig, Germany.
Maddison, D. R. 1991. The discovery and importance of multiple islands of most parsimonious trees. Systematic Zoology 40: 315–328.[Abstract]
Maddison, W. P., AND D. R. Maddison. 2000. MacClade: Analysis of phylogeny and character evolution, version 4. 03. Sinauer, Sunderland, Massachusetts, USA.
Manktelow, M. 1996. Phaulopsis (Acanthaceae): A monograph. Symbolae Botanicae Upsalienses 31: 1–184.
Manktelow, M. 2000. The filament curtain: A structure important to systematics and pollination biology in the Acanthaceae. Botanical Journal of the Linnean Society 133: 129–160.[CrossRef][Web of Science]
Manktelow, M., L. A. McDade, B. Oxelman, C. A. Furness, AND M.-J. Balkwill. 2001. The enigmatic tribe Whitfieldieae (Acanthaceae): Delimitation and phylogenetic relationships based on molecular and morphological data. Systematic Botany 26: 104–119.[Web of Science]
McDade, L. A. 1984. Systematics and reproductive biology of the Central American members of the Aphelandra pulcherrima complex (Acanthaceae). Annals of the Missouri Botanical Garden 71: 104–165.[CrossRef][Web of Science]
McDade, L. A., T. F. Daniel, C. A. Kiel, AND K. Vollesen. 2005. Phylogenetic relationships among Acantheae (Acanthaceae): Major lineages present contrasting patterns of molecular evolution and morphological differentiation. Systematic Botany 30: 834–862.[CrossRef][Web of Science]
McDade, L. A., T. F. Daniel, S. E. Masta, AND K. M. Riley. 2000a. Phylogenetic relationships within the tribe Justicieae (Acanthaceae): Evidence from molecular sequences, morphology, and cytology. Annals of the Missouri Botanical Garden 87: 435–458.[CrossRef][Web of Science]
McDade, L. A., S. E. Masta, M. L. Moody, AND E. Waters. 2000b. Phylogenetic relationships among Acanthaceae: Evidence from two genomes. Systematic Botany 25: 106–121.[CrossRef][Web of Science]
McDade, L. A., AND M. L. Moody. 1999. Phylogenetic relationships among Acanthaceae: Evidence from noncoding trnL-trnF chloroplast DNA sequences. American Journal of Botany 86: 70–80.
McDade, L. A., AND M. D. Turner. 1997. Extrafloral nectaries in Aphelandra: Anatomy, development and systematic implications. American Journal of Botany 84: 1–15.[Abstract]
Morrison, D. A. 2006. Multiple sequence alignment for phylogenetic purposes. Australian Systematic Botany 19: 479–539.[CrossRef][Web of Science]
Moylan, E. C., J. R. Bennett, M. A. Carine, R. G. Olmstead, AND R. W. Scotland. 2004a. Phylogenetic relationships among Strobilanthes s.l. (Acanthaceae): Evidence from ITS nrDNA, trnL-F cpDNA, and morphology. American Journal of Botany 91: 724–735.
Moylan, E. C., P. J. Rudall, AND R. W. Scotland. 2004b. Comparative floral anatomy of Strobilanthinae (Acanthaceae), with particular reference to internal partitioning of the flower. Plant Systematics and Evolution 249: 77–98.[CrossRef][Web of Science]
Muller, J., M. Schuller, H. Straka, AND B. Friedrich. 1989. Palynologia Madagassica et Mascarenica. Familien 182: Acanthaceae. Tropische und Subtropische Pflanzenwelt 67: 138–187.
Nees von Esenbeck, C. G. 1847. Acanthaceae. In A. de Candolle [ed.], Prodromus systematis naturalis regni vegetabilis, 46–519. Victoris Masson, Paris, France.
Nylander, J. A. A. 2004. MrModeltest, version 2. Program distributed by the author. Evolutionary Biology Centre, Uppsala University, Sweden.
Oxelman, B., M. Backlund, AND B. Bremer. 1999. Relationships of the Buddlejaceae s.l. investigated using parsimony jackknife and branch support analysis of chloroplast ndhF and rbcL sequence data. Systematic Botany 24: 164–182.[CrossRef][Web of Science]
Oxelman, B., P. Kornhall, R. G. Olmstead, AND B. Bremer. 2005. Further disintegration of Scrophulariaceae. Taxon 54: 411–425.[Web of Science]
Raj, B. 1961. Pollen morphological studies in the Acanthaceae. Grana Palynologica 3: 3–108.[Medline]
Rambaut, A. 1996–2002. Se-Al. Sequence alignment editor, version 2.0a11. Website http://evolve.zoo.ox.ac.uk/software.html?name=Se-Al [accessed Dec 2007]
Reeves, J. H. 1992. Heterogeneity in the substitution process of amino acid sites of proteins coded for by mitochondrial DNA. Journal of Molecular Evolution 35: 17–31.[CrossRef][Web of Science][Medline]
Ridley, H. N. 1924. The flora of the Malay Peninsula. IV. L. Reeve and Co. London, UK.
Ronquist, F., AND J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (Oxford, England) 19: 1572–1574.[CrossRef]
Schmidt-LeBuhn, A. N. 2003. A taxonomic revision of the genus Suessenguthia Merxm. (Acanthaceae). Candollea 58: 101–128.[Web of Science]
Schmidt-LeBuhn, A. N., M. Kessler, AND I. Hensen. 2007. Hummingbirds as drivers of plant speciation? Trends in Plant Science 12: 329–331.[CrossRef][Web of Science][Medline]
Schmidt-LeBuhn, A. N., M. Kessler, AND J. Müller. 2005. Evolution of Suessenguthia (Acanthaceae) inferred from morphology, AFLP data and ITS rDNA sequences. Organisms, Diversity & Evolution 5: 1–13.
Schönenberger, J. 1999. Floral structure, development and diversity in Thunbergia (Acanthaceae). Botanical Journal of the Linnean Society 130: 1–36.[Web of Science]
Schönenberger, J., AND P. K. Endress. 1998. Structure and development of the flowers in Mendoncia, Pseudocalyx, and Thunbergia (Acanthaceae) and their systematic implications. International Journal of Plant Sciences 159: 446–465.[CrossRef][Web of Science]
Schwarzbach, A. E., AND L. A. McDade. 2002. Phylogenetic relationships of the mangrove family Avicenniaceae based on chloroplast and nuclear ribosomal DNA sequences. Systematic Botany 27: 84–98.[Web of Science]
Scotland, R. W. 1992. Pollen morphology of Andrographideae (Acanthaceae). Review of Palaeobotany and Palynology 72: 229–243.[CrossRef][Web of Science]
Scotland, R. W. 1993. Pollen morphology of Contortae (Acanthaceae). Botanical Journal of the Linnean Society 111: 471–504.[Web of Science]
Scotland, R. W. 1998. One new and one rediscovered species of Strobilanthes Blume (Acanthaceae). Botanical Journal of the Linnean Society 128: 203–210.[Web of Science]
Scotland, R. W., P. K. Endress, AND T. J. Lawrence. 1994. Corolla ontogeny and aestivation in the Acanthaceae. Botanical Journal of the Linnean Society 114: 49–65.[CrossRef][Web of Science]
Scotland, R. W., J. A. Sweere, P. A. Reeves, AND R. G. Olmstead. 1995. Higher-level systematics of Acanthaceae determined by chloroplast DNA sequences. American Journal of Botany 82: 266–275.[CrossRef][Web of Science]
Scotland, R. W., AND K. Vollesen. 2000. Classification of Acanthaceae. Kew Bulletin 55: 513–589.[CrossRef]
Shaw, J., E. B. Lickey, E. E. Shilling, AND R. L. Small. 2007. Comparison of whole chloroplast genome sequences to choose noncoding regions for phylogenetic studies in angiosperms: The tortoise and the hare III. American Journal of Botany 94: 275–288.
Sreemadhavan, C. P. 1967. Neesiella— A new genus of Acanthaceae. Phytologia 15: 270–271.
Sreemadhavan, C. P. 1968. Indoneesiella—A substitute name in Acanthaceae. Phytologia 16: 466.
Swofford, D. L. 2000. PAUP*: Phylogenetic analysis using parsimony (*and other methods), version 4.0. Sinauer, Sunderland, Massachusetts, USA.
Tank, D. C., P. M. Beardsley, S. A. Kelchner, AND R. G. Olmstead. 2006. Review of the systematics of Scrophulariaceae s.l. and their current disposition. Australian Systematic Botany 19: 289–307.[CrossRef][Web of Science]
Tavaré, S. 1986. Some probabilistic and statistical problems on the analysis of DNA sequences. Lectures on Mathematics in the Life Sciences 17: 57–86.
Templeton, A. R. 1983. Phylogenetic inference from restriction endonuclease clevage site maps with particular reference to the humans and apes. Evolution 37: 221–244.[CrossRef][Web of Science]
Thulin, M. [ed.]. 2006. Flora of Somalia, vol. 3. Royal Botanic Gardens, Kew, Richmond, Surrey, UK.
Tripp, E. A. 2007. Evolutionary relationships within the species-rich genus Ruellia (Acanthaceae). Systematic Botany 32: 628–649.[CrossRef][Web of Science]
Tripp, E. A., T. F. Daniel, J. C. Lendemer, AND L. A. McDade. In press. New molecular and morphological insights prompt transfer of Blechum to Ruellia (Acanthaceae). Taxon.
Wagstaff, S. J., AND R. G. Olmstead. 1997. Phylogeny of Labiatae and Verbenaceae inferred from rbcL sequences. Systematic Botany 22: 165–179.[CrossRef][Web of Science]
Wasshausen, D. C. 1996. New species and new combinations in Aphelandra (Acanthaceae) from Ecuador and adjacent Peru. Nordic Journal of Botany 16: 389–407.[CrossRef][Web of Science]
Wasshausen, D. C., AND J. R. I. Wood. 2004. Acanthaceae of Bolivia. Contributions from the United States National Herbarium 49: 1–152.
Witztum, A., AND K. Schulgasser. 1995. The mechanics of seed expulsion in Acanthaceae. Journal of Theoretical Biology 176: 531–542.[CrossRef][Web of Science]
Wortley, A. H., D. J. Harris, AND R. W. Scotland. 2007. On the taxonomy and phylogenetic position of Thomandersia. Systematic Botany 32: 415–444.[CrossRef][Web of Science]
Wortley, A. H., P. J. Rudall, D. J. Harris, AND R. W. Scotland. 2005. How much data are needed to resolve a difficult phylogeny? Case study in Lamiales. Systematic Biology 54: 697–709.
Yang, Z. 1993. Maximum-likelihood estimation of phylogeny from DNA sequences when substitution rates differ over sites. Molecular Biology and Evolution 10: 1396–1401.[Abstract]
Yang, Z., AND B. Rannala. 1997. Bayesian phylogenetic inference using DNA sequences: A Markov chain Monte Carlo method. Molecular Biology and Evolution 14: 717–724.[Abstract]
CiteULike Complore Connotea Del.icio.us Digg Facebook Reddit Technorati Twitter What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
