| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Departments of Ecology and Evolutionary Biology, and Plant Sciences, University of Arizona, Tucson, Arizona 85721
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Acanthaceae • chloroplast DNA • Lamiales • Mendoncioideae • Nelsonioideae • phylogenetics • Thunbergioideae • trnL-trnF
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The precise delimitation of the family, however, has been controversial due to three small lineages that do not share these fruit characters but seem clearly allied with Acanthaceae s.s. These are Mendoncia (~60 spp.), Thunbergia (~100 spp.), and the Nelsonioideae (Nelsonia, Elytraria, and Staurogyne, together with ~100 species). Treatment of these groups has varied among students of Acanthaceae and in the synoptic works on flowering plants [e.g., Cronquist (1981) recognizes Mendonciaceae and treats the others as part of Acanthaceae, whereas Thorne (1992) treats these three lineages as subfamilies of a broader Acanthaceae].
Knowledge of suprageneric relationships within Acanthaceae benefits immensely from the work of several earlier botanists, in particular Bentham and Hooker (1876), Lindau (1895), Van Tieghem (1908), and Bremekamp (1955, 1965). These authors of course relied on morphological data and did not analyze their data phylogenetically. In these earlier works, there is also a tendency to rely on a few character systems (especially pollen). Despite such limitations, these workers have left an insightful legacy of data and ideas regarding classification of the group (see Hedrén, Chase, and Olmstead, 1995, and Scotland et al., 1995 for useful overviews). Balkwill and Getliff Norris (1988) provide an information-rich, regional (i.e., southern Africa) examination of acanth classification using morphological data. More recently, a few workers (Scotland, 1993; Hedrén, Chase, and Olmstead, 1995; Scotland et al., 1995) have examined relationships within an explicit phylogenetic framework.
There is some agreement among these recent results and the classifications presented by earlier authors, but also discord at a number of levels, and many aspects of acanth relationships remain unexamined. Delimitation of the family with regard to the three near outgroup lineages discussed above remains at issue. Within Acanthaceae s.s., the same major lineages are recognized by most workers albeit at widely varying taxonomic levels. For example, Bentham and Hooker (1876) recognized tribes Ruellieae, Justicieae, and Acantheae; Bremekamp (1965) also recognized these lineages but united the first two as subfamily Ruellioideae. Moreover, relationships among acanths have been addressed only at the highest levels [i.e., among the three or four most inclusive lineages (Hedrén, Chase, and Olmstead, 1995; Scotland et al., 1995)]. Within these major lineages, there is limited accord among existing classifications as to membership, and explicitly phylogenetic work is essentially nonexistent. This is especially true in the case of the largest and arguably most difficult group, Bremekamp's (1965) Justicieae. Bremekamp (1965) noted that his own classification of this group was unsatisfactory and should be viewed as preliminary.
We are studying phylogenetic relationships among Acanthaceae using data from morphology and nuclear and chloroplast DNA (cpDNA) sequences. These data sets will be analyzed separately and, ultimately, in conjunction. Here we report the results based on cp data, which are sequences of the noncoding regions flanking the trnL and trnF genes [see Taberlet et al. (1991) for a diagram of this region]. This research was undertaken to address four primary goals: (1) explore the utility of this region of cpDNA for phylogenetic work in Acanthaceae; (2) establish hypotheses of phylogenetic relationships among the three lineages hypothesized to be the closest relatives of Acanthaceae s.s.; (3) test the monophyly of major lineages within Acanthaceae s.s. proposed by other workers; and (4) begin to elucidate patterns of relationship within these lineages, with an emphasis on the large lineage that includes Justicia and on that which includes Aphelandra.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Fresh materials were available for all but five taxa, as indicated in the Appendix. Total genomic DNA was extracted using the modified CTAB (hexadecyltrimethylammonium bromide) method of Doyle and Doyle (1987). A fragment comprising the trnL intron, the 3' trnL exon, and the intergenic spacer between this exon and the trnF gene of the chloroplast genome (Taberlet et al., 1991) was amplified using the "c" and "f" primers designed by these same authors. Standard PCR (polymerase chain reaction) techniques were used to amplify double-stranded DNA. Sequences were generated on an ABI automated sequencer at the University of Arizona DNA sequencing facility using the same primers as in amplification. For most samples (i.e., PCR products), sequencing with one primer yielded sequence for virtually the entire fragment. However, the reverse strand was sequenced for about two-thirds of the taxa to complete the sequence and to verify the other strand.
Sequences were aligned by eye in SeqApp (Gilbert, 1992). As noted by Taberlet et al. (1991) and confirmed by others for a few plant groups (e.g., Gielly et al., 1996; Kim, 't Hart, and Mes, 1996), these sequences have a relatively high frequency of indels (see below, Table 1). Information on parsimony informative indels (i.e., those shared by two or more taxa) was added to the data matrix as presence/absence characters. Apparent short indels that were immediately adjacent to a poly-n string (e.g., a series of five Gs) were not included because experience suggests that such strings may be counted incorrectly.
|
The entire data set was analyzed with and without indels. Because including both the regions of sequence in which indels occur and the presence/absence of the same indels potentially gives added weight to these regions, the data were also analyzed with the indels but without the corresponding sequence regions. The impact of indels was further investigated by assigning them a weight of five, which is the empirically determined frequency of indels compared to substitutions (i.e., base substitutions are five times more frequent in this data set than indels). The intron and spacer regions were also analyzed separately, both with and without indels.
In addition to standard measures of fit of characters to the trees produced [i.e., consistency index (CI), retention index (RI), rescaled consistency index (RC)], the strength of support for individual branches was estimated using bootstrap values (Felsenstein, 1985) and decay indices (Bremer, 1988; Donoghue et al., 1992). Bootstrap (BS) values reported are those from the most rigorous searches that could be completed for each analysis without exceeding computer memory. To determine the impact of different search protocols on bootstrap values, we conducted an experiment using a data set from which four taxa that were always involved in polytomies were pruned (i.e., Megaskepasma, Poikilacanthus, Barleria repens, Aphelandra boyacensis). This permitted swapping to completion of bootstrap runs involving 200 "full heuristic" replicates with 20 random addition sequences, each using all four swapping options (i.e., none, NNI, SPR, and TBR); for comparison, bootstrap values from 2000 "fast stepwise addition" replicates were also obtained. The sets of bootstrap values were compared to each other, on a branch-by-branch basis, using correlation.
Decay values for each branch were determined by first using MacClade (Maddison and Maddison, 1992) to prepare a set of trees each with a single branch resolved. These trees were then loaded into PAUP as constraint trees, and the program was asked to find the shortest trees inconsistent with the constraint tree. The difference between the length of these trees and the globally shortest trees is the decay index (DI) for the branch in question.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Figure 1 presents the strict consensus of most parsimonious (MP) trees produced by the total evidence analysis (i.e., sequence data plus unweighted indels); one randomly chosen MP tree (Fig. 2) illustrates branch lengths. In terms of topology, analyses of only the intron or spacer data, as well as those treating indels in different ways (i.e., excluded, sequence regions with indels excluded, weight = 5) gave remarkably congruent results. Differences will be summarized below after presenting results from the total evidence analysis.
|
Acanthaceae s.s.: monophyly and major lineages
Acanthaceae s.s. are not supported as monophyletic by these data. However, the data do not support an alternative hypothesis but rather are unable to resolve the relationship. Within Acanthaceae s.s., there is strong support for four major lineages (Fig. 1), referred to here by informal names to avoid confusion with existing names: Acanthus lineage (BS = 100, DI = 17), Barleria lineage (including Lepidagathis: BS = 100, DI = 13), Ruellia lineage (BS = 100, DI = 26), and Justicia lineage (BS = 100, DI = 15). Among these, there is support for monophyly of the last three lineages together (BS = 100, DI = 10). There is consistent but weaker support for a sister-group relationship between the Ruellia and Justicia lineages (BS = 73, DI = 1), with the Barleria lineage as their sister group.
Relationships within major lineages
Within the Acanthus lineage, there is support for two monophyletic sublineages. One of these includes members of Aphelandreae [sensu Bremekamp (1965), Appendix; here represented by Stenandrium and four species of Aphelandra]. The other is Acantheae [sensu Bremekamp (1965), Appendix; two species each of Acanthus and Crossandra]. Monophyly of the genera for which more than one species was sampled is supported, but these data do not fully resolve relationships among the four species sampled from the large (~200 species) genus Aphelandra. There was, however, moderately strong support (BS = 86, DI = 2) for monophyly of the A. pulcherrima complex [a morphologically well-marked group with a total of ~35 species (McDade, 1984), here represented by A. campanensis and A. leonardii]. Branch lengths (Fig. 2) indicate little variation among the Aphelandra sequences.
Our data strongly support inclusion of Lepidagathis with the three Barleria species. Barleria is also supported as monophyletic. Within the Ruellia lineage, there is strong support for a close relationship between Blechum and Ruellia (BS = 98, DI = 4), but the relationships of the other three taxa are not resolved.
The Justicia lineage comprises all members of Bremekamp's (1965) Justicieae that were included here. Within this group, there is strong support for a number of monophyletic sublineages: Odontonema + Pseuderanthemum (BS = 100, DI = 8), Razisea + Stenostephanus (BS = 100, DI = 16), the Henrya sublineage (BS = 100, DI = 14), and Dicliptera + Peristrophe (BS = 100, DI = 21). In this phylogenetic context, the basal position of Odontonema + Pseuderanthemum is strongly supported (BS = 100, DI = 15), as is the monophyly of the remaining Justicieae (BS = 100, DI = 10). Dicliptera + Peristrophe are strongly supported as sister to an unresolved, weakly supported group representing Bremekamp's subtribe Justiciinae [i.e., the genus Justicia sensu Graham (1988) and its closest relatives] (BS = 100, DI = 24). The sister-group relationship between Justicia-Dicliptera and the Henrya sublineages is also supported (BS = 96, DI = 3). Within the Henrya lineage, there is strong support for monophyly of Anisacanthus, Tetramerium, and Carlowrightia (BS = 97, DI = 6); notably, these share a 149-bp deletion in the spacer. Pachystachys and Henrya are basal to these three, but their relationships are not resolved.
Comparison of results from other data sets
Analyses of subsets of the data (i.e., the intron or spacer), as well as those that differed in handling of indels, were remarkably congruent with Fig. 1. Analyses involving subsets of the data generally resulted in less resolution. Analyses of the intron region, in particular, yielded far less resolution than presented in Fig. 1 (see polytomies indicated). Analysis of the spacer region alone yielded results that were identical to Fig. 1 except that Justiciinae were not resolved as monophyletic. In two instances, analyses of subsets of the data yielded more resolution than the entire data set. First, data from the spacer placed Mendoncia + Thunbergia as sister to the Acanthus lineage but with weak support (BS = 55, DI = 1). Second, both subsets of the data resolved the position of Elytraria: the intron places Elytraria basal to Martynia and Sesamum, whereas the spacer places this taxon above Martynia and Sesamum, as sister to the other Acanthaceae. However, support for both relationships is very weak (BS < 50, DI = 1 for both).
Analysis of the sequence data alone (i.e., omitting the presence/absence indel characters) yielded results identical to Fig. 1. Giving the presence/absence indel characters a weight of five yielded 756 MP trees (length = 1105, CI = 0.744, RI = 0.894, RC = 0.665). Topologically, the strict consensus of these differs from Fig. 1 in only three ways: (1) the Ruellia lineage is resolved, with moderate support for Sanchezia as the sister group to the other four taxa (BS = 79, DI = 6), and (2) Mendoncia + Thunbergia are weakly supported as sister to the Acanthus lineage (BS = 60, DI = 1), and (3) there is very weak support for monophyly of Elytraria, Sesamum, and Martynia (BS = <50, DI = 1). Decay support for a number of branches also increased markedly as a result of giving indels a weight of five; those for which values increased by at least 50% are marked with an asterisk in Fig. 1. These may be interpreted as branches for which there is significant support from indels.
Interestingly, analysis of the data set that included the indels as presence/absence characters but excluded the regions of sequence with indels yielded topological differences in the same three areas enumerated just above. Regarding the Ruellia lineage and Mendoncia + Thunbergia, results were identical to those from the indel weighting experiment. Regarding Elytraria, results from this analysis (i.e., with indels, excluding regions of sequence with indels) placed this taxon as sister to the other Acanthaceae s.l. but with very weak support (BS < 50, DI = 1).
In sum, results from both different subsets of the data and from weighting experiments differed from the total evidence analysis (i.e., Fig. 1) almost exclusively by replacing weakly supported or unresolved portions of the phylogeny with unresolved or weakly supported alternatives. In addition, as suggested by Table 1, the intron is not sufficiently variable to resolve relationships within several lineages. The only positive (i.e., more resolved and strongly supported) difference is in the Ruellia lineage: indels offer moderate support for a lineage exclusive of Sanchezia.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The relatively high frequency of length mutations, especially in the trnL-trnF spacer, has also been previously reported (Taberlet et al., 1991; Kita, Ueda, and Kadota, 1995; Kim, 't Hart, and Mes, 1996; Mes, van Brederode, and 't Hart, 1996). Our results suggest that length mutations in the spacer are more homoplasious than those in the intron (i.e., 11 of 27 indels in the spacer evolved homoplasiously vs. six of 24 in the intron). The observation that almost all length mutations in the spacer (13 of 16 with CI of 1) are deletions, whereas those in the intron are more balanced between insertions and deletions, suggests that different evolutionary processes may be operating in these regions. However, this may not be general: Kim, 't Hart, and Mes (1996) and Mes, van Brederode, and 't Hart (1996) reported more balanced ratios of insertions to deletions in the spacer region of two groups of Crassulaceae.
Parsimony-informative variation (both substitutions and length mutations) in the trnL-trnF region (this study) is compared to that found for other chloroplast loci in Table 2 [note that the data from rbcL relies extensively on Hedrén, Chase, and Olmstead (1995)]. Some caution is warranted in drawing conclusions from these data: although the studies included much the same overall phylogenetic range of taxa, our work included more than twice as many acanth taxa. This comparison suggests that the trnL-trnF region as a whole is comparable to ndhF in terms of rate of parsimony informative substitutions per aligned nucleotide, with differences between the spacer and intron as expected from Table 1. Combined, these regions provide more than twice the rate of parsimony informative substitutions as does rbcL. In terms of informative length mutations, the trnL-trnF region far exceeds both other loci. On the whole, this region offers more phylogenetically informative variation per nucleotide sequenced than the other two chloroplast loci for which data are available. The overall low level of homoplasy (i.e., high CIs, Table 2) in these data compared to the ndhF and rbcL data should also be noted, although, again, these data are not strictly comparable due to differences in sampling. Still, this region is both variable and relatively free from homoplasy over the phylogenetic range sampled here.
|
Relationships of near outgroups
Our results do not permit unambiguous placement of Nelsonioideae (represented by Elytraria) relative to other Acanthaceae s.l. Broader sampling of Nelsonioideae and of its potential relatives (including taxa usually placed in Scrophulariaceae) may help. However, the very short branch lengths among Elytraria, Sesamum, and Martynia (Fig. 2) suggest that the trnL-trnF region may not provide variation useful for resolving relationships at this level. This is supported by the fact that relationships among these three taxa were weakly resolved in nearly every possible topology that resulted from analyses of subsets of the data and from those handling indels in different ways. Scotland et al. (1995) and Hedrén, Chase, and Olmstead (1995), also working with sequences from the chloroplast genome, included broader representation of potential nonacanth relatives of Nelsonioideae. Their results place this lineage as sister to other Acanthaceae s.l.
Mendoncia and Thunbergia are strongly supported as sister taxa by our data and also by morphological data: these plants share flowers in fascicles in leaf axils, an "epicalyx" of paired bracts subtending the flowers, highly modified rudimentary calyx, and sprawling or climbing habit (Bremekamp, 1965; L. A. McDade, personal observation). That these taxa have a closer relationship to Acanthaceae s.s. than Nelsonioideae is supported by shared loss of endosperm and reduction in number of ovules in all of these plants compared to Nelsonioideae. Mendoncia was not included in the work of Hedrén, Chase, and Olmstead (1995) or of Scotland et al. (1995), but both placed Thunbergia as either basal to a monophyletic Acanthaceae s.s. (ndhF) or as part of a polytomy with lineages belonging to Acanthaceae s.s. (rbcL).
Monophyly of Acanthaceae s.s.
As previously described, all Acanthaceae s.s. share at least one unique and unreversed morphological synapomorphy: retinaculae subtending the seeds. The explosively dehiscent capsules, although not globally unique to Acanthaceae, are also not shared by close relatives. Our data do not support monophyly of Acanthaceae s.s. It should be noted, however, that these data do not contradict this relationship (i.e., the problem is one of resolution rather than of support for an alternative hypothesis). In fact, the addition of just one character with states distributed to represent presence/absence of retinaculae yields a monophyletic Acanthaceae s.s. with Mendoncia + Thunbergia as the sister group in trees that are only one step (0.1%) longer than the MP trees presented in Fig. 1 (analysis not shown).
Identity and relationships of major lineages within Acanthaceae s.s.
The existence of four major monophyletic lineages of Acanthaceae s.s. is in accord with previous phylogenetic work (Scotland et al., 1995) and is in large part supported by morphological evidence. Members of the Acanthus lineage have four monothecous stamens. The Barleria and Ruellia lineages are marked by patterns of corolla aestivation unique in the family: quincuncial and left contort, respectively (Scotland, Endress, and Lawrence, 1994; Scotland et al., 1995). Taxonomic and morphological diversity within the very large and diverse Justicia lineage have yet to be thoroughly explored, and it is difficult to identify morphological synapomorphies with confidence. However, the taxa included in this study are marked by reduction of the androecium to two fertile stamens, with further modifications marking sublineages within this large group (see below). Members of the Justicia lineage also lack the hygroscopic hairs present on seeds of members of the Barleria and Ruellia lineages (Balkwill and Getliff Norris, 1988).
The position of the Acanthus lineage is not resolved in this analysis, but morphology strongly suggests that it is the sister group to the other three lineages. A number of morphological characters support monophyly of the Barleria, Ruellia, and Justicia lineages as a group: presence of cystoliths, articulated stems, and porate pollen. Within this group, we know of no morphological support for the relationship between the Justicia and Ruellia lineages. In fact, this hypothesis is incongruent with the preliminary morphological analysis of Scotland et al. (1995) that proposes hygroscopic hairs on the seeds as a synapomorphy for the Barleria and Ruellia lineages. However, our results agree with those of Scotland et al. (1995) based on sequence data for the chloroplast ndhF gene. It is noteworthy that neither morphological nor molecular evidence provides a basis for resolving relationships among these three lineages with a great deal of confidence. It is possible that this radiation took place quickly, leaving little evidence of the sequence of divergences.
The Acanthus lineage
This lineage corresponds to Bremekamp's (1965) subfamily Acanthoideae, marked by monothecous stamens and the absence of characters shared by other Acanthaceae s.s. (i.e., cystoliths, articulated stems, porate pollen). The two sublineages within the Acanthus lineage are Bremekamp's (1965) Acantheae and Aphelandreae. The former is marked by modification of the upper lip of the corolla: it is deeply slit in Crossandra and virtually lacking in Acanthus. No unique morphological synapomorphies have yet been identified for Aphelandreae. Bremekamp (1965) placed three other very small tribes in Acanthoideae on the basis of their four monothecous stamens: Hasselhoffieae, Rhomobochlamydeae, and Stenandriopsideae. These plants lack the modifications to the upper lip characteristic of Acantheae and the strongly zygomorphic corollas of members of both other tribes. It thus seems likely that these lineages would be placed within Acanthoideae, perhaps basal to the taxa included here.
The Barleria lineage
Perhaps the most significant difference between existing classifications and the results of the present study are with regard to the Barleria lineage. For example, Bremekamp (1965) placed Barleria in subtribe Barleriinae within his tribe Ruellieae, while recognizing Lepidagathis and its relatives as a separate tribe, Lepidagathideae. Scotland, Endress, and Lawrence (1994) pointed out that these taxa in fact share a distinctive aestivation pattern (quincuncial) and suggested that they together comprise a distinct major monophyletic lineage. Balkwill and Getliff Norris (1988) associated Lepidagathis with Hygrophila on the basis of a number of characters including presence of a rugula (groove in the interior, dorsal surface of the corolla in which the style rests) and suggested that both belong in subtribe Petalidiinae of Ruellieae (along with Dyschoriste among others).
Our results place Hygrophila and Dyschoriste consistently and with very strong support with Ruellieae, whereas Lepidagathis is placed equally robustly with Barleria. Our results thus support the hypothesis of Scotland, Endress, and Lawrence (1994), as do those of Scotland et al. (1995), that the taxa with quincuncial corolla aestivation (here Barleria and Lepidagathis) form a monophyletic group that is distinct from those with left contort corolla aestivation (i.e., the Ruellia lineage). The delimitation and membership of these two groups clearly merit additional study.
The Ruellia lineage
The five taxa included here represent a fairly broad taxonomic range within this group. Dyschoriste, Blechum, and Hygrophila were placed by Bremekamp in tribe Ruellieae but in subtribes distinct from Ruellia, and Sanchezia represents Bremekamp's (1965) tribe Trichanthereae. Our results strongly support monophyly of this diverse group as a whole and of Blechum + Ruellia. The analysis with weighted indels also supports monophyly of the Ruellieae. These relationships should be viewed as tentative until more members of the lineage can be added. On the other hand, internal branches in this lineage are short (Fig. 2), suggesting that there may not be sufficient variation in this region to resolve relationships within the Ruellia lineage with confidence.
Relationships within the Justicia lineage
Bremekamp's (1965) Justicieae are the largest and perhaps the most difficult of the major groups of Acanthaceae. Because of its diversity and morphological variation, knowledge of acanth phylogenetics will benefit greatly from an improved understanding of patterns of relationships within this lineage. Our results strongly support monophyly of a major lineage that includes all members of Bremekamp's (1965) Justicieae included in this study, as well as the existence of several sublineages. One of these corresponds to Bremekamp's subtribe Justiciinae (here represented by two species of Justicia and Poikilacanthus, plus Megaskepasma). This last genus was placed by Bremekamp in Odontoneminae, but the similarity between its pollen and that of Poikilacanthus has been noted by Daniel (1991). Other members of Bremekamp's subtribe Odontoneminae are here distributed among several sublineages as described below.
Odontonema and Pseuderanthemum (Odontoneminae sensu Bremekamp) are strongly supported as each other's closest relatives. This is expected as there are several instances in which species now placed in one genus were originally described in the other. These plants have two stamens and two staminodes, as well as a strongly stipitate capsule with a marked constriction between the two seeds in each locule, such that the capsule is hourglass or violin shaped (Daniel, 1995b). This combination of characters has also been observed in several other genera (e.g., Chileranthemum, Pulchranthus, and Oplonia), and it is likely that these belong to this lineage. Heterostyly has been documented in all of these genera except Pulchranthus and is otherwise unknown from Acanthaceae (Daniel, 1995b). These plants have not been formally classified together, although their relationships have been noted (Baum, 1982; Daniel, 1995b).
Razisea and Stenostephanus (Odontoneminae sensu Bremekamp) are representatives of a group that probably includes a number of other genera of plants with two monothecous stamens. The difficulty in separating "genera" in this group was noted by Wood (1988) and Daniel (1995a), and most of the previously recognized genera have been transferred to Stenostephanus, including Habracanthus, Hansteinia, and a number of smaller genera. Many of these were placed by Lindau (1895) in one subtribe of his tribe Isoglosseae, which otherwise corresponds to Bremekamp's Rhytiglossinae (an Old World group with no representatives in the present study). Clearly, the addition of representatives of these other groups will be necessary to verify our hypothesis that this lineage exists and is correctly placed here.
Henrya sublineage (Odontoneminae sensu Bremekamp), excepting Pachystachys (see below), includes plants whose relationship has been noted by T. F. Daniel in a number of important contributions (e.g., Daniel, 1986, 1990). Daniel (1990) has proposed that several other genera are part of this group as well (e.g., Mirandea, Mexacanthus, Aphanosperma). These plants have two bithecous stamens and no staminodes; in addition to morphological similarities, all available chromosome counts are of n = 18. This group also has geographical integrity as the ranges of these genera are centered in Mexico. Unexpected is strong support for placement of Pachystachys with this group. To our knowledge, such a relationship has not been suggested previously. Bremekamp (1965), for example, placed this genus in Odontoneminae but not in close association with any members of the Henrya lineage. These plants are primarily of the Amazon basin and seem quite unlike the remainder of this lineage (Wasshausen, 1986). Members of Pachystachys do seem to share the androecial composition of the Henrya lineage [although Wasshausen (1986) mentioned rudimentary staminodes for a couple of species]. The only vouchered chromosome count to date is of n = 18 (Daniel and Chuang, 1989). Clearly, additional data will be required to confirm the placement of Pachystachys.
In our study, Dicliptera and Peristrophe represent what is likely to be a lineage of plants that is marked by a number of morphological characters including a resupinate corolla and a unique inflorescence structure consisting of cymule units (see Wood, 1988, p. 2). This group has been recognized as distinctive compared to other members of Bremekamp's Odontoneminae and proposed as worthy of taxonomic recognition by Balkwill and Getliff Norris (1988), who indicated that Hypoestes and Periestes belong here as well. Again, a broader sample of these taxa will be necessary to test this hypothesis fully.
Finally, Bremekamp's Justiciinae are a large group of plants that have been allied on the basis of having a rugula (groove in the interior, dorsal surface of the corolla in which the style rests), no staminodes, and two anthers, the thecae of which almost always are not parallel either due to displacement along the filament or elaboration of the connective. The number of recognized genera in this group has decreased over the past 25 years as several workers pointed to the lack of definition of these genera and moved constituent species into Justicia, which is now estimated to have as many as 600 species (Graham, 1988). This process is probably not complete. Once the overall diversity within this group is understood, it may be possible to recognize smaller monophyletic lineages that will help in organizing our knowledge of this unwieldy group.
Despite morphological evidence for this lineage, our data provide only weak support for its monophyly. This is in part because Dicliptera and J. longii share one homoplasious indel that weakens support for monophyly of Justiciinae. However, the branch lengths involved (see Fig. 2) suggest that this is more an issue of absence of information than of conflicting evidence. Morphology or a more rapidly evolving locus may be necessary to provide resolution at this level. Further, as noted above, we have not yet included Old World representatives of Justiciinae, and it is thus premature to conclude that the group as a whole is monophyletic.
In sum, this is an initial step toward understanding patterns of phylogenetic relationships within Justicieae. The results presented here suggest that Bremekamp's Odontoneminae are phylogenetically heterogeneous and, more importantly (as this was already realized), provide hypotheses regarding the existence and membership of monophyletic lineages among these plants. It should be noted that not all taxic diversity assigned to Justicieae can be accommodated in the proposed lineages: among unsampled groups of Justicieae are several with characteristics that defy hypotheses about their placement. This includes Asystasia with four stamens, and a perhaps heterogeneous group with an androecium of two bithecous and two monothecous stamens (Pranceacanthus, Jurausia, Herpetacanthus, Chamaeranthemum; Wasshausen, 1984). We also do not have representatives of Bremekamp's (1965) Rhytiglossinae, but based on the morphological characteristics of these plants, hypothesize that they will be placed within our Justicia lineage. Clearly, the addition of more taxa, including members of these unsampled lineages, and more sources of evidence will permit a more comprehensive phylogenetic treatment of Justicieae.
In conclusion, our results indicate that the trnL-trnF region is useful in addressing questions of phylogenetic relationships among but not within genera in these and related plants. In terms of phylogenetic relationships, Elytraria (representing Nelsonioideae) is more distantly related to Acanthaceae s.s. than Thunbergia and Mendoncia. These last two taxa are strongly supported as sister taxa. There is strong support for monophyly of four major lineages within Acanthaceae s.s., as suggested previously by Scotland et al. (1995): the Acanthus, Barleria, Ruellia, and Justicia lineages. There is also strong support for the monophyly of the last three, and weaker evidence linking the Barleria and Justicia lineages as each others' closest relatives. More data are required before relationships among these lineages can be regarded as resolved. Within the Acanthus lineage, our results confirm the existence of monophyletic lineages representing Bremekamp's Aphelandreae and Acantheae. Lastly, within the Justicia lineage, we have developed initial hypotheses regarding the definition of sublineages, some of which correspond to earlier ideas and some of which do not. All of these hypotheses need to be tested against more data, including both more characters and more taxa.
|
FOOTNOTES |
|---|
|
|
2 Author for correspondence (e-mail:mcdade{at}u.arizona.edu
).
3 Current address: Department of Botany, Washington State University, Pullman, Washington 99164.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Baum, V. M.1982A revision of the genus Odontonema (Acanthaceae). Master's thesis, University of Maryland, College Park, MD.
Bentham, G., and J. D. Hooker.1876Genera plantarum, vol. 2. Reeve and Co., London.
Bremekamp, C. E. B.1955Notes on some acanthaceous genera of controversial position. Acta Botanica Neerlandica 4: 644–655.
———.1965Delimitation and subdivision of the Acanthaceae. Bulletin of the Botanical Survey of India 7: 21–30.
Bremer, K.1988The limits of amino acid sequence data in angiosperm phylogenetic reconstruction. Evolution 42: 795–803. [CrossRef][ISI]
Cronquist, A.1981An integrated system of classification of flowering plants. Columbia University Press, New York.
Daniel, T. F.1986Systematics of Tetramerium (Acanthaceae). Systematic Botany Monographs 12: 1–134.
———.1990Systematics of Henrya (Acanthaceae). Contributions from the University of Michigan Herbarium 17: 99–131.
———.1991A synopsis of Poikilacanthus (Acanthaceae) in Mexico. Bulletin of the Torrey Botanical Club 118: 451–458. [CrossRef][ISI]
———.1995aNew and reconsidered Mexican Acanthaceae. VI. Chiapas. Proceedings of the California Academy of Science 48: 253–282.
———.1995bRevision of Odontonema (Acanthaceae) in Mexico. Contributions of the University of Michigan Herbarium 20: 147–171.
———, and T. I. Chuang.1989Chromosome numbers of some cultivated Acanthaceae. Baileya 23: 86–93.
Donoghue, M. J., R. G. Olmstead, J. F. Smith, and J. D. Palmer.1992Phylogenetic relationships of Dipsacales based on rbcL sequences. Annals of the Missouri Botanical Garden 79: 333–345.
Doyle, J. J. and J. L. Doyle.1987A rapid isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15.
Felsenstein, J.1985Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783–791. [CrossRef][ISI]
Gielly, L., and P. Taberlet.1994Chloroplast DNA polymorphism at the intrageneric level and plant phylogenies. Comptes Rendus de l'Academie des Sciences (Paris), Sciences de la Vie (Life Sciences) 317: 685–692.
———, and ———.1996A phylogeny of the European gentians inferred from chloroplast trnL (UAA) intron sequences. Botanical Journal of the Linnean Society 120: 57–75. [CrossRef]
———, Y.-M. Yuan, P. Küpfer, and P. Taberlet.1996Phylogenetic use of noncoding regions in the genus Gentiana L.: Chloroplast trnL (UAA) intron versus nuclear ribosomal internal transcribed spacer sequences. Molecular Phylogenetics and Evolution 5: 460–466. [CrossRef][ISI][Medline]
Gilbert, D. G.1992SeqApp: a biosequence editor and analysis application. Privately published by the author.
Graham, V. A. W.1988Delimitation and infra-generic classification of Justicia (Acanthaceae). Kew Bulletin 43: 551–624. [CrossRef]
Hedrén, M., M. W. Chase, and R. G. Olmstead.1995Relationships in the Acanthaceae and related families as suggested by cladistic analysis of rbcL nucleotide sequences. Plant Systematics and Evolution 194: 93–109. [CrossRef][ISI]
Kim, J. H., H. 'T. Hart, and T. H. M. Mes.1996The phylogenetic position of East Asian Sedum species (Crassulaceae) based on chloroplast DNA trnL (UAA)-trnF (GAA) intergenic spacer sequence variation. Acta Botanica Neerlandica 45: 309–321.
Kita, Y., K. Ueda, and Y. Kadota.1995Molecular phylogeny and evolution of the Asian Aconitum subgenus Aconitum (Ranunculaceae). Journal of Plant Research 108: 429–442. [CrossRef][ISI]
Lindau, G.1895Acanthaceae. In A. Engler and K. Prantl [eds.], Die natürlichen Pflanzenfamilien 4(3b), 274–354. Engelmann, Leipzig, Germany.
Mabberley, D. J.1987The plant-book. Cambridge University Press, Cambridge.
Maddison, D. R.1991The discovery and importance of multiple islands of most-parsimonious trees. Systematic Zoology 40: 315–328. [CrossRef]
Maddison, W. P., and D. R. Maddison.1992MacClade, version 3.0. Sinauer, Sunderland, MA.
McDade, L. A.1984Systematics and reproductive biology of the Central American species of the Aphelandra pulcherrima complex (Acanthaceae). Annals of the Missouri Botanical Garden 71: 104–165. [CrossRef][ISI]
Mes, T. H. M., J. van Brederode, and H. 't Hart.1996Origin of the woody Macaronesian Sempervivoideae and the phylogenetic position of the East African species of Aeonium. Botanica Acta 109: 477–491. [ISI]
Olmstead, R. G., B. Bremer, K. M. Scott, and J. D. Palmer.1993A parsimony analysis of the Asteridae sensu lato based on rbcL sequences. Annals of the Missouri Botanical Garden 80: 700–722. [CrossRef][ISI]
Scotland, R. W.1993Pollen morphology of Contortae (Acanthaceae). Botanical Journal of the Linnean Society 111: 471–504. [CrossRef]
———, P. K. Endress, and T. J. Lawrence.1994Corolla ontogeny and aestivation in the Acanthaceae. Botanical Journal of the Linnean Society 114: 49–65. [CrossRef]
———, J. A. Sweere, P. A. Reeves, and R. G. Olmstead.1995Higher-level systematics of Acanthaceae determined by chloroplast DNA sequences. American Journal of Botany 82: 266–275. [CrossRef][ISI]
Taberlet, P., L. Gielly, G. Pautou, and J. Bouvet.1991Universal primers for amplification of three non-coding regions of chloroplast DNA. Plant Molecular Biology 17: 1105–1109. [CrossRef][ISI][Medline]
Thorne, R. F.1992Classification and geography of the flowering plants. Botanical Review 58: 225–348.
Van Tieghem, P.1908Structure du pistil et de l'ovule du fruit et de la graine des Acanthacées. Annales des Science Naturelles, Serie 9, Botanique 7: 1–24.
Wagstaff, S J., and R. G. Olmstead.1997Phylogeny of Labiatae and Verbenaceae inferred from rbcL sequences. Systematic Botany 22: 165–177.
Wasshausen, D. C.1984Pranceacanthus coccineus (Acanthaceae): A new genus and species from Amazonian Brazil. Brittonia 36: 1–7.
———.1986The systematics of the genus Pachystachys (Acanthaceae). Proceedings of the Biological Society of Washington 99: 160–185. [ISI]
Wood, J. R. I.1988Colombian Acanthaceae—some new discoveries and some reconsiderations. Kew Bulletin 43: 1–51.
This article has been cited by other articles:
|
|
 |
|
  L. A. McDade, T. F. Daniel, and C. A. Kiel Toward a comprehensive understanding of phylogenetic relationships among lineages of Acanthaceae s.l. (Lamiales) Am. J. Botany, September 1, 2008; 95(9): 1136 - 1152. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. S. Judd and R. G. Olmstead A survey of tricolpate (eudicot) phylogenetic relationships Am. J. Botany, September 1, 2004; 91(10): 1627 - 1644. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Mast, D. M. S. Feller, S. Kelso, and E. Conti Buzz-pollinated Dodecatheon originated from within the heterostylous Primula subgenus Auriculastrum (Primulaceae): a seven-region cpDNA phylogeny and its implications for floral evolution Am. J. Botany, June 1, 2004; 91(6): 926 - 942. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. C. Moylan, J. R. Bennett, M. A. Carine, R. G. Olmstead, and R. W. Scotland Phylogenetic relationships among Strobilanthes s.l. (Acanthaceae): evidence from ITS nrDNA, trnL-F cpDNA, and morphology Am. J. Botany, May 1, 2004; 91(5): 724 - 735. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Perret, A. Chautems, R. Spichiger, G. Kite, and V. Savolainen Systematics and evolution of tribe Sinningieae (Gesneriaceae): evidence from phylogenetic analyses of six plastid DNA regions and nuclear ncpGS Am. J. Botany, March 1, 2003; 90(3): 445 - 460. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Schonenberger and E. Conti Molecular phylogeny and floral evolution of Penaeaceae, Oliniaceae, Rhynchocalycaceae, and Alzateaceae (Myrtales) Am. J. Botany, February 1, 2003; 90(2): 293 - 309. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Hall, K. J. Sytsma, and H. H. Iltis Phylogeny of Capparaceae and Brassicaceae based on chloroplast sequence data Am. J. Botany, November 1, 2002; 89(11): 1826 - 1842. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Stefanovic, L. Krueger, and R. G. Olmstead Monophyly of the Convolvulaceae and circumscription of their major lineages based on DNA sequences of multiple chloroplast loci Am. J. Botany, September 1, 2002; 89(9): 1510 - 1522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. R. Mast and T. J. Givnish Historical biogeography and the origin of stomatal distributions in Banksia and Dryandra (Proteaceae) based on their cpDNA phylogeny Am. J. Botany, August 1, 2002; 89(8): 1311 - 1323. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Pires and K. J. Sytsma A phylogenetic evaluation of a biosystematic framework: Brodiaea and related petaloid monocots (Themidaceae) Am. J. Botany, August 1, 2002; 89(8): 1342 - 1359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. G. Razafimandimbison and B. Bremer Phylogeny and classification of Naucleeae s.l. (Rubiaceae) inferred from molecular (ITS, rBCL, and tRNT-F) and morphological data Am. J. Botany, July 1, 2002; 89(7): 1027 - 1041. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. M. Beardsley and R. G. Olmstead Redefining Phrymaceae: the placement of Mimulus, tribe Mimuleae, and Phryma Am. J. Botany, July 1, 2002; 89(7): 1093 - 1102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
