| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
Harvard University Herbaria, 22 Divinity Avenue, Cambridge, Massachusetts 02138 USA
Received for publication August 21, 2003. Accepted for publication February 10, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Antirrhineae • Antirrhinum • ITS • North America • Saerorhinum • Scrophulariaceae • Veronicaceae
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Coen and Meyerowitz, 1991
; Martin et al., 1991
; Cubas et al., 1999
; Hileman et al., 2003
). However, our ability to use these insights to understand plant evolution has been hampered by a poor understanding of phylogenetic relationships within Antirrhinum. A phylogeny would allow us to better exploit our knowledge of the genetics of the model organism A. majus to understand patterns of character evolution in closely related wild species. Furthermore, given the tremendous morphological and ecological diversity in this group (Kampny, 1995
), a phylogeny may help to pinpoint interesting patterns that warrant further study. Here, we investigate the phylogenetic relationships within the genus Antirrhinum with an emphasis on the New World section Saerorhinum, which has particularly high levels of morphological and ecological diversity.
Antirrhinum
The genus Antirrhinum (sensu Thompson, 1988
) consists of 36 species in three sections. Section Saerorhinum, the focus of this study, includes 15 small-flowered, mostly annual, tetraploid (n = 15–16; except A. cyathiferum with n = 13) species distributed in western North America. Section Antirrhinum, which includes the model organism A. majus, comprises 19 perennial species (n = 8) with relatively large flowers that are native to the western Mediterranean regions, with most of the species occurring as narrow endemics in the Iberian Peninsula (Tutin et al., 1972
). The two species in section Orontium also occur in the Mediterranean region and have n = 8 chromosomes, but are differentiated from section Antirrhinum by their smaller flowers and annual habit. Section Orontium differs from both sections Antirrhinum and Saerorhinum by the presence of greatly elongated calyx lobes relative to the corolla.
The species in section Saerorhinum are geographically centered in California, though some species occur as far south as Baja California Sur, as far north as southern Oregon, and as far east as western Utah. They are found on a variety of substrates including shale, serpentine, salt flats, and recently burned soils. The species are mainly weedy annuals, but A. multiflorum and A. virga are perennials with substantial stems that can grow to over 1 m in height, and A. nuttalianum is occasionally biennial (R. Oyama, personal observations; Thompson, 1988
).
In addition to the variation in substrate preference and growth habit, species in section Saerorhinum also display a high diversity of leaf morphology, flower form, flower size, and flower color. The flowers of most species of section Saerorhinum have the characteristic personate corolla found in the common garden snapdragon A. majus. This highly zygomorphic morphology includes a lower lip (or palate) that must be mechanically displaced to gain access to the interior of the flower, but it lacks the pronounced spur common in the rest of the tribe Antirrhineae. A few species in section Saerorhinum, however, have flowers that are permanently open-throated (e.g., A. coulterianum and A. ovatum), while others are cleistogamous (e.g., A. costatum, A. kingii, A. watsonii, A. kelloggii, A. nuttallianum, and A. filipes). Depending on the species, the corolla may be white, purple, pink, or yellow, and may be accentuated with spots or pigmented veins. In a few species (A. leptaleum, A. multiflorum, A. ovatum, and A. virga), part of the palate withers before anthesis.
Phylogeny and taxonomy
Recent phylogenetic studies of Scrophulariaceae sensu lato and related groups in the Lamiales have greatly changed our understanding of the assemblage of plants to which Antirrhinum is related (Olmstead et al., 1992
, 2001
; Olmstead and Reeves, 1995
; Freeman and Scogin, 1999
). These studies have identified a reasonably well-supported clade, called Veronicaceae or Antirrhinaceae (Reveal et al., 1999
; Olmstead et al., 2001
), comprised of some elements of the traditional Scrophulariaceae, including tribe Antirrhineae, plus some genera traditionally placed in other families (e.g., Plantago).
The tribe Antirrhineae, which includes both model organisms Linaria vulgaris and Antirrhinum majus, has also received attention from systematists using morphological and molecular data to explore its taxonomy and the phylogenetic relationships of its constituent taxa (Elisens, 1985a
, b
, 1986
; Thompson, 1988
; Freeman and Scogin, 1999
; Ghebrehiwet et al., 2000
). Despite this work, phylogenetic relationships among all of the genera in the tribe Antirrhineae and within the genus Antirrhinum remain uncertain. Much of this confusion is due to disagreement on the circumscription of Antirrhinum together with incomplete or uneven sampling. The inclusion of the New World species that comprise section Saerorhinum within the taxonomic genus Antirrhinum is particularly contentious (see Wettstein, 1891
; Munz, 1926
; Rothmaler, 1956
; Sutton, 1988
; Thompson, 1988
).
In the most recent and comprehensive monograph of the New World species of Antirrhinum (section Saerorhinum), Thompson (1988)
proposed a broadly inclusive circumscription of Antirrhinum encompassing the 36 species from both the New World and the Old World. In doing so, he disagreed with Rothmaler (1956)
, who had proposed a less inclusive circumscription of the genus that segregated some of the New World species into separate genera, and with Sutton (1988)
, who segregated all of the New World species into a series of smaller genera (see Table 1). To date, no phylogenetic analysis of the genus has been able to evaluate the relative merits of these various taxonomic schemes.
|
recovered a well-resolved phylogeny for section Saerorhinum using morphological data, but the tree had low levels of internal support and lacked adequate sampling to test generic or sectional monophyly (see Fig. 1A). A study of the tribe Antirrhineae by Ghebrehiwet et al. (2000)
, using morphological data and ndhF sequences, raised questions about the monophyly of the genus Antirrhinum (sensu Thompson) by recovering phylogenies in which Mohavea and Schweinfurthia were embedded within Antirrhinum s.l. (see Fig. 1B). However, the ndhF study did suggest a sister group relationship between section Antirrhinum (A. majus) and the other two sections (represented by A. orontium and A. ovatum). A similar study by Freeman and Scogin (1999)
using trnL sequences and a better sampling of Antirrhinum species also raised questions about the monophyly of Antirrhinum by recovering phylogenies placing Galvezia and Mohavea in a clade with species from section Saerorhinum. However, the trnL study could not provide much resolution of the relationships within section Saerorhinum and, because it did not sample from section Orontium, it could not resolve relationships among the three sections (see Fig. 1C). Clearly, a better phylogenetic understanding of this group is required, considering its growing importance in research into plant development.
|
; Vargas et al., 1998
). To further resolve the relationships within section Saerorhinum, we combined the ITS data with the morphological matrix of Thompson (1988)
to conduct the best sampled phylogenetic analysis for the New World Antirrhinum. This phylogeny, combined with the extensive variation in morphological, ecological, and life-history traits found in section Saerorhinum, together with the genetic knowledge gleaned from A. majus, offers great potential to improve our understanding of the evolutionary and molecular basis of diversity. |
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
), but because of alignment difficulties, only Veronica was used in the analyses.
DNA isolation and sequencing
Total genomic DNA was extracted from silica-gel-dried material or herbarium specimens using a modified 2x cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle, 1987
). The DNA was cleaned using both an ammonium acetate and a sodium acetate precipitation. The DNA was then used to amplify the ITS region including the 5.8S gene and the flanking regions of the 18S and 26S genes. The primers ITSleu (Baum et al., 1998
) and ITS4 (White et al., 1990
) were used for the PCR amplification. Conditions for PCR were as in Baum et al., (1998)
. Successful reactions were then cleaned using Qiagen PCR purification columns (Qiagen Inc., Valencia, California, USA).
Some extractions from A. vexillo-calyculatum and of A. majus required cloning in order to obtain unambiguous sequences. In those cases, the PCR reactions were run on a 1% agarose gel, the bands were visualized using ethidium bromide, cut out, and cleaned using Qiagen gel extraction columns (Qiagen). The PCR products were cloned using the Promega pGEM T-Easy cloning system (Promega Corp., Madison, Wisconsin, USA). Plasmid DNA was purified using mini-prep columns (Sigma Chemical Co., St. Louis, Missouri).
Each clone or successful PCR reaction was sequenced using BigDye, version 2 (ABI Perkin-Elmer, Foster City, California, USA) and four primers: ITSleu, ITS2 (White et al., 1990
), ITS3B (Baldwin, 1992
), and ITS4. All sequences used were read in both directions. In the case of sequencing from cloned DNA, the universal primers SP6 and T7 were substituted for ITSleu and ITS4.
Phylogenetic analyses
Sequences were assembled, edited, and manually aligned using SequencherTM (Gene Codes Corp., Ann Arbor, Michigan, USA). In most cases, there was little ambiguity in alignment. However, when there were alternatives that seemed equally plausible, we chose the alignment that minimized the number of potentially informative characters to avoid biasing the results. Sequences were submitted to GenBank with accession numbers as given in Appendix 1 (see Supplemental Data accompanying the online version of this article).
Parsimony analyses were performed using PAUP* version 4.0b10 (Swofford, 2001
). All searches for most parsimonious trees were conducted with tree bisection-reconnection (TBR) heuristic searches and 100 random taxon addition replicates. In all bootstrap analyses, support was obtained using 100 bootstrap replicates (Felsenstein, 1985
), as implemented in PAUP* version 4.0b10, with 100 random taxon addition replicates.
Parsimony analyses of the full ITS data set (e.g., with all 35 taxa) were initially performed without using insertion/deletion (indel) characters. Subsequently, we examined the data set for unambiguous indels (those greater than 1 base pair and found in two or more taxa but not overlapping with indels in other taxa). One such indel was identified, scored, added to the matrix, and given a weight of two (Fig. 2). A parsimony search with the indel characters was then performed with the same parameters as described earlier.
|
). Given that Bayesian analysis requires a model of molecular evolution to be specified, we used the program ModelTest v3.06 (Posada and Crandall, 1998
) to select an appropriate model of DNA substitution, rather than choosing one randomly. ModelTest selects the most appropriate model of molecular evolution for a given matrix of molecular sequence and reference tree (in this case one of the most parsimonious trees) under two different optimality criteria: the Akaiki information criterion (AIC); and the hierarchical likelihood ratio test (hLRT). Under the AIC, ModelTest selected a general time-reversible model (GTR) with a discrete approximation to the gamma distribution for rate variation among sites (G) and a proportion of invariant sites (I) (Yang, 1994a
, b
). Under the hLRT, ModelTest selected a more restricted model of TrN + I + G (Tamura and Nei, 1993
; Posada and Crandall, 1998
). To avoid biasing the results, we chose to use the more flexible GTR + I + G model to conduct the analyses reported here. However, in light of the known sensitivity of Bayesian analyses, we also performed analyses using the alternative model (TrN + I + G) and the parameters given by ModelTest to explore any differences between the two models (Huelsenbeck et al., 2002
; Miller et al., 2002
). The indel character was not used in any Bayesian analyses.
Tree searching using MrBayes was performed by running five coupled chains initiated from random trees (sequential heat = 0.2) for 1 000 000 generations with trees sampled every 100th generation. At the end of the run, convergence was evaluated by visual inspection of a graph of likelihood as a function of generation. A conservative burn-in period was identified, and only post-burn-in trees were saved. This analysis was repeated five times, and the majority-rule consensus trees from each run were compared to evaluate mixing. The five sets of post-burn-in trees were then pooled to form a majority rule consensus tree, and this pool was taken as the best representation of the posterior distribution of tree topology and model parameters (Huelsenbeck and Ronquist, 2001
; Miller et al., 2002
).
The phylogenetic placement of certain problematic taxa was examined using the Templeton test (Templeton, 1983
) and the Kishino–Hasegawa test (Kishino and Hasegawa, 1989
) as implemented in PAUP* 4.0b10 (Swofford, 2001
) on the ITS data. In each case, the most parsimonious tree was compared to a tree constraining the taxon of interest to its alternative placement. The Kishino-Hasegawa tests assumed the GTR + G model parameters (Yang, 1994a
, b
).
To further examine the phylogenetic relationships within section Saerorhinum, we supplemented our data with Thompson's (1988)
morphological data and conducted a combined analysis. To do so, we reduced the ITS data set to the 18 taxa analyzed by Thompson (1988)
. In light of the results from our analysis of the full ITS data set (see later and Figs. 3 and 4), two taxa not scored by Thompson (1988)
, Antirrhinum orontium and the genus Mohavea (represented by M. breviflora), were added to the matrix. Morphological character states for these additional taxa not included in Thompson's (1988)
analysis were taken from previously published information (Sutton, 1988
; Thompson, 1988
) and from observations of living plants and herbarium specimens (R. Oyama, personal observations). This was further supplemented with two additional characters for all taxa: one for flower color and one for the presence of red flecks on the palate (Appendix 2, see Supplemental Data accompanying the online version of this article).
|
|
). Parsimony analyses of the 20 taxon combined data set, focusing on section Saerorhinum, were conducted on the ITS data alone, on the morphological data alone, and on the combined data. These analyses did not include the ITS indel character, and all characters were weighted equally. |
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Parsimony analysis of the ITS data for all 35 taxa using the indel character resulted in eight most parsimonious trees with a length of 775 (CI = 0.585, RI = 0.666). In all five replicate Bayesian analyses, trees converged on a log-likelihood score of approximately –4700. The strict consensus of the eight most parsimonious trees is shown in Fig. 3 with the Bayesian posterior probabilities placed under branches leading to clades and parsimony bootstrap values above. A phylogram of one of the eight most parsimonious trees is shown in Fig. 4 to illustrate branch lengths.
A strongly supported clade that includes all species of Antirrhinum (except A. cyathiferum) and the genus Mohavea was recovered in all analyses of the full data set (see Figs. 3 and 4). This "Antirrhinum clade" is nested well within the tribe Antirrhineae and is sister to the genus Chaenorrhinum. Both the Bayesian and parsimony analyses of the ITS sequences strongly support the monophyly of section Antirrhinum. This result was unchanged when additional species from the section were added, but the lack of variation in ITS among species in section Antirrhinum resulted in no resolution within the clade. Consequently, only one placeholder for each subsection was used in the analyses presented here.
Although all eight most parsimonious trees recovered a monophyletic "Saerorhinum clade," composed of the species in section Saerorhinum but excluding A. cyathiferum and including the genus Mohavea (see Figs. 3 and 4), the bootstrap support is below 50%. Bayesian analysis using the GTR + I + G model of molecular evolution also recovers a low support value of only 52% for this clade (Fig. 3), but when the TrN + I + G model for molecular evolution is used, with parameters as determined by ModelTest v3.06 (Tamura and Nei, 1993
; Posada and Crandall, 1998
), support for a monophyletic "Saerorhinum clade" reaches 83%.
All species of sections Orontium and Saerorhinum share a unique 18-bp deletion (Figs. 2 and 3), though they are differentiated by a chromosome count of n = 8 in section Orontium and n = 15 or 16 in section Saerorhinum. Parsimony analysis of the full data set performed without using the 18-bp deletion character, placed A. orontium as sister to the other two sections of the genus. Including the 18-bp deletion as a character in the parsimony analysis and giving it a weight of two resulted in the placement of A. orontium as sister to the "Saerorhinum clade" with a bootstrap support of 58%. Bayesian analysis using the GTR + I + G model failed to discriminate between these two alternative placements for section Orontium. However, Bayesian analysis using the TrN + I + G model of molecular evolution strongly supported a sister relationship between A. orontium and the "Saerorhinum clade" (98%).
Parsimony searches constrained to force A. cyathiferum into either the Antirrhinum or the Saerorhinum clades, as was found using trnL (Freeman and Scogin, 1999
), yielded trees that were significantly longer as judged by both the Templeton and Kishino–Hasegawa tests (P < 0.05). Constraining Schweinfurthia to be in the "Saerorhinum clade," as was found with ndhF (Ghebrehiwet et al., 2000
), yielded trees that were significantly longer than the most parsimonious trees according to the Templeton test (P < 0.05) but not the Kishino– Hasegawa test. The parsimony and Bayesian analyses strongly supported the inclusion of the genus Mohavea (represented here by both of its species) within the "Antirrhinum clade." Constraining a search to enforce a clade composed of all Antirrhinum (e.g., excluding A. cyathiferum and Mohavea) resulted in trees that could be rejected by the Kishino–Hasegawa test (P < 0.01) but not by the Templeton test.
Within the "Saerorhinum clade" (section Saerorhinum plus Mohavea but excluding A. cyathiferum. See Fig. 4.), both the parsimony and Bayesian analyses found similar species level relationships. The two species of the genus Mohavea were monophyletic in all analyses (bootstrap = 100%). Similarly, A. virga and A. multiflorum always formed a monophyletic group (bootstrap = 99%). The optimal trees also suggest that the white-flowered A. subcordatum is nested within the purple-flowered species A. vexillo-calyculatum (bootstrap = 61%). These two species are then strongly supported as sister to A. ovatum (bootstrap = 90%) despite variation in chromosome numbers (A. ovatum, n = 16; A. subcordatum and A. vexillo-calyculatum, n = 15). Furthermore, this clade is supported by morphological characters (e.g., presence of leaf and calyx glands, and calyx segments).
Given the most parsimonious topology, chromosome number shows homoplasy within the "Saerorhinum clade": a minimum of three transitions between n = 16 and n = 15 need to be invoked (see Fig. 3). Although the Templeton test cannot reject monophyletic clades of n = 15 (P = 0.0707) or n = 16 (P = 0.1167) taxa, attempts to force species of section Saerorhinum into monophyletic groups based on chromosome number could be rejected by a Kishino–Hasegawa test (P < 0.05).
The reduced matrix, focusing on the "Saerorhinum clade" plus A. cyathiferum, consisted of the 696 bp of ITS sequence data and 18 morphological characters (Appendix 2, see Supplemental Data accompanying online version of this article). Partition homogeneity tests conducted in PAUP*4.0b10 indicated significant discordance between the morphological and molecular data sets (P < 0.01) (Farris et al., 1994
). Nonetheless, we conducted a combined analysis to see how conflicting signal becomes reconciled by flat-weighted parsimony.
Parsimony analysis of the ITS data alone resulted in 66 trees with a length of 227 (CI = 0.767, RI = 0.531), whose strict consensus is shown in Fig. 5A. Parsimony analysis of the morphological data for these same taxa resulted in 32 most parsimonious trees with a length of 47 (CI = 0.468, RI = 0.632). The strict consensus of these trees is shown in Fig. 5B. A parsimony analysis of the combined molecular and morphological data resulted in 18 most parsimonious trees with a length of 283 (CI = 0.693, RI = 0.519). The strict consensus of these trees is shown in Fig. 5C. The morphological data yielded the least resolved phylogeny.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Freeman and Scogin, 1999
; Ghebrehiwet et al., 2000
). Importantly, our analyses confirm the monophyly of Antirrhinum as delimited by Thompson (1988)
given the slight modification of removing one species (A. cyathiferum) and the addition of two species (Mohavea confertiflora and M. breviflora) to the circumscription of the genus. The results also point towards interesting questions of floral evolution, development, adaptation, and speciation that may now be better addressed with this phylogenetic hypothesis.
Placement of Antirrhinum s.l. within the tribe Antirrhineae
The relationships among the sampled genera of the tribe Antirrhineae are similar to the phylogeny recovered by Ghebrehiwet et al. (2000)
using morphological characters and ndhF sequence data. The major differences pertinent to the focus of this study are the placement of the genera Linaria and Schweinfurthia. The study by Ghebrehiwet et al. (2000)
found Linaria to be closely related to Antirrhinum, while our ITS data places it very distant. This might be due to the fact that Linaria is a very large and diverse genus and the two studies sampled different species. However, the more distant position recovered by the ITS data is supported by pollen characters (Elisens, 1985a
, c
, 1986
). The ndhF data (Ghebrehiwet et al., 2000
) also supported the inclusion of Schweinfurthia (from eastern Africa) within an Antirrhinum clade, whereas our ITS data support a more distant position for Schweinfurthia. This latter result is supported by differences in the fusion of the anther lobes between the two genera (Wettstein, 1891
).
The monophyly of Antirrhinum was also called into question by a study using sequences from the chloroplast marker trnL, which placed Galvezia within a clade of species from section Saerorhinum (Freeman and Scogin, 1999
). Phylogenies recovered using the chloroplast marker ndhF (Ghebrehiwet et al., 2000
) agree with our finding that Galvezia is not within Antirrhinum (sensu Thompson). One possible explanation for this is that the trnL study sampled Galvezia speciosa, whereas we used Galvezia fruticosa. One taxonomic treatment of the tribe Antirrhineae segregated Galvezia speciosa into the genus Gambelia and reserved the genus Galvezia for those species found in northern South America (Sutton, 1988
). Thus, Gambelia speciosa and the other species in the genus, which are all found in California or Baja California, could be embedded within the "Saerorhinum clade," much like the genus Mohavea. However, the ndhF phylogeny places Gambelia speciosa and Galvezia fruticosa together in a monophyletic group outside the "Antirrhinum clade," which conflicts with the trnL phylogeny but agrees with our ITS based findings (see Figs. 1 and 3). Clearly, other markers must be employed to resolve the phylogenetic placement of Gambelia. This could include sequencing ITS for Gambelia since we were unable to obtain DNA for use in this study.
The genus Antirrhinum
This study suggests that Antirrhinum, as circumscribed by Thompson (1988)
, but emended as discussed, is a monophyletic group. Support for the inclusion of the two Mohavea species within Antirrhinum and the exclusion of A. cyathiferum is corroborated by other evidence as discussed later. This study does not necessarily support the segregation of the New World species into multiple distinct genera (as in Sutton, 1988
). However, our analyses do support Thompson's (1988)
recognition and circumscription of three sections in a broadly circumscribed genus.
The sequence data do not unambiguously resolve the relationships among the three sections of the genus Antirrhinum. Parsimony analysis with all 35 taxa performed without using the single strong indel character placed section Orontium as sister to the rest of the genus, whereas inclusion of that character resulted in a sister relationship between sections Orontium and Saerorhinum (Fig. 3). Similarly, the Bayesian analysis under a TrN + I + G model of molecular evolution also supports a sister group relationship between sections Orontium and Saerorhinum, while the more flexible GTR + I + G model, which is the basis for the Bayesian results reported here, does not. The conflicting results seem to be an example of how, especially when the data are ambiguous, Bayesian analysis may be sensitive to the priors used, as well as to the model of molecular substitution (Huelsenbeck et al., 2002
). However, a sister group relationship between sections Orontium and Saerorhinum is supported by morphological characters that unite the two sections, such as similarly small flowers, unequal calyx lobes, and annual growth habit. In addition, other molecular markers such as the gene CYCLOIDEA (Hileman, 2002
) and NIA (Howarth and Baum, 2002
; Oyama, 2002
) strongly support this relationship.
Section Antirrhinum is shown to be monophyletic by all analyses, with the clade having high bootstrap support and posterior probability. The three species of section Antirrhinum shown in Figs. 3 and 4 represent the three subsections of section Antirrhinum as defined by Rothmaler (1956)
. Including ITS data for additional species from section Antirrhinum does not affect the strong support for its monophyly, but the low variation in ITS sequences among the species in the section meant that there were few informative characters and the clade formed a polytomy (P. Vargas, Reál Jardín Botánico-Madrid, unpublished data; R. Oyama, unpublished data). Our sampling of taxa is obviously inadequate to address relationships among the ca. 21 species in section Antirrhinum.
Mohavea and A. cyathiferum (= Pseudorontium cyathiferum)
The strong support in these analyses for including the two species of Mohavea within the "Saerorhinum clade" of Antirrhinum is consistent with published molecular systematic studies such as the trnL phylogeny of Freeman and Scogin (1999)
(Fig. 1C), and the ndhF phylogeny of Ghebrehiwet et al. (2000)
(Fig. 1B). However, previous taxonomists relying on morphological characters did not predict this result. The two species of Mohavea have flowers that look superficially very different from other species of Antirrhinum due to the near actinomorphy of the corolla lobes (despite strong zygomorphy of the corolla tube and palate). Additionally, Mohavea species have highly modified pollen morphology (Elisens, 1986
), and two stamens at maturity rather than the four stamens found in all other Antirrhineae.
Nonetheless, there is morphological support for the inclusion of Mohavea in the "Saerorhinum clade" close to A. coulterianum and A. filipes (Figs. 3 and 5A). The presence of red spots or flecks on the palate of the flowers is a synapomorphy that unites all four species. Furthermore, the monophyletic grouping of A. filipes and the two Mohavea species, all native to the Mohave desert, is supported by the synapomorphy of a yellow corolla. The unusual floral characteristics of Mohavea are likely the result of its specialization to pollen-collecting bees (Hileman, 2002
).
The placement of Antirrhinum cyathiferum outside the main Antirrhinum clade is well supported by ITS sequences and also by its lack of the 18-bp deletion that occurs in all New World species of Antirrhinum (and the Old World species A. orontium). Although the trnL phylogeny found 95% bootstrap support for placing A. cyathiferum in a clade with species of Antirrhinum section Saerorhinum (Freeman and Scogin, 1999
), morphological data support the ITS resolution. Rothmaler (1956)
and Sutton (1988)
excluded the species from Antirrhinum and gave it the name Pseudorontium cyathiferum. This exclusion found support in Elisens (1986)
who noted its unique pollen morphology. The position of A. cyathiferum as distant to the rest of the New World snapdragons is also supported by its chromosome count of n = 13 vs. a chromosome count of n = 15 or n = 16 for all of the other taxa in section Saerorhinum. Therefore, the position for A. cyathiferum in the trnL phylogeny suggests that either the plastid DNA reflects a misleading phylogenetic signal or that there is discordance between the nuclear and plastid genomes.
Section Saerorhinum
Our analyses found that Antirrhinum sect. Saerorhinum (sensu Thompson) is monophyletic given minor modification to its composition so as to include the two species of the current genus Mohavea and exclude A. cyathiferum. This emended group, the "Saerorhinum clade," is united by a putative tetraploidization event, that has resulted in chromosome counts of n = 15 or n = 16.
It is tempting to speculate on the parental origin of the tetraploid "Saerorhinum clade" based on its ancestral chromosome count. The most parsimonious ITS trees suggest multiple transitions between n = 16 and n = 15 within the clade, but trees involving only a single transition are not rejected by all statistical tests. A chromosome count of n = 16 might appear to be the most likely ancestral state for the Saerorhinum clade given that the sister-groups sampled here, sections Antirrhinum and Orontium, both have n = 8. In this scenario, one of the parents was likely drawn from sect. Orontium given the shared characteristics between sections Orontium and Saerorhinum, such as the annual habit, small flowers, and the 18-bp deletion event. However, these ITS data do not discriminate between an auto- or allo-tetraploid origin. Improved sampling within the tribe Antirrhineae will be necessary to resolve the origin of the Saerorhinum clade and its chromosomal evolution.
The best resolution of the phylogenetic relationships within the "Saerorhinum clade" is provided by the ITS data alone. The morphological data do not provide greater resolution, but this is likely due to the low number of characters used relative to the number of taxa. Including the morphological data with the ITS sequence in a combined analysis results in much less resolution than the topology recovered using ITS alone because there is significant conflict between the two data sets. Not surprisingly, the strict consensus trees recovered from each of the data sets are incompatible with each other.
The phylogeny recovered by the ITS data set (Figs. 3 and 5A) has implications for the evolution of certain floral characters within the "Saerorhinum clade." For example, the phylogeny suggests that open flowers evolved independently in A. ovatum (n = 16) and A. coulterianum (n = 15). This is remarkable considering the otherwise consistent corolla architecture elsewhere in Antirrhinum. Likewise, withering of the palate shows homoplasy; the most parsimonious resolution involves two independent origins, once on the lineage leading to A. ovatum and once on the lineage leading to A. leptaleum, A. multiflorum, and A. virga. Finally, there is the extreme evolution of the floral architecture in the two species of Mohavea.
It is noteworthy that section Saerorhinum includes some small clades that might be suitable for future studies in the area of evolution and developmental genetics. The sister group relationship between A. multiflorum, which is covered in sticky glandular hairs, and A. virga, which is almost completely glabrous, might represent an opportunity to explore the role of genes such as GLABROUS-1 in variation among wild species (Szymanski et al., 1998
). Likewise, the independent evolution of withered palates in two lineages (A. ovatum and the clade sister to A. cornutum) might provide an interesting system for the study of programmed cell death. Lastly, the close relationship of the white-flowered A. subcordatum and the purple-flowered A. vexillo-calyculatum makes them an attractive system to study the evolution of genes affecting floral pigmentation. In addition, the occurrence of these two species in sympatry suggests that they may be suited to studies of pollinator-mediated speciation.
Conclusions
Our study strongly supports the view of Thompson (1988)
that the genus Antirrhinum, given minor modification to the meaning, is a monophyletic group nested within the tribe Antirrhineae. In addition, section Saerorhinum (given the same modifications to the meaning) forms a clade closely related to the model organism Antirrhinum majus. The two species in the genus Mohavea are strongly supported as belonging to this "Saerorhinum clade" and, given their highly divergent floral morphologies, add yet more variation to an already extremely diverse group of plants. The relationships within section Saerorhinum are well resolved, and this study now offers a phylogenetic framework in which questions of character evolution, development, ecology, and speciation may be addressed.
|
FOOTNOTES |
|---|
2 Current address: Max-Planck Institut für chemische Ökologie, Hans-Knöll Strasse 8, 07743, Jena, Germany
3 Current address: Department of Botany, University of Wisconsin, Madison, WI 53706
4 E-mail: royama{at}ice.mpg.de
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Baldwin B. G. M. J. Sanderson J. M. Porter M. F. Wojciechowski C. S. Campbell M. J. Donoghue 1995 The ITS region of nuclear ribosomal DNA—a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82: 247-277[CrossRef][ISI]
Baum D. A. R. L. Small J. F. Wendel 1998 Biogeography and floral evolution of baobabs (Adansonia, Bombacaceae) as inferred from multiple data sets. Systematic Biology 47: 181-207[CrossRef][ISI][Medline]
Coen E. S. E. M. Meyerowitz 1991 The war of the whorls—genetic interactions controlling flower development. Nature 353: 31-37[CrossRef][Medline]
Cubas P. C. Vincent E. Coen 1999 An epigenetic mutation responsible for natural variation in floral symmetry. Nature 401: 157-161[CrossRef][Medline]
Doyle J. J. J. L. Doyle 1987 A rapid DNA isolation procedure for small quantities of fresh leaf material. Phytochemical Bulletin 19: 11-15
Elisens W. J. 1985a Monograph of the Maurandyinae (Scrophulariaceae-Antirrhineae). Systematic Botany Monographs 5: 1-97
Elisens W. J. 1985b The systematic relationship of Asarina procumbens to new world species in tribe Antirrhineae (Scrophulariaceae). Madroño 32: 168-178
Elisens W. J. 1985c The systematic significance of seed coat anatomy among New World species of tribe Antirrhineae (Scrophulariaceae). Systematic Botany 10: 282-299[CrossRef][ISI]
Elisens W. J. 1986 Pollen morphology and systematic relationships among New-World species in tribe Antirrhineae (Scrophulariaceae). American Journal of Botany 73: 1298-1311[CrossRef][ISI]
Farris J. S. M. Kallersjo A. G. Kluge C. Bult 1994 Testing significance of incongruence. Cladistics, The International Journal of the Willi Hennig Society 10: 315-319[ISI]
Felsenstein J. 1985 Confidence-limits on phylogenies—an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Freeman C. E. R. Scogin 1999 Potential utility of chloroplast trnL (UAA) gene intron sequences for inferring phylogeny in Scrophulariaceae. El Aliso 18: 141-159
Ghebrehiwet M. B. Bremer M. Thulin 2000 Phylogeny of the tribe Antirrhineae (Scrophulariaceae) based on morphological and ndhF sequence data. Plant Systematics and Evolution 220: 223-239[CrossRef][ISI]
Hileman L. C. 2002 Floral symmetry genes Cycloidea and Dichotoma: gene duplication and the evolution of stamen number in Antirrhineae (Veronicaceae). Ph.D., Harvard University, Cambridge, Massachusetts, USA
Hileman L. C. E. M. Kramer D. A. Baum 2003 Differential regulation of symmetry genes and the evolution of floral morphologies. Proceedings of the National Academy of Sciences, USA 100: 12814-12819
Howarth D. G. D. A. Baum 2002 Phylogenetic utility of a nuclear intron from nitrate reductase for the study of closely related plant species. Molecular Phylogenetics and Evolution 23: 525-528[CrossRef][ISI][Medline]
Huelsenbeck J. P. F. Ronquist 2001 MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754-755
Huelsenbeck J. P. B. Larget R. E. Miller F. Ronquist 2002 Potential applications and pitfalls of Bayesian inference of phylogeny. Systematic Biology 51: 673-688[CrossRef][ISI][Medline]
Kampny C. M. 1995 Pollination and flower diversity in Scrophulariaceae. Botanical Review 61: 350-366
Kishino H. 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][ISI][Medline]
Martin C. A. Prescott S. Mackay J. Bartlett E. Vrijlandt 1991 Control of anthocyanin biosynthesis in flowers of Antirrhinum majus. The Plant Journal 1: 37-49[ISI][Medline]
Miller R. E. T. R. Buckley P. S. Manos 2002 An examination of the monophyly of morning glory taxa using Bayesian phylogenetic inference. Systematic Biology 51: 740-753[CrossRef][ISI][Medline]
Munz P. A. 1926 The Antirrhinoideae–Antirrhineae of the New World. Proceedings of the California Academy of Science series 415: 323-397
Olmstead R. G. C. W. dePamphilis A. D. Wolfe N. D. Young W. J. Elisons P. A. Reeves 2001 Disintegration of the Scrophulariaceae. American Journal of Botany 88: 348-361
Olmstead R. G. H. J. Michaels K. M. Scott J. D. Palmer 1992 Monophyly of the Asteridae and identification of their major lineages inferred from DNA-sequences of rbcL. Annals of the Missouri Botanical Garden 79: 249-265
Olmstead R. G. P. A. Reeves 1995 Evidence for the polyphyly of the Scrophulariaceae based on chloroplast rbcL and ndhF sequences. Annals of the Missouri Botanical Garden 82: 176-193[CrossRef][ISI]
Oyama R. K. 2002 The role of pollinator-mediated reproductive isolation in plant speciation: phylogeny, ecology and floral evolution in Antirrhinum (Veronicaceae). Ph.D. dissertation, Harvard University, Cambridge, Massachusetts, USA
Posada D. K. A. Crandall 1998 MODELTEST: testing the model of DNA substitution. Bioinformatics 14: 817-818
Reveal J. L. W. S. Judd R. G. Olmstead 1999 Proposal to conserve the name Antirrhinaceae against Plantaginaceae (Magnoliophyta). Taxon 48: 182[CrossRef]
Rothmaler W. 1956 Taxonomische Monographie der Gattung Antirrhinum. Akademie-Verlag, Berlin, Germany
Schwarz-Sommer Z. P. Huijser W. Nacken H. Saedler H. Sommer 1990 Genetic-control of flower development by homeotic genes in Antirrhinum majus. Science 250: 931-936
Sutton D. A. 1988 A revision of the tribe Antirrhineae. Oxford University Press, New York, New York, USA
Swofford D. L. 2001 PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4. Sinauer, Sunderland, Massachusetts, USA
Szymanski D. B. R. A. Jilk S. M. Pollock M. D. Marks 1998 Control of GL2 expression in Arabidopsis leaves and trichomes. Development 125: 1161-1171[Abstract]
Tamura K. M. Nei 1993 Estimation of the number of nucleotide substitutions in the control region of mitochondrial-DNA in humans and chimpanzees. Molecular Biology and Evolution 10: 512-526[Abstract]
Templeton A. R. 1983 Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution 37: 221-244[CrossRef][ISI]
Thompson D. M. 1988 Systematics of Antirrhinum (Scrophulariaceae) in the New World. Systematic Botany Monographs 22: 1-142
Tutin T. G. V. H. Heywood N. A. Burges D. M. Moore D. H. Valentine S. M. Walters D. A. Webb 1972 Flora Europaea. Cambridge University Press, Cambridge, UK
Vargas P. B. G. Baldwin L. Constance 1998 Nuclear ribosomal DNA evidence for a western North American origin of Hawaiian and South American species of Sanicula (Apiaceae). Proceedings of the National Academy of Sciences, USA 95: 235-240
Wettstein R. V. 1891 Scrophulariaceae. In A. Engler and K. Prantl [eds.], Die natürlichen Pflanzenfamilien, 39–107. Verlag, Leipzig, Germany
White T. J. T. Bruns S. Lee J. Taylor 1990 Amplification and direct sequencing of fungal ribosomal RNA genes for phylogenetics. In M. A. Innis, D. Gelfand, J. Sninsky, and T. White [eds.], PCR protocols: a guide to methods and applications, 315–322. Academic Press, San Diego, California, USA
Yang Z. B. 1994a Estimating the pattern of nucleotide substitution. Journal of Molecular Evolution 39: 105-111[ISI][Medline]
Yang Z. H. 1994b Maximum-likelihood phylogenetic estimation from DNA-sequences with variable rates over sites—approximate methods. Journal of Molecular Evolution 39: 306-314[CrossRef][ISI][Medline]
This article has been cited by other articles:
|
|
 |
|
  D. G. Howarth and M. J. Donoghue Phylogenetic analysis of the "ECE" (CYC/TB1) clade reveals duplications predating the core eudicots PNAS, June 13, 2006; 103(24): 9101 - 9106. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Schwinn, J. Venail, Y. Shang, S. Mackay, V. Alm, E. Butelli, R. Oyama, P. Bailey, K. Davies, and C. Martin A Small Family of MYB-Regulatory Genes Controls Floral Pigmentation Intensity and Patterning in the Genus Antirrhinum PLANT CELL, April 1, 2006; 18(4): 831 - 851. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. MATEU-ANDRES and L. DE PACO Allozymic Differentiation of the Antirrhinum majus and A. siculum Species Groups Ann. Bot., February 1, 2005; 95(3): 465 - 473. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||
