| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
(American Journal of Botany. 2008;95:1307-1327.) doi: 10.3732/ajb.0800065 © 2008 Botanical Society of America, Inc. |
What's this? |
Systematics and Phytogeography |
2 Department of Biology, University of Missouri-St. Louis, 1 University Boulevard, St. Louis, Missouri 63121 USA 3 The Arnold Arboretum, 22 Divinity Avenue, Harvard University, Cambridge, Massachusetts 02138 USA 4 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166-0299 USA
Received for publication 19 February 2008. Accepted for publication 11 August 2008.
ABSTRACT
The family Brassicaceae comprises 3710 species in 338 genera, 25 recently delimited tribes, and three major lineages based on phylogenetic results from the chloroplast gene ndhF. To assess the credibility of the lineages and newly delimited tribes, we sequenced an approximately 1.8-kb region of the nuclear phytochrome A (PHYA) gene for taxa previously sampled for the chloroplast gene ndhF. Using parsimony, likelihood, and Bayesian methods, we reconstructed the phylogeny of the gene and used the approximately unbiased (AU) test to compare phylogenetic results from PHYA with findings from ndhF. We also combined ndhF and PHYA data and used a Bayesian mixed model approach to infer phylogeny. PHYA and combined analyses recovered the same three large lineages as those recovered in ndhF trees, increasing confidence in these lineages. The combined tree confirms the monophyly of most of the recently delimited tribes (only Alysseae, Anchonieae, and Descurainieae are not monophyletic), while 13 of the 23 sampled tribes are monophyletic in PHYA trees. In addition to phylogenetic results, we documented the trichome branching morphology of species across the phylogeny and explored the evolution of different trichome morphologies using the AU test. Our results indicate that dendritic, medifixed, and stellate trichomes likely evolved independently several times in the Brassicaceae.
Key Words: approximately unbiased test • Arabidopsis • Brassica • Brassicaceae • ndhF • PHYA • phylogeny • trichomes
The Brassicaceae are uniquely placed in plant biology as a "model family" for evolutionary developmental studies. The potential of this model hinges on reliable developmental information, genomic data, and robust phylogenetic estimates. The first two components are well developed in Brassicaceae, due largely to the wealth of developmental and genomic data from Arabidopsis thaliana (L.) Heynh. Until recently, robust phylogenetic hypotheses for the family have been lacking. However, the publication of a familywide chloroplast ndhF phylogeny (Beilstein et al., 2006
) was an important step forward in providing a framework for future phylogenetic and evolutionary studies. Monophyletic groups inferred from the ndhF phylogeny also provided the foundation for a comprehensive new tribal classification of the family (Al-Shehbaz et al., 2006) that is gradually replacing Schulz’s (Schulz, 1936) highly artificial system. In addition, the ndhF phylogenetic analysis revealed that the majority of the newly delimited tribes belong to one of three large, monophyletic groups (lineages I–III, Beilstein et al., 2006).
More recently, Bailey et al. (2006) provided a familywide phylogenetic estimate based on the internal transcribed spacer of the ribosomal RNA locus (ITS) and Koch et al. (2007) inferred phylogeny from the trnL intron/trnL-F intergenic spacer. The ITS phylogeny is nearly congruent at the tribal level to that of Beilstein et al. (2006), although the tree is less resolved and thus some tribes are represented by multiple distinct monophyletic clades (Bailey et al., 2006). Neither the ITS nor trnL intron/trnL-F intergenic spacer phylogenies provide statistically supported nodes (i.e., bootstrap values >65%) beyond the tribal level. Thus, Bailey et al. (2006) also analyzed a supermatrix of 10 genes/gene regions, while Koch et al. (2007) built a supernetwork based on sequences from four different genes/gene regions to infer relationships beyond the tribal level. Both studies recovered some clades similar to those in Beilstein et al. (2006), although the methods used preclude rigorous assessment of clade support (e.g., the supermatrix comprised mostly missing data and the supernetwork analysis does not include an index of clade support).
In this study, we assess the credibility of the three hypothesized lineages (Beilstein et al., 2006) and test the monophyly of the recently erected tribes of the family (Al-Shehbaz et al., 2006) by adding phylogenetic information from partial sequences of the gene phytochrome A (PHYA) for species of Brassicaceae previously analyzed for the chloroplast gene ndhF. Phytochrome A is one of five phytochrome genes (PHYA–PHYE) in Arabidopsis (Clack et al., 1994). PHYA is 
50% similar to PHYC, its sister gene, and to PHYB and PHYE, allowing easy identification and locus-specific amplification (Clack et al., 1994; Mathews, 2006). The extensive characterization of the function and evolution of the phytochrome gene family in Arabidopsis thaliana (Møller et al., 2002; Franklin et al., 2003a, b; Monte et al., 2003; Sharrock et al., 2003a, 2003b) and more broadly in angiosperms and other land plants (Mathews, 2006) allows highly accurate assessment of orthology vs. paralogy of phytochrome sequences. Confidence in the homology of nucleotide sites determined during the alignment process is increased due to the amino acid and structural similarities that exist among land plant phytochrome genes (Mathews et al., 1995; Mathews and Sharrock, 1997). Furthermore, sequences from phytochrome genes have been used to infer phylogeny in Poaceae (Mathews and Sharrock, 1996; Mathews et al., 2000), Fabaceae (Lavin et al., 1998), Celastraceae (Simmons et al., 2001), Phyllanthaceae (Samuel et al., 2005), Malpighiaceae (Davis et al., 2002), and Orobanchaceae (Bennett and Mathews, 2006).
Based on results from Beilstein et al. (2006), trichome morphology was identified as an important character for determining phylogenetic affinities among Brassicaceae taxa. For example, most lineage I taxa have dendritic trichomes, except for those in the tribes Descurainieae (simple trichomes) and Physarieae (stellate trichomes), while lineage II taxa have only simple trichomes or are entirely glabrous. To compare trichome morphologies in species from other lineages and tribes, in this paper we document the morphology of trichomes from species across the resultant phylogeny of Brassicaceae and the published ndhF phylogeny (Beilstein et al., 2006) using scanning electron microscopy (SEM). In addition, we recorded the trichome morphology of all species sampled in the phylogeny to test hypotheses of trichome evolution. Trichomes in Brassicaceae consist of a single cell and are morphologically diverse, especially in regard to the number and position of branches (Beilstein and Szymanski, 2004). Simple trichomes are unbranched and occur throughout the family and in species of Cleomaceae, which is sister to Brassicaceae (Hall et al., 2002). Trichomes consisting of a pronounced stalk and two or more branches are termed dendritic and likely evolved numerous times in the family (Beilstein et al., 2006). In medifixed and stellate trichomes, the stalk is greatly reduced or absent; medifixed trichomes typically have only two branches, while stellate trichomes have three or more branches that radiate from a central point. In contrast to dendritic trichomes, the chloroplast ndhF analysis suggested a single origin for medifixed and stellate trichomes (Beilstein et al., 2006). Here we document similarities between trichome morphologies among closely and distantly related species. In addition, the increased phylogenetic information provided by PHYA data and the expanded sampling of species with stellate trichomes allow a more thorough investigation of the hypothesis that these forms evolved only once in the family.
MATERIALS AND METHODS
Taxon sampling
We replicated the taxon sampling of Beilstein et al. (2006) for the nuclear gene phytochrome A (PHYA) to compare family wide phylogenetic estimates of Brassicaceae from the nucleus and chloroplast and to explore the phylogenetic resolution provided by combining the two markers (Appendix 1). Additional taxa, not included in Beilstein et al. (2006), were added to the ndhF data set to achieve maximum overlap between the two markers. We were unable to obtain reliable PHYA sequence data for a few species sampled in the ndhF study. In total, we sampled 101 species in 90 genera across the family. Numerous studies have placed the family Cleomaceae sister to Brassicaceae (Rodman et al., 1996, 1998; Hall et al., 2002, 2004). In addition, the position of Aethionema as sister to all other Brassicaceae is not in doubt, having been demonstrated by Galloway et al. (1998), Koch et al. (2001, 2007), Hall et al. (2002), and Mathews and McBreen (in press). Because the sister group and the position of the root of the family are known, we rooted our analyses using only Polanisia dodecandra (L.) DC., a member of Cleomaceae. Taxa from all clades present in the ndhF phylogeny are represented in both the PHYA and combined data sets. The sample includes members of 23 of the 25 recently proposed tribes for the family (Al-Shehbaz et al., 2006). The tribes Cochlearieae and Iberideae were not included in this study or in the earlier ndhF study (Beilstein et al., 2006) because of the lack of either vouchered material (Cochlearieae) or reliable sequence data (Iberideae).
DNA extraction, PCR amplification, cloning, sequencing, and contig assembly
Silica-dried leaf material from collecting trips to Iran and China provided additional samples not included in Beilstein et al. (2006). DNA was isolated from silica-dried leaf tissue using a modified 2x CTAB protocol (Doyle and Doyle, 1987) and purified in cesium-chloride–ethidium-bromide gradients by ultracentrifugation. Sequencing of ndhF follows Beilstein et al. (2006). PHYA fragments were PCR amplified using the PHYA specific primers a230f and a832r (Table 1) with a step-down PCR protocol (Mathews and Donoghue, 2000) (Appendix S1, see Supplemental Data with the online version of this article). Amplification produced a distinct band of 2 kb in all accessions except Brassica oleracea L. and Hirschfeldia incana (L.) Lagr.-Foss., which yielded two bands of slightly different lengths and were cloned separately. Resulting PCR fragments were cloned and sequenced following the procedure outlined in Mathews et al. (2000). Six clones each were screened from a subset of taxa used in preliminary stages of the project, and a minimum of two clones was screened from all accessions. For a few taxa, as many as 10 clones were screened, and in the case of Schizopetalon rupestre (Barn.) Reiche, six clones each from two different PCR reactions were screened to eliminate labeling or pipetting error as an explanation for the alternative placement of S. rupestre in ndhF- and PHYA-inferred phylogenies. Additional sequencing primers were designed using the program PrimaClade (Gadberry et al., 2005), which predicts primers from aligned sequence. Sequenced PHYA fragments were trimmed using the program 4Peaks version 1.7 (A. Griekspoor and T. Groothuis, http://www.mekentosj.com/4peaks) prior to assembly to eliminate portions of the sequence in which Phred quality scores consistently fell below 20. Contigs for each sequenced clone were assembled in the program SeqManII version 4.0 (DNASTAR, Madison, Wisconsin, USA) and result from double-stranded overlap of at least 85%.
|
). Intron I of PHYA was trimmed from the aligned sequences based on the position of the intron in Arabidopsis thaliana. The resulting matrix contained 1764 nucleotide sites. Initial phylogenetic analyses included all sequenced clones (number of taxa [ntax] = 203) (TreeBASE accession M3965, http://www.treebase.org). PHYA data were pruned to a single clone per sequenced taxon, unless clones failed to form a monophyletic group in the initial phylogenetic analyses. The single clone chosen was that on the shortest branch of the monophyletic group of clones of the same taxon; this sequence is more similar to the likely ancestral PHYA sequence represented by the node of the monophyletic group of clones. Data sets resulting from this initial pruning were used to infer the PHYA phylogeny (ntax = 114). However, further pruning was required to achieve complete taxon overlap between ndhF (TreeBASE accession M3966) and PHYA data sets. Thus, taxa without a corresponding ndhF sequence were eliminated from the PHYA data set, resulting in a combined ndhF/PHYA matrix of 3851 nucleotide sites (ntax = 105) (TreeBASE accession M3967).
Parsimony, likelihood (Felsenstein, 1973), and Bayesian MCMC (Yang and Rannala, 1997) phylogenetic analyses were performed on the Beowulf cluster Expedition at the University of Missouri—St. Louis (UMSL). Parsimony ratchet (Nixon, 1999) searches consisting of 20 independent replicates of 200 iterations with 15% of characters reweighted per iteration were scripted using the program PAUPRat (Sikes and Lewis, 2001) and run in the program PAUP* 4.0b10 (Swofford, 2002). Gaps were considered missing data. For likelihood and Bayesian analyses, model parameters were set to those indicated by the program Modeltest 3.6 (Posada and Crandall, 1998). Likelihood runs used PAUP* (random sequence addition, tree-bisection-reconnection [TBR] swapping, MULTREES = yes), while Bayesian analyses (2 independent runs of 10 million generations each, sampling every 1000 generations) were implemented in MrBayes 3.1 (Ronquist and Huelsenbeck, 2003). Bayesian analyses performed on the combined data set specified two partitions corresponding to ndhF and PHYA fragments and allowed model parameters of each partition to change independently (mixed model).
Maximum likelihood bootstrap (LB; Felsenstein, 1985), parsimony bootstrap (PB), and Bayesian posterior probabilities (PP) were generated to assess the support for nodes within the resulting phylogenies, whereas only PB and PP values were generated for the combined data due to the computational intensity of generating LB values. Likelihood bootstrap replicates (100) were run in parallel on the Beowulf cluster using PAUP* (random sequence addition, TBR swapping, MULTREES = yes). Parsimony bootstrap replicates (500 bootstrap replicates, 1 random addition, TBR swapping, MULTREES = yes, saving no more than 1000 trees per replicate) were implemented in PAUP*. Bayesian posterior probabilities were obtained from the majority-rule consensus of trees generated in MrBayes 3.1.
Likelihood topology tests
The approximately unbiased test (AU) (Shimodaira, 2002), as implemented in the program CONSEL (Shimodaira and Hasegawa, 2001) (10000 bootstrap replicates), was used to determine the statistical significance of differences in topologies generated by the data sets (ndhF, PHYA, and ndhF/PHYA combined) or by enforcing topological constraints to test specific evolutionary hypotheses. For example, the AU test was used to explore whether the assembled data sets contained sufficient phylogenetic signal to address the monophyly of recently erected tribes in the family (Al-Shehbaz et al., 2006). Thus, when tribes were not resolved as monophyletic in phylogenies inferred from either the PHYA or the combined ndhF/PHYA data sets, we tested whether the data were sufficient to statistically reject monophyly. This represents a conservative approach to the proposition of nonmonophyly for tribes in the family, which would require revision of the current tribal scheme, by accounting for uncertainty in phylogeny reconstruction.
Several species had disparate PHYA sequences that were not sister to each other in the PHYA trees, which required additional pruning of the PHYA data to compare it to the ndhF and combined data set trees. For instance, Idahoa scapigera (Hook.) A. Nelson & J. F. Macbr. and Sisymbriopsis yechengnica (C. H. An) Al-Shehbaz, C. H. An & G. Yang, two putative hybrid species, are represented in the PHYA and combined data sets by two clones that are not sisters, whereas the ndhF phylogeny contains only a single representative of these taxa. For both species, the putative maternal (mat) PHYA copy occurs in the same position in the phylogeny as the ndhF sequence (I. scapigera [mat] and S. yechengnica [mat]), while the putative paternal (pat) copy occurs in a different position than the ndhF sequence (I. scapigera [pat] and S. yechengnica [pat]). Thus, the putative paternal copies of I. scapigera and S. yechengnica were pruned from the PHYA parsimony and combined data sets. Similarly, most Schizopetaleae taxa are represented in the PHYA data set by two nonmonophyletic clones, and one clone each was eliminated to achieve a single-clone data set for comparison to the ndhF tree and concatenation with ndhF in the combined data set.
Following the pruning of clones from the PHYA data, we tested whether differences between the ndhF and PHYA topologies generated by full heuristic searches of the data sets were statistically significant. In addition, well-supported nodes from one data set were used as constraints in the inference of topologies under the other two data sets. For example, well-supported nodes inferred from PHYA analyses (thickened lines) were used to constrain likelihood searches of both the ndhF and ndhF/PHYA combined data sets. Furthermore, well-supported nodes from the analysis of ndhF (thickened lines) data were used to constrain likelihood searches of PHYA data. Although these analyses suggested that trees generated from the two data sets are statistically significantly different, examination of specific topological disagreement revealed that the two trees differed primarily in the placement of Schizopetaleae taxa and that the majority of other disagreements had only low bootstrap or posterior probability support (e.g., <60% PB or LB, <0.95 PP). Thus, we concatenated the data sets into a single data set for combined analyses. Furthermore, we tested the ndhF/PHYA combined data tree, and well-supported nodes (PB 
80%, PP 0.95) against the ndhF and PHYA data to insure the topology obtained reflected both markers.
For genera and species that were not monophyletic in unconstrained searches of PHYA and ndhF/PHYA combined data, we tested whether there was sufficient phylogenetic signal to reject monophyly. Topologies requiring the monophyly of relevant genera and species were generated and tested against unconstrained topologies. In addition, a constraint tree requiring monophyly of the Schizopetaleae excluding Schizopetalon rupestre was also generated for the PHYA data set to explore the effect of alternative placements of S. rupestre on the likely monophyly of other Schizopetaleae taxa.
The evolution of medifixed and stellate trichomes was examined by constraining the ndhF/PHYA combined data to require the monophyly of species that have medifixed or stellate trichomes. For example, to test whether medifixed trichomes could have resulted from a single evolutionary event, we generated a constraint tree requiring the monophyly of Erysimum capitatum (Douglas ex Hook.) Greene, Farsetia aegyptica Desv., and Rhammatophyllum erysimoides (Kar. & Kir.) Al-Shehbaz & O. Appel. Similarly, the hypothesis that stellate trichomes have a single origin was tested by generating a constraint tree in which Clypeola aspera Turrill, Fibigia suffruticosa Sweet, Physaria floribunda Rydb., and P. rosei (Rollins) O’Kane & Al-Shehbaz form a monophyletic group.
Trichome SEM
To document trichome morphology for the species studied here and to verify reports in the literature, we recorded trichome morphology for 44 of the species in the PHYA phylogeny and six species included in the previously published ndhF phylogeny (Beilstein et al., 2006), using SEM. Mature leaves from herbarium specimens were mounted directly on stubs. All stubs were sputter-coated with gold and viewed with SEM at either UMSL, Central Institute for the Deaf—Washington University (WU), or Harvard University Herbaria (HUH). Trichome images were either captured on Polaroid film (Eastman Kodak, Rochester, New York, USA) and scanned at high resolution (UMSL) or captured directly as digital images (HUH, WU). Image brightness and contrast were adjusted using Adobe Photoshop CS version 8.0 (Adobe Systems, San Jose, California, USA).
RESULTS
Characteristics of the phytochrome A and combined data sets
The PHYA sequence alignment used in all analyses consists of 1764 nucleotide sites (588 amino acid positions). The alignment spans the GAF domain, the region to which the chromophore binds (Mathews, 2006). This region varies in length near the site of chromophore binding and thus requires the introduction of one or more 3-bp indels to maintain amino acid alignment among the sampled taxa. This variation accounts for the majority of length difference among PHYA sequences. Outside the GAF domain, the alignment contains two indels of 3 bp each, and a third indel of 6 bp. Idahoa scapigera has the longest PHYA sequence (1755 bp), excluding the intron, while Lepidium alyssoides A. Gray has the shortest sequence (1716 bp). The combined data set consists of the PHYA sequence alignment detailed, plus 2087 bp of aligned ndhF data (Beilstein et al., 2006), for a total of 3851 aligned nucleotide sites (1283 amino acid positions). The results of Modeltest 3.6 indicated that the GTR + I + 
model of sequence evolution best described both the PHYA and combined data sets, whether evaluated by likelihood ratio test or AIC.
Clones originating from the same DNA accession formed a monophyletic group in phylogenetic analyses of PHYA in the majority of sampled taxa (Appendix S2, see Supplemental Data with the online version of this article), so a single clone was chosen to represent the taxon in further analyses. However, two distinct, nonmonophyletic copies of PHYA were recovered from Brassica oleracea, Caulanthus crassicaulis (Torr.) S. Wats., Hesperidanthus suffrutescens (Rollins) Al-Shehbaz, Hirschfeldia incana, Idahoa scapigera, Mostacillastrum orbignyanum (E. Fournier) Al-Shehbaz, Neuontobotrys elloanensis Al-Shehbaz, N. frutescens (Gills. ex Hook. & Arn.) Al-Shehbaz, Romanschulzia sp. O. E. Schulz, Sisymbriopsis yechengnica, Stanleya pinnata (Pursh) Britton, and Thelypodium laciniatum (Hook.) Endl. (Fig. 1). In B. oleracea and H. incana, clones varied in the length of the sequenced intron; the two B. oleracea introns differed by 427 bp, and the two H. incana introns differed by 405 bp. In contrast, intron length variation was not observed in other duplicated PHYA sequences; rather, alternative copies were cloned from PCR fragments of the same length.
|
Phylogenetic reconstructions and topology congruence
Maximum likelihood, parsimony ratchet and Bayesian phylogenetic analyses of the single clone PHYA (Fig.1), pruned PHYA (Fig. 2) and combined (Fig. 3) data sets produced topologies that largely agree with phylogenies inferred from ndhF (Fig. 2) (Table 2). In particular, the tribe Aethionemeae is sister to all other Brassicaceae, and three major lineages are recovered from PHYA and combined estimates of phylogeny (I–III, Figs. 1–3). Lineage I consists of the tribes Boechereae, Camelineae, Cardamineae, Descurainieae, Halimolobeae, Lepidieae, Physarieae, Smelowskieae, and Alysseae pro parte in phylogenies inferred from ndhF (Fig. 2) and combined (Fig. 3) data. Lineage I is not monophyletic in the maximum likelihood PHYA tree due to the placement of Alyssum canescens DC. and the tribe Cardamineae outside the lineage (Fig. 1), but these placements are poorly supported in the PHYA tree, and the monophyly of lineage I is not rejected by the PHYA data in likelihood topology tests (Table 3, Lineage I, AU P = 0.330). The tribes Brassiceae, Isatideae, and Sisymbrieae comprise a monophyletic group with the majority of Schizopetaleae species (lineage II) in ndhF, PHYA, and combined phylogenetic analyses. Similarly, all three data sets resolve the monophyly of lineage III, consisting of the tribes Anchonieae, Chorisporeae, Euclidieae, and Hesperideae.
|
|
|
|
Similarities in tribal monophyly are also reflected in topology tests generated from unconstrained heuristic searches of the ndhF, pruned PHYA, and combined data (Table 4). When phylogenetic searches of the combined data are constrained by well-supported nodes from either the ndhF or PHYA trees, the resulting topologies are not significantly different from the unconstrained tree (Table 4, ndhF well-supported nodes, AU P = 0.200; PHYA well-supported nodes, AU P = 0.409). However, the most likely ndhF (Table 4, ndhF, best) and PHYA (Table 4, PHYA, best) topologies differ significantly from trees constrained by nodes resolved in the other data set. For example, when phylogenetic searches of the PHYA data are constrained by the well-supported nodes of the ndhF phylogeny (thickened lines, Fig. 2), the likelihood of the resulting tree is significantly different from the unconstrained tree (Table 4, ndhF well-supported nodes, AU P = 0.000).
|
Aethionemeae
The PHYA, ndhF, and combined ndhF/PHYA data sets provide strong support for the sister relationship of the tribe Aethionemeae to all other tribes and taxa of Brassicaceae. The tribe, as sampled, is comprised of Aethionema saxatile and Moriera spinosa Boiss.
Alysseae
Alysseae are polyphyletic in trees inferred from all three data sets (Figs. 1–3) (Table 2). Monophyly of the tribe is not rejected by the PHYA data (Table 3, Alysseae, AU P = 0.061), but the AU test rejects the monophyly of Alysseae for the combined data (Table 5, Alysseae, AU P = 0.022). Polyphyly of Alysseae is due in part to paraphyly of Alyssum, monophyly of which is rejected by AU test (Table 3, Alyssum, AU P = 0.005). Alyssum linifolium Steph. Ex Willd., Clypeola aspera and Fibigia suffruticosa form a monophyletic group of core Alysseae in PHYA trees (Alysseae 2, Fig. 1); ndhF sequence data were not reliably amplified for A. linifolium, and thus this species was not included in the combined analysis. Alyssum canescens (Alysseae 1, Fig. 1) is sister to Arabideae in PHYA analyses (without support), is within lineage I in the ndhF tree (Fig. 2) and is sister to all other members of lineage I in the combined analysis. Farsetia aegyptica (Alysseae 3, Fig. 1) is strongly supported as sister to Lunaria annua L. in the PHYA analysis, in the same position but without support in the combined analysis and in an unresolved position in the ndhF analysis.
|
Arabideae
Arabideae are monophyletic in phylogenies inferred from ndhF, PHYA, and combined data. Within the tribe Aubrieta deltoidea and A. parviflora form a clade in all analyses. The tribe is not a member of any of the well-supported lineages defined previously, but is sister to lineage II in both PHYA and combined (Fig. 3) trees, although without support. In contrast, ndhF data place the tribe sister to a larger monophyletic group comprised of lineage II plus the unplaced tribes Eutremeae and Thlaspideae, as well as Goldbachia laevigata (M. Bieb.) DC.
Boechereae
Boechereae (lineage I) are monophyletic in all analyses (Figs. 1–3). Within the tribe, Boechera platysperma and B. shortii are monophyletic in all trees. Relationships within the tribe are largely resolved in ndhF (Fig. 2) and combined trees (Fig. 3) but not in the PHYA tree (Fig. 1).
Brassiceae
Brassiceae (lineage II) are monophyletic in all phylogenies. Brassica oleracea and Hirschfeldia incana are strongly monophyletic in PHYA analyses (Fig. 1), and together are sister to Cakile maritima Scop.; the latter relationship lacks statistical support in the PHYA tree, but is strongly supported by ndhF (Fig. 2) and combined data (Fig. 3). Brassiceae are sister to [Schizopetaleae + Sisymbrieae] in both ndhF and combined analyses.
Camelineae
Camelineae (Lineage I) are polyphyletic in the PHYA phylogeny (Figs. 1, 2). However, none of the sampled species of Camelineae is strongly supported as a member of other lineage I tribes, and the potential monophyly of Camelineae is not rejected by the PHYA data (Table 3, Camelineae, AU P = 0.070). Arabidopsis thaliana and A. lyrata (L.) O’Kane & Al-Shehbaz form a monophyletic Arabidopsis (Camelineae 1, Fig. 1) sister to species of Physaria. Camelina microcarpa Andrz. ex DC., Capsella bursa-pastoris (L.) Medik. and Catolobus pendula (L.) Al-Shehbaz are also monophyletic (Camelineae 2) and sister to other members of Physarieae, excluding Physaria. Olimarabidopsis pumila (Stephan) Al-Shehbaz, O’Kane & R. A. Price and Turritis glabra L. (Camelineae 3) are sister to the Boechereae-Halimolobeae clade, while Erysimum capitatum (Camelineae 4) is sister to members of the Descurainieae.
The polyphyly of Camelineae in the PHYA phylogenetic analyses contrasts with the strong support for their monophyly in the ndhF analysis (Fig. 2). They are also monophyletic in the combined analysis (Fig. 3), with strong support in Bayesian analyses (PP 1.0) but with lower bootstrap support (PB 59%) than in the ndhF tree.
Cardamineae
Cardamineae are monophyletic in all analyses. Within Cardamineae, ndhF (Fig. 2) and combined (Fig. 3) data place Barbarea vulgaris R. Br. and Planodes virginicum Greene in a clade sister to the clade formed by Cardamine pulchella (Hook. f. & Thomson) Al-Shehbaz & G. Yang and Iodanthus pinnatifidus (Michx.) Steudel. In contrast, relationships within Cardamineae are not statistically supported in the PHYA analysis (Fig. 1). Cardamineae are members of lineage I in ndhF and combined trees, but not in the PHYA analysis. However, monophyly of lineage I is not rejected by the PHYA data (Table 3), and the PHYA parsimony tree places Cardamineae in lineage I (Fig. 2).
Chorisporeae
Chorisporeae (lineage III) are represented by Chorispora tenella, which is sister to Dontostemon senilis (Anchonieae) in the PHYA (Fig. 1) phylogenetic analysis; their position relative to one another is unresolved in the ndhF analysis presented here (Fig. 2). The C. tenella-D. senilis clade is sister to the rest of lineage III in PHYA and combined trees and is strongly supported by combined data (PP 1.0, PB 99%), but lacks support from PHYA alone and is not recovered in the ndhF tree.
Descurainieae
Descurainieae (lineage I) are not monophyletic in the PHYA analysis, although potential monophyly of the tribe is not rejected (Table 3, Descurainieae, AU P = 0.548). In the PHYA tree (Fig. 1), Hornungia procumbens (L.) Hayek is sister to the sampled Lepidieae rather than to other members of Descurainieae. Similarly, Descurainieae are not monophyletic in the Bayesian analysis of combined data (Fig. 3), but are monophyletic in the likelihood analysis (tree not shown). The ndhF data place H. procumbens sister to [Descurainia sophia (L.) Webb + Ianhedgea minutiflora (Hook. f. & Thomson) Al-Shehbaz & O’Kane], thereby forming a monophyletic Descurainieae (Fig. 2). Regardless of the exact position of H. procumbens, all sampled Descurainieae are members of lineage I in all trees.
Euclidieae
Euclidieae sensu lato (lineage III) are strongly monophyletic in all analyses. Euclidieae s.l. includes all sampled members of Euclidieae s.s. plus Christolea crassifolia Cambes., Dilophia salsa Thomson, Shangrilaia nana Al-Shehbaz, J. P. Yue & H. Sun, and Sisymbriopsis yechengnica. Leiospora eriocalyx (Regel & Schmalh.) F. Dvorák is sister to Euclidieae s.l. in PHYA and combined phylogenies, but falls in an unresolved position in lineage III in ndhF trees.
Eutremeae
Eutremeae are monophyletic in all analyses except the pruned PHYA parsimony analysis (Fig. 2). Eutrema heterophyllum (W. W. Sm.) H. Hara and E. altaicum (C. A. Mey.) Al-Shehbaz & Warwick are sister species in PHYA (Fig. 1) and combined (Fig. 3) topologies, while ndhF data support the sister relationship of Chalcanthus renifolius Boiss. and E. altaica (Fig. 2). The tribe is derived from within a paraphyletic Thlaspideae in the PHYA likelihood tree (Fig. 1) and is sister to Thlaspideae in the combined tree, although both relationships lack statistical support.
Halimolobeae
Halimolobeae (Lineage I) are consistently monophyletic. They are sister to Boechereae with good support in PHYA (Fig. 1) and combined (PP 1.0, PB 100)(Fig. 3) analyses, but in the ndhF tree, the relationships between the two are unresolved (Fig. 2).
Heliophileae
The single accession of Heliophileae, Heliophila Burm. f. ex L. sp., forms a clade with Asta schaffneri (S. Wats.) O. E. Schulz in likelihood PHYA (Fig. 1) and combined (Fig. 3) trees; Idahoa scapigera is included in this clade in the combined analysis (Fig. 3), but without statistical support. In both the likelihood PHYA tree (Fig. 1) and the pruned parsimony PHYA tree (Fig. 2), Schizopetalon rupestre also falls in this clade with some statistical support. The ndhF topology places Heliophila sp. sister to [Noccaeeae + Conringia persica Boiss.], but without support (Fig. 2).
Hesperideae
Hesperideae (Hesperis matronalis L. and Hesperis sp. nov., lineage III) are monophyletic in all phylogenetic analyses. The tribe is sister to Bunias orientalis in PHYA (Fig. 1) and combined (Fig. 3) topologies, but with little support. The latter relationship is not supported by ndhF data (Fig. 2).
Isatideae
Isatis tinctoria L. and Myagrum perfoliatum L. comprise the strongly supported monophyletic Isatideae (lineage II) in all analyses. They are sister to all other lineage II tribes in the ndhF (Fig. 2) and combined (Fig. 3) trees, but not in the PHYA tree (Fig. 1), which is less resolved within lineage II than the ndhF and combined trees.
Lepidieae
Lepidieae (Lepidium alyssoides and L. draba L., lineage I) are monophyletic in all analyses. The tribe is sister to Hornungia procumbens (Descurainieae 2) in phylogenies inferred from PHYA (Fig. 1) and combined (Fig. 3) data. In contrast, in the ndhF tree Lepidieae are sister to Cardamineae (Fig. 2). Neither placement is statistically supported.
Noccaeeae
Noccaeeae are monophyletic and are strongly supported as sister to Conringia persica in all analyses. The relationship of (Conringia + Noccaeeae) to other tribes of the family is unresolved.
Physarieae
Physarieae (lineage I) are monophyletic in phylogenies inferred from ndhF and combined data, but not in phylogenies inferred from PHYA data (Figs. 1 and 2). There, Physaria floribunda and P. rosei are resolved as sister (Physarieae 1), but are more closely related to Camelineae 1 than to Dimorphocarpa wislizenii (Engelm.) Rollins, Nerisyrenia johnstonii J. D. Bacon, and Synthlipsis greggii A. Gray (Physarieae 2), but with little support. Lineage I tribes Camelineae, Boechereae, Halimolobeae, and Physarieae form a well-supported clade in the ndhF (Fig. 2) and combined (Fig. 3) trees, with Physarieae sister to the other three tribes. The potential monophyly of Physarieae is not rejected by PHYA data (Table 3, Physarieae, AU P = 0.423).
Schizopetaleae
Schizopetaleae (lineage II) are monophyletic in phylogenies inferred from ndhF and combined data. The tribe is closely related to sampled members of Sisymbrieae in all analyses (Figs. 1–3). In the phylogeny inferred from PHYA data, all Schizopetaleae except Schizopetalon rupestre (Schizopetaleae 2, Fig. 1) form a large monophyletic group (Schizopetaleae 1, Fig. 1), but S. rupestre is sister to the clade formed by Heliophila sp. and Asta schaffneri. Furthermore, all sampled Schizopetaleae, except Hesperidanthus jaegeri and Streptanthus squamiformis Goodman, have two copies of PHYA (i and ii in Schizopetaleae 1, Fig. 1) that form reciprocally monophyletic groups of sequences. AU test results (Table 3) reject monophyly for the two sampled PHYA copies of Caulanthus crassicaulis (AU P = 0.015), Hesperidanthus suffrutescens (AU P = 0.005), Mostacillastrum orbignyanum (AU P = 0.000), Neuontobotrys elloanensis (AU P = 0.034), N. frutescens (AU P = 0.029), and Stanleya pinnata (AU P = 0.002). In contrast, monophyly for the two copies of Romanschulzia sp. (AU P = 0.084), and Thelypodium laciniatum (AU P = 0.367) is not rejected.
Hesperidanthus and Neuontobotrys are each represented by two species in the PHYA, ndhF, and combined data sets. Hesperidanthus is monophyletic in the combined tree with high posterior probability (PP 1.0), but with low bootstrap support (PB < 50%) (Fig. 3). The clone 1 copies of Hesperidanthus jaegeri and H. suffrutescens are monophyletic in the PHYA tree, although without statistical support (Fig. 1). The relationship between the two species is unresolved in the ndhF tree, and the two are not monophyletic in the pruned PHYA tree (Fig. 2). In contrast, Neuontobotrys elloanensis and N. frutescens are never monophyletic in any analysis. Instead, N. frutescens forms a well-supported monophyletic clade with Mostacillastrum orbignyanum and Schizopetalon rupestre in the ndhF tree (Fig. 2) and is monophyletic with M. orbignyanum in the combined tree (PP 1.0, PB 92%) (Fig. 3).
When PHYA data are pruned to a single copy per accession for comparison with ndhF, removing the Schizopetaleae 2 copy for all species with two copies, Streptanthus squamiformis, from which only a single copy was recovered, falls outside the Schizopetaleae, yet remains firmly placed within lineage II (Fig. 2). In contrast, Schizopetalon rupestre and Streptanthus squamiformis are sister to other Schizopetaleae in the combined tree (supported in Bayesian analyses only) (Fig. 3).
Sisymbrieae
Sisymbrium altissimum L. is supported as sister to S. linifolium Nutt., in ndhF (Fig. 2) and combined (Fig. 3) trees, and together they form a monophyletic Sisymbrieae (lineage II), sister to Schizopetaleae. In contrast, the two species are not sisters in the PHYA tree, but form a grade leading to Schizopetaleae (Figs. 1 and 2). However, PHYA data do not reject monophyly for Sisymbrieae (Table 3, Sisymbrieae, AU P = 0.630). All data sets place Sisymbrieae, however circumscribed, in lineage II.
Smelowskieae
Smelowskieae (lineage I) are monophyletic in all analyses. All trees support the sister relationship of Smelowskia tibetica (Thomson) Lipsky and S. calycina (Stephan ex Willd.) C. A. Mey (Figs. 1–3). Smelowskia annua Rupr. is sister to the clade formed by the other two species.
Thlaspideae
Thlaspideae are monophyletic in ndhF and combined analyses, but not in PHYA analyses, although monophyly of the tribe is not rejected by the PHYA data (Table 3, Thlaspideae, AU P = 0.500). In the PHYA phylogeny, Alliaria petiolata (M. Bieb.) Cavara & Grande and Thlaspi arvense L. are sister taxa (Thlaspideae 1, Fig. 1), but Pseudocamelina campylopoda Bornm. & Gauba ex Bornm. (Thlaspideae 2) is sister to the clade that includes Eutremeae, Thlaspideae 1, and Goldbachia laevigata (support for most of these relationships is weak).
Unplaced taxa
Asta schaffneri, Biscutella didyma L., Cremolobus subscandens Kuntze, Idahoa scapigera, and Lunaria annua are not included in any of the tribes previously described due to the lack of phylogenetic resolution in ndhF analyses (Fig. 2). In the PHYA tree, the two copies of Idahoa scapigera occur in different positions, and each placement receives some statistical support (Fig. 1). One copy is sister to Cremolobus subscandens, a relationship similar to that in the ndhF tree (Fig. 2). The second copy of I. scapigera PHYA forms a clade with Asta schaffneri, Heliophila sp., and Schizopetalon rupestre, a relationship not recovered in the ndhF tree. The branch lengths of all of these taxa are relatively long, while the branches supporting relationships among these taxa are relatively short (Fig. 1). In addition, monophyly for the two copies of I. scapigera is not rejected by PHYA data (Table 3, Idahoa, AU P = 0.111). The conflicting signal for the placement of A. schaffneri, C. subscandens, Heliophila sp., and I. scapigera is apparent from the lack of support for the monophyly of this group in the combined analysis (Fig. 3). A similar situation occurs in efforts to place Lunaria annua, which is sister to Farsetia aegyptica in the PHYA and combined trees, but is sister to B. didyma, with low support, in ndhF trees.
Trichome SEM and evolutionary hypothesis testing
Trichome producing species sampled in both the phylogenetic and SEM studies were classified as having either simple, dendritic, medifixed, or stellate trichomes (Fig. 3). Simple trichomes grow away from the epidermal surface and are unbranched. Dendritic trichomes have a distinct stalk that grows away from the epidermal surface, and tree-like branches, some of which continue to grow vertically, away from the epidermal surface. In medifixed and stellate trichomes, the stalk is greatly reduced, or in some cases absent, and the branches grow parallel to the epidermal surface. Medifixed trichomes have only two branching arms that grow away from the point of attachment to the epidermis. In contrast, stellate trichomes can have as many as 30 branches and these emanate from a central point. In some stellate trichomes, the branches are themselves branched, but the trichome remains nearly radially symmetric.
To document more fully the trichome morphology across Brassicaceae, we used SEM to examine the trichomes of several species not sampled in the current phylogenetic analyses, but which are robustly resolved in tribes based on the previously published ndhF tree (Beilstein et al., 2006). These taxa include Anelsonia eurycarpa (A. Gray) J. F. Macbr. & Payson and Phoenicaulis cheiranthoides Nutt. (Boechereae), Dontostemon senilis (Anchonieae), Lobularia maritima (L.) Desv. (Alysseae), Stenopetalum nutans F. Muell. (Camelineae), and Sterigmostemum acanthocarpum (Fisch. & C. A. Mey) Kuntze. In addition, 45 taxa included in the current phylogenetic study were examined.
Species from different lineages and tribes produce trichomes of similar morphology. For example, simple and dendritic trichomes occur in species from all three lineages (Figs. 4–10). Medifixed and stellate trichomes are less common, although they also occur in species from different lineages and tribes (Fig. 3). The hypothesis that medifixed trichomes evolved once in the family is rejected by the combined data in AU topology tests (Table 5, Medifixed trichomes evolved once, AU P = 0.000). Similarly, the combined data reject a single origin for stellate trichomes (Table 5, Stellate trichomes evolved once, AU P = 0.000). In addition to likelihood hypothesis testing, we also explored the evolution of trichomes by reconstructing ancestral character states using likelihood, which indicated that simple, dendritic, medifixed, and stellate trichomes each evolved more than once in Brassicaceae (Appendix S3, see Supplemental Data with the online version of this article).
|
|
Species lacking trichomes
All the sampled Aethionemeae, Cardamineae, Eutremeae, and Noccaeeae lack trichomes. In addition, the majority of sampled Schizopetaleae are glabrous, with Schizopetalon rupestre (dendritic trichomes) being a notable exception.
Simple trichomes
Species with simple trichomes include lineage I taxa Smelowskia tibetica (Smelowskieae), which also has dendritic trichomes (Fig. 6B), and Lepidium alyssoides (Lepidieae) (Fig. 6H); and lineage II taxon Sisymbrium altissimum (Sisymbrieae) (Fig. 9B). Numerous lineage III species have simple trichomes, including Dontostemon senilis (Anchonieae) (Fig. 7A); Chorispora tenella and Diptychocarpus strictus Trautv. (Chorisporeae) (Fig. 7F, G); and Christolea crassifolia, Desideria linearis (N. Busch) Al-Shehbaz, and Sisymbriopsis yechengnica (Euclidieae) (Fig. 8A, B, H). In addition, Biscutella didyma and Cremolobus subscandens (Fig. 9C, D) are not included in any of the lineages or tribes and have simple trichomes.
|
|
|
|
|
Lineage III species with dendritic trichomes include Matthiola farinosa, M. integrifolia, Oreoloma violaceum, and Sterigmostemum acanthocarpum (Anchonieae) (Fig. 7B–E); Hesperis matronalis (Hesperideae) (Fig. 7H); Euclidium syriacum (L.) R. Br., Neotorularia korolkowii (Regel & Schmalh.) Hedge & J. Léonard, Sisymbriopsis mollipila (Maxim.) Botsch., Strigosella africana Botsch., and Tetracme pamirica Vassilcz. (Euclidieae) (Fig. 8C–E, G, I).
Taxa not included in any of the three lineages but which have dendritic trichomes include Arabis alpina L., Aubrieta deltoidea, and Baimshania pulvinata Al-Shehbaz (Arabideae) (Fig. 10A–C); and Alyssum canescens (Alysseae) (Fig. 10G).
Medifixed trichomes
Species with medifixed trichomes include lineage I taxa Erysimum capitatum and Stenopetalum nutans F. Muell. (Camelineae) (Fig. 4E, F), lineage III taxon Rhammatophyllum erysimoides (Euclidieae) (Fig. 8F), and Farsetia aegyptica and Lobularia maritima (Alysseae) (Fig. 10H, I).
Stellate trichomes
Species with stellate trichomes include lineage I taxa Physaria floribunda and P. rosei (Physarieae) (Fig. 6F, G), and Clypeola aspera, Fibigia suffruticosa, Alyssum linifolium (Fig. 10D–F) (Alysseae), which do not fall within any of the three major lineages.
DISCUSSION
Data from the nuclear marker PHYA further our understanding of phylogenetic relationships in Brassicaceae by increasing confidence in the lineages and tribes inferred from the chloroplast marker ndhF. Aethionemeae are sister to all other Brassicaceae, as in earlier studies (Galloway et al., 1998; Koch et al., 2001). More importantly, data from ndhF and PHYA support the recognition of three lineages in the family, each consisting of several tribes (lineages I–III, Figs. 1–3). These lineages are the only well-supported groups above the level of tribe in any family level phylogenetic study to date. Furthermore, the support for all three lineages is unique to this study. The three lineages occur but receive <50% bootstrap support in the trnL/trnF phylogenetic analysis (Koch et al., 2007). Both lineages I and II are resolved in the supermatrix analysis of Bailey et al. (2006), but only lineage I receives consensus bootstrap support >50%. Confidence in the monophyly of 13 tribes (Aethionemeae, Arabideae, Boechereae, Brassiceae, Cardamineae, Euclidieae, Eutremeae, Halimolobeae, Hesperideae, Isatideae, Lepidieae, Noccaeeae, and Smelowskieae) is increased as a result of the ndhF–PHYA analyses. In contrast, the monophyly of four tribes (Camelineae, Descurainieae, Physarieae, and Schizopetaleae) differs between the ndhF and PHYA phylogenies, and thus these tribes require future phylogenetic study.
Tribal delimitations
Most tribes in PHYA and combined phylogenetic analyses are monophyletic and thus do not disagree with phylogenies inferred from ndhF data alone. In contrast, several tribes are not monophyletic in the PHYA and combined phylogenetic analyses, suggesting that the taxonomy of these tribes requires careful reconsideration.
Lineage I
Camelineae, Boechereae, Halimolobeae, and Physarieae are each monophyletic in ndhF and combined phylogenies, and together they form a well-supported clade, with Physarieae sister to the other three tribes. Physarieae are monophyletic in the ITS phylogenetic analyses (Bailey et al., 2006), and members of the tribe produce pollen with more than three colpi per pollen grain, a form unique in the family. PHYA data do not reject the potential monophyly of Physarieae. In contrast, Camelineae are not monophyletic in either the ITS or supermatrix tree of Bailey et al. (2006), although the species of Camelineae sampled are not resolved as members of other tribes. While Camelineae are not supported as monophyletic in PHYA trees, Camelina microcarpa, Capsella bursa-pastoris, and Catolobus pendula form a strongly supported monophyletic group in ndhF, PHYA, and combined trees (Figs. 1–3). Similarly, the genus Arabidopsis is monophyletic in PHYA and all other family level phylogenetic studies (Bailey et al., 2006; Beilstein et al., 2006; Koch et al., 2007).
The failure of Camelineae to form a monophyletic group in PHYA and ITS phylogenies contrasts with the strong support for the monophyly of the tribe in phylogenies generated from ndhF data. Incongruence between nuclear (PHYA, ITS) and chloroplast (ndhF) phylogenies could result either from incomplete lineage sorting of nuclear gene alleles in the case of PHYA or incomplete ribosomal gene conversion in ITS. Alternatively, a history of hybridization and introgression between members of Camelineae, Physarieae, or other lineage I taxa could lead to discordant plastid and nuclear phylogenies. However, because the monophyly of Camelineae is not rejected by PHYA data (Table 3), additional sampling may still confirm the monophyly of the tribe. Whatever process is leading to the different phylogenetic results between sampled nuclear and chloroplast markers, the tribe requires additional data to elucidate relationships among its members and thus to infer the closest relatives of Arabidopsis.
Lineage II
The monophyly of lineage II, comprising Brassiceae, Isatideae, Schizopetaleae, and Sisymbrieae, is well established in the ndhF, PHYA, and combined ndhF/PHYA phylogenies, although the markers differ in regard to the monophyly of Schizopetaleae and Sisymbrieae. The placement of Schizopetalon rupestre outside lineage II makes Schizopetaleae paraphyletic in PHYA phylogenies, but it is supported as monophyletic in ndhF phylogenies. Neither the supermatrix nor ITS data set (Bailey et al., 2006) includes S. rupestre, precluding comparisons. The tribe, excluding S. rupestre, is monophyletic in PHYA trees. Thus, S. rupestre is the only statistically significant point of disagreement between ndhF and PHYA phylogenies for the tribe (Table 3). Except for Pringlea antiscorbutica R. Br. ex Hook. f. (not sampled here), which is restricted to islands in the South Indian Ocean, species in the tribe are distributed only in the Americas (Al-Shehbaz et al., 2006). Floral morphology is the most diverse of any tribe in the family and includes variation in filament length, presence vs. absence of a gynophore, channeled or crisped petals, and erect sepals that form a floral tube, especially in Streptanthus and Caulanthus (Al-Shehbaz et al., 2006). The species of Schizopetalon are restricted to southern reaches of the Andes and produce flowers with highly divided petals and a corolla tube formed by the erect sepals. Thus, both the distribution and floral morphology of S. rupestre suggest the species is a member of the Schizopetaleae. In contrast, species of Schizopetalon differ from other sampled Schizopetaleae taxa by producing dendritic, rather than simple, trichomes (Fig. 9A). Either the ndhF or PHYA sequence could be a sequencing error, but additional accessions of S. rupestre, and other species of the genus, are required to test this possibility. A better understanding of the limits of Schizopetaleae (sensu Al-Shehbaz et al., 2006) requires additional sampling of Schizopetalon Sims, and the putative sister genus Mathewsia Hook. & Arn.
Sisymbrieae include about 40 species, all of which are now placed in Sisymbrium. Sisymbrieae have terete fruits and simple trichomes (Fig. 9B) and are distributed primarily in Eurasia and Africa (Al-Shehbaz et al., 2006).The ndhF and combined data fully agree with the ITS and trnL-F sequence data (Warwick et al., 2002, 2006) that suggested reduction of Schoenocrambe (formerly Schoenocrambe linifolia (Nutt.) Greene) (Warwick and Al-Shehbaz, 2003) to synonymy of Sisymbrium, making S. linifolium the only member of the genus and tribe native to North America. Note that other species formerly placed in Sisymbrium, including North American taxa, have been transferred to genera of Schizopetaleae (Warwick et al., 2006).
The PHYA data indicate a history of duplication events in lineage II taxa. Two monophyletic groups of PHYA sequences were found among species in the tribe Schizopetaleae (excluding Schizopetalon rupestre) (1 and 2, Fig. 1), and topology tests forcing monophyly for clones recovered from individual species are significantly less likely than the unconstrained tree for most Schizopetaleae species (Table 3). In contrast to other sampled Schizopetaleae, both clones of Neuontobotrys elloanensis are in the same monophyletic group and thus could be evidence of either a species-specific duplication event or of additional duplication events in the history of Schizopetaleae that were either lost or not recovered from other species of the tribe. When Schizopetaleae PHYA clade 1 sequences are removed for comparisons with ndhF trees, Hesperidanthus jaegeri, from which only a single PHYA copy was recovered, falls outside the Schizopetaleae but remains firmly placed within lineage II. Similarly, when Schizopetaleae PHYA clade 2 is removed, the species Streptanthus squamiformis, also represented by a single clone, falls outside Schizopetaleae, but remains a member of lineage II, indicating that paralogous copies of PHYA are sampled from S. squamiformis and H. jaegeri and thus explaining the nonmonophyly of Schizopetaleae (excluding S. rupestre). The paralogous copies may be the result of differential gene copy loss in S. squamiformis and H. jaegeri or may reflect a failure to amplify additional, orthologous PHYA copies from these taxa.
In another example, Brassica oleracea and Hirschfeldia incana, members of the Brassiceae, are represented in the PHYA phylogeny by two nonmonophyletic clones. In this case, each B. oleracea clone is sister to a clone of H. incana (Fig. 1) (online Appendix S2). The presence of at least two copies of PHYA in B. oleracea and H. incana is consistent with evidence from comparative mapping and chromosome painting experiments that indicate a genome triplication event early in the history of the tribe Brassiceae (17–18 Mya) (Lagercrantz, 1998; Lysak et al., 2005; Parkin et al., 2005). Interestingly, the branch lengths of these clones are the longest of any of the sampled taxa, suggesting an accelerated rate of evolution. Conversely, the two PHYA clones of Cakile maritima, also a member of Brassiceae, form a monophyletic group sister to the duplicated copies of B. oleracea and H. incana and have branch lengths similar to those of other sampled taxa (Fig. 1). In chromosome painting studies (Lysak et al. 2005), C. maritima shows evidence of the triplication event that characterizes other Brassiceae. Thus, if C. maritima contains more divergent copies of PHYA, they were not among the sequenced clones, and the sequenced copies of C. maritima PHYA are apparently evolving more slowly than those of B. oleracea and H. incana. Alternatively, C. maritima may have undergone gene loss following the triplication event. Gene loss has been documented following polyploidization in both B. rapa and B. oleracea (Town et al., 2006; Yang et al., 2006).
Lineage III
Lineage III is a primarily Asian radiation whose members have been largely omitted from other phylogenetic studies of Brassicaceae. The lineage contains Anchonieae, Chorisporeae, Euclidieae, and Hesperideae in all analyses; support is higher in the combined analysis than with either gene alone. However, Anchonieae sensu Al-Shehbaz et al. (2006) is not monophyletic because Chorispora tenella (Chorisporeae) and Dontostemon senilis (Anchonieae 2) form a strongly supported clade in the PHYA and combined trees (Figs. 1–3), not immediately related to Anchonieae 1. Diptychocarpus strictus (Chorisporeae) also falls in this clade in the published ndhF tree (Beilstein et al., 2006), but is not included in the current analyses. All three species have exclusively simple trichomes (Fig. 7A, D. senilis; 7F, C. tenella; not pictured, D. strictus). Conversely, Anchonieae 1 produce dendritic trichomes (Fig. 7B–D) and form a strongly supported group in all analyses. In the ndhF analysis of Beilstein et al. (2006), Sterigmostemum acanthocarpum is a member of this clade and also has dendritic trichomes (Fig. 7E). Bunias orientalis (Anchonieae) is strongly supported as a member of lineage III in all trees, but is not supported as sister to other Anchonieae species, although it also has dendritic trichomes (Beilstein et al., 2006). Warwick et al. (2007), in a comprehensive sample of ITS sequences from 101 species in Anchonieae, Euclidieae, Chorisporeae, and Hesperideae, recovered two distinct monophyletic lineages of Anchonieae. One includes species in the genus Dontostemon (although D. senilis was not included in the study), while the other includes species of Matthiola and Oreoloma. However, Warwick et al. (2007) did not include Bunias orientalis. Despite the strong support for the sister relationship of D. senilis and C. tenella in both PHYA and combined trees, monophyly of Anchonieae is not rejected by either PHYA or combined data (Tables 2, 4). Nevertheless, the observation that ndhF, PHYA, and ITS data place members of the tribe in distinct, nonmonophyletic lineages makes the monophyly of the tribe highly suspect.
Phylogenies inferred from ndhF, PHYA, and combined data support the expansion of the tribe Euclidieae to include Christolea crassifolia, Dilophia salsa, Leiospora eriocalyx, and Shangrilaia nana. Al-Shehbaz et al. (2006) indicated that the latter three were likely members of Euclidieae based on the presence of a mixture of simple and branched trichomes, incumbent cotyledons, and two-lobed stigmas. However, the species were only provisionally placed, pending additional molecular data. Inclusion of Christolea crassifolia in the Euclidieae is also required to maintain the monophyly of Euclidieae s.l.; C. crassifolia is sister to Dilophia salsa in all phylogenies, but with only weak support (Figs. 1–3). Warwick et al. (2007) also found support for Euclidieae s.l., as well as identifying an additional lineage in the tribe (Euclidieae II). The latter group (Warwick et al., 2007) includes several genera not sampled here, but included in the tribe in Al-Shehbaz et al. (2006) based on morphology.
Taxa not included in lineages I–III
In addition to the Aethionemeae, which are sister to all other Brassicaceae, several tribes are placed outside the three major lineages. The tribes Eutremeae, Thlaspideae, and the species Goldbachia laevigata form a monophyletic group in PHYA and combined phylogenies, but with support only in the Bayesian analysis of combined data (PP 0.99) and not in the parsimony bootstrap analysis (PB 56%). Thlaspideae themselves are not monophyletic in PHYA analyses, due to the placement of Pseudocamelina campylopoda as sister to the clade formed by the Eutremeae and Thlaspideae. However, monophyly of the tribe is not rejected by PHYA data (Table 3), and its monophyly is well supported in ndhF and combined trees. The position of Goldbachia laevigata is unresolved in the ITS phylogeny (Warwick et al., 2007) and thus does not contradict its placement with the tribes Thlaspideae and Eutremeae here. Although the positions of Eutremeae and Thlaspideae relative to one another are unresolved in the ndhF tree (Fig. 2), species in the two tribes share the same base chromosome number (x = 7) and palmately veined leaves (Warwick et al., 2007). Thus, evidence from morphology, cytology, and phylogeny supports the sister relationship, but confidence in the relationship requires additional phylogenetic study, which should include species in the genus Goldbachia.
Alysseae are not monophyletic in ndhF, PHYA, or combined analyses, and taxa currently classified as Alysseae occur in three different regions of the PHYA (Fig. 1) and combined (Fig. 3) trees. In Beilstein et al. (2006), the tribe (sensu Schulz 1936) was represented by Alyssum canescens, Farsetia aegyptica, and Lobularia maritima, which did not form a monophyletic group. However, monophyly of the tribe was not rejected by the SH test (Beilstein et al., 2006), so Al-Shehbaz et al. (2006) retained the tribe pending further study. Sampling within Alysseae is expanded in the current study by inclusion of Alyssum linifolium, Clypeola aspera, and Fibigia suffruticosa, which form a monophyletic group in PHYA analyses, but are not closely related to either A. canescens or F. aegyptica (reliable PHYA sequence was not obtained for L. maritima). Bailey et al. (2006) also found evidence to segregate L. maritima from other Alysseae. Furthermore, F. aegyptica and L. maritima are united by having medifixed trichomes (Fig. 10H, I), while Fibigia suffruticosa, C. aspera, and A. linifolium produce stellate trichomes (Fig. 10D–F); the trichomes of A. canescens are dendritic (Fig. 10G). Despite the polyphyly of the Alysseae in ndhF, PHYA, and ITS phylogenies, the monophyly of the tribe is not rejected in topological tests of PHYA data (Table 3). However, the monophyly of the tribe is rejected by the AU test of the combined data (Table 5), and trichome morphology in combination with the lack of monophyly in trees inferred from all three data sets suggests that, as circumscribed in Al-Shehbaz et al. (2006), it consists of three independent lineages. Warwick et al. (2008) recently recircumscribed Alysseae using ITS data and an expanded sample of species in Alyssum as well as other putative Alysseae taxa. Their results agree with those presented here; the core Alysseae are defined by the genera Clypeola, Fibigia, and several species of Alyssum, including A. linifolium, while both Farsetia aegyptica and A. canescens (transferred to the genus Ptilotrichum) fall outside Alysseae (Warwick et al., 2008).
Noccaeeae are monophyletic and sister to Conringia persica in all analyses. In the ITS tree (Bailey et al., 2006) Conringia perfoliata (C. A. Mey.) N. Busch. is monophyletic with species of Noccaeeae. Thus, phylogenetic evidence suggests Noccaeeae could be expanded to include C. persica and C. perfoliata and perhaps other species of Conringia. The relationship of Conringia plus Noccaeeae to other tribes of the family is not well resolved and lacks support in ndhF phylogenies (Fig. 2) (Beilstein et al., 2006) and in the PHYA tree (Fig. 1). The combined tree (Fig. 3) resolves the same clade as that found in the PHYA tree (Fig. 1), and the relationship receives significant Bayesian support (PP 0.98) but low bootstrap support (PB < 50%). Thus, the relationship of Noccaeeae to other tribes of the family requires additional phylogenetic study.
The relationships of several species whose placement in the ndhF analyses was either unresolved or received low support remain problematic in PHYA and combined ndhF/PHYA analyses. For example, Biscutella didyma is well resolved as a member of the large Brassicaceae clade sister to the Aethionemeae in all trees, but its position within this clade is unresolved. In contrast, Asta schaffneri, Heliophila sp., Idahoa scapigera, and Schizopetalon rupestre form a monophyletic group in PHYA analyses (Fig. 1), although neither the ndhF nor combined tree shows this relationship. The branches leading to each of these species are relatively long compared with the length of the branch supporting the relationship (Fig. 1), suggesting the possibility that the relationship is due to long-branch attraction. Thus, the putative association of these taxa with one another requires further phylogenetic exploration.
Trichome SEM and evolution
Trichome morphology is labile in Brassicaceae. In particular, distantly related species often share the same trichome branching pattern, while closely related species can have dramatically different branching patterns (Fig. 3) (online Appendix S3). For example, branching patterns are identical in Arabidopsis thaliana (Fig. 4A) and Olimarabidopsis pumila (Fig. 4D), relatively closely related members of Camelineae, as well as in the more distantly related Strigosella africana (Fig. 8D) (Euclidieae) and Aubrieta deltoidea (Fig. 10B) (Arabideae). Similarly, highly branched, dendritic trichomes occur in species from numerous tribes, including Alysseae (Fig. 10H), Anchonieae (Fig. 7B–D), Boechereae (Fig. 5A, D), Descurainieae (Fig. 6A), Euclidieae (Fig. 8E), Schizopetaleae (Fig. 9A), and Smelowskieae (Fig. 6C, D), among others. Conversely, Smelowskia calycina and S. tibetica are sister species (Figs. 1–3), although S. calycina (Fig. 6C) has highly branched dendritic trichomes and S. tibetica (Fig. 6B) has both simple and dendritic trichomes. The transition between simple and branched trichomes has also occurred frequently in Euclidieae (Fig. 8). Thus, the information on trichome branching added here substantiates previous analyses, which suggest that branching likely evolved numerous times in the family (Beilstein et al., 2006) and that nearly identical branching patterns in trichomes from distantly related species are the result of convergent evolution (online Appendix S3).
Previous analyses suggested that stellate and medifixed trichomes may each have a single origin within Brassicaceae, a hypothesis that was not rejected by the SH topology tests of the ndhF data (Beilstein et al., 2006). However, neither medifixed- nor stellate-trichome-producing species form a monophyletic group in any of our analyses. For example, Erysimum capitatum (Fig. 4E) (Camelineae), Farsetia aegyptica (Fig. 10H) (Alysseae), and Rhammatophyllum erysimoides (Fig. 8F) have medifixed trichomes and belong to different tribes and lineages. In contrast to ndhF analyses of trichome evolution, the combined data reject the hypothesis that medifixed trichomes had a single origin (Table 5). The sample of species with medifixed trichomes is the same in the current and previous studies, suggesting that addition of PHYA sequence data increased the disparity between the most likely tree and the constrained tree enough that the two hypotheses are significantly different by the AU test. The sample of species with stellate trichomes is expanded in the current study by addition of phylogenetic data for Alyssum linifolium, Clypeola aspera, and Fibigia suffruticosa (Alysseae) (Fig. 10D–F), and Physaria rosei (Physarieae) (Fig. 6G). The previously published ndhF analysis included only P. floribunda (Fig. 6F) (Physarieae) and Alyssum canescens (Alysseae) (Beilstein et al., 2006). Trichomes in A. canescens, however, are classified as dendritic in the current study because SEM studies of A. canescens trichomes show that they have a pronounced stalk and that the trichome branches do not radiate from a central point (Fig. 10G). Despite the reclassification of A. canescens, more species with stellate trichomes are included in the current analysis, and the AU test results reject the single origin hypothesis (Table 5).
Conclusions
The PHYA analysis presented here is the most highly resolved and well-supported phylogeny of a nuclear-coding gene of the plant family Brassicaceae to date. In addition, ndhF and PHYA are protein-coding genes, allowing sequence data to be aligned at the amino acid level and thus providing confidence in the homology of analyzed characters. Both the PHYA and combined trees confirm the monophyly of the majority of the recently delimited tribes (Al-Shehbaz et al., 2006) and support recognition of three lineages in the family, each of which is comprised of several tribes. Furthermore, the interpretation of results benefits from independent, thorough phylogenetic analyses of ndhF and PHYA data, thus providing a greater understanding of the resolution afforded by each marker and permitting detailed examination of topological disagreements between the individual markers and between results from the single gene and the combined analysis. Topological disagreements between ndhF and PHYA highlight the need for future phylogenetic study in Camelineae, which contains Arabidopsis thaliana, as well as in Descurainieae, Physarieae, and Schizopetaleae.
Appendix 1. Taxa used in this study, GenBank accession number for ndhF sequence, GenBank accession number for PHYA sequence; and voucher information. Greenhouse-grown specimens cultivated at the Missouri Botanical Garden or elsewhere are noted after the voucher information. I-A Exp = Iranian–American Expedition (collection date follows). Voucher specimens are deposited in the following herbaria: Arnold Arboretum, Harvard University = A, Kunming Institute of Botany = KUN, Missouri Botanical Garden = MO, Tehran University = TUH, University of California = UC, University of Utah = UT, and University of Wisconsin = WIS. Some species have been placed in synonymy or transferred to other genera since the publication of the ndhF tree of Beilstein et al. (2006). When species names in GenBank differ between Beilstein et al. (2006) and the current work, the species designation of Beilstein et al. (2006) follows the GenBank accession number for the ndhF sequence. The ndhF and PHYA sequences of Arabidopsis thaliana were downloaded from GenBank and not generated during this study, thus collection number and voucher information are not given. Sequence alignments and trees produced during this study can be found on TreeBASE (htto://www.treebase.org, study accession S2106).
|
FOOTNOTES
1 The authors thank members of the Kellogg Laboratory at the University of Missouri-St. Louis (UMSL), S. Fuentes, P. Stevens, two anonymous reviewers, and M. Simmons for helpful comments that improved this manuscript. The American Society of Plant Taxonomists, the E. Desmond Lee and Family Laboratory of Plant Systematics (UMSL), Federated Garden Clubs of Missouri, International Center for Tropical Ecology (UMSL), and National Science Foundation (DEB-0408104 to M.A.B. and E.A.K.) provided funding for this project. Fieldwork by I.A.A.-S. was supported by National Geographic Grants 7405-03 and 7452-03. M.A.B. thanks S. Aliscioni, N. Deginani, D. Eakman, A. Marticorena, N. Whiteman, and M. Windham for assistance with collections, Dr. Gomez-Campo for providing seed, and N. Nagalingum for help with SEM figure layout. Scanning electron microscopy was done at UMSL, Central Institute of the Deaf (Washington University, St. Louis), and Harvard University Herbaria. 
5 Author for correspondence (mbeilstein{at}oeb.harvard.edu)
LITERATURE CITED
Al-Shehbaz, I. A., M. A. Beilstein, AND E. A. Kellogg. 2006. Systematics and phylogeny of the Brassicaceae (Cruciferae): An overview. Plant Systematics and Evolution 259: 89–120.[CrossRef][Web of Science]
Bailey, C. D., M. A. Koch, M. Mayer, K. Mummenhoff, S. L. O’Kane Jr., S. I. Warwick, M. D. Windham, AND I. A. Al-Shehbaz. 2006. Toward a global phylogeny of the Brassicaceae. Molecular Biology and Evolution 23: 2142–2160.
Beilstein, M. A., I. A. Al-Shehbaz, AND E. A. Kellogg. 2006. Brassicaceae phylogeny and trichome evolution. American Journal of Botany 93: 607–619.
Beilstein, M. A., AND D. B. Szymanski. 2004. Cytoskeletal requirements during Arabidopsis trichome development. In B. Hussey [ed.], The plant cytoskeleton in cell differentiation and development. Blackwell, Oxford, UK.
Bennett, J. R., AND S. Mathews. 2006. Phylogeny of the parasitic plant family Orobanchaceae inferred from phytochrome A. American Journal of Botany 93: 1039–1051.
Clack, T., S. Mathews, AND R. A. Sharrock. 1994. The phytochrome apoprotein family in Arabidopsis is encoded by 5 genes: The sequences and expression of PHYD and PHYE. Plant Molecular Biology 25: 413–427.[CrossRef][Web of Science][Medline]
Davis, C. C., C. D. Bell, P. W. Fritsch, AND S. Mathews. 2002. Phylogeny of Acridocarpus-Brachylophon (Malpighiaceae): Implications for tertiary tropical floras and Afroasian biogeography. Evolution 56: 2395–2405.[Web of Science][Medline]
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.
Felsenstein, J. 1985. Confidence limits on phylogenies: An approach using the bootstrap. Evolution 39: 783–791.[CrossRef][Web of Science]
Felsenstein, J. 1973. Maximum likelihood and minimum-steps methods for estimating evolutionary trees from data on discrete characters. Systematic Zoology 22: 240–249.[Abstract]
Franklin, K. A., S. J. Davis, W. M. Stoddart, R. D. Vierstra, AND G. C. Whitelam. 2003a. Mutant analyses define multiple roles for phytochrome C in Arabidopsis photomorphogenesis. Plant Cell 15: 1981–1989.
Franklin, K. A., U. Praekelt, W. M. Stoddart, O. E. Billingham, K. J. Halliday, AND G. C. Whitelam. 2003b. Phytochromes B, D, and E act redundantly to control multiple physiological responses in Arabidopsis. Plant Physiology 131: 1340–1346.
Gadberry, M. D., S. T. Malcomber, A. N. Doust, AND E. A. Kellogg. 2005. Primaclade—A flexible tool to find primers across multiple species. Bioinformatics (Oxford, England) 21: 1263–1264.
Galloway, G. L., R. L. Malmberg, AND R. A. Price. 1998. Phylogenetic utility of the nuclear gene arginine decarboxylase: An example from the Brassicaceae. Molecular Biology and Evolution 15: 1312–1320.[Abstract]
Hall, J. C., H. H. Iltis, AND K. J. Systsma. 2004. Molecular phylogenetics of core brassicales, placement of orphan genera Emblingia, Forchhammeria, Tirania, and character evolution. Systematic Botany 29: 654–669.[CrossRef][Web of Science]
Hall, J. C., K. J. Systsma, AND H. H. Iltis. 2002. Phylogeny of Capparaceae and Brassicaceae based on chloroplast sequence data. American Journal of Botany 89: 1826–1842.
Koch, M., C. Dobes, C. Kiefer, R. Schmickl, L. Klimes, AND M. A. Lysak. 2007. Supernetwork identifies multiple events of plastid trnF(GAA)pseudogene evolution in the Brassicaceae. Molecular Biology and Evolution 24: 63–73.
Koch, M., B. Haubold, AND T. Mitchell-Olds. 2001. Molecular systematics of the Brassicaceae: Evidence from coding plastidic matK and nuclear Chs sequences. American Journal of Botany 88: 534–544.
Lagercrantz, U. 1998. Comparative mapping between Arabidopsis thaliana and Brassica nigra indicates the Brassica genomes have evolved through extensive genome replication accompanied by chromosome fusions and frequent rearrangements. Genetics 150: 1217–1228.
Lavin, M., E. Eshbaugh, J. M. Hu, S. Mathews, AND R. A. Sharrock. 1998. Monophyletic subgroups of the tribe Millettieae (Leguminosae) as revealed by phytochrome nucleotide sequence data. American Journal of Botany 85: 412–433.[Abstract]
Lysak, M. A., M. Koch, A. Pecinka, AND I. Schubert. 2005. Chromosome triplication found across the tribe Brassiceae. Genome Research 15: 516–525.
Maddison, D. R., AND W. P. Maddison. 2002. MacClade 4.05 OS X. Analysis of phylogeny and character evolution. Sinauer, Sunderland, Massachusetts, USA.
Mathews, S. 2006. Phytochrome-mediated development in land plants: Red light sensing evolves to meet the challenges of changing light environments. Molecular Ecology 15: 3483–3503.[CrossRef][Medline]
Mathews, S., AND M. J. Donoghue. 2000. Basal angiosperm phylogeny inferred from duplicate phytochromes A and C. International Journal of Plant Sciences 161: S41–S55.[CrossRef][Web of Science]
Mathews, S., M. Lavin, AND R. A. Sharrock. 1995. Evolution of the phytochrome gene family and its utility for phylogenetic analyses of angiosperms. Annals of the Missouri Botanical Garden 82: 296–321.[CrossRef][Web of Science]
Mathews, S., AND K. S. McBreen. In press. Phylogenetic relationships of B-related phytochromes in the Brassicaceae: Redundancy and the persistence of phytochrome D. Molecular Phylogenetics and Evolution.
Mathews, S., AND R. A. Sharrock. 1996. The phytochrome gene family in grasses (Poaceae): A phylogeny and evidence that grasses have a subset of the loci found in dicot angiosperms. Molecular Biology and Evolution 13: 1141–1150.[Abstract]
Mathews, S., AND R. A. Sharrock. 1997. Phytochrome gene diversity. Plant, Cell & Environment 20: 666–671.[CrossRef]
Mathews, S., R. C. Tsai, AND E. A. Kellogg. 2000. Phylogenetic structure in the grass family (Poaceae): Evidence from the nuclear gene phytochrome B. American Journal of Botany 87: 96–107.
Møller, S. G., P. J. Ingles, AND G. C. Whitelam. 2002. The cell biology of phytochrome signalling. New Phytologist 154: 553–590.[CrossRef][Web of Science]
Monte, E., J. M. Alonso, J. R. Ecker, Y. L. Zhang, X. Li, J. Young, S. Austin-Phillips, AND P. H. Quail. 2003. Isolation and characterization of phyC mutants in Arabidopsis reveals complex crosstalk between phytochrome signaling pathways. Plant Cell 15: 1962–1980.
Nixon, K. C. 1999. The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15: 407–414.[CrossRef][Web of Science]
Parkin, I. A. P., S. M. Gulden, A. G. Sharpe, L. Lukens, M. Trick, T. C. Osborn, AND D. J. Lydiate. 2005. Segmental structure of the Brassica napus genome based on comparative analysis with Arabidopsis thaliana. Genetics 171: 765–781.
Posada, D., AND K. A. Crandall. 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics (Oxford, England) 14: 817–818.[CrossRef]
Rodman, J. E., K. G. Karol, R. A. Price, AND K. J. Sytsma. 1996. Molecules, morphology, and Dahlgren’s expanded order Capparales. Systematic Botany 21: 289–307.[CrossRef][Web of Science]
Rodman, J. E., P. S. Soltis, D. E. Soltis, K. J. Sytsma, AND K. G. Karol. 1998. Parallel evolution of glucosinolate biosynthesis inferred from congruent nuclear and plastid gene phylogenies. American Journal of Botany 85: 997–1006.[Abstract]
Ronquist, F. R., AND J. P. Huelsenbeck. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (Oxford, England) 19: 1572–1574.[CrossRef]
Samuel, R., H. Kathriarachchi, P. Hoffmann, M. H. J. Barfuss, K. J. Wurdack, C. C. Davis, AND M. W. Chase. 2005. Molecular phylogenetics of Phyllanthaceae: Evidence from plastid matK and nuclear PHYC sequences. American Journal of Botany 92: 132–141.
Schulz, O. E. 1936. Cruciferae. In A. Engler, and H. Harms [eds.], Die natürlichen Pflanzenfamilien, 227–658. Engelmann, Leipzig, Germany.
Sharrock, R. A., T. Clack, AND L. Goosey. 2003a. Differential activities of the Arabidopsis phyB/D/E phytochromes in complementing phyB mutant phenotypes. Plant Molecular Biology 52: 135–142.[CrossRef][Web of Science][Medline]
Sharrock, R. A., T. Clack, AND L. Goosey. 2003b. Signaling activities among the Arabidopsis phyB/D/E-type phytochromes: A major role for the central region of the apoprotein. Plant Journal 34: 317–326.[CrossRef][Web of Science][Medline]
Shimodaira, H. 2002. An approximately unbiased test of phylogenetic tree selection. Systematic Biology 51: 492–508.[CrossRef][Web of Science][Medline]
Shimodaira, H., AND M. Hasegawa. 2001. CONSEL: For assessing the confidence of phylogenetic tree selection. Bioinformatics (Oxford, England) 17: 1246–1247.[CrossRef]
Sikes, D. S., AND P. O. Lewis. 2001. Beta software, version1. PAUPRat: PAUP* implementation of the parsimony ratchet. Distributed by the authors. Department of Ecology and Evolutionary Biology, University of Connecticut, Storrs, USA.
Simmons, M. P., C. C. Clevinger, V. Savolainen, R. H. Archer, S. Mathews, AND J. J. Doyle. 2001. Phylogeny of the Celastraceae inferred from phytochrome B gene sequence and morphology. American Journal of Botany 88: 313–325.
Swofford, D. L. 2002. PAUP*: Phylogenetic analysis using parsimony (*and other methods), beta version 4.0. Sinauer, Sunderland, Massachusetts, USA.
Town, C. D., F. Cheung, R. Maiti, J. Crabtree, B. J. Haas, J. R. Wortman, E. E. Hine, R. Althoff, T. S. Arbogast, L. J. Tallon, M. Vigouroux, M. Trick, AND I. Bancroft. 2006. Comparative genomics of Brassica oleracea and Arabidopsis thaliana reveal gene loss, fragmentation, and dispersal after polyploidy. Plant Cell 18: 1348–1359.
Warwick, S. I., AND I. A. Al-Shehbaz. 2003. Nomenclatural notes on Sisymbrium (Brassicaceae). Novon 13: 265–267.[CrossRef][Web of Science]
Warwick, S. I., I. A. Al-Shehbaz, R. A. Price, AND C. Sauder. 2002. Phylogeny of Sisymbrium (Brassicaceae) based on ITS sequences of nuclear ribosomal DNA. Canadian Journal of Botany 80: 1002–1017.[CrossRef]
Warwick, S. I., C. Sauder, AND I. Al-Shehbaz. 2006. Molecular phylogeny, morphology and cytological diversity of Sisymbrium (Brassicaceae). In A. K. Sharma, and A. Sharma [eds.], Plant genome: Biodiversity and evolution, 219–250. Oxford and IBH Publishing, New Delhi, India.
Warwick, S. I., C. A. Sauder, AND I. A. Al-Shehbaz. 2008. Phylogenetic relationships in the tribe Alysseae (Brassicaceae) based on nuclear ribosomal ITS DNA sequences. Canadian Journal of Botany 86: 315–336.[CrossRef]
Warwick, S. I., C. A. Sauder, I. A. Al-Shehbaz, AND F. Jacquemoud. 2007. Phylogenetic relationships in the tribes Anchonieae, Chorisporeae, Euclidieae, and Hesperideae (Brassicaceae) based on nuclear ribosomal ITS DNA sequences. Annals of the Missouri Botanical Garden 94: 56–78.[CrossRef][Web of Science]
Yang, T. J., J. S. Kim, S. J. Kwon, K. B. Lim, B. S. Choi, J. A. Kim, M. Jin et al.. 2006. Sequence-level analysis of the diploidization process in the triplicated FLOWERING LOCUS C region of Brassica rapa. Plant Cell 18: 1339–1347.
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?
This article has been cited by other articles:
|
|
 |
|
  T. L. P. Couvreur, A. Franzke, I. A. Al-Shehbaz, F. T. Bakker, M. A. Koch, and K. Mummenhoff Molecular Phylogenetics, Temporal Diversification, and Principles of Evolution in the Mustard Family (Brassicaceae) Mol. Biol. Evol., January 1, 2010; 27(1): 55 - 71. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. K. Conner, H. F. Sahli, and K. Karoly Tests of adaptation: functional studies of pollen removal and estimates of natural selection on anther position in wild radish Ann. Bot., June 1, 2009; 103(9): 1547 - 1556. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Mummenhoff, A. Polster, A. Muhlhausen, and G. Theissen Lepidium as a model system for studying the evolution of fruit development in Brassicaceae J. Exp. Bot., April 1, 2009; 60(5): 1503 - 1513. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
