| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
3Department of Biology, University of Missouri–St. Louis, 1 University Boulevard, St. Louis, Missouri, 63121 USA; 4Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri, 63166-0299 USA
Received for publication July 27, 2005. Accepted for publication January 7, 2006.
ABSTRACT
To estimate the evolutionary history of the mustard family (Brassicaceae or Cruciferae), we sampled 113 species, representing 101 of the roughly 350 genera and 17 of the 19 tribes of the family, for the chloroplast gene ndhF. The included accessions increase the number of genera sampled over previous phylogenetic studies by four-fold. Using parsimony, likelihood, and Bayesian methods, we reconstructed the phylogeny of the gene and used the Shimodaira–Hasegawa test (S–H test) to compare the phylogenetic results with the most recent tribal classification for the family. The resultant phylogeny allowed a critical assessment of variations in fruit morphology and seed anatomy, upon which the current classification is based. We also used the S–H test to examine the utility of trichome branching patterns for describing monophyletic groups in the ndhF phylogeny. Our phylogenetic results indicate that 97 of 114 ingroup accessions fall into one of 21 strongly supported clades. Some of these clades can themselves be grouped into strongly to moderately supported monophyletic groups. One of these lineages is a novel grouping overlooked in previous phylogenetic studies. Results comparing 30 different scenarios of evolution by the S–H test indicate that five of 12 tribes represented by two or more genera in the study are clearly polyphyletic, although a few tribes are not sampled well enough to establish para- or polyphyly. In addition, branched trichomes likely evolved independently several times in the Brassicaceae, although malpighiaceous and stellate trichomes may each have a single origin.
Key Words: Arabidopsis • Brassica • Brassicaceae • ndhF • phylogeny • Shimodaira–Hasegawa test • trichomes
The mustard family (Brassicaceae or Cruciferae) forms a monophyletic group sister to Cleomaceae (Koch et al., 2001
, 2003
; Hall et al., 2002
). Nearly all members of the family have six stamens in a tetradynamous pattern (two short and four long), a cruciform corolla (i.e., in the form of a cross, hence the older family name), and a distinct capsular fruit (silique: a 2-locular fruit with parietal placentation and a partition dividing it in halves). Species in the family exhibit several highly variable fruit and embryo characters that have been used extensively in classification. The first comprehensive treatment of the family was that of de Candolle (1821)
, who based his classification on fruit type (longer than wide vs. wider than long) and seed embryos (position of the radicle in relation to cotyledons in the seed). Schulz (1936)
proposed the latest and most widely used tribal classification of the family. Employing many of the elements of de Candolle (1821)
, Schulz relied heavily on fruit characters and seed morphology to delimit tribes and subtribes.
Trichome type has received less attention than fruit morphology and seed anatomy as a potentially informative character for delimiting tribes in the family. Prantl (1891)
, however, broke from tradition when he segregated species on the basis of unbranched (simple) vs. branched trichomes, and he remains the only taxonomist to propose the use of trichome type to diagnose taxa at the tribal level. More recently, trichome variation has been used to delineate both genera and species in Brassicaceae (Rollins and Banerjee, 1975
, 1976
, 1979
; Lichvar, 1983
; Jacquemoud, 1988
; Al-Shehbaz, 1989
, 1990
, 1994a
, b; Ancev, 1991
; Mulligan, 1995
). Plants in the Brassicaceae range from completely glabrous to densely hairy and, as noted by Prantl (1891)
, the hairs may be simple or branched. Branched trichomes in the family exhibit diverse morphologies. Trichomes that consist of a distinct, primary axis (stalk) and two (forked) or more (dendritic) branches are most common. In some genera, the stalk of the trichome is greatly reduced, or absent, and the branches radiate from a central point. Stalkless trichomes consisting of two main branches are termed malpighiaceous, and those with three or more branches are stellate. The use of trichomes as a taxonomic character is complicated by the presence, in some genera, of glandular, multicellular trichomes. However, distinct differences suggest the two types of structures are not homologous. Glandular trichomes are almost always multicellular and exude secondary compounds, whereas eglandular trichomes are comprised of only a single cell and are not secretory. Here we concentrate on eglandular trichomes, which occur with greater frequency.
Brassicaceae includes two important model systems. Arabidopsis thaliana (L.) Heynh. is the most widely studied plant model species and the first flowering plant to have its entire genome sequenced (The Arabidopsis Genome Initiative, 2000
). Studies in A. thaliana have addressed an impressive spectrum of questions and have refined our understanding of numerous topics ranging from ecology to cellular biology (Somerville and Meyerowitz, 2002). The second model system is the agriculturally important Brassica oleracea complex (B. oleracea L., B. rapa L., B. nigra (L.) W. D. J. Koch, and their three reciprocal hybrids), which has provided insight into the genetics of flowering time (Schranz et al., 2002
), hybridization, and gene silencing, (Pires et al., 2004
), among many other phenomena. Surprisingly, despite clear family-level morphological characters and an overwhelming accumulation of information on A. thaliana and the Brassica oleracea complex, we know comparatively little about the evolutionary history of the family.
Why have phylogenetic studies of the mustard family lagged behind other modes of inquiry? One major reason is the historical use of fruit and seed morphology to classify the 3500+ species into 350 genera and 19 tribes (Schulz, 1936
). Both structures have proven highly labile in evolutionary time; all molecular phylogenetic data show that species with similar fruits and seeds may be unrelated, whereas species with dramatically different fruits and seeds may be very closely related (Koch et al., 2001
, 2003
). The tribes and genera sampled in those studies are mostly poly- or paraphyletic, including Sisymbrieae, the tribe containing A. thaliana (Koch et al., 2001
). As a result, the existing classification provides little guidance for sampling in a phylogenetic study. No previous phylogenetic study of the family has included more than 25 distinct genera, representing 1/14 of all described genera. In contrast, the current study increases the phylogenetic sampling in the family to include nearly 1/3 of all genera and 17 of 19 tribes. Thus, the results presented here provide an important contribution to our understanding of Brassicaceae evolution.
The objectives of the current study are (1) to estimate phylogeny in the family, (2) to test the potential monophyly of the tribes of the family (thus re-examining the usefulness of fruit and seed-shape characters for defining monophyletic groups), (3) to evaluate trichome-branching pattern as a potentially informative morphological character, and (4) to provide an essential framework for future studies in the Brassicaceae.
MATERIALS AND METHODS
Taxa
We sampled 114 accessions of Brassicaceae (Appendix) for the chloroplast gene ndhF, and recorded the tribe (sensu Schulz, 1936
), and trichome type (simple, forked/dendritic, stellate, malpighiaceous) of each species. This sample includes 17 of 19 tribes (sensu Schulz, 1936
) encompassing species in 101 currently accepted genera plus two in the outgroup Cleomaceae (Hall et al., 2002
). Leaf material from the majority of species was collected in silica gel specifically for this project, with collecting trips in North and South America, and central and east Asia. Several species were grown from seeds obtained from the Brassicaceae seed bank of Dr. Cesar Gomez-Campo (Universidad Politécnica de Madrid, Spain). DNA for three accessions was isolated from herbarium specimens. Sequence data for A. thaliana were taken from the full chloroplast sequence in GenBank (accession number NC000932). The monotypic tribes Pringleae and Chamireae (Schulz, 1936
) were not included because freshly collected material was unavailable.
Molecular methods and phylogenetic analysis
DNA from silica-dried and fresh material was extracted using a modified CTAB protocol (Doyle and Doyle, 1987
) and purified in cesium-chloride–ethidium-bromide gradients in an ultracentrifuge. Using protocols optimized for Brassicaceae, the chloroplast gene ndhF was PCR amplified using primers designed for this study (see Supplemental Data accompanying the online version of this article) in combination with those of Sweeney and Price (2000)
. The ndhF gene was sequenced using techniques outlined in Giussani et al. (2001)
. Purified PCR products were sequenced on an ABI Prism 377 automated sequencer (Applied Biosystems, Vienna, Austria) at the University of Missouri–St. Louis with dye terminator chemistry. Double-stranded sequences (minimum overlap=85%) were trimmed at high stringency using DNA STAR-SeqMan II version 4.03 (Lasergene Navigator, Madison, Wisconsin, USA) and aligned at the amino acid level by eye in MacClade 4.05 for OS X (Maddison and Maddison, 2002
). Sequences are deposited in GenBank (Appendix 1).
Phylogeny was estimated using parsimony, maximum-likelihood, and Bayesian methods. Fifteen replicates of 200 parsimony ratchet iterations were implemented using PAUPMacRat (Sikes and Lewis, 2001
) in PAUP* version 4.0b10 (Swofford, 2002
), with 15% of characters re-weighted at each iteration, and the strict consensus of the resulting trees was computed using PAUP*. Sequence evolution models for maximum-likelihood and Bayesian analyses were evaluated using Akaike information criteria (AIC) and hierarchical likelihood ratio test (LRT), with the aid of ModelTest 3.06 (Posada and Crandall, 1998
). Likelihood runs were implemented in PAUP* (TVM+I+
, random sequence addition, tree-bisection-reconnection (TBR) swapping, MULTREES=yes). Bayesian inference used MrBayes 3.1 (Ronquist and Huelsenbeck, 2003
) and a slightly more complex model of evolution (GTR+I+, two independent runs each of 4 chains, 5000000 generations, sampling every 1000 trees). Convergence of chains and burn-in for each Bayesian run was determined independently by plotting –log likelihood, tree length, and the shape parameter of the gamma distributed rate variation (alpha) against the number of generations. Sampled trees whose –log likelihood, tree length, or shape parameter had yet to reach stationarity were discarded (332 trees and 226 trees, respectively). The remaining trees from each run were combined into a single data set (9442 trees), and a majority-rule consensus was computed using PAUP*.
Support for nodes within the resulting phylogenies was explored by parsimony bootstrap (PAUP*, 500 replicates each with 1000 random sequence additions, TBR swapping, saving no more than 500 trees per replicate) and likelihood bootstrap (100 replicates run in parallel using PAUP* for UNIX on the Beowulf Cluster Expedition at the University of Missouri–St. Louis (1 random sequence addition, TBR swapping, MULTREES=yes). These values were compared with Bayesian posterior probabilities obtained from the majority-rule consensus of trees obtained in MrBayes 3.1.
Shimodaira–Hasegawa test
To evaluate trees resulting from alternative reconstruction methods (parsimony, likelihood, and Bayesian approaches), to determine the likelihood of monophyly for the tribes of Schulz (1936)
, and to test scenarios of trichome evolution, we used the Shimodaira–Hasegawa test (S–H test) (Shimodaira and Hasegawa, 1999
) to compare 30 different phylogenetic hypotheses (Table 1). To test the monophyly of Schulz's (1936) tribes, we used MacClade 4.05 to construct constraint trees with all the sampled tribes as monophyletic simultaneously (Schulz, 1936
, Table 1), and individually (one constraint tree for each sampled tribe, e.g., Matthioleae, Table 1). Thirteen taxa included in this study were described after Schulz's 1936 publication, and these taxa were designated as "new taxa," placed in one of Schulz's tribes based on morphology, and used in the construction of additional constraint trees (e.g., tribes new taxa, Matthioleae new taxa). Similarly, to test scenarios of trichome evolution, we constructed constraint trees in which each trichome morphology evolved only once (e.g., simple, dendritic, malpighiaceous, stellate), trichome branching evolved only once (branching), trichomes evolved only once (trichome), and in which each trichome type defined a distinct monophyletic clade (trichome clades). Following the construction of constraint trees, we used PAUP* with the original data set to infer likelihood phylogenies for each designated constraint under the TVM+I+ model using the same parameters as for the unconstrained search. Finally, the most likely topologies inferred under the constraints, as well as the parsimony, unconstrained likelihood, and Bayesian tree topologies were input into PAUP* where an S–H test was used to determine whether the constraint trees were statistically worse than the most likely tree (1000 bootstrap replicates to generate a distribution by resampling estimated log likelihoods [RELL method]).
|
ndhF sequence data
The aligned data matrix consists of 2085 characters across 116 taxa (GenBank numbers DQ288726–DQ288840). Sequence for Arabidopsis thaliana ndhF was obtained from GenBank (NC000932). Arabis alpina L. has the longest ndhF sequence (2079 base pairs [bp]), but most taxa produce sequences of 2070 bp. The longest indel in the data set is three codons long and accounts for the extended sequence length of A. alpina. The shortest ndhF sequences (2064 bp) occur in Aethionema saxatile (L.) R.Br., Catolobus pendula (L.) Al-Shehbaz, Dimorphocarpa wislizenii (Englem.) Rollins, Moriera spinosa Boiss., Physaria floribunda Rybd., and Sisymbrium linifolium Nutt. The ndhF sequences of Arabidopsis lyrata (L.) O'Kane & Al-Shehbaz, Aubrieta parviflora Boiss., and Myagrum perfoliatum L. are shorter than 2070 bps due to problems obtaining high quality sequence at either the 3' or 5' end of the gene. The sequences of these three species are still included in the final data matrix because 85% or more of the sequence is double stranded (e.g., sequencing in A. lyrata resulted in 1999 double-stranded base pairs, or 96.6% of 2070 total base pairs, although the 14 bp from the 5' end and 57 bp from the 3' end are considered as missing data). Idahoa scapigera (Hook.) A. Nelson & J.F. Macbr. sequences do not form a continuous open reading frame throughout the gene. Stop codons were identified consistently at base position 1643. In multiple sequencing attempts from two different accessions of I. scapigera, including cloning the entire amplified region, we never discovered a functional copy of ndhF. Because the nonfunctional copies were recovered repeatedly, we infer that they are not PCR artifacts.
Brassicaceae ndhF sequences are A-T rich (29.6 and 40%, respectively). Sequence divergence (pairwise distances) among ingroup taxa with open reading frames for ndhF sequences range from 0%, between Mostacillastrum elongatum O.E. Schulz and Schizopetalon rupestre (Barn.) Reiche, to 7.9% between Moriera spinosa and Chorispora tenella. The greatest pairwise distance in the data set is 8.4%, between the outgroup Cleome rutidosperma DC. and both Diptychocarpus strictus (Fisch. ex M. Bieb.) Trautv. and C. tenella. Sequence divergence between C. rutidosperma and either putatively nonfunctional copy of I. scapigera is 8.6–8.7%.
Phylogenetic analyses
Tree topologies resulting from parsimony ratchet, likelihood, and Bayesian analyses are statistically not significantly different (Figs. 1, 2; Table 1). The parsimony ratchet replicates yield 942 equally parsimonious trees from the 3000 trees produced by 15 replicates of 200 iterations. The strict consensus of these trees has a length of 2715 steps, a consistency index = 0.31, excluding uninformative characters, and a retention index = 0.64. The evaluation of 64 models of evolution for use in likelihood and Bayesian analyses indicates that the least complex model of evolution favored by the data is dependent upon whether the likelihood scores of models are compared by AIC or LRT. The TVM+I+ model, which differs from the most complex model (GTR+I+ ) by having four rather than six substitution rates, is favored by AIC, while the GTR+I+ is favored by LRT. The TVM+I+ model was used to produce a likelihood tree with a –ln L = 19262.1044 (Fig. 1). MrBayes 3.1 does not permit the selection of the TVM+I+ model, so we specified the GTR+I+ model for Bayesian analyses (Fig. 2). All generated trees are congruent, regardless of the method of construction or model specified.
|
|
Lineage I
Arabidopsis thaliana to Alyssum canescens DC. (Figs. 1, 2). Lineage I is a well-supported monophyletic group (100/91/78) including 40 accessions and characterized by the presence of forked and dendritic trichomes in the majority of species sampled (Fig. 2). Within lineage I is a strongly supported subgroup formed by clades A through D (Arabidopsis thaliana–Physaria floribunda; 100/100/ 98). Clade A [A. thaliana to Erysimum capitatum (Douglas ex Hook.) Greene] is characterized by forked and dendritic trichomes, with Erysimum capitatum having malpighiacious trichomes. Species in clade A represent four tribes and eight genera; the clade is strongly supported as monophyletic (100/99/99). Within clade A, A. thaliana and A. lyrata form a monophyletic group (100/79/85) and together are sister to a monophyletic group containing Camelina microcarpa Andrz. ex DC., Camelina laxa C.A. Mey., Capsella bursa-pastoris (L.) Medik., and Catolobus pendula (100/80/100). The genus Camelina is strongly supported as monophyletic (100/100/100), as is the group formed by Capsella bursa-pastoris and Catalobus pendula (100/94/95). Other members of clade A include Turritis glabra L., O. pumila, and E. capitatum; the placement of the latter two species in relation to other members of the clade is unresolved. Clade A is sister to Stenopetalum nutans F. Muell in all most parsimonious trees, although this placement is without support.
Clade B [Anelsonia eurycarpa (A. Gray) J.F. Macbr. & Payson to Polyctenium fremontii (S. Wats.) Greene] contains species with forked or dendritic trichomes; the group is strongly supported as monophyletic (100/95/89). Within clade B, the genus Boechera (Á. Löve & D. Löve) is paraphyletic; the closest relative to A. eurycarpa is Boechera platysperma (A. Gray) Al-Shehbaz (100/87/87), and together these two species are sister to the clade comprising B. laevigata (Muhl. ex Willd.) Al-Shehbaz and B. shortii (Fernald) Al-Shehbaz (100/86/86). Phoenicaulis cheiranthoides Nutt. and Nevada holmgrenii (Rollins) N.H. Holmgren form a monophyletic group in clade B (100/87/84), and this group is sister to the Anelsonia–Boechera group (100/87/89). Polyctenium fremontii and Cusickiella quadricostata (Rollins) Rollins are sequentially sister to the remainder of the clade, respectively.
Clade C [Pennellia longifolia (Benth.) Rollins to Halimolobus montanum (Griseb.) O. E. Schulz] includes species with forked or dendritic trichomes; the group is strongly supported as monophyletic (100/96/88). Within the clade, Pennellia brachycarpa Beilstein & Al-Shehbaz and P. longifolia are sister (100/100/100), while the relationships to Halimolobus and Mancoa are unresolved.
Clade D contains Dimorphocarpa wislizenii (Engelm.) Rollins and Physaria floribunda. Physaria floribunda has stellate trichomes, while D. wislizenii trichomes are dendritic; the clade is strongly supported as monophyletic (100/100/99). Clade D is sister to the well-supported group containing clades A–C (100/100/98).
Lineage I also includes four other distinct clades and the taxa Alyssum canescens and Hornungia procumbens, the relationships among which are largely unsupported. Clade E (Barbarea vulgaris R.Br. to Nasturtium officinale R. Br.) encompasses a series of glabrous species; the group is monophyletic (100/100/99). There is considerable structure within clade E, which consists of two primary groups, each with good support. One group (100/89/79) contains members of tribe Arabideae [B. vulgaris, Planodes virginicum (L.) Greene, Leavenworthia crassa Rollins] and tribe Lunarieae (Selenia dissecta Torr. & A. Gray). The second group (100/98/97) contains Cardamine pulchella (Hook. f. & Thoms.) Al-Shehbaz & G. Yang, Iodanthus pinnatifidus (Michx.) Steudel, and N. officinale; all three species are members of the tribe Arabideae.
Clade F is a strongly supported (100/97/99) group consisting of Lepidium alyssioides A. Gray and L. draba L. Both species have simple trichomes and angustiseptate fruits (flattened perpendicular to the partition) with one-seeded locules.
Species of clades G and H [Descurainia sophia (L.) Webb to Sophiopsis annua (Rupr.) O.E. Schulz] were assigned to two different tribes (Sisymbrieae and Lepidieae) by Schulz (1936)
and share forked or dendritic trichomes. Clade G (100/88/87) includes Descurainia sophia and Ianhedgea minutiflora (Hook. f. & Thoms.) Al-Shehbaz & O'Kane. Clade H (100/92/91) places Sophiopsis annua with Hedinia tibetica (Thoms.) Ostenf. and Smelowskia calycina (Stephan ex Willd.) C.A. Mey, the latter two as sister taxa (100/92/94). The position of Alyssum canescens, a member of tribe Alysseae, is unresolved in relation to clades A–H. This species has stellate trichomes and is firmly placed in lineage I (100/91/78).
Lineage II
Thelypodium laciniatum (Hook.) Endl. to Myagrum perfoliatum L. (Figs. 1, 2). The majority of species in lineage II lack trichomes, although Neuontobotrys elloanensis Al-Shehbaz and Sisymbrium frutescens Gill. ex Hook. have simple trichomes, and those of Schizopetalon rupestre are dendritic (Fig. 2). The lineage includes 18 accessions, comprises three distinct clades (I–K, Fig. 1) and is strongly supported as monophyletic (100/98/98). Clade I (T. laciniatum to Sisymbrium linifolium (Nutt.) Nutt.) is strongly supported as monophyletic (100/100/100). Schizopetalon rupestre, Mostacillastrum elongatum O.E. Schulz, and Sisymbrium frutescens are a strongly supported monophyletic group (100/100/99), as is the group formed by S. altissimum L. and S. linifolium (100/100/100), which together are sister to the rest of clade I. The latter two species are either glabrous or have simple trichomes and are typical members of tribe Sisymbrieae.
Clade J [Brassica oleracea to Hirschfeldia incana (L.) Lagr.-Foss.] is comprised of three representative species of the tribe Brassiceae; the clade is sister to clade I (100/100/100). All three species of clade J are glabrous. Isatis tinctoria and Myagrum perfoliatum L. form clade K (100/97/93) and are sister to all other members of lineage II (100/98/98). Both species are glabrous and traditionally have been assigned to different tribes (Fig. 2).
Lineage III
Braya rosea Bunge to Dontostemon senilis Maxim. (Figs. 1, 2). Support for the monophyly of lineage III is slightly weaker than that of the other major lineages (100/76/68). Trichomes across the lineage are simple, dendritic, or malpighiaceous. Most of the 24 sampled species in lineage III are contained in one of four clades (Q–T), although D. senilis, Bunias orientalis L, and Leiospora eriocalyx (Regel & Schmalh.) F. Dvorák form a polytomy with these clades. There is strong support for the monophyly of clade Q (100/99/99), the largest clade in the lineage, which contains species assigned to six different tribes and consists of two primary groups. The first group contains the species B. rosea, Shangrilaia nana Al-Shehbaz, J.P. Yue & H. Sun, Christolea crassifolia, and Dilophia salsa Thoms. and is supported as monophyletic (100/91/92). Within this first group, Braya rosea and S. nana are sister taxa (100/100/100), and B. rosea has forked trichomes while the other three species have simple hairs. The second group (Solms-laubachia zhongdianensis J.P. Yue, Al-Shehbaz & H. Sun to Tetracme pamirica Vassilcz.) is also supported as monophyletic (100/91/92). Within this second group, Malcolmia africana (L.) R. Br. and Neotorularia korolkowii (Regel & Schmalh.) Hedge & J. Léonard are sister taxa (100/75/89), and together they are sister to T. pamirica, a relationship present in all most-parsimonious trees but otherwise lacking phylogenetic support; all three species have dendritic trichomes. The group consisting of the remaining five species, Solms-laubachia zhongdianensis, Desideria linearis (N. Busch) Al-Shehbaz, Sisymbriopsis mollipila (Maxim.) Botsch., Rhammatophyllum erysimoides (Kar. & Kir.) Al-Shehbaz & O. Appel, and Euclidium syriacum (L.) R. Br., is monophyletic (100/73/81); the trichomes of S. zhongdianensis and D. linearis are simple, those of R. erysimoides are malpighiaceous, and those of E. syriacum and S. mollipila are dendritic. The genus Sisymbriopsis is also represented in clade Q by S. yechengnica (C.H. An) Al-Shehbaz, C.H. An & G. Yang, a species with simple trichomes that is sister to the remaining taxa within the clade.
Clade R consists of the species Matthiola integrifolia Kom., Oreoloma violaceum Botsch., Sterigmostemum acanthocarpum (Fisch. & C.A. Mey.) Kuntze, and Matthiola farinosa Bunge ex Boiss.; all four species have forked/dendritic trichomes. Oreoloma violaceum and S. acanthocarpum are sister taxa (100/100/99), and the two species together are sister to M. integrifolia, although the latter relationship is not as well supported (94/75/87). The species Hesperis matronalis L. and Hesperis sp. nov. (clade S) have forked trichomes and uniseriate, glandular papillae and form a strongly supported monophyletic Hesperis (100/100/100). Clade T consists of Chorispora tenella (Pallas) DC. and Diptychocarpus strictus (100/100/100); both species have been assigned to the tribe Matthioleae, although C. tenella is glabrous and D. strictus has dendritic trichomes.
In addition to lineages I–III, several smaller monophyletic groups appear in the ndhF phylogeny. Three glabrous species, Chalcanthus renifolius Boiss., Taphrospermum altaicum C.A. Mey, and Eutrema heterophylum (W.W. Sm.) H. Hara, form clade L, a well-supported monophyletic group representing two tribes (100/95/90). Thlaspi arvense L. falls within a strongly supported monophyletic clade M, which also includes Alliaria petiolata (M. Bieb.) Cavara & Grande and three other species (100/92/92). Parlatoria rostrata Boiss. & Hohen. is sister to A. petiolata (100/100/100), although A. petiolata has simple trichomes and P. rostrata is glabrous. Pseudocamelina camplyopoda Bornm. & Gauba ex Bornm. and Graellsia saxifragaefolia Boiss. are closely related to T. arvense (100/99/97); all three species are glabrous and have been assigned to different tribes. Three species of Noccaea Moench. are strongly supported as monophyletic and together with Conringia persica Boiss. form clade O (100/100/100). Clade N (100/100/100) includes five species with dendritic trichomes (Arabis alpina to Baimshania pulvinata Al-Shehbaz) (Figs. 1, 2). Within the clade, Aubrieta deltoidea (L.) DC. and A. parviflora are strongly supported as sister taxa (100/100/100). The two species of Aubrieta are sister to Arabis alpina and Draba altiaca Bunge (94/73/70), although the support for this relationship is not as strong. Baimshania pulvinata is also a member of clade N. Farsetia aegyptica Desv. and Lobularia maritima (L.) Desv. constitute clade P (100/98/100), have been assigned to the tribe Alysseae, and have malpighiaceous trichomes. Clade U is comprised of Moreira spinosa and Aethionema saxatile (Figs. 1, 2) and is strongly supported as sister the remainder of the family Brassicaceae (100/100/100). Both species have been assigned to the tribe Lepidieae and are entirely glabrous.
The phylogenetic position of nine accessions included in the study remains unresolved. Heliophila sp. is the only representative of the exclusively South African tribe Heliophileae and its trichomes are simple. Menonvillea hookeri Rollins and Cremolobus subscandens Kuntze both have simple trichomes and are members of the tribe Cremolobeae. Two cloned ndhF fragments from the monotypic, North American endemic Idahoa scapigera are supported as monophyletic, but otherwise are placed in an unresolved position. Similarly the species Asta schaffneri (S. Wats.) O.E. Schulz, Biscutella didyma L., Goldbachia laevigata (M. Bieb.) DC., Iberis sempervirens L., Ionopsidium acaule Rchb., and Lunaria annua L. show no statistically supported relationship to other sampled taxa.
S–H test, tribal classification, and trichome evolution
Twelve of the 19 tribes of Brassicaceae were represented by two or more genera in our study. These were used to produce constraint trees in an S–H test to evaluate the validity of different phylogenetic hypotheses (Table 1). Topologies in which the monophyly of the tribes Arabideae, Hesperideae, Lepidieae, Matthioleae, and Sisymbrieae were enforced differed significantly from the most likely tree (P < 0.05), whether or not they included species or genera identified after Schulz's (1936) treatment (new taxa). Conversely, topologies in which the monophyly of tribes Alysseae, Brassiceae, Cremolobeae, Drabeae, Euclidieae, Lunarieae, and Streptantheae were enforced did not differ significantly from the most likely tree, regardless of the inclusion of new taxa. However, topologies in which monophyly was required for all tribes of the family simultaneously (Schulz 1936
, tribes new taxa; Table 1) did differ significantly from the most likely tree.
Seven scenarios of trichome evolution were also evaluated using the S–H test. Topologies in which monophyly was required for all trichome-producing taxa (trichomes, Table 1) and in which each trichome type formed a monophyletic group (trichome clades, Table 1) were statistically significantly different than the most likely tree. Similarly, topologies forcing taxa with simple or dendritic trichomes into monophyly were also statistically significantly different than the most likely tree. Conversely, topologies in which malpighiaceous and stellate trichomes evolved only once did not differ significantly from the most likely tree.
DISCUSSION
The sample of Brassicaceae included in this study is the most extensive phylogenetic sampling of the family to date and represents a four-fold increase in generic sampling and a three-fold increase in tribal coverage over previous studies. The chloroplast ndhF gene provides sufficient signal to divide the sampled taxa into three lineages, and 92 of the 113 species sampled fall into one of 21 well-supported monophyletic clades. None of the lineages reflect either the tribal delimitations of Schulz (1936)
or trichome morphology. In addition, lineage III is a novel grouping overlooked by previous phylogenetic studies due to lack of appropriate sampling.
Results obtained using ndhF are largely consistent with the trees produced from other molecular markers. The genus Aethionema forms the basal lineage (clade U; Fig. 1) in this and all previous molecular phylogenies in which it has been included (Galloway et al., 1998
; Koch et al., 2001
, 2003
; Hall et al., 2002
). Moreira spinosa is a spine-forming species and is most closely related to Aethionema saxatile. Both taxa are centered in the Irano-Turanian region (Hedge and Rechinger, 1968
), suggested as a possible site of origin for the Brassicaceae (Hedge, 1976
). Moriera was united with Aethionema by some authors (e.g., Hayek, 1911
), and our data are consistent with that conclusion. The lineages of the Brassicaceae (I–III) lack defining morphological features that would permit efficient identification. In contrast, the monophyletic clades (A–U), 15 of which are included in one of the three lineages, have uniform trichome branching morphologies or, in some cases, stable fruit and seed morphologies. These clades largely form the basis for a new tribal classification of the family (Al-Shehbaz et al., in press
).
Previous tribal classification
Our data confirm the difficulty of using fruit and seed characters as indicators of relationship. The tribe Sisymbrieae, with its long slender fruits, is polyphyletic, as are the tribes Arabideae, Matthioleae, and Hesperideae, which were defined on the basis of the position of the embryo radicle in relation to the folded cotyledons (Fig. 2). The tribe Lepidieae is delineated on the basis of fruits that are flattened perpendicular to the partition (angustiseptate), and the polyphyly of this tribe (Fig. 2) indicates that the evolution of angustiseptate fruits is much more complex than current taxonomy suggests.
Aethionema saxatile and Moriera spinosa are members of the tribe Lepidieae, and their basal phylogenetic position implies that angustiseptate fruits may have evolved at, or near, the origin of the family. The fruits of most Cleomaceae are longer than wide, although the fruits of some Cleomella species are slightly wider than long. However, a more focused exploration of fruit evolution is required to untangle the evolution of fruit shape in both Brassicaceae and Cleomaceae.
The genus Thlaspi, also assigned to the tribe Lepidieae, is polyphyletic, with Noccaea (and other genera not sampled here) being split from it. Meyer (1973)
retained striate-seeded species in the genus Thlaspi, and segregated species lacking striations into Noccaea (clade N), a distinction supported in this study and other phylogenetic work (Zunk et al., 1999
; Koch and Mummenhoff, 2001
). Interestingly, T. arvense is a close relative of Alliaria petiolata and Parlatoria rostrata, two additional species with striated seeds, although no other members of clade M have striations. The genus Lepidium s.l. (including Cardaria) is monophyletic (clade F) and is characterized by a reduction in stamen number from six to four, and sometimes two (Bowman et al., 1999
; Mummenhoff et al., 2001
). Other phylogenetic studies show that species of Coronopus and Stroganowia are also included in Lepidium (Al-Shehbaz et al., 2002
).
Brassica oleracea and other members of the tribe Brassiceae share fruits that are broken laterally into two segments (heterocarpic) and/or cotyledons that are folded together around the radicle (conduplicate), characters that are not present elsewhere in the family (Al-Shehbaz, 1985b
). Three members of tribe Brassiceae form clade J and topologies that force all five sampled Brassiceae into monophyly are statistically indistinguishable from the most likely tree (Table 1). Support for the monophyly of Brassiceae is evident in other phylogenetic work (Warwick and Black, 1993
, 1994
, 1997a
, b
). Despite the putative monophyly of the tribe, Conringia persica and Chalcanthus renifolius are not included in clade J, but are well-supported members of other clades in the phylogeny. Conringia persica lacks conduplicate cotyledons, an observation that supports its segregation from other Brassiceae (Al-Shehbaz, 1985a
).
The tribes Alysseae, Cremolobeae, Drabeae, Euclidieae, Lunarieae, and Streptantheae are not monophyletic in any of the most parsimonious, most likely, or Bayesian trees resulting from phylogenetic analyses in this study. Despite the placement of members of these tribes in distinct, well-supported, monophyletic groups in the phylogeny presented here, constraint trees forcing these tribes into monophyly are statistically not significantly different from the unconstrained, most likely tree. It is important to note, however, that these tribes are not as heavily sampled as the tribes Arabideae, Hesperideae, Lepidieae, Matthioleae, and Sisymbrieae. Thus, including additional taxa from these tribes in future phylogenetic studies may strengthen the proposition of their para- or polyphyly.
Five tribes of the Brassicaceae were represented in our study by a single accession, making an assessment of the monophyly of these tribes impossible. Stenopetalum nutans (Stenopetaleae) is restricted to Australia and is strongly supported as a member of lineage I. The sole member of the exclusively South African Heliophileae, Heliophila sp., is unplaced in relation to the major lineages. The taxa Romanschulzia sp. (Romanschulzieae), Schizopetalon rupestre (Schizopetaleae), and Stanleya pinnata (Pursh.) Britton (Stanleyeae) are members of clade I and are part of a monophyletic group of 11 taxa found only in the New World. Most species of this group were relatively recently assigned to the tribe Thelypodieae (Al-Shehbaz, 1985b
) and share stamens of nearly equal length, a gynophore, and petals with a distinct claw. Schulz (1936)
and, later Takhtajan (1997)
, believed that the tribes Romanschulzieae, Schizopetaleae, Stanleyeae, and Streptantheae were the most primitive in his familial classification based on the presence of equal length stamens and a gynophore in the Capparaceae, although this relationship is not supported by any molecular data. More recently, phylogenetic results indicate that some South American taxa assigned to the tribe Sisymbrieae (Schulz, 1936
) should be included in an expanded Thelypodieae (Warwick et al., 2002
), a result confirmed here by the placement of four South American species, traditionally assigned to the Sisymbrieae, in the monophyletic group of 11 taxa detailed previously.
The monotypic tribes Chamireae and Pringleae were not included in our study. Warwick et al. (2002)
found evidence to indicate that Pringlea antiscorbutica R. Br. ex Hook. f., which is endemic to several islands in the southern Indian Ocean, is closely related to the South American Sisymbrieae. Thus, Pringleae would likely fall within clade I in the phylogeny presented here. The tribe Chamireae may be closely related to the tribe Heliophileae and both are restricted to southern Africa (Mummenhoff et al., 2005
).
Trichome characters
Trichome morphology correlates with phylogeny better than does fruit morphology, although trichome branching also has a complex pattern of evolution in the Brassicaceae. It is important to note that our analyses of trichome evolution are limited to phylogenetically sampled accessions. In some cases, genera sampled here contain species with alternative trichome morphologies. As a result, the scenarios of trichome evolution tested here are over-simplifications, but still provide insight into general trends.
Species representing the basal branch of Brassicaceae (clade U) are entirely glabrous, and branched trichomes probably arose after the divergence of this clade from the remainder of the family. Trichomes in the Cleomaceae, sister to Brassicaceae, are exclusively simple. It is unlikely that trichome branching evolved only once in the Brassicaceae because the topology forcing branched trichome taxa into monophyly is statistically significantly worse than the unconstrained, most likely tree. As a result, branched trichomes across the family are more likely the result of more than one evolutionary event. Perhaps the best example of an ostensibly independent origin of branching occurs in lineage II. The lineage is characterized by species that are either glabrous or have only simple trichomes, except in the case of Schizopetalon rupestre. The true pattern of trichome evolution across the family may represent numerous innovations of trichome branching, but ultimately careful developmental and molecular genetic studies are needed to make a more confident assessment of trichome evolution.
In contrast to the general phenomenon of trichome branching, results of the S–H test indicate that stellate trichomes may have a single evolutionary origin (P = 0.775, Table 1). Stellate trichomes occur in lineage I in Alyssum and Physaria and are strongly correlated with arid habitats. Species of Alyssum are distributed in the Mediterranean, while the genus Physaria is distributed in the southwestern United States (Rollins and Banerjee, 1975
, 1976
, 1979
). The genus Physaria was recently united with Lesquerella (Physaria is the earlier name) and forms the polycolporate clade with the genera Dimorphocarpa, Dithyrea, Lyrocarpa, Nerisyrenia, Paysonia, and Synthlipsis (Al-Shehbaz and O'Kane, 2002
), though none of these genera include species with stellate trichomes. However, pollen grains in the polycolporate clade (clade D) have more than three colpi, and this character is a synapomorpy for the clade (O'Kane and Al-Shehbaz, 2002
). Stellate trichomes also occur in Alyssum canescens. The genus Alyssum contains numerous species with stellate trichomes, and future studies addressing stellate trichome evolution should consider these species as well.
Results of the S–H test also indicate that malpighiaceous trichomes may have arisen only once in the Brassicaceae (P = 0.113, Table 1). Despite this fact, taxa with malpighiaceous trichomes are members of distinct, well-supported groups within the most parsimonious, most likely and Bayesian trees. Erysimum capitatum and the Australian endemic Stenopetalum nutans are both members of lineage I; two Mediterranean species, Farsetia aegyptaica and Lobularia maritime, form clade O; and the central Asian Rhammatophyllum erysimoides is a member of clade P in lineage III. It is interesting, therefore, that the S–H test results do not support the conclusion that these taxa evolved malpighiaceous trichomes independently of one another. Such results suggest that the S–H test is relatively conservative and is sensitive to the number (or perhaps proportion) of taxa designated to fall within a particular monophyletic group.
ndhF phylogeny and comparative biology
Organismal phylogenies are important tools for the interpretation of morphology and the assessment of paralogy vs. orthology in gene families (Daly et al., 2001
; Fiebig et al., 2004
; Malcomber and Kellogg, 2004
). The overwhelming accumulation of developmental and genetic information in the model species Arabidopsis thaliana and Brassica oleracea in combination with a well-resolved Brassicaceae phylogeny provide a framework for inquiries in evolutionary developmental genetics.
The genetic pathways that control trichome branching in A. thaliana have been extensively studied and include the genes ZWICHEL (ZWI), STICHEL (STI), and ANGUSTIFOLIA (AN) (Hülskamp, 2000
; Schwab et al., 2000
). These genes are each apparently part of independent, partially redundant pathways. Analysis of double mutant combinations of ZWI, STI, and AN in A. thaliana indicates that the loss of any two of these genes leads to the production of exclusively simple trichomes (Hülskamp, 2000
), a mechanism that could explain the loss of trichome branching more generally.
The proposed mechanism of trichome loss in A. thaliana can be evaluated in light of the phylogenetic results. Taxa in clade E, which contains representatives of the genera Barbarea, Planodes, Leavenworthia, Selenia, Cardamine, Iodanthus, and Nasturtium, are glabrous. Species of these genera form the so-called Cardamine alliance and share an affinity for aquatic to semi-aquatic habitats (Franzke and Hurka, 2000
; Mitchell and Heenan, 2000
; Sweeney and Price, 2000
). Molecular genetic studies of the Cardamine alliance, therefore, can be used to test the applicability of the proposed mechanism of A. thaliana trichome loss to other monophyletic groups in the Brassicaceae and to evaluate the connection between aquatic habitats and trichome loss.
In combination with phylogenetic information, molecular genetic findings from A. thaliana also make it possible to address the potential homology of glandular and eglandular trichomes. Glandular trichomes occur in clades Q–S of lineage III and are also characteristic of the outgroup taxa Cleome rutidosperma and Polanisia dodecandra. Species in lineage III can have both glandular and eglandular trichomes. In A. thaliana, where only eglandular trichomes occur, the genes TRYPTICHON, GLABRA1 and GLABRA2 interact in the initiation of trichomes (Schiefelbein, 2003
). Analyses of orthologous genes in the species of lineage III may reveal whether the glandular and eglandular trichomes of Brassicaceae are truly homologous.
Conclusions
The phylogeny presented here is an important step in developing a more robust evolutionary history of the family Brassicaceae. Sequence data from the chloroplast gene ndhF provide well-supported phylogenetic estimates that complement and extend previous molecular work on the family. Greatly expanded taxon sampling has facilitated the identification of novel groups and a broad assessment of the taxonomic value of fruit and seed characters and trichome branching patterns. The tribal classification proposed by Schulz (1936)
and still in widespread use is shown to be a poor reflection of relationship; at least five of 12 tribes represented by two or more genera in the study are clearly polyphyletic. In many cases, well-defined molecular clades in the phylogeny do not have obvious morphological synapomorphies, which makes it difficult to place genera that lack molecular data into the clades and lineages of the current phylogeny. However, the provisional clades delimited here provide a valuable framework by which morphology can be reevaluated in the light of phylogeny (Al-Shehbaz et al., in press
). In terms of larger goals, the considerable phylogenetic structure inferred from ndhF provides an important opportunity for reciprocal illumination between the fields of anatomy and development and molecular genetics.
Taxon; ndhF GenBank accession; Voucher specimen, Collection locale; Herbarium.
Aethionema saxatile (L.) R. Br.; DQ288726; Beilstein 03-177, USA, MO, cultivated; MO. Alliaria petiolata (M. Bieb.) Cavara & Grande; DQ288727; Beilstein 02-91, USA, MI; MO. Alyssum canescens DC.; DQ288728; Bartholomew et al. 8657, China, Xinjiang; MO. Anelsonia eurycarapa (A. Gray) J.F. Macbr. & Payson; DQ288729; Beilstein 01-72, USA, CA; MO. Arabidopsis lyrata (L.) O'Kane & Al-Shehbaz; DQ288730; Beilstein s.n., USA, MO; MO. A. thaliana (L.) Heynh.; NC000932. Arabis alpina L.; DQ288731; Beilstein s.n., USA, MO, cultivated; MO. Asta schaffneri (S. Wats.) O. E. Schulz; DQ288733; Fuentes-Soriano 48, Mexico, Nuevo Leon; MO. Aubrieta deltoidea (L.) DC.; DQ288734; Al-Shehbaz s.n., cultivated; MO. A. parviflora Boiss.; DQ288735; I-A Exp., 23 May 2004, Iran; UC & TUH.
Baimshania pulvinata Al-Shehbaz; DQ288736; Al-Shehbaz 20026, China, Yunnan; MO. Barbarea vulgaris R. Br.; DQ288737; Beilstein 01-04, USA, MO; MO. Biscutella didyma L.; DQ288738; Beilstein 01-82, USA, MO; MO. Boechera laevigata (Muhl. ex. Willd.) Al-Shehbaz; DQ288739; Beilstein 01-06, USA, MO; MO. B. platysperma (A. Gray) Al-Shehbaz; DQ288740; Beilstein 01-57, USA, NV; MO. B. shortii (Fernald) Al-Shehbaz; DQ288741; Al-Shehbaz s.n., USA, MO; MO. Brassica oleracea L.; DQ288742; Beilstein s.n., broccoli cv.; MO. Braya rosea Bunge; DQ288743; Bartholomew et al. 8447, China, Xinjiang; MO. Bunias orientalis L.; DQ288744; I-A Exp., 28 May 2004, Iran; UC & TUH.
Cakile maritima L.; DQ288745; Beilstein 01-76, USA, CA; MO. Camelina laxa C. A. Mey.; DQ288747; I-A Exp., 29 May 2004, Iran; UC & TUH. C. microcarpa Andrz. ex DC.; DQ288746; Beilstein 01-22, USA, NM; MO. Capsella bursa-pastoris (L.) Medik.; DQ288748; S. Mathews 492, USA, MO; MO. Cardamine pulchella (Hook. f. & Thoms.) Al-Shehbaz & G. Yang; DQ288749; Solomon et al. 20021, Yunnan, China; MO. Catolobus pendula (L.) Al-Shehbaz; DQ288732; Bartholomew et al. 8569, China, Xinjiang; MO. Caulanthus crassicaulis (Torr.) S. Wats.; DQ288750; Beilstein 01-50, USA, UT; MO. Caulostramina jaegeri (Rollins) Rollins; DQ288751; Beilstein 01-74, USA, CA; MO. Chalcanthus renifolius Boiss.; DQ288752; I-A Exp., 26 May 2005, Iran; UC & TUH. Chorispora tenella (Pallas) DC.; DQ288753; Beilstein 01-85, USA, MO cultivated; MO. Christolea crassifolia Cambes.; DQ288754; Bartholomew et al. 8302, China, Xinjiang; MO. Cleome rutidosperma DC.; DQ288755; Torke 217, French Guiana, Cayenne; MO. Conringia persica Boiss.; DQ288756; I-A Exp., 20 May 2004, Iran; UC & TUH. Cremolobus subscandens Kuntze; DQ288757; Beck 7270, Bolivia, Chapare; MO. Cusickiella quadricostata (Rollins) Rollins; DQ288758; Beilstein 01-66, USA, CA; MO.
Descurainia sophia (L.) Webb; DQ288759; Beilstein 01-19, USA, NM; MO. Desideria linearis (N. Busch) Al-Shehbaz; DQ288760; Bartholomew et al. 8461, China, Xinjiang; MO. Dilophia salsa Thoms.; DQ288761; Bartholomew et al. 8456, China, Xinjiang; MO. Dimorphocarpa wislizenii (Englem.) Rollins; DQ288763; Beilstein 01-12, USA, OK; MO. Diptychocarpus strictus Trautv.; DQ288762; I-A Exp., 24 May 2004, Iran; UC & TUH. Dontostemon senilis Maxim.; DQ288764; Bartholomew et al. 8642, China, Xinjiang; MO. Draba altaica Bunge; DQ288765; Bartholomew et al. 8448, China, Xinjiang; MO.
Erysimum capitatum (Douglas ex Hook.) Greene; DQ288766; Beilstein 01-20, USA, NM; MO. Euclidium syriacum (L.) R. Br.; DQ288767; I-A Exp., 2 June 2004, Iran; UC & TUH. Eutrema heterophyllum (W. W. Sm.) H. Hara; DQ288768; Bartholomew et al. 8490, China, Xinjiang; MO.
Farsetia aegyptiaca Desv.; DQ288769; Beilstein 01-88, USA, MO, cultivated; MO.
Glaucocarpum suffrutescens (Rollins) Rollins; DQ288770; Beilstein 01-54, USA, UT; MO. Goldbachia laevigata (M. Bieb.) DC.; DQ288771; Bartholomew et al. 8300, China, Xinjiang; MO. Graellsia saxifragaefolia Boiss.; DQ288772; I-A Exp., 26 May 2004, Iran; UC & TUH.
Halimolobus montanum (Griseb.) O. E. Schulz; DQ288773; Beilstein 03-107, Argentina, Cordoba; MO. Hedinia tibetica (Thoms.) Ostenf.; DQ288774; Bartholomew et al. 8254, China, Xinjiang; MO. Heliophilasp. Burm. f. ex L.; DQ288775; Burge 1031, South Africa; MO. Hesperis sp. nov. Al-Shehbaz; DQ288777; I-A Exp., collected May 2004, Iran; UC & TUH. H. matronalis L.; DQ288776; Beilstein 01-86, USA, MO cultivated; MO. Hirschfeldia incana (L.) Lagr.–Foss.; DQ288778; Beilstein 03-117, Argentina, Cordoba; MO. Hornungia procumbens (L.) Hayek; DQ288779; Bartholomew et al. 9546, China, Xinjiang; MO.
Ianhedgea minutiflora (Hook. f. & Thoms.) Al-Shehbaz & O'Kane; DQ288780; Solomon et al. 21646, Tajikistan, Badakhson; MO. Iberis sempervirens L.; DQ288781; Beilstein 03-92, USA, MO cultivated; MO. Idahoa scapigera (Hook.) A. Nelson & J. F. Macbr.; DQ288782; Baum 365, USA, WA; A. I. scapigera (Hook.) A. Nelson & J. F. Macbr.; DQ288783; Baum s.n., USA, WI, cultivated; WIS. Iodanthus pinnatifidus (Michx.) Steudel; DQ288784; Beilstein 01-01, USA, MO; MO. Ionopsidium acaule Rchb.; DQ288785; Beilstein 03-178, USA, MO cultivated; MO. Isatis tinctoria L.; DQ288786; Beilstein 02-89, USA, MO cultivated; MO.
Leavenworthia crassa Rollins; DQ288787; Beck 40, USA, TN; MO. Leiospora eriocalyx (Regel & Schmalh.) F. Dvorak; DQ288788; Bartholomew et al. 8430, China, Xinjiang; MO. Lepidium alyssoides A. Gray; DQ288789; Beilstein 01-51, USA, UT; MO. L. draba L.; DQ288790; Beilstein 01-24, USA, NM; MO. Lobularia maritima (L.) Desv.; DQ288791; Beilstein 01-87, USA, MO cultivated; MO. Lunaria annua L.; DQ288792; Al-Shehbaz s.n., USA, MO cultivated; MO.
Malcolmia africana (L.) R. Br.; DQ288793; Beilstein 01-46, USA, UT; MO. Mancoa hispida Wedd.; DQ288794; Beilstein 03-151, Argentina, Jujuy; MO. Matthiola farinosa Bunge ex Boiss.; DQ288796; I-A Exp., 21 May 2004, Iran; UC & TUH. M. integrifolia Kom.; DQ288795; Solomon et al. 21374, Tajikistan, Badakhshon; MO. Menonvillea hookeri Rollins; DQ288797; Sweeney 0265, Chile, Santiago; MO. Moriera spinosa Boiss.; DQ288798; I-A Exp., 20 May 2004, Iran; UC & TUH. Mostacillastrum elongatum O. E. Schulz; DQ288799; Beilstein 03-14, Argentina, Tucuman; MO. Myagrum perfoliatum L.; DQ288800; I-A Exp., 2 May 2004, Iran; UC & TUH.
Nasturtium officinale R. Br.; DQ288801; Beilstein 01-39, USA, NV; MO. Neotorularia korolkowii (Regel & Schmalh.) Hedge & J. Léonard; DQ288803; Bartholomew et al. 8220, China, Xinjiang; MO. Neuontobotrys elloanensis Al-Shehbaz; DQ288802; Beilstein 03-165, Chile, Region II; MO. Nevada holmgrenii (Rollins) N. H. Holmgren; DQ288829; Windham 2186, USA, MO; UT. Noccaea cochleariforme (DC.) Á. Löve & D. Löve; DQ288804; Beilstein 01-21, USA, NM; MO. N. sp. Moench; DQ288805; I-A Exp., 26 May 2004, Iran; UC & TUH. N. sp. Moench; DQ288806; I-A Exp., 26 May 2004, Iran; UC & TUH.
Olimarabidopsis pumila (Stephan) Al-Shehbaz, O'Kane & R. A. Price; DQ288807; Beilstein s.n., USA, MO cultivated; MO. Oreoloma violaceum Botsch.; DQ288808; Bartholomew et al. 8596, China, Xinjiang; MO.
Parlatoria rostrata Boiss. & Hohen.; DQ288809; I-A Exp., 26 May 2004, Iran; UC & TUH. Pennellia brachycarpa Beilstein & Al-Shehbaz; DQ288811; Beilstein 03-148, Argentina, Jujuy; MO. P. longifolia (Benth.) Rollins; DQ288810; Fuentes-Soriano 78, Mexico, Chichuahua; MO. Phoenicaulis cheiranthoides Nutt.; DQ288812; Beilstein 01-37, USA, NV; MO. Physaria floribunda Rydb.; DQ288813; Beilstein 01-17, USA, NM; MO. Planodes virginicum Greene; DQ288814; Al-Shehbaz s.n., USA, MO; MO. Polanisia dodecandra (L.) DC.; DQ288815; Stevens s.n., USA, MO; MO. Polyctenium fremontii (S. Wats.) Greene; DQ288816; Beilstein 01-42, USA, ID; MO. Pseudocamelina campylopoda Bornm. & Gauba ex Bornm.; DQ288817; I-A Exp., 23 May 2004, Iran; UC & TUH.
Rhammatophyllum erysimoides (Kar. & Kir.) Al-Shehbaz & O. Appel; DQ288818; Bartholomew et al. 9134, China, Xinjiang; MO. Romanschulzia sp. O. E. Schulz; DQ288819; Fuentes-Soriano 54, Mexico, Nuevo Leon; MO.
Schizopetalon rupestre (Barn.) Reiche; DQ288820; Beilstein 03-168, Chile, Region IV; MO. Selenia dissecta Torr. & A. Gray; DQ288822; Beck 32, USA, MO cultivated; MO.
Shangrilaia nana Al-Shehbaz, J. P. Yue & H.Sun; DQ288823; Al-Shehbaz & J P. Yue s.n., China, Yunnan; KUN. Sisymbriopsis mollipila (Maxim.) Botsch.; DQ288824; Bartholomew et al. 8335, China, Xinjiang; MO. S. yechengnica (C. H. An) Al-Shehbaz, C. H. An & G. Yang; DQ288825; Bartholomew et al. 9569, China, Xinjiang; MO. Sisymbrium altissimum L.; DQ288826; Beilstein 01-26, USA, NM; MO. S. frutescens Gill. ex Hook.; DQ288827; Beilstein 03-171, Argentina, La Rioja; MO. S. linifolium Nutt.; DQ288821; Beilstein 01-49, USA, UT; MO. Smelowskia calycina (Stephan ex Willd.) C. A. Mey; DQ288828; Al-Shehbaz s.n., China, Xinjiang; MO. Solms-laubachia zhongdianensis J. P. Yue, Al-Shehbaz & H. Sun; DQ288830; Al-Shehbaz s.n., China, Xinjiang; MO. Sophiopsis annua (Rupr.) O. E. Schulz; DQ288831; Bartholomew et al. 8271, China, Xinjiang; MO. Stanleya pinnata (Pursh) Britton; DQ288832; Beilstein 01-28, USA, CO; MO. Stenopetalum nutans F. Muell.; DQ288833; Maconochie 2417, Australia, N. Territory; MO. Sterigmostemum acanthocarpum (Fisch. & C. A. Mey.) Kuntze; DQ288834; I-A Exp., 20 May 2004, Iran; UC & TUH. Streptanthus squamiformis Goodman; DQ288835; Beilstein 01-11, USA, OK; MO.
Taphrospermum altaicum C. A. Mey.; DQ288836; Bartholomew et al. 8485, China, Xinjiang; MO. Tetracme pamirica Vassilcz.; DQ288837; Solomon et al. 21386, Tajikistan, Badakhson; MO. Thelypodium laciniatum (Hook.) Endl.; DQ288838; Beilstein 01-65, USA, CA; MO. Thlaspi arvense L.; DQ288839; Beilstein 01-25, USA, NM; MO. Turritis glabra L.; DQ288840; I-A Exp., 2 June 2004, Iran; UC & TUH.
|
1 The authors thank members of the Kellogg Lab at the University of Missouri–St. Louis (UMSL), S. Fuentes, P. Stevens, and two anonymous reviewers 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 provided funding for this project. M.A.B. thanks S. Aliscioni, N. Deginani, D. Eakman, A. Marticorena, N. Whiteman, and M. Windham for assistance with collections, and Dr. Gomez-Campo for providing seed. 
2 Author for correspondence (mab347{at}studentmail.umsl.edu
)
LITERATURE CITED
Al-Shehbaz I. A.. 1985a. The genera of Brassiceae (Cruciferae; Brassicaceae) in the southeastern United States. Journal of the Arnold Arboretum 66: 279-351.[Web of Science]
Al-Shehbaz I. A.. 1985b. The genera of Thelypodieae (Cruciferae; Brassicaceae) in the southeastern United States. Journal of the Arnold Arboretum 66: 95-111.[Web of Science]
Al-Shehbaz I. A.. 1989. Dactylocardamum (Brassicaceae), a remarkable new genus from Peru. Journal of the Arnold Arboretum 70: 515-521.[Web of Science]
Al-Shehbaz I. A.. 1990. A revision of Weberbauera (Brassicaceae). Journal of the Arnold Arboretum 71: 221-250.[Web of Science]
Al-Shehbaz I. A.. 1994a. Petroravenia (Brassicaceae), a new genus from Argentina. Novon 4: 191-196.[CrossRef]
Al-Shehbaz I. A.. 1994b. Three new South American species of Draba (Brassicaceae). Novon 4: 197-202.
Al-Shehbaz I. A. O'Kane S. L. Jr. 2002. Lesquerella is united with Physaria (Brassicaceae). Novon 12: 319-329.[CrossRef][Web of Science]
Al-Shehbaz I. A. Mummenhoff K. Appel O.. 2002. Cardaria, Coronopus, and Stroganowia are united with Lepidium (Brassicaceae). Novon 12: 5-11.
Al-Shehbaz I. A. Beilstein M. A. Kellogg E. A.. In press. Systematics and Phylogeny of the Brassicaceae (Cruciferae): an overview. Plant Systematics and Evolution..
Ancev M. E.. 1991. Genus Alyssum in Bulgarian flora. In S. I. Kozhukharov, B. A. Kuznov [eds.] Evolution of flowering plants and florogenesis, 2. Asteraceae, Brassicaceae, Poaceae, Cyperaceae 85-117 Bulgarian Academy of Sciences, Sofia, Bulgaria.
The Arabidopsis Genome Initiative.. 2000. Analysis of the genome of the flowering plant Arabidopsis thaliana. Nature 408: 796-815.[CrossRef][Medline]
Bowman J. L. Bruggemann H. Lee J.-Y. Mummenhoff K.. 1999. Evolutionary changes in floral structure within Lepidium L. (Brassicaceae). International Journal of Plant Sciences 160: 917-929.[CrossRef][Web of Science][Medline]
de Candolle A. P.. 1821. Cruciferae. Systema Naturale 2: 139-700.
Daly D. C. Cameron K. M. Stevenson D. W.. 2001. Plant systematics in the age of genomics. Plant Physiology 127: 1328-1333.
Doyle J. J. Doyle J. L.. 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochemical Bulletin 19: 11-15.
Fiebig A. Kimport R. Preuss D.. 2004. Comparisons of pollen coat genes across Brassicaceae species reveal rapid evolution by repeat expansion and diversification. Proceedings of the National Academy of Sciences, USA 101: 3268-3291.
Franzke A. Hurka H.. 2000. Molecular systematics and biogeography of the Cardamine pratensis complex (Brassicaceae). Plant Systematics and Evolution 224: 213-234.[CrossRef][Web of Science]
Galloway G. L. Malmberg R. L. Price R. A.. 1998. Phylogenetic utility of the nuclear gene arginine decarboxylase: an example from the Brassicaceae. Molecular Biology and Evolution 15: 1312-1320.[Abstract]
Giussani L. M. Cota-Sanchez J. H. Zuloaga F. O. Kellogg E. A.. 2001. A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. American Journal of Botany 88: 1993-2012.
Hall J. C. Systsma K. J. Iltis H. H.. 2002. Phylogeny of Capparaceae and Brassicaceae based on chloroplast sequence data. American Journal of Botany 89: 1826-1842.
Hayek A. V.. 1911. Entwurf eines Cruciferen–Systems auf phylogenetischer Grundlage. Beihefte zum Botanischen Centralblatt 27: 127-335.
Hedge I. A.. 1976. A systematic and geographical survey of the Old World Cruciferae. In J. G. Vaughn, A. J. MacLeod, B. M. G. Jones [eds.] The biology and chemistry of the Cruciferae 1-355 Academic Press, New York, New York, USA.
Hedge I. A. Rechinger K. H.. 1968. Cruciferae, Flora Iranica, 1–372 Akademische Druck und Verlagsanstalt, Graz, Austria.
Hülskamp M.. 2000. How plants split hairs. Current Biology 10: R308-R310.[CrossRef][Web of Science][Medline]
Jacquemoud F.. 1988. Monographie du genre Sterigmostemum M. Bieb. (Cruciferae–Hesperideae). Boissiera 40: 1-161.
Koch M. Mummenhoff K.. 2001. Thlaspi s. str. (Brassicaceae) versus Thlaspi s. l.: morphological and anatomical characters in the light of ITS nrDNA sequence data. Plant Systematics and Evolution 227: 209-225.[CrossRef][Web of Science]
Koch M. Haubold B. Mitchell-Olds T.. 2001. Molecular systematics of the Brassicaceae: evidence from coding plastidic matK and nuclear Chs sequences. American Journal of Botany 88: 534-544.
Koch M. Al-Shehbaz I. A. Mummenhoff K.. 2003. Molecular systematics, evolution and population biology in the mustard family (Brassicaceae). Annals of the Missouri Botanical Garden 90: 151-171.[CrossRef][Web of Science]
Lichvar W. R.. 1983. Evaluation of Draba oligosperma, Draba pectinipila and Draba juniperina complex (Cruciferae). Great Basin Naturalist 43: 441-443.[Web of Science]
Maddison D. R. Maddison W. P.. 2002. MacClade 4.05 OS X. Analysis of phylogeny and character evolution Sinauer, Sunderland, Massachusetts, USA.
Malcomber S. T. Kellogg E. A.. 2004. Heterogeneous expression patterns and separate roles of the SEPALLATA gene LEAFY HULL STERILE1 in grasses. Plant Cell 16: 1692-1706.
Meyer F. K.. 1973. Conspectus der "Thlaspi": Arten Europas, Afrikas and Vorderasiens. Feddes Repertorium 84: 449-470.
Mitchell A. D. Heenan P. B.. 2000. Systematic relationships of New Zealand endemic Brassicaceae inferred from nrDNA ITS sequence data. Systematic Botany 25: 98-105.[CrossRef][Web of Science]
Mulligan G. A.. 1995. Synopsis of the genus Arabis (Brassicaceae) in Canada, Alaska and Greenland. Rhodora 97: 109-163.[Web of Science]
Mummenhoff K. Bruggemann H. Bowman J. L.. 2001. Chloroplast DNA phylogeny and biogeography of the genus Lepidium (Brassicaceae). American Journal of Botany 88: 2051-2063.
Mummenhoff K. Al-Shehbaz I. A. Bakker F. T. Linder H. P. Mühlhausen A.. 2005. Phylogeny, morphological evolution, and speciation of endemic Brassicaceae genera in the Cape flora of southern Africa. Annals of the Missouri Botanical Garden 92: 400-424.[Web of Science]
O'Kane S. L. Jr Al-Shehbaz I. A.. 2002. Paysonia, a new genus segregated from Lesquerella (Brassicaceae). Novon 12: 379-381.[CrossRef][Web of Science]
O'Kane S. L. Jr Al-Shehbaz I. A.. 2003. Phylogenetic position and generic limits of Arabidopsis (Brassicaceae) based on sequences of nuclear ribosomal DNA. Annals of the Missouri Botanical Garden 90: 603-612.[CrossRef][Web of Science]
Pires J. C. Zhao J. W. Schranz M. E. Leon E. J. Quijada P. A. Lukens L. N. Osborn T. C.. 2004. Flowering time divergence and genomic rearrangements in resynthesized Brassica polyploids (Brassicaceae). Biological Journal of the Linnean Society 82: 675-688.[CrossRef]
Posada D. Crandall K. A.. 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14: 817-818.
Prantl K.. 1891. Cruciferae. In A. Engler, K. Prantl [eds.] Die natürlichen Pflanzenfamilien 145-206 Wilhelm Engelmann, Leipzig, Germany.
Rollins R. C. Banerjee U. C.. 1975. Atlas of the trichomes of Lesquerella (Cruciferae). Publication of the Bussey Institute Harvard University 1975: 3-48.
Rollins R. C. Banerjee U. C.. 1976. Trichomes in studies of the Cruciferae. In J. G. Vaughn, A. J. McLeod, B. M. G. Jones [eds.] The biology and chemistry of the Cruciferae 145-166 Academic Press, New York, New York, USA.
Rollins R. C. Banerjee U. C.. 1979. Trichome patterns in Physaria. Publication of the Bussey Institute Harvard University 1979: 65-77.
Ronquist F. R. Huelsenbeck J. P.. 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574.
Schiefelbein J.. 2003. Cell-fate specification in the epidermis: a common patterning mechanism in the root and shoot. Current Opinion in Plant Biology 6: 74-78.[CrossRef][Web of Science][Medline]
Schranz M. E. Quijada P. Sung S. B. Lukens L. Amasino R. Osborn T. C.. 2002. Characterization and effects of the replicated flowering time gene FLC in Brassica rapa. Genetics 162: 1457-1468.
Schulz O. E.. 1936. Cruciferae. In A. Engler, H. Harms [eds.] Die natürlichen Pflanzenfamilien 227-658 Wilhelm Engelmann, Leipzig, Germany.
Schwab B. Folkers U. Ilgenfritz H. Hülskamp M.. 2000. Trichome morphogenesis in Arabidopsis. Philosophical Transactions of the Royal Society of London, B, Biological Sciences 355: 879-883.[CrossRef]
Shimodaira H. Hasegawa M.. 1999. Multiple comparisons of log–likelihoods with applications to phylogenetic inference. Molecular Biology and Evolution 16: 1114-1116.[Web of Science]
Sikes D. S. Lewis P. O.. 2001. PAUPRat: PAUP* implementation of the parsimony ratchet, beta software, version 1 Distributed by the authors at website, http://www.ucalgary.ca/~dsikes/software2.htm.
C. R. Somerville, E. M. Meyerowitz [eds.]. 2002. The Arabidopsis book. [Online], website http://www.aspb.org/publications/arabidopsis/ American Society of Plant Biologists, Rockville, Maryland, USA.
Sweeney P. W. Price R. A.. 2000. Polyphyly of the genus Dentaria (Brassicaceae): evidence from trnL intron and ndhF sequence data. Systematic Botany 25: 468-478.[CrossRef][Web of Science]
Swofford D. L.. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods), version 4 Sinauer, Sunderland, Massachusetts, USA.
Takhtajan A.. 1997. Diversity and classification of flowering plants Columbia University Press, New York, New York, USA.
Warwick S. I. Al-Shehbaz I. A. Price R. A. Sauder C.. 2002. Phylogeny of Sisymbrium (Brassicaceae) based on ITS sequences of nuclear ribosomal DNA. Canadian Journal of Botany 80: 1002-1017.[CrossRef]
Warwick S. I. Black L. D.. 1993. Molecular relationships in subtribe Brassicinae (Cruciferae, tribe Brassiceae). Canadian Journal of Botany 71: 906-918.[Web of Science]
Warwick S. I. Black L. D.. 1994. Evaluation of the subtribes Moricandiinae, Savignyinae, Vellinae, and Zillinae (Brassicaceae, tribe Brassiceae) using chloroplast DNA restriction site variation. Canadian Journal of Botany 72: 1692-1701.[Web of Science]
Warwick S. I. Black L. D.. 1997a. Molecular phylogenies from theory to application in Brassica and allies (Tribe Brassiceae, Brassicaceae). Opera Botanica 132: 159-168.
Warwick S. I. Black L. D.. 1997b. Phylogenetic implications of chloroplast DNA restriction site variation in subtribes Raphaninae and Cakilinae (Brassicaceae, tribe Brassiceae). Canadian Journal of Botany 75: 960-973.[Web of Science]
Zunk K. Mummenhoff K. Hurka H.. 1999. Phylogenetic relationships in tribe Lepidieae (Brassicaceae) based on chloroplast DNA restriction site variation. Canadian Journal of Botany 77: 1504-1512.[Web of Science]
This article has been cited by other articles:
|
|
 |
|
  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]
|
|
|
|
 |
|
 M. A. Lysak, M. A. Koch, J. M. Beaulieu, A. Meister, and I. J. Leitch The Dynamic Ups and Downs of Genome Size Evolution in Brassicaceae Mol. Biol. Evol., January 1, 2009; 26(1): 85 - 98. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. A. Beilstein, I. A. Al-Shehbaz, S. Mathews, and E. A. Kellogg Brassicaceae phylogeny inferred from phytochrome A and ndhF sequence data: tribes and trichomes revisited Am. J. Botany, October 1, 2008; 95(10): 1307 - 1327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Mandakova and M. A. Lysak Chromosomal Phylogeny and Karyotype Evolution in x=7 Crucifer Species (Brassicaceae) PLANT CELL, October 1, 2008; 20(10): 2559 - 2570. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Busch, J. Sharma, and D. J. Schoen Molecular Characterization of Lal2, an SRK-Like Gene Linked to the S-Locus in the Wild Mustard Leavenworthia alabamica Genetics, April 1, 2008; 178(4): 2055 - 2067. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Bosch, K. Heo, M. K. Sliwinski, and D. A. Baum An exploration of LEAFY expression in independent evolutionary origins of rosette flowering in Brassicaceae Am. J. Botany, March 1, 2008; 95(3): 286 - 293. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D. C. Ortiz, E. A. Kellogg, and H. V. D. Werff Molecular phylogeny of the moonseed family (Menispermaceae): implications for morphological diversification Am. J. Botany, August 1, 2007; 94(8): 1425 - 1438. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. E. Schranz, A. J. Windsor, B.-h. Song, A. Lawton-Rauh, and T. Mitchell-Olds Comparative Genetic Mapping in Boechera stricta, a Close Relative of Arabidopsis Plant Physiology, May 1, 2007; 144(1): 286 - 298. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Koch, C. Dobes, C. Kiefer, R. Schmickl, L. Klimes, and M. A. Lysak Supernetwork Identifies Multiple Events of Plastid trnF(GAA) Pseudogene Evolution in the Brassicaceae Mol. Biol. Evol., January 1, 2007; 24(1): 63 - 73. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. D. Bailey, M. A. Koch, M. Mayer, K. Mummenhoff, S. L. O'Kane Jr, S. I. Warwick, M. D. Windham, and I. A. Al-Shehbaz Toward a Global Phylogeny of the Brassicaceae Mol. Biol. Evol., November 1, 2006; 23(11): 2142 - 2160. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. A. Kellogg Progress and challenges in studies of the evolution of development J. Exp. Bot., October 1, 2006; 57(13): 3505 - 3516. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
