Arabidopsis family ties: Molecular phylogeny and age estimates in Brassicaceae

Abstract

The Brassicaceae family is of great scientific interest because it contains the plant model organism Arabidopsis thaliana. Currently, contemporary plant research activities expand to other Brassicaceae taxa. Despite that, the phylogeny of this family is only partly understood. The present study deepens our understanding of a family-wide phylogeny by using two new approaches in phylogenetic family-wide research. We used a molecular marker from the mitochondrial genome and utilised a relaxed molecular dating method. Our data generally confirms a recent tribal alignment of Brassicaceae. We present for the first time a biogeographical scenario for the broad-scale Brassicaceae evolution. We suggest that Brassicaceae most likely evolved some 19 mya in or near the eastern Mediterranean region from a common ancestor of its sister family Cleomaceae. The early Brassicaceae formed a lineage adapted to more open/drier habitats than its capparoid progenitors. The early Brassicaceae evolution was very rapid and the radiation was most likely driven by climatic changes that created open habitats and the well-documented expansion of open grass-dominated ecosystems. Moreover, our dating suggests that the radiation events correlate with an ancient genome duplication in the early history of the family, which is evidenced by recent genomic studies.

Full text

425
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
INTRODUCTION
Modern advances in plant biology are unthinkable
without Arabidopsis thaliana (L.) Heynh., or thale cress.
This model organism, together with about 3,700 other
species in about 338 genera, belongs to the mustard family
(Brassicaceae or Cruciferae), a monophyletic group dis-
tributed worldwide. In addition to this “supermodel”, the
family also comprises many ornamentals (e.g., Matthiola),
numerous weeds (e.g., Thlaspi arvense, Capsella bursa-
pastoris), and numerous economic plants (e.g., cabbage,
canola, mustard). The Brassica oleracea L. complex has
become another, well-established model system in plant
research (Koch & al., 2003; Koch & Mummenhoff, 2006).
Using the broad-based knowledge from Arabidopsis and
Brassica, research activities has expanded to the study of
genome evolution (Lysak & al., 2006), evolution of self-
incompatibility systems (
Sherman-Broyles & al., 2007
),
or genetic basis of morphological characters (Bowman,
2006) in distantly related genera of Brassicaceae.
In order to fully understand the evolution of vari-
ous morphological, physiological, or molecular traits and
their convergences, a sound phylogenetic framework is
of paramount importance. Indeed, a robust, family-wide
phylogeny has long been missing. However, a major step
towards such phylogeny has been achieved by Beilstein
& al. (2006) who analysed the chloroplast (cp) gene ndhF
in 113 species of 101 genera in 17 of the 19 tradition-
ally circumscribed tribes of the family. In their study, 97
of 114 ingroup accessions are attributed to 21 strongly
supported clades some of which are grouped into three
monophyletic lineages of higher order and with strong to
moderate support. As in previous, though smaller-scale,
phylogenies within Brassicaceae (summarized in Koch
& al., 2003), the work of Beilstein & al. (2006) had em-
phasized three characteristic outcomes: (1) Most of the
traditionally recognized and widely used tribes sensu
Schulz (1936) are not monophyletic. (2) Many morpho-
logical characters independently evolved several times
in Brassicaceae. For example, Beilstein & al. (2006) re-
vealed that branched trichomes are homoplasious in the
family, as they appeared in several unrelated tribes. As
shown by Al-Shehbaz & al. (2006), the traditional use of
such morphological characters has undoubtedly lead to
highly unnatural classification systems. (3) Except for
the well-supported sister-group relationship of the genus
Aethionema to the rest of the family, there is a lack of reso-
lution along the backbone of the phylogenetic trees. These
Arabidopsis family ties: molecular phylogeny and age estimates in
Brass icaceae
Andreas Franzke1,4, Dmitry German2, Ihsan A. Al-Shehbaz3 & Klaus Mummenhoff
4
1 Present address: Botanical Garden, University of Heidelberg, Im Neuenheimer Feld 340, 69120
Heidelberg, Germany. afranzke@hip.uni-heidelberg.de (author for correspondence)
2 South-Siberian Botanical Garden, Altai State University, Lenin Str. 61, 656049, Barnaul, Russia
3 Missouri Botanical Garden, P.O. Box 299, St. Louis, Missouri 63166–0299, U.S.A.
4 FB Biologie/Chemie, Botanik, Universität Osnabrück, Barbarastraße 11, 49069 Osnabrück, Germany
The Brassicaceae family is of great scientific interest because it contains the plant model organism Arabidopsis
thaliana. Currently, contemporary plant research activities expand to other Brassicaceae taxa. Despite that, the
phylogeny of this family is only partly understood. The present study deepens our understanding of a family-
wide phylogeny by using two new approaches in phylogenetic family-wide research. We used a molecular
marker from the mitochondrial genome and utilised a relaxed molecular dating method. Our data generally
confirms a recent tribal alignment of Brassicaceae. We present for the first time a biogeographical scenario
for the broad-scale Brassicaceae evolution. We suggest that Brassicaceae most likely evolved some 19 mya in
or near the eastern Mediterranean region from a common ancestor of its sister family Cleomaceae. The early
Brassicaceae formed a lineage adapted to more open/drier habitats than its capparoid progenitors. The early
Brassicaceae evolution was very rapid and the radiation was most likely driven by climatic changes that cre-
ated open habitats and the well-documented expansion of open grass-dominated ecosystems. Moreover, our
dating suggests that the radiation events correlate with an ancient genome duplication in the early history of
the family, which is evidenced by recent genomic studies.
KEYWORDS: Brassicaceae, molecular systematics, mtDNA, nad4 intron 1, phylogeny
PHYLOGENETICS

426
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
findings of Beilstein & al. (2006) confirm the conclusions
reached in previous molecular studies (Galloway & al.,
1998; Zunk & al., 1999; Koch & al., 2000, 2001, 2007;
Hall & al., 2002) using different genome markers. Pri-
marily based on the phylogenetic analysis of Beilstein &
al. (2006), Al-Shehbaz & al. (2006) proposed a new tribal
alignment for the Brassicaceae and assigned 260 genera,
some of which tentatively, to 25 tribes, including 7 new.
In the present paper, we test this new tribal concept
(Al-Shehbaz & al., 2006) by the analysis of at least two
representatives of each new tribe for the first intron of the
mitochondrial (mt) gene for NADH subunit 4 (nad4). We
chose this particular phylogenetic marker for the following
reasons: As this marker stems from a different organelle it
is independent from the cpDNA marker used in the work
of Beilstein & al. (2006). A nuclear gene marker that is
easy to handle and promising is not currently available.
The widely used ITS sequences, with many sequences
also available for Brassicaceae taxa, are too variable for
a family-wide analysis resulting in ambigious alignment
and noisy datasets as was recently shown by Bailey & al.
(2006). The generated parsimony tree from the ITS se-
quences showed a considerably high amount of homoplasy
(consistency index [CI] = 0.16). The use of other nuclear
markers established in smaller-scale phylogenetic analyses
within the Brassicaceae, like the pistillata gene (Bailey
& Doyle, 1999), arginine decarboxylase gene (Galloway
& al., 1998), the chalcone synthetase or the alcohol de-
hydrogenase (Koch & al., 2000, 2001), is generally prob-
lematic due to paralogy within gene families. Too
little is
known about the number of loci in such gene families
within different lineages of the family. Charlesworth
& al. (1998) demonstrated that species of the
genus
Leavenworthia have three alcohol dehydrogenase
loci,
whereas Arabidopsis thaliana has only one. The increase
of gene loci can be generated by genome duplications.
Ancient polyploidy events, which lead to the duplication
and triplication of genomes (and loci), generally played
an important role in the evolution of Brassicaceae. Based
on the genome analysis of Arabidopsis, several authors
(e.g., The Arabidopsis Genome Initiative, 2000;
Vision &
al., 2000; S
imillion & al., 2002; Blanc & al., 2003; Bow-
ers & al., 2003;
Schranz & al., 2006
) suggested that the
evolution of Brassicaceae is characterized by at least one
complete ancient genome duplication. Lysak & al. (2005)
provided evidence that members of the tribe Brassiceae
evolved from an ancestor with chromosomal triplication.
Moreover, all phylogenies based on these nuclear markers,
however, did not provide any basal resolution.
Yang & al. (1999), who were the first to conduct a
phylogenetic analysis of Brassicaceae using the first intron
of the mt gene for nad4, included representatives of eight
genera. They showed that the nucleotide substitution rate in
the first intron of nad4 is about 23 times slower than in the
nuclear ITS sequences. This low substitution rate should
potential ly le ad to re solution of the backbone of the Bra ssi-
caceae phylogeny and a reduction of homoplasy in family-
wide analyses. It is well established that the mitochondrial
genome is inherited through the maternal lineage (Hu &
al., 2008 and references therein), thus, resulting trees rep-
resent maternal phylogenies. Whereas frequent recombina-
tion characterizes f lowering plant mitochondrial genomes,
some mitochondrial gene arrangements may, in contrast,
be conserved between streptophyte algae and early land
plant clades (Wahrmund & al., 2008). One example is the
first intron in nad4, which we used in our analysis and
was reported to be present in all flowering-plant species
examine d so fa r (Itchoda & al., 2002). In a ddition to coding
sequences, it has been shown that mitochondrial introns
also represent valuable sources of polymorphic markers
because they seem to have fixed nucleotide and indel mu-
tations more readily than coding sequences (Laroche &
al., 1997; Bakker & al., 2006). Along with markers from
the nuclear and chloroplast genome, recent analyses pro-
vided evidence that mtDNA sequences represent a third
independent genomic estimate of angiosperm relationships
useful for resolving a variety of phylogenetic questions
among both ancient and more recently evolving lineages
(Wahrmund & al., 2008; Knoop, 2004; Barkmann & al.,
2004 and references therein).
As a broad-based phylogenetic framework for Bras-
sicaceae is still in its infancy, the ndhF phylogeny of
Beilstein & al. (2006), and the new tribal classification
(Al-Shehbaz & al., 2006) on which it was based, were im-
portant first steps. In the next step, we test this new tribal
classification using an independent molecular marker
from the mitochondrial genome.
The present study includes 49 out of the 338 genera of
Brassicaceae, representing a broad spectrum of the taxo-
nomic variation in the family and all tribes of the latest
phylogenetic classification (Al-Shehbaz & al., 2006; Beil-
stein & al., 2006). The selection of four representatives
of the tribe Euclidieae and two for each of the remaining
24 tribes sensu Al-Shehbaz & al. (2006) is appropriate to
address the recognition of relationships among the major
phylogenetic lineages within the family. This is the first,
broader-scale phylogenetic analysis of the family using
mitochondrial DNA, and it follows two additional family-
wide studies (Bailey & al., 2006; Koch & al., 2007), which
will be considered in the discussion.
In addition we provide a more advanced time es-
timate, using an elaborated (relaxed) molecular dating
method (Drummond & al., 2006) to infer key dates in
the Brassicaceae evolution. Previo us studies dealing with
age estimations in Brassicaceae were based on rather
simplistic approaches. Interestingly, nearly all those
studies were based on the age estimations presented
by Yang & al. (1999). As the focus in the present paper

427
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
is on the overall Brassicaceae phylogeny, we dated the
following nodes of our mtDNA analysis: (1) The age of
the Cleome/Brassicaceae split to estimate the age of the
family. Recent molecular phylogenetic studies on the or-
der Capparales (Hall & al., 2002; Sánchez-Acebo, 2005)
revealed a sister group relationship of Brassicaceae and
Cleomaceae ( = Capparaceae subfamily Cleomoideae).
(2) The basal polytomy of Brassicaceae excluding the
tribe Aethionemeae. (3) Lineage I sensu Beilstein & al.
(2006), including the tribes Camelineae, Boechereae, Ha-
limolobeae, Physarieae, Cardamineae, Lepidieae, Descu-
rainieae, and Smelowskieae. This lineage was identified
in all higher-level Brassicaceae phylogenies (Bailey &
al., 2006; Beilstein & al., 2006; Koch & al., 2007). By
integrating data from phylogenetics, molecular dating,
genomics, biogeography, and palaeoecology/climatology
we propose a reasonable scenario for the large-scale evo-
lution of Brassicaceae.
MATERIALS AND METHODS
Data collection. —
We sampled 49 accessions of
Brassicaceae for the mitochondrial nad4 intron (Appen-
dix). Sequence data for three Brassicaceae taxa (Arabidop-
sis thaliana (L.) Heynh., Brassica nigra (L.) Koch, Car-
damine scutata Thunb.), and those of the three outgroup
species (Moringa oleifera Lam., Moringaceae; Capparis
lanceolaris DC., Capparaceae; Cleome viscosa L., Cleo-
maceae) were taken from GenBank (Y08501, AF095247,
AF095252, AF451593, AF451591, AF451588).
The ingroup samples include four representatives
of the tribe Euclidieae and two each of the remaining
24 tribes sensu Al-Shehbaz & al. (2006). The intron se-
quences of Diptychocarpus strictus (Fisch. ex M. Bieb.)
Trautv. and Chorispora tenella (Pall.) DC. were very
short (515 and 499 base pairs, respectively) in compari-
son with the remaining analysed taxa. Here, the intron
length varied from 1,268 bp (Aethionema arabicum) to
1,651 bp (Heliophila linearis), with an arithmetic mean of
1,460 bp. The nad4 intron sequences were aligned by eye
and regions of alignment ambiguity were removed prior
to phylogenetic analysis. The aligned data matrix consists
of 1,491 characters.
DNA extraction and PCR amplification. —
DNAs
from Moriera spinosa, Nevada holmgrenii, Descurainia
stricta, Ianhedgia minutif lora, Myagrum perfoliatum,
Noccaea cochleariformis, Taphrosp e rmum altaicum,
Oreoloma violaceum and a species of Heliophila were
kindly provided by Mark Beilstein and correspond to the
taxa used in the ndhF phylogeny (Beilstein & al., 2006).
The remaining DNAs were extracted mainly from her-
barium material using the DNA isolation kit NucleoSpin
Plant (Macherey-Nagel, Germany) and the included buffer
C1. The nad4 intron was PCR amplified using primers
that were used by Yang & al. (1999), and in most cases
in combination with internal primers designed for this
study: internal forward primer 1 (IF) (5′-AAGGGGT
GCTCCTAGGTGTG-3′), internal forward primer 2 (IIF)
(5′-ACAAGGGCCGACGACGACGGAAG-3′), internal
reverse primer 1(IIR) (5′-CCGTTACCTGAATTCGC
GCA-3′), internal reverse primer 2 (IR) (5′-GTACGT
GAGACTTCCGCATC-3′). Twenty-five micro-liters PCR
reactions were performed with 1 μg DNA in a master mix
containing 1
×
PCR buffer (10 mM TRIS-HCl, 50 mM KCl,
0.08% Nonidet P40, 2.5 mM MgCl
2
,
pH 8.8), 0.15 mM
dNTP (Biozym), 10 pmol of each primer and 0.625 U of
Taq polymerase (Fermentas, Germany) using a PTC-200
(Biozym, Germany) thermal cycler. The temperature pro-
file included an initial denaturing step of 4 min at 94°C; 40
cycles of 1 min at 94°C for DNA denaturation, 45 s at 55°C
for primer annealing, and 2 min at 72°C for primer exten-
sion; a final extension of 20 min at 72°C; and a 4°C soak.
The PCR products were then run on 1% agarose gels. Suc-
cessful reactions were purified using the NucleoSpin Ex-
tract II Kit of Macherey-Nagel (Germany). These products
were sequenced at the University of Osnabrück on an ABI
Prism 377 and an ABI 310 automated sequencer (Applied
Biosystems, Germany) with dye terminator chemistry. The
primers of Yang & al. (1999) were used as the external
sequencing primers often along with two nested prim-
ers designed in this study: nested forward primer (NF2
= 5′-TTCCCGAAAGCGTGCCAATC-3′) and nested
re verse primer (NR1 = 5′-CTGATATGCTGCCTTGAT
CT-3′). The internal parts of the nad4 in tr on were se quence d
using internal forward primer 1 (IF = 5′-AAGGGGT
GCTCCTAGGTGTG-3′), internal forward primer 2 (IIF =
5′-ACAAGGGCCGACGACGGAAG-3′), internal reverse
primer internal reverse primer 1 (IIR = 5′-CCGTTACCT
GAATTCGCGCA-3′), and internal reverse primer 2 (IR
= 5′-GTACGTGAGACTTCCGCATC-3′).
Phylogenetic analyses. —
Phylogeny was esti-
mated using parsimony and Bayesian methods. For the
parsimony analysis, 15 replicates of 200 parsimony ratchet
iterations (Nixon, 1999) were implemented using PRAP
(Müller, 2004) in PAUP* version 4.0b10 (Swofford, 2000),
with 15% of characters re-weighted at each iteration with
10 random addition cycles. Multiple maximally-parsi-
monious trees were summarized with a strict consensus
tree. Moringa oleifera served as the outgroup taxon in
the parsimony analysis. In order to assess the amount of
homoplasy, the consistency index CI (Kluge & Farris,
1969) were computed in PAUP*. Support for clades was
estimated by parsimony jackknifing in PAUP*, emulating
Jac resampling and eliminating 37% of the characters in
each of 10,000 repl icates, with tree bisection-recon nection
(TBR) swapping and MulTrees on. We performed a single-
search replicate saving two trees per jackknife replicate.

428
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
Such search strategies were found to represent a reason-
able balance between computational effort and reliable
results (Freudenstein & al., 2004).
The model of DNA substitution for the Bayesian anal-
ysis of phylogeny t hat best fit ted the data was determined
under MrModelTest 2.2 software (Nylander, 2004) and the
Akaike information criterion (AIC). Bayesian inference of
phylogeny was performed using MrBayes 3.1 (Ronquist
& Huelsenbeck, 2003). Following MrModelTest, the sym-
metrical model of sequence evolution (SYM) was em-
ployed in MrBayes, with an allowance for a gamma (G)
distribution of rates and a proportion of invariant sites (I).
Two simultaneous analyses were run, each for ten million
generations, on four parallel chains, with a temperature for
the heated chain set to 0.5. The values sampled for differ-
ent parameters were examined using Tracer 1.3 (Rambaut
& Drummond, 2003). For each MrBayes analysis, every
100th tree was sampled and, after analysing the param-
eter values in Tracer, 25,000 initial trees were discarded
as “burn-in”.
Molecular dating. —
A basic prerequisite for the
molecular dating method is a sound external calibration,
most often based on fossil evidence. Unfortunately, reli-
able Brassicaceae fossils for calibrating nodes are prob-
lematic. Known pollen fossils from the Miocene cannot
be attributed to distinct Brassicaceae taxa, whereas deter-
minable fossils from the Pleistocene only reflect minimum
ages and are not useful in a family-wide context. The old-
est known Brassicaceae macrofossil available (Rorippa,
2.5 –5 mya ; Mai, 1995) can b e us ed t o indicat e the Rorippa/
Cardamine split. This rationale was used by Koch & al.
(2000) for molecular dating in the Brassicaceae. As nad4
intron 1 analysis yielded virtually no sequence difference
between Rorippa and Cardamine (only three nucleotide
substitutions out of 1,491 positions) this approach appears
problematical. Relaxed dating analyses (following Drum-
mond & al., 2006) using this calibration yielded an un-
reasonably high age for the Brassicaceae (mean: 64 myr;
95% highest posterior density [HPD]: 6.34 myr to 187.59
myr). Indeed, molecular dating results relying on recent
calibration points in combination with very low sequence
divergence are generally prone to erroneous dating (Kay
& al., 2006). Therefore, calibration estimates will need to
come f rom node estimates obtained from fossil-calibr ated
published phylogenies. We chose a calibration point based
on an age estimate of a node within the order Brassicales.
Wikström & al. (2001) estimated the split between the
Moringaceae and Brassicaceae to be ca. 70 mya. There is
at least “indirect” fossil evidence, that this node estimate
has a correct order of magnitude. Dressiantha, which is
the oldest known fossil flower (Turonian, Upper Creta-
ceous: ca. 90 mya) that has been clearly assigned to the
Brassicales, shows strong aff inities to Moringa (
Gandolfo
& al., 1989
).
Because a likelihood ratio test using MrModeltest 2.2
(Nylander, 2004) and PAUP* (Swofford, 2000) signifi-
cantly rejected a global molecular clock for our dataset, we
chose a relaxed clock approach to infer the ages of several
nodes in Brassicaceae evolution. The molecular dating
analyses were performed in BEAST v1.4 (Drummond
& Rambaut, 2003) following a relatively new approach
of Drummond & al. (2006) under a relaxed-clock model
using the BEAST user interface BEAUti to create the
BEAST input file. In this approach, a relaxed-clock model
is incorporated into a Bayesian phylogenetic inference.
The analyses were performed with a relaxed molecu-
lar clock model drawn uncorrelated rates from a log-nor-
mal distribution. We created a uniform prior on the root
height with 70 mya upper and 70.001 mya as lower bound
in order to calibrate the split Moringa versus Brassicaceae
to 70 mya (see above). Additionally, Moringa was con-
strained to be treated as the outgroup taxon. We used the
Yule speciation process as tree prior and needed a user–
defined starting tree. As the SYM model is not included
in BEAST v1.4, we chose the more general “general time
reversible model” GTR with a gamma (G) distribution
with four gamma categories of rates and a proportion of
invariant sites (I). After tuning the operators using the
auto-optimization option, the final analysis was done with
the MCMC chain length set to ten millions. These settings
resulted in ESS values, determined in Tracer, for all es-
timated parameters and node ages above 100, indicating
a sufficient posterior distribution quality. The sequences
of Diptychocarpus strictus and Chorispora tenella were
removed from the alignment in the dating analysis due to
their very short sequence length.
RESULTS
Phylogenetic analysis of nad4 intron sequence
data. —
The aligned data matrix (excluding areas of am-
biguous alignments) consists of 1,491 characters across
52 Brassicaceae species representing all tribes sensu Al-
Shehbaz & al. (2006). Of the 278 variable characters, 129
were potentially parsimony informative. The alignment of
the nad4 intron 1 sequences, which is available from the
authors
upon request
, required the incorporation of nu-
merous larger gaps. Preliminary analysis (data not shown)
revealed that the indel distribution does not contain
phy-
logenetic information.
The parsimony approach resulted
in 212 MPTs of 347 steps with a remarkable low level of
homoplasy in the data (CI 0.905), especially by consid-
ering the high number of taxa studied and the familial
phylogenetic level of the analysis. The Bayesian tree gen-
erated from the same dataset was an almost exact match
of the parsimony tree, with the exception of slight differ-
ences in branching within some terminal nodes (Fig. 1).

429
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
Noticeable differences include the positions of Heliophila
and Diptychocarpus/Chorispora, but the Diptychocarpus/
Chorispora ambiguity probably resulted from their very
short sequences and thus limited number of phylogeneti-
cally informative characters (Fig. 1). These differences,
however, did not have any impact on the phylogenetic
conclusions addressed in this study.
Molecular dating. —
Our relaxed bayesian molecu-
lar-clock dating yielded the following age estimates. The
split between Cleome (Cleomaceae) and Brassicaceae,
which indicates the age of the latter family, was estimated
at 18.541 mya (95% HPD: 1.219 to 45.142 mya). The split
between the most basal lineage (tribe Aethionem eae) and
the remaining Brassicaceae was estimated at 14.549 mya
(95% HPD, 1.024 to 35.388 mya), whereas the age of the
rest of Brassicaceae was estimated at 11.379 myr (95%
HPD, 0.861 to 27.516 myr). Finally we estimated the age
of lineage I sensu Beilstein & al. (2006) to be 7.548 myr
(95% HPD, 0.474 to 18.761 myr). Rounded dating values
are incorporated in Fig. 1 and will be discussed below.
DISCUSSION
mt nad4 intron 1 as phylogenetic marker. —
To
determine tribal relationships in Brassicaceae we have
used mtDNA sequences, which have recently been used
to study plant evolutionary relationships either alone or
in combination with data from other genomic compart-
ments (Barkmann & al., 2004 and references therein). One
reason for the utility of plant mtDNA may be the lower
substitution rates that characterize this genome, which
seem to provide characters with low levels of homoplasy
(Barkmann & al., 2004 and references therein). However,
using mtDNA as phylogenetic marker can be problem-
atic due to the
risk
of mistaking
paralogy
for orthology
as the transfer of nucleic acids from the mitochondria
to the nucleus has been reported (Laroche & al., 1997).
In a pilot study of Yang & al. (1999), the first intron of
the mitochondrial gene for NADH subunit 4 (nad4) was
tested for its phylogenetic use in Brassicaceae. Yang & al.
(1999) sequenced at least three clones from each of the 16
Brassicaceae accessions studied and for 3 accessions of
three species, respectively. All clones of the same acces-
sion and all accessions of one species revealed the same
sequence. In addition to the cloning experiments of Yang
& al. (1999), our experiments do not show any evidence
for paralogue copies of nad4 intron 1 sequences in Bras-
sicaceae taxa. Judged from direct sequencing, we did not
find any evidence (e.g., double bands upon PCR amplifi-
cation, polymorphic nucleotide positions) for duplicated
loci or paralogs. The phylogenetic value of nad4 intron
1 sequences for systematic purposes in Brassicaceae is
also
confirmed
by the present
study itself. The
results of
our family-wide phylogenetic analysis corresponds quite
well with published nuclear and cpDNA phylogenies (for
detailed discussion see below). This finding indirectly
argues for the orthology of the amplified nad4 intron se-
quences including the unusual short sequences of Dipty-
chocarpus and Chorispora tenella. We also performed
phylogenetic analysis without the sequence information of
Diptychocarpus and Chorispora and found no topological
differences.
Recognition of tribes sensu Al-Shehbaz & al.
(2006) with mt nad4 intron 1. —
This discussion is
mainly aimed to compare our findings with the broad-
scale nuclear ITS analysis of Bailey & al. (2006). It in-
cludes all tribes sensu Al-Shehbaz & al. (2006) minus the
tribe Chorisporeae. Although Bailey & al. (2006) per-
formed a “prel imi nar y supermat rix” analysis of dat a from
alcohol dehydrogenase 1, atpB, chalcone synthase, ITS,
matK, ndhF, pistillata intron 1, rbcL, leafy, and trnL-F
for 65 taxa, their supermatrix analysis is tentative at best
because the matrix contains 67% missing data. Fifteen
Brassicaceae tribes sensu Al-Shehbaz & al. (2006) were
recognized as monophyletic entities in our phylogenetic
analyses (see Fig. 1). They are Aethionemeae, Camelin-
eae, Boechereae, Cardamineae, Descurainieae, Lepidieae,
Alysseae, Chorisporeae, Arabideae, Thlaspideae, Noc-
caeeae, Cochlearieae, Hesperideae, Anchonieae, and
Heliophileae. However, as shown by Bailey & al. (2006),
some of these tribes (e.g., Alysseae, Camelineae, Lepi-
dieae, and Arabideae) are likely to be polyphyletic. The
tribe Chorisporeae, monophyletic in our study, was not
analyzed by Bailey & al. (2006) and Euclidieae as delim-
ited by Al-Shehbaz & al. (20 06) is not monophyletic in ou r
tree. Furthermore, Warwick & al. (2007), have also shown
that both Anchonieae and Euclidieae are polyphyletic. In
her study, Warwick & al. (2007) analysed 22 of the 25
gener a assigned by Al-Shehbaz & al. (2006) to Euclidieae
and showed that the tribe is split into two well resolved
clades. Euclidieae s.str (Euclidieae I sensu Warwick &
al., 2007) contained the type genus Euclidium W.T. Ai-
ton along with 13 other genera (e.g., Braya, Tetracme).
Whereas Euclidieae II (sensu Warwick & al., 2007), des-
ignated Malcolmieae by Al-Shehbaz & Warwick (2007),
included the type species of Malcolmia, i.e., M. maritima
(L.) W.T. Aiton and its closest Mediterranean relatives,
e.g., M. littorea (L.) R. Br., and Maresia nana (DC.) Batt.
Our mtDNA analysis strongly supports the maintainance
of Euclidieae (Euclidieae I sensu Warwick & al., 2007)
as a distinct tribe and the recognition of the Malcolmieae
Al-Shehbaz & Warwick (Euclidieae II sensu Warwick &
al., 2007).
Sampled representatives of the tribes Smelowskieae,
Physarieae, Halimolobeae, Brassiceae, Isatideae, and
Eutremeae sensu Al-Shehbaz & al. (2006) appeared un-
resolved within clades of higher order in our analysis

430
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
(Fig. 1). All six tribes were monophyletic in Bailey & al.
(2006) and other more smaller-scale works cited therein.
We suspect that the lack of resolution in our analysis is
the result of the relatively young age of the taxa sampled
and low substitution rates of the chosen mitochondrial
marker, which is approximately 23 times slower (
Ya ng
& al., 1999
) than the ITS marker used by Bailey & al.
(2006). Members of the tribe Sisymbrieae also appeared
unresolved in the current study. This may reflect the
“weak paraphyly” of this tribe as found by Bailey & al.
(2006). Also, both representatives of Schizopetaleae were
not monophyletic in the present study. This is in contrast
to the results of Bailey & al. (2006) and Warwick & al.
(2002).
The tribe Euclidieae clearly appeared polyphyletic in
our phylogenetic analysis. Malcolmia littorea (L.) R. Br.
forms, along with Maresia nana (DC.) Batt., a group un-
related to a clade consisting of Braya rosea Bunge and
Tetracme quadricor nis (Steph.) Bunge. Also, in the analy-
sis of both Bailey & al. (2006) and Warwick & al. (2007),
the Euclidieae sensu Al-Shehbaz & al. (2006) appeared
to be polyphyletic with Braya belonging to one of the
divergent lineages.
Larger tribal groupings within Brassicaceae. —
The clearest deep split within Brassicaceae is the sister-
group relationship between the tribe Aethionemeae and
the rest of the family. This split was confirmed in studies
using different genes and gene sets from the nuclear and/
or the chloroplast genome (Galloway & al., 1998; Zunk &
al., 1999; Koch & al., 2000, 2001; Hall & al., 2002; Bailey
& al., 2006; Beilstein & al., 2006).
As shown in the present phylogenetic analysis, the
mitochondrial DNA genome also provides evidence for
the very basal position of the tribe Aethionemeae, indicat-
ing an Old-World origin of Brassicaceae. When Cleome is
used as the sole outgroup (versus including it with Cap-
paris [Capparaceae] and Moringa [Moringaceae] as the
three outgroups), the sister relationship of Aethionemeae
to the rest of Brassicaceae is supported by higher boot-
strap values (99% versus 83%). The tribe Aethionemeae
consists of the genus Aethionema (ca. 50 spp. centered
in Turkey, with fewer species extending eastward into
Turkmenistan and westward into Spain and Morocco) and
the doubtfully distinct Moriera (monotypic: Afghan istan,
Iran, Turkmenistan). As the family is divided basally into
the species-poor tribe Aethionemeae (ca. 50 spp.) versus
the rest of Brassicaceae (ca. 3,650 spp.), future evolution-
ary studies should include comparative analyses of these
disparate sister groups.
In addition to the sister relationship between the Aeth-
ionemeae and the remaining Brassicaceae, Beilstein & al.
(2006) and Al-Shehbaz & al. (2006) proposed three larger
tribal groupings. Their lineage I, which included the tribes
Camelineae, Boechereae, Halimolobeae, Physarieae, Car-
damineae, Lepidieae, Descurainieae, and Smelowskieae,
is consistent with our mtDNA tree (Fig. 1). A correspond-
ing clade, although not supported, was also confirmed by
Bailey & al. (2006) and Koch & al. (2007). All members
of this lineage analysed so far are characterized by a trnF
(GAA) pseudogene (Koch & al., 2007), however, neither
Halimolobeae nor Physarieae were included in the last
st udy. Furthermore, most of tribes in the lineage are char-
acterized by a base chromosome number of x = 8, and the
notable exception is Boechereae with x = 7.
Lineage II sensu Beilstein & al. (2006) and Al-Sheh-
baz & al. (2006) is monophyletic and includes the tribes
Schizopetaleae, Sisymbrieae, Brassiceae, and Isatideae.
However, the lineage was not supported by our data,
though its representative tribes were placed in a moder-
ately supported clade together with representatives of the
tribes Iberideae, Eutremeae, Thlaspideae, and Noccaeeae
(Fig. 1). The lineage received very low support in the ITS
analysis of Bailey & al. (2006). Except for the tribe Noc-
caeeae, this larger tribal grouping corresponds to a weakly
supported clade in Beilstein & al. (2006) and Al-Shehbaz
& al. (2006).
Lineage III sensu Beilstein & al. (2006) and Al-
Shehbaz & al. (2006) consists of the tribes Euclidieae,
Anchonieae, Hesperideae, and Chorisporeae. This lin-
eage also appeared in the parsimony analysis of our data,
but in the Bayesian analysis only the monophyly of Eu-
clidieae, Anchonieae, and Hesperideae is strongly sup-
ported (Fig. 1). Bailey & al. (2006) also determined a clade
uniting Anchonieae and Hesperideae with Cochlearieae,
Fig. 1. Comparison of phylogenetic relationships among tribes of Brassicaceae sensu Al- Shehbaz & al. (2006) and a
mtDNA phylogeny based on the first intron of NADH subunit 4 sequences. Shown is the 50% majority-rule consensus
trees of 75,001 post-burn-in trees from a Bayesian analyses and the strict consensus of 212 most parsimonious trees
(347 steps, CI = 0.905), respectively. Posterior probabilities > 95 are given above the branches. Numbers below branches
represent jackknife values > 50% of the maximum parsimony analysis. The numbers in brackets refer to the number of
genera assigned to the tribes followed by the distribution area of the tribes with minor distribution areas given in brackets.
“Lineage I” refers to a tribal grouping detected by Beilstein & al. (2006) and other studies. *Al- Shehbaz & al. (2006) did not
recognize Euclidieae I and II, represented in our analysis by Braya/Tetra cme and Malcolmia littorea/Maresia, respectively,
as was suggested by Warwick & al. (2007). For further details see Discussion. Age estimates for nodes based on a relaxed
molecular clock approach are indicated by arrows. For complete taxon names see Appendix. The 95% HPD for the dating
of several clades are given in brackets.

431
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
Phylogenetic relationships Mitochondrial DNA phylogeny
sensu Al-Shehbaz & al. (2006)
Moringaceae Moringa
Capparaceae Capparis
Cleomaceae Cleome
Aethionemeae (2) Aethionema
Middle East Moriera
Camelineae (12–13) Arabidopsis
Eurasia (America, Australia) Camelina
Boechereae (7) Boechera
N America Nevada
Halimolobeae (5) Halimolobus
New World Penellia
Physarieae (7) Physaria
N America (S America) Lyrocarpa
Cardamineae (10) Cardamine
All continents except Antarctica Selenia
Lepidieae (3-5) Lepidium
All continents except Antarctica Stroganowia
Descurainieae (6) Descurainia
Americas, Eurasia, Africa Ianhedgea
Smelowskieae (2) Smelowskia
E/C Asia (N America) Hedinia
Alysseae (17) Alyssum
Eurasia, N Africa Berteroa
Schizopetaleae (20) Strephanthus
New World (S Indian Ocean) Thelypodium
Sisymbrieae (1) Sisymbrium incl.
Eurasia, Africa (N America) Schoenocrambe
Brassiceae (20) Mediterranean, Brassica
SW Asia (S Africa, N America) Cakile
Isatideae (8) SW Asia , Isatis
Middle East (Europe, N Africa) Myagrum
Eutremeae (1) Eutrema
Asia (N America) Taphrospermum
Thlaspideae (7) Thlaspi s.str.
Europe, SW Asia Alliaria
Arabideae (6) Arabis
Eurasia and N America (S America) Draba
Noccaeeae (3) Noccaea
Eurasia, N Africa (New World) Microthlaspi
Iberideae (1) Europe Iberis
(NW Africa, Turkey, SW/C Asia) Iberis
Cochlearieae (1) Europe, Cochlearia incl.
NW Africa, N America, Asia Ionopsidium
Heliophileae (1–2) Heliophila
S Africa Heliophila
Malcolmia
Euclidieae (25) Maresia
Eurasia, N/E Africa (N America) Braya
Tetracme
Anchonieae (12) Anchonium
Eurasia, N/E Africa (N America) Oreoloma
Hesperideae (1) Hesperis
Middle East, Europe (N Africa, C Asia) Hesperis
Chorisporeae (3) Diptychocarpus
Asia Chorispora
97
99
99
99
96
100
99
84
99
93
98
99
98
99
71
62
98
99
96
99
89
98
56
99
99
100
99
100
100
95
100
100
99
100
90
/
Lineage I 19 mya
(95% HPD: 1–45 mya)
11 mya
(95% HPD: 1–28 mya)
8 mya
(95% HPD:
0.5–19 mya)
15 mya
(95% HPD: 1–35 mya)
“Bayes”
“Bayes”
84
63
83
70
/
58
93 /
/
99
/
98
100
66
81
83
85
100
/
*

432
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
Heliophileae, and Thlaspideae. Their analysis did not in-
clude members of the Chorisporeae.
To summarize: The mtDNA phylogeny (Fig. 1) gen-
erally confirms the tribal concept of Al-Shehbaz & al.
(2006). Based on our current knowledge (Al-Shehbaz &
al., 2006; Bailey & al., 2006; Beilstein & al., 2006; Koch
& al., 2007; Warwick & al. 2007, 2008) the synoptic
conclusions of this issue are: (1) Aethionemeae is sister
to the rest of the family. (2) Lineage I sensu Beilstein &
al. (2006) was identified in all higher-level Brassicaceae
phylogenies. This lineage comprises the tribes Cam-
elineae, Boechereae, Halimolobeae, Physarieae, Car-
damineae, Lepidieae, Descurainieae, and Smelowskieae
sensu Al-Shehbaz & al. (2006). (3) Schizopetaleae,
Sisymbrieae, Brassiceae, and Isatideae are somewhat
closely related, most probably together with Iberideae,
Eutremeae, and Thlaspideae. (4) Euclidieae, Anchonieae
and, most probably, Hesperideae form a phylogenetic
lineage.
Backbone phylogeny of Brassicaceae. —
Schol-
ars of the Brassicaceae agree that the tribe Aethionemeae
is the basalmost lineage in the family. The deep phylogeny
within the rest of the family, however, remains unclear.
The present study agrees with the earlier findings (e.g.,
Al-Shehbaz & al., 2006; Bailey & al., 2006, Beilstein &
al., 2006; Koch & al., 2007) that the phylogenetic rela-
tionships among the larger tribal groupings were either
undetected or with poor support.
Cochlearieae was the second basalmost tribe in the
supermatrix analysis of Bailey & al. (2006), but it ap-
peared in a more derived position in the ITS tree of that
work. The tribe was also assumed the second basalmost
position in the super-network reconstruction of Koch & al.
(2007). But based solely on cpDNA sequences, Cochlearia
appeared in a derived position in a highly supported clade
that included Lobularia lybica (tribe Alysseae).
Al-Shehbaz & al. (2006) and Bailey & al. (2006) sug-
geste d that t he lack of resolution was probably due to m ajor
radiation(s) in the early evolutionary history of the family.
However, the high amount of homoplasy in the data of
Beilstein & al. (2006) and Bailey & al. (2006), as indicated
by the low consistency indices of 0.31 and 0.16, respec-
tively, could also be a potential source for low resolution.
Our mtDNA phylogeny also does not resolve the back-
bone phylogeny in Brassicaceae excluding Aethionemeae
(Fig. 1), and we also argue that this lack of resolution is
due to sudden radiation. However, because of the relatively
low amount of homoplasy in our data (CI = 0.9), we can
rule out homoplasy as the prominent factor for low basal
resolution within the family (see Rokas & Carroll, 2006
for the influence of excess homoplasy on the resolution of
ancient divergences). Taking into account the results of
Bailey & al. (2006) and Beilstein & al. (2006), we suggest
that the inclusion of more taxa into the mtDNA dataset
would dramatically increase the number of those taxa fall-
ing in the basal polytomy of Brassicaceae.
Molecular dating. —
Our age estimation for the
whole Brassicaceae based on the mitochondrial DNA is
19 myr, and compared to previous studies, this age esti-
mation is rather young. However, the 95% HPD interval
overlaps with published ranges for the age of the family.
The highest published time estimations for the minimal
age of the family are 30–60 myr (Koch & al., 2000) and
40 myr (Koch & al., 2001). Based on estimates for the
divergence of Brassica from Arabidopsis of ca. 20 mya
(
Yang & al., 1999;
Koch & al., 2001), Schranz & al.
(2006) approximate the divergence of Brassicaceae from
its siste r family Cleomaceae to be 41 mya. The same order
of magnitude is implied by the analysis of Ermolaeva & al.
(2003), who suggested that the genome duplication in Ara-
bidopsis thaliana occurred some 38 to 30 mya. But this
study is also based on the estimation of the Arabidopsis-
Brassica split of previous works (e.g.,
Yang & al., 1999;
Koch & al., 2000, 2001).
Henry & al. (2006) gave a comparable minimal age
range and suggested the ancient polyploidization event
in the early Brassicaceae evolution took place 24 to 40
mya. The basal polytomy of Brassicaceae is most likely
due to an ancient rapid radiation event, the age of which
we estimate at 11 mya. This split reflects the date of the
last common ancestry of Arabidopsis and Brassica spe-
cies, the two genera most widely used as Brassicaceae
model organisms. Similar age estimates (14–20 myr) of
that split were published (
Yang & al., 1999;
Koch & al.,
2000, 2001). The recognition of lineage I is consistent
with recent larger scale phylogenetic studies (Bailey &
al., 2006; Beilstein & al., 2006; Koch & al., 2007). The
basic chromosome number of this group is x = 8. How-
ever, chromosomal rearrangements led to the reduced
genome of Arabidopsis thaliana with n = 5 (Lysak & al.,
2006). Based on the comparative genome mappings be-
tween two Arabidopsis and one Capsella species (Came-
lineae), Koch & Kiefer (2005) estimated that the ances-
tors of Caps ella and Arabidopsis diverged approximately
10–14 mya. Although they did not explicitly mention the
most recent ancestor of Capsella and Arabidopsis, their
age estimate is in the same order of magnitude as ours
(8 myr) for lineage I sensu Beilstein & al. (2006) that
includes both Capsella and Arabidopsis. Interestingly,
the age range of this lineage (8–14 myr), as determined
by molecular data, falls within the Upper Miocene (ca.
11–5 mya), as do the oldest reliable fossil records for the
Brassicaceae (Muller, 1981).
Biogeographic scenario for Brassicaceae evolu-
tion. —
Based on our current knowledge of a family-
wide phylogeny and age estimations, we propose a likely
biogeographic scenario that reflects the large-scale evo-
lution of the family. Hayek (1911) was among the first

433
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
to suggest that Brassicaceae were derived from Cappa-
raceae subfamily Cleomoideae. Recent molecular phylo-
genetic studies on the order Capparales (Hall & al., 2002;
Sánchez-Acebo, 2005) revealed three well-supported line-
ages: Capparaceae subfamily Capparoideae and the sister
taxa subfamily Cleomoideae and Brassicaceae. Hall & al.
(2002) suggested the recognition of these three as distinct
families (Capparaceae, Cleomaceae, Brassicaceae), a posi-
tion now widely accepted (e.g., APG II, 2003). The sister
relationship of Brassicaceae and Cleomaceae is also sup-
ported by our data, and their relationship to Capparaceae
is taken into account herein.
Species of both Cleomaceae and Capparaceae are
widespread in the (sub)tropical and warm-temperate re-
gions of the world, but some occur in dry habitats. Hall
& al. (2002) suggested a paleotropical origin for Cappa-
raceae, but the origin of Cleomaceae remained open. Re-
cently, an Old-World orig in (Af rica, Mediterranea n) of the
Cleomaceae has been suggested (Inda & al., 2008). The
bulk of extant Cleomaceae species grow in the tropics.
This would imply that the last common ancestor of Bras-
sicaceae and Cleomaceae was adapted to a warm, humid
(sub)tropical climate. We dated the split between Cleome
and the Brassicaceae, and thus the age of this family, to
19 myr (see above), which indicates a Lower Miocene
(ca. 24–16 mya) origin. As shown above, the deepest split
of the tribe Aethionemeae from the rest of Brassicaceae
was about 16 mya. When comparing this split (15 mya)
with the estimated age for the whole Brassicaceae (19
myr), the diversification of Aethionemeae versus rest must
have occurred during the very early evolutionary history
of the family at the end of Miocene. The vast majority
of Aethionemeae occurs in Turkey and perhaps not too
far from the ancestral area of the family. The centers of
greatest diversity of Brassicaceae in the Old World are
the Irano-Turanian and adjacent Mediterranean regions
(Hedge, 1976; Al-Shehbaz, 1984), and almost all of Turkey
falls within these two regions. In fact, Turkey is one of the
richest crucifer countries in the world and has about 560
species (Al-Shehbaz & al., 2007). On the basis of early
phylogenetic branching discussed above and current dis-
tributional data, we agree with Hedge (1976) and Rollins
(1993) in considering the Irano-Turanian as the cradle of
the family. One might suggest that Brassicaceae evolved
and diversified elsewhere and that the present–day high
diversification in the Irano-Turanian region represents the
remnants of a wider distribution. However, there is no line
of evidence that supports that.
The Middle Eocene (49–37 mya) was the time of
greatest expansion of the palaeotropical flora into the
Northern Hemisphere, including the Turkish palynofloras
(Akgün & al., 2002). The tropical climate was more or
less uniform until almost the end of Eocene (ca. 34 mya).
The climate then became cooler and perhaps drier in
Early Oligocene (ca. 34–29 mya), leading to a prevalent
warm subtropical climate and flora and continuing until
Late Oligocene-Miocene (Akgün & al., 2007). An eco-
logical analysis of Late Oligocene-Miocene palynomo rph
assemblages from Turkey (Akgün & al., 2007) revealed
several ecologically diverse paleo-associations. Based
on our estimated birth date for Brassicaceae (19 mya),
we suggest a lower Miocene origin of the family from a
“capparoid” ancestor adapted to warm and humid habi-
tats in the east Mediterranean area. Recent Brassicaceae
species occupy various habitats but primarily open, drier
ones (Appel & Al-Shehbaz, 2003), and it is likely that
their ancestors did the same. No palynofloras are known
from the early Burdigalian period (ca. 20–16 mya) of
Turkey to demonstrate the existence of open vegetation.
However, typical open vegetation elements (e.g., Cheno-
podiaceae, Poaceae) were reported from the latest Bur-
digalian (Akgün & al., 2007) and indicate also suitable
habitats for cruciferous plants. Fossil Brassicaceae pollen
in this palynoflora might not be documented because
plants of the family are not wind-pollinated, and their
pollen is not readily and widely dispersed as are those of
Chenopodiaceae and Poaceae. A mosaic pattern of sub-
tropical and drier open areas, indicated by the palynoflora
of the latest Burdigalian (ca. 16 mya), supports the idea
that the typically open and dry-adapted Brassicaceae are
the radiation products of humid-adapted Capparaceae/
Cleomaceae ancestors. Unfortunately, there is no fos-
sil evidence of the three families in the Mediterranean
region, but capparaceous Bartonian fossils are known
from the Middle Eocene (ca. 41–37 mya) of the British
Isles (Chandler, 1960) and the Upper Eocene (ca. 37–34
mya) of Central Europe (Mai, 1995).
Radiation of Brassicaceae. —
The Oligocene (ca.
34–23 mya) brought cooler and drier climate forcing the
evergreen Eocene flora of the Northern Hemisphere to
move southwards, and many more deciduous and her-
baceous families appeared (Wolfe, 1969, 1978; Tiffney,
1985). The most important biotic change, in our context,
is the origin and spread of grass-dominated ecosystems
during the Oligocene and Miocene (Webb & Opdyke,
1995; Jacobs & al., 1999; Strömberg, 2005), indicating that
climatic factors most probably promoted the establishment
of new open habitats. Our molecular analysis dates the
radiation of Brassicaceae following the split of Aethione-
meae to early Miocene time (11 mya). We further argue
that the diversification then into the main lineages of the
family was coupled with the increase of open habitats. If
so, this would be a different evolutionary pattern from
what Stömberg (2005) reached for open-habitat grasses of
North America. Based on phytoliths, she suggested that
considerable taxonomic diversification of grasses predates
their ecological expansion into open-habitat in Late Oli-
gocene/Early Miocene.

434
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
Brassicaceae have effective dispersal capacities, as
evidenced by their spread from the Old-World center of
origin into the current worldwide distribution. Long-
distance dispersal events, intercontinental dispersal in-
cluded, have been well documented for several Brassi-
caceae genera (Franzke & al., 1998; Bleeker & al., 2002;
Mummenhoff & al., 2004; Koch & Kiefer, 2006; Koch &
al., 2006; Mummenhoff & Franzke, 2007). We suggest,
that general climatic changes during the Miocene created
newly opened habitats/niches, that acted as an extrinsic
motor for rapid radiation.
Intrinsically, the genomic constitution of early
Brassicaceae undoubtedly was important in creating
the capacity to radiate. Henry & al. (2006) proposed
an evolutionary scenario that explains the shape of the
derived Arabidopsis thaliana genome by three ancient
polyploidization events, each followed by a phase of dip-
loidization. The first two genome doublings occurring
in early eudicot and angiosperm evolution, respectively,
were supposed to have been key events in the origin
and diversification of the angiosperms (De Bodt & al.,
2005). Especially, the duplications of MADS-box genes
involved in flower development may have been important
for the invention and diversification of flowers (
Zahn
& al., 2005). A mechanism driving diversification after
genome duplication had been explained with a loss of al-
te rnative copies of duplicated genes leading to reproduc-
tive isolation (Werth & Windham, 1991; Lynch & Force,
2000). Recently, Scannell & al. (2006) provided strong
evidence indeed for this mode of specification in yeast.
A third genome duplication leading to the Arabidopsis
genome was dated to early times of Brassicaceae evolu-
tion, 24 to 40 mya (Henry & al., 2006). As this time
frame is in the same order of magnitude to that which
we estimated for the Brassicaceae radiation (20 mya), our
data supports the idea of De Bodt & al. (2005) suggest-
ing that this (third) polyploidization was also a key event
for Brassicaceae evolution. There is some evidence that
this genome duplication took place in early Brassicaceae
evolution excluding the tribe Aethionemeae (Schranz &
Mitchell-Olds, 2006). Galloway & al. (1998) conducted
a phylogenetic analysis of 13 taxa of Brassicaceae for the
two polyploidy–derived arginine decarboxylase (Adc)
genes and detected the two Adc paralogs in all members
except Aethionema
.
Blanc & al. (2003) extrapolated the
polyploidy event after the phylogenetic split between
Aethionema and the remaining Brassicaceae. If this
third genome doubling, which occurred after the early
Aethionema split, created and increased the capacity for
(adaptive) radiation, the number of taxa is expected to
be drastically unequal in number in both sister lineages.
Indeed, the tribe Aethionemeae presently consists of ca.
50 species, whereas the rest of Brassicaceae contains ca.
3,650 species.
Brassicaceae genome duplications and current
taxonomic confusion. —
It has been clearly demon-
strated (e.g., Al-Shehbaz, 1984; Mummenhoff & al., 1997,
2005; Appel & Al-Shehbaz, 2003; Koch & al., 2003; Beil-
stein & al., 2006) that the morphological characters in
Brassicaceae are highly homoplasious, thus leading to
highly artificial taxonomic concepts. It is therefore vir-
tually impossible to use morphology alone to establish
a family-wide phylogeny (Al-Shehbaz & al., 2006), and
previous attempts to do so (e.g., Hayek, 1911; Schulz,
1936) have failed considerably. Why this “variation-
on-the-theme–diversity” in Brassicaceae, instead of
a “firework-of-themes–diversity” as we can see in the
angiosperms in general, for both radiations were, as out-
li ned above, supposed to be d riven by genome doublings?
Maere & al. (2005) studied modelling analysis for the
evolution of the different functional categories in genome
duplications. They indicated that gene decay occurring
after the duplication event in Brassicaceae were higher
and less biased toward functional classes compared to
the genome duplication/diploidization events in early an-
giosperm evolution. We hypothesize, that differences in
the quality and quantity of gene decay in Brassicaceae
possibly contributed to the highly homoplastic distribu-
tion of morphological traits across the family. Following
this train of thoughts one even could state that the fuzzy
family-wide morphology then reflects the basal polytomy
within Brassicaceae.
Our data clearly indicated that the unresolved back-
bone was due to rapid radiation of Brassicaceae and thus
might serve as another example for the “Bushes in the
tree of life” sensu Rokas & Carroll (2006), i.e., clades
that are putatively u nresolvable even though conventional
sequence data were (vastly) increased. If so, the lack of
resolution has to be taken into account in future research
that bases on phylogenetic relationship (e.g evo-devo
studies). Nevertheless, the taxon sampling for existing
Brassicaceae datasets should be expanded to assign not
yet sampled taxa to the major lineages within the cruci-
fers. Genomic stu dies of representatives of these lineages
should t hen be analysed to test the hy pothesis if diver sifi-
cation in early Brassicaceae evolution was also driven by
the loss of alternative copies of duplicated genes followed
by reproductive isolation.
ACKNOWLEDGEMENTS
We thank U. Coja and C. Gieshoidt for technical assistance,
A. Drummond, A. Rambaut, S.S. Renner, M. Pirie for help with
molecular dating methods, Thomas Couvreur and two anomy-
mous reviewers for their valuable comments, and Herbaria and
Institutions that supplied plant material. This work was funded
by the German Research Foundation (DFG).

435
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
LITERATURE CITED
Akgün, F., Akay, E. & Erdoúan, B. 2002. Tertiary terrestrial
to shallow marine deposition in central Anatolia: a palyno-
logical approach. Turk. J. Earth Sci. 11: 127–160.
Akgün, F., Kayseri, M.S. & Akkiraz. M.S. 2007. Palaeocli-
matic evolution and vegetational changes during the Late
Oligocene-Miocene period in western and Central Anatolia
(Turkey). Palaeogeogr. Palaeoclimatol. Palaeoecol. 253:
56–106.
Al-Shehbaz, I.A. 1984. The tribes of Cruciferae (Brassicaceae)
in the southeastern United States. J. Arnold Arbor. 65:
343–373.
Al-Shehbaz, I.A., Beilstein, M.A. & Kellogg, E.A. 2006. Sys-
tematics and phylogeny of the Brassicaceae (Cruciferae):
an overview. Pl. Syst. Evol. 259: 89–120.
Al-Shehbaz, I.A., Mutlu, B. & Dönmez, A.A. 2007. The
Brassicaceae (Cruciferae) of Turkey, updated. Turk. J.
Bot. 31: 327–336.
Al-Shehbaz, I.A. & Warwick, S.I. 2007. Two new tribes
(Dontostemoneae and Malcolmieae) in the Brassicaceae.
Harvard Pap. Bot. 12: 429–433.
APG II. 2003. An update of the Angiosperm Phylogeny Group
classification for orders and families of flowering plants:
APG II. Bot. J. Linn. Soc. 141: 399 –436.
Appel, O. & Al-Shehbaz, I.A. 2003. Cruciferae. Pp. 75–174 in:
Kubitzki, K. & Bayer, C. (eds.), The Families and Genera
of Vascular Plants, vol. 5, Flowering Plants. Dicotyledons.
Capparales, Malvales and Non-betalain Caryophyllales.
Springer, Berlin, Heidelberg, New York.
Bailey, C.D. & Doyle, J.J. 1999. Potential phylogenetic util-
ity of the low-copy nuclear gene pistillata in dicotyledon-
ous plants: comparison to nrDNA ITS and trnL intron in
Sphaerocardamum and other Brassicaceae. Molec. Biol.
Evol. 13: 20–30.
Bailey, C.D., Koch, M.A., Mayer, M., Mummenhoff, K.,
O’Kane, S.L., Jr., Warwick S.I., Windham, M.D. & Al-
Shehbaz, I.A. 2006. Towards a global nrDNA ITS phy-
logeny of the Brassicaceae. Molec. Biol. Evol. 23: 2142–
2160.
Bakker, F.T., Breman, F. & Merckx, V. 2006. DNA sequence
evolution in fast evolving mitochondrial DNA nad1 exons
in Geraniaceae and Plantaginaceae. Ta xon 55: 887– 896.
Barkmann, T., Lim, S.-H., Salleh, K.M. & Nais, J. 2004.
Mitochondrial DNA sequences reveal the photosynthetic
relatives of Rafflesia, the world’s largest flower. Proc. Natl.
Acad. Sci. USA. 101: 787–792.
Beilstein, M.A., Al-Shehbaz, I.A. & Kellogg. E.A. 2006.
Brassicaceae phylogeny and trichome evolution. Amer. J.
Bot. 93: 607–619.
Blanc, G., Hokamp, K. & Wolfe, K.H. 2003. A recent poly-
ploidy superimposed on older large-scale duplications in
the Arabidopsis genome. Genome Res. 13: 137–144.
Bleeker, W., Weber-Sparenberg, C. & Hurka, H. 2002.
Chloroplast DNA variation and biogeography in the genus
Rorippa Scop. (Brassicaceae). Pl. Biol. 4: 10 4 –111.
Bowers, J.E., Chapman, B.A., Rong, J.K. & Paterson, A.H.
2003. Unravelling angiosperm genome evolution by phy-
logenetic analysis of chromosomal duplication events.
Nature 422: 433–438.
Bowman, J.L. 2006. Molecules and morphology: comparative
developmental genetics of the Brassicaceae. Pl. Syst. Evol.
259: 199–215.
Chandler, M.E.J. 1960. Plant remains of the Hengistbury and
Barton Beds. Bull. Brit. Mus. (Nat. Hist.), Geol. 4: 191–238.
Charlesworth, D., Liu, F.L. & Zhang, L. 1998. The evolution
of the alcohol dehydrogenase gene family by loss of in-
trons in plants of the genus Leavenworthia (Brassicaceae).
Molec. Biol. Evol. 15: 552–559.
De Bodt, S., Maere, S., Van de Peer, Y. 2005. Genome dupli-
cation and the origin of angiosperms. Trends Ecol. Evol.
20: 591–597.
Drummnod, A.J., Ho, S.Y.W., Phillips, M.J. & Rambaut, A.
2006. Relaxed phylogenetics and dating with confidence.
PLoS Biol. 4(5): e88, doi:10.1371/journal.pbio.0040088.
Drummond, A.J. & Rambaut, A. 2003. BEAST v1.0. Avail-
able from http://evolve.zoo.ox.ac.uk/beast/.
Ermolaeva, M.D., Wu, M., Eisen, J.A. & Salzberg, S.L. 2003.
The age of the Arabidopsis thaliana genome duplication.
Pl. Molec. Biol. 51: 859–866.
Franzke, A., Pollmann, K., Bleeker, W., Kohrt, R. & Hurka,
H. 1998. Molecular systematics of Cardamine and allied
genera (Brassicaceae). ITS and non-coding chloroplast
DNA. Folia Geobot. Phytotax. 33: 225–240.
Freudenstein, J.V., Van den Berg, C., Goldman, D.H., Kores,
P.J., Molvray, M. & Chase, M.W. 2004. An expanded
plastid DNA phylogeny of Orchidaceae and analysis of
jackknife branch support strategy. Amer. J. Bot. 91: 149–
157.
Galloway, G.L., Malmberg, R.L. & Price, R.A. 1998. Phy-
logenetic utility of the nuclear gene arginine decarboxy-
lase: an example from Brassicaceae. Molec. Biol. Evol.
15: 1312–1320.
Gandolfo, M.A., Nixon, K.C. & Crepet, W.L. 1989. A new
foss il f lower fro m the Tu ronia n of New Jers ey Dressiantha
bicarpellata gen. et sp. nov. (Capparales). Amer. J. Bot.
85: 964–974.
Hall, J.C., Sytsma, K.J. & Iltis, H.H. 2002. Phylogeny of Cap-
paraceae and Brassicaceae based on chloroplast sequence
data. Amer. J. Bot. 89: 1826–1842.
Hayek, A. 1911. Entwurf eines Cruciferensystems auf phylo-
genetischer Grundlage. Beih. Bot. Centralbl. 27: 127–335.
Hedge, I.C. 1976. A systematic and geographical survey of
the Old World Cruciferae. Pp. 1–45 in: Vaughan, J.G.,
Macleod, A.J. & Jones, B.M.G. (eds.), The Biology and
Chemistry of the Cruciferae. Academic Press, London.
Henry, Y., Bedhomme, M. & Blanc, G. 2006. History, proto-
history and prehistory of the Arabidopsis thaliana chromo-
some complement. Trends Pl. Sci. 11: 267– 273.
Hu, Y., Zhang, Q., Rao, G. & Sodmergen 2008. Occurrence
of plastids in the sperm cells of Caprifoliaceae: biparental
plastid inheritance in angiosperms is unilaterally derived
from maternal inheritance. Pl. Cell Physiol. 49: 958–968.
Inda, L., Torrecilla, P., Catalán, P. & Ruiz-Zapata, T. 2008.
Phylogeny of Cleome L. and its close relatives Podandro-
gyne Ducke and Polanisia Raf. (Cleomoideae, Cleoma-
ceae) based on analysis of nuclear ITS sequences and
morphology. Pl. Syst. Evol. 274: 111–126.
Itchoda, N., Nishizawa, S., Nagano, H., Kubo, T. & Mikami,
T. 2002. The sugar beet mitochondrial nad4 gene: an intron
loss and its phylogenetic implication in the Caryophyllales.
Theor. Appl. Genet. 104: 209–213.
Jacobs, B.F., Kingston, J.D. & Jacobs, L.L. 1999. The origin

436
TAXON 58 (2) • May 2009: 425–437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
of grass dominated ecosystems. Ann. Missouri Bot . Gard.
86: 590–643.
Kay, K.M., Whittall, J.B. & Hodges, S.A. 2006. A survey of
nrITS substitution rates across angiosperms supports an
approximate molecular clock with life history effects. BMC
Evol. Biol. 6: 36, doi:10.1186/1471-2148-6-36.
Kluge, A.G. & Farris, J.S. 1969. Quantitative phyletics and
the evolution of anurans. Syst. Zool. 18: 1–32.
Knoop, V. 2004. The mitochondrial DNA of land plants: pe-
culiarities in phylogenetic perspective. Curr. Genet. 46:
123–139.
Koch, M., Al-Shehbaz, I.A. & Mummenhoff, K. 2003. Mo-
lecular systematics, evolution, and population biology in
the mustard family (Brassicaceae): a review of a decade of
studies. Ann. Missouri Bot. Gard. 90: 151–171.
Koch, M., Dobeš, C., Kiefer, C., Schmickl, R., Klimeš, L.
& Lysák, M.A. 2007. Supernetwork identifies multiple
events of plastid trnF(GAA) pseudogene evolution in the
Brassicaceae. Molec. Biol. Evol. 24: 63–73.
Koch, M., Haubold, B. & Mitchell-Olds, T. 2000. Compara-
tive evolutionary analysis of chalcone synthase and alco-
hol dehydrogenase loci in Arabidopsis, Arabis, and related
genera (Brassicaceae). Molec. Biol. Evol. 17: 1483–1498.
Koch, M., Haubold, B. & Mitchell-Olds, T. 2001. Molecu-
lar systematics of the Brassicaceae: evidence from coding
plastidic matK and nuclear Chs sequences. Amer. J. Bot.
88: 534–544.
Koch, M.A. & Kiefer, C. 2006. Molecules and migration: bio-
geographical studies in cruciferous plants. Pl. Syst. Evol.
259: 121–142.
Koch, M.A. & Kiefer, M. 2005. Evolution among cruciferous
plants: a lectu re from the comparison of the genetic maps of
three diploid species — Capsella rubella, Arabidopsis lyrata
subsp. petraea, and A. thaliana. Amer. J. Bot. 92: 761–767.
Koch, M., Kiefer, C., Ehrich, D., Vogel, J., Brochmann,
C. & Mummenhoff, K. 2006. Three times out of Asia
Minor — the phylogeography of Arabis alpina L. (Brassi-
caceae). Molec. Ecol. 15: 825–839.
Koch, M.A. & Mummenhoff, K. 2006. Evolution and phy-
logeny of the Brassicaceae. Plant Syst. Evol. 259: 81–83.
Laroche, J., Li, P., Maggia, L. & Bousquet, J. 1997. Molecu lar
evolution of angiosperm mitochondrial introns and exons.
Proc. Natl. Acad. Sci. USA. 94: 5722–5727.
Lynch, M. & Force, A.G. 2000. The origin of interspecies
genomic incompatibility via gene duplication. Amer. Natu-
ralist 156: 590 – 605.
Lysak, M.A., Berr, A., Pecinka, A., Schmidt, R., McBreen,
K. & Schubert, I. 2006. Mechanisms of chromosome
number reduction in Arabidopsis thaliana and related
Brassicaceae species. Proc. Natl. Acad. Sci. U.S.A. 103:
5224 –5229.
Lysak, M.A., Koch, M.A., Pecinka, A. & Schubert, I. 2005.
Chromosome triplication found across the tribe Brassiceae.
Genome Res. 15: 516–525.
Maere, S., De Bodt, S., Raes, J., Casneuf, T., Van Montagu,
M., Kuiper, M. & Van de Peer, Y. 2005. Modeling gene
and genome duplications in eukaryotes. Proc. Natl. Acad.
Sci. U.S.A. 102: 5454–5459.
Mai, D.H. 1995. Tertiä re Vegetatio n sges chich te Eur opas.
Fischer, Jena.
Muller, J. 1981. Fossil pollen records of extant angiosperms.
Bot. Rev. 47: 1–42.
Müller, K. 2004. PRAP — calculation of Bremer support for
large data sets. Molec. Phylog. Evol. 31: 780–782.
Mummenhoff, K., Al-Shehbaz, I.A., Linder, P., Bakker,
F. & Mühlhausen, A. 2005. Phylogeny, morphological
evolution, and speciation of endemic Brassicaceae genera
in the Cape Flora of southern Africa. Ann. Missouri Bot.
Gard. 92: 400–424.
Mummenhoff, K., Bowman, J.L., Linder, P., Friesen, N. &
Franzke, A. 2004. Molecular evidence for bicontinental
hybridogenous genomic constitution in Lepidium sensu
stricto (Brassicaceae) species from Australia and New
Zealand. Amer. J. Bot. 91: 252–259.
Mummenhoff, K. & Franzke, A. 2007. Gone with the bird,
Late Tertiary and Quaternary intercontinental long-dis-
tance dispersal and allopolyploidization in plants. Syst.
Biodiv. 5: 255–260.
Mummenhoff, K., Franzke, A. & Koch, M. 1997. Molecular
data reveal convergence in fruit characters, traditionally
us ed i n the cla ssif icati on of Thlaspi s.l. (Brassicaceae) Bot.
J. Linn. Soc. 125: 183–199.
Nixon, K.C. 1999. The parsimony ratchet, a new method for
rapid parsimony analysis. Cladistics 15: 407–414.
Nylander, J.A.A. 2004. MrModeltest, version 2. Programm
distributed by the author. Evolutionary Biologie Centre,
Uppsala University.
Rambaut, A. & Drummond, A.J. 2003. Tracer, version 1.3.
Available from http://tree.bio.ed.ac.uk/software/tracer/.
Rokas, A. & Carroll, S.B. 2006. Bushes in the Tree of Life.
PLoS Biol. 4: e352, doi:10.1371/jour nal.pbio.0040352.
Rollins, R.C. 1993. The Cruciferae of Continental North Amer-
ica. Stanford University Press, Stanford.
Ronquist, F. & Huelsenbeck, J.P. 2003. MrBayes 3: Bayesian
phylogenetic inference under mixed models. Bioinformat-
ics 19: 1572–1574.
Sánchez-Acebo, L. 2005. A phylogenetic study of the New
Wor ld Cleome (Brassicaceae-Cleomoideae). Ann. Missouri
Bot. Gard. 92: 179–201.
Scannell, D.R., Byrne, K.P., Gordon, J.L., Wong, S. &
Wolfe, K.H. 2006. Multiple rounds of speciation associ-
ated with reciprocal gene loss in polyploid yeasts. Nature
440: 341–345.
Schranz, M.E, Lysak, M.A. & Mitchell-Olds, T. 2006. The
ABC’s of comparative genomics in the Brassicaceae: build-
ing blocks of crucifer genomes. Tre nds Pl. Sci . 11: 535 –54 2.
Schranz, M.E & Mitchell- Olds, T. 2006. Independent ancient
polyploidy events in the sister families Brassicaceae and
Cleomaceae. Pl. Cell 18: 1152–1165.
Schulz, O.E. 1936. Cruciferae. Pp. 227–658 in: Engler, A. &
Harms, H (eds.), Die natürlichen Pflanzenfamilien, vol.
17B. Engelmann, Leipzig.
Sherman-Broyles, S., Boggs, N., Farkas, A., Liu, P., Vreba-
lov, J., Nasrallah, M.E. & Nasrallah, J. 2007. S Locus
genes and the evolution of self-fertility in Arabidopsis
thaliana. Pl. Cell 19: 94–106.
Simillion, C., Vandepoele, K., Van Montagu, M.C.E., Za-
beau, M. & Van de Peer, Y. 2002. The hidden duplication
past of Arabidopsis thaliana. Proc. Natl. Acad. Sci. U.S.A.
99: 13627–13632.
Strömberg, C.A.E. 2005. Decoupled taxonomic radiation and
ecological expansion of open-habitat grasses in the Ceno-
zoic of North America. Proc. Natl. Acad. Sci. U.S.A. 102:
11980–11984.

437
Franzke & al. • Phylogeny and age estimates in Brassicaceae
TAXON 58 (2) • May 2009: 425–437
Swofford, D.L. 2000. PAUP*, version 4.0b10. Sinauer, Sun-
derland.
The Arabidopsis Genome Initiative. 2000. Analysis of the
genome sequence of the f lowering plant Arabidopsis thal-
iana. Nature 408: 796–815.
Tiffney, B.H. 1985. Perspectives on the origin of the f loristic
similarity between eastern Asia and eastern North Amer-
ica. J. Arnold Arbor. 66: 73–94.
Vision, T.J., Brown, D.G. & Tanksley, S.D. 2000. The ori-
gins of genomic duplications in Arabidopsis. Science 290:
2114–2117.
Wahrmund, U., Groth-Malonek, M. & Knoop, V. 2008. Trac-
ing plant mitochondrial DNA evolution: rearrangements of
the ancient mitochondrial gene cluster trnA-trnT-nad7 in
liverwort phylogeny. J. Molec. Evol. 66: 621–629.
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. Canad. J. Bot.
80: 1002–1017.
Warwick, S.I., Al-Shehbaz, I.A. & Sauder, C.A. 2008. Phy-
logenetic relationships in the tribe Alysseae (Brassicaceae)
based on nuclear ribosomal ITS DNA sequences. Canad.
J. Bot. 86: 315–336.
Warwick, S.I., Sauder, C., Al-Shehbaz, I.A. & Jacquemoud,
F. 2007. Phylogenetic relationships in the Brassicaceae
tribes Anchonieae, Choripsoreae, Euclideae, and Hesper-
ideae based on nuclear ribosomal ITS DNA sequences.
Ann. Missouri Bot. Gard. 94: 56–78.
Webb, S.D. & Opdyke, N.D. 1995. Global climatic influence
on Cenozoic land mammal faunas. Pp. 184–208 in: Ken-
nett, J. & Stanley, S. (eds.), Effects of Past Global Change
on Life. National Academic Press, Washington, D.C.
Werth, C.R . & Windham, M.D. 1991. A model for divergent,
allopatric speciation of polyploid pteridophytes resulting
from silencing of duplicate-gene expression. Amer. Natu-
ralist 137: 515–526.
Wikström, N., Savolainen, V. & Chase, M.W. 2001. Evolution
of the angiosper ms: calibrating the family tree. Proc. Roy.
Soc. London, Ser. B, Biol. Sci. 268: 2211–2220.
Wol fe, J. A. 1969. Neogene floristic and vegetational history of
the Pacific Northwest. Madrono 20: 83–110.
Wolfe, J.A . 1978. A paleobotanical interpretation of Terti-
ary climates in the Northern Hemisphere. Amer. Sci. 66:
694–703.
Yang, Y.W., Lai, K.N., Tai, P.Y. & Li, W.H. 1999. Rates of nu-
cleotide substitution in angiosperm mitochondrial DNA se-
quences and dates of divergence between Brassica and the
other angiosperm lineages. J. Molec. Evol. 48: 597–604.
Zahn, L.M., Kong, H., Leebens-Mack, J.H., Kim, S., Soltis,
P.S., Landherr, L.L., Soltis, D.E., dePamphilis, C.W.
& Hong, M. 2005. The evolution of the SEPALLATA
subfamily of MADS-box genes: a preangiosperm origin
wi th mul tiple du plications t hro ughou t angiospe rm h ist ory.
Genetics 169: 2209–2223.
Zunk, K., Mummenhoff, K., Hurka, H. 1999. Phylogenetic
relationships in tribe Lepidieae (Brassicaceae) based on
chloroplast DNA restriction site variation. Canad. J. Bot.
77: 1504–1512.
Appendix. Species, vouchers, and GenBank accession numbers of taxa analysed for nad4 intron 1 in this study.
Taxon, location, voucher specimen, GenBank accession
Aethione ma arabicum ( L.) A ndr z. e x DC ., Tur key, Duran 4950, EU931343. Alliaria petiolata (M. Bieb.) Cavara & G rande, Ger many, OSBU 14252, EU931344.
Anchonium elichrysifolium (DC.) Boiss., Turkey, Mummenhoff 1540, EU931346. Arabis alpina L., Spain, OSBU 6814, EU931347. Alyssum saxatile L.
Netherla nds (cultivated plant), WAG 0044558, EU931345. Berteroa incana (L.) DC. Austria, Mummenhoff 1536, EU931348. Boeche ra drummondii (A. G ray)
A. Löve & D. Löve, U.S.A., OSBU 14958, EU931349. Braya rosea Bunge, Russia, OSBU 13175, EU931350. Cakile maritima L., U.S.A. (cultivated plant),
Wheeler GKI 70, EU931351. Camelina microcarpa And rz. ex DC., Russia, OSBU 13852, EU931352. Chorispora tenella (Pall.) DC., Russia, OSBU 13772,
EU931353. Cochlearia danica L., Germany, Mummenhoff 1697, EU931354. Descurainia stricta (Phil.) Reiche, Chile, Ta ylor & Po ol 1156 3 (M O), EU931355.
Diptychocarpus strictus (Fisch. ex M. Bieb.) Trautv., Iran, TU H 35369, EU931356. Draba altaica (C.A. Mey.) Bunge, Kyrgyzstan, OSBU 16144, EU931357.
Eutrema heterophyllum (W.W. Smith) H. Hara, China, Bartholomew & al. 8490 (MO), EU931358. Halimolobos diffusa (A. Gray) O.E. Schulz, U.S.A.,
OSBU 1141, EU931359. Hedinia alt aica Pobed., Mongolia, OSBU 10413, EU931360. Heliophila linearis (Thunb.) DC., South Africa, Li nder, Hardy & Moline
7596 , EU931361. Heliophila L. sp., South Africa, Burge 1031 (MO), EU931362. Hesperis pendula DC., Jordan , OSBU 13646, EU931363. Hesperis sibirica
L., Russia, OSBU 13609, EU931364. Hornungia petraea (L.) Reichenb., Morocco, OSBU 13661, EU931365. Ianhedgea minutiflora (Hook. f. & Thoms.)
Al-Shehbaz & O’Kane, Tajikistan, Solomon & al. 21646 (MO), EU931366. Iberis amara L., Germany, Mummenhoff 1695, EU931367. Iberis procumbens
Lange, Portug al, OSBU 15071, EU931368. Isatis brevip es (Bunge) Jafry, Kyrgyzst an, OSBU 15810, EU931370. Ionopsidiu m acaule (Desf.) Reichenb., U.S.A.
(cultivated plant), Baum 373 (A), EU931369. Lepidium latifolium L., Kyrgyzstan, OSBU 15614, EU931371. Lyrocarpa coulteri Hook. & Harvey ex Harvey
var. apiculata, Mexico, MO 4805250, EU931372. Malcolmia littorea (L .) R. Br., Port ugal , OSBU 15114, EU931373. Maresia nana (DC.) Batt., Egypt, WAG
0193060, EU931374. Moriera spinosa Boiss., Iran, TUH 33732, EU931375. Myagrum perfoliatum L. Iran, Iranian-American expedition s.n. (UC, TUH),
EU931376. Nevada holmgrenii (Rollins) N. Holmgren, U.S.A., Windham 2186 ( MO), EU931377. Noccaea sp., Ger many, OSBU 16216, EU931379. Noccaea
cochleariformis (DC.) A. Löve & D. Löve, U.S.A., Beilstein 01–21 (MO), EU931378. Oreoloma violaceum Botsch., China, Bartholomew & al. 8596 (MO),
EU931380. Pennellia micrantha (A. Gray) Nieuwl., U.S.A., OSBU 9469, EU931381. Physaria brassicoides Rydb., U.S.A., OSBU 1157, EU931382. Schoe-
nocrambe linearifolia (A. Gray) Rollins, U.S.A., OSBU 9468, EU931383. Selenia dissecta Torr. & A. Gray, U.S.A., OSBU 9452, EU931384. Sisymbrium
loeselii L., Kyrgyzstan, OSBU 15567, EU931385. Smelowskia calycina (Steph.) C.A. Mey., Russia, OSBU 13343, EU931386. Streptanthus sparsiflorus
Rollins, U.S.A., OSBU 9460, EU931387. Stroganowia sagittata Kar. & Kir., Kazakhstan, OSBU 14801, EU931388. Taphr osperm um a ltai cum C.A. Mey.,
China, Bartholome w & al. 8485 (MO), EU931389. Tet racm e qu adr ic orn is (Steph.) Bunge, Kyrgyzstan, OSBU 15902, EU931390. Thelypodium lasiophyllum
(Hook. & Arn.) Greene, U.S.A., OSBU 1090, EU931391. Thlaspi arvense L., Germany, Mummenhoff 1594, EU931392.