Mechanisms of chromosome number reduction in
Arabidopsis thaliana and related Brassicaceae species
Martin A. Lysak*
, Alexandre Berr
, Ales Pecinka
, Renate Schmidt§
, Kim McBreen
, and Ingo Schubert
*Jodrell Laboratory, Royal Botanic Gardens, Kew, Richmond TW9 3AB, United Kingdom; Leibniz-Institute of Plant Genetics and Crop Plant Research (IPK),
Corrensstrasse 3, 06466 Gatersleben, Germany; §Max-Planck-Institut fu¨ r Molekulare Pflanzenphysiologie, 14424 Potsdam, Germany; and Allan Wilson
Centre for Molecular Ecology and Evolution, Massey University, Palmerston North, New Zealand
Edited by Maarten Koornneef, Wageningen University and Research Centre, Wageningen, The Netherlands, and approved February 1, 2006 (received for
review December 14, 2005)
Evolution of chromosome complements can be resolved by ge-
nome sequencing, comparative genetic mapping, and comparative
chromosome painting. Previously, comparison of genetic maps and
gene-based phylogenies suggested that the karyotypes of Arabi-
dopsis thaliana (n 5) and of related species with six or seven
chromosome pairs were derived from an ancestral karyotype with
eight chromosome pairs. To test this hypothesis, we applied
multicolor chromosome painting using contiguous bacterial artifi-
cial chromosome pools of A. thaliana arranged according to the
genetic maps of Arabidopsis lyrata and Capsella rubella (both n
8) to A. thaliana, A. lyrata, Neslia paniculata, Turritis glabra, and
Hornungia alpina. This approach allowed us to map the A. lyrata
centromeres as a prerequisite to defining a putative ancestral
karyotype (n 8) and to elucidate the evolutionary mechanisms
that shaped the karyotype of A. thaliana and its relatives. We
conclude that chromosome ``fusions'' in A. thaliana resulted from
(i) generation of acrocentric chromosomes by pericentric inver-
sions, (ii) reciprocal translocation between two chromosomes (one
or both acrocentric), and (iii) elimination of a minichromosome that
arose in addition to the ``fusion chromosome.'' Comparative chro-
mosome painting applied to N. paniculata (n 7), T. glabra (n 6),
and H. alpina (n 6), for which genetic maps are not available,
revealed chromosomal colinearity between all species tested and
allowed us to reconstruct the evolution of their chromosomes from
a putative ancestral karyotype (n 8). Although involving differ-
ent ancestral chromosomes, chromosome number reduction fol-
lowed similar routes as found within the genus Arabidopsis.
chromosome painting genome homeology karyotype evolution
genome colinearity phylogeny
The specific chromosome complement of a eukaryotic organ-
ism, the karyotype, may vary as to the number, size, and
shape of chromosomes even between closely related taxa. Size
and or shape of chromosomes may change by means of recip-
rocal translocations, inversions, insertions, or deletions. At the
diploid level, chromosome numbers may vary because of meiotic
missegregation in translocation heterozygotes or by so-called
fusion fission events (1, 2). Such mechanisms combined with
doubling of entire complements (polyploidy) have generated the
extant karyotypes in the course of evolution. Because karyotypes
are dynamic structures, the reconstruction of ancestral karyo-
types on the basis of well described extant karyotypes is neces-
sary to understand the evolutionary direction of karyotypic
variation. In principle, there are two strategies to uncover
putative ancestral karyotypes and to reconstruct the evolution-
ary events that led to the present karyotypes, both based on
comparison of chromosome colinearity shared by extant species.
One strategy is comparative genetic mapping, as exemplified for
grasses (3); the other is comparative chromosome painting
(CCP), successfully applied to vertebrates (for review, see refs.
4 and 5).
Among species of the Brassicaceae family, chromosome num-
bers vary from n 4 to 128 (6). Within the genus Arabidopsis and
its close relatives, basic chromosome numbers vary between 5, 6,
7, and 8, with most species having 8 chromosome pairs. Based on
comparative genetic maps between Arabidopsis thaliana (n 5)
and Arabidopsis lyrata (n 8; refs. 7 and 8) as well as between
A. thaliana and Capsella rubella (n 8; ref. 9) and on DNA
sequence-derived phylogenetic trees (10), an ancestral karyo-
type with 8 chromosomes has been inferred. It has been sug-
gested that the karyotype of A. thaliana descended from the
ancestral karyotype by two reciprocal translocations, three chro-
mosome fusions, and at least three inversions (8, 11).
Here, we studied chromosomal colinearity between A. thali-
ana and four cruciferous species with different chromosome
numbers (A. lyrata, n 8; Neslia paniculata, n 7; Turritis glabra,
n 6; and Hornungia alpina, n 6) by CCP with bacterial
artificial chromosome (BAC) contigs covering all chromosome
arms of A. thaliana. The CCP results were compared with the
genetic maps in the context of phylogenetic relationships based
on multiple gene sequences. For A. lyrata, chromosome painting
confirmed the pattern of genome colinearity derived from the A.
thaliana A. lyrata genetic maps (7, 8). Moreover, we present new
insights as to the centromere positions within the ancestral
karyotype and the role of specific inversion and translocation
events during chromosome number reduction from n 8 to n
5 in A. thaliana. Applying CCP to the three species with six or
seven chromosome pairs for which genetic maps are not avail-
able, we uncovered the extent of chromosomal colinearity
between these species and A. thaliana. We identified events
associated with the chromosome number reduction as compared
with the ancestral karyotype.
Results
Chromosome Painting Confirmed and Specified the Evolutionary
Events Toward the A. thaliana Karyotype. To address the chromo-
somes that correspond to the linkage groups (LGs) of A. lyrata
(7, 8), multicolor CCP was applied to its pachytene chromo-
somes. A. thaliana BAC contigs arranged according to the eight
LGs (Fig. 1B) labeled one bivalent each. The position of all
centromeres and of the five nucleolus organizing regions
(NORs) (on the short arms of chromosomes 1, 3­5, and 7) has
been determined for the A. lyrata karyotype (Fig. 1 B and D­F).
Because the LGs of A. lyrata largely correspond to those of the
more distantly related C. rubella (11), the karyotypes of both
species should be similar to the ancestral karyotype (AK) of the
entire Arabidopsis lineage.
Conflict of interest statement: No conflicts declared.
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: BAC, bacterial artificial chromosome; CCP, comparative chromosome paint-
ing; LG, linkage group; AT, chromosome of Arabidopsis thaliana; AK, ancestral karyotype;
NOR, nucleolus organizing region.
Data deposition: The sequences reported in this paper have been deposited in the
GenBank database (accession nos. AF167571, M64119, AY198405, X52320, AP000423, and
DQ310510­DQ310545).
To whom correspondence should be addressed at: Jodrell Laboratory, Royal Botanic
Gardens, Kew, Richmond, Surrey TW9 3AB, United Kingdom. E-mail: m.lysak@kew.org.
 2006 by The National Academy of Sciences of the USA
5224­5229 PNAS March 28, 2006 vol. 103 no. 13 www.pnas.org cgi doi 10.1073 pnas.0510791103
With the aim of visualizing the components of the ancestral
karyotype with eight chromosome pairs (AK1 to -8) on the five
A. thaliana chromosomes, multicolor painting was performed
with BAC pools arranged according to entire or partial LGs of
A. lyrata (Figs. 1 C­F and 2A). Probes painting A. thaliana
chromosomes 1 (AT1) hybridized also to the two A. lyrata
chromosomes that correspond to LGs 1 and 2 matching with
comparative mapping results. The BAC contig covering the
distal part of the bottom arm of AT1 and homeologous to LG2
labeled the entire AK2 chromosome of A. lyrata (Fig. 1D).
Subdividing the probe for LG2 in three differently labeled
subcontigs revealed an inversion comprising the top (short) arm
of AK2 (Fig. 2A). This inversion has been detected by genetic
mapping in C. rubella (9) and very recently also in A. lyrata (12).
Because neither centromeric nor telomeric sequences of chro-
mosome AK2 were detected at the point of ``fusion'' within AT1
(13), we conclude that a pericentric inversion transformed AK2
into an acrocentric chromosome. This inversion event was
followed by a reciprocal translocation between the short arm end
of AK1 and the centric end of AK2, resulting in AT1 (Fig. 2B).
The second very small translocation product, consisting mainly
of telomeric repeats from AK1 and AK2 and of the AK2
centromere, became lost most likely because it failed to pair
properly during meiosis and lacked essential genes.
Probes designed according to A. lyrata LGs 3, 4, and 5 labeled
chromosomes AT2 and AT3 (Fig. 1E). Thus, AT2 comprises the
terminal part of the top arm of AK5, followed by the proximal
part of the bottom arm, the centromere, and the top (short) arm
of AK3 and by the entire AK4 (without a centromere). AT3 is
homeologous to the bottom arm, the centromere, and part of the
top arm of AK5, and a large part of the bottom arm of AK3 (Fig.
1C). Our data show that the centromeres of AT2 and AT3
retained the same position as in AK3 and AK5 and confirm that
chromosomes AT2 and AT3 have been formed by a reciprocal
translocation between AK3 and AK5, as indicated previously (7,
8, 11). Within AK4, no pericentric inversion involving the top
arm was found by FISH with pools subdividing this region.
However, a paracentric inversion constituting the entire short
arm, followed by a pericentric one, could have generated an
acrocentric AK4 that fused with AK3 5 to AT2 by a reciprocal
translocation without interrupting colinearity to LG4. The small
second translocation product consisting of telomeres and the
centromere of AK4 was lost (Fig. 2C).
Probes designed according to A. lyrata LGs 6, 7, and 8 labeled
chromosomes AT4 and AT5 (Fig. 1F). AT4 is homeologous to
the top (short) arm of AK6 and to the bottom arm of AK7.
Chromosome AT5 comprises the bottom (long) arm of AK6, the
top (short) arm of AK7, and the entire AK8 (without a centro-
mere). The first step toward the recent chromosomes AT4 and
AT5 was a reciprocal translocation apparently between the
centromeres of AK6 and AK7 (Fig. 2D). The inversion in the
long arm of AT4 was originally detected by genetic mapping (8,
9) and could be confirmed by CCP. This apparently pericentric
inversion (14) occurred simultaneously with or after the trans-
location between AK6 and AK7. Another inversion within the
bottom arm of AT5 in comparison with C. rubella linkage group
H (8, 9, 11) originally generated an acrocentric chromosome
from AK8. Finally, a reciprocal translocation ``fused'' this acro-
centric with AK6 7 to AT5 (Fig. 2D). Thus, this fusion followed
the same scheme as the fusion that contributed to the formation
of AT1 (Fig. 2B).
CCP Reveals the Modes of Chromosome Number Reduction also for
Species Lacking Genetic Maps. To elucidate the karyotype evolu-
tion from the putative ancestral karyotype toward the karyo-
types of N. paniculata, T. glabra, and H. alpina (all n 8), BAC
pools arranged according to the eight LGs (or particular regions
thereof) of A. lyrata C. rubella were used for CCP.
Fig. 3A1 shows the reconstructed karyotype of N. paniculata
(n 7). Probes designed according to the reference LGs 1­3 and
6­8 each labeled an individual pachytene bivalent or mitotic
chromosome pair (Fig. 3 A3­A5). CCP with the probe corre-
Fig. 1. Evolution of the A. thaliana (n 5) karyotype based on comparative chromosome painting in A. lyrata (n 8). (A) The events leading to chromosome
number reduction in the course of evolution from an ancestor with n 8 chromosomes toward the karyotype of A. thaliana. Inversion (I), translocation (T), and
fusion events (F) as well as divergence time estimates are from Koch and Kiefer (11). Only consecutive paracentric and pericentric inversions (Ipa Ipe) within AK4
are newly proposed. Events that require a distinct chronological order are arranged together. (B) Idiogram of the A. lyrata karyotype. The chromosomes
correspond to the ancestral karyotype, with the NORs (black spheres) actually found in A. lyrata. (C) Idiogram of the A. thaliana karyotype indicating the
composition of chromosomes AT1 to -5 derived from the ancestral karyotype (B) and the events involved in chromosome number reduction shown in A. (D­F)
Pachytene chromosomes of A. thaliana (AT1 to -5) and A. lyrata (AK1 to -8) painted by A. thaliana BAC contigs arranged according to the linkage groups of A.
lyrata and pseudocolored according to B; empty squares arrowheads indicate centromeric regions. (Scale bar: 5 m in D­F.)
Lysak et al. PNAS March 28, 2006 vol. 103 no. 13 5225
PLANTBIOLOGY
sponding to LG 8 revealed an acrocentric chromosome with a
short arm apparently consisting of 45S rDNA (Fig. 3A6). The
probes corresponding to LGs 4 and 5 each labeled one arm of the
same chromosome (Fig. 3A4). FISH with differently labeled
BAC pools of LGs 4 and 5 revealed a presumably pericentric
inversion within AK4, rendering it acrocentric before a recip-
rocal translocation that fused both LGs and generated a dis-
pensable minichromosome consisting mainly of a centromere
and telomeric sequences (Fig. 3A2). Because we could not detect
an inversion within AK5, a paracentric inversion comprising its
short arm, before a pericentric inversion, could have generated
an acrocentric AK5 (without interrupting colinearity to LG5)
that fused with modified AK4 (Fig. 3A2), resembling the fusion
of AK3 and AK4 to AT2 (Fig. 2C).
In T. glabra (n 6), probes designed according to the LGs 1,
4, 6, and 7 each labeled an individual pachytene bivalent (Figs.
3B1 and 4). The probes for LGs 2 and 8 painted a single
chromosome (AK2 8). This chromosome with a centromere
within the part derived from AK2 apparently resulted from a
pericentric inversion rendering AK8 acrocentric, followed by a
reciprocal translocation between AK2 and the acrocentric AK8
(data not shown). The small dispensable translocation product
became lost. A second fusion chromosome corresponds to AK3
and AK5. It originated (after a pericentric inversion comprising
the short arm of AK3) by means of a reciprocal translocation
between the AK3 centromere and the short arm end of AK5. As
revealed by differential painting of subregions from the short
arms of both ancestral chromosomes, a paracentric inversion
occurred within the fusion chromosome AK3 5 (Fig. 3B2).
Furthermore, differential painting revealed a paracentric inver-
sion in the bottom (long) arm of the chromosome corresponding
to AK7 (Fig. 4).
In H. alpina (n 6), probes designed according to the LGs 1,
3, 4, and 7 each labeled one individual pachytene bivalent.
Apparently, AK8 became acrocentric by means of a pericentric
inversion before a reciprocal translocation led to fusion with
AK6. A pericentric and subsequently a paracentric inversion
happened later on within the long arm of the fusion chromosome
involving regions derived from both ancestral chromosomes
(Fig. 3C2). AK2 and AK5 are combined by translocation of one
arm of AK5 to one end and of the second arm of AK5 to the
other end of AK2 (Fig. 3C1 and Fig. 5, which is published as
supporting information on the PNAS web site).
The Phylogenetic Tree Indicates Recurrent Chromosome Number Re-
duction from an Ancestral Karyotype with Eight Chromosome Pairs.
Phylogenetic analysis of four independent gene loci supported a
largely congruent phylogeny. To visualize the extent of this
compatibility between gene trees, the four gene trees were
combined into a single supernetwork by using the z-closure rule
(ref. 15, Fig. 4, and Fig. 6, which is published as supporting
information on the PNAS web site). The inferred phylogenetic
relationships resemble those published previously (10, 16), with
H. alpina being most distantly related to A. thaliana. The similar
karyotypes of Arabidopsis and Capsella lineages support the view
that the A. lyrata Capsella karyotypes resemble the hypothetical
ancestral karyotype with eight chromosome pairs. Our data
imply that the putative ancestral karyotype (n 8) is older than
the split between the Hornungia and the Arabidopsis-Capsella-
Turritis clades. Furthermore, placing reconstructed karyotypes
onto the phylogenetic tree suggests that chromosome number
reduction was an independent and recurrent process in each of
the four lineages analyzed, mostly including different ancestral
chromosomes (Fig. 4).
Discussion
Multicolor chromosome painting of pachytene chromosomes of
A. thaliana with probes arranged according to the linkage groups
of A. lyrata and C. rubella (considered to represent largely the
ancestral karyotype of A. thaliana and its relatives) have allowed
us to detect all inversion, translocation, and fusion events that
apparently contributed to the evolution of the A. thaliana
karyotype, as inferred from comparative genetic maps (7­9, 11).
More importantly, chromosome painting using a contiguous
BAC tiling path for A. thaliana allowed us to localize the
centromeres and border chromosomal inversions and transloca-
tions more precisely than what was deducible from comparative
maps with a low or medium marker density. For the following
events, the chronological order of chromosome rearrangements
Fig. 2. Reconstruction of A. thaliana fusion chromosomes. (A) Painting of
AK2 chromosome of A. lyrata and AT1 chromosome of A. thaliana with
differently labeled BAC pools covering LG 2 of A. lyrata shows that a pericen-
tric inversion has occurred before AK2 became fused with AK1 during the
evolution toward A. thaliana. Empty squares arrowheads indicate centro-
meric regions. (Scale bar: 5 m.) (B) The origin of fusion chromosome AT1 by
a pericentromeric inversion within AK2 followed by a reciprocal translocation
with breakpoints at the short arm end of AK1 and close to the centromere of
(acrocentric) AK2. (C and D) The presumed origin of A. thaliana chromosomes
AT2 to -5. (C) Chromosome AT3 has originated because of a reciprocal trans-
location between AK3 and AK5. The fusion chromosome AT2 arose by means
of a translocation between AK3 5 and an acrocentric AK4. A paracentric (Ipa)
and a pericentric (Ipe) inversion within AK4 are presumed to explain the origin
of AT2. (D) AT4 was formed by means of a reciprocal translocation between
AK6 and AK7 and subsequent pericentric inversion; additionally, it gained an
NOR. Shown is pericentric inversion of AK8 into an acrocentric and subsequent
translocation with the short arm of AK6 7 generated AT5. The minichromo-
somes arising as a second translocation product in B­D are dispensable and got
lost. The ancestral chromosomes are presumed to bear NORs (empty spheres)
at the same positions as in A. lyrata; NORs actually found in A. thaliana are
depicted as black spheres. Inversions (I), translocations (T), and fusions (F) are
enumerated as in Fig. 1.
5226 www.pnas.org cgi doi 10.1073 pnas.0510791103 Lysak et al.
leading to chromosome number reduction in A. thaliana is fixed
(Figs. 1 A and C and 2 and ref. 11): inversion 1 3 fusion 3,
resulting in AT1; translocation 1 3 inversion 2 (and 3) 3 fusion
1 resulting in AT4 and AT5; and translocation 2 3 para- and
pericentric inversions in AK4 3 fusion 2, resulting in AT2 and
AT3. However, it remains uncertain which group of linked
events occurred earlier or later. It is also not clear whether AK8
was originally an acrocentric as in N. paniculata where the short
arm apparently consists only of 45S rDNA (Fig. 3A6) or a
submetacentric as in A. lyrata and C. rubella, before participating
in the fusions toward the karyotypes of A. thaliana, T. glabra, and
H. alpina. Both assumptions imply repeated pericentric inver-
sions within AK8.
Furthermore, our study suggests that fusion events reducing
the basic chromosome number from eight to five are based on
reciprocal translocations between meta- submetacentric and
acrocentric chromosomes, the latter generated by pericentric
inversions. The minichromosomes that, in addition to the ``fusion
chromosomes,'' resulted from these translocation events have
presumably been lost. This scenario explains the elimination of
centromeres toward the A. thaliana karyotype as well as a
potential loss of terminal NORs.
An important finding in our study was that CCP reliably
detected chromosomal colinearity between A. thaliana and more
distantly related species for which (comparative) genetic maps
are not available (Figs. 3 and 4). CCP has enabled us to elucidate
chromosome rearrangements and reductions that have led to
present day karyotypes. Our investigation of three such species
revealed mainly the same mechanisms to be responsible for
chromosome number reduction as found for A. thaliana. Re-
markably, of the 52 breakpoints contributing to the 26 rear-
rangements considered in our work, 84.6% involved centromeric
(24 breakpoints) or terminal (20 breakpoints) positions where
repetitive sequences are clustered. Concordant with our obser-
Fig. 3. Idiograms and painted chromosomes of N. paniculata (n 7), T. glabra (n 6), and H. alpina (n 6). Chromosomes are colored according to their
colinearity to the ancestral karyotype (AK1 to -8; see Fig. 1) in A1, A3­A5, and B1 and C1, or according to the differently labeled BAC subcontigs used in CCP
(A2­C2). (A3­A5) CCP in N. paniculata performed with A. thaliana probes arranged according to the AK1 to -8 chromosomes of A. lyrata. Pairs of painted mitotic
chromosomes are inserted in A3­A5. The presumed origin of the fusion chromosome AK4 5 and painting of this chromosome are shown in A2. (A6) The probe
corresponding to chromosome AK8 labels the long arm of a chromosome that on the other side of the heterochromatic pericentromere region displays only 45S
rDNA (green) and thus represents a nearly acrocentric homeolog of AK8. (B1 and C1) Idiograms of reconstructed karyotypes of T. glabra (B1) and H. alpina (C1).
(B2 and C2) The presumed origin and corresponding painting of the T. glabra fusion chromosome AK3 5 (B2) and the H. alpina fusion chromosome AK6 8 (C2).
Empty squares arrowheads indicate centromeric regions, black spheres represent actual NORs, and empty spheres hypothetical NORs (see Fig. 2). (Scale
bars: 5 m.)
Lysak et al. PNAS March 28, 2006 vol. 103 no. 13 5227
PLANTBIOLOGY
vations, in telomerase-deficient A. thaliana plants terminal
NORs participate in chromosome rearrangements 10 times
more often than expected at random (17).
In general, chromosome rearrangements exhibit a low level of
homoplasy because the same chromosomal patterns are unlikely
to occur independently in different phylogenetic lineages. When
shared by different species, such rearrangements represent ``rare
genomic changes'' (18) indicative of phylogenetic relatedness
and common ancestry. Ancestral chromosome patterns shared
by H. alpina and the remaining species support the assumption
of a common ancestor for the whole species group. Thus,
chromosomal colinearity has an immense potential to infer
phylogenetic relationship and or to refine sequence-based phy-
logenetic frameworks. Because CCP was successfully adapted
also to species of the tribe Brassiceae (19), it should be applicable
to clarify chromosome evolution even for taxa of the Brassi-
caceae family remotely related to A. thaliana.
Materials and Methods
Chromosome Preparation. Chromosome spreads were prepared
from: A. thaliana (accessions Columbia and C24), A. lyrata subsp.
lyrata (Bash Bish, MA; cf. 14), H. (Pritzelago) alpina (Austrian
Alps Mountains, Alpengarten Rannach, Austria), N. paniculata
(Botanic Garden, Copenhagen, Denmark; code no. 28, index no.
783, 2001) and T. (Arabis) glabra (Stangerode, Harz Mountains,
Germany). See Table 1, which is published as supporting infor-
mation on the PNAS web site, for detailed information on
species used in this study. Whole inflorescences were harvested
from plants grown in a greenhouse, fixed in 6:3:1 ethanol:chlo-
roform:acetic acid overnight and stored in 70% ethanol at 20°C
until use. Selected inflorescences were rinsed in distilled water
and citric buffer (10 mM sodium citrate, pH 4.8). Flower buds of
suitable size were excised and digested in 0.3% pectolyase,
cellulase, and cytohelicase (all Sigma) in citric buffer at 37°C for
3­5 h, transferred into citric buffer, and kept at 4°C until use.
Individual flower buds were put on a clean slide under a
stereomicroscope and disintegrated by a needle in a drop of citric
buffer. Subsequently, 10 l of 45% acetic acid was added to the
suspension, and the drop was stirred by a needle on a hot plate
(50°C) for 0.5­2 min. During the stirring, 10­40 l of 45% acetic
acid was added. Subsequently, chromosomes were fixed by
adding 100­200 l of ethanol:acetic acid (3:1). The slide was
tilted to remove the fixative, dried by a hair dryer, and inspected
under phase contrast. Suitable slides were postfixed in 4%
formaldehyde dissolved in distilled water for at least 10 min and
air-dried.
Probe Labeling, FISH, and Image Processing. The BAC clones used
for FISH were obtained from the Arabidopsis Biological Re-
source Center (Columbus, OH). Isolation of BAC DNA and
selection of chromosome-specific clones suitable for chromo-
some painting was performed as described (14, 20). Table 2,
which is published as supporting information on the PNAS web
site, shows BAC contigs used for CCP in Figs. 1­3; the detailed
list of BACs used for chromosome painting is available upon
request. The BAC clone T15P10 (AF167571) bearing 45S rRNA
genes was used for in situ localization of NORs. For comparative
painting, pools of BAC contigs were arranged according the
comparative genetic maps between A. thaliana C. rubella (9), A.
thaliana A. lyrata subsp. petraea (7), or A. thaliana A. lyrata
subsp. lyrata (8). The designation of LGs and chromosomes of A.
lyrata, C. rubella, and the putative AK was as in Kuittinen et al.
(7) and Koch and Kiefer (11). Thus, chromosomes AK1 to AK8
correspond to LG 1 to 8 of A. lyrata (7) and A to H of C. rubella
(9), respectively. BAC DNA was labeled by nick translation (21)
with biotin-dUTP, digoxigenin-dUTP, dinitrophenyl (DNP)-
dUTP, or with Cy3-dUTP (20). For chromosome painting,
labeled BACs were pooled, precipitated, and resuspended in
20­40 l of hybridization mix [20% or 50% (A. thaliana)
formamide, 10% dextran sulfate, 2 SSC, 50 mM sodium
phosphate (pH 7.0)] per slide. Labeled probe and chromosomes
were denatured together on a heat block at 80°C for 2 min and
incubated in a moist chamber at 37°C for 36­60 h. Posthybrid-
ization washing was performed in 50% formamide (A. thaliana)
or 20% formamide (all other species) in 2 SSC at 42°C.
Fluorescent detection was as follows: biotin-dUTP was detected
by avidin conjugated with Texas Red (Vector Laboratories),
goat anti-avidin conjugated with biotin (Vector Laboratories),
and again with avidin conjugated with Texas Red; digoxigenin-
dUTP by mouse anti-digoxigenin (Roche Applied Science,
Mannheim, Germany) and goat anti-mouse conjugated with
Alexa Fluor 488 (Molecular Probes); and DNP-dUTP by rabbit
anti-DNP (Sigma) and goat anti-rabbit conjugated with Cy5
(Jackson ImmunoResearch). Cy3-dUTP was observed directly.
Chromosomes were counterstained with 2 g ml DAPI in
Vectashield (Vector Laboratories). Fluorescence signals were
analyzed with an Axioplan 2 epifluorescence microscope (Zeiss)
by using 100 1.4 Zeiss plan apochromat objective and cooled
charge-coupled device (CCD) camera (Spot 2e, Diagnostic
Instruments, Sterling Heights, MI). Images were acquired sep-
arately for all five fluorochromes by using the appropriate
excitation and emission filters (AHF Analysentechnik, Lagen-
feld, Germany). The five monochromatic images were pseudo-
colored and merged by using ADOBE PHOTOSHOP 6.0 software
(Adobe Systems, San Jose, CA). Eight pseudocolors were ap-
plied to depict the eight linkage groups chromosomes of an
ancestral karyotype.
DNA-Based Phylogenetic Reconstruction. Nucleotide sequences of
five genes were obtained for each of the taxa genetically mapped
and or analyzed by CCP, as well as for the taxa Boechera stricta,
Camelina microcarpa, Crucihimalaya wallichii, Lepidium apeta-
lum, Olimarabidopsis pumila, Pachycladon novae-zelandiae and
Fig. 4. Idiograms and Supernetwork for the five investigated species and for
C. rubella. Length of edges in the network is related to evolutionary distance
(29). Note that Capsella bursa-pastoris gene sequences were used to infer the
phylogenetic position of C. rubella. A. lyrata and C. rubella chromosomes are
assumed to resemble the ancestral karyotype (AK1 to -8). Colinear chromo-
somes chromosome regions are depicted in the same color. Red and black
brackets indicate paracentric and pericentric inversions, respectively.
5228 www.pnas.org cgi doi 10.1073 pnas.0510791103 Lysak et al.
Physaria gracilis. Phylogenetic analysis of these genes was un-
dertaken to help evaluate an unrooted species phylogeny for
these species. A. thaliana sequences were obtained from Gen-
Bank (accession nos. M64119, AY198405, X52320, and
AP000423). All sequences unique to this study have been
deposited in GenBank (accession nos. DQ310510-DQ310545).
Genomic DNA was isolated from leaves by using the plant
Dneasy minikit (Qiagen, Valencia, CA). Five genes were PCR
amplified and sequenced: (i) the single copy gene encoding
cytosolic glyceraldehyde-3-phosphate dehydrogenase (G3Pdh)
(22); (ii) the single copy gene encoding FRIGIDA (FRI) (23); (iii)
the multicopy internal transcribed spacer (ITS), including ITS1,
5.8S, and ITS2 of the nuclear ribosomal DNA sequence units
(24); and (iv and v) the two chloroplast loci, ribulose-1,5-
bisphosphate carboxylase oxygenase large subunit gene (rbcL,
25) and trnL­trnF spacer (26). Because both rbcL and the
trnL-trnF regions potentially evolve as a unit, we concatenated
these as a single CpDNA marker.
Sequence contigs were prepared by using SEQUENCHER 4.2
(Gene Codes, Ann Arbor, MI), and visually aligned with MAC-
CLADE 4.06 (27) by using the A. thaliana gene structure and
sequence as a reference. Gene trees were estimated by maximum
likelihood (ML) by using PAUP 4.0 [options: exhaustive search;
Multrees (save multiple trees), and TBR (tree bisection and
reconnection) branch swapping]. Substitution models were se-
lected by MODELTEST 3.5 (28) under the Akaike information
criterion, with parameters estimated during ML exhaustive
searches. It was necessary to constrain Hornungia, Lepidium, and
Physaria to a monophyletic group in the ITS tree. These four
trees have been combined by using the Z-closure super network
procedure (15), as implemented in SPLITSTREE 4.1 (29).
We thank J. Bruder, M. Ku¨hne, and R. Schubert for technical assistance
and P. Lockhart and R. Rieger for critical reading of the manuscript. This
work was supported in part by a grant from the German Research
Foundation (DFG) (to M.A.L. and I.S.). The Allan Wilson Centre is
acknowledged for supporting K.M. during this study.
1. Schubert, I., Rieger, R. & Ku¨nzel, G. (1991) in Chromosome Engineering in
Plants: Genetics, Breeding, Evolution (Part A), eds. Gupta, P. K. & Tsuchiya, T.
(Elsevier, Amsterdam), pp. 113­140.
2. Schubert, I. & Rieger, R. (1994) in Chromosomal Alterations: Origin and
Significance, eds. Obe, G. & Natarajan, A. T. (Springer, Berlin), pp. 380­394.
3. Moore, G., Devos, K. M., Wang, Z. & Gale, M. D. (1995) Curr. Biol. 5, 737­739.
4. Wienberg, J. (2004) Curr. Opin. Genet. Dev. 14, 657­666.
5. Murphy, W. J., Pevzner, P. A. & O'Brien, S. J. (2004) Trends Genet. 20,
631­639.
6. Appel, O. & Al-Shehbaz, I. A. (2003) in The Families and Genera of Vascular
Plants, eds. Kubitzki, K. & Bayer, C. (Springer, Berlin), Vol. 5, pp. 75­174.
7. Kuittinen, H., de Haan, A. A., Vogl, C., Oikarinen, S., Leppa¨la¨, J., Koch, M.,
Mitchel-Olds, T., Langley, C. H. & Savolainen, O. (2004) Genetics 168,
1575­1584.
8. Yogeeswaran, K., Frary, A., York, T. L., Amenta, A., Lesser, A. H., Nasrallah,
J. B., Tanksley, S. D. & Nasrallah, M. E. (2005) Genome Res. 15, 505­515.
9. Boivin, K., Acarkan, A., Mbulu, R. S., Clarenz, O. & Schmidt, R. (2004) Plant
Physiol. 135, 735­744.
10. Koch, M., Bishop, J. & Mitchell-Olds, T. (1999) Plant Biol. 1, 529­537.
11. Koch, M. A. & Kiefer, M. (2005) Am. J. Bot. 92, 761­767.
12. Hansson, B., Kawabe, A., Preuss, S., Kuittinen, H. & Charlesworth, D. (2006)
Genet. Res., in press.
13. The Arabidopsis Genome Initiative (2000) Nature 408, 796­815.
14. Lysak, M. A., Pecinka, A. & Schubert, I. (2003) Chromosome Res. 11, 195­204.
15. Huson, D. H., Dezulian, T., Kloepper, T. & Steel, M. A. (2004) Lect. Notes
Comput. Sci. 3240, 388­399.
16. Koch, M., Haubold, B. & Mitchell-Olds, T. (2001) Am. J. Bot. 88, 534­544.
17. Siroky, J., Zluvova, J., Riha, K., Shippen, D. E. & Vyskot, B. (2003) Chromo-
soma 112, 116­123.
18. Rokas, A. & Holland, P. W. H. (2000) Trends Ecol. Evol. 15, 454­459.
19. Lysak, M. A., Koch, M. A., Pecinka, A. & Schubert, I. (2005) Genome Res. 15,
516­525.
20. Pecinka, A., Schubert, V., Meister, A., Kreth, G., Klatte, M., Lysak, M. A.,
Fuchs, J. & Schubert, I. (2004) Chromosoma 113, 258­269.
21. Ward, P. B. (2002) in FISH: A Practical Approach, Practical Approach Series,
eds. Beatty, B., Mai, S. & Squire, J. (Oxford Univ. Press, Oxford), Vol. 260, pp.
5­28.
22. Strand, A. E., Leebens-Mack, J. & Milligan, B. G. (1997) Mol. Ecol. 6, 113­118.
23. Le Corre, V., Roux, F. & Reboud, X. (2002) Mol. Biol. Evol. 19, 1261­1271.
24. White, T. J., Bruns, T., Lee, S.& Taylor, J. (1990) in PCR Protocols: A Guide
to Methods and Applications, eds. Innis, M. A., Gelfand, D. H., Snisky, J. J. &
White, T. J. (Academic, San Diego), pp. 315­322.
25. Chase, M. W., Soltis, D. E., Olmstead, R. G., Morgan, D., Les, D. H., Mishler,
B. D., Duvall, M. R., Price, R. A., Hills, H. G., Qiu, Y.-L., et al. (1993) Ann.
Missouri Bot. Garden 80, 528­580.
26. Taberlet, P., Gielly, L., Pautou, G. & Bouvet, J. (1991) Plant Mol. Biol. 17,
1105­1109.
27. Maddison, D. R. & Maddison, W. P. (2000) MACCLADE 4: Analysis of Phylogeny
and Character Evolution (Sinauer, Sunderland, MA).
28. Posada, D. & Crandall, K. A. (1998) Bioinformatics 14, 817­818.
29. Huson, D. H. & Bryant, D. (2006) Mol. Biol. Evol. 23, 254­267.
Lysak et al. PNAS March 28, 2006 vol. 103 no. 13 5229
PLANTBIOLOGY