Journal of Experimental Botany, Vol. 54, No. 380,
Plant Reproductive Biology Special Issue, pp. 1±10, January 2003
DOI: 10.1093/jxb/erg034
Meiotic cytology and chromosome behaviour in wild-type
Arabidopsis thaliana
Susan J. Armstrong1 and Gareth H. Jones
School of Biosciences, The University of Birmingham, Birmingham B15 2TT, UK
Received 12 June 2002; Accepted 12 September 2002
Abstract
This article reviews the historical development of
cytology and cytogenetics in Arabidopsis, and
summarizes recent developments in molecular cytogenetics, with special emphasis on meiotic studies.
Despite the small genome and small chromosomes
of Arabidopsis, considerable progress has been
made in developing appropriate cytogenetical techniques for chromosome analysis. Fluorescence in
situ hybridization (FISH) applied to extended meiotic pachytene chromosomes has resulted in a standardized karyotype (idiogram) for the species that
has also been aligned with the genetical map. A
better understanding of ¯oral and meiotic development has been achieved by combining cytological
studies, based on both sectioning and spreading
techniques, with morphometric data and developmental landmarks. The meiotic interphase, preceding prophase I, has been investigated by marking
the nuclei undergoing DNA replication with BrdU.
This allowed the subclasses of meiotic interphase
to be distinguished and also provided a means to
time the duration of meiosis and its constituent
phases. The FISH technique has been used to analyse in detail the meiotic organization of telomeres
and centromeric regions. The results indicate that
centromere regions do not play an active role in
chromosome pairing and synapsis; however, telomeres pair homologously in advance of general
chromosome synapsis. The FISH technique is currently being applied to analysing the pairing and
synapsis of
interstitial
chromosome
regions
through interphase and prophase I. FISH probes
also allow the ®ve bivalents of Arabidopsis to be
identi®ed at metaphase I and this has permitted an
1
analysis of chiasma frequencies in individual bivalents, both in wild-type Arabidopsis and in two
meiotic mutants.
Key words: Arabidopsis, chiasmata, chromosomes,
cytogenetics, cytology, FISH, immunocytology.
Introduction
Meiosis is a highly conserved process in eukaryotes,
occupying a central role in the life cycles of all sexually
reproducing organisms. Two successive rounds of chromosome segregation follow a single round of DNA replication, producing four haploid products. The segregation of
homologous pairs of chromosomes at the ®rst division is
dependent on their prior pairing, synapsis and recombination at earlier stages. The second meiotic division serves to
separate the two sister chromatids of each chromosome.
Subsequent fertilization of male and female gametes
restores the diploid state. An understanding of the meiotic
process is pivotal to furthering research on reproduction,
fertility, genetics and breeding, and in plants, has
implications for crop production.
Analysis of meiosis in the ¯owering plant, Arabidopsis
thaliana L. (2n=10), is a growing area of research.
Arabidopsis has a number of positive attributes for meiotic
studies, including a large pool of tagged meiotic mutants
and the molecular tools available to characterize them
(Caryl et al., 2003). As cytology and cytogenetics are
central to meiotic studies, this review aims to bring
together existing information on these important aspects of
meiosis in wild-type Arabidopsis.
The nuclear genomes of most angiosperms are typically
several-fold to 100 times larger than that of Arabidopsis
(Bennett and Smith, 1991). The current estimate of the
To whom correspondence should be addressed. Fax: +44 (0)121 414 5925. E-mail: [email protected]
ã Society for Experimental Biology 2003
2 Armstrong and Jones
genome size of Arabidopsis is 125 MB, comprising 25 498
genes. One of the reasons for the small genome is that,
unlike most angiosperms, less than 10% of its genome
consists of repeated sequences, principally comprising the
telomeres, the pericentromeric heterochromatin and the
rDNA genes associated with chromosomes 2 and 4.
Publication of the complete genome sequence of
Arabidopsis has provided the foundation for the comparison of conserved processes in eukaryotes, identifying a
wide range of plant-speci®c gene functions and establishing rapid ways to identify genes, including those associated with meiosis (Arabidopsis Genome Initiative, 2000).
Development of cytogenetics in Arabidopsis
The relatively small nuclear genome of Arabidopsis is at
odds with the general requirement of the cytogeneticist for
large chromosomes and good chromosome morphology.
Many of the cytogenetic advances of the twentieth century,
including chromosome behaviour in meiosis, were based
on species that met these requirements, such as Liliaceous
plants, Amphibia and Orthopteran insects. Because of its
small chromosomes and consequent poor chromosome
morphology, the earliest Arabidopsis cytogenetic studies
only con®rmed the number of chromosomes in the
complement (Jones and Heslop-Harrison, 2000, and references therein). Early attempts at karyotype analysis were
only moderately successful. Giemsa C-banding led to
improvements in chromosome identi®cation together with
recognition that the haploid genome contained two
nucleolus organizing regions (NORs) and, ultimately, an
alignment of the cytological chromosome map with the
genetic linkage map (Steinitz-Sears and Lee Chen, 1970;
Ambros and Schweizer, 1976; Schweizer et al., 1987).
These early cytogenetic studies of Arabidopsis relied on
squashing techniques and conventional cytological stains,
applied to root-tip meristem preparations or pollen mother
cells. Much clearer images of mitotic chromosomes of
Arabidopsis were eventually obtained by staining with the
DNA ¯uorochrome DAPI, and the locations of NORs (45S
rDNA), 5S rDNA and pericentromeric heterochromatin
blocks have been con®rmed by ¯uorescence in situ
hybridization (FISH) (Maluszynska and Heslop-Harrison,
1991; Murata et al., 1997; Brandes et al., 1997).
The requirement for clearer chromosome morphology in
Arabidopsis meiosis led to an adaptation of a spreading
method originally developed for preparing tomato meiotic
pachytene chromosomes (Zhong et al., 1996). This
improved methodology employs enzyme digestion of
pollen mother cells (PMCs) or embryo-sac mother cells
(EMCs) followed by acid disassociation and air-drying of
cells directly onto slides, combined with DAPI staining
(Ross et al., 1996; Armstrong et al., 1998; Armstrong and
Jones, 2001). The development of molecular cytogenetic
techniques together with improved cytology has led to
Arabidopsis becoming a useful model for the investigation
of chromatin organization both in meiosis and in the
interphase cell (Lysak et al., 2001; Armstrong et al., 2001),
and for the characterization of meiotic mutants (Ross et al.,
1997; Caryl et al., 2003).
The Arabidopsis karyotype
Despite the improvement in methods for the analysis of
mitotic chromosomes (see previous section), the mitotic
metaphase chromosomes of Arabidopsis are very condensed and their morphology is not well suited for detailed
karyotyping and physical mapping of chromosome landmarks. By contrast, extended pachytene bivalents of
Arabidopsis, prepared by the spreading technique, are 20±
25 times longer than the condensed mitotic chromosomes
and allow much greater precision for molecular cytogenetic analysis (Fransz et al., 1998). Each meiotic
bivalent is made up of two fully synapsed homologous
chromosome with DAPI bright pericentromeric regions
(Ross et al., 1996). Two of the bivalents (chromosomes 2
and 4) have a second DAPI bright region associated with
the nucleolar organizing regions (NORs). Measurements
based on bivalent length, arm ratios and features of the
DAPI bright regions, combined with FISH using 5S and
45S rDNA probes have produced a de®nitive pachytene
karyotype (Fransz et al., 1998; Fig. 1). In agreement with
earlier karyotyping exercises the chromosomes are numbered according to their corresponding linkage groups
(Schweizer et al., 1987), despite the non-concurrence with
their physical rankings.
Figure 1 illustrates the karyotype of Arabidopsis
(Ws). Chromosomes 1, 3 and 5 constitute a metacentric/
sub-metacentric group of chromosomes while chromosomes 2 and 4 are both short acrocentric chromosomes
Fig. 1. Karyotype of pachytene chromosomes of A. thaliana.
Pericentromeric heterochromatin (grey), NOR (stippled) and 5SrDNA
(black).
Arabidopsis meiotic cytology 3
with distinctly unequal arms. Chromosome 1 is the
largest metacentric/sub-metacentric chromosome and
lacks any rDNA sites. Chromosome 2 is a small
acrocentric chromosome having a short-arm NOR
corresponding to a 45S rDNA site. Chromosome 3 is
the shortest metacentric/sub-metacentric chromosome;
this chromosome is variable between accessions with
respect to possession, location and size of a 5S rDNA
site. Chromosome 4 is another NOR-bearing acrocentric chromosome, similar in overall size and organization to chromosome 2, but distinguished by
possession of an invariant 5S rDNA site located
proximally on the short arm. Chromosome 5 is the
second longest metacentric/sub-metacentric chromosome, additionally distinguished from chromosome 1
by possession of an invariant large 5S rDNA site
located proximally on the shorter arm (Fig. 1).
In addition to its value for karyotyping purposes,
Fransz et al. (2000) demonstrated the powerful nature
of the FISH technique applied to spread pachytene
bivalents for high resolution physical mapping at the
cytological level. In this work an integrated cytogenetic
map was produced for the short arm of chromosome 4,
showing detailed positions of various multicopy and
unique sequences relative to euchromatin and heterochromatic segments. Moreover, they demonstrated that
a recombination cold spot was associated with a
condensed heterochromatin knob and, conversely, high
recombination regions were correlated with more
exended euchromatin.
Meiosis
Floral development
Meiosis occurs in the immature developing ¯ower bud of
Arabidopsis. The ¯ower has a simple structure, typical of
the Brassicaceae, with a calyx of four free sepals and a
corolla of four petals. There are four long medial staments
and two shorter lateral stamens. The superior gynaecium
has two carpels whose locules are separated by a false
septum (Smyth et al., 1990).
In¯orescences in Arabidopsis require a methodical
approach if the meiotic stages are to be located and
correctly identi®ed. Floral development in Arabidopsis has
been analysed in detail, based on morphometric characters
of ¯ower buds and their component parts, with 12 stages of
¯oral development recognized (Smyth et al., 1990). In
order to facilitate the rapid and ef®cient preparation of
meiocytes for cytogenetic analysis, ¯oral stage, as de®ned
by Smyth et al. (1990) has been linked with morphometric
data, developmental landmarks and meiotic stage
(Armstrong and Jones, 2001; Armstrong et al., 2001;
Table 1). The gross anatomy of anthers and ovules was
determined from observations on resin-embedded semithin sections and lightly squashed whole mounts of anthers
and/or gynoecia at appropriate developmental stages.
Figure 2 illustrates the progression of meiosis in semithin sections. Male and female meiosis in Arabidopsis, as
in other angiosperms, are asynchronous. Prophase I of
meiosis in embryo-sac mother cells commences as the
pollen mother cells arrive at the tetrad stage.
Table 1. Summary of the relationship between ¯oral development and cytological landmarks
Floral stagea
Bud size
(mm)
8. Anther and
®lament regions
already separated
(stage 7). Locules
present in the long
stamens. Primordia of
petals visible
Cytology
Meiotic stage
<0.3
Anthers
differentiated into
®ve tissue layers.
The innermost cells,
the PMCs are
surrounded by the
tapetum, the middle
cell layer, the
endothecium, and
the epidermis
Meiotic interphase
in the PMCs
9. Petal primordia
stalked at base
0.3±0.4
Tapetal cells
become binucleate
at zygotene/early
pachytene stages
Meiosis in PMCs
10. Petals level with
the short stamens.
Stamens green
<0.5
0.3±0.4
Gynoecium cylinder
deeply slotted.
0.5
Closure of cylinder,
appearance of
stigmatic papillae
Microspores.
Meiotic interphase
in EMC.
Meiosis in EMCs
11. Anthers green,
but changing to
yellow as gynoecium enlarges
<0.7
0.6±0.8
Stigmatic papillae
develop
a
As described by Smyth et al. (1990).
Gynoecium length
during meiosis
Meiosis in EMCs
4 Armstrong and Jones
Fig. 2. Semi-thin sections through young ¯ower buds of A. thaliana, stained with toluidine blue, showing complete anther locules containing
PMCs in meiotic interphase (A), pachytene (B) and tetrads (C) surrounded by four cell layers. (Bar=10 mm).
Anther and PMC development
The smallest ¯ower buds that could be practically
dissected from in¯orescences and processed for spreading
were about 0.2±0.3 mm long and coincided with late stage
7/early stage 8. Before this stage the archesporial cells
undergo periclinal divisions to give rise to the primary
parietal and primary sporogenous layers (Misra, 1962).
Following the cessation of asynchronous archesporial
mitotic divisions in anthers, the microsporocytes or pollen
mother cells (PMCs) develop more or less synchronously
through meiosis. By ¯oral stage 8, the ®ve layers of the
stamen have differentiated. PMCs are surrounded by four
layers: the tapetum, the middle layer, the endothecium and
the epidermis, each of which is one cell thick (Owen and
Makaroff, 1995; Fig. 2). The PMCs are about three times
larger than tapetal cells. They have very thin walls at this
stage and contain a large centrally located nucleus with a
prominent nucleolus (Fig. 2A). On the other hand, the
densely staining tapetal cells are smaller, and the nucleus
and nucleolus are correspondingly smaller. Light squashes
of anthers at this stage result in the extrusion of PMCs
which tend to remain attached to each other, forming
columns of cells. This feature is presumed to be due to the
presence of intercellular cytoplasmic channels connecting
the meiocytes at early stages of development (Owen and
Makaroff, 1995). This is a common feature of pollen
mother cells at early stages of development, and has been
described in several different plant species (Esau, 1977).
During later stages of meiosis these channels are closed
coincident with the deposition of callose on PMC walls
and extruded PMCs are consequently unconnected and
easily separated. During ¯oral stage 9 the remaining
meiotic stages take place. In sections of Arabidopsis
anthers the leptotene stage is dif®cult to distinguish due to
the extreme ®neness of the chromosome threads, but this
stage is associated with movement of the nucleolus to the
periphery of the nucleus. More de®nite chromatin threads
are visible by zygotene due to the progressive condensation of the chromosomes and their synapsis into bivalent
structures. At this stage the chromosomes are characteristically polarized to one side of the PMC nucleus, and
deposition of callose onto PMC walls is evident. The
tapetal cells become binucleate at zygotene±pachytene
stages (Fig. 2B) as a result of a single synchronized mitotic
division without subsequent cytokinesis. By pachytene, the
chromosomes in the PMCs are fully synapsed, which in
sections can, on closer observation, be visualized as thick
threads (Fig. 2B). From pachytene onwards the PMCs have
Arabidopsis meiotic cytology 5
thick callose walls and the subsequent stages of divisions I
and II are easily distinguished, culminating in the appearance of tetrads (Fig. 2C).
Ovule and EMC development
The gynoecium of Arabidopsis consists of two fused
carpels that develop as a single cylinder. A total of 40±60
ovules are produced in four rows (Mans®eld and Bowman,
1994). In the Arabidopsis accession Ws there are about 40
ovules in each gynoecium (Armstrong and Jones, 2001).
There is a clear relationship between meiotic progression and gynoecium length and morphology (Armstrong
and Jones, 2001; Table 1). The smallest gynoecia (length
0.3±0.4 mm) have open pistils with a distinctly slotted
appearance at their ends and appear at ¯oral stage 9/10.
Gynoecia at this stage contain megasporocytes or embryosac mother cells (EMCs) that are at meiotic interphase.
Larger gynoecia (length 0.5±0.8 mm) are characterized by
closure of the pistil end and appearance of stigmatic
papillae (still within ¯oral stage 10). Meiotic stages from
prophase I onwards are found in gynoecia in this length
range. There is considerable synchrony in meiotic development between the different ovules within an in¯orescence.
In female meiosis, the single-celled archesporangium
functions directly as the EMC and forms a distinctive cell
even in the meiotic interphase and early meiotic stages. In
sections, the EMC is seen to be more than three times
larger than the surrounding cells of the nucellum
(Armstrong and Jones, 2001; Webb and Gunning, 1990;
Bajon et al., 1999). In lightly squashed preparations, the
EMC is again seen to be clearly larger than any of the
surrounding cells of the ovule. It contains a large nucleus
with a prominent nucleolus and a large volume of
cytoplasm with organelles.
Observations on sectioned and lightly squashed anthers
and gynoecia are useful for determining the timing and
progression of meiosis in relation to ¯oral development,
but they are not appropriate for detailed observations on
chromosome organization and behaviour during meiosis.
Conventional squashing and staining procedures using
classical stains such as orcein or carmine are also
unsuitable for analysis of Arabidopsis meiosis. This is
principally because they fail to produce a suf®ciently
intense staining of the very ®ne extended chromosomes
that characterize the prophase I stages of meiosis in
Arabidopsis. Enzyme digestion/acid disassociation and
air-drying of meiocytes onto slides, combined with DAPI
staining gives superior resolution of Arabidopsis chromosomes, particularly in the critical prophase I stages, and
forms the basis for the following description of meiosis.
Male and female meiosis in Arabidopsis are very similar
in most respects, with the exception of the tetrad stage. The
following description is therefore based on the progression
of meiosis in PMC, with references to female meiosis
where necessary.
This description provides reference observations for
interpreting the meiotic stages in Arabidopsis. It is beyond
the scope of this article to provide ®gures to illustrate all
the key stages of meiosis mentioned in the text. However, a
series of images detailing the progression of meiosis in
Arabidopsis PMCs is presented in an earlier paper (Ross
et al., 1996; see also Fig. 1 in Caryl et al., 2003).
The meiotic interphase
Descriptions of meiosis commonly commence at the onset
of prophase I, with the ®rst appearance of distinct
chromosome threads at leptotene, but there is a widespread
and increasing recognition that meiotic interphase (sometimes termed premeiotic interphase) encompasses many
important events despite its uniform and relatively undifferentiated appearance (Dover and Riley, 1977; Loidl,
1990; Zickler and Kleckner, 1998, 1999). Following the
demonstration by Taylor and Macmaster (1954) that
meiotic DNA replication occurs during meiotic interphase,
and not prophase I as was originally thought, the meiotic Sphase has been extensively analysed in a wide range of
eukaryotes. A common feature of this S-phase is its
extended duration compared to somatic/mitotic S-phases
(John, 1990). The meiotic S-phase also serves as a useful
mid-interphase marker, separating the unreplicated G1
phase from the replicated G2 phase. Several other important events are likely to occur during this interphase,
including chromatin modi®cation and the expression of
many meiotic genes whose products are required for the
successful completion of meiosis. Despite the general
realization of the importance of the meiotic interphase, the
analysis of meiosis in Arabidopsis has, to date, barely
considered this phase. Immunocytological approaches are
now being used to investigate both the S-phase and the
appearance of key meiotic proteins in Arabidopsis PMCs.
The duration of meiotic G2, and ensuing meiotic stages,
may be determined by marking the S-phase and subsequently sampling the marked population in a time-course
experiment. Previous experimental investigations involving meiotic S-phase labelling in plants relied on tritiated
thymidine, a radioisotopically labelled precursor of DNA,
combined with autoradiographic detection of incorporated
label (Bennett et al., 1971; Holm, 1977). An alternative
labelling technique that avoids the time-consuming autoradiographic detection is to mark the cells in S-phase with
the thymidine analogue bromodeoxyuridine (BrdU) that
can be detected immunocytologically using anti-BrdU
antibody carrying a ¯uorescent tag.
This approach has been applied to the labelling of
Arabidopsis meiotic S-phase. By immersing cut stems
bearing in¯orescences in BrdU solution uptake of BrdU
could be detected in a subset of meiotic interphase nuclei
6 Armstrong and Jones
within 1 h and a high level of labelling within 2 h. In a
time-course experiment, cut stems were immersed in BrdU
for 2 h and then transferred to water. In¯orescences were
sampled at intervals, ®xed and processed for detection of
BrdU in cytological preparations. By this means the
duration of meiotic G2 and all subsequent stages as far as
tetrads has been determined, as well as the duration of Sphase itself (SJ Armstrong, GH Jones, unpublished data).
In addition, these experiments allowed positive identi®cation of S-phase meiocytes, as those labelled after short
periods of exposure to BrdU, and also G2 cells as those that
became labelled as cells passed from S-phase to G2 at
progressively later time points. By a process of elimination
it was possible to identify G1 interphase cells as a category
of cells, with distinctive morphology, that were not
labelled at short sampling times after the pulse-label. G1
cells were expected to be found in the next smallest buds to
those containing S labelled cells. They were smaller than
those of S-phase and G2 but nevertheless were still larger
than the surrounding somatic cells. They typically had a
relatively large nucleolus and condensed pericentromeric
heterochromatin. By contrast the S-phase cells were larger
and the pericentromeric heterochromatin was decondensed
as evidenced by lack of DAPI bright regions. G2 cells
showed a very similar nuclear organization to S-phase in
terms of nuclear size, nucleolus size and location, and
chromatin condensation. The overall duration of meiosis in
Arabidopsis measured at 20 °C, from S-phase to the
formation of tetrads, is 36 h.
Prophase I
Early leptotene proper is distinctive and is characterized by
extensive stretches of unsynapsed chromosome axes
having bead-like chromomere differentiation along their
length and the reappearance of condensed pericentromeric
heterochromatin. However, the transition from late G2 to
early leptotene is gradual and inde®nable, with the
appearance of short stretches of chromosome axes, that
probably correspond to axial elements at the electron
microscope level, and progressive condensation of the
heterochromatin. This gradual transition makes precise
determination of the relative durations of G2 and leptotene
dif®cult. In early leptotene the nucleolus is a large
structure, taking up as much as one-third of the nuclear
volume. It occupies a central position in the nucleus, but
towards the end of the leptotene it moves progressively
towards the nuclear periphery, where it remains throughout
the rest of prophase I. Zygotene, by de®nition, is the stage
of chromosome synapsis, and early zygotene is therefore
characterized by the ®rst indications of synapsis. The
zygotene stage in Arabidopsis characteristically fails to
produce well spread preparations. This appears to be due to
a characteristic polarization and clumping of the chromosomes towards one side of the nucleus and is accompanied
by a tendency for aggregation of the pericentromeric
heterochromatin into a variable number of clumps. At
these early meiotic prophase I stages in PMCs, the
organelles are characteristically distributed. Typically,
they accumulate at the opposite pole to the chromatin
during late leptotene/zygotene. This arrangement was not
observed in the EMC, which could be due to organizational
differences between the single cell (EMC) compared to the
many PMCs in the central column of the anther.
At pachytene, the chromosomes are much shorter
compared to the earlier prophase stages (Fransz et al.,
1998). All chromosomes are seen to be fully synapsed
double structures, with obvious chromomere differentiation. At the electron microscope level the synapsed
homologues are seen to be associated by typical tripartite
synaptonemal complexes (SCs) running the entire length
of each of the ®ve bivalents (Albini, 1994). By late
pachytene, the bivalents are quite well separated from each
other and can often be fully traced. The only exception to
this is a tendency for the NORs, located subterminally on
the short arms of chromosomes 2 and 4, to be associated
and it appears that they stay together through the rest of
prophase I, up to and including diakinesis.
The transition from pachytene to diplotene involves
the gradual and progressive separation or desynapsis of the
homologues along most of their length, characterized at
the EM level by the disruption and eventual disappearance
of the SC. Once the chromosomes are fully desynapsed,
except at chiasmata, the early diplotene nuclei are ®lled
with extended and rather fuzzy single chromosome
threads, corresponding to the so-called diffuse stage
described in many organisms including angiosperms.
From this point onwards the ®ve bivalents gradually
condense to give discrete bivalent structures, associated by
chiasmata, the degree of condensation increasing progressively from late diplotene, through diakinesis to its
maximum at metaphase I.
Metaphase I to telophase II
At metaphase I ®ve condensed bivalents are present in
wild-type meiosis. They show co-orientation on the
metaphase I spindle, and each bivalent contains from one
to three chiasmata (see below). Anaphase I achieves the
separation of homologous whole chromosomes, whose
sister chromatids remain associated at centromere regions.
This stage is of brief duration and is rarely observed in
Arabidopsis. Telophase I leads to the dyad stage in which
the PMC contains two polar groups of chromosomes that
partially decondense without achieving the full interphase
condition. Partitioning of the cytoplasm does not directly
follow the ®rst nuclear division and is deferred until after
the second division. During metaphase I and anaphase I the
cytoplasmic organelles are excluded from the spindle and
subsequently form a distinct band or saddle between the
Arabidopsis meiotic cytology 7
daughter nuclei that persists into the second division.
Metaphase II and anaphase II follow the normal course,
with separation of chromatids into daughter nuclei. Late
anaphase II cells contain the expected four groups of ®ve
chromatids within a common cytoplasm, which gradually
re-form into tetrads of haploid interphase nuclei.
An overall conclusion from this survey is that meiosis in
Arabidopsis follows a conventional pattern. Apart from an
interesting difference in the mode of telomere clustering
(described below) meiosis is comparable to that in other
plant species that have been extensively analysed cytogenetically, such as maize, rye and tomato. Meiosis also
appears to be very similar in PMCs and EMCs of
Arabidopsis. The major difference concerns the fate of
the haploid meiotic products. In anthers, the haploid cells
(microspores) all develop into pollen grains, the male
gametophytes. In ovules, on the other hand, the haploid
meiotic products form a linear tetrad of megaspores, three
of which abort leaving a single megaspore which undergoes mitotic division to give rise to the embryo-sac, the
female gametophyte.
Molecular cytogenetics of meiosis
Superimposed on this basic descriptive framework of
Arabidopsis meiosis, molecular cytogenetic approaches
have been applied to extend understanding of key meiotic
events and processes. An illustration of this, as described
earlier, is the use of BrdU labelling of replicating DNA,
and detection by anti-BrdU antibody, to identify and mark
cells in meiotic S-phase.
Immunocytology is also being applied to determine the
time of expression of meiotic genes and the intranuclear
locations of the proteins they encode. Several Arabidopsis
meiotic genes have now been cloned and subsequently
expressed in a bacterial in vitro system to yield
recombinant protein products. These can then be used to
generate antibodies to meiotic proteins which can be
employed in immunolocalization studies. For example,
expression of the Asy1 protein in Arabidopsis and
Brassica oleracea has been investigated by this means.
This protein is required for normal meiotic chromosome
synapsis since an insertional mutant (asy1) has been shown
to be completely asynaptic (Caryl et al., 2000). The protein
is initially detected during meiotic interphase as numerous
punctate foci distributed over the chromatin. As prophase I
proceeds, the signal becomes increasingly continuous and
is closely associated with the axial elements of unsynapsed
chromosomes and with the synaptonemal complexes of
synapsed chromosomes, but is not present in the chromatin
loops extending from these structures. Immunogold
labelling in conjunction with electron microscopy established that the protein localizes to regions of chromatin that
are closely associated with the axial/lateral elements,
rather than being a component of these structures. These
observations indicate that the protein is required for
morphogenesis of the synaptonemal complex, possibly
by de®ning regions of chromatin that associate with the
axial elements (Armstrong et al., 2002). Using a similar
approach the expression and localization of several other
meiotic proteins are currently being investigated.
FISH is also being used to label chromosomal structures,
or just chromosomal sites, in order to analyse the progress
of chromosome pairing and synapsis. This approach has
been applied successfully to the analysis of these events in
several other organisms (reviewed in Zickler and
Kleckner, 1998, 1999). In Arabidopsis, FISH was used to
mark centromeres and telomeres, and their intranuclear
arrangements were analysed from meiotic interphase
through to late prophase I (Armstrong et al., 2001;
Fig. 3). Centromere regions of Arabidopsis are unpaired,
widely dispersed and peripherally located in nuclei during
meiotic interphase, and they remain unpaired and unassociated throughout leptotene. Eventually they associate
pairwise during zygotene, as part of the nucleus-wide
synapsis of homologous chromosomes. Telomeres, on the
other hand, show a persistent association with the
nucleolus throughout meiotic interphase. Variation in
telomere signal number indicates that telomeres undergo
pairing during this interval, preceding the onset of general
chromosome synapsis. During leptotene the paired telomeres lose their association with the nucleolus and become
widely dispersed. As the chromosomes synapse during
zygotene, the telomeres reveal a loose clustering within
one hemisphere, which may represent a degenerate or relic
bouquet con®guration. It is proposed that in Arabidopsis
the classical leptotene/zygotene bouquet is absent and is
replaced functionally by nucleolus-associated telomere
clustering. As an extension to this approach, FISH probes
are currently being used to mark selected chromosomal
sites in order to analyse the pairing and synaptic behaviour
of non-centromeric and non-telomeric regions.
Chiasma analysis in Arabidopsis
Although chiasmata can occasionally be clearly resolved
in diplotene and diakinesis of Arabidopsis, chiasma
scoring at these stages has associated problems due to
the dif®culty of distinguishing relational twists from
genuine chiasmata. In addition, the persistent association
of the NORs with the nucleolus at this stage makes it
virtually impossible to identify chiasmata in the short arms
of chromosomes 2 and 4. Metaphase I is therefore
preferable for chiasma scoring in this species, despite
the highly condensed state of the bivalents. FISH with the
repeated 5S and 45S rDNA probes can identify each of
the ®ve bivalents at metaphase I. Using this approach the
chiasma frequencies of marked individual bivalents of the
Ws accession of Arabidopsis have been analysed, as well
as two meiotic mutants of Arabidopsis, one asynaptic
8 Armstrong and Jones
Fig. 3. FISH of telomere (red) and pericentromeric heterochromatin (CENs) (green) probes to wild-type A. thaliana PMC nuclei. (A) S-phase
nucleus showing extended CENs and unpaired telomeres clustered around the nucleolus. (B) early leptotene nucleus showing dispersed CENs and
paired telomeres. (C) Pachytene/very early diplotene nucleus showing paired CENs and telomeres; arrow indicates interstitial telomere locus
adjacent to CEN on chromosome 1. (D) Diplotene/diakinesis nucleus showing ®ve bivalents. (E) Metaphase I nucleus showing ®ve fully
condensed and aligned bivalents. (F) Tetrad stage showing four haploid nuclei; note telomeres have reassociated with the reformed nucleoli
(Bar=10 mm).
(asy1) and one desynaptic (dsy1) (Sanchez-Moran et al.,
2001). The mutants were isolated from T-DNA transformed lines created in the Ws background genotype (Ross
et al., 1997). A wild-type mean chiasma frequency of 9.24
per cell was determined from a sample of 50 cells, a value
that is consistent with estimated genetic recombination
values. Individual bivalent chiasma frequencies varied
according to chromosome size; chromosome 1 had the
highest mean chiasma frequency (2.14) while the acrocentric chromosomes had the lowest frequencies, 1.54 and
1.56 for chromosomes 2 and 4, respectively. Chiasma
analysis of female meiosis is more demanding since each
ovule contains but a single EMC and, consequently, the
estimate of female chiasma frequency is based on fewer
cells. A mean EMC chiasma frequency of 8.5 was obtained
from a sample of ten diakinesis and metaphase I cells
(Armstrong and Jones, 2001). Chiasma frequency has been
shown to differ signi®cantly between the sexes in several
different species (John, 1990), but there is no consistency
in the direction of the difference. In some cases male
Arabidopsis meiotic cytology 9
chiasma frequency is higher, while in others the female has
a higher frequency; in yet other cases no difference was
found. The observation of a lower chiasma frequency in
Arabidopsis EMCs compared to PMCs is consistent with
genetic recombination differences reported by Vizir and
Korol (1990).
Conclusion
Arabidopsis is proving to be a convenient and powerful
model system for the analysis of meiosis. The availability
of good cytology combined with developments in molecular cytogenetics and molecular biology put it on a par
with other `higher' eukaryotic models such as Drosophila,
C. elegans and mouse and, in addition, extend the range of
model organisms to the plant kingdom. One of the aims of
current meiosis research is to understand whether and to
what extent meiotic processes and controls are conserved
across all eukaryotes and, in this context, the inclusion of a
plant model is very important. Meiosis is a highly complex
process involving the action and interaction of very many
genes. As a result of recent research in a number of model
systems there is now a better understanding of these genes,
the processes they determine and their regulation.
However, a full understanding of the meiotic process is
still a distant prospect and this ®eld will continue to be a
fascinating and rewarding ®eld of endeavour for many
years to come.
Acknowledgement
We are grateful to the BBSRC for their ®nancial support of
much of the work reported in this review.
References
Albini SM. 1994. A karyotype of the Arabidopsis thaliana genome
derived from synaptonemal complex analysis at prophase I of
meiosis. The Plant Journal 5, 665±672.
Ambros P, Schweizer D. 1976. The Giemsa C-banded karyotype of
Arabidopsis thaliana (L) HEYN. Arabidopsis Information
Service 13, 167±171.
Arabidopsis Genome Initiative. 2000. Analysis of the genome
sequence of the ¯owering plant Arabidopsis thaliana. Nature 408,
796±815.
Armstrong SJ, Caryl AP, Jones GH, Franklin FCH. 2002. Asy1,
a protein required for meiotic chromosome synapsis, localizes to
axis-associated chromatin in Arabidopsis and Brassica. Journal
of Cell Science 115, 3645±3655.
Armstrong SJ, Franklin FCH, Jones GH. 2001. Nucleolusassociated telomere clustering and pairing precede meiotic
chromosome synapsis in Arabidopsis thaliana. Journal of Cell
Science 114, 4207±4217.
Armstrong SJ, Fransz P, Marshall DF, Jones GH. 1998. Physical
mapping of DNA repetetive sequences to mitotic and meiotic
chromosomes of Brassica oleracea var. alboglabra by
¯uorescence in situ hybridization. Heredity 81, 666±673.
Armstrong SJ, Jones GH. 2001. Female meiosis in wild-type
Arabidopsis thaliana and in two meiotic mutants. Sexual Plant
Reproduction 13, 177±183.
Bajon C, Horlow C, Moyamayer JC, Sauvanet A, Robert D.
1999. Megasporogenesis in Arabidopsis thaliana L.: an
ultrastructural study. Sexual Plant Reproduction 12, 99±119.
Bennett MD, Chapman V, Riley R. 1971. The duration of meiosis
in pollen mother cells of wheat, rye and Triticale. Proceedings of
the Royal Society of London, Series B 178, 259±275.
Bennett MD, Smith JB. 1991. Nuclear DNA amounts in
angiosperms. Philosophical Transactions of the Royal Society
of London, Series B 274, 227±274.
Brandes A, Thompson H, Dean C, Heslop-Harrison JS. 1997.
Multiple repetitive DNA sequences in the paracentromeric
regions of Arabidopsis thaliana L. Chromosome Research 5,
238±246.
Caryl AP, Armstrong SJ, Jones GH, Franklin FCH. 2000. A
homologue of the yeast HOP1 gene is inactivated in the
Arabidopsis meiotic mutant asy1. Chromosoma 109, 62±71.
Caryl AP, Jones GH, Franklin GCH. 2003. Dissecting plant
meiosis using Arabidopsis thaliana mutants. Journal of
Experimental Botany 54, 000±000.
Dover GA, Riley R. 1977. Inferences from genetical evidence on
the course of meiotic chromosome pairing in plants.
Philosophical Transactions of the Royal Society of London,
Series B 277, 313±326.
Esau K. 1977. Anatomy of seed plants, 2nd edn. New York: Wiley.
Fransz P, Armstrong S, Alonso-Blanco C, Fischer TC, TorresRuiz RA, Jones GH. 1998. Cytogenetics for the model system
Arabidopsis thaliana. The Plant Journal 13, 867±876.
Fransz P, Armstrong S, de Jong JH, Parnell LD, van Drunen C,
Dean C, Zabel P, Bisseling T, Jones GH. 2000. Integrated
cytogenetic map of chromosome 4S of A. thaliana: structural
organization of heterochromatic knob and centromere region.
Cell 100, 367±376.
Holm PB. 1977. The premeiotic DNA replication of euchromatin
and heterochromatin in Lilium longi¯orum (Thunb.). Carlsberg
Research Communications 42, 103±151.
John B. 1990. Meiosis. Cambridge: Cambridge University Press.
Jones GH, Heslop-Harrison JS. 2000. Classical and molecular
cytogenetics of Arabidopsis. In: Wilson ZA, ed. Arabidopsis: a
practical approach. Oxford: Oxford University Press, 105±124.
Loidl J. 1990. The initiation of chromosome pairing: the
cytological view. Genome 33, 759±778.
Lysak MA, Fransz PF, Ali HBM, Schubert I. 2001. Chromosome
painting in Arabidopsis. The Plant Journal 28, 689±697.
Maluszynska J, Heslop-Harrison JS. 1991. Localization of
tandemly repeated DNA sequences in Arabidopsis thaliana. The
Plant Journal 1, 159±166.
Mans®eld SG, Bowman J. 1994. Ovules. In: Bowman J, ed.
Arabidopsis. An atlas of morphology and development. New
York: Springer-Verlag, 299±301.
Misra RC. 1962. Contribution to the embryology of Arabidopsis
thaliana (Gay and Monn.) Agra University Journal of Research
11, 191±199.
Murata M, Heslop-Harrison JS, Motoyoshi F. 1997. Physical
mapping of the 5S ribosomal RNA genes in Arabidopsis thaliana
by multi-colour ¯uorescence. The Plant Journal 12, 31±37.
Owen HA, Makaroff CA. 1995. Ultrastructure of
microsporogenesis and microgametogenesis in Arabidopsis
thaliana (L.) Heynh. ecotype Wassilewskija (Brassicaceae).
Protoplasma 185, 7±21.
Ross KJ, Fransz P, Armstrong SJ, Vizir I, Mulligan B, Franklin
FCH, Jones GH. 1997. Cytological characterization of four
meiotic mutants of Arabidopsis isolated from T-DNA
transformed lines. Chromosome Research 5, 551±559.
Ross KJ, Fransz P, Jones GH. 1996. A light microscopic atlas of
10 Armstrong and Jones
meiosis in Arabidopsis thaliana. Chromosome Research 4, 507±
516.
Sanchez-Moran E, Armstrong SJ, Santos JL, Franklin FCH,
Jones GH. 2001. Chiasma formation in Arabidopsis thaliana and
in two meiotic mutants. Chromosome Research 9, 121±128.
Schweizer D, Ambros P, Grundler P, Varga F. 1987. Attempts to
relate cytological and molecular chromosome data of Arabidopsis
thaliana to its genetic linkage map. Arabidopsis Information
Service 25, 27±34.
Smyth DR, Bowman JL, Meyerowitz EM. 1990. Early ¯ower
development in Arabidopsis. The Plant Cell 2, 755±767.
Steinitz-Sears LM, Lee-Chen S. 1970. Cytogenetic studies in
Arabidopsis thaliana. Canadian Journal of Genetics and
Cytology 12, 217±223.
Taylor JH, MacMaster RD. 1954. Autoradiographic and
microphotometric studies of deoxyribose nucleic acid during
microgametogenesis in Lilium longi¯orum. Chromosoma 6, 489±
521.
Vizir I, Korol AB. 1990. Sex differences in recombination
frequency in Arabidopsis. Heredity 65, 379±383.
Webb MC, Gunning BES. 1990. Embryo-sac development in
Arabidopsis thaliana. I. Megasporogenesis including the
microtubular cytoskeleton. Sexual Plant Reproduction 7, 153±
163.
Zhong X, de Jong JH, Zabel P. 1996. Localization of repetitive
sequences on metaphase and pachytene chromosomes of tomato
using ¯uorescence in situ hybridization. Chromosome Research 4,
24±28.
Zickler D, Kleckner N. 1998. The leptotene-zygotene transition of
meiosis. Annual Review of Genetics 32, 619±697.
Zickler D, Kleckner N. 1999. Meiotic chromosomes: integrating
structure and function. Annual Review of Genetics 33, 603±704.