PII: S1534-5807(02)00122-3


Developmental Cell, Vol. 2, 135–142, February, 2002, Copyright  2002 by Cell Press

ReviewBuilding Beauty:
The Genetic Control of Floral Patterning

about the mechanisms underlying this process. Be-
cause at this point there is a very large number of original
publications in this field, we have cited reviews for most

Jan U. Lohmann1 and Detlef Weigel1,2,3
1Plant Biology Laboratory
The Salk Institute for Biological Studies
La Jolla, California 92037 of the work published before the mid-1990s.
2 Department of Molecular Biology
Max Planck Institute for Developmental Biology The ABCs of Flower Development
72076 Tübingen Contemporary work on floral patterning began with the
Germany study of a series of mutants in which floral organs de-

velop normally, but in the inappropriate whorl. Such
mutants had been collected from garden snapdragon,

Floral organ identity is controlled by combinatorial ac- Antirrhinum majus, by Hans Stubbe, and from the mus-
tion of homeotic genes expressed in different territories tard relative Arabidopsis thaliana by Maarten Koornneef.
within the emerging flower. This review discusses recent In the late 1980s, three groups, headed by Enrico Coen
progress in our understanding of floral homeotic genes, in the United Kingdom, Elliot Meyerowitz in the United
with an emphasis on how their region-specific expres- States, and Heinz Saedler in Germany, recognized the
sion is regulated. value of these mutants as homeotic mutants, and used

them to initiate molecular and genetic studies of floral
Although flowers appear in a stunning diversity of forms, patterning. The initial genetic studies quickly led to pro-
from the intricate and beautiful to the simple and incon- posal of the ABC model, now considered a milestone
spicuous, their basic plan is remarkably invariant across in plant developmental biology (Bowman et al., 1991;
all species. The flowers of dicots, which represent one Coen and Meyerowitz, 1991). Based on phenotypic and
of the two major subdivisions of flowering plants and genetic analyses, the model states that development of
include the reference plant Arabidopsis thaliana, are the four types of floral organs is governed by overlapping
organized into four concentric rings of organs, termed activities of three classes of regulatory genes. Termed
whorls (Figures 1A and 1B). The outer two whorls are A, B, and C, each class of genes is active in two adjacent
occupied by sterile organs, with the normally green se- whorls (Figure 1C). Activity of A class genes alone leads
pals that protect the emerging flower bud in the first to formation of sepals in the first whorl, while combining
whorl and the often showy and colorful petals that can their activity with that of B class genes promotes the
serve to attract pollinators in the second whorl. The formation of petals in the second whorl. Similarly, the
inner two whorls are devoted to reproduction, the central combination of B and C class activity is required for
purpose of flower formation. Stamens, the male repro- stamen formation in the third whorl, while C class genes
ductive organs that produce pollen, are found in the by themselves control formation of carpels in the fourth
third whorl, while the central fourth whorl is occupied whorl.
by carpels, the female reproductive organs, which are To account for mutant phenotypes, the ABC model
normally fused to form the gynoecium (Figure 1A). After included another tenet, namely that A and C class activ-
fertilization, the gynoecium develops into the fruit har- ity are mutually exclusive and repress each other, since
boring the seeds. Organ number in the different whorls A and C class mutants are essentially mirror images of
is typically fixed; in Arabidopsis, there are four sepals, each other. In A class mutants, C class activity expands
four petals, six stamens, and two carpels (Figure 1B). into all whorls, with sepals being replaced by carpels,

Flowers develop from primordia that arise on the and petals by stamens. Conversely, in C class mutants,
flanks of the shoot apical meristem, a self-regulating A class activity expands into whorl three and four. In
population of undifferentiated cells that forms the addition, the flower becomes indeterminate in C class
growth point of the plant. Initially, the floral primordium mutants, that is, it no longer produces a limited number
is organized in a similar manner to the shoot apical of organs, and new flowers form inside the original
meristem, with a central group of stem cells. For simplic- flower, giving rise to a flower consisting of (sepals, pet-
ity, the young floral primordium, before the emergence als, petals)n. Expression of B class genes is not affected
of floral organ primordia, is often called a floral meristem. by mutations in either A or C class genes. Therefore,
After a few days, sepal primordia arise, followed by petal inactivation of B class genes causes second whorl or-
and stamen primordia. The floral meristem is consumed gans to adopt the same fate as first whorl organs, and
by the formation of the central carpels, which either third whorl organs the same as fourth whorl organs,
arise fused or fuse shortly after they emerge. giving rise to flowers consisting of sepals, sepals, car-

During floral patterning, several processes need to pels, carpels.
occur coordinately, including the proper positioning of The original genes of the B and C classes turned out
floral organs and specification of their identity in a posi- to be orthologs in Antirrhinum and Arabidopsis (Table
tion-dependent manner. Among these, most is known 1). C class activity was initially represented by a single
about the genetic and molecular control of floral organ gene, PLENA (PLE) in Antirrhinum and its ortholog AGA-
identity, and here we summarize what has been learned MOUS (AG) in Arabidopsis. B class activity requires a

pair of related genes in both species, DEFICIENS (DEF)/
GLOBOSA (GLO) in Antirrhinum and APETALA3 (AP3)/3 Correspondence: weigel@weigelworld.org



Developmental Cell
136

Figure 1. The Basics of Flower Development

(A) Mature Arabidopsis flower with sepals (se),
petals (pe), stamens (st), and carpels (ca).
(B) Floral formula indicating whorls one to
four (w1–4).
(C) Diagram of ABC model, indicating do-
mains of ABC gene activities.

PISTILLATA (PI) in Arabidopsis. In contrast, the canoni- and promoter studies revealed that regulation occurs
mainly at the level of transcription, as the promoters ofcal A class gene APETALA2 (AP2) from Arabidopsis has

no direct counterpart in Antirrhinum, where A class activity homeotic genes are predominantly active in those
whorls where their function is required. An exception iswas only represented by dominant mutations ovulata and

macho, which later turned out to be gain-of-function alleles the A class gene AP2, which is expressed uniformly in
all whorls. AP2 is also unusual in that it is the only floralof the C class gene PLE (Weigel and Meyerowitz, 1994;

Theissen et al., 2000; Zhao et al., 2001a). homeotic gene that does not encode a MADS domain
transcription factor. Subsequently, it was discoveredAlthough the ABC model proposed that the homeotic

genes are only active in specific whorls, genetic analysis that one MADS box gene, APETALA1 (AP1), has dual
roles: it acts during early stages of flower developmentalone could not tell how their activity was regulated.

Cloning of the ABC genes with subsequent expression redundantly with other factors to specify floral identity,

Table 1. Early Floral Patterning Genes

Arabidopsis Antirrhinum Gene Product

Meristem identity
LEAFY (LFY) FLORICAULA (FLO) DNA binding, plant-specific
APETALA1 (AP1) SQUAMOSA (SQUA) MADS domain

B class regulators
UNUSUAL FLORAL ORGANS (UFO) FIMBRIATA (FIM) F box
SUPERMAN (SUP) OCTANDRA (OCT)? Zinc finger
? CHORIPETALA ?
? DESPENTEADO ?

C class regulators
WUSCHEL ? Homeodomain

General ABC repressors
CURLY LEAF (CLF) ? Polycomb group (Enhancer of zeste)
INCURVATA2 (ICU2) ? ?
EARLY BOLTING IN SHORT DAYS (EBS) ? ?
EMBRYONIC FLOWER1 (EMF1) ? Plant-specific, nuclear?
EMBRYONIC FLOWER2 (EMF2) ? Polycomb group (Suppressor of zeste 12)

General ABC activators
POLYPETALA (POLY) ?

ABC genes
A class

APETALA1 (AP1) — MADS domain
APETALA2 (AP2) ? AP2 domain
AINTEGUMENTA ? AP2 domain
LEUNIG (LUG) ? Tup1-like corepressor, WD40 repeats
STERILE APETALA (SAP) ? Plant-specific, nuclear?
? STYLOSA (STY) ?
? FISTULATA (FIS) ?

B class
APETALA3 (AP3) DEFICIENS (DEF) MADS domain
PISTILLATA (PI) GLOBOSA (GLO) MADS domain

C class
AGAMOUS (AG) PLENA (PLE) & FARINELLI (FAR) MADS domain
CRABS CLAW (CRC) ? YABBY domain
SPATULA (SPT) ? bHLH domain
HUA1 ? Plant-specific, nuclear?
HUA2 ? RNA binding domain

ABC cofactors
SEPELLATA1-3 ? MADS domain

Question marks indicate that an orthologous mutant has not been described; the dash indicates that the most closely related gene does not
have the same function.



Review
137

and it contributes during later stages to A function. Con- to that of LFY. However, although LFY is an important
regulator of AP1, AP1 activation is merely delayed, notsistent with these two roles, AP1 RNA is initially ex-

pressed throughout the flower, but becomes restricted abolished, in lfy mutants, indicating that redundant fac-
tors contribute to AP1 activation (Liljegren et al., 1999).to the A domain during later stages (Weigel and Meyero-

witz, 1994; Theissen et al., 2000; Zhao et al., 2001a).
However, ap1 mutants, while defective in sepal and petal …Then Comes B…
development, do not have as clear a homeotic pheno- The picture of initial activation is more complex for B
type as ap2 mutants. Moreover, the homeotic function and C class genes, which require region-specific regula-
of AP1 does not seem to be conserved in Antirrhinum tors for their expression. The investigation of B class
(Theissen et al., 2000). Several other Arabidopsis and genes AP3 and PI as possible LFY targets seemed most
Antirrhinum genes that contribute to A function have promising, as their expression is much more reduced
now been described; they are discussed in more detail in strong lfy mutants than that of the A class gene AP1
in the section on regulation of C function. or the C class gene AG. However, despite this observa-

Cloning of the ABC genes also allowed for validation tion, it is still unclear whether LFY is a direct activator
of the ABC model using gain-of-function experiments of B class genes. The first indication for interaction of
with transgenic plants. With the exception of AP1, ec- LFY with region-specific coregulators in the activation
topic expression of ABC genes leads to the formation of ABC genes came from an analysis of another gene
of flowers that have phenotypes opposite to those ob- required for B class gene expression, UNUSUAL FLO-
served in the respective loss-of-function mutants. For RAL ORGANS (UFO). Unlike LFY, which is expressed
example, constitutive overexpression of both AP3 and throughout the young flower, UFO is expressed tran-
PI leads to the formation of flowers in which the first siently in the flower in a domain similar to that of AP3
whorl is occupied by petals instead of sepals and the and PI (Figure 2). In addition, UFO is expressed in the
fourth whorl carpels are replaced by stamens (Krizek shoot apical meristem in a pattern that mimics that in
and Meyerowitz, 1996). Results from these experiments the floral meristem, being excluded from the center and
not only confirmed the predictions made by the ABC the periphery of the meristem (Lee et al., 1997). The
model concerning organ identity, but also corroborated interaction of UFO and LFY was most strikingly demon-
the idea that regulation of ABC gene activity occurs strated by their ability to activate AP3 and PI outside
mainly at the level of transcription. the flower, when both UFO and LFY are ectopically ex-

pressed (Parcy et al., 1998; Honma and Goto, 2000).
Overall, based on these observations, it seems that re-The ABCs Begin with A…

A question that is central to our understanding of floral gion-specific expression of B class genes results from
the interplay of LFY, which provides floral specificity,patterning is how the pattern of ABC gene expression

is set up. Formally, the formation of individual flowers with UFO, which provides regional specificity within
meristems.is downstream of floral induction, the process that un-

derlies the transition from vegetative to reproductive Despite their strong gain-of-function effects, neither
LFY nor UFO is absolutely required for B class genedevelopment. One of the genes integrating the multiple

endogenous and environmental signals that regulate the expression. A candidate for another, possibly direct,
activator of B class genes is AP1, which functions nottiming of floral induction is the meristem identity gene

LEAFY (LFY), the Arabidopsis ortholog of FLORICAULA only as a homeotic gene, but also as a floral identity
gene. Ectopic AP3 expression has been observed both(FLO) from Antirrhinum (Blázquez and Weigel, 2000).

Expression of ABC genes is much reduced or absent in in plants that express AP1 ectopically and in plants that
express an activated form of AP1, AP1:VP16, in thelfy and flo mutants, in which flowers are replaced by

shoot-like structures, but until recently it was unclear normal AP1 domain (Sessions et al., 2000; Ng and Yanof-
sky, 2001). A direct role of AP1 in regulating AP3 iswhether ABC genes were directly controlled by LFY and

FLO (Weigel and Meyerowitz, 1994; Theissen et al., 2000; further supported by the finding that AP1 binds to the
AP3 promoter and that the binding site is required forZhao et al., 2001a).

Both FLO and LFY are expressed uniformly in young normal activity of this promoter (Hill et al., 1998; Tilly et
al., 1998).floral primordia as soon as these arise. The first hint that

they might be direct regulators of floral homeotic genes In contrast to B class activators LFY and AP1, UFO
is not a DNA binding protein, but belongs to the familycame from the observation that constitutive ectopic ex-

pression of LFY not only causes plants to flower early, of F box proteins, many of which have been shown to
provide substrate specificity to a class of E3 ubiquitinas expected from its role in floral induction, but also

induces ectopic expression of the A class gene AP1 ligases known as SCFs (Samach et al., 1999). UFO inter-
acts both in vitro and in vivo with another common(Parcy et al., 1998). Induction of AP1 by LFY does not

require protein synthesis, as shown with plants that con- SCF subunit, the SKP1 homolog ASK1, supporting the
proposal that UFO acts by controlling the ubiquitinationstitutively express a hormone-regulated version of LFY

(Wagner et al., 1999). Furthermore, fusion of LFY to a of AP3 and PI regulators (Samach et al., 1999; Zhao et
al., 2001b). The most common effect of ubiquitinationheterologous activation domain allows it to activate a

reporter gene that is under the control of AP1 cis-regula- is the targeting of proteins for proteasome-dependent
degradation, and it is conceivable that UFO promotestory sequences in yeast (Parcy et al., 1998), providing

further evidence that the interaction is direct. In wild- degradation of an AP3/PI repressor, but ubiquitination
can also regulate protein activity in other ways (e.g.,type, AP1 is activated shortly after LFY throughout the

emerging floral primordium, in a pattern very similar Kaiser et al., 2000). An answer to the question of how



Developmental Cell
138

with a broader role of the two genes, their expression
is not restricted to the outer whorls of the developing
flower (Jofuku et al., 1994; Conner and Liu, 2000). Both
genes encode apparent transcription factors—AP2 a
member of a plant-specific class of DNA binding pro-
teins and LUG a WD40 repeat protein with similarity to
transcriptional repressors such as Tup1 from yeast or
Groucho from Drosophila. Several regulatory elements
that mediate repression of AG by AP2 and LUG have
been identified (Bomblies et al., 1999; Deyholos and
Sieburth, 2000), but it is not known whether repression
by AP2 and LUG is direct. The same is true for AINTEGU-
MENTA (ANT) and STERILE APETALA (SAP; Table 1),
both of which act redundantly with AP2 in repressing
AG and promoting organ identity in the outer whorls
(Elliott et al., 1996; Klucher et al., 1996; Byzova et al.,
1999; Krizek et al., 2000). Like LUG, ANT and SAP are
expressed outside the flower and have other defects in
addition to those resulting from AG misexpression. The
most notable other role of ANT is in controlling organ
initiation and organ size (Elliott et al., 1996; Klucher et
al., 1996; Krizek, 1999; Mizukami and Fischer, 2000).

Another important negative regulator of AG is CURLY
LEAF (CLF). clf mutant flowers have carpelloid sepals
in the first whorl and staminoid petals in the second
whorl, phenotypes reminiscent of AG derepression
(Goodrich et al., 1997). In addition to ectopic expression
in the flower, AG RNA is expressed widely in vegetative
tissue of clf mutants. This vegetative expression of AG
causes clf mutants to flower early, even though AG nor-
mally has no role in controlling flowering time. CLF itself
is expressed throughout the plant and encodes a Poly-
comb group gene with closest similarity to Enhancer of
zeste from Drosophila. Although Polycomb complexes
have not yet been detected in plants, it is thought that
the CLF product, like its animal counterparts, is involved

Figure 2. Flow Chart of Early Floral Patterning in chromatin remodeling (Goodrich et al., 1997). Like
Upstream regulators LFY, WUS, and UFO are expressed in specific Polycomb group proteins in animals, the primary role of
domains, which, together with repression of AP1 by AG, results in CLF in the flower is maintenance, rather than establish-
the ABC pattern. How the SEP pattern is regulated is not known. ment, of AG repression (Goodrich et al., 1997). Further-
ABC gene products and SEP proteins, all of which are MADS domain

more, there is weak ectopic AP3 expression in clf mu-proteins, assemble into higher order, most likely quaternary, com-
tants, pointing to a more general role of CLF inplexes, which specify different organ identities. It is not known
repressing homeotic genes (Goodrich et al., 1997; Ser-whether AP1 assembles into higher order complexes.
rano-Cartagena et al., 2000).

CLF acts redundantly with INCURVATA2 (ICU2) in re-
pressing AG in both flowers and vegetative tissue. icu2UFO acts may come from the investigation of two genes,
and clf single mutants have similar phenotypes, but dou-CHORIPETALA and DESPENTEADO, which mediate the
ble mutants show a much more severe phenotype, witheffects of the UFO homolog FIMBRIATA in Antirrhinum
carpelloid features on leaves along with ap2-like flowers(Wilkinson et al., 2000).
(Serrano-Cartagena et al., 2000). It will therefore be inter-
esting to learn whether ICU2 also encodes a Polycomb

…Followed by C group protein. Two other genes with more general roles
Arguably the most complete picture of ABC gene regula- in repressing a wide array of developmental regulators,
tion has emerged for the C class gene AG and its Antir- including homeotic genes, are EMBRYONIC FLOWER1
rhinum counterpart PLE. In line with the tenet of the (EMF1) and EMF2 (Aubert et al., 2001; Yoshida et al.,
ABC model that A and C function are mutually inhibitory, 2001). EMF2 is also a member of the Polycomb group
initial studies focused on repression of AG and PLE in genes and encodes a homolog of SU(Z)12 from Dro-
the periphery of the flower. The importance of transcrip- sophila (Yoshida et al., 2001). Yet another repressor of
tional repression was confirmed with the observation AG and AP3 is EARLY BOLTING IN SHORT DAYS (EBS),
that AG RNA expands into the outer whorls of the A but, in contrast to the other genes discussed so far,
class mutants ap2 and leunig (lug). However, AG is not expression of homeotic genes is only increased within
only activated in a larger domain, but also earlier and their normal domains in ebs mutants (Gómez-Mena et
more strongly in these mutants, suggesting that they al., 2001).

Given the large number of pleiotropic loci involvedare not merely region-specific repressors. Consistent



Review
139

in repression of Arabidopsis ABC genes, it is not too AG (Lohmann et al., 2001). Thus, similar to the example
surprising that several Antirrhinum mutations, such as of LFY interacting with UFO to activate AP3 and PI, LFY
stylosa (sty) and fistulata (fis), cause ectopic expression interacts with WUS, which is expressed in a specific pat-
of PLE, along with other complex phenotypes. It is not tern in both shoot and floral meristems, to activate AG.
known whether these loci correspond to any of the Ara-
bidopsis genes described above, but their unique pheno- Refining the Floral ABCs
types suggest that they define a different set of repressors Like other cascades of transcriptional regulation during
(McSteen et al., 1998; Motte et al., 1998). Similarly, the development, fine-tuning and maintenance are impor-
Antirrhinum mutant polypetala, in which PLE as well as tant aspects of ABC gene regulation. An interesting case
DEF expression are reduced, has no obvious counterpart is that of the B class genes AP3 and PI, whose initial
in Arabidopsis (McSteen et al., 1998). expression extends from whorls two and three, where

Because none of the cloned negative regulators of both have a homeotic function, into adjacent whorls,
AG are expressed in a region-specific fashion, it appears with some expression of AP3 in whorl one and of PI in
that AG expression is globally repressed throughout the whorl four. After initial activation, the products of both
plant and that this repression is overcome by region- genes are required to maintain their own expression. At
specific activators in the center of wild-type flowers. As least for AP3, this autoregulation is likely to be direct,
with A and B class genes, the LFY transcription factor as the AP3 promoter contains CArG boxes that are
is an important upstream regulator of AG. The first indi- bound by AP3/PI heterodimers in vitro and that are re-
cation that AG was directly regulated by LFY came from quired for promoter activity in vivo (Riechmann et al.,
analysis of plants carrying an activated form of LFY, 1996; Hill et al., 1998; Tilly et al., 1998). In the case of
LFY:VP16. When expressed in the normal LFY domain, PI, the mechanism of autoregulation is less clear. Even
LFY:VP16 causes phenotypes similar to those of trans- though deletion studies have defined an AP3/PI-respon-
genic or mutant plants with ectopic AG expression. More sive element in the PI promoter, it does not contain a
significantly, expression of LFY:VP16 in vegetative tis- CArG box, nor is it bound by AP3/PI heterodimers
sue is sufficient for AG activation, similar to the activa- (Honma and Goto, 2000). This contrasts with the situa-
tion of AP3 and PI by the combination of LFY and UFO tion in Antirrhinum, where both the DEF and GLO pro-
(Parcy et al., 1998). LFY binds to AG regulatory se- moters contain CArG boxes bound by DEF/GLO hetero-
quences in vitro, and the LFY binding sites are required dimers (Theissen et al., 2000; Zhao et al., 2001a).
for both the normal AG expression pattern and the re- Another level of B class gene regulation is provided
sponse to LFY:VP16 in vivo (Busch et al., 1999), provid- by SUPERMAN (SUP), which is required to maintain the
ing strong evidence that LFY is indeed a direct regulator inner boundary of AP3 expression. SUP itself is under
of AG. control of the floral meristem identity gene LFY, which

Since AG is activated only in a subset of LFY-express- activates SUP through AP3/PI-dependent and -inde-
ing cells, region-specific coregulators must be required pendent pathways (Sakai et al., 2000).
either to repress AG in whorls one and two or to enhance Finally, an important crossregulatory interaction oc-
its activation in whorls three and four. Two recent publi- curs between AP1 and AG. As mentioned before, AP1
cations support the latter idea by showing that the ho- has dual functions—an early role as a floral identity gene
meodomain protein WUSCHEL (WUS) contributes to ac-

and a later role as an A class homeotic gene. These
tivation of AG in the center of flowers (Lenhard et al.,

dual functions are reflected in its expression pattern,
2001; Lohmann et al., 2001; Figure 2). WUS was first

with AP1 initially being expressed throughout the floral
identified because of its role in maintaining a stem cell

primordium and later becoming restricted to presump-
population in the center of shoot apical and floral meri-

tive whorls one and two. Repression of AP1 in the center
stems. Because of their shoot meristem defects, wus

of the flower is AG dependent (Theissen et al., 2000;
mutants rarely make flowers, but the occasional flowers

Zhao et al., 2001a), although it remains to be seen
that are formed mostly lack stamens and carpels, the

whether this is a direct effect of AG. Crossregulation oforgans specified by AG. WUS is activated before AG in
AP1 by the C class gene AG, conforming to the thirdflowers and its RNA accumulates in a domain that is
tenet of the ABC model that C class activity represseseventually included in the AG expression domain (Mayer
A class activity, provides an economical way of estab-et al., 1998). Although wus mutants can make a few
lishing the ABC pattern, as independent region-specificstamens, WUS is required for normal AG activation, as
regulators are only required for AG.plants with reduced WUS expression also have a re-

duced AG expression domain (Lohmann et al., 2001).
Beyond the ABCsConversely, ectopic WUS expression leads to ectopic
One of the most satisfying findings of early experimentsactivation of AG, demonstrating that WUS is also suffi-
with floral homeotic mutants was that plants lacking allcient to drive AG expression in flowers (Lenhard et al.,
three classes of ABC gene activities formed flowers that2001; Lohmann et al., 2001). WUS binds to sites adjacent
had only leaf-like organs (Bowman et al., 1991), confirm-to the LFY binding sites in the AG enhancer, and both
ing Goethe’s (1790) assertion made two centuries earlieract together to activate transcription from AG regulatory
that floral organs are modified leaves. It was disappoint-sequences in a yeast transactivation assay (Lohmann
ing, therefore, that overexpression of ABC genes, aloneet al., 2001). Since LFY and WUS can bind DNA indepen-
or in combination, failed to convert leaves into floraldently, activation is likely due to synergistic effects on
organs. Only recently has the missing piece of the puzzlethe basal transcription machinery. Mutating the WUS
been found. It turns out that at least B and C class genesbinding sites strongly reduces the activity of the AG

enhancer, confirming that WUS is a direct activator of cannot function without a trio of MADS box genes, the



Developmental Cell
140

SEPALLATA genes, whose combined knockout pheno- 2001) as well as the KNAT2 homeobox gene (Pautot et
type resembles that of plants without B and C function al., 2001; Table 1).
(Pelaz et al., 2000). Conversely, overexpressing SEP
genes in combination with ABC genes leads to spectac- Summary
ular transformation of vegetative leaves into floral or- The regulatory system governing early floral patterning
gans (Honma and Goto, 2001; Pelaz et al., 2001). The is well conserved in the two reference plants Arabi-
molecular basis of these effects is that that ABC gene dopsis and Antirrhinum, which represent the two major
products form higher order complexes with SEP pro- subdivisions of higher dicots. Consistent with the many
teins, which provide activation domains for those MADS similarities between Arabidopsis and Antirrhinum, the
domain proteins that cannot activate transcription on role of ABC genes is largely conserved in other dicots
their own (Honma and Goto, 2001; Figure 2). A second as well, and even in monocots such as grasses (e.g.,
way in which formation of higher order complexes may Ambrose et al., 2000; Ma and dePamphilis, 2000). This
contribute to synergistic effects on the regulation of observation not withstanding, there are variations in the
target genes is by increasing DNA binding affinity (Egea- manner in which B function genes contribute to the
Cortines et al., 1999). development of petals and stamens, as deduced from

Having found conditions in which ABC genes can in- recent work on basal dicots (Kramer and Irish, 1999).
duce floral organ fate throughout the plant should Other differences in the regulatory systems are due to
greatly facilitate the identification of their target genes. gene duplication and loss, which has resulted in various
So far, little is known about such target genes, not very degrees of redundancy and subfunctionalization. Exam-
different from the situation for many developmental reg- ples are the multiple AG orthologs in Antirrhinum, petu-
ulators in animals (Pradel and White, 1998). One of the nia and cucumber, which differ in their ability to induce
most promising reports for Arabidopsis has been the reproductive organ fate (Tsuchimoto et al., 1993; Kater et
one from Sablowski and Meyerowitz (1998), who used a al., 1998; Davies et al., 1999), or the second whorl-specific
hormone-dependent version of AP3 to search for direct phenotype of a mutation in the petunia B class gene
target genes. Subsequent analysis of the NAP gene, green petals (gp; van der Krol et al., 1993). A more signifi-
which was identified with this method, revealed why cant discrepancy is that there is no evidence for AP2
the power of genetics is limited when it comes to a orthologs controlling C class activity in other species
comprehensive picture of homeotic target genes: NAP (Maes et al., 2001). Thus, AP2 may have acquired its
expression is not confined to petals and stamens, where role in AG regulation relatively recently during the evolu-
AP3 is active, and modulating NAP activity in vivo has tion of Arabidopsis.
complex effects that do not obviously hint to a role of Although there has been significant progress in under-
NAP in mediating AP3 activity. standing the mechanisms of floral patterning, there are

There are, however, some target genes that have or- still many outstanding issues. The most significant is
gan-specific effects and that have been identified by probably how the prepattern, which results in region-
genetic analyses. One example is that of the SHAT-

specific expression of homeotic activators such as UFO
TERPROOF (SHP) genes, which are regulated by AG,

and WUS, is generated. The answer to this question will
and which in turn control region-specific patterning

hopefully come from the rich body of work that deals
within the carpel, an AG-dependent organ (Liljegren et

with the origin, structure, and function of shoot meri-
al., 2000). The SHP genes are closely related to AG, and

stems (Brand et al., 2001). Downstream of the homeotic
it will be interesting to learn whether the carpel-specific

genes, it seems likely that systematic global expression
patterning function of the SHP genes originated only

profiling will enable comprehensive identification of tar-
after the duplication event that gave rise to AG and SHP

get genes. For both the upstream and downstream
genes, or whether there was an ancestral version of AG

events, the major challenge remaining will be to decipherthat controlled all these functions.
the logic of regulatory interactions that underlie the for-Interestingly, in addition to its early function in speci-
mation of flowers.fying carpel identity, AG itself is required for the pat-

terning of specific carpel structures. Although ag single
Acknowledgmentsmutants lack carpels, because of expansion of A func-

tion into the center of the flower, removing A function
We thank José Dinneny, Julin Maloof, Javier Palatnik, Norman

in ap2 ag double mutants leads to the formation of car-
Warthmann, Phil Wigge, and Xuelin Wu for reading of the manuscript

pelloid leaves in these flowers. The fact that these or- and discussion. We apologize to those whose original research we
gans do not have the full inventory of pattern elements did not cite for space constraints. Our work on floral patterning is
found in normal carpels indicates both that AG is re- supported by a fellowship from the HFSPO (J.U.L.), grants from the

NIH (R01 GM62932), NSF (IBN-0078273), and U.S. Department ofquired for patterning within the carpel and that other
Energy (DE-FG03-98ER20317.), and by the Max Planck Institute, ofgenes must act in parallel with AG in this process. Two
which D.W. is a Director.genes that have such functions are CRABS CLAW (CRC)

and SPATULA (SPT; Table 1), and, consistent with ge-
Referencesnetic studies, activation of CRC and SPT is at least

partially independent of AG (Bowman and Smyth, 1999;
Ambrose, B.A., Lerner, D.R., Ciceri, P., Padilla, C.M., Yanofsky, M.F.,

Heisler et al., 2001). Carpel patterning also involves fac- and Schmidt, R.J. (2000). Molecular and genetic analyses of the
tors that do not have necessarily carpel-specific effects silky1 gene reveal conservation in floral organ specification between
(e.g., Sessions et al., 1997). Other factors that act in eudicots and monocots. Mol. Cell 5, 569–579.
parallel with AG and contribute to C function are the Aubert, D., Chen, L., Moon, Y.H., Martin, D., Castle, L.A., Yang, C.H.,

and Sung, Z.R. (2001). EMF1, a novel protein involved in the controlHUA1 and HUA2 (Chen and Meyerowitz, 1999; Li et al.,



Review
141

of shoot architecture and flowering in Arabidopsis. Plant Cell 13, Honma, T., and Goto, K. (2001). Complexes of MADS-box proteins
are sufficient to convert leaves into floral organs. Nature 409,1865–1875.
525–529.Blázquez, M.A., and Weigel, D. (2000). Integration of floral inductive

signals in Arabidopsis. Nature 404, 889–892. Jofuku, K.D., den Boer, B.G.W., Van Montagu, M., and Okamuro,
J.K. (1994). Control of Arabidopsis flower and seed development byBomblies, K., Dagenais, N., and Weigel, D. (1999). Redundant
the homeotic gene APETALA2. Plant Cell 6, 1211–1225.enhancers mediate transcriptional repression of AGAMOUS by

APETALA2. Dev. Biol. 216, 260–264. Kaiser, P., Flick, K., Wittenberg, C., and Reed, S.I. (2000). Regulation
of transcription by ubiquitination without proteolysis: Cdc34/Bowman, J.L., and Smyth, D.R. (1999). CRABS CLAW, a gene that
SCFMet30-mediated inactivation of the transcription factor Met4. Cellregulates carpel and nectary development in Arabidopsis, encodes
102, 303–314.a novel protein with zinc finger and helix-loop-helix domains. Devel-

opment 126, 2387–2396. Kater, M.M., Colombo, L., Franken, J., Busscher, M., Masiero, S.,
Van Lookeren Campagne, M.M., and Angenent, G.C. (1998). MultipleBowman, J.L., Smyth, D.R., and Meyerowitz, E.M. (1991). Genetic
AGAMOUS homologs from cucumber and petunia differ in theirinteractions among floral homeotic genes of Arabidopsis. Develop-
ability to induce reproductive organ fate. Plant Cell 10, 171–182.ment 112, 1–20.
Klucher, K.M., Chow, H., Reiser, L., and Fischer, R.L. (1996). TheBrand, U., Hobe, M., and Simon, R. (2001). Functional domains in
AINTEGUMENTA gene of Arabidopsis required for ovule and femaleplant shoot meristems. Bioessays 23, 134–141.
gametophyte development is related to the floral homeotic geneBusch, M.A., Bomblies, K., and Weigel, D. (1999). Activation of a
AP2. Plant Cell 8, 137–153.floral homeotic gene in Arabidopsis. Science 285, 585–587.
Kramer, E.M., and Irish, V.F. (1999). Evolution of genetic mechanismsByzova, M.V., Franken, J., Aarts, M.G., de Almeida-Engler, J., Engler,
controlling petal development. Nature 399, 144–148.G., Mariani, C., Campagne, M.M., and Angenent, G.C. (1999). Arabi-
Krizek, B.A. (1999). Ectopic expression of AINTEGUMENTA in Arabi-dopsis STERILE APETALA, a multifunctional gene regulating inflo-
dopsis plants results in increased growth of floral organs. Dev.rescence, flower, and ovule development. Genes Dev. 13, 1002–
Genet. 25, 224–236.1014.

Chen, X., and Meyerowitz, E.M. (1999). HUA1 and HUA2 are two Krizek, B.A., and Meyerowitz, E.M. (1996). The Arabidopsis homeotic
genes APETALA3 and PISTILLATA are sufficient to provide the Bmembers of the floral homeotic AGAMOUS pathway. Mol. Cell 3,

349–360. class organ identity function. Development 122, 11–22.

Coen, E.S., and Meyerowitz, E.M. (1991). The war of the whorls: Krizek, B.A., Prost, V., and Macias, A. (2000). AINTEGUMENTA pro-
genetic interactions controlling flower development. Nature 353, motes petal identity and acts as a negative regulator of AGAMOUS.
31–37. Plant Cell 12, 1357–1366.

Conner, J., and Liu, Z. (2000). LEUNIG, a putative transcriptional Lee, I., Wolfe, D.S., Nilsson, O., and Weigel, D. (1997). A LEAFY co-
corepressor that regulates AGAMOUS expression during flower de- regulator encoded by UNUSUAL FLORAL ORGANS. Curr. Biol. 7,
velopment. Proc. Natl. Acad. Sci. USA 97, 12902–12907. 95–104.

Davies, B., Motte, P., Keck, E., Saedler, H., Sommer, H., and Lenhard, M., Bohnert, A., Jürgens, G., and Laux, T. (2001). Termina-
Schwarz-Sommer, Z. (1999). PLENA and FARINELLI: redundancy tion of stem cell maintenance in Arabidopsis floral meristems by
and regulatory interactions between two Antirrhinum MADS-box interactions between WUSCHEL and AGAMOUS. Cell 105, 805–814.
factors controlling flower development. EMBO J. 18, 4023–4034. Li, J., Jia, D., and Chen, X. (2001). HUA1, a regulator of stamen and
Deyholos, M.K., and Sieburth, L.E. (2000). Separable whorl-specific carpel identities in Arabidopsis, codes for a nuclear RNA binding
expression and negative regulation by enhancer elements within protein. Plant Cell 13, 2269–2281.
the AGAMOUS second intron. Plant Cell 12, 1799–1810. Liljegren, S.J., Gustafson-Brown, C., Pinyopich, A., Ditta, G.S., and
Egea-Cortines, M., Saedler, H., and Sommer, H. (1999). Ternary Yanofsky, M.F. (1999). Interactions among APETALA1, LEAFY, and
complex formation between the MADS-box proteins SQUAMOSA, TERMINAL FLOWER1 specify meristem fate. Plant Cell 11, 1007–
DEFICIENS and GLOBOSA is involved in the control of floral archi- 1018.
tecture in Antirrhinum majus. EMBO J. 18, 5370–5379.

Liljegren, S.J., Ditta, G.S., Eshed, Y., Savidge, B., Bowman, J.L., and
Elliott, R.C., Betzner, A.S., Huttner, E., Oakes, M.P., Tucker, W.Q.J., Yanofsky, M.F. (2000). SHATTERPROOF MADS-box genes control
Gerentes, D., Perez, P., and Smyth, D.R. (1996). AINTEGUMENTA, seed dispersal in Arabidopsis. Nature 404, 766–770.
an APETALA2-like gene of Arabidopsis with pleiotropic roles in ovule

Lohmann, J.U., Hong, R., Hobe, M., Busch, M.A., Parcy, F., Simon,
development and floral organ growth. Plant Cell 8, 155–168.

R., and Weigel, D. (2001). A molecular link between stem cell regula-
Goethe, J.W.v. (1790). Versuch die Metamorphose der Pflanzen zu tion and floral patterning in Arabidopsis. Cell 105, 793–803.
erklären (Gotha, Germany: C.W. Ettinger).

Ma, H., and dePamphilis, C. (2000). The ABCs of floral evolution.
Gómez-Mena, C., Piñeiro, M., Franco-Zorrilla, J.M., Salinas, J., Cell 101, 5–8.
Coupland, G., and Martı́nez-Zapater, J.M. (2001). early bolting in

Maes, T., Van De Steene, N., Zethof, J., Karimi, M., D’Hauw, M.,
short days: an Arabidopsis mutation that causes early flowering and

Mares, G., Van Montagu, M., and Gerats, T. (2001). Petunia AP2-
partially suppresses the floral phenotype of leafy. Plant Cell 13,

like genes and their role in flower and seed development. Plant Cell
1011–1024.

13, 229–244.
Goodrich, J., Puangsomlee, P., Martin, M., Long, D., Meyerowitz,

Mayer, K.F.X., Schoof, H., Haecker, A., Lenhard, M., Jürgens, G.,
E.M., and Coupland, G. (1997). A Polycomb-group gene regulates

and Laux, T. (1998). Role of WUSCHEL in regulating stem cell fate
homeotic gene expression in Arabidopsis. Nature 386, 44–48.

in the Arabidopsis shoot meristem. Cell 95, 805–815.
Heisler, M.G., Atkinson, A., Bylstra, Y.H., Walsh, R., and Smyth, D.R.

McSteen, P.C., Vincent, C.A., Doyle, S., Carpenter, R., and Coen,(2001). SPATULA, a gene that controls development of carpel margin
E.S. (1998). Control of floral homeotic gene expression and organtissues in Arabidopsis, encodes a bHLH protein. Development 128,
morphogenesis in Antirrhinum. Development 125, 2359–2369.1089–1098.
Mizukami, Y., and Fischer, R.L. (2000). Plant organ size control:Hill, T.A., Day, C.D., Zondlo, S.C., Thackeray, A.G., and Irish, V.F.
AINTEGUMENTA regulates growth and cell numbers during organo-(1998). Discrete spatial and temporal cis-acting elements regulate
genesis. Proc. Natl. Acad. Sci. USA 97, 942–947.transcription of the Arabidopsis floral homeotic gene APETALA3.
Motte, P., Saedler, H., and Schwarz-Sommer, Z. (1998). STYLOSADevelopment 125, 1711–1721.
and FISTULATA: regulatory components of the homeotic control ofHonma, T., and Goto, K. (2000). The Arabidopsis floral homeotic
Antirrhinum floral organogenesis. Development 125, 71–84.gene PISTILLATA is regulated by discrete cis-elements responsive

to induction and maintenance signals. Development 127, 2021– Ng, M., and Yanofsky, M.F. (2001). Activation of the Arabidopsis B
class homeotic genes by APETALA1. Plant Cell 13, 739–754.2030.



Developmental Cell
142

Parcy, F., Nilsson, O., Busch, M.A., Lee, I., and Weigel, D. (1998). A and Development, M.T. McManus and B. Veit, eds. (Sheffield: Shef-
field Academic Press), pp. 89–142.genetic framework for floral patterning. Nature 395, 561–566.

Zhao, D., Yu, Q., Chen, M., and Ma, H. (2001b). The ASK1 genePautot, V., Dockx, J., Hamant, O., Kronenberger, J., Grandjean, O.,
regulates B function gene expression in cooperation with UFO andJublot, D., and Traas, J. (2001). KNAT2: evidence for a link between
LEAFY in Arabidopsis. Development 128, 2735–2746.Knotted-like genes and carpel development. Plant Cell 13, 1719–

1734.

Pelaz, S., Ditta, G.S., Baumann, E., Wisman, E., and Yanofsky, M.F.
(2000). B and C floral organ identity functions require SEPALLATA
MADS-box genes. Nature 405, 200–203.

Pelaz, S., Tapia-Lopez, R., Alvarez-Buylla, E.R., and Yanofsky, M.F.
(2001). Conversion of leaves into petals in Arabidopsis. Curr. Biol.
11, 182–184.

Pradel, J., and White, R.A. (1998). From selectors to realizators. Int.
J. Dev. Biol. 42, 417–421.

Riechmann, J.L., Krizek, B.A., and Meyerowitz, E.M. (1996). Dimer-
ization specificity of Arabidopsis MADS domain homeotic proteins
APETALA1, APETALA3, PISTILLATA, and AGAMOUS. Proc. Natl.
Acad. Sci. USA 93, 4793–4798.

Sablowski, R.W., and Meyerowitz, E.M. (1998). A homolog of NO
APICAL MERISTEM is an immediate target of the floral homeotic
genes APETALA3/PISTILLATA. Cell 92, 93–103.

Sakai, H., Krizek, B.A., Jacobsen, S.E., and Meyerowitz, E.M. (2000).
Regulation of SUP expression identifies multiple regulators involved
in Arabidopsis floral meristem development. Plant Cell 12, 1607–
1618.

Samach, A., Klenz, J.E., Kohalmi, S.E., Risseeuw, E., Haughn, G.W.,
and Crosby, W.L. (1999). The UNUSUAL FLORAL ORGANS gene of
Arabidopsis thaliana is an F-box protein required for normal pat-
terning and growth in the floral meristem. Plant J. 20, 433–445.

Serrano-Cartagena, J., Candela, H., Robles, P., Ponce, M.R., Pérez-
Pérez, J.M., Piqueras, P., and Micol, J.L. (2000). Genetic analysis
of incurvata mutants reveals three independent genetic operations
at work in Arabidopsis leaf morphogenesis. Genetics 156, 1363–
1377.

Sessions, A., Nemhauser, J., McColl, A., Roe, J., Feldmann, K., and
Zambryski, P. (1997). ETTIN patterns the Arabidopsis floral meristem
and reproductive organs. Development 124, 4481–4491.

Sessions, A., Yanofsky, M.F., and Weigel, D. (2000). Cell-cell signal-
ing and movement by the floral transcription factors LEAFY and
APETALA1. Science 289, 779–782.

Theissen, G., Becker, A., Di Rosa, A., Kanno, A., Kim, J.T., Münster,
T., Winter, K.U., and Saedler, H. (2000). A short history of MADS-
box genes in plants. Plant Mol. Biol. 42, 115–149.

Tilly, J.J., Allen, D.W., and Jack, T. (1998). The CArG boxes in the
promoter of the Arabidopsis floral organ identity gene APETALA3
mediate diverse regulatory effects. Development 125, 1647–1657.

Tsuchimoto, S., van der Krol, A.R., and Chua, N.-H. (1993). Ectopic
expression of pMADS3 in transgenic petunia phenocopies the petu-
nia blind mutant. Plant Cell 5, 843–853.

van der Krol, A.R., Brunelle, A., Tsuchimoto, S., and Chua, N.-H.
(1993). Functional analysis of petunia floral homeotic MADS box
gene pMADS1. Genes Dev. 7, 1214–1228.

Wagner, D., Sablowski, R.W.M., and Meyerowitz, E.M. (1999). Tran-
scriptional activation of APETALA1 by LEAFY. Science 285,
582–584.

Weigel, D., and Meyerowitz, E.M. (1994). The ABCs of floral homeotic
genes. Cell 78, 203–209.

Wilkinson, M., de Andrade Silva, E., Zachgo, S., Saedler, H., and
Schwarz-Sommer, Z. (2000). CHORIPETALA and DESPENTEADO:
general regulators during plant development and potential floral
targets of FIMBRIATA-mediated degradation. Development 127,
3725–3734.

Yoshida, N., Yanai, Y., Chen, L., Kato, Y., Hiratsuka, J., Miwa, T.,
Sung, Z.R., and Takahashi, S. (2001). EMBRYONIC FLOWER2, a
novel Polycomb group protein homolog, mediates shoot develop-
ment and flowering in Arabidopsis. Plant Cell 13, 2471–2481.

Zhao, D., Yu, Q., Chen, C., and Ma, H. (2001a). Genetic control of
reproductive meristems. In Meristematic Tissues in Plant Growth