Review articles
Drawing lines and borders:
how the dehiscent fruit of
Arabidopsis is patterned
José R. Dinneny and Martin F. Yanofsky*
Summary
The advent of fruits marked a key innovation in the
evolution of flowering plants and helped generate a
diverse array of mechanisms for seed dispersal. In the
model plant, Arabidopsis thaliana, seed dispersal occurs
through a process known as ‘‘pod-shatter’’ in which the
fruit structure falls to pieces upon light mechanical
pressures. This dispersal mechanism is dependent on
the careful patterning of tissues in the fruit, which
perform diverse functions that enable the fruit to open
at maturation. Using the genetic power of Arabidopsis,
many of the molecular components that help specify
these tissues have been identified. Studies of the
interactions among these genes have revealed a regula-
tory network that limits processes such as cell–cell
separation and lignification to discreet regions of the
fruit. Knowledge of these processes in a model fruit
creates a foundation on which to build an understanding
of the evolution of fruit form in other species and provides
tools to engineer shatter-resistant seed pods to prevent
crop loss in plants of agronomic importance such as
canola. BioEssays 27:42–49, 2005.
ß 2004 Wiley Periodicals, Inc.
Introduction
Fruits can develop into exquisite structures that sense and
respond to interactions with animals or the elements, which act
as triggers for the release of seeds. This added level of
Division of Biological Sciences, University of California San Diego, CA
*Correspondence to: Martin F. Yanofsky, Division of Biological
Sciences, University of California San Diego, La Jolla, CA 92093-
0116. E-mail: Marty@ucsd.edu
DOI 10.1002/bies.20165
Published online in Wiley InterScience (www.interscience.wiley.com).
Abbreviations: GFP, Green Fluorescent Protein; NPA, N-1-
naphthylphthalamic acid; ARF, Auxin Response Factor; ETT, ETTIN ;
SPT, SPATULA; STY1,2, STYLISH 1, STYLISH 2; SHI, SHORT
INTERNODES; GA, gibberellic-acid; KAN1,2, KANADI 1, KANADI 2 ;
CRC, CRABS CLAW ; ANT, AINTEGUMENTA; LUG, LEUNIG; enb,
endocarp layer b; ena, endocarp layer a; FUL, FRUITFULL; SHP1,2,
SHATTERPROOF 1, SHATTERPROOF 2; ALC, ALCATRAZ; IND,
INDEHISCENT; RPL, REPLUMLESS; GUS, beta-glucuronidase,
QTL, Quantitative Trait Locus.
42 BioEssays 27.1
regulation controlling seed dispersal has helped plants acquire
a new level of mobility, despite their sedentary lifestyle. The
mechanisms by which fruits act to disperse seeds has been
well described.(1,2) Fruit-mediated seed-dispersal mechan-
isms can be roughly divided into animal-assisted and plant-
assisted strategies. Fleshy or sticky fruits, or those that
develop Velcro-like burrs, use animals to carry seeds away
from the mother plant. Many plants have also developed
mechanisms to self-disperse seeds by highly modifying their
fruit structure. The fruits of Acer and Fraxinus, for example,
develop extended wing-like structures that allow seeds to glide
away. Certain plants have also developed fruits that forcibly
eject seeds using mechanical forces that accumulate during
maturation. The squirting cucumber, Ecballium elaterium,
builds up high amounts of pressure in its fruit, which is filled
with seeds and a mucilaginous material. When disturbed, the
fruit detaches from the stem and squirts a fast jet of seeds.
Fruits of Impatiens, or touch-me-nots, are so named because
of the explosive manner in which they release their seeds
when touched. The model plant, Arabidopsis thaliana, has
also developed a mechanism to forcibly open its seed pod, or
silique, using mechanical forces that build up as the fruit dries
out in a process known as dehiscence.(3) While not as
explosive as dehiscence in Impatiens, this process effectively
discloses the mature seeds, which can then be released and
scattered by rain or wind.
The use of Arabidopsis has facilitated the cloning and
characterization of genes involved in patterning the fruit.
Members of the Brassicaceae family, including Arabidopsis,
develop solitary fruits that form from the fusion of two modified
leaves, termed carpels.(2,4–7) In Arabidopsis, the carpel pri-
mordia emerge from the central region of a flower primordium
as a congenitally fused cylinder of tissue that forms the
gynoecium, the pre-fertilization structure that develops into the
fruit. Unlike mature fruit, many tissues in the gynoecium are in
a quasi-undifferentiated state. Tissues required for fertilization
are completely differentiated in a mature flower when anthers
shed their pollen onto the stigma, whereas tissues important
for fruit dehiscence require signals after fertilization to
complete their development.(8)
This review will discuss recent work in Arabidopsis on
events that pattern the overall architecture of the gynoecium
BioEssays 27:42–49, ß 2004 Wiley Periodicals, Inc.
 Review articles

and the dehiscence zone in the mature fruit. We will also

discuss possible extensions of this work into the evolution of
fruit form in the Brassicaceae and suggest applications for
information garnered from studying fruit dehiscence to
biotechnology-facilitated crop improvement.
Prelude to the fruit: patterning the gynoecium

The apical–basal axis

The distribution of tissues in the fruit is a direct consequence of
patterning events before fertilization that divide the gynoecium
into distinct regions. Along the apical–basal axis, the
gynoecium is divided into apical, medial and basal regions
(Figs. 1A, 2).(9) At the apex of the gynoecium, stigmatic tissue
develops that differentiates papillary cells on which the pollen
adheres and germinates. Connecting the stigma to the ovary is
the style. Inside the style, a specialized tissue termed the
transmitting tract develops. Cells in the transmitting tract
exude sugars, proteins and signals that feed and guide the
rapidly growing pollen tube to the ovules, which will later form
seeds after fertilization.(10) The ovary comprises the medial
region and houses the ovules. At the base of the ovary is a
short stalk, termed the gynophore, that attaches the ovary to
the flower.
The plant hormone, auxin, appears to play an important role
in the regional specification of tissues in the apical–basal axis
of the gynoecium.(11) Nemhauser et al. proposed a model for
auxin-directed tissue specification whereby a gradient of auxin
exists during gynoecium development, with the highest levels
at the apical end and the lowest levels at the basal end
(Fig. 2A). Peak levels of auxin at the apex, the presumed
source of auxin, are proposed to promote stigma and style
development, while moderate levels in the medial region
promote ovary development and the lowest levels in the basal Figure 1. Diagramatic view of the gynoecium and mature fruit
region promote gynophore formation. Although direct visua- of Arabidopsis. A: Exterior structures of a prefertilization-stage
lization of the auxin gradient has not been performed, a GFP gynoecium. B: Transverse cross section of the ovary region of
(Green Florescent Protein)-marker line, which has been used gynoecium. C: Transverse cross section of a mature fruit.
D: Transverse cross-section of the replum/valve region of a
to report peak levels of auxin response, is expressed in the
mature fruit treated with the lignin-specific stain, phloroglucinol,
apical cells of the gynoecium.(12) The transport of auxin by which reveals the lignified layer of the valve margin. E: Staining
efflux carriers seems to be important in establishing and/ with alcian blue and saffranin O reveals the light-blue colored
or maintaining this gradient as flowers treated with the separation layer on the replum side of the valve margin. Notice
polar-auxin-transport inhibitor, N-1-naphthylphthalamic acid that the valve on the left side of the replum has already
separated from the replum at the separation layer. V, valve; R,
(NPA),(13) develop gynoecia with reduced ovary development
replum.
and have enlarged gynophores and styles (Fig. 2B).(11,14)
Based on these patterning defects, Nemhauser et al. pro-
posed that the suppression of polar-auxin transport by NPA
causes the pooling of auxin in the apical region and a sub- Nemhauser et al. to propose that ETT may act to mediate a
sequent depletion of auxin in the basal region. This change in moderate response to auxin in the ovary and broaden the size
auxin distribution alters the size of the three domains in the of the medial region in the gynoecium.
gynoecium, with an increase in the apical and basal regions Formation of the apical region of the gynoecium is largely
and a decrease in the size of the medial region. Similar effects dependent on SPATULA, a member of the bHLH (basic-helix–
are seen when the Auxin Response Factor (ARF), ETTIN loop–helix) family of transcription factors that is expressed in
(ETT), is mutated (Table 1).(9) In addition, weak ett alleles are apical tissues including the stigma, style and transmitting tract
enhanced by NPA treatment.(11) Together, these data led and is necessary for their development.(15,16) ETT ’s role in

BioEssays 27.1 43
Review articles

Figure 2. Models depicting the proposed distribution of auxin and its role in pattering the gynoecium. A: Nemhauser et al. proposed that
auxin acts as a morphogen, controlling the distribution of tissues in the gynoecium in a concentration-dependent manner. Auxin levels,
depicted in red, are proposed to be highest near the apical region of the gynoecium and promote stigma and style development. Moderate
concentrations of auxin promote ovary development including the formation of the valves and ovules (not diagramed), whereas, low levels
of auxin promote gynophore development. B: NPA treatment inhibits polar auxin transport and is proposed to result in the pooling of auxin in
the apical region of the gynoecium, the presumed source of auxin. NPA treatment would also reduce the breadth of the auxin gradient,
reducing the size of regions of the gynoecium exposed to moderate levels of auxin and increasing the size of regions exposed to low levels of
auxin. This alteration in the auxin gradient would then result in a change in the distribution of tissues in the gynoecium, with an increase in
apical and basal tissues and a reduction in medial tissues.

promoting the formation of the medial region is, to a large
extent, mediated through the repression of SPT expression in
the ovary as loss-of-function mutations in SPT suppress the
ectopic development of apical tissues observed in ett
mutants.(16,17) Whereas NPA treatment enhances the ett
mutant phenotype, treatment of the spt gynoecium with NPA
can actually rescue some defects in style growth seen in spt
mutants.(11) These results led Nemhauser et al. and Heisler
et al. to propose that SPT may promote auxin signaling,
enabling a peak auxin response in the apical region of the
gynoecium necessary for the formation of style, stigma and
transmitting tract. Thus, treatment of spt gynoecia with NPA,
which presumably causes the pooling of auxin in apical cells,
would rescue the style defects seen in spt mutants by

Table 1. Summary of genes controlling gynoecium and fruit development in Arabidopsis

Gene name Transcription factor classification Role in gynoecium or fruit development

ETTIN (ETT) Auxin Response Factor (ARF) Promotes ovary formation by inhibiting SPT expression
SPATULA (SPT) basic Helix-Loop-Helix (bHLH) Promotes apical-tissue development including the style, stigma and
transmitting-tract
STYLISH 1,2 (STY1,2) SHI-type Zinc-finger Promotes style development
KANADI 1,2 (KAN 1,2) GARP-type Represses expansion of internal-medial tissue development into the
lateral regions of the gynoecium
CRABSCLAW (CRC) YABBY-type zinc-finger with HMG motif Works with KAN genes to repress expansion of internal-medial tissue
development into the lateral regions of gynoecium
AINTEGUMENTA (ANT) AP2-type Promotes the growth of medial tissues with LUG
LEUNIG (LUG) Tup1 corepressor Promotes the growth of medial tissues with ANT
FRUITFULL (FUL) MADS-box Represses expression of the valve-margin identity genes in the valves;
Promotes the lignification of the enb layer
SHATTERPROOF 1,2 (SHP 1,2) MADS-box Promotes valve-margin development partly through activation of IND
and ALC expression; Promotes the lignification of the enb layer
ALCATRAZ (ALC) basic Helix-Loop-Helix (bHLH) Controls the development of the valve-margin separation layer;
Promotes the lignification of the enb layer
INDEHISCENT (IND) atypical basic Helix-Loop-Helix (bHLH) Controls the development of the valve-margin separation layer and
lignified layer; Promotes the lignification of the enb layer
REPLUMLESS (RPL) BEL-subfamily homeodomain Represses expression of valve-margin identity genes in the replum

44 BioEssays 27.1

overriding deficits in auxin signaling resulting from the loss-of-
SPT activity. Alternatively, we would like to propose that
SPT may itself play a role in establishing high auxin levels in
apical tissues by inhibiting polar auxin transport. This hypoth-
esis is consistent with SPT ’s role in apical tissue development
and fits well with the observation that NPA can rescue spt
defects. In the future, it will be important to develop techniques
to visualize the localization and concentration of auxin during
gynoecium development to verify the auxin-gradient model.
More recently, another set of factors that promote style
development have been identified, STYLISH 1 and 2, which
encode SHI-like zinc-finger transcription factors.(18) Loss of
STY1,2 function results in a partial loss of style and stigma
tissues. These phenotypes are enhanced by the spt mutation.
Conversely, overexpression of STY1 causes the ectopic deve-
lopment of style-like epidermal cells and the accumulation of
vascular tissue patches along the valves, a characteristic nor-
mally found in the style. STY1,2 also have roles in leaf deve-
lopment where loss of STY1,2 function results in serrated leaves
while plants that overexpress STY1 develop ectopic leaf blade
tissue along their petioles (leaf stalk). Whether these functions in
regulating leaf growth are related to the growth defects in style
development remains to be determined. SHORT INTER-
NODES, the founding member of the SHI family of transcription
factors, was first identified through its ability to suppress
response to the plant hormone, gibberellic-acid (GA).(19) While
GA has been implicated in postfertilization development,(8,20) GA
regulation is not known to affect gynoecium patterning, thus it will
be interesting to see if STY1,2 have similar activities.
The mediolateral axis
By taking a cross section through the gynoecium, one can see
the bilateral symmetry of tissues resulting from the fusion of
the two carpels (Fig. 1B). This bilateral symmetry becomes
apparent very early during gynoecium development when
medial tissues begin to rapidly divide forming ridges at the
point where the carpels are fused together.(5,7) These medial
ridges eventually grow out and fuse together, post-genitally, to
form the septum (technically, a false-septum in Brassica-
ceae).(2) Prior to fertilization, tissues in the septum will
differentiate transmitting tract tissue to guide pollen tubes
from the style to the ovules. Simultaneously during septum
fusion, tissues flanking the septum, termed the placenta,
begin to initiate small radial outgrowths that will develop into
ovules.(21) The valves form from the lateral carpel walls and
enclose the ovules within the ovary. Valve tissues are
separated medially, at the point of fusion, by a structure
termed the replum. While not easily identifiable morphologi-
cally, the valve-margin tissues develop between the replum
and the valves and can be distinguished from the valves by
molecular markers before fertilization.(22)
A number of factors important for the formation of lateral
organs were first identified by their role in gynoecium
Review articles
development. This observation suggests that the gynoecium
may function something like a ‘‘canary in a coalmine’’ whereby
the gynoecium may be especially sensitive to minor defects in
growth and patterning that do not strongly affect other organs.
Members of the KANADI (KAN) and YABBY families of
transcription factors regulate the polar localization of tissues in
lateral organs.(23–25) All lateral organs in Arabidopsis, includ-
ing carpels, develop with an inherent polarity. The interior side
of the lateral organ that is closest to the shoot meristem is
termed the adaxial side, while the side opposite the meristem
is termed the abaxial side. In leaves, these transcription
factors regulate the distribution of the various tissue types that
distinguish the abaxial and adaxial sides. In the gynoecium,
these functions translate into a control of the polarity of the
medial region as well as the distribution of tissues along the
mediolateral axis. Loss-of-KAN1 function results in only minor
defects in gynoecium patterning including the occasional
development of ectopic style and ovules in place of the
replum.(25,26) This phenotype is further enhanced when com-
bined with mutations in CRABS CLAW (CRC), the founding
member of the YABBY family.(15,27,28) kan1 crc double mutants
develop a dramatic phenotype in which medial tissues of the
gynoecium lose their polarity resulting in the conversion of the
replum into septum and placental tissues and the formation of
ovules on the outside of the gynoecium. Further loss-of-KAN
function, in kan1,2 double mutants, results in the complete
spread of these internal medial tissues throughout the gyno-
ecium.(25) KAN1 is apparently expressed initially on the
external side of the replum,(26) whereas CRC is expressed
throughout the external epidermis of the gynoecium. Thus,
KAN1,2 and CRC appear to function together to exclude the
development of internal gynoecium tissues from the exterior.
Other factors with dual roles in organogenesis and
gynoecium patterning include AINTEGUMENTA (ANT) and
LEUNIG (LUG). Both of these genes have roles in lateral-
organ development. ANT is important for promoting the growth
of lateral organs by maintaining tissues in a meristematic
state,(29 –32) whereas loss-of-LUG function results in serrated
leaves, indicating a role for LUG in leaf blade develop-
ment.(33,34) LUG also plays a role in gynoecium development
and loss-of-function mutations result in a reduction in carpel
fusion near the style and the development of horn-like
structures from the apical region of the valves. Loss-of-ANT
function also has some minor effects on style fusion. Together,
however, ANT and LUG play a crucial role in the development
of all medial tissues in the gynoecium as ant lug double
mutants completely lack replum, style, septum and placental
tissues.(35) Whether these defects reflect a specific role for
ANT and LUG in medial-tissue development, or whether they
represent more general symptoms related to defects in lateral-
organ growth, these data highlight the important role lateral-
organ factors play in the morphogenesis of the gynoecium. It
will be interesting to determine how ANT and LUG help to
BioEssays 27.1 45
Review articles
promote the growth of specific regions of the gynoecium and at
what point these pathways might converge with those that
control organ identity.(15,36,37)
Post-fertilization fruit development
Introduction to the Arabidopsis fruit
The process of pod-shatter requires a patterning mechanism
to draw a line of dehiscence in the ovary. Correct spatial
regulation of where this line is drawn and when dehiscence
occurs is crucial for successful seed dispersal. While all of the
tissue layers present in the mature fruit are already formed in
the gynoecium before fertilization,(3,38,39) tissues of the valve
and valve margin region require as yet unknown signals
produced by post-fertilization processes to acquire their final
differentiated state. Certain mutations and hormone treat-
ments can mediate partial fruit maturation in the absence of
fertilization and suggest that these signals may emanate from
the developing seeds.(39–41)
Several changes in histology occur during fruit maturation
(Fig. 1B,C). The valve consists of approximately six cell
layers.(3,38,39) The outermost layer is termed the outer epider-
mis. Before fertilization, this cell layer contains small cells with
undifferentiated stomata. After fertilization, the outer-epider-
mal cells elongate and the stomata complete their develop-
ment providing pores in the epidermis for gas exchange.
Internal to the outer epidermis are three layers of mesophyll
tissue, comprising photosynthetic cells, that expand after
fertilization and form auxiliary vascular strands. The next-
most-internal layer after the mesophyll is a layer of lignified
tissue termed the endocarp layer b, or enb. During fruit
maturation, cells in this tissue layer form deposits of lignin, a
phenolic polymer that adds rigidity to the cell wall. The
epidermal layer on the inside surface of the ovary, termed
the endocarp layer a, or ena, expands greatly after fertilization
and then degenerates during fruit maturation.
At the valve margins, two tissue layers differentiate into the
dehiscence zone where the valves will separate from the
replum. On the valve-side of the margin, a patch of cells one to
two layers thick become lignified (Fig. 1C,D). This lignified
layer is continuous with the enb and will form the spring-like
tension mechanism that drives the separation of the valves
from the replum during pod-shatter.(3) On the replum side of
the valve margin, a layer of isodiametrically shaped cells
develop, termed the separation layer (Fig. 1C,E). Separation-
layer cells degenerate the middle lamella between adjacent
cell walls and separate from each other during dehiscence.
This process can be easily visualized using marker lines that
track the expression of middle-lamella-degrading enzymes
such as polygalacturonase.(42,43)
Regulation of valve-margin development
Genes controlling the post-fertilization differentiation of
tissues during fruit development can be divided into two
46 BioEssays 27.1
categories, those that promote valve-margin development and
those that repress it (Fig. 3). The first gene identified to have a
role in this process was the MADS-box transcription factor,
FRUITFULL (FUL).(38) Loss-of-FUL activity has dramatic
effects on the post-fertilization development of fruits, resulting
in a severe reduction in fruit size. In contrast to wild-type fruit,
which have space to accommodate the growth of developing
seeds, the seeds of ful mutant fruit are highly compacted due to
the failure of the fruit to elongate, hence the name fruitfull. FUL
is first expressed during early flower development in the
central dome of tissue that will emerge to form the gynoecium.
Later, FUL expression resolves and becomes limited to the
valve regions. Two patches of expression are also present
near the medial regions of the presumptive style and mark
tissues that will later develop into a distinct aggregation of
vascular elements.
Loss-of-FUL activity has a strong effect on the distribution
of cell types in the valve regions. In a ful mutant, cells in the
mesophyll tissue layers become lignified late in fruit develop-
ment and are much smaller than in wild type.(38,44) In weaker
ful alleles, these tissues are less lignified and stain similarly to
the separation layer (A. Roeder unpublished results). A
possible mechanism for these tissue defects was proposed
Figure 3. Genetic pathway controlling fruit development in
Arabidopsis. The distribution of tissues in the fruit is controlled
by two sets of genes that either promote valve margin
development (SHP1,2, IND and ALC) or restrict it from
surrounding regions (FUL, RPL). Valve margin development
is controlled by SHP1,2, IND and ALC. Both IND and ALC
appear to act largely downstream of SHP1,2, however, IND and
ALC also appear to be independently controlled by other
factors. SHP1,2 and IND control separation- and lignified-layer
development, while ALC is only necessary for separation-layer
formation. These four genes are limited to the valve margin
through repression by FUL in the valves(44,46) and by RPL in the
replum.(22) (A. Roeder unpublished results.)

by Ferrándiz et al. when it was discovered that valve-margin
identity genes are ectopically expressed in ful valves. This
indicated that FUL negatively regulates the expression of
valve-margin identity genes in the valves and suggested that
this ectopic activity may have caused the partial conversion of
valve tissues into valve margin. This hypothesis led to the
prediction that loss-of-function mutations in valve margin
identity genes should rescue ful valve development.
At the time, the only known factors with mutant phenotypes
affecting valve margin development were the redundant
MADS-box genes, SHATTERPROOF1 and 2 (SHP1,2).(45)
SHP1,2 are initially expressed broadly in all medial tissues
including the replum, valve margins, septum and ovules and
later become excluded from the replum as the gynoecium
matures. Together, shp1,2 mutations lead to the loss of valve-
margin lignification and separation layer formation. Fur-
thermore, ectopic expression of SHP genes leads to the
development of valve-margin-like characteristics in the valves,
including ectopic lignification of mesophyll cells and reduction
in valve expansion, similar to weak ful mutants. Surprisingly,
however, removal of SHP activity in ful mutants does not
substantially rescue the defects seen in ful valves, which
remain ectopically lignified in ful shp1,2 mutants.(44) Only the
development of stomata in the valve epidermis, which is
absent in ful mutants, is rescued. These results suggested that
other factors may be causing the ful defects. Additional
observations also suggested that other unknown factors
control valve margin development. First, shp1,2 mutants can
still develop valve-margin-like tissues near the apical end of
the fruit, leading to partial dehiscence. Second, the enhancer-
trap marker line, YJ80, which is expressed in the valve margin,
is still expressed at the apex of shp,1,2 fruit.(46)
Characterization of the basic-helix–loop–helix transcrip-
tion factor, ALCATRAZ (ALC), led to the discovery of a novel
gene class controlling dehiscence.(47) Unlike mutations in
shp1,2 which affect most tissues of the valve margin, alc
affects only a select set of these tissues. alc mutants have a
well-developed lignified layer but lack separation-layer tis-
sues. These defects may not be the sole cause of indehis-
cence, however, as the cells that replace the separation layer
can tear apart in alc mutants. Instead, it has been proposed
that a ‘‘lignified bridge’’ that forms between the lignified layer of
the valve margin and the lignified tissues of the replum vas-
cular bundle may prevent complete separation of the valves
from the replum. ALC is broadly expressed initially in the valve
and valve-margin region, and only later becomes restricted to
the valve margin after fertilization. ALC is also ectopically
expressed in the valves of ful mutants.(46) Removal of ALC
activity in ful mutants, however, does not abolish the ectopic
valve lignification even when combined with shp1,2, although
fruit size is moderately rescued in ful alc shp1,2 mutants.
Again, these results suggested that other valve-margin factors
may also contribute to the ful mutant phenotype.
Review articles
Of all the mutations that affect dehiscence, loss-of-
INDEHISCENT (IND) function has the strongest effect on
valve margin development.(46) IND encodes a member of an
atypical class of bHLH transcription factors in plants. In strong
alleles of ind, both the lignified layer and separation layer are
eliminated throughout the fruit. Unlike, SHP1,2 and ALC, IND
is strictly expressed in the valve-margin region just prior to
pollination. Further demonstration of the central role IND plays
in valve-margin development came when it was discovered
that the ind mutation was able to rescue many aspects of the
ful-mutant phenotype and could suppress the ectopic lignifica-
tion of ful valves. Mutations in IND are also able to substantially
rescue fruit elongation, expansion of the valve epidermal
cells and stomata development. Additional loss of other valve-
margin factors with ind further rescues the fruit elongation
defects of ful, with a near complete rescue of fruit elongation in
ind alc shp1,2 ful quintuple mutants. Together, these data
demonstrate that FUL’s main role in valve development is to
suppress the expansion of valve-margin-identity gene expres-
sion into the valves.
While the fruits of ind alc shp1,2 ful quintuple mutants are
rescued to a large extent, they also tend to be bumpy.
Bumpiness is often seen when valve-margin-identity factors
are misexpressed in valve tissues.(45) Thus, this bumpiness
could reflect the activity of yet unknown factors involved in
valve-margin development. Alternatively, this phenotype may
be caused by another defect only seen in the quintuple mutant,
the loss of enb lignification. In addition to being expressed in
the valves or valve margins, FUL, IND, ALC, SHP1,2 also
appear to be expressed together in the enb layer.(45,46) Their
role in enb development, however, has remained elusive.
Using phloroglucinol to visualize lignin deposition, the enb
layer is clearly absent when all five genes are mutated.
Interestingly, no other combination of mutations results in this
phenotype indicating that all genes are required for enb
lignification. How these antagonistic factors function together
is unclear, but these results suggest that the nature of their
interaction may be context dependent.
The model that has emerged so far is that valve-margin
development is controlled by several factors whose expres-
sion is limited to the valve margin through negative regulation
by FUL in the valves (Fig. 3). Valve-margin development is
also restricted from the replum, and raises the question of what
factors might repress valve-margin development in these
tissues? Identification of the replumless (rpl) mutation, which
results in a loss of replum development, helped to characterize
this part of the pathway.(22) RPL encodes a member of the BEL
subfamily of homeodomain transcription factors and is
expressed very early in the medial region of the gynoecium.
Expression can also be detected in the replum and septum
tissues in a mature gynoecium using the RPL promoter to drive
the expression of a reporter gene. In rpl mutants, Roeder et al.
found that the cells that replaced the replum were narrow in
BioEssays 27.1 47
Review articles
diameter and stained similarly to separation-layer cells. Using
marker lines that report the pattern of SHP and IND
expression, Roeder et al. confirmed that the replum in rpl
mutants had been converted into valve margin. Furthermore, it
was determined that ectopic SHP1,2 activity in the replum was
the cause of these defects, as loss-of-SHP function in a rpl-
mutant background rescued replum development. Together,
these data indicated that RPL promotes replum development
through the negative regulation of SHP genes and that RPL
itself is not necessary for replum development, per se, but
required to prevent the misspecification of replum tissues. This
role is similar to the part FUL plays in valve development.
In summary, work on fruit dehiscence in Arabidopsis has
elucidated a genetic network that draws a line of dehiscence in
the fruit, limiting specialized cellular processes necessary for
pod shatter to a narrow stripe of tissue termed the valve margin.
Valve-margin development is controlled by a group of genes
(SHP1,2, ALC and IND) with partially overlapping functions that
are limited to the valve margin through negative regulation by
FUL in the valves and RPL in the replum. In the future, it will be
interesting to determine how the expression patterns of these
genes are established and how these processes relate to the
prefertilization patterning of the gynoecium.
Prospects
Variation in fruit shape in Brassicaceae
While the silique is a characteristic fruit form of the Brassica-
ceae, many variations exist in the family, some of which
suggest possible genetic mechanisms based on our current
understanding of fruit development.(2,48) Canola (Brassica
napus), for example, develops a fruit without a replum and
structurally resembles fruits of moderate rpl alleles.(49)
Members of the genera, Coronopus and Crambe, develop
indehiscent fruit which also tend to be globe-shaped instead of
elongated, suggesting the possibility that FUL function may be
disrupted. Perhaps the most-specialized fruit in the Brassica-
ceae develop in Cakile, which grows along the ocean shore
and develops fruit well adapted for distribution by water.(2)
Whereas most fruit in the Brassicaceae develop dehiscence
zones that separate the valves from the replum, Cakile
develops an abscission zone positioned transversely in the
fruit, creating two corky segments that can be dispersed by
water. It would be interesting to determine if this new abscision
zone is patterned by the same factors that control valve margin
formation in Arabidopsis, or whether it develops by some other
mechanism, perhaps similar to the processes that control the
abscission of floral organs. Other variation also exists inclu-
ding the development of elongated styles (Brassica napus)
and explosive fruit (Cardamine). Thus, the Brassicaceae
family provides several avenues to study natural variation in
fruit form and patterning, making it a field that is sure to bear
fruit.
48 BioEssays 27.1
Agricultural applications
The fruit of Arabidopsis has proven a useful tool for dissecting
the genetic networks that control pattern formation in plants.
This information can serve a second purpose, as well, by
leading to innovative approaches for crop improvement.
Canola seeds, which are harvested for oil, can often be lost
due to premature pod shatter, particularly under adverse
weather conditions. With our new understanding of the genes
that control dehiscence, it may be possible to modify crops to
inhibit pod shatter and prevent such seed loss. RNAi methods,
for example, could be used to reduce the expression of SHP,
IND or ALC to inhibit valve margin formation.(50) Alternatively,
ectopic expression of FUL could be used to inhibit the activity
of the valve-margin-identity genes.(44) These genes could also
be used to identify and characterize naturally occurring genetic
variation in the form of Quantitative Trait Loci (QTL) affecting
dehiscence. Since completely indehiscent fruit present their
own problems, making seed harvesting more difficult, QTLs of
moderate effect may represent more useful tools for fine tuning
the dehiscence process.
Acknowledgments
We apologize to our colleagues whose work has not been sited
in full due limitations of space. We thank Adrienne Roeder for
permission to communicate unpublished results. We thank
members of the Yanfosky laboratory for helpful comments and
discussion during the preparation of this review and two
anonymous reviewers for useful suggestions. Our work on
fruit development is supported by the National Science
Foundation.
References
1. van der Pijl L. 1972. Principles of Dispersal in Higher Plants, 2nd Edition
(New York, Heidelberg, Berlin: Springer-Verlag).
2. Judd WS, Campbell CC, Kellogg EA, Stevens PF. 1999. Plant
Systematics: A Phylogenetic Approach (Sunderland: Sinauer Associ-
ates, Inc.).
3. Spence J, Vercher Y, Gates P, Harris N. 1996. ‘Pod shatter’ in Arabi-
dopsis thaliana, Brassica napus and B. juncea. J Microsc 181:195–203.
4. Smyth DR, Bowman JL, Meyerowitz EM. 1990. Early flower development
in Arabidopsis. Plant Cell 2:755–767.
5. Bowman JL. 1993. Arabidopsis: An Atlas of Morphology and Develop-
ment (New York: Springer-Verlag New York, Inc.).
6. Alvarez J, Smyth DR. 2002. CRABS CLAW and SPATULA genes regulate
growth and pattern formation during gynoecium development in
Arabidopsis thaliana. Int J Plant Sci 163:17–41.
7. Bowman JL, Baum SF, Eshed Y, Putterill J, Alvarez J. 1999. Molecular
genetics of gynoecium development in Arabidopsis. Curr Top Dev Biol
45:155–205.
8. Vivian-Smith A, Koltunow AM. 1999. Genetic analysis of growth-
regulator-induced parthenocarpy in Arabidopsis. Plant Physiol 121:
437–451.
9. Sessions RA, Zambryski PC. 1995. Arabidopsis gynoecium structure in
the wild type and in ettin mutants. Development 121:1519–1532.
10. Johnson MA, Preuss D. 2004. Plotting a course: multiple signals guide
pollen tubes to their targets. Dev Cell 2:273–281.
11. Nemhauser JL, Feldman LJ, Zambryski PC. 2000. Auxin and ETTIN in
Arabidopsis gynoecium morphogenesis. Development 127:3877–3888.
12. Benkova E, Michniewicz M, Sauer M, Teichmann T, Seifertova D, et al.
2003. Local, efflux-dependent auxin gradients as a common module for
plant organ formation. Cell 115:591–602.

13. Geldner N, Friml J, Stierhof YD, Jurgens G, Palme K. 2001. Auxin
transport inhibitors block PIN1 cycling and vesicle trafficking. Nature
413:425–428.
14. Okada K, Ueda J, Komaki MK, Bell CJ, Shimura Y. 1991. Requirement of
the auxin polar transport system in early stages of Arabidopsis floral bud
formation. Plant Cell 3:677–684.
15. Alvarez J, Smyth DR. 1999. CRABS CLAW and SPATULA, two
Arabidopsis genes that control carpel development in parallel with
AGAMOUS. Development 126:2377–2386.
16. Heisler MG, Atkinson A, Bylstra YH, Walsh R, Smyth DR. 2001.
SPATULA, a gene that controls development of carpel margin tissues
in Arabidopsis, encodes a bHLH protein. Development 128:1089–1098.
17. Alvarez J, Smyth DR. 1998. Genetic pathways controlling carpel
development in Arabidopsis thaliana. J Plant Res 111:295–298.
18. Kuusk S, Sohlberg JJ, Long JA, Fridborg I, Sundberg E. 2002. STY1 and
STY2 promote the formation of apical tissues during Arabidopsis
gynoecium development. Development 129:4707–4717.
19. Fridborg I, Kuusk S, Moritz T, Sundberg E. 1999. The Arabidopsis dwarf
mutant shi exhibits reduced gibberellin responses conferred by over-
expression of a new putative zinc finger protein. Plant Cell 11:1019–1032.
20. Jacobsen SE, Olszewski NE. 1993. Mutations at the SPINDLY locus of
Arabidopsis alter gibberellin signal transduction. Plant Cell 5:887–896.
21. Robinson-Beers K, Pruitt RE, Gasser CS. 1992. Ovule development in
wild-type Arabidopsis and two female-sterile mutants. Plant Cell 4:1237–
1249.
22. Roeder AH, Ferrándiz C, Yanofsky MF. 2003. The role of the
REPLUMLESS homeodomain protein in patterning the Arabidopsis fruit.
Curr Biol 13:1630–1635.
23. Sawa S, Watanabe K, Goto K, Liu YG, Shibata D, et al. 1999.
FILAMENTOUS FLOWER, a meristem and organ identity gene of Arabi-
dopsis, encodes a protein with a zinc finger and HMG-related domains.
Genes Dev 13:1079–1088.
24. Siegfried KR, Eshed Y, Baum SF, Otsuga D, Drews GN, et al. 1999.
Members of the YABBY gene family specify abaxial cell fate in Arabi-
dopsis. Development 126:4117–4128.
25. Eshed Y, Baum SF, Perea JV, Bowman JL. 2001. Establishment of
polarity in lateral organs of plants. Curr Biol 11:1251–1260.
26. Kerstetter RA, Bollman K, Taylor RA, Bomblies K, Poethig RS. 2001.
KANADI regulates organ polarity in Arabidopsis. Nature 411:706–709.
27. Bowman JL, Smyth DR. 1999. CRABS CLAW, a gene that regulates
carpel and nectary development in Arabidopsis, encodes a novel protein
with zinc finger and helix-loop-helix domains. Development 126:2387–
2396.
28. Eshed Y, Baum SF, Bowman JL. 1999. Distinct mechanisms promote
polarity establishment in carpels of Arabidopsis. Cell 99:199–209.
29. Elliott RC, Betzner AS, Huttner E, Oakes MP, Tucker WQ, et al. 1996.
AINTEGUMENTA, an APETALA2-like gene of Arabidopsis with pleio-
tropic roles in ovule development and floral organ growth. Plant Cell 8:
155–168.
30. Klucher KM, Chow H, Reiser L, Fischer RL. 1996. The AINTEGUMENTA
gene of Arabidopsis required for ovule and female gametophyte
development is related to the floral homeotic gene APETALA2. Plant
Cell 8:137–153.
31. Krizek BA. 1999. Ectopic expression of AINTEGUMENTA in Arabidopsis
plants results in increased growth of floral organs. Dev Genet 25:224–
236.
Review articles
32. Mizukami Y, Fischer RL. 2000. Plant organ size control: AINTEGUMENTA
regulates growth and cell numbers during organogenesis. Proc Natl
Acad Sci USA 97:942–947.
33. Liu Z, Meyerowitz EM. 1995. LEUNIG regulates AGAMOUS expression in
Arabidopsis flowers. Development 121:975–991.
34. Conner J, Liu Z. 2000. LEUNIG, a putative transcriptional corepressor
that regulates AGAMOUS expression during flower development. Proc
Natl Acad Sci USA 97:12902–12907.
35. Liu Z, Franks RG, Klink VP. 2000. Regulation of gynoecium marginal
tissue formation by LEUNIG and AINTEGUMENTA. Plant Cell 12:1879–
1892.
36. Yanofsky MF, Ma H, Bowman JL, Drews GN, Feldmann KA, et al. 1990.
The protein encoded by the Arabidopsis homeotic gene AGAMOUS
resembles transcription factors. Nature 346:35–39.
37. Pinyopich A, Ditta GS, Savidge B, Liljegren SJ, Baumann E, et al. 2003.
Assessing the redundancy of MADS-box genes during carpel and ovule
development. Nature 424:85–88.
38. Gu Q, Ferrandiz C, Yanofsky MF, Martienssen R. 1998. The FRUITFULL
MADS-box gene mediates cell differentiation during Arabidopsis fruit
development. Development 125:1509–1517.
39. Vivian-Smith A, Luo M, Chaudhury A, Koltunow A. 2001. Fruit develop-
ment is actively restricted in the absence of fertilization in Arabidopsis.
Development 128:2321–2331.
40. Ohad N, Margossian L, Hsu YC, Williams C, Repetti P, et al. 1996. A
mutation that allows endosperm development without fertilization. Proc
Natl Acad Sci USA 93:5319–5324.
41. Chaudhury AM, Ming L, Miller C, Craig S, Dennis ES, et al. 1997.
Fertilization-independent seed development in Arabidopsis thaliana.
Proc Natl Acad Sci USA 94:4223–4228.
42. Peterson M, Sander L, Child R, Van Onckelen H, Ulvskov P, et al. 1996.
Isolation and characterization of a pod dehiscence zone-specific
polygalacturonase from Brassica napus. Plant Mol Biol 31:517–527.
43. Jenkins ES, Paul W, Craze M, Whitelaw CA, Weigand A, et al. 1999.
Dehiscence-related expression of an Arabidopsis thaliana gene encod-
ing a polygalacturonase in transgenic plants of Brassica napus. Plant
Cell Environ 22:159–167.
44. Ferrándiz C, Liljegren SJ, Yanofsky MF. 2000. Negative regulation of the
SHATTERPROOF genes by FRUITFULL during Arabidopsis fruit devel-
opment. Science 289:436–438.
45. Liljegren SJ, Ditta GS, Eshed Y, Savidge B, Bowman JL, et al. 2000.
SHATTERPROOF MADS-box genes control seed dispersal in Arabidop-
sis. Nature 404:766–770.
46. Liljegren SJ, Roeder AH, Kempin SA, Gremski K, Ostergaard L, et al.
2004. Control of fruit patterning in Arabidopsis by INDEHISCENT. Cell
116:843–853.
47. Rajani S, Sundaresan V. 2001. The Arabidopsis myc/bHLH gene
ALCATRAZ enables cell separation in fruit dehiscence. Curr Biol
11:1914–1922.
48. Rollins RC. 1993. The Cruciferae of Continental North America-
Systematics of the Mustard Family from the Arctic to Panama (Stanford:
Stanford. Press).
49. Meakin PJ, Roberts JA. 1990. Dehiscence of fruit in oilseed rape (Brassica
napus L.)1. Anatomy of pod dehiscence. J Exp Bot 41:955–1002.
50. Chuang CF, Meyerowitz EM. 2000. Specific and heritable genetic
interference by double-stranded RNA in Arabidopsis thaliana. Proc Natl
Acad Sci USA 97:4985–4990.
BioEssays 27.1 49