Dispatch
917
Plant morphogenesis: Designing leaves
David Jackson
Leaf shape and architecture vary greatly throughout the
plant kingdom, and even within an individual plant
during different phases of growth. Now, the
development of a compound leaf architecture in tomato
has been shown to be associated with the expression
of the knotted1 homeobox gene in leaf primordia.
Address: Plant Gene Expression Center, USDA/ARS, University of California, Berkeley, 800 Buchanan Street, Albany, California 94710, USA.
Current Biology 1996, Vol 6 No 8:917–919
© Current Biology Ltd ISSN 0960-9822
Leaves exhibit a wide range of form and function: as well as
the flattened photosynthetic organs we are most familiar
with, modifications of leaf morphology produce the rigid
spines of cacti, the cup-shaped insect traps of pitcher
plants, colorful flower parts such as petals and some of the
tendrils used for attachment by climbing plants. One fundamental choice in leaf architecture is whether to be simple
or compound (see Fig. 1a–c). Simple leaves can have
simple shapes — such as round, oval or strap-like — or they
can be lobed, like an oak leaf. In contrast, compound leaves
are subdivided into smaller units called leaflets.
Given the immense variation in leaf architecture, how do
we define a leaf? The only defining feature may be that all
leaves are initiated as dorsiventral or flattened primordia
from the shoot apical meristem [1]. Despite its complexity,
a compound leaf is considered to be a single leaf because it
is initiated as a single primordium from the shoot apical
meristem; the leaflets are initiated from the leaf primordium later in development [2]. The shoot apical meristem is the group of indeterminate stem cells at the apex of
the growing shoot that serves to initiate organs throughout
the life of the plant [3]. In contrast, leaves usually have a
restricted, or determinate, potential for growth (certain
plants, such as some ferns, have indeterminate leaves,
however). In flowering plants, the shoot apical meristem
also initiates axillary shoot meristems that will become
leafy branches or flowers, but axillary shoot primordia differ
from leaf primordia because they are radially symmetrical.
In a variety of plants with simple leaves, knotted1 (kn1) and
closely related members of the kn1 homeobox gene family
are expressed in the shoot apical meristem and in axillary
meristems, but not in leaves (Fig. 2). In fact, kn1 expression is down-regulated in a group of cells on the flank of
the meristem which will form the next leaf primordium.
These observations suggest that the role of kn1 might be to
maintain cells of the shoot apical meristem in an indeterminate state [4]. Support for this hypothesis came from
studies of dominant Kn1 mutations in maize that cause
ectopic expression of kn1 in leaf primordia. This ectopic
expression is associated with the sporadic growths (‘knots’)
that result from extra cell divisions in the leaf, perhaps
indicating a reduction in determinacy in the leaf [4].
Hareven et al. [5] have discovered a possible new role for
kn1 in the development of compound leaves. The compound leaf of tomato has about nine leaflets borne on a
Figure 1
(a)
(b)
(c)
(d)
© Current Biology 1996
Different leaf forms (architectures). (a) Simple leaf (sunflower). (b)
Simple, lobed leaf (oak). (c) Compound leaf (tomato). (d) ‘Super
compound’ leaf from a tomato plant overexpressing the maize knotted1
gene (reproduced with permission from [5]).
918
Current Biology 1996, Vol 6 No 8
Figure 2
(a)
(b)
2
Shoot
apical
meristem
1
(c)
*
Midvein
Midvein
Expression pattern of kn1 mRNA in maize and
Tkn1 mRNA in tomato. (a) Expression of kn1
in the maize shoot apical meristem, shown in
longitudinal section. Note that kn1 is not
expressed in the leaf primordia, denoted 1
and 2 [1,5] and is down-regulated in a group
of cells on the flank of the shoot apical
meristem where the next leaf will initiate
(marked *). (b) Transverse section through a
maize leaf primordium; note the lack of kn1
expression. (c) Transverse section through a
tomato leaf primordium; note that Tkn1 is
expressed at the margins of the leaf
primordium, where leaflets will initiate, and in
the provascular strands. Leaf primordia are
shown in green, vascular strands in mauve
and kn1 expression in purple.
© Current Biology 1996
central midrib. Upon overexpression of the maize kn1
gene, however, the leaves become ‘super compound’, with
each leaf having up to 2000 leaflets (Fig. 1d). The analysis
by Hareven et al. [5] suggests that compound leaves
develop by a mechanism fundamentally different from
that in simple leaves, and that the kn1 gene may have been
recruited to function in the compound developmental
program.
A major advantage of tomato for studying the control of
leaf architecture is the large collection of mutants that have
defects in this process [6]. Three recessive mutations —
entire, potato leaf and trifoliate — reduce the compound
nature of the leaf, although only one mutation, the semidominant Lanceolate (La), is known to convert the tomato
leaf to a fully simple architecture. Another dominant mutation, Petroselinum (Pts), increases the compound nature of
the leaf such that each leaflet is further subdivided. Overexpression of kn1 in either the potato leaf, trifoliate or Pts
mutant backgrounds gives largely additive effects on leaf
architecture. Interestingly, the simple leaf architecture of
La plants is epistatic to the effects of the kn1 transgene; the
double mutant in this case showed an alteration of leaf
shape (lobing) but did not become compound [5].
Could one of the tomato leaf mutants correspond to the
tomato homologue of the maize kn1 gene? For example,
Pts has the phenotype expected for a kn1 overexpressor,
and La could conceivably be a dominant-negative or antimorph allele of kn1. Previous dosage studies are consistent
with the idea that La acts as an antimorph [7]. To address
this question, a kn1-related gene was isolated from tomato
and mapped. This gene, named Tkn1, did not map to the
same location as La or Pts, although there are probably at
least two other kn1-related genes in tomato [5,8,9].
However, Tkn1 did map close to entire, a gene characterized
by recessive mutants whose leaves appear simple because
their leaflets are fused. Tkn1 cDNAs from entire plants had
the same sequence as wild-type, although this does not
rule out the possibility that the entire gene encodes Tkn1, as
the effects of the mutation could be regulatory, affecting
for example the domain of expression of the gene rather
than its coding sequence.
The normal expression pattern of Tkn1 is strikingly different from that of its putative homologues in either maize or
Arabidopsis. In these plants, which have simple leaves, kn1
is normally expressed only in the shoot apical meristem and
is down-regulated as a leaf is initiated (Fig. 2a,b) [4,8,10,11].
Tkn1, however, is expressed in both leaf primordia and the
shoot apical meristem of tomato (Fig. 2c). Is the initiation of
leaflets from the leaf primordium of a compound leaf therefore somewhat analogous to the initiation of leaves by the
shoot apical meristem? Should the compound tomato ‘leaf’
be more accurately described as a shoot?
Several observations refute this hypothesis. First, the
tomato leaf is initiated as a dorsiventral structure, unlike
shoots which are radially symmetrical at inception. Second,
the leaflets of a tomato leaf are initiated basipetally — first
near the tip and later towards the base of the leaf; organ
initiation by shoots is in the opposite direction (acropetal).
Third, the leaves of flowering plants are initiated in close
association with an axillary meristem; the leaflets on a
compound leaf do not have axillary meristems. Perhaps
one should consider the tomato leaf as sharing characteristics of both shoot and leaf identities, an idea that has been
proposed for other compound leaves [12].
What then is the role of kn1 homologues in the development of compound leaves? In tobacco or Arabidopsis —
plants with simple leaves — kn1 overexpression causes the
Dispatch
leaves to become lobed but not compound [8,13,14]. Even
in tomato, overexpression of kn1 in sepals (simple leaf
homologues in the flower) or in the simple leaves of La
mutant plants does not make these organs compound.
Thus, kn1 overexpression appears to accentuate a developmental program specific for compound leaves. Presumably
it does this by maintaining leaf cells in a developmental
state in which they can respond to the ‘compound’ signal.
An alternative possibility is that lobed and compound
leaves might not be fundamentally different, but that compoundness is a severe version of lobing. Evidence to
support this hypothesis comes from overexpression of kn1
or a kn1-like gene in simple-leaved plants: this induces
lobed leaves which, like compound leaves, also have some
shoot characteristics [13,14]. For example, stipules and
meristems, normally found only at the base of the leaf, are
produced ectopically at the bases of the lobes [14].
Elucidation of the role of Tkn1 in tomato leaf development
awaits the isolation of loss-of-function mutations. In Arabidopsis, the shootmeristemless (stm) gene encodes the probable kn1 homologue [11]. Recessive stm mutants fail to
initiate leaves as a result of a failure in either the initiation
or the maintenance of the shoot apical meristem. It might
therefore prove necessary to isolate weak alleles or to use
conditional expression in order to allow leaf initiation to
occur and then to test the role of Tkn1 in compound leaf
development.
The tomato experiments open up many interesting questions. For example, is Tkn1 down-regulated to allow initiation of a leaf and then reactivated in a group of cells at
the margin of the leaf where the leaflets are initiated?
Alternatively, Tkn1 expression may be constitutive during
leaf initiation and early development in tomato. What are
the mechanisms controlling the differential regulation of
kn1 during the development of simple and compound
leaves? As overexpression of kn1 is not sufficient to
convert a simple leaf into a compound leaf, what are the
other factors involved? Is kn1 normally expressed during
the development of other compound leaves or of lobed
leaves? What about fern leaves which differentiate
acropetally, and so have even more similarities to shoots
than do the compound leaves of tomato? Overexpression
of kn1 in a number of species has elegantly demonstrated
our abilities to manipulate leaf form, and grocery stores
may one day carry designer vegetables resulting from
such endeavors. However, further understanding of the
underlying mechanisms controlling leaf architecture
awaits the isolation of leaf-shape genes from compoundleaved plants like tomato and pea, and the determination
of the interactions between these genes and kn1 in leaf
primordia.
Acknowledgements
I thank members of the Hake lab, Seung Y. Rhee and James Keddie for
discussion of this article.
919
References
1. Kaplan DR: Comparative developmental evaluation of the
morphology of unifacial leaves in the monocotyledons. Bot Jahrb
Syst 1975, 95:1–105.
2. Dengler NG: Comparison of leaf development in normal (+/+),
entire (e/e), and Lanceolate (La/+) plants of tomato, Lycopersicon
esculentum ‘Ailsa Craig’. Bot Gaz 1984, 145:66–77.
3. Sussex I M: Developmental programming of the shoot meristem.
Cell 1989, 56:225–229.
4. Smith L, Greene B, Veit B, Hake S: A dominant mutation in the
maize homeobox gene, Knotted-1, causes its ectopic expression
in leaf cells with altered fates. Development 1992, 116:21–30.
5. Hareven D, Gutfinger T, Parnis A, Eshed Y, Lifschitz E: The making of
a compound leaf: genetic manipulation of leaf architecture in
tomato. Cell 84:735–744.
6. Stevens AM, Rick CM: Genetics and breeding. In The Tomato Crop.
Edited by Atherton JG, Rudick J. New York: Chapman and Hall; 1986.
7. Stettler RF: Dosage effects of the lanceolate gene in tomato. Am J
Bot 1964, 51:253–264.
8. Lincoln C, Long J, Yamaguchi J, Serikawa K, Hake S: A knotted1-like
homeobox gene in Arabidopsis is expressed in the vegetative
meristem and dramatically alters leaf morphology when
overexpressed in transgenic plants. Plant Cell 1994,
6:1859–1876.
9. Kerstetter R, Vollbrecht E, Lowe B, Veit B, Yamaguchi J, Hake S:
Sequence analysis and expression patterns divide the maize
knotted1-like homeobox genes into two classes. Plant Cell 1994,
6:877–1887.
10. Jackson D, Veit B, Hake S: Expression of maize KNOTTED-1
related homeobox genes in the shoot apical meristem predicts
patterns of morphogenesis in the vegetative shoot. Development
1994, 120:405–413.
11. Long JA, Moan EI, Medford JI, Barton MK: A member of the
KNOTTED class of homeodomain proteins encoded by the
SHOOTMERISTEMLESS gene of Arabidopsis. Nature 1996,
379:66–69.
12. Lacroix CR, Sattler R: Expression of shoot features in early leaf
development of Murraya paniculata (Rutaceae). Can J Bot 1994,
72:678–687.
13. Sinha NR, Williams RE, Hake S: Overexpression of the maize
homeobox gene, KNOTTED-1, causes a switch from determinate
to indeterminate cell fates. Genes Dev 1993, 7:787–795.
14. Chuck G, Lincoln C, Hake S: KNAT1 induces lobed leaves with
ectopic meristems when overexpressed in Arabidopsis. Plant Cell
1996, in press.