York
GEORGE COUPLAND has green fingers. With a little genetic tinkering, he has
produced plants that flower to order. Doused with the correct chemical trigger,
they spring into bloom within days. Unfortunately, the chemical is a
steroid鈥攏ot the sort of thing you鈥檇 want to spread liberally on your
roses. But, with a few modifications, the work of Coupland and his colleagues at
the John Innes Centre in Norwich has enormous commercial potential. If
other plants could be persuaded to flower on cue, farmers would be able to
control when their crops were ready for harvesting and flower growers could
produce a continuous supply of blooms.
Coupland鈥檚 cooperative weeds are an offshoot of a growing body of knowledge
about the process of flowering. In the past few years, researchers have pieced
together a map of the genes that link flowering to environmental signals. They
have identified a genetic switch that determines whether a bud turns into a
flower or a normal shoot. They have even begun to understand how plants choose
where flowers will appear on the shoot.
Advertisement
The fruits of this research鈥攆rom trees that bloom when they are just a
few centimetres high to embryonic plants that begin the flowering process within
the seed鈥攕how that a change in the right genes can dramatically alter when
and how plants bloom.
The story of genetic research into flowering stretches back to 1920, when two
American biologists reported that a mutation in a single gene can completely
alter the flowering habits of tobacco. The mutation transforms tobacco from a
plant that flowers as soon as it reaches a certain size to one that flowers only
if the days are the right length. Since then, other mutations have helped to
identify many genes associated with floral control circuits. The common wall
cress Arabidopsis thaliana, used by Coupland and many other plant
geneticists, contains around 30 such genes, so it is hardly surprising that a
detailed picture of the flowering process is only now emerging.
Geneticists have identified four main groups of flowering genes in
Arabidopsis. The first pushes the plant towards blooming during the long
summer days. The second delays flowering when the days are short. Another relays
information about the temperature and tips the balance towards producing blooms
when the weather warms up in spring. A final group responds to the plant鈥檚
internal state and presses the flowering button more insistently as plants grow
older and bigger. This allows factors such as rainfall, soil quality and
disease, which affect a plant鈥檚 growth rate, to influence flowering.
鈥淚ntuitively, I would feel that there are several different routes to
flowering,鈥� says Coupland, 鈥渁nd I would imagine that they could well counteract
each other.鈥� Whatever the conditions, though, Arabidopsis will always
flower eventually. Under natural conditions, it may bloom any time from around
three weeks to nearly a year after germination. But as Coupland鈥檚 research
shows, a sufficiently strong push on a single pathway can hot-wire the
decision-making circuits, creating plants that flower on demand.
Early bloom
Coupland鈥檚 group used a gene that promotes flowering when the days are long,
encouraging Arabidopsis to flower in summertime. Up to a point, the
more daylight hours there are when the plant is growing, the quicker it will
come into bloom. Plants carrying only mutant copies of the gene, however, lose
this response. Length of day has no effect on how quickly they flower, hence the
gene鈥檚 name鈥擟ONSTANS or CO.
In normal plants, long days stimulate the CO gene to produce more of
a protein called CO, which in turn stimulates other flowering genes. Coupland鈥檚
genetically engineered weed produces a hybrid protein, which consists of CO
attached to a steroid receptor. The receptor prevents CO from doing its work by
keeping it in the body of the cell, away from the nucleus where it would
normally activate other genes. Dousing the plants with the right steroid,
however, turns the receptor into a taxi. 鈥淲hen the steroid binds to the
receptor, the receptor moves to the nucleus,鈥� says Coupland. It鈥檚 not quite
sunshine in a bottle, but it has the same effect.
A similar principle could be used to make crops flower in response to the
weather forecast, rather than the weather itself. 鈥淵ou can unlink flowering from
environmental signals,鈥� says Coupland. 鈥淭here are cases where it would be
advantageous to have plants flowering slightly earlier than normal.鈥� In Canada,
for example, where the short summer means that the oil seed rape crop does not
always have time to ripen, forcing plants to flower a couple of weeks early
might make all the difference.
This assumes that other plants can be engineered in the same way as
Arabidopsis, which is by no means certain. 鈥淚t is not yet possible to tell
how far you can extend the research to other species,鈥� says Coupland. His team
is now testing whether the CO gene can induce tobacco plants to flower.
But if the work on plant flowering is to progress, researchers need to study
genes that are common to many flowering plants.
And herein lies a problem. The DNA sequences of the genes involved largely
remain a mystery. This is the case even in the much-studied Arabidopsis.
The weed鈥檚 entire genome is currently being sequenced but, as yet, there is no
way that geneticists can be sure that the genes they are studying in
Arabidopsis are also found in other plant species.
However, one group of genes has emerged that several plants share. When
activated by CO or by one of the other flowering control pathways these act like
a 鈥渇loral switch鈥� making a shoot grow into a flower. At the Salk Institute for
Biological Studies in La Jolla, California, Detlef Weigel has been studying one
of these genes, called LEAFY. Arabidopsis plants that lack a
working copy of the gene make abnormal flowers containing leaf-like organs. And
the familiar snapdragon, Antirrhinum鈥攚hich is only distantly
related to Arabidopsis鈥攈as a similar gene that switches the plant
from making a shoot to producing a flower.
Working with Ove Nilsson from the Swedish University of Agricultural Sciences
in Ume氓, Weigel鈥檚 team introduced extra copies of LEAFY into
Arabidopsis. At least four other genes are needed in the floral switch, but
the researchers wanted to see whether they could push the whole mechanism into
the 鈥渙n鈥� position by artificially activating LEAFY alone. Sure enough,
the genetically engineered plants did flower earlier than normal.
Instant bonsai
Next, the team added extra copies of LEAFY into aspen trees. The
result varied from plant to plant probably depending, say the researchers, on
how much protein the new genes were producing. 鈥淲e have examples were the plants
just get a few inches tall and then make flowers and die,鈥� says Weigel. 鈥淲e also
have ones that make it into the greenhouse and grow as bushes.鈥� The fact that
the LEAFY gene has the same effect in plants as diverse as
Arabidopsis and aspen suggests that it could be used as a universal 鈥渙n鈥�
button to induce early flowering in all plants.
Weigel and Nilsson鈥檚 instant bonsai could bring tree breeders out of the
woods鈥攍iterally. 鈥淲hat we have so far certainly wouldn鈥檛 be useful in the
field,鈥� says Weigel. 鈥淏ut where it would be useful would be in a breeding
programme. If you can make plants flower earlier, then you can make breeding
quicker.鈥� Of course, tiny trees growing in the laboratory could not tell you
much about their ability to thrive in a real forest.
But molecular biologists can now pinpoint individual plants with particular
characteristics such as disease resistance, by looking at their DNA. So breeding
with pint-sized trees is feasible. 鈥淵ou breed in the characteristic you want
using the tiny plants,鈥� says Weigel. 鈥淭hen you simply remove the introduced gene
by standard genetic crosses and you have normal trees again.鈥�
There may be ways to shift flowers in space as well as time. Not only do
Weigel and Nilsson鈥檚 aspens flower very early, they also make isolated flowers
rather than many blooms clustered into catkins. Research with
Arabidopsis and Antirrhinum may explain what is going on. Both
plants normally produce flowers on tall flowering stems which keep on growing
until the plant dies. Blooms appear on all sides of this spike, but something
keeps the floral switch in the 鈥渙ff鈥� position at the growing tip.
Switched-on flowers
In January, Desmond Bradley and his colleagues, also at the John Innes
Centre, reported that at least one of the genes involved in this process is the
same in both Arabidopsis and Antirrhinum. In
Arabidopsis, the gene is called TERMINAL FLOWER 1, or
TFL1, because mutations in the gene produce short flowering spikes that end
in a bloom. TFL1 is normally very active in cells near the tip of the
flowering stem, where it inhibits LEAFY and other floral switch
genes.
The equivalent gene plays a similar role in Antirrhinum. However, by
introducing extra copies of the LEAFY gene into Arabidopsis,
Weigel and Nilsson showed that they could override TFL1鈥檚 protection of
the shoot tip. Arabidopsis with extra LEAFY genes produce a
bloom at the tip of the flowering stem. In the genetically engineered aspen
trees, this process may have reached its limit. Catkins are just squashed up
flowering stems, and in the miniature aspens the tip of the catkin appears to
turn into a flower before any side-blooms have formed.
Following this work, genetic flower arranging could be just around the
corner. 鈥淥ne slightly frivolous example would be to make novelty ornamental
plants,鈥� says Weigel. 鈥淚f you could turn all the lateral shoots into lateral
flowers, for example, you could have a rose which not only had a flower at the
apex of each branch, but also flowers in all of the leaf axils.鈥� A less
colourful but more practical application would be crops tailored to produce more
fruits or seeds, or to make these more accessible for harvesting.
Genetic flower arranging could also be used to change the shape of a plant.
Normally, growth occurs only at the shoot tip and at the bud at the base of each
leaf. Each growing point can form either a flower or a shoot, but not both, so
manipulating the position of flowers can alter a plant鈥檚 shape.
鈥淵ou might want to change the architecture of a plant for growing in a
particular environment, making it more bushy or more elongated,鈥� suggests
Bradley. Aspen plants carrying extra copies of LEAFY, for example, do
not grow into trees because at some stage in their development the shoot tip
turns into a flower. When this happens early, the plant cannot make enough
leaves so it dies. In plants where this occurs later, shoot tips turn into
flowers, encouraging side branches to grow from buds lower down the plant,
creating a bush rather than a tree.
Controlling flowering may not be just about 鈥渟witching on鈥� flowers though.
Turn off vegetative growth, and flower production may occur by default. At the
University of California at Berkeley, Renee Sung has found two genes responsible
for early shoot growth. Arabidopsis plants that lack working copies of
these develop a 鈥渓ive fast, die young鈥� attitude that would make a rock star look
like a shrinking violet. 鈥淚n mutants,鈥� says Sung, 鈥渢he shoot apex begins some
kind of activity in the embryo.鈥� Exactly what is happening becomes clear shortly
after sowing. The seedlings open their two seed leaves (the cotyledons) and
then, instead of making a normal shoot, they bolt and produce a tiny flowering
stem. A month or so later, consumed by flowers, the mutants die.
Such leaps in our understanding of the genetics of flowering are set to
change farming and horticulture. But Coupland, Weigel and their colleagues are
not the first to goad plants into flowering. Traditionally, apple growers shook
recalcitrant trees awake by beating their trunks with sticks. Plants often bloom
when things seem to be going badly wrong, in a last-ditch attempt at
reproduction.