is essential. If just one extra ‘accidental’ cell division occurred per year, shoot meristems could be a thousand times larger than normal within a decade,15
and trees that live for many decades would become grotesque super-organisms.
A commonplace example illustrating the urgent need to regulate when to grow
a leaf, and when to do some other more specialized job like building a flower, is
Ancient genes, new pl Ants a 73
the case of cauliflowers. The head of a cauliflower shows what happens when
meristems go wrong, as masses of flower buds fail to undergo normal floral
development, causing over-proliferation of cells and the formation of the thick
creamy ‘curd’. It is possible to induce a similar mutation in the flowering plant Arabidopsis and make it produce weird miniature ‘cauliflower-like’ structures on its flowering shoot.16 The signals that determine when the meristem ‘decides’ it is time to produce daughter cells destined to become leaves are not fully understood, but we know they involve the powerful plant growth hormone auxin
(from the Greek auxein, meaning to grow).17 Whatever the mechanism, leaf formation is initiated at specific positions in the shoot tip and at regular intervals, as revealed by an often helical sequence of bulges that form around the
meristem. This ensures the appearance of the elegant and efficient arrangement
of leaves that minimizes competition for light between them.
Leaf formation is initiated when the meristem transitions from indeterminate
growth to a determinate growth programme. In indeterminate growth mode,
plants can become giant trees like those of the primeval swamp forests, but in
determinate growth mode things become finite. Determinate growth is the norm
for all animals, including humans. Our bodies call it a day and stop growing at
maturity and in this sense we are just like plants when they make leaves. This
transition requires switching off genes maintaining self-renewing stem cells in the shoot meristem, and activating those directing the development of leaves. The
KNOX family of genes maintains cell production in the shoot apex by coding for a suite of powerful transcription factors. These are small proteins that bind to
each other and specific segments of the DNA molecule to switch genes on and off
at specific times and places. In this way, transcription factors provide plants with a set of logical molecular operators for regulating patterns of gene expression that control their development. If factor X binds to factor Y, then the resulting combination may bind to the DNA molecule in a particular region and switch a single gene, or
whole collections of genes, on or off. The default setting for KNOX genes is
‘activated’, meaning they normally produce transcription factors blocking the
development of new leaves. During leaf development, proteins coded by another
set of genes— ARP genes—block the transcription of KNOX genes, allowing recruitment of stem cells into the developmental pathway for making leaves.
Because of this negative interaction, ARP gene proteins are said to ‘repress’ the
74 a Ancient genes, new pl Ants
activity of KNOX genes. KNOX genes derive their name from the knots they produce in leaves when plants are engineered to express them in leaves instead
of the normal shoot meristem position. Knots appear because cells undergoing
excessive divisions cause the rapidly expanding leaf blade to pucker and buckle.18
Originally discovered in the crop plant maize,19 the KNOX–ARP genetic toolkit for instructing shoots to make leaves was discovered later to be doing much the
same role in the flowering plant Arabidopsis. This tells us something important.
Because according to the green tree of life, these two species last shared a com-
mon ancestor over 140 million years ago. The logical inference we can draw from
this is that the KNOX–ARP two-component core genetic toolkit for making leaves was inherited from a common ancestor that lived at that time, or even earlier. The question is how much further back in time can we trace the evolutionary history
of this genetic toolkit for building leaves?
Jane Langdale is a British professor of molecular biology addressing exactly this sort of question by scrutinizing the evolutionary history and function of genes
involved for everything from mosses and hornworts to ferns and flowering plants.
Langdale is unusual for starting out her career as a human geneticist before being won over by the charms of the plant kingdom. She could not resist being seduced
by big science questions about the origin and diversification of our land floras, and about the rich opportunities for seeing further by marrying evidence from
fossils and DNA molecules. Working with the spikemoss Selaginella, her team reported the exciting discovery a decade ago that the KNOX–ARP genetic toolkit for building leaves had already originated by some 350 million years ago, far earlier in plant evolutionary history than anyone had previously suspected. The evidence
came from elegant gene-swap experiments. At the time these were regarded as
cutting edge, but they have since become routine and the cornerstone of unpick-
ing the antiquity of genetic programmes governing plant development.20
The trick leading to this discovery was to engineer lines of Arabidopsis in which the ARP gene has been deleted from the genome, called knock-out lines in the language of geneticists. These lines conspicuously failed to produce normal leaves arranged in compact rosettes around the base of the plants. Unable to repress
KNOX activity and allow stem cells in the shoot meristem to transition towards leaf formation, these lines developed strangely deformed leaves with thick stems
and lacking proper leaf-blades.21 However, substitute the ARP gene from
Selaginella into these Arabidopsis lines and more-or-less normal leaves developed.22
Ancient genes, new pl Ants a 75
In essence, these experiments demonstrated that the ARP gene important for making microphyll leaves in a lycophyte functions interchangeably with its counterpart in a flowering plant for making megaphyll leaves. The magic of doing science is felt keenly with such breakthrough moments that open the door to understanding the origin of leaves themselves.
In short, then, what we have seen so far is that two lineages of land plants,
separated by more than 350 million years of evolution, possess common genetic
toolkits for building different types of leaves (microphylls and megaphylls). In
fact, the discovery has greater evolutionary significance because the common
ancestor of lycophytes and flowering plants lived before leaves originated. Our
best guess suggests this was an ancestral leaf-less twig-like plant. If correct, then the KNOX–ARP module inherited by both lineages was part of the core genetic toolkit of the earliest leafless vascular land plants. This offers us an insight into how evolution works. Lycophytes and flowering plants might have evolved different mechanisms for repressing KNOX activity to make leaves. Instead, evolution co-opted a pre-existing genetic module, which originally regulated how the naked
shoots of simple early land plants grew, and redeployed it to control the developmental programme for making leaves.
We should not be surprised to learn that the evolution of leaves involved more
genetic components than the ancient KNOX–ARP module, and new discoveries
are revealing once again the deep roots of that machinery. Leaves are programmed
to develop in size and shape in three dimensions—the width, thickness, and
length of a leaf blade are all genetically predetermined. Then there is also the
essential requirement for the developing leaf to synthesize the correct sorts of
&
nbsp; cells and tissues in the right place and at the right time. The surface of a leaf facing upwards towards the sky, for example, is specially built for intercepting sunlight.
In consequence, it has densely packed cells filled with chloroplasts for capturing light energy. The lower leaf surface is built with stomata, the microscopic gas valves that facilitate capture of carbon dioxide from the atmosphere for photosynthesis.
Leaves also require the synthesis of vascular tube networks as the structure
develops for transporting water and sugars.
Many of these complex aspects of leaf development are under the control of an
ancient family of genes (snappily named HD-ZIP genes) that originated in green algae long before land plants evolved.23 Based on their exceptional antiquity and great utility when it comes to governing plant development, HD-ZIP genes were
76 a Ancient genes, new pl Ants
probably part of the basic genetic toolkit of early land plants. In these plants, they were probably originally involved in orchestrating the formation of vascular tissue in stems. In modern lycophytes, for example, HD-ZIP genes regulate the development of vascular plumbing in their stems and microphyll leaves. This suggests that lycophytes gained the distinguishing vascular strand running through their leaves by recruiting similar genetic machinery that was already responsible for producing vascular tissues in stems. The great plant morphologist F.O. Bower would have
been pleased with this. Some 80 years ago, he proposed that microphyll leaves
arose through modification of spiny outgrowths protecting plants from attack by
herbivorous animals.24 Now the genes have spoken, it seems he may have had a
point. Only hundreds of millions of years later did HD-ZIP genes acquire new roles in specifying the identity of cells and tissues forming specialized megaphyll leaf surfaces.25 The story of HD-ZIP genes illustrates how, under the right selection pressures, pre-existing components of a charophyte algal genetic toolkit can be
redeployed for making novel structures that algae lack—leaves.
Complementing our emerging picture of the genetic machinery for building
leafy shoots are discoveries from the world of molecular genetics that explain
how roots evolved. As with leaf evolution, the fossil record establishes the framework for our thinking.26 When early terrestrial floras colonized Earth’s continental surfaces, fossils suggest they had nothing in the way of proper roots. Instead, the first land plant rooting structures were basic filamentous cells called rhizoids.
Rhizoids probably originated over 470 million years ago in freshwater algae,27
with the oldest example of a fossil land plant having simple, shallow, rhizoid-
based rooting systems dating to the early Devonian, 411 million years ago. By
analogy with the function of rhizoids in living plants, it seems reasonable to
assume they performed a similar job in these early land-based photosynthetic
pioneers, nourishing shoots by delivering flows of water and dissolved nutrients
from sediments and soils.28
By the time short herbaceous vegetation appeared later in the Devonian, spi-
dery rhizomatous systems had developed that helped stabilize river banks and
create soils by providing an organic matrix for binding everything together and
resisting erosion.29 Some of the best examples of these simple structures captured in the rock record belong to an extinct fossil plant called Asteroxylon (Figure 10).30
The appearance of the parts of Asteroxylon that were below ground is markedly at odds with that of its leafy shoots. This distinction immediately tells us that
Ancient genes, new pl Ants a 77
A
B
C
D
E
F
Figure 10 Early Devonian land plants and their rhizoid rooting systems. (A) Asteroxylon mackiei, (B) Horneophyton lignieri, (C) Nothia aphylla. D to F, Rooting structures. (D) Longitudinal section of the rooting system of A. mackiei showing dichotomized branching. (E) Transverse section of a corm bearing rhizoids of H. lignieri. (F) Transverse section of a rhizome of N. aphylla showing a ridge on the ventral surface that bears the rhizoids. Bar = 4 cm in (A), 3 cm in (B) and (C), 1 mm in (D), 0.45 mm in (E), and 1.5 mm in (F.)
Devonian plants possessed diverse genetic toolkits for generating specialized tissues composed of organized communities of cells needed for their very different
above- and below-ground parts.
In the terms of the plant development scheme described earlier for shoots,
we know that roots arise from another cluster of stem cells located at the root
78 a Ancient genes, new pl Ants
tip. The root meristem has a somewhat simpler task than its above-ground
counterpart does, in that it only has to give rise to the tissues of the main root.
In other words, a wonderfully complex plant can develop all of the intricate
tissues it needs from a handful of stem-cell clusters in shoots and roots
(Figure 9). Several different groups of vascular plants independently evolved
rooting systems, including seed plants, ferns, horsetails, and their allies, and they took on a great diversity of forms. Remarkably, a tiny fossil root tip, just 2 mm across, belonging to a giant lycophyte tree, has been discovered in Carboniferous rocks (ca. 320 million years old).31 This extraordinary fossil preserves beautiful details of cells in an actively growing root meristem and shows us that this
ancient root tip is organized much like that of living gymnosperm roots. With
the later appearance of trees like Archaeopteris, increasingly deep and elaborate rooting systems took hold through the Devonian as diverse forest ecosystems
tethered themselves to the continents. The evolution of large and complex root
structures adorned with fine root hairs must have provided these trees with
improved mechanical anchorage into soils as well as a greater capacity for
extracting water and scavenging nutrients required to support their prodigious
growth.
Thanks to insights from these sorts of fossils, the task of reconstructing the
genetic instructions early land plants might have used to make their simple root-
ing systems now becomes a little more straightforward. It amounts to unpicking
the genetic pathways controlling the development of rhizoids in mosses and
liverworts, and comparing them with those that flowering plants deploy to make
hairy roots. Oxford University-based geneticist Liam Dolan has dedicated his
scientific career to this worthy quest. He is concerned, some might say obsessed, with all things rooty—from genes to fossils. He provides a sort of complementary below-ground counterpart to Langdale’s work on shoot evolution at Oxford.32
Dolan has discovered that the pair of genes regulating the formation of multicel-
lular rhizoids in moss is the same as those making the fine single-celled root
hairs adorning the roots of the flowering plant Arabidopsis.33 The story leading to this discovery starts with what we know about where root hairs in Arabidopsis roots originate. The cells on the surface of the root are organized in alternate
rows of hair-forming cells and non-hair cells. Each root hair is a single cell,
which produces a tip-growing tubular protuberance that becomes a hair. Two
Ancient genes, new pl Ants a 79
RSL genes control the process. The two genes are very similar to each other in terms of the amino acids they code for, suggesting that one is a copy of the
other that arose by a gene duplication event. Activation of these genes pro-
duces proteins that accumulate in the hair cell and trigger additional genes to
grow a root hair. Delete either gene from Arabidopsis and the resulting plants develop nake
d, hairless roots. Mosses grow filamentous multicellular rhizoids
instead of root hairs under the guidance of a pair of genes related to those used by Arabidopsis, and if you delete them in knock-out lines, the moss develops stunted rhizoids.34
In the next step, Dolan’s group transferred one of the moss genes into Arabidopsis lines lacking their own copy of the equivalent gene; they then gain the ability to develop normal, hairy roots. Note that ‘gene-swap’ experiments from an ancient
to a more recently evolved plant lineage is the same strategy as that which
Langdale adopted in her work on the evolution of leaves, but with a different
group of plants. Langdale used a lycophyte, and Dolan used a representative of
the non-vascular bryophyte group, mosses.35 What these experiments demon-
strate is that the genetic machinery mosses use for building multicellular rhizoids can also control the development of single-celled root hairs in flowering plants.
Given these two lineages last shared a common ancestor over 400 million years
ago, the implication is that these genes were present in the earliest land plants and enabled them to grow rhizoids.
As is so often the case in science, when you delve into the details, things turn
out to be more complicated than they first appeared, and there is actually a net-
work of genes acting at different stages of root hair and rhizoid formation. Genes in the moss rhizoid network are able to act interchangeably with those of the
flowering plant root hair network,36 but there are fundamental differences in how the genes are wired together which allow the growth hormone auxin to work in
different ways.37 In moss, auxin promotes rhizoid development by acting on indi-
vidual genes independently of the others in the network. In a flowering plant,
however, it acts on one set of genes in the network, which in turn regulate the
activity of another set.38 What we have, then, are similar genes controlling the
development of root hairs and rhizoids that are wired up differently to produce
different structures. Intriguingly, auxin also promotes the growth of rhizoids in charophyte algae.
80 a Ancient genes, new pl Ants
Making Eden Page 12
