Philologists studying the history and development of languages envy us biologists. They face the problem of exchange of words and grammatical structures between even unrelated languages, and have to sort out the resulting confusion. As genes tend to remain confined to the species in which they originate, and their descendants, we don’t have to worry about this sort of thing.
Or do we? Genomes may be less ‘watertight’ than we have supposed. Genetic engineers use bacteria and viruses to transfer genes between very different organisms and the genes frequently function perfectly in their new homes. What we can do in a few years, nature may have done countless times in the past three billion.
Organisms, like languages, remain classifiable, but some groups are more difficult to sort than others. For example, in the past 10 years four highly respected plant taxonomists, Arthur Cronquist, Armen Takhtajan, Robert Thorne and Rolf Dahlgren, have each drawn up systems for the Angiosperms, the flowering plants. They largely agree over the major families, but there is much squabbling over which families belong to which higher-level taxa, such as orders and superorders.
Advertisement
Ever since scientists accepted the theory of evolution, most taxonomy has attempted to reflect phylogeny, the pattern of descent of organisms. We have assumed that genomes change only internally – by mutation and the transposition and duplication of genes, for example – with minimal input from outside. If this is so, we can construct a ‘one true tree of life’. We place on adjacent ‘twigs’ organisms that share a recent common ancestor, such as chimps and ourselves. On remote ‘branches’ we place those such as plants and insects whose common ancestor lived very long ago indeed. It should all be very neat and unambiguous.
For many groups, such as the large vertebrates we tend to study most, this process of sorting is fairly easy. Large vertebrates have ‘closed’ genomes; their evolution is largely of internal origin. In such complex, finely tuned organisms, the introduction of new genetic material is more likely to be disruptive, so that the altered genetic constitution is not passed on. Also, early in their development, animals tend to set aside the cells that will produce eggs and sperm. As a result, most imported genes enter ordinary body cells and cannot be passed on. This process may have evolved because animals that kept extraneous genes out of the germ line were more successful.
In other organisms, almost any dividing cell may give rise to reproductive cells – as in plants. Intact plants can be grown from dividing cells taken even from root tissue. As a result, imported genes will be passed to their offspring much more often. A genome which frequently accepts material from outside and remains viable, so passing the material to descendants, might be called an ‘open’ genome.
Biologists already accept the openness of the genome of prokaryotes – organisms having cells with no nucleus such as bacteria. These organisms frequently exchange genetic material. So resistance to antibiotics, for instance, can appear in bacteria never before exposed to these chemicals. We now classify prokaryotes by cell-wall structure and ribosomal RNA, while accepting that their evolutionary history is more like a network than a tree. Perhaps we should also reconsider the attempt to construct phylogenetic trees for some more complex organisms and think of their evolution in terms of networks too.
The idea is not fantastic; researchers have already discovered that genes can be transferred between eukaryotic organisms – those with complex nucleated cells – such as protozoans, algae, fungi, plants and animals. Three mechanisms are known: hybridisation, so-called lateral transfer by microorganisms and the incorporation of symbiotic organisms.
Hybridisation is the fusion of sex cells, or gametes, from different species. It happens frequently in all kinds of organisms, but is particularly common in temperate plants and orchids. Normally, hybridisation yields offspring that are infertile because the chromosomes, the structures that house an organism’s genes, cannot be matched in pairs. But in plants gametes that have failed to halve their chromosome numbers can fuse to form new species. With a doubled chromosome set, the plant remains fertile. The salt-marsh grass Spartina townsendii recently originated in this way. Such things tend not to happen in animals, in which such changes of chromosome number are usually lethal.
More frequent than the creation of new species is the exchange of genes through hybrid backcrossing with their parents. A group of species can form a ‘hybridisation chain’ in which each may hybridise with its immediate neighbours but not with others further along the chain. In a group of recent origin which has suffered few extinctions, species at the extreme ends of the chain may look quite different, yet genes are being passed between these via the intermediate species. Bill Burtt at the Royal Botanic Garden, Edinburgh, has long believed that angiosperms are difficult to classify because ancestral forms were able to hybridise across larger morphological gaps than present ones can. Hybridisation chains, more active in new groups than old ones, provide a mechanism. Extinctions break such chains, stopping gene exchange. And any population which becomes geographically isolated from the other members of the chain will also break from it.
Hybridisation can involve only close relatives, but other mechanisms can effect gene exchange between quite unrelated organisms. Although such events are much rarer, they have more dramatic effects when they occur. Bacteria and viruses, of the kinds used by genetic engineers, mediate the ‘lateral transfer’ of genes. These naturally occurring microorganisms temporarily splice their genomes into those of their hosts. When they move on, they occasionally incorporate adjacent host genes which they can subsequently transfer to other hosts.
We tend to think of such parasites only as pathogens, bringing disease and death to their hosts. But if a parasite is to survive as a species, its host must live long enough for the parasite to reproduce – and if the host species become extinct, so do their parasites. Well-adapted parasites – the majority – do minimal harm to their hosts. As they trouble us little, we study them little. We concentrate on the ill-adapted pathogenic minority, which have only recently begun infecting us or our domesticated animals and plants.
Low pathogenicity in those parasites that insert their DNA into that of their hosts suggests they tend not to insert their own genetic material too close to genes vital to the survival of the host cell. Close proximity of inserted genes to those of the host concerned with DNA replication or energy production might prove fatal to the host, and thus to the parasites.
On the other hand, the parasite may actually benefit by insertions which interfere with the host’s ‘morphogenetic’ genes, those involved in development, or by insertions near genes that are responsible for making poisons. By affecting morphogenetic genes, the parasite may cause galls to form in which it may then proliferate. If it activates poison-producing genes it will protect itself from predation.
As a result of these two opposing tendencies, most cases of lateral transfer are likely to involve genes not essential to the immediate survival of individual cells – though they may be highly influential to the host organism as a whole. However, most phylogenetic trees based on molecular data use only that class of genes that are vital to the immediate survival of individual cells, for the excellent reason that these are found in all organisms and do the same jobs. Those small differences that do exist between species tend to be found in parts of the DNA that are functionally ‘neutral’, and are due to point mutations that accumulate at similar rates in diverging lineages. These changes may therefore give a misleadingly clear picture of phylogeny.
Lateral transfer is most likely between related organisms because these tend to share the same parasites. We may find ‘lateral transfer chains’ that are analogous to hybridisation chains. As each strain of microorganism is likely to infect several related species, or even a whole genus or family, these lateral transfer chains may be able to extend across much larger taxonomic ranges than hybridisation chains. The links would not be hybrids, but microorganisms capable of infecting adjacent groups in the chain, and of transferring genes between them. Because of the large numbers of species involved, these chains would be difficult to break. Mass extinctions or long periods of time are generally needed to do so.
If lateral transfer were to occur between very different organisms, cases should be easy to discover because they would stand out. For instance, the presence of the same rare gene in both humans and apes arouses little interest; but should someone then discover that termites but no other insects also have this rare gene, we would take notice. This type of transfer may be caused by parasites that are not too choosy about which species they inhabit. In addition, bacteria exchange genes among themselves, so that one bacterium infecting an animal may pick up an animal gene and later transfer it to a plant parasite. This parasite then inserts the animal gene into the plant acting as its host. It has been suggested that haemoglobin, the molecule that carries oxygen in vertebrate blood, found its way into the roots of leguminous plants in this way.
Some 20 years ago, Frits Warmolt Went at the Desert Research Institute, Nevada, invoked lateral transfer to explain why various unrelated plants in New Zealand grow in similar ways. In 1990 Jared Diamond of the University of California at San Diego suggested in Nature that this phenomenon was an adaptation by the plants to grazing patterns of the recently extinct flightless birds, the moas. Though this explanation does not rule out the possibility that the plants acquired the genes responsible by lateral transfer, it emphasises that laterally transferred genes, like any others, will remain in populations only if they are adaptive or at least neutral.
When the process of lateral transfer is invoked as an evolutionary mechanism, most biologists oppose it, claiming that it is unnecessarily exotic. Yet the ubiquity of the organisms capable of causing it suggests it should be a common happening. The controversy resembles that surrounding theories about the biological consequences of asteroids hitting the Earth – it happens, but how important is it?
The incorporation of symbiotic organisms is almost certainly the rarest process of gene transfer because it involves taking on board whole genomes rather than just the odd gene, suggesting its effects are likely to be profound. It was first invoked by Lynn Margulis of Boston University in the 1960s to explain why the mitochondria and chloroplasts of eukaryotes resemble bacteria. Her explanation is now widely accepted.
Many plants have intimate bacterial and fungal symbiotes and at least one case is known where these are passed on in the seeds. The plant concerned is Psychotria bacteriophila, a coffee relative with a nitrogen-fixing bacterial partner. Flowering plants share with bacteria and fungi an immense chemical inventiveness. Cases are known where the same unusual chemical is made by a plant group and also a microorganism. An example is betalain, the red pigment of beetroot and other plants in a group of related families that include cacti and mesembryanthemums. They are also found in some basidiomycete fungi, relatives of the mushrooms. Molecular biology, by finding and sequencing the genes responsible, should help us to decide how this happened. If the incorporation of symbiotes was involved, we should expect to find other fungal genes in these plants.
At what stages in evolution may these processes have been important in eukaryotes? Fossil evidence suggests that many new groups arise when there are ecological vacancies to fill, as a result of mass extinctions, say, or the colonisation of new terrestrial habitats. The first occupiers of a vacant ‘niche’ will not have the sophisticated adaptations of later ones, but competition between different organisms for the same niche generates pressure for refinements. Groups with open genomes would have a great advantage in this situation because an exchange of genes would give to several species in a taxon the same advantage that sex gives to individual members of a species – it allows the testing of combinations of genes that originated in separate organisms. Such a group of species would be a more innovative taxon than one with closed genomes and its members able to adapt faster. As adaptations become more specialised, further exchange of genes between species is more likely to be deleterious, and pressure for genome closure will return.
So we would expect to find open genomes in rapidly expanding groups of recent origin, and in those whose members have experienced relatively few extinctions – because extinction would tend to break the hybridisation and lateral transfer chains responsible for most gene exchange. We would expect to meet problems in trying to classify such groups, especially if we rely on chemical and morphological criteria, at the expense of the fossil record and geographical information.
Among plants we find these characteristics most marked in the angiosperms or flowering plants. This is an enormously disparate group. They range from tiny duckweeds and delicate herbs to cacti and mighty trees, with reproductive structures – their flowers and fruits – of immense variety. By contrast other major taxa, such as conifers and the species-rich ferns, are rather uniform in structure.
Botanists divide angiosperms into some 400 families. Most of these families are distinct enough, but there is dispute over their interrelations and phylogeny. Even within some families, there are problems over subdivision.
The problem can be understood if we imagine only a few of the modern families surviving a hypothetical future mass extinction. Suppose the survivors were a small number of the most successful and distinct groups such as grasses, orchids, composites and legumes. Future botanists would then be able to separate their descendants cleanly, with little dispute. Were they then to stumble on a time-capsule filled with a complete range of seeds from our own time, their reactions to what grew from them might be similar to those of present-day zoologists faced with the remarkable diversity of animals we now know to have been alive 600 million years ago (see ‘Soft-centred fossils’, New ÐÓ°ÉÔ´´, 11 August 1990).
In general, the more characters we study, the clearer any phylogenetic pattern behind them should become. But with angiosperms, things are seldom as simple as this. A taxonomist who rates morphology highly may place the cabbage family Cruciferae and the poppy family Papaveraceae close together. One who prefers to rely on chemistry will say they are unrelated and instead allies the Cruciferae with the otherwise very dissimilar nasturtium family Tropaeolaceae. What are we to make of this?
As Cronquist has pointed out, most evolutionary theorists have been zoologists and have given relatively little thought to the very different problems of plants. Nevertheless botanists have long been aware of the difficulties and have offered many explanations for them. For example, it seems certain that morphological details are more vital to an animal’s survival than that of a plant. Change the shape of a leaf or petal and it may affect a plant’s survival little. Make corresponding changes to animal parts such as legs or wings, and the result is likely to be fatal. There is almost certainly much neutral evolution in plants, allowing more rapid changes. This has been used to explain the disparity in angiosperms – but then why are other plant groups so much more uniform?
The way in which plants interact with animals is more likely to be due to selection than to neutral evolution. Plants defend themselves with spines and poisons. Many produce flowers which are pollinated by animals, so must succeed in attracting them; and many rely on animals to disperse their seeds. But if we use such characteristics to classify plants we still find similar characters appearing in groups we believe, on other grounds, to be unrelated.
SAME BUT DIFFERENT
Parallel evolution is often invoked to explain this phenomenon – related groups are assumed to have produced similar mutations after their divergence. If similar characters appear in distantly related taxa, we invoke convergent evolution – similar solutions to similar problems – which produced, for example, the streamlined shape of fish and dolphins. If this fails to stand up, we say the taxa have retained a trait possessed by a common ancestor, the rest of whose descendants have lost it.
While such explanations are obviously often correct, when met too often they begin to look like a set of props being put under the branches of an old and moribund tree – in this case, a phylogenetic one? If a philologist found a similar pattern in a group of languages, she would invoke extensive borrowings. We may begin to suspect that the plants or their not-too-remote ancestors had open genomes, and that borrowings may also be occurring in evolution.
Why might angiosperms have been especially prone to such borrowings? Low extinction rates and the absence of mass extinction events would tend to encourage genomic openness. Angiosperms came through the only major mass extinction since their origin – the so-called K/T event that ended the Cretaceous 65 million years ago – relatively unscathed, losing only 75 per cent of their species.
This may seem devastating, yet it was a lot less than many taxa suffered – none of the dinosaurs survived, for example. Among flowering plants, most genera and almost all families survived. If the K/T event was caused by a major asteroid impact, it will have produced a severe but quite short period of cold (or heat, according to the authority consulted) combined with total darkness. Those surviving were probably those capable of weathering such conditions through dormancy. Large animals, which tend to deal with adversity by migrating, died out, because the whole planet was affected. Small animals, which hibernate or produce resistant stages, survived. Freshwater biotas, which need a capacity for dormancy because even in normal times ponds often dry out, sail through mass extinctions. They include many ancient and primitive groups that have died out elsewhere – dragonflies, lungfish and branchiopod Crustacea such as fairy shrimps and water fleas.
Angiosperms often produce seeds in vast numbers, capable of long survival in the soil till conditions improve. Others have bulbs or tubers capable of prolonged dormancy. Some are deciduous, which constitutes another form of dormancy. Most gardeners will tell you of tender plants, which, though supposedly without such a capacity, sprouted again after a cold winter. The fact that tropical rainforest plants seem to have survived the K/T event as well as any others, and that they show the same difficulties in classification as temperate ones, suggests they share an immunity to extinction and may have genomes just as open.
The fossil history of angiosperms is patchy. The most taxonomically important parts, the flowers, rarely fossilise. We rely largely on leaves and pollen. What is found tells of increasing diversity over the past 100 million years and relatively little extinction of major taxa. The wide distribution of even ‘advanced’ families such as Campanulaceae that lack plausible means of long-distance dispersal, combined with our knowledge of plate tectonics, implies that many of these are older than fossils suggest and that diversification of the group may have occurred earlier and more rapidly than is often supposed. A rich and diverse group called Bennettitales which had some intriguing resemblances to angiosperms had died out before the K/T event, and there is evidence that early ferns and conifers were more diverse than later ones and so were more prone to extinction than angiosperms, for whatever reasons.
If angiosperm genomes have remained open, this may simply be due to low extinction rates. Sufficient closely related subtaxa survived for long enough for hybridisation and lateral transfer chains to remain active for a long time. Ferns and conifers were substantially decreased by extinctions early in their history, so their remaining taxa were probably too far apart for the gene-exchange mechanisms to remain active.
To test these undeniably speculative ideas, we must study plant genomes in great depth. Plants often defend themselves with an impressive chemical arsenal, and we should look for similar unusual chemicals appearing in apparently unrelated groups, to see whether the methods of synthesis are the same in each. We already know of the presence of erucic acid and mustard oils in both Cruciferae and Tropaeolaceae, for example. Again, the alkaloid sparteine is made, using the same pathway, both by the pea-related lupins and the papaveraceous Chelidonium, the greater celandine. Using such examples, we may find, sequence and compare the genes responsible. If the transfer of the sizable blocks of genes needed to make such complex molecules is at all frequent, that of smaller units may be assumed to be commonplace.
If we are trying to sort out an open-genome group, we should return to our philologist with her languages, and study her methods. Philologists use words that tend not to be borrowed from other languages to elucidate relationships. They compare across languages words for things universal to human experience – such as those for Sun, body parts and kinship terms.
Can biologists find their equivalent? We may indeed have them, in the ‘housekeeping’ genes for DNA replication and energy production that all cells must have, and which we have argued will not often be transferred. Even more helpful may be the genes of mitochondria and chloroplasts – for although these have their own DNA and DNA replication systems, they appear to lack viruses capable of infecting them, and therefore of transferring genes in or out of them. If this is so, they will have completely closed genomes with respect to other species, though they may exchange DNA with their hosts’ nuclei.
Of even greater interest may be the morphogenetic ‘homeobox’ genes which guide development in animals and which are involved in the development of animal embryos and which have only recently been found in plants. These, like housekeeping genes, must be present in all plants, but as they are not essential to the survival of individual cells, they may be subject to lateral transfer. They might be useful as a means of measuring its frequency.
However neat a phylogenetic tree is produced by studying housekeeping and mitochondrial genes, this may not be the whole story. We must also take into account the network – the transferred genes. Like linguists, we may have to accept that for many taxa, although the framework has a well-defined origin, many of the detailed features came from elsewhere. We may find that we have to reconsider plans to insert genes for herbicide resistance into crop plants, lest these defeat their purpose by transferring to weeds. We would also gain a new view of life. If the entire world’s gene-pool becomes accessible to all through the processes described here, then microorganisms we have thought of as nothing but hateful pathogens become vital instruments in evolution.
Chris Heron is a science graduate who has worked in horticulture for 12 years and is now at the University Botanic Gardens, Cambridge.