THE WORLD’S growing population depends ultimately on plants for food. They in turn depend on supplies of nutrients, especially nitrogen compounds, in the soil. In world agriculture today, the natural processes for replenishing nitrogen used up by crops are too slow to sustain the productivity needed, and the shortfall is made up by chemical fertilisers, prepared industrially by taking nitrogen from the atmosphere.
A few types of bacteria, the nitrogen-fixing bacteria, can use nitrogen from the atmosphere, but plants cannot. In 1971, Ray Dixon, at the then Agricultural Research Council Unit of Nitrogen Fixation, University of Sussex, successfully transferred the genes responsible for nitrogen fixation by Klebsiella to Escherichia coli, so creating a new genus of nitrogen-fixing bacteria. His experiment first raised the serious possibility of using genetic manipulation to generate new kinds of nitrogen-fixing crop plants. How has that dream fared? The crucial material possessed by nitrogen-fixing bacteria is the enzyme nitrogenase, which converts atmospheric nitrogen to ammonia; it is coded for by a complex of genes known as nif, and to fuel the process bacteria need energy. Some, such as the cyanobacteria, use solar energy by combining nitrogen fixation with photosynthesis. Others, such as the soil bacteria Azotobacter and Klebsiella, use organic food; yet others enter into symbiotic relationships with plants, receiving some of the products of the plant’s photosynthesis and in return donating nitrogen to the plant.
Ecologically and agriculturally, the most important nitrogen-fixing systems are symbioses. Familiar examples are the partnerships between leguminous plants (such as peas, lupins and clover) and a few genera of soil bacteria collectively known as the rhizobia. In these symbioses, the rhizobia infect the roots of the host plant, causing nodules to form. Within the nodules, the bacteria multiply and then become dormant, though continuing to fix nitrogen. Each group of leguminous plants has its own species of rhizobium and is not colonised by other types. So far, scientists have discovered only one non-legume, the tropical plant Parasponia, which is colonised by a rhizobium.
Advertisement
Throughout the world, the use of chemical nitrogen fertiliser has tended to displace traditional biologically based agriculture; crop rotations, intercropping and nitrogen-fixing green manures such as clover have fallen into disuse, sometimes even forgotten. Yet they often have local advantages in cost, accessibility and environmental impact. So, especially in developing countries, scientists and technologists are reintroducing, developing and modifying such procedures: introducing new self-fertilising crops such as soya beans or peanuts, planting nitrogen-fixing trees such as Leucaena or alder between crops, extending the use of Azolla or cyanobacteria as green manure, reverting to crop rotations, composting and so on. A lot can still be done to spread and improve existing technologies, often without any genetic manipulation beyond old-fashioned selection and breeding of good varieties of plant and microbe, but their effects are likely to be marginal at best. To ensure food supplies for the burgeoning generations of the next millennium more dramatic advances are needed.
Dixon’s success caught the imagination of an interested public in ways which were sometimes disconcerting. Many people thought that one had but to pop some nif genes into wheat or rice and, bingo, world starvation would end. Even some scientists were naive: soon after our first report was published, I remember, as assistant director of the Sussex unit, having to compose a somewhat embarrassed refusal to a professor of botany who wished to occupy our precious laboratory space for three weeks – the time he thought it would take to put nif into a plant.
Well, we were all a bit starry-eyed, but we always knew that the problems of manipulating nitrogen fixation would be complicated, and that researchers would have to forget about practical applications for a long time, while they ferreted into the basic science underlying the process. Even so, we had no real idea of the wealth of science that would be – would have to be – uncovered before we could even begin to think of putting such manipulations into practice. The new discoveries that nitrogen fixation research has led to, in genetics, biochemistry, physiology and chemistry, make a fascinating story. Eighteen years after Dixon’s experiment, scientists are now in a position to suggest at least three feasible and rational approaches to the long-term upgrading of biological nitrogen fixation.
The first two are the result of a burst of intensive basic research on the genetics and biochemistry of the legume symbiosis during the early and mid-1980s. With insight gained from this research, it is now possible to modify the symbiotic bacteria, the rhizobia, by manipulating the genes in their chromosomes. For example, the various strains of the rhizobium that colonises the soya bean differ in their efficiency at converting nitrogen from the atmosphere, and therefore in their ability to support a good crop of beans.
Harold Evans and his group at Oregon State University have worked out one reason for the difference. Nitrogenase always forms hydrogen as a by-product. Some strains of rhizobia possess an enzyme system called hydrogenase (coded for by a complex of genes known as hup) which traps the hydrogen and recycles it, to recover energy among other benefits. The less efficient strains of the soya bean rhizobium lack hup genes and just let the hydrogen escape, thereby wasting energy. (It is a surprising truth that, from many a soya-bean plantation, there is a gentle waft of hydrogen into the atmosphere.) Evans’s group has cloned the hup genes and transferred them into strains of rhizobia which do not have them; tests which allow for the effects of weather, soil, competitors, plant and bacterial genotype show that possession of hup makes the symbiosis measurably, though not spectacularly, more efficient.
On another tack, the ‘wild’ rhizobia that live in soil are often highly competitive – they colonise plants much more readily than do ‘tame’ rhizobia from laboratory collections. Unfortunately, as far as crop plants are concerned, the wild strains are also usually poor nitrogen fixers: ‘ineffective’ in the technical jargon. Competitiveness arises from a complex of properties and it is difficult to identify all the genes involved. However, it is now possible, at least in principle, to bypass that particular problem: by selecting the most competitive of the wild strains, and then manipulating into them genes that will enhance their effectiveness. Other research has pointed the way to other manipulations of agriculturally desirable properties that can be performed on rhizobia. In the future, no doubt comparable operations will be feasible with other classes of nitrogen-fixing bacteria, such as Azospirillum, Bacillus, Klebsiella and Azotobacter, associated with certain grass and cereal roots.
A second, more ambitious, outcome of the research into legume symbiosis is that scientists can consider creating symbioses between existing bacterial colonisers and new plants. Back in 1975, Tom LaRue, then of the National Research Council’s Regional Prairie Laboratory in Saskatoon, Canada, gave a talk at a symposium on nitrogen fixation at which he showed a faked drawing, as if from an old botany flora, of a barley plant with nodules on its roots: this was Triticum leguminosum, a species, in his words, so far found only in research grant applications. There was much laughter, not all of it entirely comfortable. For his joke epitomised the ideas of many research workers.
In the spring of last year, Ted Cocking’s team at the University of Nottingham reported an important advance towards nodulated cereals. They treated the roots of rice seedlings for a short time with enzymes, such as one from the gut of a snail, which digest cell walls. This removed the walls of a few of the living, growing cells at the tip of the root, making them susceptible to infection by rhizobia, which the cell walls would normally exclude. Two-day-old rice seedlings, so treated, were inoculated with rhizobia and returned to a growth chamber. The seedlings continued to grow and, in about a month, the roots had formed nodule-like structures which contained rhizobia. They have now achieved the same with wheat. The nodules, only about 3 millimetres across, have close similarities to legume nodules.
The report on the rice nodules stated that ‘nitrogenase activity . . . was the limit of sensitivity of the assay procedure’: a pessimist would have said none was detected. But it is early days, and there are many reasons why such activity might have been suppressed in these first experiments. But even if researchers can achieve nitrogenase activity in due course, there are still all sorts of problems to solve on the way to a useful nodulating cereal. Among these are how to stabilise the association and make it hereditary, how to regulate the fixation process, how to get the plant to make substances essential to the functioning of the nodules, such as leghaemoglobin and a variety of seemingly essential substances called nodulins (one of which, for example, helps the plant to recognise its bacterial symbiont). But the Nottingham group has rekindled serious interest in creating hitherto unknown symbioses between cereals and high-efficiency nitrogen-fixing bacteria.
The third, even more radical, approach is to try to endow plants with their own nitrogen-fixing capacity. Yet, if symbioses can work so well, you may ask, why bother? In terms of pure science, the answer is simple: evolution has not so far produced a nitrogen-fixing plant and we want to know why. But there are good practical reasons for trying, too.
Apart from the obvious one of increasing the range of plant systems that can fix nitrogen, a more subtle reason concerns the competitiveness of the ‘wild’ rhizobia. These have no special interest in being highly effective; they compete their way into their ecological niche and then fix just enough nitrogen to get by. Consequently, a serious problem in using biological nitrogen fixation in agriculture is how to ensure that good, effective strains of symbiotic microbe (which farmers can buy as commercial preparations) colonise the crop rather than their relatively ineffective but highly competitive wild cousins. The technology would be greatly eased if scientists could relieve plants entirely of their dependence on the vagaries of microbes.
Doing away with microbes
There are two ways in which one might teach plants to fix their own nitrogen. The first is to create a nitrogen-fixing organelle inside the plant’s cells. Biologists now believe that one such organelle, the chloroplast, is descended from a microbial symbiont, probably a photosynthetic cyanobac terium, which originally lived inside the plant and has now come largely under the control of the plant’s genes. Could one mimic chloroplast evolution to establish a ‘diazoplast’: a comparable symbiont which could fix nitrogen?
In 1984, Alan Paton and his colleague Sriyani Aloysius, of the University of Aberdeen, published evidence that opened up an exciting vista. They found that plant cells can take up specialised forms of bacteria called L-forms, and that L-forms remain alive and active within the cell cytoplasm, doing no harm and even multiplying along with the plant cell. Some nitrogen-fixing bacteria, such as strains of Azotobacter, can assume L-forms. If one could arrange a balanced inter dependence between cell and bacterium, and also transmission down the generations of plant, one would have something approaching a ‘diazoplast’. The problems of doing this are formidable, however, and Paton’s group is concentrating for the moment on simpler associations, especially on defining the interactions between plant and the L-forms of bacteria more precisely.
The second way, perhaps the earliest dream, is to take the nif genes from an appropriate bacterium and engineer them, together with their regulatory apparatus, into plant cells. As with the novel nodulated plants, basic research has made scientists aware of several features of biological nitrogen fixation which they must enable a plant to overcome if it is to emulate a nitrogen-fixing microbe.
Finding the tools for the job
Some of these obstacles seem, at present, relatively minor. For example, nitrogenases contain atoms of iron and most have another metal as well, usually molybdenum but sometimes vanadium. So nitrogen-fixing plants would need the genetic and biochemical machinery for taking up and processing these metal atoms. A second concern is the amount of energy consumed in nitrogen fixation. But in biological terms it is nothing special – similar to that normally expended when a plant assimilates nitrate. Another problem which will loom large is that of oxygen exclusion: oxygen deactivates the nitrogenase proteins, and nitrogen-fixing bacteria have developed a variety of ways to keep the oxygen of air away from their nitrogenases. Above all, nitrogen-fixing plants would have to possess nif genes, and be able to use and regulate them.
In 1988, Alf Puhler and his team at the University of Bielefeld, in West Germany, identified the complete sequence of DNA in the nif cluster of Klebsiella pneumoniae, establishing the presence of a surprising 20 genes, all in a row . Most nitrogen-fixers do not have their nif genes so conveniently linked, so K. pneumoniae is a promising source of DNA for such a manipulation. Indeed, this is the bacterium that Dixon and others have used to make several new types of nitrogen-fixing bacteria. Molecular biologists have now worked out how nif is regulated in Klebsiella in sufficient detail for a rational scheme for transferring these nitrogen-fixing genes into plant cells to be proposed.
A major problem, however, is that bacterial genes do not make sense to plants (nor to animals for that matter – my pet dream is of a nitrogen-fixing goat that could convert all our junk paper and cellulose waste direct to meat). Though higher organisms and bacteria use essentially the same genetic code, they have different reading systems: the on-off switches (promoters), the numbers of genes read at a time and the ways of handling the genetic message all differ.
One approach would be to clone bacterial genes and to link them to promoters which the plant can recognise. Then one could introduce the manipulated gene(s) into plant cells so that they become part of the plant’s own hereditary material. Research groups in the US, Britain, Belgium and Germany showed in the early 1980s that this sort of gene transfer can be done. They successfully placed in tobacco plants a bacterial gene which specifies resistance to the antibiotic kanamycin, to which both plants and bacteria are sensitive. The novel gene was introduced into the plant by means of a bacterium carrying a plasmid, a small loop of DNA, that naturally integrates into plant DNA, causing benign tumours. The plasmid’s promoters are ones that tobacco plants recognise. The plants read the gene correctly, and not only did they resist kanamycin but so did their progeny. But to try to introduce in this manner 20 genes, which must produce their various products in subtly linked ways, would be laborious.
At the Nitrogen Fixation Laboratory at the University of Sussex, Mandy Dowson Day is now working on this problem. Her long-term plan is to put only the genes essential for regulation under plant control, and to let these genes regulate the rest of the nif cluster in the ‘bacterial’ way. This is plausible because the plant chloroplast retains some genetic relics of its origin as a bacterium and so ought to be a promising target for a nif gene package. The oxygen produced by that organelle would not be welcome, but a mutated chloroplast, or a related organelle which does not photosynthesise, might serve. It may also be necessary to introduce the rpoN gene, which regulates the nif genes .
Decisions on such issues, however, are premature until Day knows whether it is even possible for a plant cell or organelle to make nitrogenase proteins. As a first step, Day plans to introduce into plant cells combinations of just two nif genes, chosen because together they can produce one of the proteins which make up nitrogenase. Such experiments should tell her whether it is possible for the plant to make the proteins.
The practical benefits that may arise from novel nitrogen-fixing systems are still in the pipeline. But if scientists can create autonomous nitrogen-fixing crop plants and bring them into use, what risks and benefits might ensue? The environmental problems of nitrate pollution of water will become worse as inputs of nitrogen increase, be they biological or chemical. That is part of the fallout of the population explosion. Whether biological nitrogen will be more manageable than chemical nitrogen, as many expect, is largely a matter of guesswork (and prejudice). But romantic fears of the planet being colonised by rampant nitrogen-fixing weeds, formed by inadvertent transfer of nif genes among plants, may be dismissed. As with all such gene transfers precautions are necessary, but even the worst scenario is bland. There would be no rampant colonisation because if nitrogen is in abundant supply, some other nutrient (most often phosphorus) will limit growth instead. This is what happens with clover, soya beans, alder and the rest of the present nitrogen-fixing systems.
In terms of economics, the market for nitrogen-fixing crop plants would probably be small in the developed world because the direct benefits to farmers would be at best marginal. While there would be a modest saving in the costs of fertiliser plus a ‘greener’ image in terms of run-off and nitrate pollution, farmers would have to balance these two virtues against the inconvenience of changing techniques, and a small decrease in optimum, though not necessarily actual, yields of crops. The fertiliser industry need not be seriously concerned. While sales for industrial nitrogen fertiliser would not increase as fast as they otherwise might, and in the long term might decline, sales of other chemical fertilisers would increase in proportion to world demand for food.
It is in developing countries that the economic benefits would be felt. For the Third World, nitrogen-fixing plants could mean that the substantial cost of importing chemical nitrogen fertilisers would decline and perhaps vanish. The effect would be comparable to the Green Revolution, without its financial penalty. Farmers would earn a decent living and governments could deploy their hard currency in other ways.
And what of social gains in such countries? Would their peoples benefit? Would malnutrition and starvation then become things of the past? I pass.
* * *
Why bother with nitrogen fixation?
IN JULY 1988 the world’s population topped 5 billion. Assuming global catastrophe – war, famine or plague – does not intervene, that number must double by the early decades of the next century, because the children who will be the parents of the next few billion are with us already, and they will breed their offspring before the majority of us die. Setting all other problems of the population explosion aside, and they are terrible, these people must all be fed or they will fight.
Food production depends primarily on plants finding adequate amounts of nitrogen compounds in the soil. Nitrogen fixation comes into the question because, though farmers can supply more than 80 per cent of the necessary nitrogen compounds by recycling nitrogen already in the biosphere (as composts, organic manures and such), there is wastage. The nitrogen cycle shows that, on a global scale, soil nitrogen has at present to be topped up with about 12 per cent new nitrogen each year.
The only important source of such nitrogen is the atmosphere. Nitrogen fixation is the conversion of atmospheric nitrogen into forms that plants can use, a conversion principally performed by nitrogen-fixing bacteria or by industrial chemists using the Haber process. Industry now provides the biosphere with something approaching 30 per cent of its new nitrogen, as 50 to 60 million tonnes of fertiliser a year.
Such fertiliser is cheap in agricultural terms, yet as population growth accelerates, formidable problems arise. There are the energy and staffing costs of building and running new Haber installations. The fertiliser needs to be packaged and transported from where it is made to where it is needed.
So far, food production has kept up with population growth, and this is due largely to the use of chemical nitrogen-fertiliser. For example: India has become a net exporter of food through the use of chemical fertiliser, though its import costs the country dear in hard currency. The Haber process is responsible for keeping more than a third of the world’s population alive, and mostly fed, today.
Over the next few decades the input of new nitrogen into the world’s agricultural soils must increase in proportion to the increasing population. Such an increased input will have worrisome effects on the environment, and there is an argument about whether the new nitrogen should be of chemical or biological origin. It is largely irrelevant. Increased use of chemical nitrogen fertiliser is inescapable; a parallel return to greater exploitation of biological nitrogen-fixing systems, still responsible for providing more than 60 per cent of the new nitrogen, seems common sense.
* * *
2: How the nif genes of Klebsiella work
THE chromosome of Klebsiella pneumoniae is a circle of coiled DNA. All the 20 genes needed to make nitrogenase (the nitrogen-fixing enzyme system) and to regulate its manufacture, are located together in a single cluster. They are regulated in sub-clusters, or operons, groups of one to six genes which are transcribed together.
Scientists have now worked out the biological functions of about two-thirds of the nif genes, which they identify by letters. Thus H, D and K code for the actual proteins of nitrogenase; several of the others (Q, B, V, N, E, M) code for products which process those proteins, getting molybdenum and iron into the right place; F and J specify proteins which the cell must have for nitrogenase to work. Two are regulator genes: A codes for a protein which tells the cell to use the other operons; while L’s protein countermands this instruction.
For nifA to initiate transcription of the other nif operons, another gene product is necessary, that of rpoN (sometimes designated ntrA). Both nifA and L are themselves regulated by so-called ntr genes, which are part of the cell’s machinery for sensing whether the environment is running short of fixed nitrogen. The nif genes are unique to nitrogen-fixers; ntr and rpo genes are common to many types of bacteria.
Next week Professor Jeff Leigh writes about the chemistry of nitrogen fixation.