HUMANS are notoriously good at wiping species off the face of the planet. In 100,000 years we鈥檝e driven untold numbers of mammals, birds, plants and reptiles to extinction. And all that can be said in our defence is that, smallpox excepted, there has not been malice aforethought.
That could all be about to change. According to Austin Burt, an evolutionary biologist at Imperial College in London, a new genetic technology may make it possible to engineer extinctions. The technology is targeted at species that, to many people, would be better extinct: insects that carry fatal diseases. Wipe them out and you could save millions of people from malaria and hundreds of thousand more from dengue fever, sleeping sickness and the like. 鈥淚鈥檓 not convinced it should be done,鈥� says Burt. 鈥淏ut on paper, at least, it stands a good chance of working.鈥�
Burt鈥檚 ideas, detailed this week in an article in Proceedings of the Royal Society B (DOI: 10.1098/rspb.2002.2319), are a deadly twist on the long-standing idea of using genetic engineering to neutralise disease-carrying insects. But instead of tweaking the mosquito to make it less able to transmit malaria 鈥� the aim of most of the work in this fast-growing field 鈥� Burt鈥檚 work is more apocalyptic.
Advertisement
His proposal hinges on the strange properties of DNA sequences called homing endonuclease genes. HEGs have found a cunning way of evading the normal rules of heredity, exploiting a loophole to get extra copies of themselves into the next generation. This 鈥渟elfish gene鈥� behaviour means HEGs are molecular parasites 鈥� which adds a touch of poetic justice to the idea of using them to wipe out insects that make their living by sucking blood and spreading disease.
In organisms that have paired chromosomes, a gene present on only one chromosome normally gets passed on to exactly half the organism鈥檚 offspring. Unless, that is, the gene is an HEG. An HEG on one chromosome in a cell can use that cell鈥檚 repair mechanisms to get itself copied onto the chromosome鈥檚 partner. If the cell in question is a cell that makes eggs or sperm, this copying means that all the eggs or sperm will contain a copy of the HEG 鈥� and so all the offspring get a copy. In this way HEGs can spread through a population very quickly indeed.
HEGs pull this off by sabotaging their host鈥檚 DNA in a way that can only be repaired through the creation of a new HEG (see Diagram). Each HEG is the gene for a homing endonuclease, an enzyme that recognises a specific sequence of double-stranded DNA and then cuts the chromosome at that point. When the enzyme has done its job, the cell tries to repair the damaged chromosome using the equivalent sequence on the other chromosome as a template. And it turns out that on the other chromosome that sequence is split in two, with the HEG in the middle. In repairing the damage, the cell stitches a new copy of the HEG into the damaged DNA.
The key to using HEGs for pest control is to make sure they achieve their selfish ends at the expense of a gene responsible for some vital aspect of the host鈥檚 health. This is not something that comes naturally to HEGs. Genes that harm their hosts tend to have poor long-term prospects, and the HEGs found so far in yeast, plants and lichens seem, like most sensible parasites, to do very little harm. But an HEG shaped by deliberate design rather than evolution could be nasty enough to wipe out a species. The trick is to make the HEG recognise a DNA sequence in the middle of a vital gene, and so destroy its function.
Imagine a gene that is necessary for a fertilised egg to develop into an adult mosquito 鈥� let鈥檚 call it ogrowup. Without a copy of ogrowup, the fertilised egg dies. Now imagine an HEG engineered to target ogrowup. Imagine, too, that the HEG is only activated when cells start to divide to make eggs or sperm. When that happens, the HEG copies itself into the second chromosome. All the eggs or sperm 鈥� which carry a single set of chromosomes 鈥� will have a copy of the HEG and no functioning ogrowup gene.
When you release a mosquito containing such an HEG into the wild, nothing much happens at first. The HEG mosquito breeds with an unmodified mosquito and its offspring are healthy, thanks to the ogrowup gene they got from their unmodified parent. But when this second generation comes to breed, the HEGs go to work again; again, none of the eggs or sperm carry a working ogrowup gene. And in the next generation the same thing will happen. By this time the HEG will be spreading like wildfire and the likelihood of two HEG carriers breeding becomes higher. And when two HEG carriers mate with one another they produce no offspring.
With a lot of HEG carriers around, carrying the gene becomes a big selective disadvantage. But by the time that happens it will be too late. The gene will be halfway round the world before natural selection can get its boots on. With a head start like that, the fatal effects of the HEG get the upper hand and the insect carrying it could die out completely. According to Burt鈥檚 calculations, an HEG that starts off in just 1 per cent of the population but gets into 95 per cent of its carriers鈥� sperm and eggs could be killing four-fifths of the population鈥檚 offspring within 12 generations 鈥� in the tropics, that鈥檚 as little as 36 weeks. A couple of HEGs, or one that got into 99.9 per cent of the eggs and sperm, could result in 99.8 per cent of the offspring dying, probably enough to wipe out the population completely.
Burt is not the first person to consider messing around with mosquito genes in order to tackle malaria. Chris Curtis of the London School of Hygiene and Tropical Medicine, has been publishing on the subject since the late 1960s, and recently the field has been positively swarming with ideas. At the end of February, a number of would-be mosquito modifiers went to Seattle to talk with the Bill and Melinda Gates Foundation, which is considering investing a serious amount of money in such technologies.
Rather than eradicate mosquitoes, almost all of this work focuses on altering them. The idea is to find genes that make mosquitoes less likely to transmit malaria, and then spread these genes through the population. There has been some success in the first part of the problem. In the past two years it has become possible to modify the main malaria mosquitoes in Africa and India. And in 2002, researchers at Case Western Reserve University in Ohio and the University of Bayreuth in Germany reported that mosquitoes with an anti-malarial gene were 80 per cent less likely to pass on a form of malaria to rats (Nature, vol 417, p 452). Now that the genomes of humans, mosquitoes and the malaria parasite itself have been sequenced, scientists should be able to make it almost impossible for malaria parasites to infect humans. 鈥淚鈥檓 sure they will succeed鈥�, says Curtis.
But he is not sure that that will be enough. The difficulty, according to Curtis and others, lies in the second part of the problem 鈥� spreading those genes through the population at large. It is not enough to release a few anti-malarial mosquitoes. If being resistant to malaria were an advantage to mosquitoes, natural selection would have solved the malaria problem long ago. So engineered mosquitoes will tend to be a bit less fit than wild ones, a prediction largely borne out by recent research by another group at Imperial College (Science, vol 299, p 1225).
To solve this problem, the resistance genes need to be hitched to a 鈥渄river鈥� 鈥� a piece of DNA that spreads for some other reason. Various drivers have been discussed, including transposons and parasites that live within the mosquitoes鈥� cells, but they all share a significant drawback. 鈥淭he crunch problem,鈥� says Curtis, 鈥渋s how you make sure that the thing you want driven remains linked to the driving system.鈥� If you think of the driver as a locomotive and the things you want driven as the carriages, he says, then if the coupling between them breaks, the locomotive will drive off into the distance while the carriages start to roll backwards. There鈥檚 always a risk that a new mutation will uncouple the driver and its carriages, and even if the chances of this happening are very small, it鈥檚 still a fatal flaw. Work by some of Curtis鈥檚 colleagues suggests that if the engineered mosquitoes are just 20 per cent less fit than wild ones, and even if the chance of uncoupling is as low as one in a million, the locomotive always runs away and the resistance genes die out. To Curtis, that looks like the end of the line. 鈥淚f we don鈥檛 have a reasonable prospect of driving those genes into wild populations, there鈥檚 no point.鈥�
Others in the field are not as pessimistic as Curtis, and think they can keep drivers and driven together. But Burt鈥檚 HEG approach avoids the problem altogether. The HEGs don鈥檛 need a driver 鈥� they are locomotives designed to steam through other genes rather than pull freight. Mutations that alter the HEGs鈥� effectiveness will be lost from the population, but the original version will keep spreading happily. 鈥淓volutionary stability is a key feature of the system,鈥� says Burt.
Another advantage is reversibility. Once resistance genes are let loose in the population there is no clear way of recalling them; they will be there for good, along with any unintended consequences. The HEG approach, though, leaves room for second thoughts.
The genetic code allows the same message to be spelled out in somewhat different ways, so it鈥檚 possible to encode the same protein using two different DNA sequences. To continue the example used earlier, it would be possible to design a version of the ogrowup gene that produced the correct protein but didn鈥檛 contain the DNA sequence recognised by the HEG. So if it were decided that some clemency was needed, insects carrying the tweaked ogrowup gene could be released. The new gene, carrying with it the great advantage of immunity to the HEG, would spread quickly and easily without any need for a driver. A useful extension to this precaution would be to design resistant versions of the target gene for closely related species, in case the HEGs should jump from one species to another.
What looks like prudent reversibility from our standpoint, though, will look like a get-out clause for the mosquito. If people can design an HEG-resistant version of the ogrowup gene, then what鈥檚 to stop mosquitoes evolving one too? 鈥淧ests always develop resistance, and you have to think about that,鈥� says Burt. One way to combat resistance would be to use a suite of HEGs aimed at a number of sites within the crucial gene; another would be to aim at several different genes. Just as it is harder to evolve resistance to a number of drugs at once, so it would be harder to evolve resistance to several HEGs. In time, resistance would arise regardless. But that only matters if the time required is available. If the HEGs wipe out the whole population, the problem does not arise.
Burt is not suggesting that HEGs should be inserted into mosquitoes straight away. For a start, the technology isn鈥檛 ready yet. HEGs have been transferred into some insect species but not mosquitoes. Another missing piece is the ability to design homing endonucleases that recognise specific DNA targets. Those available today only recognise DNA sequences in their original hosts. But advances in protein design should crack that problem fairly soon. David Baker of the University of Washington in Seattle, an expert on predicting the shapes and properties of proteins, has recently started working on the problem with some of his colleagues. It鈥檚 possible, he says, that 鈥渨e could be designing nucleases with novel specifications within six months.鈥� And if their initial approach doesn鈥檛 work, Baker says, another surely will.
This is why Burt thinks that using HEGs will soon become possible. In which case, he says, research into wiping out a few of the species that do the greatest harm to humans 鈥� and debate about whether that would be a good idea 鈥� should start now.
One obvious issue is the public acceptability of releasing genetically modified mosquitoes. It鈥檚 hard to assess this at present, but it may be tested soon. At Oxford University, a team led by Luke Alphey is working on a way of engineering mosquitoes so that they cannot produce female offspring, except under certain lab conditions. This technique could be used to create large numbers of male mosquitoes that can only have male offspring, a new version of the well-established 鈥渟terile insect technique鈥�. Let loose a large number of these sterile males and you swamp the local wild mosquito population. Alphey鈥檚 approach is the least radical of the GM mosquito techniques and so likely to be the first to come before the regulators. If successful, it could pave the way for the release of lethal HEG-carrying mosquitoes.
For the time being, Burt plans to develop a non-lethal HEG from yeast for fruit flies, to see how well it can spread through a lab population. If that works, then it will be time to think about moving to mosquitoes. It鈥檚 at this point, says Burt, that a really serious debate will have to be had. 鈥淚t鈥檚 an ethical and moral question,鈥� he says. 鈥淵ou鈥檇 be extinguishing a species, and lots of people spend their lives trying to save species. But when I talk to people about it I have yet to hear anyone who thinks [eradicating malaria-carrying mosquitoes] would be terribly bad. And nobody is crying for smallpox.鈥� More importantly, no one is crying for loved ones lost to smallpox any more, and, barring a crime against humanity, no one ever will again. But millions of people lose a family member to malaria every year.