Insects readily evolve resistance to insecticidal proteins that are introduced into genetically modified crop plants. Continuous directed evolution has now been used to engineer a toxin that overcomes insect resistance. See Article p.58
The genetic engineering of crops to express proteins that are toxic to insects is a safe and cost-effective alternative to chemical pesticides1. The insecticidal toxins most commonly used in agriculture are the Cry proteins from the ubiquitous soil bacterium Bacillus thuringiensis (Bt). Since becoming commercially available in 1996, crops that produce Bt toxins have been widely adopted, and more than 420 million hectares have been planted around the world2. However, insect resistance quickly emerged as a major threat to the long-term success of such crops2. On page 58 of this issue, Badran et al.3 present an elegant method for the continuous evolution of engineered Bt toxins, and describe a toxin that targets a new receptor on insect cells and thus overcomes existing resistance.
Bt toxins form crystalline inclusion bodies that, when ingested by insects, are solubilized and activated by gut protease enzymes4. The toxins then bind to specific receptors on insect midgut cells and form membrane pores that destroy the cells, killing the insect. A variety of receptors are targeted by different Bt toxins, including alkaline phosphatase, ATP-binding cassette transporters and cadherin-like proteins. The affinity and specificity of these toxin–receptor interactions underlie one of the biggest advantages of Bt toxins as pesticides: unlike broad-spectrum chemical insecticides, Bt toxins kill only specific families of insects4, effectively suppressing pest populations without damaging their natural enemies5 or endangering human health1.
Alongside economic and environmental gains6, the rapid adoption of crops engineered to produce Bt toxins has led to powerful selection pressures for resistant insects. The first field observation of substantial resistance was reported just 6 years after Bt crops were commercially introduced; since then, resistance to newly introduced toxins has appeared as little as 2 years after initial commercial availability2. Overall, observations accumulated over the past 20 years have repeatedly shown that insects can rapidly overcome most of the Bt-toxin crops that were designed to control them, highlighting the fierce arms race between humans and insects for crop consumption.
The evolution of insect resistance is often mediated by mutation, deletion or reduced expression of midgut-cell receptors4. Badran et al. addressed the problem of receptor-mediated resistance by engineering a widely used Bt toxin, Cry1Ac, to tightly bind to a receptor that it does not naturally target, the cadherin-like receptor from the common insect pest Trichoplusia ni (TnCAD). To rapidly isolate variants of Cry1Ac that have the desired characteristics, the authors used phage-assisted continuous evolution (PACE), a highly efficient method for the directed evolution of proteins.
In PACE, viruses that infect bacteria (called bacteriophage, or just phage) are made to multiply in a constant supply of host bacteria. Both the phage and the bacteria are engineered to ensure that phage infectivity depends on a specific characteristic of an evolving protein7. This is achieved by coupling the desired activity of the protein to the expression of a gene that is essential for phage infectivity. The target protein for engineering, which is encoded by the phage, continuously evolves over multiple phage generations, and is under powerful selection for activity. The process is speeded up by increasing the mutation rate in the bacterial host, so that extensive genetic variability is screened in a short time.
Badran et al. adapted the PACE technique to evolve a tight protein–protein interaction between Cry1Ac and TnCAD. Their method (which is based on a bacterial two-hybrid system) was designed such that a stronger interaction between the evolving protein (Cry1Ac) and the binding target (a TnCAD-derived fragment) leads to increased transcription of a gene that allows for greater phage infectivity. After 22 days of continuous phage proliferation, representing more than 500 generations of replication and selection, the authors isolated multiple evolved variants of Cry1Ac. The stability of variants containing consensus mutations (mutations that appeared in several different Cry1Ac variants) was further improved by removing mutations that lead to protein destabilization. The resulting Cry1Ac variants exhibited high affinity for TnCAD, without losing their ability to bind to the native Cry1Ac receptor, and were able to efficiently kill Cry1Ac-resistant as well as susceptible insects (Fig. 1).
The incorporation of engineered Bt toxins such as these Cry1Ac variants into genetically modified crops would be a welcome addition to the limited pesticide arsenal. The evolutionary arms race will continue, of course, and it will probably be just a few years until insects evolve resistance to these new toxins as well. Nevertheless, the ability to engineer multiple toxins against target receptors of choice may prove instrumental in the future. As ever-more pest species adapt to existing toxins, innovative tools and strategies to combat the evolution of resistance must be pursued to maintain the global food supply. This is especially important given the expected growth of the human population to 9.7 billion by 2050 (ref. 8), increasing the demand for crops9.
A popular strategy to delay resistance involves 'pyramids' — crops that produce two or more toxins targeting the same pest, making the emergence of resistant insects much less likely10. Toxins that bind to previously untargeted insect receptors will be favourable additions to such pyramids, because they would be expected to reduce the probability of cross-resistance (when an insect that is resistant to one toxin is also resistant to another). Future work, however, may have to search for even more durable strategies, such as targeting regions on evolutionarily conserved essential receptors.