Such sloppy semantics have eclipsed an important distinction between tolerance and resistance recognized by evolutionary biologists. Tolerance refers to the ability of an organism to withstand a certain amount of damage from environmental insults without suffering the loss of fitness that one would expect. Resistance, on the other hand, acts through a biochemical change that blocks or reduces the action of a pesticide, herbivore, or disease (for example), and resistant organisms hardly notice the attack. For example, herbicide-resistant Roundup Ready soybean plants are engineered with genes for proteins that sidestep the destructive pathway of glyphosate, and exposure leaves them undamaged. A plant exposed to glyphosate might evolve tolerance by producing more seeds, or larger leaves, to compensate for the damage that glyphosate can cause. Both resistant and tolerant organisms will enjoy higher fitness than their non-resistant or non-tolerant counterparts; however, to understand the dynamics of tolerant organisms over time it is important to understand the different means by which this common ending is reached.

Although Roundup Ready soybeans may best be described as herbicide resistant, what about the weedy species sympatric to this transgenic crop? Might tolerance evolve in a species subject to frequent application of an herbicide? This question has been investigated by the authors of two recent articles in The Proceedings of the National Academy of Sciences that discuss the development of herbicide tolerance in tall morning glories (Ipomoea purpurea), a weedy species in the southeastern United States^1,2.

To measure tolerance, the fitness of a genotype is measured under undamaging (control) and damaging environmental conditions¹. In a randomized block experiment, Baucom and Mauricio² compared the relative fitness of 32 maternal genotypes of tall morning glories under two conditions: no herbicide applied; and RoundUp herbicide applied at a rate of 1.121 kg/ha. To assess fitness, the investigators collected fruits and viable seeds from the plants, and they noted the fraction of leaves showing ill effects from glyphosate application. They regressed the relative fitness of each genotype on the environmental condition (i.e., no herbicide or herbicide) and used the slope to indicate the level of tolerance. In this analysis, a slope of zero implied complete tolerance, which was operationally defined by the authors as the ability of the plant to reproduce in spite of herbicide damage. For example, in the figure below each line represents a different genotype’s "fitness norm of reaction"² after the effect of block was removed; the x-axis represents the two environments (0 = no herbicide; 1 = herbicide treatment). Lines with negative slope (lower relative fitness under the herbicide treatment) have low tolerance; positive slope indicates overcompensation; and a slope of zero means that the genotype is completely tolerant of the herbicide. Because the slopes are different across genotypes, this also indicates that genetic variation for tolerance exists in these morning glories.

Figure 1. Relationship between relative fitness and treatment environment for the 32 maternal lines. Residuals of fitness were used after the effect of block was removed. On the x axis, 0 = glyphosate absent, and 1 = glyphosate present. Slopes of the lines represent tolerance. (Source: Baucom RS & Mauricio R (2004) PNAS 101, 13386-13390; copyright Proceedings of the National Academy of Sciences USA)

By correlating the relative fitness under the different environments with the overall tolerance of the genotype, the researchers assessed the costs and benefits of tolerance.² In Figure 2, relative fitness values under control and herbicide-application treatments (in graphs A and B, respectively) are correlated with tolerance.

Figure 2. Costs and benefits of tolerance. (A) Costs of tolerance indicated by a significant negative genetic correlation between fitness and tolerance in the presence of glyphosate as measured by a standardized selection gradient. In both graphs, the y axis depicts the residuals of relative fitness after the effects of block had been removed, and the x axis depicts the level of tolerance for each maternal line standardized to a mean of zero and a variance of one. (Source: Baucom RS & Mauricio R (2004) PNAS 101, 13386-13390; copyright Proceedings of the National Academy of Sciences USA)

The negative slope in A indicates that there is a fitness cost to tolerance in a glyphosate-free environment; conversely, the positive slope in B signifies a fitness advantage associated with tolerance under the herbicide treatment (as one might expect). A specific example of this fitness cost under the control condition is that the most glyphosate-tolerant line produced 35% fewer viable seeds than the most glyphosate-susceptible line².

In other words, as long as there is sufficient variation in the genes for tolerance (as there seems to be for I. purpurea), one would expect strong selection for tolerance in the presence of the herbicide, and selection against tolerance when the herbicide is absent.²

For risk assessors, the distinction between resistance and tolerance is an important one, because the two traits are likely to evolve in different ways and therefore have different effects¹. In a study modeling the evolutionary dynamics of disease resistance and disease tolerance in a long-lived host, Roy and Kirchner³ found that resistance genes eventually settled at a polymorphic equilibrium, whereas tolerance genes were driven to fixation. Variation still exists in the traits governing tolerance to glyphosate in the morning glory because Roundup represents a relatively new selective force on this plant¹. Because theory predicts fixation of tolerance under constant exposure to a selective force³, managers might be encouraged by Baucom and Mauricio’s discovery of fitness tradeoffs associated with tolerance when the selective force is absent. This means that the evolution of tolerance might be delayed, in this case, by halting application of Roundup or possibly by applying an herbicide that acts through a different mechanism¹. In the context of a field of Roundup Ready crops, this recommendation necessarily implies rotation to a different crop plant.

Monitoring specifically for tolerance in weedy species will be important to better understand this phenomenon, particularly where selective forces are high, such as in Roundup Ready fields. Further complicating future management of tolerance is the possibility that tolerance and resistance traits are genetically linked, so management schemes may have to account for the evolution of both in certain species. Since the ecology of agricultural landscapes is changing rapidly, including new selective pressures associated with farming herbicide-resistant crops, agronomic practices should be guided by adaptive management. As Roy cautions¹, if we don’t learn how to effectively retard the evolution of resistance and tolerance in pest species, we will compromise the efficacy of herbicides and ultimately face a widespread agricultural problem.

1. Roy BA (2004) PNAS 101, 13974-13975

2. Baucom RS & Mauricio R (2004) PNAS 101, 13386-13390

3. Roy BA & Kirchner JW (2000) Evolution 54, 51-63

Probably the most obvious way that two or more transgenes can be introduced into one plant is to add them sequentially by either (a) crossing a plant containing one transgene with individuals harboring another transgene, or (b) by retransforming a transgenic plant with additional transgenes. Many examples can be drawn to illustrate the success of these strategies¹. For example, sequential crossing has been used to pyramid different pest or disease-resistance genes into crops, to produce novel proteins such as secretory IgA antibodies in plants, to introduce new biochemical pathways for biodegradable plastic biosynthesis, or to effect complex manipulations to existing pathways for lignin biosynthesis or fruit ripening. Retransformation has also been used in species as diverse as Arabidopsis and trees to successfully introduce or manipulate biochemical pathways controlling flower color or salinity tolerance, and pathways involved in producing starch, fatty acids, and lignin. However, iterative procedures such as these are severely limited by the fact that the introduced transgenes can segregate in subsequent generations, as they are not linked, but sited at different, random loci in the plant’s genome. Significant time, effort, and cost will be associated with maintaining the larger progeny populations that will be needed to identify plants that retain all transgenes. Nevertheless, retransformation can have considerable advantages for species in which segregation is not a problem, because vegetative propagation methods are usually employed. However using vegetative propagation means that different selectable marker genes used during rounds of retransformation cannot be eliminated by out-crossing and will accumulate in the plants, potentially creating a significant hurdle for regulatory approval and public acceptance. Recently, several systems that enable the removal of selectable marker genes² or that allow for transformation without the use of selectable markers have been developed that may help to overcome this limitation.

Perhaps the most promising approach taken to date to introduce multiple transgenes into plants has been the strategy of co-transformation. Multiple transgenes on different T-DNAs (harbored together or separately within one or several Agrobacterium strains) or multiple DNA fragments³ (if using biolistics) are inoculated or bombarded together onto plant tissues. The technique is quick, works with a variety of species, and has the major advantage that transgenes tend to co-integrate at the same chromosomal position and will therefore be inherited together in most progeny. Co-transformation has already been used commercially to produce maize with stacked herbicide tolerance and insect resistance traits. It was also used to engineer ‘Golden rice’, increase isoflavone levels in soybean, produce rice resistant to multiple insects, concurrently improve several wood quality traits in trees, and to allow Arabidopsis to produce a biodegradable plastic co-polymer. Some reports suggest that co-transformed transgenes tend to integrate as high copy number repeats that might promote transgene silencing or complicate the regulatory approval process. Not enough work has yet been done to determine whether this will be a real problem when working with large numbers of transgenes, but finding transgenics with stable expression and relatively simple integration patterns is not a problem when small numbers of transgenes are involved. Techniques for resolving loci with complex integration patterns, via recombination, have been tested in wheat and may offer a future solution to this problem. Similarly, more work is needed to develop methods that would promote coordinate expression of multiple co-transformed genes. Existing literature, though sparse, suggests that as the number of co-transformed genes increases, so does the difficulty in identifying plants expressing all transgenes at effective levels.

Another strategy that has been used with great success to stack traits in commercial GM crops has been to physically link two or more genes by positioning them contiguously on DNA that will transfer as a single entity into a plant (i.e., on a single T-DNA for Agrobacterium-mediated transformation). At a research level, as many as four or five genes enabling biodegradable plastic production have been linked within one T-DNA and introduced into Arabidopsis or oilseed rape. The practical problems associated with assembling large multi-gene cassettes are being addressed by the development of several new vector systems in which individual gene cassettes constructed in unidirectional shuttle vectors can be easily transferred to special plant transformation vectors that have been engineered to contain an array of rare endonuclease cleavage sites matching those of the shuttle vectors. Large numbers of genes might also be transferred by assembling several T-DNA cassettes, each of moderate size (harboring just two or three linked genes), and by co-transforming these different T-DNAs. However even physically linked genes are not necessarily going to be coordinately expressed at similar levels. In addition, factors such as the number of transgenic loci, the number of insertions at a given locus, and the stability of each locus can influence transgene expression. Repeated sequences within transgenes, either in coding or regulatory regions, can also lead to variability in expression by activating gene silencing. Current opinion is divided on whether flanking linked genes with matrix-associated regions can improve coordinated expression and/or reduce transgene silencing.

Several alternative novel strategies to express multiple genes from a single promoter have been developed and may present the best prospects for ensuring coordinated expression of introduced genes in plants.

Plastid transformation holds tremendous promise as a way of easily expressing polycistronic constructs (indeed entire bacterial operons) in plants⁴. Extremely high levels of protein production can be achieved. However, relatively few crop plants can yet be transformed in this way, and the system will never be suitable for proteins requiring certain post-translational modifications or targeting to other sub-cellular locations.

Internal ribosome entry sites (IRESs) are viral sequences that can directly recruit ribosomes to internal positions within mRNAs and initiate translation in a cap-independent manner. IRESs have been used in gene therapy research in animal systems and a few reports describe their use in polycistronic constructs in plants. Limitations of the strategy are that, although cistrons expressed in this way are coordinately regulated, they are expressed at different levels due to the relative inefficiency of internal initiation and IRES-mediated expression can be lower than normal cap-dependent expression.

The expression of multiple proteins can also be coordinated by expressing them initially in a single open reading frame as a polyprotein that could subsequently be processed into component polypeptides. Several research groups have tested the value of connecting protein sequences via short linker sequences that are substrates for plant proteinases. However this approach depends on temporal and spatial coincidence between the expression of the polyprotein and the endogenous plant proteinase that will process it, and this limits the applications for which this strategy will be useful. Several groups have constructed chimeric polyproteins in which a processing protease (the potyvirus NIa protease) is encoded within the polyprotein itself. By separating coding sequences by the protease’s heptapeptide cleavage sequence, discrete polypeptide products can be processed post-translationally. Polyprotein cleavage is efficient but expression levels can be low, and targeting of component proteins to different subcellular compartments is complicated if polyprotein translocation occurs more rapidly than polyprotein cleavage. An alternative polyprotein strategy separates component polypeptides by the 20 amino acid 2A peptide of certain picornaviruses. This novel peptide mediates an intra-ribosomal ‘skip’ during translation so that a peptide bond is not formed at the end of the 2A sequence, yet translation continues. Thus 2A peptides can effect co-translational polyprotein ‘cleavage’ allowing unimpeded subsequent targeting of individual polypeptides to different sub-cellular compartments⁵. Chimeric polyproteins incorporating 2A have been widely tested in eukaryotic systems including mammalian, human, insect, yeast, and fungal cells, and several reports now describe their successful use in plants.

An increasing diversity of conventional and novel techniques are being used to stack genes and/or traits in transgenic plants and some have already been used in commercial production. All methods have individual advantages and limitations and further refining and supplementation of this ‘toolkit’ is still necessary in order to provide more durable and cleaner transgenic technologies for the future. Advancing the technologies for multiple gene manipulation in plants is a major challenge facing 21st century plant biotechnology. Meeting this challenge is essential if the full opportunities for GM crops to effect revolutionary change in agriculture, industry, nutrition, and even medicine are to be realized.

The author’s research is supported by the UK Biotechnology and Biological Sciences Research Council (BBSRC).

1. Halpin C. (2005) Gene stacking in transgenic plants – the challenge for 21^st century plant biotechnology. Plant Biotech. J. 3, 141-155

2. Hare PD & Chua NH. (2002) Excision of selectable marker genes from transgenic plants. Nat. Biotechnol. 20, 575-580

3. Altpeter F et al. (2005) Particle bombardment and the genetic enhancement of crops: myths and realities. Mol. Breeding 15, 305-327

4. Bock R & Khan S. (2004) Taming plastids for a green future. Trends Biotechnol. 22, 311-318

5. El Amrani A et al. (2004) Coordinate expression and independent subcellular targeting of multiple proteins from a single transgene. Plant Physiol.135, 16-24

SUSTAINABLE AND DURABLE INSECT PEST RESISTANCE IN TRANSGENIC CROPS
Paul Christou

Insect pests constitute a major constraint in crop production. Approximately 15% of agricultural production is lost every year to insects, and consequently farmers spend billions of US$ yearly to provide effective control using chemical insecticides. A number of major crops, for example, cotton, cannot be cultivated without heavy chemical inputs in the form of insecticides. Often such insecticides are highly toxic to non-target animals (and humans), they have serious negative environmental impacts, and increasing numbers of insect pests are capable of evolving resistance against chemical insecticides. Biological control has been pursued; however, results are, in most cases, marginal in row crop situations. We are thus faced with a problem in terms of being able to control damaging insect pests using conventional strategies. Recombinant DNA technology provides a complementary/alternative strategy to combat insect pests.

Transgenic plants expressing Bt toxins have been used successfully to provide resistance against a number of insect pests for several years, with insect resistance being the second most widely used trait in transgenic crops (after herbicide resistance) in world agriculture. One potential problem with Bt-based insecticides is that they are very widely used, with up to 90% of microbiological insect control products based on topically-applied Bt toxins. For this reason, there is concern that insects might evolve resistance to Bt toxins and indeed the diamondback moth (Plutella xylostella) has evolved resistance in some open field populations in response to repeated exposure to foliar sprays containing Bt proteins, while laboratory selection experiments with other insect pests have shown that recessive mutant alleles can confer resistance to multiple Bt toxins.

It is worth noting, however, that the evolution of resistant insects against transgenic plants expressing Bt toxins has yet to be seen in the field, defying dire predictions based on computer simulations and various deterministic models, and despite the fact that insect resistance crops have been cultivated on millions of hectares worldwide for many years. This, of course, does not mean that insects will not evolve resistance under field conditions at some point in the near or distant future. It is always prudent to devise strategies to assure the long term usefulness of transgenic crops expressing insecticidal genes. It will indeed be very unfortunate if the effectiveness of such crops that took years to develop is compromised through inappropriate deployment strategies. Such strategies encompass multiple components ranging from the nature of the genetic construct and its expression profile, combining multiple transgenes in the same plant, deployment strategies to discourage the development of homozygous resistant insect populations, etc.

A major problem with single Bt genes is that the target insect spectrum is limited to coleopterans and lepidopterans, and in some cases, even some lepidopterans may not be susceptible to specific Bt genes.

One strategy to address potential limitations in conventional transgenic insect pest control involves the stacking or pyramiding of multiple transgenes in the same transgenic plant. We reported previously the simultaneous introduction of cry1Ac, cry2A, and gna (snowdrop lectin from Galanthus nivalis) into rice¹. The transgenes provided very significant protection against three of the most important insect pests of rice: rice leaf folder (Cnaphalocrocis medinalis), yellow stemborer (Scirpophaga incertulas), and brown planthopper (Nilaparvata lugens). The triple transformants showed synergistically greater resistance to these insects than plants expressing single transgenes, and in our experiments, we observed that the greatest reduction in insect survival and the greatest reduction in plant damage occurred in plants expressing all three transgenes. Bioassays using the triple transgenic plants showed complete eradication of the rice leaf folder and yellow stemborer, and 25% reduction in the survival of the brown planthopper. We therefore concluded that combinations of genes, whose products disrupt different biochemical or physiological processes in the same insect, provide a multi-mechanism defense.

We have now devised an alternative strategy to extend/complement transgene pyramiding in which a conventional insecticidal molecule such as the Bt toxin cry1Ac is fused to the non-toxic ricin B-chain². The recognition of toxin binding sites in the insect midgut is an important factor determining the spectrum of Bt toxin activity. Several groups investigating the mechanism of toxin recognition have identified N-acetyl galactosamine residues as an important component of Bt toxin-binding receptors. We therefore selected the non-toxic ricin B subunit as a fusion partner for the Bt toxin, since it binds such residues with very high affinity. We postulated that a fusion protein comprising an N-terminal Bt toxin and a C-terminal lectin polypeptide would provide a novel binding domain that would allow the hybrid protein to bind to a wider repertoire of receptors than the control toxin, in this case, cry1Ac. Furthermore, the fact that single alleles in homozygous form can confer resistance to Bt toxins suggests that each toxin interacts with a single receptor, and loss or modification of this receptor leads to resistance in otherwise susceptible insects. Therefore, by increasing the number of binding domains on each toxin, the likelihood of resistance evolving in target populations is reduced, as mutations affecting several different receptors are highly unlikely to occur simultaneously because of the mathematical compounding of mutation infrequencies.

The results of insect bioassays in our experiments showed that transgenic rice and maize plants expressing the fusion protein were synergistically more toxic to a range of insect pests than those expressing cry1Ac alone. Furthermore, the fusion protein conferred resistance to a broader spectrum of insect pests than those normally susceptible to cry1Ac. In our greenhouse experiments, we were able to demonstrate effective control of the striped stemborer (Chilo suppressalis), the cotton leaf worm (Spodoptera littoralis) [Figure 1], and the leafhopper Cicadulina mbila. The aphid, Rhopalosiphum padi, was not affected in any way.

Figure 1. Survival of Spodoptera littoralis on T1 transgenic maize plants. (Source: Mehlo et al (2005) PNAS 102, 7812-7816; copyright Proceedings of the National Academy of Sciences USA)

Our experiments thus demonstrated that: (a) insects that are normally susceptible to Bt genes are even more susceptible to the fusion protein between Bt and the ricin B chain, suggesting that we may not require a high dose deployment strategy for transgenic crops expressing such molecules; (b) a major lepidopteran pest that is resistant to Bt now becomes susceptible; (c) a homopteran pest that is outside the normal host range of Bt is now highly susceptible to the hybrid molecule; and (d) not all insects become susceptible to the fusion protein molecule. The last point is extremely important as it demonstrates a degree of selectivity in the mode of action of the fusion protein. This also provides a starting point for further engineering experiments to construct highly specific toxins targeted against particular insect pests but not others.

Standard Bt cotton, maize, and potato crops have been released commercially and have been readily adopted by farmers, while Bt rice, although yet to be commercialized, has been tested successfully in several field trials. Rice plants expressing Bt genes are expected to be commercialized in China in the very near future and are already being cultivated illegally by farmers. Experience has shown the benefits of such enhanced crops in terms of increased yields, reduced chemical inputs and, as a knock-on effect, improved farmer and consumer health. Sustainable resistance against insect pests is the cornerstone of any sensible deployment strategy that utilizes transgenic plants expressing insecticidal proteins, either alone or, preferably, within an integrated pest management system. It is in the context of resistance management that crops expressing fusion proteins such as those described in our work could be the most beneficial.

Bt crops in the U.S. must be co-maintained with refuges that decrease the selection pressure on target pests and reduce the theoretical likelihood of establishing resistance, while in China, refuges in Bt cotton crops are provided by an alternative host plant that supports the major pest species, Helios armigera. This type of mandatory refuge management system may be difficult to implement for crops such as rice and maize in developing countries with many smallholder farms, and where substitute hosts for insect pests are not available. Fusion proteins such as those we have developed have the potential of providing strong and sustainable resistance requiring multiple counter-adaptive mutations, but would require only a single toxin transgene.

Lessons learned following the widespread use of chemical pesticides for the control of insect pests over the past several decades call for reason and caution in how we deploy transgenic plants expressing insecticidal genes presently and in the future, and a requirement to learn about the basic mechanisms. Bt transgenic plants are remarkably specific in their activity, with little or no effect on non-target organisms. Field tests with Bt rice, maize, and other crops have revealed no negative impact on biodiversity and indeed a positive impact resulting from the reduction in pesticide use³. Another strong advantage of conventional Bt crops is that there is no credible evidence for toxicity or allergenicity in humans.

The environmental, economic, and social benefits of insect resistant crops are well known, despite claims to the contrary by opponents of biotechnology applications that particularly benefit developing countries. For Bt cotton, farmers in India have seen a 70% reduction in the use of chemical insecticides translating to a very substantial savings in expenditure on insecticides by subsistence farmers who are the individuals least likely to afford them. Up to 85% increases in cotton yields have been seen by Indian farmers who have adopted the technology. Similarly, in China, farmers who are cultivating Bt cotton have been able to reduce the number of insecticide applications to one third, compared to farmers that grow conventional cotton⁴. In addition, the number of farmers in China reporting pesticide poisoning symptoms in conjunction with cultivation of Bt cotton was reduced from 50,000 to less than 2,500.

Wider application of this technology will necessitate testing a broad range of insects encompassing agriculturally important pests and also beneficial insects and other organisms to ascertain the full efficacy and safety of these fusion proteins. Prudent and necessary safety evaluations and environmental impact assessments of transgenic insect resistant crops need to be performed in a rational and scientific manner. It is necessary to ascertain effects on non-target organisms; however, we need to keep such evaluations in the appropriate biological perspective. If, for example, there are some minor detrimental effects on the growth rate of the green lacewing, this author believes the plight of hundreds of millions of people in the developing world who are on the brink of starvation or those subsistence farmers who are forced to use toxic chemicals to combat insects should also be a component of the assessment exercise!

As a final thought, an increasing number of resources are being diverted towards the development of ever more sensitive methods to detect so called "unintended effects on non-target organisms" in insect resistant transgenic crops. So far, most, if not all, of the results are negative, especially when compared to current agricultural practices, a control many politically-minded researchers do not usually perform. Is it not perhaps time to apply scientific reasoning and come to the logical conclusion that if a battery of tests fails to demonstrate any substantial detrimental effects, perhaps the answer might be that these plants do not pose any real threat to the environment, rather than pursuing further research to develop even more sensitive methods in the hope that such detrimental effects might be detected?

I would like to thank Jonny Gressel for helpful comments on the final draft of this manuscript and the Rockefeller Foundation for continuous support over the past 15 or so years.

1. Bano-Maqbool S, Riazuddin S, Loc NT, Gatehouse AMR, Gatehouse JA & Christou P (2001) Expression of multiple insecticidal genes confers broad resistance against a range of different rice pests. Molecular Breeding 7, 85-93

2. Mehlo L, Gahakwa D, Nghia P-T, Loc N-T, Capell T, Gatehouse J, Gatehouse A & Christou P (2005). An alternative strategy for sustainable pest resistance in genetically enhanced crops. Proceedings of the National Academy of Sciences USA 102, 7812-7816

3. James C (2004). Global Status of Commercialized Biotech/GM Crops: 2004. ISAAA, Ithaca NY

4. Huang J, Hu R, Rozelle S & Pray C (2005) Insect-resistant GM rice in farmers´ fields: Assessing productivity and health effects in China. Science 308, 688-690

A WIDE CROSS-TOLERANCE OF TRANSGENIC RICE CONTAINING HUMAN CYP2B6 TO VARIOUS CLASSES OF HERBICIDES
Sakiko Hirose and Hiroyuki Kawahigashi*

Pesticides have played very important roles against food shortages and plant diseases, and our modern life would be vastly different without their use. Weed infestation adversely affects crop production by reducing yields and decreasing market prices of the crop. For cost-effective land use, crop yield and price must be maximized, and the cost of weed control minimized. Herbicides substantially lighten the farmer’s heavy physical workload and improve crop yield and product quality. Herbicides are now widely used for crop cultivation and for management of lawns, railroad rights-of-way, highway margins, and other purposes, although there remains the possibility of non-target effects of pesticides to the environment caused by agricultural runoff.