Comments must be received on or before March 23, 2004 and may be submitted by postal mail/commercial delivery or by e-mail. If you use postal mail/commercial delivery, please send four copies of your comment (an original and three copies) to:

Docket No. 03-031-2
Regulatory Analysis and Development
PPD, APHIS, Station 3C71
4700 River Road, Unit 118
Riverdale, MD 20737-1238.

Effects on Plants
To some extent, the results of the FSE reinforce these concerns. As used in the trials, the herbicides applied to GMHT beet (glyphosate: Roundup Biactive from Monsanto) and spring sown oilseed rape (glufosinate ammonium: Liberty from Bayer CropScience) did give better weed control than the conventional herbicides used on the non-GM crops, and this did lead to lower weed biomass at the end of the season, and fewer weed seeds being caught in seed rain traps. There were also fewer seeds in the seed bank in the following year in the next crop (mostly cereals) in the rotation^1,2. In maize the opposite was true. Glufosinate applied to the GM variety gave poorer control of weeds than the conventional regimes, which, in 75% of the test crops, were based on atrazine applied mostly pre-emergence. As atrazine is a persistent, relatively broad spectrum residual herbicide, it is not surprising that it gave better control than one that was applied some time after the crop and weeds had emerged. Indeed, conventional maize crops had the fewest weeds, lowest weed biomass, and produced fewest weed seeds of all the crops tested. The recent announcement within the European Union that atrazine would be banned in the near future has cast doubt on the validity of the maize results; however, as this intention was not known at the time the experiment was devised, the results remain valid for the comparisons that were made. Further work may need to be done to re-examine the comparison when farmers have selected alternative conventional products in a couple years time, after the ban has been implemented.

Effects on Invertebrates
Although the results as far as plant survival is concerned are fairly clear cut, the effects of the two treatments on invertebrates were less clear. For the majority of organisms measured, including carabid and staphylinid beetles, spiders, slugs, snails, true bugs, pests, and predators, there was no difference between the conventional and GMHT crops^{3, 4}. Within some of those groups, a few species that rely heavily on the weeds for food or habitat did respond to the differences in weed and seed abundance. For example there were more seed-eating carabid beetles (Harpalus rufipes) in conventional beet and oilseed rape and GMHT maize in July and/or August when more seeds were available; in contrast there were more detritivorous springtails (Collembola) in all three GMHT crops later in the season because there was more decaying plant material to feed on following the later control of large weeds in these fields. This in turn led to more springtail-feeding carabid beetles (Loricera pilicornis) in the GMHT crops³.

The results that contributed most to the headlines in the national and international press, however, concerned butterflies and bees. Bees were significantly reduced in GMHT beet, but not in the other two crops, largely because of the good control of thistles by glyphosate⁴ compared to conventional herbicides. However, the number of bees recorded in conventional beet was very low compared to the number seen in rape crops (2.26 compared to 36.9 per km of transect), so the importance of this statistic is questionable. The picture is similar for butterflies. Although there were significantly fewer butterflies both within and around GMHT beet and oilseed rape crops, the numbers in the former were much less^4,5. Beet is not a favored crop for these aerial insects unless fields are infested with thistles and other flowering plants. Of the butterfly species affected in oilseed rape, more than half were pest species (Pieris). So again, the significance of these results on non-pest species depends on the availability of other food sources within the farmed landscape and is certainly not as draconian as suggested by the headlines.

Implications for Farm Management
In contrast to the perceived harmful effects, there were some considerable benefits in terms of farm management to be gained from using the GMHT technology. The number of spray applications made to the GMHT crops, and therefore also the amount of energy used to apply them, was significantly reduced in beet (by 55%) and oilseed rape (by 12%), but not in maize⁶. The number and quantity of active ingredients were substantially reduced in beet and maize, but not in spring rape. The cost of the herbicides applied to the GMHT crops, especially beet, is likely to be considerably less than the current cost of conventional herbicides. Other studies⁷ have shown a potential benefit of £150/ha (approx. USD$110/acre) to be gained from growing GMHT sugar beet compared to conventional crops. These are important considerations for farmers and would contribute to the decision-making process if the crops were to become available.

Mitigation of Adverse Environmental Effects
Are the FSE results bad enough to persuade the government to ban GMHT crops for general release? More than anything else, what these studies have revealed for the first time on a large scale is that intensive agriculture is having a gradual, but nevertheless inexorable, effect on the diversity of life within the UK farmed landscape. These studies showed that there were greater differences between crops than between the herbicide treatments within any individual crop. Conventional maize for example was the least diverse crop, but both conventional and GMHT oilseed rape were more diverse than the others⁸. The declines in seedbanks that have been highlighted will continue whether or not GM crops are introduced, although there is evidence that the rate of decline of seeds in the seed bank would be significantly enhanced by the technology¹. However, it is not the introduction of GMHT crops that needs to be addressed, but the way in which society wants food to be produced in the future.

• 1) A return to low intensity farming—this would require more land to be brought into production to compensate for the reduced yields that would inevitably ensue; or
• 2) Further intensification on the most productive areas of land—this would allow less productive areas to be returned to wildlife.

Both approaches would require some financial encouragement, and indeed both approaches could be used in tandem. However, some scenarios could be achieved only with GM technology. For example, taking a relatively low-intensity approach, we have devised a method within GMHT sugar beet that allows weeds to survive longer between the rows early in the season to the benefit of some wildlife. Glyphosate is applied as a band spray at first application (by using a cheap modification of the nozzles on an existing spray boom) at the 2–4 leaf stage of beet plants, followed by a later overall application that removes weeds when they become competitive. The optimum timing of the latter spray depends on the density and species composition of weeds in the field. None of the conventional herbicides used in sugar beet would be able to achieve this effect, as they give poor control of weeds with more than 2–4 leaves. This approach encourages insects and other invertebrates to remain in the crop for longer⁹, reduces the number of pests by camouflaging the beet plants and encouraging predators¹⁰, and can offer greater food resources and habitat for ground-nesting birds at a time of year when such resources are scarce in the arable environment. Yields from band-sprayed plots were as good as those from plots receiving conventional herbicides, and the cost of the herbicides was much cheaper, so there is a great financial incentive to growers to pursue this strategy.

Another alternative to increase the number of seeds available for granivorous birds later in the season has been to apply glyphosate once overall early in the season. This allows later emerging, but non-competitive weeds to mature and produce seeds. Of course both methods produce crops that look untidy, at least for part of the season, and this will require growers to re-think their approach to crop production—a pristine crop is not necessarily an environmentally-friendly one. However, the power of GMHT technology allows growers to have confidence that yields will not be adversely affected even if some weeds remain in the crops, and that would make it easier to persuade them to adopt these low intensity approaches.

The second approach, to increase the intensity of cropping, is likely to be more acceptable to growers, but probably not environmentalists. Given that crops do need to be grown to produce food, some compromise is necessary. Is it better to provide a larger area of poor diversity or many small areas of rich diversity? This question needs to be tested on a landscape scale.

The high intensity philosophy is to increase production in the most fertile parts of the fields, namely the field centres. Headlands would not be cropped, but would instead be made into set-aside areas with either natural regeneration, or, if necessary, sown with mixtures of wild flower seeds or species to encourage wildlife. Strips of untreated areas would also be left through the fields themselves to mitigate the losses of seeds caused by either conventional or GMHT cropping. However, the influences of this latter approach on populations of farmland birds cannot be examined until large scale ecological studies are set up.

Conclusion
The adverse effects reported to be due to GMHT beet and oilseed rape crops were due to the more efficient control of weeds by the respective broad-spectrum herbicides applied to those crops. In the same way, the better control of weeds by use of the broad-spectrum residual herbicide atrazine caused the same effects in conventional maize crops. More creative use of herbicides such as glyphosate and glufosinate-ammonium could reverse these effects within crops, but more intensive agricultural production of the most fertile land could allow less productive land to be used for the benefit of wildlife.

1. Heard MS et al. (2003a) Weeds in fields with contrasting conventional and genetically modified herbicide-tolerant crops. 1. Effects on abundance and diversity. Phil. Trans. R. Soc. Lond. B. 358: 1819-1832.

2. Heard MS et al. (2003b) Weeds in fields with contrasting conventional and genetically modified herbicide-tolerant crops. 2. The effects on individual species. Phil. Trans. R. Soc. Lond. B. 358: 1833-1846.

3. Brooks DR et al. (2003) Invertebrate responses to the management of genetically modified herbicide-tolerant and conventional spring crops. 1. Soil surface active invertebrates. Phil. Trans. R. Soc. Lond. B. 358: 1847-1862.

4. Haughton AJ et al. (2003) Invertebrate responses to the management of genetically modified herbicide-tolerant and conventional spring crops. 2. Within-field plant epigeal and aerial arthropods. Phil. Trans. R. Soc. Lond. B. 358: 1863-1878.

5. Roy DB et al. (2003) Invertebrates and vegetation of field margins adjacent to crops subject to contrasting herbicide regimes in the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Phil. Trans. R. Soc. Lond. B. 358: 1879-1898.

6. Champion GT et al. (2003) Crop management and agronomic context of the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Phil. Trans. R. Soc. Lond. B. 358: 1801-1818.

7. May M. (2003) Economic consequences for UK farmers of growing GM herbicide tolerant sugar beet. Ann. Appl. Biol 142: 41-48.

8. Hawes C et al. (2003) Responses of plants and invertebrate trophic groups to contrasting herbicide regimes in the Farm Scale Evaluations of genetically modified herbicide-tolerant crops. Phil. Trans. R. Soc. Lond. B. 358: 1899-1915.

9. Dewar AM et al. (2003) A novel approach to the use of genetically modified herbicide-tolerant crops for environmental benefit. Proceedings of the Royal Society: Biological Sciences 270 (1513): 335-340.

10. Dewar AM et al. (2000) Delayed control of weeds in glyphosate-tolerant sugar beet and the consequences on aphid infestation and yield. Pest Management Science 56(4): 345-350.

BIOENGINEERING OF THE ROOT-SOIL INTERFACE: A HAIRY STORY
Marcel Bucher

Terrestrial plants are sessile organisms that stay at the same habitat during their entire life cycle, and as such, their survival is crucially dependent on rapid adaptation to environmental changes. The above-ground chlorophyll-containing parts provide mainly sucrose, derived from photosynthetic carbon fixation basipetally to sustain root growth and development. The root system serves many tasks—it anchors the plant, absorbs water and minerals from the soil, and delivers certain growth regulators.

Despite its importance for plant productivity, the root is a largely unexplored frontier for genetic engineering¹. Root processes influence rhizosphere chemistry, nutrient acquisition, and interactions with beneficial and pathogenic organisms. Roots have the remarkable ability to secrete a vast array of low and high molecular weight molecules into the rhizosphere in response to biotic and abiotic stresses—in a process termed rhizosecretion. There is interest in root bioengineering for numerous reasons. Engineered roots may be better adapted to nutrient poor or saline soils, be better hosts to beneficial organisms, resist soil-borne pathogens more effectively, exhibit increased competitiveness with weeds, remediate toxic waste from contaminated locations, or be a cost-effective alternative for the production of bioactive metabolites and proteins for molecular farming.

Today, engineering of rhizosecretion has become possible using molecular genetic tools. We are interested in investigating whether root hair cells (also called trichoblasts) are suitable target cells for genetic engineering of the root secretory machinery to improve attributes such as plant nutrition.

Why root hair cells?
To exploit the soil optimally for often scarce minerals, most terrestrial plants respond with the formation of a well-developed mycorrhiza, i.e., a symbiosis between the root and soil-borne arbuscular-mycorrhizal fungi, thus increasing the below-ground absorptive surface area. This allows improved plant nutrition due to the ability of the large extraradical mycelium to exploit soil volumes that are otherwise not accessible to the non-colonized plant root. A key process in this symbiosis is the transfer of phosphorus as inorganic phosphate (Pi) from the fungus to the host plant in exchange for carbohydrates derived from the plant. Phosphate transfer at the fungus-plant interface is mediated by mycorrhiza-specific Pi transporter proteins². Another strategy to combat nutrient scarcity is the formation of an increased root-shoot ratio and an enlargement of the absorptive surface area relative to the root volume³. The increase in absorptive root surface area can be attributed in part to the enhanced production of root hairs, their length being especially important for Pi uptake⁴. The formation of cluster roots (bottlebrush-like clusters of rootlets densely covered with root hairs) in certain plant species such as white lupin is another type of response to Pi deficiency. Cluster roots have been functionally linked with efficient mechanisms for the chemical mobilization of sparingly available Pi sources in the rhizosphere, involving the excretion of large amounts of organic chelators and enzymes (e.g., acid phosphatase)³.

Root hairs are the tubular extensions of trichoblast cells at the root epidermis and make up between 70% to 90% of the total root surface area. They play a dominant role in a number of root functions. For example, membrane proteins responsible for the uptake of certain nutrients have been localized to the root epidermal layer, including root hair cells and are involved in the transport of nutrients such as Pi, potassium, nitrate, and iron (Fe). Increases in both root hair length and density in response to Fe and Pi deficiencies have been reported (Figure 1A), and the study of root hairless mutants revealed an essential role of root hairs in Pi uptake from the soil solution. Moreover, root hairs are instrumental in the anchorage of the plants in the soil, in water uptake, and in the establishment of the Rhizobium symbiosis in legumes.

Cloning of a root hair-specific gene
Having a regulatory genetic unit to direct expression of a foreign gene to root hair cells is a prerequisite to engineer specifically the root-soil interface. During environmental stress conditions, expression of a gene and its encoded protein at the site where it is needed would be an economic process for the plant, which could be of advantage over constitutive gene expression in all tissues. Moreover, expression of the transgene in root hairs avoids accumulation of the encoded protein in, for example, the edible parts of a crop plant.

The differential hybridization technique is a classical method used to identify mRNAs in comparative studies. A tomato cDNA library from root hair cells has been synthesized. Subsequently, a differential screening resulted in the cloning of several cDNAs, the corresponding transcripts of which were highly abundant in root hairs and were found to encode extensin-like proteins from cell walls^4,5. One cDNA corresponding to LeExt1 was selected for detailed expression analysis, and genomic fragments from the promoter region were cloned. To visualize promoter activity in planta, the promoter fragments were fused to the E. coli ß-glucuronidase (GUS) reporter gene. The chimeric genes were then introduced into plants by Agrobacterium-mediated T-DNA transfer, and promoter activity was eventually assessed by histological analysis of GUS activity in the transgenic plants. In situ hybridization of LeExt1 mRNA and promoter/GUS expression studies provided evidence for a direct correlation between LeExt1 expression and root hair growth (Figure 1B⁶). Root hair-specific activity of LeExt1/GUS was comparable in roots of transgenic plants from diverse species such as tomato, potato, tobacco, strong-spined Medick (Medicago truncatula) and cassava.

Secretion of a consensus phytase from root hair cells
Secretion is a basic function of all plant cells and organs and is especially developed in plant roots. Besides secretion of low molecular weight molecules, plant roots have the capacity for protein secretion. Proteins are secreted when targeted to the lumen of the endoplasmic reticulum by signal peptide-mediated translocation, followed by migration through the exocytotic pathway. This allows rhizosecretion to be modulated by transferring chimeric genes encoding recombinant secretory proteins into plants.

In soils rich in organic matter, 20-50% of the organic Pi may be bound in phytate (myo-inositol hexakisphosphate). Phytase activity, which hydrolyzes Pi esters in phytate, in root exudates of potato, a plant with a relatively low Pi acquisition efficiency, is generally missing⁷. We hypothesized that the development of crop plants with an increased capacity to secrete phytase may enhance acquisition of orthophosphate originating from organic Pi, and may lead to improvements in fertilizer management in agriculture. We therefore aimed at transferring this activity to potato via Agrobacterium-mediated gene transfer. As a candidate gene, we chose a synthetic phytase gene that encoded a consensus phytase exhibiting an increased thermostability and improved stability towards proteolytic degradation. This gene was originally designed to produce a phytase protein that can be added to feed pellets for an increased Pi availability in feed for animal production⁸. The consensus phytase gene was under the control of the root hair-specific LeExt1 gene promoter, and therefore the encoded consensus phytase was expected to accumulate exclusively in root hairs and was efficiently secreted from transgenic potato roots due to a secretory signal peptide N-terminally fused to the consensus phytase protein. When phytate, the natural substrate for phytase activity, was present in the nutrient solution, a typical pattern of phytate degradation with a transient accumulation of lower molecular weight inositol phosphates was observed (Figure 1C and D). Moreover, transgenic potato plants accumulated 40% more Pi in their leaves and tubers than wild type when phytate was added to the soil substrate in cultures from the greenhouse.

It remains to be analyzed whether secretion of a phytate-degrading activity in these designer plants will eventually lead to an improved phosphate acquisition efficiency and therefore a greater yield especially in organic soils. Phytate is poorly digested by farm animals with simple stomachs as opposed to ruminants. The undigested phytate is excreted with the feces into the environment, i.e., in agricultural land where organic fertility sources such as animal manure and compost have been used, and phytate is degraded due to soil microbial phytase activity leading to an increased soil Pi content. Consequently, Pi loss from the topsoil is a significant water pollution problem mainly in areas where animal production is concentrated. Phytase-secreting plants thus could be used to decontaminate these soils from phytate and eventually reduce the phosphate concentration in the surface runoff water while simultaneously feeding the plant with phosphate. To evaluate the effectiveness of this approach, field tests in suitable agroecosystems would be necessary.

Summary and perspectives
Overall, engineering the root-soil interface was made possible via cell-specific expression in root hair cells of a recombinant gene encoding a secretory phytase, which resulted in an improved Pi nutrition. In future work, engineering rhizosecretion could be used for the development of plants which are more resistant to adverse soil conditions or which help to clean-up contaminated soils. Moreover, it could be possible to develop plants secreting bioactive proteins via their root hairs that are relevant for medical purposes.

Acknowledgements
I would like to thank Prof. Nikolaus Amrhein, ETH Zurich, for his continuous support. I am grateful to Dr. Philip Zimmermann, ETH Zurich, for photographs and Dr. Theresa Fitzpatrick for critical reading of the manuscript.

References

1. Bucher M. (2001) in Plant Roots: The Hidden Half, eds. Waisel Y, Eshel A & Kafkafi U. (Marcel Dekker, New York), Vol. 3, pp. 279-294.

2. Rausch C et al. (2001) Nature 414: 462-70.

3. Marschner H. (1995) Mineral nutrition of higher plants. Academic Press, London.

4. Bates TR and Lynch JP. (2000) Am J Bot 87: 964-970.

5. Bucher M et al. (1997) Plant Mol Biol 35: 497-508.

6. Bucher M et al. (2002) Plant Physiol 128: 911-23.

7. Zimmermann P et al. (2003) Plant Biotechnol J 1: 353-360.

8. Lehmann M et al. (2000) Protein Eng 13: 49-57.

POULTRY RESEARCH IN THE POST-GENOME ERA
Eric Wong

The completion of a draft sequence of the chicken genome provides another powerful tool for poultry scientists. The challenge will be to utilize this information for improving avian growth, reproduction, and health. Clearly, the integration of all of the current resources (genetic and physical maps, QTL markers, EST libraries and microarrays, whole genome sequence, and proteomics) will be key to unraveling the molecular mechanisms that control complex biological systems.

The genome sequence should facilitate studies of functional genomics that aim to identify genes and their regulatory elements, the encoded gene products and the gene expression patterns for various metabolic processes. For the molecular biologist, no longer will chicken genes need to be cloned by the laborious process of screening lambda or BAC libraries; instead, the genes can now be "cloned" in silico. In the rapidly developing field of proteomics, the genome sequence will be essential to predict the amino acid sequences of encoded proteins/peptides. With a complete genome sequence, the search for candidate genes that are in close proximity to a marker linked to a desirable trait is greatly simplified.

Poultry nutritionists and physiologists will also benefit from genomic and proteomic technologies. Microarrays and proteome analysis will reveal groups of genes that are coordinately regulated in a metabolic pathway. The complete sequence of the chicken genome will facilitate the search for avian genes. For example, the identification of chicken genes that are involved in nutrient transport can be completed by a search of the avian genome for homologues to known transporter genes from other species. This type of genome scan has revealed that chickens and mammals share many transporters, however some mammalian transporters do not appear to be present in the avian genome.

Avian health is another area that should benefit from the genome sequence. Understanding the cellular genes regulating a host-pathogen interaction will be essential for developing strategies to enhance disease resistance. The development of cDNA microarrays for expression profiling of immune genes will reveal key genes or groups of genes involved in host response to infection. The identification of these genes will lead to the development of healthier birds, which in turn will have a direct impact on food safety issues.

On December 18, the University of California, Davis, announced its recall of about 30 tomato seed samples distributed during the past seven years to researchers in the United States and abroad. Tests showed that the samples did not contain the intended variety, but rather a type engineered to express the PG gene, which improves the thickness of tomato paste. In 1996 Petoseed Company (since acquired by Seminis Vegetable Seeds) and Zeneca Plant Science commercialized a similar genetically modified (GM) tomato, a variety approved by the U.S. Food and Drug Administration and the U.S. Department of Agriculture.

A different type of GM plant mixing incited controversy in another part of Northern California. Echoing the Nuclear Free Zone movement a generation ago, anti-GM crop activists sponsored a ballot measure that will let Mendocino County residents decide whether to ban GM organisms (GMOs). If ballot Measure H is passed, then the county would become the first in the country to prohibit farmers from planting GM crops and raising GM livestock.

The California Plant Health Association, an agricultural industry group, did not care for the ballot's scare tactic statements, such as the assertion that "GMOs will irreversibly contaminate native plants and trees." The organization filed a lawsuit in December requesting the deletion of portions of the ballot measure. The judge, however, decided that he should not be micromanaging ballot language and that the CPHA has the opportunity to express opposing views on the ballot.

Supporters of the ban, including organic farmers, argue that they need Measure H to prevent their crops from being tainted with GM crops. Although Mendocino's biggest cash crop is reportedly marijuana, the county does have a strong organic farming presence. What Mendocino County seems to lack is the presence of GM crops. Nevertheless, a GM ban would benefit organic farmers who could use it as a marketing tool. A vote on Measure H is slated for March 2.

State and local venues are becoming popular battlefields for anti-GM crop activists and biotech lobbyists. However, the anti-GM movement targeted the federal government in a pending Hawaiian skirmish.

On November 12 a coalition of environmental groups filed suit in the Honolulu federal district court seeking an order for the USDA to develop an Environmental Impact Statement on the environmental and health risks associated with the open-air testing of biopharm crops, plants engineered to produce pharmaceuticals and chemicals for industrial uses. Characterizing the USDA's efforts as a "laissez faire regulatory approach," the plaintiffs alleged that the agency's regulation of field testing is inadequate to assure that biopharm crops do not contaminate soil or food supplies, harm humans or wildlife, or cross-breed with wild plants or conventional crops. This lack of oversight, they argued, violates the National Environmental Policy Act and the Endangered Species Act.