January 2007

.pdf version

The FY 2007 Biotechnology Risk Assessment Research Grants Program request for applications is now available on the web at The program supports environmental assessment research related to the introduction of genetically engineered organisms into the environment. Electronic submission is required this year and the deadline for submissions is February 15, 2007.

Stephen O. Duke and Antonio L. Cerdeira

Glyphosate is the most important herbicide since 2,4-D, and biotechnology has magnified its importance. It has a unique target site in the shikimate pathway, 5-enolpyruvyl-shikimate-3-phosphate synthase (EPSPS). Since transgenic, glyphosate-resistant crops (GRCs) were introduced over ten years ago, adoption of this technology in the U.S. and South America has been dramatic. Of the 100 million hectares of transgenic crops grown annually worldwide in 2005, almost 90% were glyphosate-resistant alone or had glyphosate-resistance genes stacked with Bt toxin-based insect resistance genes. Adoption rates of GR soybean and cotton in the U.S. are shown in Figure 1. Almost all of this transgenic crop area is composed of four crops: canola, cotton, maize, and soybean. About 60% of the global soybean crop is now glyphosate resistant.

Although no significant negative environmental effects have been documented in the vast areas growing these crops, there is concern about possible environmental impacts of this technology. There are potential positive and negative effects of the herbicide, the transgene, and the herbicide/transgenic crop weed management system. We discuss all three aspects, with some reference to weed management products and practices replaced by this technology. We recently reviewed this topic in much more detail, where many of the specific points of this brief review are referenced (Cerdiera and Duke, 2006).

Effects on Herbicide Use
Glyphosate is not a low use rate herbicide; however, it is considered a low risk herbicide in terms of toxicity and environmental effects. Some studies have claimed that the volume of herbicide use is greater with GRCs. However, others have concluded that no significant change in the amount of herbicides used has occurred with the adoption of GRCs in the USA. The impact of GRCs on herbicide use varies with the crop, but there are no cases so far in which use rate has substantially increased. However, the amount of glyphosate and other herbicides used with a GRC grown repeatedly will increase with time, as weed pressure increases (see below). Both increased amounts of glyphosate and other herbicides are being used in these cases.

The decreasing cost of glyphosate due to loss of patent protection has made higher application rates economical in some cases. Heavy adoption of GR soybeans in the U.S. contributed to dramatic reductions in costs of some other soybean herbicides due to competition. Thus, indirectly, GR soybeans have helped make it more economical for farmers to use higher rates of certain other herbicides.

A more important consideration than use rate is actual environmental impact. In most situations, glyphosate replaces herbicides that are significantly more toxic and that persist much longer in soil and water. Several studies of the total environmental impact of glyphosate versus the herbicides that it replaces have found the impact is less with GRCs.

Effects on Fossil Fuel Use
A major expense and source of pollution in weed management are the fossil fuels used in tillage and herbicide application. This factor is seldom considered when evaluating environmental impacts of herbicide use. GRCs have reduced the need for tillage (discussed below) and, sometimes, the number of herbicide applications. In a recent European study using a life-cycle assessment approach, Bennett et al. (2004) concluded that the major environmental advantage of growing GR sugarbeet would be lower emissions from herbicide manufacturing, transport, and field operations, thus reducing contributions to global warming, smog, ozone depletion, ecotoxicity of water, and acidification and nitrification of soil and water.

Effects on Soil and Water
Glyphosate is not a significant soil contaminant when used at recommended doses. Glyphosate is rapidly degraded by soil microbes and is inactivated through its strong adsorption to soil, reducing its potential biological activity. The half-life of glyphosate in soil varies from just over a week to several months, which is shorter than many of the herbicides that it replaces. It is commonly metabolized to aminomethylphosphonic acid (AMPA), which has higher mobility in soil.

Potential direct effects of GRCs and their management on soil biota include changes in soil microbial activity due to direct effects of glyphosate, differences in the amount and composition of root exudates of GRCs versus non-GRCs, and alteration in microbial populations because of the effects of management practices for transgenic crops, such as changes in other herbicide applications and tillage. Most studies find no effects or very limited, transient effects of glyphosate on soil microflora. Effects that have been detected are field site and season dependent, and the changes in microbial communities associated with GRCs are more variable and transient than those caused by other agricultural practices such as crop rotation, tillage, use of certain other herbicides, and irrigation. So far, no agriculturally significant effect of glyphosate on soil microorganisms has been documented.

The soybean nitrogen-fixing symbiont Bradyrhizobium japonicum possesses a glyphosate-sensitive EPSPS, and glyphosate can inhibit or kill the microbe. Furthermore, glyphosate is preferentially translocated to soybean nodules. Applying glyphosate to GR soybeans can reduce nodulation, nodule size, and leghemoglobin content of nodules. In the field, effects are transient, and no evidence indicates that crop yield is affected.

Top soil loss due to tillage is perhaps the most environmentally destructive effect of row crop agriculture. GRCs facilitate reduced tillage systems, which contribute to reductions in soil erosion from water and wind, fossil fuel use, air pollution from dust, soil moisture loss, and soil compaction. Reduced tillage also improves soil structure, leading to reduced runoff and pollution of surface waters with sediment, nutrients, and pesticides. U.S. soybean farmers have reduced tillage since GR soybeans were introduced, translating to a large annual savings on tillage costs. In a five-year period in the U.S., during which the planting of GR soybeans increased from only a few percent to about 70% (Fig. 1), there was a dramatic increase in the adoption of no-tillage and reduced tillage management (Fig. 2). Most of this change was associated with the GR soybean production. Similarly, there has been a rise in no-tillage soybean production in Argentina with the adoption of GR soybeans. Dramatic reductions in soil erosion were documented where GR soybeans were grown using no-tillage practices.

Even though glyphosate is highly water soluble, it does not leach into ground water in most soils due to its strong adsorption to soil components. Soil and sediments in bodies of water are the main sinks for glyphosate residues from surface water, greatly reducing further transport. Once in surface water, it dissipates more rapidly than most other herbicides. It is one of the few herbicides that has been approved for aquatic vegetation use by the US EPA. Models have predicted less herbicide pollution of water in GR maize and wheat than in non-transgenic varieties of these crops.

Risk to aquatic organisms is negligible or small at application rates <4 kg/ha and only slightly greater at application rates of 8 kg/ha (application rates much higher than recommended rates). A life-cycle assessment technique used to compare conventional sugarbeet agricultural practices with risks that might be expected if GR sugarbeet were grown suggests that growing this GRC would be less harmful to water ecology for the GRC than for the conventional crop (Bennett et al., 2004). In general, what we know about glyphosate’s movement to surface water suggests less impact of GRCs on aquatic vegetation than conventionally-grown crops.

Food and Feed Safety
Glyphosate might directly alter food safety if it or its metabolic products are found in the edible portions of the crop at levels above tolerance levels, or if the transgene itself could alter food safety, either directly or indirectly. For regulatory approval, transgenic crops are scrutinized to a far greater level than conventional crops.

Glyphosate has a very low level of toxicity in mammals (Williams et al., 2000), and it is not retained to a significant extent in animal tissues. It is not a restricted use pesticide and is sold for home use. Residues of glyphosate and AMPA have been found below tolerance levels in GR soybean seed. Non-toxic levels of glyphosate and its degradation products can be expected to be found in edible parts of GRCs.

The transgene(s) itself could theoretically be toxic due to direct toxicity, antinutritive effects, or allergenic effects or because the gene could cause a change in the metabolic pathways of the crop, changing levels of already existing metabolites or introducing a new metabolite. None of these effects has been found with GRCs. Extensive feeding studies have shown that animal feed from several GRCs has no effect on compositional and nutritional safety or animal health with a variety of animals grown for human consumption.

Plant and Animal Life
There is concern that the more complete weed control usually obtained with GRCs may reduce biodiversity in agricultural areas. Because glyphosate has low toxicity, effects on animals are most likely to be due to indirect effects on plant life. No-tillage agriculture that is favored by GRCs results in weed species shifts and in more vegetation on the field before and after the period of crop production, resulting in improved habitat for other organisms. If weed-enhanced animal biodiversity is desired, strategies have been devised for optimizing this with GRCs without a yield loss.

The environmental effects of GR soybeans was compared to the effects with non-transgenic soybeans for over 1400 U.S. Midwest farms (Nelson and Bullock, 2003). Toxicity of glyphosate and other herbicides for rats was used in their estimates. Unlike most other studies, relative toxicity of herbicide choices available to farmers was considered. Simulation models suggested that GR soybean technology is more environmentally friendly, especially with regard to mammalian toxicity, than other technologies for U.S. farms.

Effects of sublethal glyphosate exposures to non-target plants near GRC fields are not well studied. Glyphosate can be a plant growth stimulant at concentrations that one might expect from agricultural drift. Sugarcane farmers in several parts of the world use glyphosate at low doses to increase sugar content. However, at non-lethal doses, glyphosate can impair plant defenses against pathogens.

The influence of glyphosate on plant diseases on GRCs is variable (Table 1). Mechanisms of these effects are not well studied, and there is no clear picture yet as to whether there is a net reduction or increase in plant disease on GRCs.

Effects on Weed Problems
Five weed species have evolved resistance to glyphosate in GRC fields (see Considering the high selection pressure caused by glyphosate in GRCs, this problem will probably increase. Although glyphosate is non-selective and broad-spectrum, it cannot control all plant species or biotypes at recommended dose rates. Thus, weed species or biotypes with some natural resistance have filled niches vacated in GRC agroecosytems. This has necessitated higher rates of glyphosate application and/or inclusion of other herbicides with glyphosate. Reduction of tillage that has been encouraged by the adoption of GRCs also causes shifts in weed species. Finally, GRCs can be a problem as volunteers the following season in a different GRC field. To combat these developing weed problems, farmers are applying other herbicides with glyphosate. Most changes in vegetation resulting from GRC use at this time are problems in agroecosystems with no effects in natural areas.

Introgression of Transgenes
Introgression (sometimes called gene flow) is the movement of a gene or genes from donor plants to sexually compatible recipient plants of a different genotype by crossing, followed by backcrossing of the hybrid with the recipient population until the gene is stabilized in the population. For full introgression, several backcrosses are required. Hybrids between species or between crops and weedy variants of the crop are often unfit, but the herbicide enhances the survival of unfit crosses that might not survive under normal competitive situations, allowing the survivors to backcross with the non-crop parent, resulting in eventual introgression of the herbicide resistance transgene into the wild population.

Among the commercial GRCs grown in North America, only canola and maize have weedy relatives in North America with which they could interbreed. Transgene flow from GR canola to Brassica rapa L. has been documented in commercial fields. GR gene movement from test plots of non-commercial GR creeping bentgrass to wild species of Agrostis has been documented. Bentgrass is not a troublesome weed, but if it were glyphosate resistant, it might become a problem in GRCs.

Gene flow from GRCs to non-GRCs of the same crop species is more likely than outcrossing with other species. Non-GRCs contaminated with transgenes may not be accepted by some markets, depending on the degree of contamination. For some crops, such as soybean, outcrossing is not considered a significant problem, but for rice, maize, and canola considerable outcrossing can occur. Gene flow from GR canola to non-transgenic canola has occurred in Canada.

GR transgenes would provide no survival benefit in a natural environment. However, when a glyphosate resistance gene is coupled with another transgene that would provide a natural ecosystem fitness advantage (e.g., disease, insect, drought, or temperature extreme resistance) a potential problem could arise. A transgene conferring an added fitness advantage could change the balance between plant species within an ecosystem, a potentially undesirable outcome that might also affect non-plant species. When both genes are used, the use of the herbicide in the presence of the hybrid will favor backcrossing until the gene conferring the fitness advantage has introgressed. At this time, insect and glyphosate resistance transgenes are coupled in commercially available transgenic maize and cotton.

Movement of fitness-enhancing transgenes into wild populations is the only non-reversible risk of transgenic crops. Thus, preventing movement of transgenes to wild populations is highly desirable. But, relatively little effort is being expended to develop 'fail safe' methods to do this.

Our generalizations may not apply to all situations, as risks and benefits are geography and time dependent. In the context of the replaced herbicides and agronomic practices, the apparent health and environmental benefits of the glyphosate/GRC combination are significant in most studied cases. Glyphosate is more environmentally and toxicologically benign than many herbicides that it replaces. Perhaps the most significant contribution of GRCs to the environment has been its influence on adoption of reduced- or no-tillage agriculture. The preponderance of evidence indicates that GRC-derived feed and foods are safe and are nutritionally equivalent to their non-transgenic counterparts. Little or no direct impact or risk is expected from glyphosate resistance transgenes when they introgress into wild populations. However, when stacked with genes that impart a fitness advantage in the wild, the glyphosate-resistance trait, in the presence of glyphosate use, can increase the chances of potentially harmful gene flow.

References and further reading

Bennett R et al. (2004) Environmental and human health impacts of growing genetically modified herbicide-tolerant sugar beet: a life-cycle assessment. Plant Biotechnol. J. 2, 273-278

Cellini FA et al. (2004) Unintended effects and their detection in genetically modified crops. Food Chem. Toxicol. 42, 1089-1125

Cerdeira AL & Duke SO. (2006) The current status and environmental impacts of glyphosate-resistant crops: A review. J. Environ. Qual. 35, 1633-1658

Geisy JP et al. (2000) Ecotoxicological risk assessment of Roundup herbicide. Rev. Environ. Contam. Toxicol. 167, 35-120

Nelson DS & Bullock GC. (2003) Simulating a relative environmental effect of glyphosate-resistant soybeans. Ecol. Econom. 45, 189-202

Williams GM et al. (2000) Safety evaluation and risk assessment of the herbicide Roundup and its active ingredient, glyphosate, for humans. Regul. Toxicol. Pharmacol. 31, 117-165

Stephen O. Duke
Research Leader
Natural Products Utilization Research Unit
Agricultural Research Service, USDA

Antonio L. Cerdeira
Brazilian Dept. Agricuture
Agricultural Research Service

Raffaella Tinghino

Food allergies are adverse reactions to an otherwise harmless food or food component that involve an abnormal response of the body’s immune system to a specific protein(s) in foods. Food allergies are caused by a wide variety of foods. Theoretically, any food that contains protein would be capable of eliciting an allergic reaction, although foods vary widely in their likelihood of provoking allergic sensitization. Approximately 2–5% of the adult population is affected, with the prevalence among children being even greater.

The incidence of food allergy around the world is increasing in line with other forms of allergic disease. Under healthy conditions, the immune system reacts towards innocuous dietary antigens by inducing antigen-specific systemic non-responsiveness, termed oral tolerance; whereas under pathological conditions, such as food allergy, it is believed that a dysregulation in the mechanisms of oral tolerance induction prevails. Most food-allergic reactions are mediated by specific IgE antibodies, and the development of an IgE-mediated response to an allergen is the result of a series of molecular and cellular interactions involving antigen presenting cells, T cells and B cells. The production of IgE is promoted by Th2 cells and their cytokines (Fig. 1).1

Most confirmed food allergies are associated with a relatively limited range of products (peanuts, crustaceans, fish, milk, eggs, tree nuts, wheat, and soybeans)2, although there are many other foods and food products that have also been implicated. Soybean is one of the major sources of protein in human and animal nutrition and it has also been well characterized as an important allergenic source. At least 16 IgE-binding soy proteins with molecular masses from 7.5 to 97 kDa may be involved in clinical allergy.

Advances in biotechnology have resulted in an increasing number of genetically engineered foods, and among these, soybean is one of the most widespread. The predominant genetically engineered soybean grown in the world is Roundup Ready, which is resistant to glyphosate, the active ingredient in Roundup agricultural herbicides. This resistance was obtained by introduction of the glyphosate-tolerant cp4 epsps coding sequence, derived from the common soil bacterium Agrobacterium sp. strain CP4, into the soybean genome using particle acceleration transformation. Globally, genetically modified soybean made up 60% of all transgenic crops grown in 2005.3

The application of modern biotechnology to food production presents new opportunities and challenges for human health. The potential benefits to the public health sector include altering the nutrient content of foods, decreasing their allergenic potential, and improving the efficiency of food production systems. On the other hand, the potential effects on human health of the consumption of food produced through genetic modification must be carefully examined.

With the development of genetically engineered crop plants there has been a growing interest in the approaches available to assess the potential allergenicity of novel gene products. This allergenicity might derive from changes in endogenous protein levels, expression of known allergens in different foods, and the expression of novel proteins that may be allergenic. Although strategies exist for such assessments, improvements should be considered, especially in cases in which the gene of interest is derived from a source with no history of allergenicity. A report in 2001 of a joint FAO/WHO expert consultation on allergenicity of foods derived from biotechnology suggests an integrated and stepwise, case-by-case approach (Fig. 2), which incorporates consideration of the serological identity of the protein of interest with known human allergens, examination of the amino acid homology with, and/or structural similarity to, allergenic proteins, and measurement of the stability of the protein in a simulated gastric fluid.4

The same document gave an indication that, for additional assessment of the potential allergenicity of expressed proteins, informative data can be generated using animal models. These data could be used in concert with the approaches summarized above.

Development of a murine model of soybean specific IgE sensitization
Recently, we developed a murine model of IgE mediated food allergy based on oral administration of soybean protein extract. The purpose of our study was to develop and characterize the immune response induced in the animal model by intragastrical immunization with wild-type soybean extract (wt-SE), or with a genetically modified soybean extract (gm-SE). Then, we used this model to compare the immunological response obtained both with wt-SE and with gm-SE to assess the potential allergenicity of gm-SE with respect to its natural counterpart.5

The Balb/c mouse has been widely utilized to evaluate the sensitizing potential of novel proteins, because it favors the development of Th2-type immune responses and the production of IgE antibody. Although oral administration is the preferred route of exposure, it has been demonstrated that such a regimen may lack the sensitivity required for effective identification of the inherent sensitizing potential. This is probably attributable, at least in part, to the fact that oral exposure to the antigen generally results in the development of tolerance. Therefore, the major general obstacle towards establishing food allergy models is the strong innate tendency of animals to develop immunological tolerance to ingested antigens. This is particularly relevant in the case of soybean allergen, as the usual mouse diet includes soybean proteins. Consequently, for our experiments we used Balb/c mice fed on soy-free feed and born in our animal facilities from females fed on soy-free feed. Recent studies have developed murine models of IgE-mediated food allergy based on oral coadministration of antigen with an optimal dose of cholera toxin (CT) to generate a robust allergen-specific antibody response.6

In our study, we used a similar approach to develop our murine model, combining it with the use of Balb/c mice of the second (F2) offspring generation bred on a soy protein-free diet. We generated the first murine model of IgE-mediated soybean sensitization, in which a soybean-specific IgE response was induced by oral immunization, and reported the characterization of the antigen-specific cellular and humoral responses. To this aim, two allergenic extracts from wt-soybean and from gm-soybean seeds were prepared and also characterized. We immunized Balb/c mice with both wt-SE and gm-SE and obtained high levels of specific IgE and IgG1 vs. low levels of IgG2a. This pattern is indicative of a Th2-type response induced by the oral immunization with CT as an adjuvant.

Comparison of immunological responses of wt-SE and gm-SE
In our model we obtained a comparable level of IgE and IgG1 antibodies between those produced by gm-SE-sensitized mice and those obtained from wt-SE-sensitized mice. Moreover, using a specific IgE ELISA inhibition test, we observed that IgE antibodies produced after immunization with wt-SE were able to inhibit completely the binding of IgE obtained by oral administration with gm-SE. Similar results were obtained in the opposite case. These data strongly suggest that gm-SE shared the same allergenic components with wt-SE. In particular, the complete IgE inhibition (up to 100%) obtained in the assay performed on gm-SE as an antigen, wt-SE as an inhibitor, and sera from gm-SE sensitized mice sustains the hypothesis that the exogenous CP4-EPSPS protein is not able to induce an IgE response under our conditions.

T lymphocytes and cytokines have been demonstrated to play a pivotal role in the induction of IgE response. Therefore, to further assess the suitability of our animal model for testing the allergenicity of genetically engineered foods, we evaluated T cell responses to the SEs. We found that spleen cells from mice treated with gm-SE exhibited the same proliferative responses to wt-SE in vitro stimulation compared with homologous antigen stimulation and vice versa. Antigen stimulation of spleen cells from mice sensitized either with wt-SE or with gm-SE induced significantly increased and comparable production of IL-4, IL-5 and IFN-g.

Taken together, our data on the humoral and cellular response demonstrate that in sensitized mice, we observed a predominantly Th2-type immune response, with increased food-specific IgE and IgG1 antibodies and concomitant production of IL-4 and IL-5. Similar antibodies and cytokine profiles have been found in human beings with food allergy. We could also show that gm-SE induced an immunological response comparable with that induced by wt-SE. Moreover, results obtained in our model by specific IgE ELISA inhibition and by antigen-specific T cell proliferation demonstrated that wt-SE and gm-SE shared B and T epitopes. Finally, the absence of humoral and cellular response against control proteins (irrelevant proteins) in the same assays confirms that our model is specific for the components of SE. In conclusion, considering that there is no single predictive bioassay available to assess the allergenic potential of proteins in humans4, the murine IgE sensitization model described provides valuable information regarding the allergenicity of modified soybean derived from biotechnology, and it could be a suitable system for the in vivo testing of genetically modified foods.

Agricultural biotechnology offers the promise to produce crops with improved agronomic characteristics and enhanced consumer benefits. Foods produced through agricultural biotechnology are already reaching the consumer marketplace, and there is a growing concern that introducing foreign genes into food plants may have an unexpected and negative impact on human health. As a result, the safety evaluation of transgenic food must be subject to a careful and complete safety assessment before commercialization. Of particular interest with genetically engineered organisms is the risk of allergic reactions. It is possible that the manipulations of genetic engineering may increase the potential risk of food allergy. Nevertheless, biotechnology may be able to modify foods to remove or change the proteins that cause allergy, offering the potential of making nutritious foods available to people who presently cannot eat them.

The safety evaluation of transgenic food is relatively easy when the allergenicity of the gene source is known. Nordlee et al. showed that an allergen (2S-albumin) from a food known to be allergenic (Brazil nut) can be transferred into another food (soybeans) by genetic engineering.7 On the contrary, the safety assessment approach could be a very complex problem when the expressed protein comes from a source that is not known to be allergenic, as in the case of Roundup Ready soybean, or when gene products are down-regulated for hypoallergenic purposes. Although strategies exist for such assessments, improvements should be considered to develop testing strategies to examine the allergenicity of genetically engineered food. In this context, animal models would be of considerable benefit if they facilitated a more direct evaluation of the ability of proteins to induce allergic responses in vivo.

Note: Animals were housed and treated according to the local guidelines for animal care (D.L. 116/92, which has implemented in Italy the requirements of the European Directive 86/609/EEC on laboratory animal welfare).


1. Sampson HA. (1999) Food allergy. part 1: immunopathogenesis and clinical disorders. J Allergy Clin Immunol 103, 717–28.

2. Codex alimentarius commission twenty-third session, FAO, Rome, 28 June–3 July 1999.

3. James C. Preview: Global Status of Commercialized Biotech/GM Crops: 2005. ISAAA Briefs No. 34, ISAAA, Ithaca, NY, 2005

4. Report of a joint FAO/WHO expert consultation on allergenicity of foods derived from biotechnology, 22–25 January 2001, FAO, Rome, Italy.

5. Gizzarelli F, Corinti S, Barletta B. Iacovacci P, Brunetto B, Butteroni C, Afferni C, Onori R, Miraglia M, Panzini G, Di Felice G, & Tinghino R. (2006) Evaluation of allergenicity of genetically modified soybean protein extract in a murine model of oral allergen-specific sensitization. Clinical and Experimental Allergy 36, 238–248

6. Li X-M et al. (2000) A murine model of peanut anaphylaxis: T- and B-cell responses to a major peanut allergen mimic human responses. J Allergy Clin Immunol 106,150–8

7. Nordlee JA et al. (1996) Identification of a brazil-nut allergen in transgenic soybeans. N Engl J Med 334, 688-692

Raffaella Tinghino
Istituto Superiore di Sanità
Rome Italy

Patricia Obregon

In 2004 an estimated forty-two million human immunodeficiency virus (HIV) infections were found throughout the world, and more than 95% of the cases and deaths from AIDS occurred in developing countries. The continuing spread of the epidemic and the high rate of infected people in that area of the world have raised the need to establish urgent preventive measures and extend antiretroviral therapy access (HAART). However, an ongoing serious concern is that these crucial measures are only accessible to a minor number of people who need them, and the introduction of HAART has been unable to slow the progress of HIV in these countries. Therefore, the development of an efficient and cost-effective HIV vaccine has become an urgent need, and the advance of new strategies will be necessary to halt the spread of HIV/AIDS.

Transgenic plants have emerged as a promising technology to create recombinant biopharmaceutical proteins and vaccines. They offer a spectrum of exclusive advantages, and so their potential to be used as bioreactors for the production of therapeutic molecules is a current area of research intensively explored.

• Plant systems produce full-length mammalian proteins that appear to be processed similarly to their native counterpart with appropriate folding, assembly, and post-translational modifications. In fact, a wide variety of complex and valuable foreign proteins can be expressed efficiently in transgenic plants.1

• Production of recombinant proteins in plants offers economic advantages. It has been estimated that the cost of producing proteins in transgenic plants may be 100 fold lower than in transgenic animals or mammalian cell cultures, and the possibility of using the plant tissue as a carrier for oral delivery could also diminish the expensive step of recombinant protein purification. The use of plants as an expression system for recombinant protein production would be at least as economical as traditional industrial facilities (by fermentation processes or bioreactor systems).

• Some important potential advantages of producing recombinant proteins in plants for vaccine development include: production as well as storage can occur near the site of use; the heat stable formulations eliminate the need for refrigeration; and the need for hypodermic delivery (needles) can be also eliminated.

These criteria indicate that a plant-based production system is a promising technology for the generation of easily distributed, affordable HIV vaccines. Additionally, one of the most obvious benefits of using plants as protein expression systems is the potential for scale up. Potentially vast amounts of recombinant protein could be produced simply by increasing the planting area.

One of the major obstacles to recombinant protein production in plants is the low level of protein expression. At least 45 antigens have been successfully expressed in plants but their levels of expression are low, between 0.0005 – 0.3 % of total soluble protein (TSP).1 Moreover, because a protein purification step is needed for the final product of a plant-based production system, the final protein yield is therefore diminished. In this regard it has been estimated that a protein expression level compatible with purification technologies could be represented by 1% TSP.2 Therefore, improving the foreign protein production yield in plants is a crucial objective that will have a significant impact on the economic feasibility of plants as bioreactors. Different strategies — use of novel promoters, codon optimization, improvement of protein stability, targeting of recombinant proteins to intracellular compartments, and improvement of downstream processing technologies — have been used to improve production. However, increasing the yield of foreign protein in plants is still a major goal of plant biotechnology, for which further optimization strategies are required.

Experimental design and engineering of HIV-1 antigen-antibody fusion molecule
One of the most important HIV antigens likely to form part of any HIV vaccine is the HIV-1 p24 capsid protein. It belongs to the group of proteins known as Gag proteins and constitutes one of the major structural proteins of HIV. Studies have shown cross-clade antibody responses against conserved epitopes of Gag in HIV infected individuals, and the absence of anti-Gag antibodies is indicative of disease progression. Moreover, T-cell immune responses are probably the most important protective mechanism against HIV, and the HIV-1 p24 antigen is the reported target of T-cell immune responses in infected individuals.3,4 Therefore, efforts are being made to develop more efficient and feasible expression strategies for production and therapeutic use of recombinant p24. The HIV-1 p24 protein has been recently expressed in tobacco plants by different strategies. However, although research demonstrates that HIV-1 p24 can be successfully expressed in tobacco plants, consistently low expression levels of the p24 antigen in stably transformed plants are reported.5

Mammalian immunoglobulin (Ig) is the only class of molecules reported to reliably reach high expression levels in transgenic plants (IgG antibodies, 1% TSP, and the secretory immunoglobulins, IgA, 5%-8% TSP).6,7 There is a significant difference in protein expression levels between monomeric antigens and polymeric Ig in plants; the reason for this difference is still not determined. Nevertheless, we decided to examine the potential of antibody sequences to enhance recombinant antigen expression levels in transgenic plants. Furthermore, since our studies intended to explore vaccine candidates in transgenic plants, we decided to investigate the design and engineering of an antigen-antibody fusion molecule capable of retaining the immunogenicity of the antigen fusion partner while incorporating the functional components of the antibody fusion partner.

We established two molecular approaches for the expression and production of the HIV-1 p24 antigen in transgenic tobacco plants. First, as a control, we engineered the unmodified single HIV-1 p24 gene (p24) to be expressed under the control of the constitutive CaMV 35S promoter and its translation product targeted to the plant endomembrane system. Second, we engineered the HIV-1 p24 antigen-antibody fusion molecule. Immunoglobulins are polymeric molecules constituted by four monomeric chains: two identical heavy chains and two identical light chains. Each chain, heavy and light, possesses two different regions: one variable region involved in the recognition of antigen; and one constant region required for assembly of the chains. Moreover, the heavy chain constant region is also involved in activation of immune effector functions. Thus, in this second strategy, the HIV-1 p24 antigen was fused to the Cα2 and Cα3 constant region domains of a human immunoglobulin (IgA) heavy chain (p24/Cα2-Cα3) and its expression in tobacco plants was investigated.

After sequencing analysis, both p24 genetic constructs were separately cloned into a pMON530 plant expression vector. A specific mouse IgG 5’ leader sequence had been previously included upstream of each transgene into the vector in order to direct the recombinant protein to the plant endomembrane system. Subsequently, each genetic construct was individually transferred into Agrobacterium tumefaciens strain LBA4404 by electroporation. Transformation of Nicotiana tabacum (var. Xantii) plants was made by co-cultivation with Agrobacterium transformants, and the effect of the random nature of Agrobacterium-mediated gene insertion on the protein expression levels was minimized by meticulously selecting the highest expresser transgenic plants of each construct.

Enhancing HIV-1 p24 antigen expression by IgA heavy chain fusion partner
Two genetic constructs, p24 and p24/Cα2-Cα3, were engineered for expression in tobacco plants. As a first step, after generation and selection on antibiotics of transgenic plants transformed with either one or the other chimeric construct, the expression of both plant-derived recombinant proteins, p24 or p24/Cα2-Cα3, was analyzed. By using transgenic plant protein extract and specific anti-HIV-1 p24 antibodies in ELISA and Western blot, the accumulation of correctly folded recombinant full-length HIV-1 p24 was demonstrated in both cases.

The domains Cα2 and Cα3 are responsible for the dimerization of α-heavy chains in immunoglobulin A (IgA) molecules under natural conditions. In our study, the expression of the full-length p24/Cα2-Cα3 fusion molecule was also confirmed, and its assembly into dimer molecular form indicates that the IgA Cα2-Cα3 domains fragment retains its native capability to assemble when expressed in plants as the p24/Cα2-Cα3 fusion partner.

The HIV-1 p24 gene DNA sequence was identical in both constructs, and the plant codon usage was not optimized in either case. Thus, since the p24 antigen was efficiently expressed in tobacco plants by both strategies, the next step was to investigate the level of HIV-1 p24 expression in each case. An important difference in the p24 protein expression levels was observed when the HIV-1 p24 gene was expressed after fusion with the human IgA Cα2-Cα3 heavy chain sequence. A significant increase of up to 13-fold in the overall expression levels of p24 antigen in p24/Cα2-Cα3 transgenic plants was achieved (1.4% TSP) compared to those of p24 antigen when the HIV-1 p24 gene was expressed alone (0.1% TSP) in transgenic tobacco plants.

Plants are very efficient at producing immunoglobulins, probably because the endomembrane system of plant and mammalian cells are organized in an identical manner. In addition, plant chaperones homologous to mammalian chaperones have been described within the plant endoplasmic reticulum (ER), and their interaction with Ig chains determines the efficiency of protein folding and assembly.8 However, although both Ig heavy and light chains can be expressed individually in plants, enhancement of recombinant Ig expression levels has been reported when light and heavy chains are co-expressed in transgenic plants.6 These results suggest that the assembly status of the molecule is a determinant of stability, and accordingly, the observation of p24/Cα2-Cα3 homodimers during our study suggests that the addition of Cα2 and Cα3 domains in the p24/Cα2-Cα3 fusion molecule may confer some structural advantages in terms of recombinant protein stability expressed in plants.

At the same time, sub-cellular targeting plays an important role in determining the yield of recombinant protein, as it strongly influences the processes of protein folding, assembly, and post-translational modifications. Antibodies targeted to the secretory system usually accumulate to significantly higher levels than those of antibodies expressed in the cytosol. Moreover, the stability of antibodies is lower in the apoplast than in the lumen of ER. In our study, sub-cellular trafficking of both p24 and p24/Cα2-Cα3 proteins was also analyzed by immunoprecipitation and pulse-chain experiments in transgenic tobacco protoplasts. Our results demonstrated that HIV-1 p24 recombinant protein is efficiently secreted to the extra-cellular space when it is expressed alone. Conversely, HIV-1 p24 fused to human heavy chain Cα2-Cα3 domains is retained inside the cell. Proteins that accumulate in the secretory system are secreted into the apoplast in the absence of further targeting information. IgA antibodies accumulate predominantly within the plant endomembrane system and, in part, are targeted to vacuoles. The presence of a cryptic sorting signal in the IgA Cα3 tailpiece has been identified as an element responsible for this vacuolar targeting.9 However, despite this sub-cellular targeting, IgA is expressed at high levels in plants and, indeed, at higher levels than other (IgG) antibodies, which are secreted to the extracellular space. In accord with these observations, we confirmed that p24/Cα2-Cα3 fusion molecule contains that same sorting signal, and taken together, our results indicate that the addition of IgA Cα2-Cα3 sequence may divert the recombinant HIV-1 p24 antigen to a different sub-cellular compartment than the HIV-1 p24 expressed alone.

For immunogenicity testing of the p24 antigen in the context of an antigen-antibody fusion molecule, eight groups of five BALB/c (H-2d) mice were subcutaneously immunized at day 0 with different doses (3 µg , 10 µg, and 30 µg) of either purified plant-based p24/Cα2-Cα3 fusion protein or E. coli p24-His (as a positive control). A group injected with PBS buffer was used as a negative control. In all cases, alum was included as an adjuvant. Mice were boosted at 3 and 8 weeks, and samples collected at 0, 3, 8, and 11 weeks. Importantly, serum analyses revealed that plant-derived p24 is immunogenic in mice when expressed as the p24/Cα2-Cα3 fusion molecule, under a dose-dependant response, with highest titers after priming with 10 µg of recombinant protein. Furthermore, T-cell epitopes were conserved in plant-derived HIV-1 p24, as T-cell responses were elicited in mice against both plant-derived as well as bacteria-derived recombinant p24 antigen.

In terms of vaccine development, we foresee some applications where it may be preferable to retain the Ig sequence on the final recombinant fusion protein. IgA is the most abundantly Ig produced in the body. It is localized on both sera and mucosal surfaces. The main route for the HIV infection to be contracted in more than 90% of HIV-infected individuals is via the mucosal surfaces of the genital tract or rectum, or through breastfeeding, and studies have shown that an important implication of secretory IgA, which constitutes the main class of antibody in this area, is a protective mucosal immune response against HIV. More recently, the so-called IgA Fc α-receptors have been defined as the most likely candidate to initiate potent effector immune functions upon binding to serum IgA through heavy chain constant domains.10 In this context, dimers of the p24/Cα2-Cα3 fusion molecule may bind to Fc α-receptors to trigger a specific immune response.

We have demonstrated that Ig fusion partners can be used as an alternative strategy for enhancing recombinant antigen expression in plants. There are still other factors to be considered before this technology can be ready for practical use. However, the antigen-antibody fusion strategy might lead to a new technology with important implications for both the economic viability of using plants as bioreactors for recombinant protein production and the development of a strategy to design new vaccines with enhanced specific immunological properties against HIV and other diseases.


1. Arntzen C et al. (2005) Plant-derived vaccines and antibodies: potential and limitations. Vaccine 23(15), 1753-6

2. Kusnadi AR et al. (1998) Processing of transgenic corn seed and its effect on the recovery of recombinant betaglucuronidase. Biotechnol. 60, 44-52

3. Dyer WB et al. (2002) Correlates of antiviral immune restoration in acute and chronic HIV type 1 infection: sustained viral suppression and normalization of T cell subsets. Aids Res human Retroviruses 18, 999-1010

4. McMichel AJ et al. (2001) Cellular immune responses to HIV. Nature 410, 980-7

5. Zhang GG et al. (2002) Production of HIV-1 p24 protein in transgenic tobacco plants. Mol. Biotechnol. 20, 131-136

6. Hiatt A et al. (1989) Production of antibodies in transgenic plants. Nature 342, 76-78

7. Ma JK et al. (1995) Generation and assembly of secretory antibodies in plants. Science 268, 716-719

8. Nuttall J et al. (2002) ER-resident chaperone interactions with recombinant antibodies in transgenic plants. Eur J Biochem 269, 6042-51

9. Hadlington et al. (2003) The C-terminal extension of a hybrid immunoglobulin A/G heavy chain is responsible for its Golgi-mediated sorting to the vacuole. Mol. Biol. Cell 14, 2592-2602

10. Otten MA et al. (2004) The Fc receptor for IgA (Fc-RI, CD89), Immunol. Lett. 92, 23-31

Patricia Obregon
Dept. Of Cellular and Molecular Medicine
St. George’s Hospital University of London
London SW17 0RE

ISB News Report
1900 Kraft Drive #103
Corporate Research Center
Virginia Tech
Blacksburg, VA 24061

The material in this News Report is compiled by NBIAP's Information Systems for Biotechnology, a joint project of USDA/CSREES and the Virginia Polytechnic Institute and State University. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture, or Virginia Tech. The News Report may be freely photocopied or otherwise distributed without charge.

ISB welcomes your comments and encourages article submissions. If you have a suitable article relevant to our coverage of the agricultural and environmental applications of genetic engineering, please e-mail it to the Editor for consideration.

Ruth Irwin, Editor (

To have the News Report automatically e-mailed to you, send an e-mail message to with your request.
Connect to for internet access to ISB News Reports, textfiles, and databases.

Information Systems for Biotechnology, Virginia Tech, 1900 Kraft Drive, Suite 103, Blacksburg, VA 24061, tel: 540-231-3747, fax: 540-231-4434, e-mail: