"Biotechnology is a key component of our growing agricultural economy," said Conner. "USDA's program will help the biotechnology sector become better stewards by focusing on the implementation of best management practices so that problems can be prevented."

USDA's biotechnology initiative complements a program called, "Excellence Through Stewardship," which is already underway in the biotechnology industry. While industry's program is focused on quality management to ensure product integrity of biotech-derived plant products throughout the product life cycle, USDA will emphasize the quality of the process for safely introducing these GE organisms in compliance with federal regulations.

The Biotechnology Quality Management System was developed to be as inclusive as possible so that a broad array of participants could participate. It will consist of two program levels, based on domestically and internationally recognized quality management systems. Specifically,

USDA's Animal and Plant Health Inspections Service (APHIS) intends to oversee the Biotechnology Quality Management System program in partnership with USDA's Agricultural Marketing Service (AMS), which will manage the audit component of the program and accredit third party auditors. Audits will verify that participants have procedures in place, and that they are performed correctly to meet the regulatory requirements for any given GE field trial or movement. As part of the program's emphasis on preventive measures, participants will be encouraged to correct deficiencies discovered in an audit before compliance problems develop.

The Biotechnology Quality Management System and its associated audits will complement, not replace, APHIS' current regulatory compliance and inspection process by focusing on planning and good management practices that can improve a participant's ability to meet regulatory requirements. The current inspection program will continue to cover specific permits and notifications to ensure compliance with regulations.

APHIS, in partnership with AMS, will implement the voluntary system through an agency notice and participation in the program will not be a regulatory requirement. APHIS also will work proactively to provide outreach and guidance to those companies and researchers that choose to participate and develop these quality management programs.

Currently there are several audit-based, quality verification systems in operation throughout USDA's marketing and regulatory programs mission area, such as AMS' Process Verified Programs. APHIS regulates the confined field release, interstate movement and importation of GE organisms. APHIS currently ensures compliance with regulations through inspections conducted at critical stages; consistent and appropriate enforcement actions; and comprehensive record keeping and reporting requirements. Quality management systems are intended to improve regulatory compliance by fostering a company's commitment to sound controls, quality management practices, and effective compliance with federal regulatory requirements.

It may not be entirely surprising that the release of a glyphosate tolerant alfalfa crop has been received as highly controversial. Roundup Ready^® alfalfa is different from other GE crops in that its pollen is easily dispersed by honey bees over distances that can exceed three kilometers. Transgenic pollen is likely to fertilize flowers of untransformed alfalfa plants that grow on an estimated 20 million acres throughout the United States. The resulting contamination of seed with foreign DNA may compromise the perceived quality, sale, and export market of alfalfa. Furthermore, introduction of yet another Roundup Ready^® large-acreage crop results in an inevitable increase in glyphosate usage while encouraging the further establishment of glyphosate resistant weed populations.

All these issues resulted in a recently imposed injunction, in effect revoking the 2005 approval of Roundup Ready^® alfalfa.¹ Although preparation of an environmental impact statement, to be released about a year from now, may convince the court to issue permanent injunctive relief, there will continue to be concerns about the production of Roundup Ready^® alfalfa. It seems, therefore, imperative to develop alternative approaches to genetic engineering that make it possible to genetically modify crops while addressing most of the controversial issues associated with the original GE alfalfa crop.

A recent study published in Transgenic Research demonstrates that the quality of alfalfa can be enhanced without incorporating marker genes or other types of foreign DNA into the crop.² One aspect of this new approach is the employment of a new marker-free transformation procedure. This method was developed by first incubating two-day old seedlings for 16 hr at 4°C. After excision of cotyledons at the apical nodes comprising meristematic tissues (Fig. 1A), the resulting explants were infected with a highly virulent Agrobacterium C58/pMP90 donor strain carrying the reporter β-glucuronidase (gus) gene (Fig. 1B).

Unique methods were then used to enhance contact between acceptor and donor cells. Instead of applying a conventional agitation or vacuum infiltration step, emerged seedlings were vigorously vortexed for an extensive period of time (~30 min). This procedure did not cause any irreversible damage to treated seedlings. Indications for the extent of transient DNA transfer were obtained by assaying for β-glucuronidase (gus) activity.

Infection from Agrobacterium resulted in high levels of transient transformation. The cut and vortex-infected seedlings were inserted vertically into hormone-free media for a short recovery time, during which new shoots arose from the cut surfaces of about 60% of explants. After 14 days, the explants had developed into rooted seedlings that were planted in soil and transferred to the greenhouse. Subsequent analyses of upper new leaves of five-week old plants demonstrated 7% of these leaves stably expressed the gus gene in most or all cells ("all-blue" leaves), whereas an additional 17.5% expressed this gene in part of the tissue (Fig.1C).

Extensive gus expression in upper leaves was expected to be indicative for transformation of meristematic and germ line cells. This theory was confirmed by allowing transformed plants to mature and set seed in the greenhouse. Subsequent analyses of T1 progenies demonstrated successful transgene transmission. Although segregation ratios in some cases deviated from the 3:1 ratio expected for fully transformed T0 plants, all of the independent families tested contained at least some siblings expressing the gus gene. DNA gel blot analyses of randomly-chosen T1 plants confirmed the integrity of transmitted T-DNAs, and estimated the average copy number of this element at 1.6 (Fig. 1D and data not shown). Collectively, the data demonstrated that alfalfa can be transformed without selectable marker genes.

This new method provides several additional advantages in addition to avoiding the stable integration of bacterial selectable marker genes. First, it limits the time, materials, and resources required for complex in vitro manipulations, while also eliminating the risk of somaclonal variation that is associated with both hormone treatment and callus formation. Second, the method substantially reduces the amount of time from transformation to seed set from about seven weeks for conventional systems⁵ to five weeks. Third, the in planta transformation method has been applied successfully to a commercial variety, whereas the conventional methods require very specific highly regenerable genotypes such as RegenSY that have little commercial value.³

To demonstrate that the new transformation method could be used for production of intragenic plants displaying an enhanced quality trait, a silencing construct targeting the native caffeic acid o-methyltransferase (Comt) gene was positioned within an alfalfa-derived transfer DNA.⁴ Alfalfa plants were transformed as described above and allowed to mature in the greenhouse. Polymerase chain reaction (PCR)-based genotyping of 1,000 five-week old plants identified 2.4% that contained the modified P-DNA. Stem sections were isolated from intragenic progeny plants derived from eight randomly chosen original transformants and assayed for lignin accumulation. This analysis demonstrated reduced lignin levels in three of eight cases (Fig. 1E and data not shown). Studies performed by others have already shown that these reduced lignin levels enhance the value of alfalfa as feed for dairy cattle.⁵

A gardener at Colonial Williamsburg picks tobacco budworm larvae from his crops as he explains to visitors that these insects have been eating plants since the early days. As farmers around the world have battled voracious insect pests, the agrochemical industry has come to their aid with various insecticides. Spraying crops with these chemicals has worked to kill the invaders and increase crop production until the insects develop resistance. When this happens farmers switch to another chemical to rid their fields of these pests. Meanwhile the environmentalists are concerned about the effects of insecticides on other wildlife and humans.

In the 1990's the agricultural industry took a new approach by creating transgenic crops like Bt cotton and Bt corn. The Bacillus thuringiensis bacteria have genes that produce proteins toxic to these insect pests. Specific Bt genes have been genetically engineered into various crops so the plants make enough Bt toxin to kill insects that feed on them. The question remains: Will insects eventually develop resistance to Bt toxin?

The EPA recommends a "high-refuge" strategy that requires farmers growing Bt cotton to also grow adjacent fields of non-Bt cotton. Insects feeding on normal cotton will not be pressured to develop resistance and will mate with insects from Bt cotton fields. If resistance has emerged in the Bt fields, the offspring of these matings are unlikely to produce resistant progeny since the mutation is recessive. The development of Bt resistance is also the interest of scientists who want to assist farmers with pest management. Knowing which genes have been changed in Bt resistant insects will enable the agricultural industry and governmental agencies to devise new strategies to maintain the efficacy of transgenic crops.

Our research group uses a genetic approach to identify and characterize Bt resistance genes. Since there is no Bt resistance in field populations of Heliothis virescens (Hv), we study a Bt-resistant laboratory strain called YHD2, developed by Fred Gould at North Carolina State University. Hv collected from the field were reared on a diet containing very low doses of Bt toxin Cry1Ac. Individuals that survived this treatment were mated, and offspring were selected on the toxin. This mating and selection process yielded a YHD2 strain that is 10,000 times more resistant to Cry1Ac than the susceptible strain. This highly resistant YHD2 strain has multiple Bt resistance genes; these genes are fixed, and resistance is recessive.

Our genetic approach uses two crosses prepared as single-pair matings. The first cross is between YHD2 females and susceptible males to yield F1 families. Individuals from the F1 families, selected on low dose Cry1Ac diet for seven days, grow slowly because of the recessive nature of Bt resistance. A second backcross is constructed between F1 and YHD2 individuals, with offspring selected on toxin in a growth inhibition assay. Backcross families showing a bimodal distribution in size after seven days on toxin are analyzed for resistance. Individuals from female informative families are used to map the resistance gene on a particular linkage group, since there is no crossing-over during meosis when the female backcross parent is the F1. The order of DNA markers on a particular linkage group can be determined using individuals from a male informative backcross family, since crossing-over occurs in the F1 male. Southern blots are prepared using DNA from small and large offspring from female and male informative backcross families plus parents and grandparents. DNA markers already assigned to the 31 linkage groups are used as probes to identify the linkage group that contains the Bt resistance gene, and then to order DNA markers on that linkage group. The goal is to find DNA markers on either side of the Bt resistance gene close enough to positionally clone the resistance gene. Success of this approach depends on the availability of DNA markers on each of the 31 linkage groups and polymorphism for these markers in the resistant and susceptible strains.

Our approach to identify genes involved in Bt resistance mechanisms utilizes candidate genes. Cadherins and N-aminopeptidases bind Bt toxins, midgut proteases activate Bt toxins, alkaline phosphatases may function as oligomeric toxin receptors, and N-acetylglycosamine transferases glycosylate GPI-anchored midgut membrane receptor proteins. Our research shows that the cadherin-like gene on linkage group 9 (LG9) in YHD2 has a retrotransposon inserted in the fifth cadherin repeat (Fig. 1)¹. As a result the cadherin-like protein translated from this mutant gene (r1) is truncated and is never deposited as an integral protein in the midgut membrane, where it would normally bind the activated Bt toxin. Examining two cDNA libraries, one constructed from the YHD2 Bt resistant strain and one made from the susceptible Hv strain, led to this discovery. By designing a number of degenerate primers to the cadherin-like gene whose protein product is known to bind Bt toxin in Manduca sexta, gene fragments were PCR amplified and used to probe these two libraries. Sequencing of positive clones from the libraries revealed a large foreign DNA insertion in the cadherin-like gene from YHD2. In addition the size of the messenger RNA for the cadherin-like gene was larger in the Bt resistant YHD2 strain compared to the susceptible strain. Further studies by Juan Luis Jurat-Fuentes showed the r1 mutant fromYHD2 does not express the full-length cadherin-like protein.²

In the pink bollworm, Pectinophora gossypiella, Morin et al. located three independent mutations in the cadherin-like gene that conferred Bt resistance to insects found in field populations of Bt cotton³. Xu et al. discovered a stop codon in the cadherin gene sequence of a GYBT Bt resistance strain of Helicoverpa armigera.⁴ In a study to determine the frequency of resistance to Cry1Ac in field populations of Hv, Fred Gould set up 2000 matings of Hv field-collected males with YHD2 females. The offspring from each mating were selected on Bt toxin. Any mating that produced offspring, half of which would grow on Bt toxin, was considered to contain a Bt resistant allele in the field-collected male parent. The study found three such males in field populations of Hv, which were frozen for later analysis.⁵ Years later when the cadherin-like gene had been implicated in a Bt resistance mechanism, our laboratory sequenced the gene from one of the frozen, field-collected males. This male contains two alleles for the cadherin-like gene, one similar to the susceptible and one containing a large insertion of another type (r2) in the membrane-proximal region of the gene. Offspring from the cross between this male (r2,s) and the YHD2 female (r1/r1) can grow on Bt toxin if they received the r2 allele from their father and the r1 allele from their mother (unpublished results). Thus, evidence is mounting that the cadherin-like gene is involved in resistance to Bt toxin in Lepidoptera.

The r1 mutation in the cadherin gene accounts for up to 80% of Bt resistance in YHD2. This fact and the analysis of other crosses suggest multiple Bt resistance genes exist in this strain. In order to study these genes, it was necessary to separate the cadherin-like r1 gene from the rest of the Bt resistance genes. The idea was to create a YEE strain lacking the cadherin r1 mutation, a YFO strain containing only the cadherin r1 mutation, and a reconstituted YHD3 strain having all the Bt resistance genes. To construct these three strains, a YHD2 female was mated to a susceptible male, and F1 offspring were selected on a low dose Cry1Ac toxin diet. F1 females were backcrossed to YHD2 males and the BC offspring were selected on Bt toxin for seven days, weighed, and transferred to normal diet to complete development. The smallest individuals from the BC were used to create the YEE strain. The presence of the r1 mutation in the cadherin gene can be determined by PCR using forward and reverse primers designed around the insertion site of the retrotransposon in the cadherin gene. Matings were set up between individuals from YEE families, offspring were selected on Bt toxin, and the parents of these matings were analyzed for the presence of the r1 mutation. Offspring from parents lacking the r1 mutation and growing well on toxin were used to create the next generation of the YEE strain. This process continued until no r1 mutation was observed. Matings and selection on toxin have continued for 58 generations to fix the Bt resistant gene(s). To create YFO, BC individuals of medium size were mated with susceptible individuals, offspring were selected on a low dose of Bt toxin and the YFO parent of these crosses was tested for the presence of the r1 mutation. This process of mating YFO resistant to susceptible individuals was continued for a total of six generations in order to remove other Bt resistance genes. After this, YFO individuals were mated and selected on toxin, and parents were analyzed for presence of the r1 allele. Offspring from families whose parents contained the largest number of r1 alleles were used in constructing the next generation. Matings continued until the r1 allele was fixed in YFO and then to maintain the strain. Mating the largest BC offspring, selecting for maximal growth on Bt toxin, and gradually increasing the toxin concentration for 58 generations created YHD3. In addition three analogous strains were created in like manner and designated RER (no cadherin r1), RFO (only cadherin r1) and REE (all resistant genes).

Analysis of the Bt resistance genes in these strains has been accomplished in two ways. Individuals of a resistant strain are crossed to the susceptible individuals to create F1 families, then backcrosses are set up between F1 and the resistant individuals. Offspring from both crosses are selected for resistance on Cry1Ac. As explained earlier, BC families in which the F1 parent is a female and her offspring exhibit a bimodal distribution of weights after toxin treatment can be used to map the resistance gene. BC families where the F1 parent is a male are used to order DNA markers on the linkage group containing the resistant gene. The second way of analyzing resistance is to cross two resistant strains to form the F1 and then backcross F1 to the most resistant strain.

Using this genetic approach, we have identified a resistance gene on LG2 and have closely ordered DNA markers on either side of it. The initial cross between YFO and YHD3 followed by a backcross to YHD3 gave us this result. An initial cross between RER and the susceptible strain also confirmed this fact. YEE appears to have two resistance genes. Crosses between resistant strains are being constructed to analyze the other gene, possibly on LG10. A Hv Bac library provides an excellent resource tool for walking to the resistance gene from closely proximinal DNA markers. The recently sequenced Bombyx mori genome assists our efforts to find genes closely linked to Bt resistance, as our use of this genetic approach moves forward to identify Bt resistance genes.

Two reports^1,2 describe the characterization of experimental transgenic maize plants containing a wheat genomic DNA fragment encoding the Glu1-Dx5 gene. This gene encodes a high molecular weight glutenin, a storage protein that is a component of gluten, the complex polymer that gives wheat flour dough its elasticity. The motivation for transferring this gene to corn was to determine the feasibility of producing grain with altered flour properties that could allow development of improved or novel maize-based food products. In addition, production of individual wheat proteins in corn could be a valuable approach for studying gluten sensitivity because it allows proteins to be evaluated on an individual basis, something that is difficult to do with wheat-derived proteins.

In their initial evaluation of these plants, the authors unexpectedly determined that the wheat genomic fragment exhibited maternal inheritance in all four of the events that were characterized.¹ The second report² characterizes the mechanism that confers maternal inheritance to this transgene and demonstrates that this transgene can be used to prevent the transfer of another transgene (an herbicide resistance gene) to other plants through pollen. The ability to prevent the transfer of a transgene through pollen could have application in commercial production of transgenic crops.

In an effort to understand the biological mechanism conferring maternal inheritance to transgene loci containing the wheat genomic fragment, the authors examined the pollen of the transgenic plants. Because the transgene was inherited maternally, these plants were necessarily heterozygous at the transgene locus. Corn is a diploid species, but the pollen is haploid. In a transgenic corn plant that is heterozygous at a transgene locus, half the pollen would normally be expected to contain the transgene and the other half would not.

In the transgenic corn described in these studies, transgene DNA is present in the pollen, but the pollen exhibits reduced viability. In some cases, two classes of pollen were visually discernible, and one of these classes was not viable. The authors hypothesize that the pollen containing the transgene is among that which is non-viable. This leads to the hypothesis that pollen containing the transgene locus fails to develop normally and is non-viable, although it is still shed, while non-transgenic pollen from the same plant develops normally, is shed, and is viable. If this hypothesis is correct, the transgene locus functions as a pollen-specific gametocide, explaining its maternal inheritance.

It is not clear why this DNA fragment functions as a gametocide in corn. In wheat, the Glu1-Dx5 gene produces a storage protein that accumulates only in the endosperm of seeds. In the transgenic corn described in this study, the Glu1-Dx5 gene produces a protein of expected size in the endosperm, but also mediates the pollen-specific gametocide function that results in maternal inheritance of the transgene. The Glu1-Dx5 protein gene product is not detectable in the pollen of transgenic corn. It is possible that the gametocidic function is contained on the wheat genomic fragment but is not related to production of the Glu1-Dx5 transgene product. Intriguingly, when the Glu1-Dx5 gene is controlled by a maize endosperm-specific promoter, the transgene does not exhibit the gametocidic function, indicating that the Glu1-Dx5 coding sequence is not sufficient to cause the gametocidic function, and that the gametocidic activity is conferred by DNA present in the wheat genomic fragment but lacking in the construct containing the corn promoter, which includes several kilobases of DNA upstream of the Glu1-Dx5 translational start site, including the promoter of this gene.

Many important crop species, including maize, sunflower, and alfalfa, produce seeds mainly by cross pollination. With segregated grain markets, cross pollination can be a problem because it can occur between production fields of different types of grain. This can result in the inadvertent transfer of traits from one production field to another where these traits may not be desired.

Several methods have been proposed to control unwanted outcrossing, including use of mutants³ or management strategies such as detasseling.⁴ Maternally inherited transgenes are another solution to this problem of outcrossing and can be produced by chloroplast transformation.⁵ Unfortunately, this technology is currently only available for a few species and has not been routinely successful in corn. A pollen-specific gametocide such as the one described² could be deployed so that it is genetically linked to a transgene of interest. In this system, pollen containing the transgene is not viable, and therefore cannot pass the transgene on to progeny. Non-transgenic pollen remains viable and is produced in quantities sufficient for fertilization of the crop. This approach could potentially be used to confer maternal inheritance to any transgene, regardless of the trait conferred by the transgene.

Several problems remain with this system as described. One problem is that because the transgene locus is maternally inherited, the transgene must be deployed as a heterozygote. Heterozygous plants would produce seed that segregate for the transgene locus, so only half of the seeds produced would be transgenic. There are several ways to overcome this deficiency. One approach would be to use a mechanism of selection for those plants containing the transgene, for example by incorporating a gene for herbicide resistance into the transgene locus. Another approach could involve incorporating a gene into the transgene locus that makes transgenic seeds distinguishable to mechanical sorters, for instance a seed color gene.

A problem with the use of Glu1-Dx5 as a transgenic pollen-specific gametocide is that it is not completely effective, so in some cases low levels of transgenic pollen are shed in some events. The Glu1-Dx5 transgene has been useful for demonstrating the feasibility of using a pollen-specific gametocide to control pollen dispersal, but it does not function well enough to be used commercially. Other transgenes that function as pollen-specific gametocides exist,⁶ and may be better suited to this purpose.