Selectable markers are commonly used in plant genetic transformation; however, marker genes usually are not needed once transgenic plants have been identified. The presence of marker genes, especially antibiotic markers and herbicide resistance markers in transgenic crop plants, may elicit environmental and consumer concerns. In addition, the existence of marker genes in transgenic crops could evoke additional, lengthy risk assessments for release of crops that contain useful novel traits.

Markers that promote plant regeneration have been recently described. However, continuous expression of these markers may interfere with normal plant growth and development. Removal of this type of marker from plant tissues is necessary unless expression is under good control. Furthermore, current transformation technologies permit only the introduction of a very limited number of genes into plants. Retransformation of the same line is needed for multiple trait modifications. New selectable markers are thus needed with each transformation to pyramid the same crop variety with different desirable traits. The number of selectable marker genes that are suitable for each crop species is usually very limited. This is especially true for transformation of recalcitrant species. Marker excision can allow reuse of a marker after each transformation step. Marker elimination will not only appease some potential environmental and consumer concerns, it will also remove technical barriers for plant genetic transformation^1,2.

Co-transformation and technology development
Transformation selection uses a marker gene to identify the transformants that contain the gene of interest or to eliminate non-transformed tissues. Conventionally, the gene of interest and the marker gene are tightly linked in the transformation vectors and are transferred and integrated into the plant genome together to ensure selection. Nevertheless, the marker gene can permit selection even if it is distant from the gene of interest.

Co-transformation was developed for marker removal. The gene of interest and the marker gene are transferred separately via different approaches, and the two genes are inserted at loci that are not linked. The marker is segregated later in the next generation. Co-transformation can be conducted by different approaches, such as the use of separate transformation vectors in the same Agrobacterium, or the use of different Agrobacterium strains. Co-transformation using mixed Agrobacterium strains or vectors is technically simple and is effective for transformation and marker removal. Nevertheless, this approach may not be suitable for species in which transformation efficiency is low.

Various technologies have been developed for co-transformation improvement. For example, two copies of T-DNA can be placed on the same transformation vector. In this case, the transgene and the marker gene are located independently on each of the T-DNAs, which are separated by intervening DNA sequences. The two T-DNAs can also be located adjacent to each other using two extra T-DNA border regions, creating two almost contiguous T-DNAs.

Instead of using multiple T-DNAs, a single T-DNA with a unique design can also be used for co-transformation. Lu and co-workers³ developed a system called double right-border binary vector. The vector contained only one copy of T-DNA but carried two copies of the T-DNA right border sequences, which flanked a selectable marker gene. The gene of interest was inserted next, and then one copy of the left border sequence was placed at the end. The vector enabled two separate insertions: one developed from the first right border and contained the selectable gene along with the target gene, and the other came from the second right border carrying only the target gene.

Another system, described by Huang et al.⁴, broke the traditional T-DNA based principle of gene transfer. Contrary to conventional vector design, they repositioned the selectable marker in the backbone of a regular binary vector, leaving only the gene of interest within the T-DNA region. They reasoned that the marker gene would be transferred into the plant cell either along with the T-DNA that was initiated at the first border but failed to terminate at the second border, or as part of the T-strand that was initiated at the second border. The attempt was based on previous observations that vector backbone sequences frequently transfer along with T-DNA into plant cells. They tested this design in maize and found that co-transformation efficiency was higher than when the vector was comprised of multiple T-DNAs containing the gene of interest and the marker separately. Marker-free transgenic plants were successfully recovered in subsequent generations.

Co-transformation is technically simple and is an effective approach for segregating the transgene and the marker gene, and for subsequent marker removal. Nevertheless, co-transformation efficiency should be reasonably high and the gene locations should be distantly separated. Such demands may not always be possible. A major limitation with co-transformation is that marker excision occurs only in the next generation, and thus the time course is relatively long. Also, the methodology is not suitable for woody tree plants and for vegetatively propagated plants.

Site specific excision and the system control
The Cre-lox site-specific recombination system of bacteriophage P1 can be used as a strategy for marker gene removal. The system consists of two components: the Cre recombinase and two 34 bp loxP recognition sequences. The expression of the Cre protein causes recombination between the two loxP sites and results in excision of the DNA sequence flanked by the sites⁵. Cre recombinase is introduced into the transgenic plants carrying the selectable marker gene either by a second round of transformation or by out-crossing. In addition to Cre/lox, several other site-specific DNA excision systems derived from other organisms may be adapted for marker removal in transgenic plants. These include the R/RS system from Zygosaccharomyces rouxii, Flp/frt from Saccharomyces cerevisiae, and Gin/gix from bacteriophage Mu.

Site-specific excision is an effective approach for marker gene removal. However, the system has some limitations. For instance, re-transformation is laborious, expensive, and time consuming. Out-crossing is not suitable for vegetatively propagated plants and woody tree species. In addition, the continuous and high level of expression of the cre recombinase gene may result in phenotypic aberrations in some plant species.

Some new techniques that aim to control recombinant systems have been developed. The cre gene expression system can be controlled by specific gene promoters, such as inducible gene promoters, including chemical inducible promoters and heat shock protein systems. The marker gene is allowed to process for selection. Once selection is complete and upon induction of the promoter, the cre gene expresses, which then activates the lox sequences and the marker gene is excised. This new technology is effective for marker gene removal in different crop species and can also be applied to generate marker-free woody tree species via vegetative propagation.

Expression control of recombinases no doubt has significantly extended the site-specific excision system and made the system more useful and practical. In addition to using a specific promoter to control the cre gene, transient expression of cre recombinase in some approaches may be sufficient to activate the excision and lead to marker gene removal. However, the system still needs improvement for wider use.

Transposon elements and marker gene removal
Transposition systems such as Ac/Ds have been used as a strategy to eliminate marker genes. The technique is based on the fact that the DNA sequence located within Ds repeats can be mobilized to excise along with the Ds mobile element. The gene of interest or the marker sequence can be placed within the mobile sequence. Upon activation by Ac transposase, the transgene or marker gene residing inside of the transposable Ac element can be excised and subsequently reinserted elsewhere in the plant genome. Thus, the transgene and the marker gene may be separated by transposon elements. The use of transposition to relocate the transgene to a new chromosomal locus is a useful approach for generating marker-free plants. However, genetic crossing or segregation is required to separate the gene of interest and the marker gene, and thus it is time consuming and entails a long process.

In rare cases due to transposition errors, the excised DNA may fail to reinsert into the host genome and be lost. Based on this observation, Ebinuma et al.⁶ described a vector in which the marker gene was constructed within the Ac/Ds transposable system. Marker-free transgenic tobacco and woody aspen plants were obtained via the reinsertion failure. Although the system was effective and marker elimination occurred in the same generation, marker removal efficiency via this strategy is poor, due to the low incidence of occurrence.

Marker gene removal from chloroplasts
Chloroplast DNA, in general, is maternally inherited. However, there exists some concern that marker genes may transfer to relevant weeds in the environment. Marker genes may also move to microorganisms, and the marker gene may function in bacteria, as the expression system is directed by prokaryotic regulatory elements. Therefore removal of marker genes after transgenic plants are recovered is still highly desirable in transplastomic plants.

The well characterized site-specific Cre/lox system has also been used for plastid marker gene removal. In a study by Corneille et al. in 2001⁷, the negative selectable marker codA gene, which was used to identify marker removal events, was flanked by two lox sites placed in plastids. Nuclear-encoded but plastid targeted Cre recombinase was introduced via either Agrobacterium-mediated retransformation or by pollination. Efficient elimination of the marker gene from plastids occurred via either approach. Plastid marker removal using the Cre/lox system was also demonstrated in other studies. It has been noticed that marker removal in plastids occurs rather rapidly, which is in contrast to the slow process of marker establishment in plastids during transformation.

Another approach for marker removal from plastids is via homologous DNA recombination. Iamtham and Day⁸ described a system in which several marker genes, including addA, bar, and gus, were conventionally flanked by different recombination sites. These marker genes were sequentially eliminated later via loop-out recombination. This system, however, is difficult to regulate because selection of homoplastomic transformants appeared to be unpredictable. Klaus et al.⁹ described a system in which the marker gene is placed outside of the flank regions used for homologous recombination. Recombination via either left or right flanks resulted in co-integration of the vector, but a further round of recombination resulted in the loss of the marker gene. Thus, the marker gene was only transiently co-integrated into the genome. The marker gene expression for transplastomic selection was still provided effectively via this manner⁹. The advantage of this method is that marker-free plants can be generated directly in the first generation.

Summary

Marker removal is an important and integral aspect of plant transgenic research today, and many different approaches and technologies have been demonstrated. Routine marker gene removal from transgenic plants will eventually become more feasible in a wider array of plant species as research in the area continues to proceed rapidly.

REDUCING GOSSYPOL IN COTTONSEED MAY IMPROVE HUMAN NUTRITION
Keerti S. Rathore

As world population continues to rise, there is a growing requirement for food and fiber, especially in developing countries. In addition, agriculture is increasingly tapped to supplement world energy needs. Since the area of arable land is limited, meeting these needs must involve improving crop yields sustainably and utilizing global agricultural output more efficiently.

The terpenoid class of secondary metabolites produced by many plant species has important ecological functions either as attractants (e.g., linalool) or as defense compounds^1,2 (e.g., bitter triterpenoid cucurbitacins, pungent diterpenoid polygodial, gossypol, and related compounds in cotton). However, some terpenoids and other defense compounds are toxic to humans and animals. Over the course of human history, man has learned to either avoid consumption of toxic plants/parts, or to inactivate/neutralize toxic compounds before ingestion (e.g., cassava, kidney beans). In some cases, the inedible plant product is used as feed for those domestic, ruminant animals that suffer little or no ill effects from consuming these products. Cottonseed, produced in abundance as a byproduct of fiber production, represents such a case.

Cotton is a cash crop for more than 20 million farmers in developing countries in Asia and Africa. For every 1 kg of fiber, cotton plants produce approximately 1.65 kg of seeds: e.g., global cotton cultivation in 2005 yielded 23.5 million metric tons (MMT) of lint and over 40 MMT of seed. Cottonseed contains 21% oil and 23% protein of relatively high quality. The global annual production of cottonseed could potentially provide the total protein requirements of nearly 500 million people for a year (50 g/day rate) if the seed were safe for human consumption. However, the utility of cottonseed for food is hampered by the presence of gossypol, located in seed glands. Gossypol is a terpenoid that is cardio- and hepatotoxic to humans and monogastric animals; unfortunately, this toxicity limits cottonseed to use primarily as ruminant cattle feed, either as whole seeds, or as meal following oil extraction. Even though cottonseed does find some indirect use in human nutrition in the form of cattle feed, ruminant animals are highly inefficient in terms of feed conversion. Thus, the potential of cottonseed to help meet food requirements of the burgeoning world population provides a great impetus for eliminating gossypol from cottonseed.

Classical breeding and its limitations
Sub-epidermal, lysigenous glands containing one or more terpenoids are present in all parts of the cotton plant except stele tissue, young roots, xylem, and seed coat³. Gossypol is the predominant terpenoid in glands of seed-kernel and petal, while glands in other organs contain varying levels of gossypol and related terpenoids. Discovery of a glandless trait in a mutant cotton variety (Hopi Moencopi) provided plant breeders with a tool to eliminate glands and therefore gossypol from seed⁴. Several national and international programs were launched to transfer this useful trait into commercial varieties to produce gossypol-free cottonseed. These programs yielded cottonseed that could be fed to more efficient feed-utilizing, monogastric animals and was even deemed safe for human consumption.

In nutrition studies, glandless cottonseed compares favorably as a source of protein with other traditional food sources⁴. However, these glandless cotton varieties were a commercial failure. Under field conditions, glandless plants were more prone to insect damage, because they constitutively lacked gossypol and other protective terpenoids⁵; consequently, these cotton plants were not accepted by farmers. A few wild, diploid Gossypium species exhibit a glandless-seed and glanded-plant phenotype. Attempts to introgress this highly desirable trait into commercially popular tetraploid cotton have had limited success. Consequently, the potential of cottonseed to contribute to human nutrition has not yet been realized.

Biotechnology to the rescue
As mentioned, glands containing various terpenoids in aerial plant parts protect against herbivory by insects. Many terpenoids are also induced in response to fungal or bacterial challenge and are believed to function as phytoalexins^2,3. These compounds are derived from the primary sesquiterpene skeleton, (+)-δ-cadinene. (+)-δ-cadinene synthase (d-CS), a cyclase enzyme, catalyzes conversion of farnesyl diphosphate to (+)-d-cadinene, the first committed step in biosynthesis of gossypol and related terpenoids⁶. The ability to block this committed step solely in seeds could produce cottonseeds with reduced gossypol levels while maintaining wild-type levels of gossypol and other defensive terpenoids in the rest of the plant. Attempts to use the antisense gene suppression mechanism targeting δ-CS to eliminate gossypol from cottonseed were unsuccessful, or resulted in a small reduction in gossypol levels, or provided ambiguous results^6-8. It took a more powerful gene silencing mechanism and a strong, highly seed-specific promoter to achieve a meaningful reduction in seed-gossypol.

Sunilkumar et al.⁸ used RNA interference (RNAi)-mediated suppression of the δ-CS gene under control of a cotton-seed promoter to significantly reduce δ-CS activity in developing embryo, which resulted in a seed-specific reduction of toxic gossypol. The efficacy, strength, and tissue specificity of the a-globulin promoter had been demonstrated in an earlier study⁸. This promoter becomes active at 15 days post-anthesis (dpa) in developing cotton embryos. At this stage, promoter activity is low and localized in the middle portion of the zygotic embryo. There is a slow increase in activity until 20 dpa; however, after this stage, it rises rapidly and spreads throughout the embryo. The promoter remains active until maturity. On the other hand, δ-CS transcripts are first detected at 23 dpa and reach steady-state level from 30 to 50 dpa⁶. Thus, the α-globulin promoter is active for quite some time prior to appearance of δ-CS transcripts in developing embryo and continues to be active until embryo maturity. This promoter was therefore used to direct generation of hairpin transcripts targeting the δ-CS gene in cotton.

An efficient means to induce heritable, RNAi in plants is by stable integration and expression of a transgene construct encoding self-complementary transcripts that can fold back to generate double-stranded RNA molecules ('hairpin' RNA or hpRNA). The efficiency with which a target gene is silenced is significantly improved when the loop portion of these hpRNAs consists of a functional intron⁹ (ihpRNA). To generate ihpRNA transcripts, a 604 bp sequence from a δ-cadinene synthase cDNA clone obtained from Gossypium hirsutum was used as the trigger sequence to assemble the transformation construct. This sequence bore a high degree of homology to several previously published sequences of d-CS genes from diploid and tetraploid cottons and was expected to silence most members of the δ-CS, multigene family.

Following Agrobacterium-mediated transformation, several lines were obtained that produced seeds containing significantly reduced levels of gossypol. This reduced seed-gossypol phenotype was correlated with presence of transgene, as well as with substantial reductions in levels of target transcripts and enzyme activity in the developing transgenic embryo. Most relevantly, levels of gossypol and related terpenoids involved in defense against insects and diseases were not reduced in foliage, floral organs, and roots of transgenic plants, as compared to wild-type controls.

Three different transgenic lines were monitored up to the T2 generation, and the reduced seed-gossypol trait was found to be stable and heritable. One of the lines produced seeds with gossypol levels as low as 0.2 µg/mg, showing approximately a 98% reduction in the level of the toxin (seeds from greenhouse-grown, wild-type, cv. Coker 312 cotton plant contain approx. 10 µg gossypol/mg of kernel). The United States Food and Drug Administration (FDA) and United Nations Food and Agriculture Organization and World Health Organization (FAO/WHO) permit up to 0.45 µg/mg (450 ppm) and 0.6 µg/mg (600 ppm), respectively, of free gossypol in edible cottonseed products⁴. Thus, by using tools of modern molecular biology, gossypol was reduced in cottonseed to a level considered safe for human consumption.

An intriguing property of RNAi is that it can spread from cell-to-cell and long distances from its point of origin. This phenomenon, initially observed in C. elegans, is also believed to occur in plant systems. On the basis of these observations, doubts have been cast on the feasibility of achieving tissue- or cell-specific silencing of desired gene(s) in plants¹⁰. If the RNAi-induced silencing signal were to spread out of developing embryos, or if the components of the silencing mechanism remaining in mature seed were to spread to the plant upon its germination, the seed (kernel)-specificity offered by the a-globulin promoter would be lost, and the plant would suffer from the same weakened resistance that was observed in glandless cotton.

However, results presented in the PNAS report indicate strict confinement of silencing⁸. Single seed analysis showed that null-segregant seeds, which developed side-by-side with RNAi-silenced embryos in the boll of a T0 transgenic line, had wild-type levels of gossypol. Thus, they were not affected by the silenced status of adjacent embryos. In addition, leaves, floral organs, and roots of transgenic plants, grown from silenced seeds, exhibited wild-type levels of terpenoids. This observation suggests that spreading of the RNAi-induced silencing signal does not occur in cotton, and consequently, gossypol reduction remains confined to cottonseed.

The study also provided additional evidence from unrelated research showing that RNAi-mediated silencing targeting the expression of GFP in cotton does not spread and is strictly limited to tissues expressing hairpin transcripts. These results suggest that RNAi-mediated gossypol reduction will be restricted to the developing embryo and the kernel in mature seed. Thus, defensive capabilities offered to remaining plant organs by gossypol and related terpenoids should not be compromised in RNAi lines.

Conclusions
The availability of reduced-gossypol cottonseed offers the possibility that the cotton plant, in addition to providing fiber, can also be a source of nutrition, either directly as food in the form of seed/seed products or as feed for more efficient feed-utilizing, monogastric animals. Of course, RNAi lines will have to be field tested, and pass the regulatory approval process, before the technology becomes available and is accepted by cotton farmers.

Plant breeding will continue to help meet requirements of humanity for food, fiber, and fuel. The results from the RNAi study suggest that tools of modern molecular biology offer an alternative means to address problems that are not always possible to solve using traditional breeding methods. This research also shows that biotechnology, in addition to improving productivity of agricultural crops, can help us better utilize agricultural products by either modifying their nutritional composition or eliminating harmful compounds from edible organs of the plant.

Malaria causes 300-500 million clinical cases and over one million deaths per year worldwide. Forty percent of the world's population lives in areas where malaria is transmitted, and every 30 seconds a child dies of malaria. In humans, malaria is caused by a number of parasite species including Plasmodium falciparum, which spends half of its lifecycle in humans and the other half in mosquito, such as Anopheles gambiae.

Transmission of malarial parasite from mosquito to human occurs through a mosquito bite. When an infected mosquito bites a human, plasmodium, which is present in mosquito saliva, enters the body and migrates to the liver. In liver, the parasite matures and then infects red blood cells, where a second stage of maturation occurs to form gametocytes. When a non-infected mosquito bites an infected human, gametocytes are transmitted to the mosquito gut, where they further develop into ookinetes. Ookinetes then traverse the gut wall by a receptor-mediated process and migrate to and infect salivary glands, thus completing the cycle.

Malaria can be effectively controlled using insecticide-treated bed nets or spraying insecticides (e.g., DDT) to kill mosquitoes or with drugs such as chloroquine to control plasmodium. However, resistance to insecticides has developed in mosquitoes and resistance to drugs has developed in plasmodium. Thus new strategies for effective control of parasite in either the human or mosquito host are urgently needed.

One approach that has been proposed is development of transgenic mosquitoes with reduced capacity to transmit malarial parasite. However, in order for this method to be effective, it is important to understand how transgenes can be introgressed or introduced into mosquito populations in the field. Furthermore, the genetically modified mosquito must be able to successfully compete with the nontransgenic mosquito in the environment and also must not negatively impact the environment or ecosystem.

In 2002, Ito et al. reported development of transgenic Anopheles stephensi mosquitoes that had impaired ability to transmit the murine malarial parasite, Plasmodium berghei. Screening a library of random dodecapeptides revealed a peptide, named SM1, that specifically bound to epithelial cells of the midgut and the salivary gland. SM1 is hypothesized to bind to a receptor, present on midgut epithelial cells, that is used by plasmodium to traverse the midgut. Thus SM1 binding interferes with binding of parasite to midgut wall. Expression of SM1 was placed under control of the carboxypeptidase promoter, which directs secretion of SM1 into midgut in response to ingestion of a blood meal.

Effectiveness of this strategy was tested in mosquito feeding trials. Transgenic mosquitoes were allowed to feed on parasite infected mice. The number of parasites per transgenic mosquito decreased 80% compared to nontransgenic mosquitoes. Furthermore, the ability of infected mosquitoes to transmit plasmodium to mice was greatly diminished.

These transgenic mosquitoes represented a promising step in development of a novel malaria control strategy. However, one important issue that remained to be tested was whether there was a fitness cost of the transgene to the genetically modified mosquito. Moreira et al. (2004) showed that transgenic mosquitoes expressing SM1 showed no significant reduction in fitness parameters, such as mosquito survival, fecundity (eggs laid per female), and fertility (proportion of eggs that hatched into larvae) relative to nontransgenic mosquitoes. Fitness was also examined in cage experiments in which 250 heterozygous transgenic mosquitoes were crossed with 250 nontransgenic mosquitoes. No consistent deviation from expected transgene frequency was observed, indicating that the SM1 transgene did not impose a fitness load.

Interestingly, an effect on fitness of transgenic mosquitoes may be transgene specific. Transgenic mosquitoes expressing the phospholipase A2 (PLA2) gene in midgut, under control of the same carboxypeptidase promoter, also showed an 87% reduction in plasmodium formation and a greatly impaired transmission of parasite to uninfected mice, as seen with SM1 transgenic mosquitoes. However, PLA2 transgenic mosquitoes showed a reduced fitness, i.e., laid fewer eggs. In cage experiments the frequency of transgenic mosquitoes rapidly decreased such that by the fifth generation there was almost a complete loss of PLA2 transgenic mosquitoes.

As a final test, the fitness of SM1 transgenic mosquitoes feeding on infected mice was compared to nontransgenic mosquitoes (Marrelli et al., 2007). Transgenic mosquitoes were more fit, i.e., had higher fecundity and lower mortality, than nontransgenic mosquitoes. In cage experiments, transgenic mosquitoes gradually replaced nontransgenic mosquitoes when they were maintained on mice infected with gametocyte producing parasites, but not on mice infected with gametocyte deficient parasites. By the ninth generation, transgenic mosquito frequency reached about 70%. These results demonstrate that when feeding on plasmodium infected blood, transgenic mosquitoes have a selective advantage over nontransgenic mosquitoes, which is likely due to their reduced parasite load.

These results have significant implications for devising novel malaria control strategies by genetically modifying mosquitoes. Transgenic mosquitoes feeding on infected mice may have a fitness advantage over nontransgenic mosquitoes. This would result in the transgenic mosquito gradually becoming the predominant species and thus lead to reduction in rate of malaria transmission.

In recent years, federal district courts have found environmental statute violations in the way that the Animal and Plant Health Inspection Service (APHIS) regulates genetically engineered (GE) plants. A federal judge recently fashioned an unusual remedy for a violation: he placed a permanent injunction on an APHIS-approved cultivation of a GE crop.

The case began in June 2005 when APHIS issued a Finding of No Significant Impact and approved Monsanto Company's petition requesting nonregulated status for GE Roundup Ready^® alfalfa. Opponents to deregulation stressed the possibility that bee pollination could transfer the GE alfalfa's glyphosate tolerance gene to conventional alfalfa. Nevertheless, APHIS concluded that growers of conventional or organically-grown crops could emplace reasonable quality control measures to ensure that their crops did not include any GE alfalfa.

Alfalfa growers, the Sierra Club, and other farmer and consumer associations filed a lawsuit, alleging that APHIS' deregulation of GE alfalfa violated the National Environmental Policy Act. Cultivation of GE alfalfa would result in spread of the glyphosate tolerance gene to natural alfalfa, they contended, an event that would create a significant environmental impact.

Charles R. Breyer, a judge in the US District Court for the Northern District of California, agreed with the plaintiffs. APHIS had effectively concluded, according to the judge, that any environmental impact would be insignificant, because organic and conventional farmers bore the responsibility to prevent genetic contamination. Despite APHIS' conclusion, Judge Breyer could find no evidence that the agency had investigated if farmers could actually protect their crops from genetic contamination.

On February 13, 2007, the judge held that APHIS had failed to take a "hard look" at the potential environmental impacts of its deregulation decision, a step required by the National Environmental Policy Act. He granted plaintiffs' motion for summary judgment on the claim that APHIS must prepare an Environmental Impact Statement (EIS).

On March 2, the plaintiffs filed a request for a permanent injunction to block APHIS' deregulation of the GE alfalfa until the agency preformed its environmental review. The judge granted plaintiffs' request on May 3. First, the judge vacated APHIS' June 2005 determination of nonregulated status for the GE alfalfa. Then, the judge instructed the agency to prepare an EIS and reconsider the deregulation petition. APHIS must complete its EIS and again decide to deregulate before farmers can plant Roundup Ready alfalfa.

Meanwhile, farmers had planted 220,000 acres of GE alfalfa before the ban. Judge Breyer decided that the alfalfa may be grown, harvested, and sold under certain conditions. For instance, farmers must apply APHIS-approved procedures to clean farm equipment used in GE alfalfa production to minimize the risk of the spread of GE alfalfa seed and hay. Harvested GE alfalfa must be stored in designated and clearly labeled containers. And in the most controversial condition, APHIS must gather information about the locations of GE alfalfa seed production sites and GE alfalfa hay fields and reveal this information to the public. This would enable producers of conventional or organically-grown alfalfa to decide if they should test their crops for contamination.

APHIS has requested that the judge amend the conditions, including the widespread disclosure of specific locations of GE alfalfa fields. Previous disclosures of GE crop locations, the agency noted, triggered vandalism and intimidation of farmers.

USDA spokeswoman Rachel Iadicicco told the Associated Press that the court-imposed environmental study could take up to two years to complete. Monsanto Company announced that the company is reviewing its options, including the possibility of an appeal.

EPO Makes A Meal of Soy Patent, While Court Issues Toxic Verdict
In another unusual May 3 decision, the European Patent Office (EPO) revoked Monsanto's patent EP301749B1 with claims for the genetic modification of soybean plants. The EPO took this action 13 years after the patent's grant.

In July 1988, Agracetus filed the patent application, which describes particle bombardment methods for genetically altering soybean plants. The EPO granted the patent in March 1994 with claims to genetic engineering methods, and soybeans and seeds that contain a genetic alteration. Monsanto acquired Agracetus in 1996 and became the owner of the soybean patent.

For years, opponents fought against the patent, alleging that it gave Monsanto de facto control over all GM soybeans. The patent's adversaries realized one victory in 2003 when the EPO struck a claim to a method of genetically altering any kind of plant with particle bombardment. The agency decided that the patent lacks sufficient disclosure for such a broad claim and limited claims to soybean plants.

Now, the EPO has revoked the soybean claims on the basis that the claims lacked novelty. An EPO spokesman said that the decision is final with no further appeals available. Since the patent would have expired in 2008, elimination of the soybean claims should yield limited practical effects. However, the legal basis for the decision may significantly impact the agbiotech industry. The EPO will publish an explanation of its decision by the end of the year.

Battles over Bacillus thuringiensis technology continue. In July 2002, Syngenta filed a lawsuit claiming that Monsanto and other companies infringed at least one of US Patent Nos. 6,075,185; 6,320,100; and 6,403,865. These patents include claims to synthetic Bt toxin genes designed for increased expression in corn and claims to transgenic corn plants resistant to insects.

In December 2004, Judge Sue L. Robinson of the Delaware District Court held that defendants had not infringed Syngenta's '185 and '100 patents as a matter of law. These patents focus on methods for optimizing codons for more efficient expression of Bt insecticidal proteins in corn. The judge decided that the codon usage of defendants' products does not fall within the scope of the '185 and '100 patent claims. A jury then found the '865 patent invalid on the grounds of obviousness and lack of written description.

Syngenta appealed the jury verdict to the US Court of Appeals for the Federal Circuit. On May 3, the Federal Circuit affirmed.

The relevant claims of the '865 patent cover transgenic corn plants that produce a Bt protein encoded by a recombinant gene that has a G+C content of at least about 60%. The key prior art reference presented to the jury for an obviousness consideration was a published patent application of Kenneth A. Barton and Michael J. Miller, US Patent Application Publication No. 2001/0003849. The document teaches that Bt genes have a high proportion of codons rich in A+T, while plants generally have codons rich in G+C. Barton and Miller describe a method for enhancing Bt toxin expression in GE plants by selecting codons that reflect the G+C bias.

While conceding that the general notion of substituting codons rich in G+C may have been obvious, Syngenta insisted that the idea to modify the coding sequence of the Bt toxin gene to increase the G+C content to more than 60 percent would not have been obvious. In one line of argument, Syngenta asserted that the patent application focused on GE tobacco plants and that the same codon substitution strategy could not reasonably be expected to succeed in corn.

The court pointed out, however, that the application includes a scorched earth statement: "there is good reason to believe and expect that the increased efficiency of expression achieved in tobacco through the use of the method and coding region of the present invention will be equally applicable in other plant species, as it is in tobacco." The Federal Circuit found substantial evidence to support the jury's verdict on obviousness.

No Savior for Seed Saver
For the third time, the Federal Circuit rendered a decision on the patent dispute between farmer Homan McFarling and Monsanto. The litigation's origin dates to 1998 when McFarling purchased Monsanto's Roundup Ready soybean seeds. He also signed the company's Technology Agreement that required him to use the seed for planting a commercial crop in a single season. Under the contract, McFarling could not supply seed to any other person for planting, save any crop produced from the seed for replanting, and supply saved seed to anyone for replanting. Yet McFarling saved 1500 bushels of Roundup Ready soybeans from his 1998 harvest and planted them the following year. He saved over 3000 bags of soybeans from his 1999 harvest for his next crop.

When Monsanto discovered that McFarling had saved the GE seeds, the company sued the farmer in the Eastern District of Missouri, alleging patent infringement and breach of contract. The district court ruled that the farmer could not use seed saved from crops grown with the patented soybeans. McFarling appealed to the Federal Circuit and lost.

Back in district court, Monsanto moved for summary judgment on its claims for patent infringement and breach of the Technology Agreement. The court ruled in favor of Monsanto.

McFarling appealed to the Federal Circuit, claiming that the district court erred when it ruled against his patent misuse defense, his antitrust counterclaim, and his defense under the Plant Variety Protection Act. The Federal Circuit upheld the district court's decision holding McFarling liable for breach of contract and dismissing McFarling's counterclaims and defenses. The Federal Circuit remanded the case for a determination of Monsanto's actual damages.

In a petition for a writ of certiorari to the US Supreme Court, McFarling pressed his patent misuse defense and antitrust counterclaim. The Court denied the petition.

Once again in district court, a jury returned a damages verdict of $40 per bag of saved seed, adding up to about $375,000 owed to Monsanto. McFarling appealed.

At the Federal Circuit, McFarling argued that the amount of the damages award should be limited to $6.50 per bag. This is the Technology Fee that Monsanto charged licensees who purchased Roundup Ready seeds under its Technology Agreement.

The Federal Circuit disagreed. Under the Monsanto license agreement, soybean farmers paid the company a Technology Fee and promised to refrain from planting Roundup Ready seed saved from a previous season's crop. The promise ensured that farmers would purchase Roundup Ready seed from an authorized distributor seed company, which also charged a fee for soybeans. Monsanto effectively split its royalty fee into a $6.50 direct payment and a payment of $19 to $22 to the seed companies that promoted and distributed Monsanto's products.

In addition to these royalty benefits, the court found advantages of the contract, including an increased yield and decreased cost of weed control. Altogether, the benefits justified the jury's assessment about a reasonable royalty due to Monsanto, the court held.

The decision may signal the end of the case. This time, the Federal Circuit did not send an issue back to the district court.