SCOPE 44 - Introduction of Genetically Modified Organisms into the Environment

15.1 INTRODUCTION

In the past decade scientific and public debate over the use of gene splicing ultimately led to the development of guidelines and methodology to control and assess potential biohazards. Over the past several years, the attention has focused on assessment of the deliberate release of genetically modified organisms into the environment with small-scale experimental field testing viewed as a necessary part of the process in order to provide useful baseline data. Among the issues relevant to environmental assessment of the impact of the deliberate release of genetically modified microbes, plants, and animals is the question of the organism's survival, growth and multiplication, safety, and possible dissemination of its genetic information. These issues are currently being addressed by the scientific community, the various regulatory agencies, and environmental groups as deliberate release experiments are undertaken or planned for the future.

The genetically engineered microbes seem to generate the most concern because they are more difficult to monitor and the chance of horizontal transfer of genetic information is greater. Nonetheless, the vast amount of scientific information that has been generated over the years has demonstrated their potential for applications to benefit mankind. Past experience with a number of microbial pest control agents suggests that in many cases the introduction of genetically manipulated insecticidal microbes probably will be without consequence to the environment.

Numerous genetic-engineering studies of the entomocidal bacteria are already under way. Both Bacillus thuringiensis and B. sphaericus are well recognized as having the greatest potential for use in insect population control and they have received considerable attention by commercial interests. Both bacteria are characterized by the presence of a parasporal crystal composed of polypeptides. The insecticidal crystal proteins (ICP) of B. thuringiensil are toxic to many lepidoptera (caterpillars), diptera (mosquitoes and blackflies), and coleoptera (beetles) larvae, depending upon the strain, while the toxins of B. sphaericus are limited in range to mosquito larvae (Aronson et al., 1986; Andrews et al., 1987).

The genetics of B. thuringiensis has been advanced considerably over the past 10 years such that a number of important factors relevant to its use in genetic-engineering studies are now well documented in the literature. These advances may be summarized as follows: (1) the ICP genes are carried on large transmissible plasmids, (2) the ability to produce insecticidal crystals can be transferred to other bacilli by transconjugation, (3) the ICP genes have been cloned from isolates toxic to lepidopteran, dipteran, and coleopteran insects and expressed in other bacteria, (4) nucleotide and deduced amino acid sequences and protoxin subdomains are known and have been compared between subspecies of lepidopteran-, dipteran-, and coleopteran-killing subspecies of B. thuringiensis, and (5) possible regulatory mechanisms of ICP gene expression have been described (Andrews et al., 1987; Dean, 1984; Carlton and Gonzalez, 1985; Hofte et al., 1987; Sekar et al., 1987; Herrnstadt et al., 1987; Whiteley et al., 1984, 1985, 1987; Whiteley and Schnept, 1986; Aronson et al., 1986; Ellar et al., 1985). The ICP genes of B. sphaericus have also been cloned and sequenced (Ganesan et al., 1983; Hindley and Berry, 1987; Baumann et al., 1987).

The highly selective action of these two entomocidal bacteria offers an environmentally safe method of insect control. Bacillus thuringiensis is one of several microbial control agents that has enjoyed registration with the Environmental Protection Agency (EPA). This agency is now issuing guidance to applicants on developing the required data to reflect advances in risk assessment techniques for genetically engineered microbial pesticides. The EPA is responsible for determining whether the pesticide will cause, or significantly increase the risk of, unreasonable adverse effects to nontarget organisms or the environment.

All of the information gained from studies on the molecular biology and genetics of B. thuringiensis and B. sphaericus over the past 10-15 years has led to attempts, and undoubtedly will lead to future attempts, at providing more effective insect pest control agents based on B. thuringiensis and B. sphaericus ICPs for release into the field. Either natural gene transfer among the strains of B. thuringiensis to yield broader spectrum isolates (Battisti et al., 1985; Gonzalez et al., 1982; Klier et al., 1983) or genetic engineering of the ICP gene itself are available techniques to achieve such a goal (Andrews et al., 1987). Reports have begun to appear of transferring truncated or full-length sequences of ICP genes into crops such as tobacco, tomatoes, cabbage, cotton (Adang et al., 1987; Andrews et al., 1987; Barton et al., 1987; Fischkoff et al., 1987; Anonymous, 1988a) and into other bacteria associated with crop plants (e.g. Pseudomonas fluorescens and Clavibacter xyli subsp. cynodontis; see Crop Genetics International, 1987; de Marsac et al., 1987; Obukowicz et al., 1986a, 1986b; Obukowicz, 1987; Watrud et al., 1985) to give unique delivery systems that reportedly will be available longer for insect ingestion.

The reports are indicative of future strategies in directing the modification of the activity of the ICPs. Such modifications should result in the production of new forms of ICPs, a broadened host range, creation of biocides possessing activities against two or more different target insects, increased persistence, ability to recycle, increased plant tolerance and resistance to insect pests, elimination of processing, manufacturing, transportation and farmer application of toxin in some instances, and perhaps increased toxin yield. Present and future strategies include the use of transforming/transducing/mating systems, recombinant-DNA manipulation/mutagenesis, molecular cloning of toxin genes (e.g. transfer of two or more genes for distinct insecticide activity in the same host cell or transfer of the ICP gene into other prokaryotes). Transfer of the toxin gene into insect viral genomes and/or transfer of the toxin gene into the insect's food source (e.g. algae, tobacco, potato, soybean, corn, cabbage, cotton, tomato, etc.) are additional strategies. The improvement of bio-insecticides to control insect pests should further decrease the dependence on chemical insecticides.

The application of intra- and interspecies transconjugation should be useful in attempts at broadening the host-range specificity by introducing several crystal gene-containing plasmids of different origins into the same recipient strain. This achievement could lead to the development of a multipurpose biocide possessing activities against lepidopteran, dipteran, and/or coleopteran insects. Strains of B. thuringiensis having broader spectra have already been obtained using a transconjugation system (Battisti et al., 1985; Gonzalez et al., 1982; Klier et al., 1983). Klier et al. (1983) transferred the toxin gene of B. thuringiensis subsp. thuringiensis (berliner 1715), which had been transferred into B. subtilis, to an acrystalliferous B. thuringiensis mutant of subspecies kurstaki and also to B. thuringiensis subsp. israelensis. In the latter transfer, the resulting transcipient strain produced both types of insecticidal protein toxins and was active against both lepidopteran and dipteran larvae. Gonzalez et al. (1982) have also demonstrated the production of more than one protein toxin in the same cell from two different subspecies of B. thuringiensis (thuringiensis and kurstaki) that are each toxic to specific lepidopteran larvae.

If expression problems can be overcome, and in some cases they have been solved, the genes for the B. thuringiensis protein toxin might be introduced into a variety of atypical systems to improve the total efficiency of this biocide in the field. A few such efforts have recently been publicized (Crop Genetics International, 1987; Fischkoff et al., 1987; Graf, 1985; Kronstad and Whiteley, 1986; Obukowicz et al., 1986a, 1986b; Obukowicz, 1987; Oeda et al., 1987; Prefontaine et al., 1987; Watrud et al., 1985; Anonymous, 1988a).

For example, the gene might be introduced into bacteria commensurable with plants. Seeds could be watered with these modified bacteria, and the seeds and resultant seedlings may be afforded some measure of protection against susceptible insect pests as well as the growing plant. Both the Monsanto Company and Mycogen have cloned the protein gene of B. thuringiensis subsp. kurstaki into Pseudomonas species. The Monsanto Company's biocide, a recombinant P. fluorescens, which in the natural state is presumably found on the roots of corn plants, is intended to colonize corn roots for control of the black cutworm, Agrotis ipsilon. Corn seeds can be coated with the preparation. To ensure that the toxin gene is not transferred to another microbial species, the company's scientists inactivated the transposase enzyme to prevent the movement of the transposon, which is associated with the toxin gene. Mycogen's approach involved creating a `biological package', or microcapsule, for the toxin to protect it from environmental stresses. The company `fixed' or killed the Pseudomonas organism for use in the field. The killed vegetative cells presumably protect the toxin, thus preventing premature deactivation. Mycogen suggests that using killed cells should make the product easier to formulate and handle. More recently, Crop Genetics International (1987) has inserted the ICP gene from B. thuringiensis supsp. kurstaki into the chromosome of the endophytic bacterium, C. xyli subsp. cynodontis, for control of the European corn borer, one of the largest uncontrolled insect pests of corn (Calvin et al., 1988). The endophytic microorganism is capable of colonizing corn internally.

Other bacteria, such as natural pond microflora, might be suitable hosts for insertion of mosquitocidal crystal protein genes and field release. The suitability of the cyanobacteria for molecular cloning purposes has improved over the past 5 years with the identification and construction of useful shuttle-cloning vectors. These cloning systems in which the mosquitocidal protein gene from B. thuringiensis could be introduced into a blue-green algae and maintained on an autonomous replicon should be helpful in developing algae-based mosquito controls having persistance and ability to recycle.

Plant Genetic Systems scientists have succeeded in inserting the mosquitocidal gene of B. thuringiensis subsp. israelensis into a blue-green algae (personal communication). Additionally, de Marsac et al. (1987) have reported transferring the mosquitocidal gene of B. sphaericus into the cyanobacterium, Anacystis nidulans. The fragment was initially cloned and expressed in E. coli and B. subtilis using pHV33 as a shuttle vector, and then cloned into the blue-green algae with a pUC303 shuttle vector. The cyanobacteria grow on the upper layers of aquatic habitats, persist relatively long in the environment, and presumably should be more effective in controlling mosquito larvae. These reports are good indicators of potential development of biocides with amplified persistance in mosquito habitats.

Other recent efforts at developing novel biocides involve incorporating the toxin genes directly into the genetic background of plants, thereby allowing the plant to synthesize its own insecticidal material. The success of such efforts should result in bypassing the need for repeated applications of the agent, and eliminate the processing, manufacturing, and transportation of insecticidal products. Such modifications have been made possible since both efficient DNA vector constructs are now available for use in transporting foreign genes into higher plant cells. In vitro systems for genetic engineering of plants have also been developed for a number of crop plants including corn, cotton, soybean, cabbage, tobacco, tomato, potato, lettuce, carrot, sugar beet, etc.

The vectors for transferring foreign genes such as the ICP gene into higher plant cells or tissues currently are based on Agrobacterium tumefaciens or the cauliflower mosaic virus, although other sources of vector materials being studied include A. rhizogenes, other DNA caulimoviruses, geminiviruses (e.g. bean golden mosaic virus), potato leafroll virus, and liposomes (Andrews et al., 1987). The Rohm and Haas Company (plant genetic systems), Agrigenetics Advanced Science Company, Agricetus, and the Monsanto Company have reported that the B. thuringiensis toxin gene of subsp. kurstaki has been inserted into tobacco where it is expressed (Freedman, 1985; Adang, 1987; Andrews et al., 1987). Both leaves and the tobacco plant tissues synthesize the insecticidal toxin. Barton et al. (1987) have also reported successful transfer and expression of the B. thuringiensis ICP gene in tobacco where it provides the plants with resistance to lepidopteran pests. Tobacco plants carrying the toxin gene should be resistant to tobacco budworm, cabbage looper, and the tobacco hornworm.

Other important agricultural insects susceptible to the B. thuringiensis subsp. kurstaki toxin are the corn earworm, cotton bollworm, and the beet armyworm. Agracetus has reportedly developed cotton plants that harbor the ICP gene and are resistant to the major lepidopteran pests of this important crop. Under laboratory conditions, the transgenic cotton plants deter feeding and cause death of the pest larvae that do feed. Subsequent greenhouse testing confirmed the resistance and also demonstrated that the ICP gene was passed on to the seed and subsequent generations (Anonymous, 1988a). A few other publicized reports indicate that work is underway on transferring the toxin gene to cabbage and potato.

A significant step in developing insect-tolerant transgenic plants other than tobacco and cotton was recently reported by Fischkoff et al. (1987). Two truncated genes from B. thuringiensis subsp. kurstaki were incorporated into a plant expression vector for Agrobacterium-mediated transformation. The genes were transferred into tomato plants where they were expressed. Expression of the ICP genes conferred insect tolerance on the transgenic tomato plants and their progeny. Larvae (Manduca sexta, Heliothis zea, and H. virescens) were killed within 48 hours with very little evidence of feeding damage to the leaf. Some of the transgenic plants were able to kill 100% of the M. sexta and H. virescens larvae. The investigators also obtained preliminary evidence that some toxin activity was detectable in tomato fruit.

The future development and field use of such novel microbial insecticide products by genetic manipulation will require the availability of efficient and appropriate cloning/transfer/promoter systems that incorporate the technical advantage of high-level expression of a variety of B. thuringiensis ICP genes. Transfer of important genes to select procaryotes and to higher plants appears to be a facet of intense research that is now beginning to yield significant results. Application to the regulatory agencies for environmental release permits have concomitantly increased.

15.3 DELIBERATE RELEASE AND ENVIRONMENTAL ASSESSMENT OF A BIO-INSECTICIDAL BACTERIUM-A CASE STUDY

The Environmental Protection Agency (EPA), the United States Department of Agriculture (USDA), and the Food and Drug Administration (FDA) have established protocols and procedures for implementing their policies for prerelease reviews of proposals involving genetically engineered organisms. These agencies have identified general data requirements and possess the authority to request additional data when necessary to evaluate individual proposals. The USDA and EPA may combine the wide-ranging expertise of the various federal agencies in their reviews and also coordinate with state regulatory officials. The Federal Insecticide, Fungicide and Rodenticide Act (FIFRA) prohibits the distribution, sale, and use of pesticides that have not been registered with EPA, including microbial pesticides. The EPA must review all data submitted on each microbial pesticide before determining whether it should be registered. The Agency has issued guidelines to applicants on developing the required data and continues to revise the guidelines to reflect advances in risk-assessment techniques for genetically engineered microorganisms.

In order to gather product performance and other data necessary for the application, producers may obtain an Experimental Use Permit (EUP) to conduct field studies. The EPA and USDA have indicated that as they gain knowledge and experience from case-by-case reviews, they expect to develop broadly applicable procedures and guidelines. The agencies contend that, for the most part, existing laws available for the regulation of products developed by traditional techniques should be adequate for regulating genetically engineered organisms. Also, all federal agencies are required under the National Environmental Policy Act (NEPA) of 1969 to prepare an analysis before taking a major action that may significantly affect the environment. The agencies first perform a preliminary assessment of the possible consequences of an action to determine whether to prepare an Environmental Impact Statement (EIS) or a Finding Of No Significant Impact (FONSI). If the Environmental Assessment (EA) indicates a significant environmental impact, the Agency must prepare a detailed EIS. The EPA strategy is to regulate genetically engineered microbial products using existing authority with some additional rulemaking for its chemical control statute (Anonymous, 1988b). Unlike the USDA and the EPA, the FDA at this writing does not plan to promulgate new rules for regulation.

In December of 1987 Crop Genetics International (CGI) of Hanover, Maryland, transmitted an application to the EPA and to the USDA for approval to field test a recombinant endophytic bacterium, Clavibacter xyli subspecies cynodontis (Cxc), that is capable of colonizing the vascular system of corn. The organism was engineered to harbor the ICP gene of B. thuringiensis subsp. kurstaki HD73 (Btk) and express the ICP active against the larval stages of the European corn borer (ECB), Ostrinia nubilalis (Hubner). Field testing was planned for May to October of 1988, at the CGI Research Farm, Ingleside, Maryland, and at the USDA Agriculture Research Center, East Farm, Beltsville, Maryland. Additionally, CGI planned to conduct a field test of the Cxc/Btk recombinant bacterium with the Institut National de la Recherche Agroaomique (INRA) at Montfavet, France, in 1988.

This field release has been selected as the case study because CGI not only chose to claim limited confidentiality but also chose to leave essentially all of their submittal open to the public. Additionally, field releases of genetically engineered bacteria seem to generate the most concern and represent rather unique monitoring issues. This particular field test was designed to ascertain specific information on the fate of a genetically designed bio-insecticidal bacterium in the environment.

The main objectives of the field study was to ascertain the effectiveness of Cxc/Btk (InCide) as a corn associated insecticidal agent for control of the ECB and to obtain further knowledge on the behavior of the engineered microbe in the environment. The EUP application included data-addressing environmental persistence and propagation, genetic stability, effects on nontarget plants, and toxicology (human health effects). The Hazard Evaluation Division (HED) of the EPA presented their final scientific position on CGI's EUP application for the field test on 21 May 1988 (a preliminary scientific position was completed on 25 March 1988, forwarded to an EPA workgroup, other Federal agencies-NIH, NSF, FDA, USDA and made available for public comment).

The Animal and Plant Health Inspection Service (APHIS) reviewed CGI's permit application for a permit for release to the environment of a regulated article, and on 20 May 1988 a document giving notice that the USDA intended to issue a permit for release into the environment of a regulated article under regulations issued pursuant to the Federal Plant Pest Act and the Plant Quarantine Act was signed. The document also contained an EA and FONSI. After reviewing the application, APHIS concluded that the proposed field test of the genetically engineered microorganism did not present a risk of introduction or dissemination of a new plant pest and would not have a significant impact on the quality of the human environment. Several supplemental permit conditions, as well as the standard permit conditions, were also included. The EA and FONSI were prepared in accordance with NEPA, regulations of the Council on Environmental Quality for implementing the procedural provisions of NEPA, USDA regulations for implementing NEPA, and APHIS Guidelines for implementing NEPA.

Prior to the approvals of EPA and APHIS the Institution Biosafety Committee (IBC) at the Beltsville Agriculture Research Center reviewed the proposed field studies and agreed that the field study was of low risk and was not likely to exert an adverse effect on the environment; nor was the genetically engineered organism likely to be lethal to humans or to nontarget organisms, provided that the study was conducted in accordance with the containment, monitoring, and contingency procedures and modifications detailed in the documents submitted to the IBC. The IBC gave its approval on 12 May 1988 contingent on the issuance of an EUP from the EPA under regulations issued pursuant to FIFRA, the issuance of a permit for introduction of a genetically engineered organism from APHIS under regulations issued pursuant to the Federal Plant Pest Act (FPPA) and the Plant Quarantine Act (PQA), and fulfillment of any requirements of NEPA. 

The findings and approvals for the field tests were based on CGI's scientific data which was submitted in six volumes of documentation and on in-depth discussions, formal reviews, and miscellaneous other documentation requested by the reviewing organizations. Materials in the volumes provided background information on the wild-type endophytic microorganism and compared the biological characteristics of the endophyte and the recombinant-DNA organism, descriptions of the techniques required to achieve transformation, a description of the ICP gene and the integration vectors used to insert the desired genetic material, including antibiotic resistance genes, a description of the resulting recombinant constructions containing the ICP gene, extensive information addressing the toxicology of the genetically engineered organism, and a description of the proposed field tests and the experiments to be conducted with the recombinant-DNA organism.

Initially, CGI requested that APHIS concur with a contention that C. xyli subsp. cynodontis was not a regulated article; however, APHIS, after a search of the scientific literature, concluded that the organism was capable of inciting `disease or damage' to bermudagrass, and therefore was a <sup>,</sup>regulated article' under the provisions for issuing a permit for the introduction of an article regulated by them. Hence, this is the reason why CGI had to receive a permit from both agencies before conducting the field test. The `Coordinated Framework for Regulation of Biotechnology,' published in the Federal Register on 26 June 1986, established a structure for Federal biotechnology regulation and provided for coordination between agencies when jurisdiction is shared. The `Coordinated Framework' provided that, when an application is for release into the environment of a microorganism intended to act as a pesticide subject to the provisions of FIFRA, EPA is to serve as the lead agency. In the review process APHIS and EPA fully coordinated their separate reviews under their respective authorities.

This particular case study represents one example of what is involved in the preparation for, and evaluation of, deliberate release of a genetically engineered bacterium. Most of the following discussion of this example is excerpted from the EA document and the materials submitted by CGI to EPA, APHIS, and the IBC (BARC Biosafety Committee, 1988). The major source of the discussion, however, is exerpted from the APHIS document (USDA, APHIS, Washington, DC).

The finding by APHIS that there would be no significant impact on the environment from the field test was based on a number of factors documented by CGI in their submitted materials and summarized by APHIS as follows:

1. The inserted ICP gene is lost from Cxc at the rate of once in every 1.25 x 10⁴ bacterial cells per generation, and both the revertant (which has lost the ICP gene) and the naturally occurring Cxc grow faster (14 percent growth rate differential in laboratory medium) than the recombinant bacterium (in fact, it was shown that the ICP gene will eventually be lost from the recombinant bacterium).

2. The genetic alterations are not expected to enhance any plant pathogenic property of the recombinant organism compared to the parental strain of Cxc that occurs naturally in Maryland where the test was to take place.

3. Although Cxc can under limited circumstances be transferred by mechanical means (e.g. cutting tools) to other plants, it is not transferred easily by other mechanisms in the field. Transfer to other plants by mechanical transfer is minimized by the field test design and field test protocol which include buffer zones and tool disinfestation. In addition, regular monitoring for the recombinant bacterium will ensure that if it spreads to plants at the edge of the test plot it will be detected.

4. Dissemination of Cxc can occur in seed, so all seed not used for research purposes (in containment) will be destroyed, preventing transfer by this mechanism.

5. Data were provided by CGI to demonstrate that the probability of transfer of the ICP gene from the recombinant bacterium to other microorganisms is extremely remote.

6. The recombinant bacterium has a relatively low order of toxicity to susceptible insects and since the field test plots are very small, the introduction of the recombinant bacterium would pose no significant impact on insect populations.

7. No threatened or endangered insect species are present in Maryland, so the introduction of the recombinant bacterium would pose no threat to these insects.

8. The inherent properties of Cxc and the recombinant bacterium indicated that there would be no human health risks since the organisms do not grow at human body temperature and they have been demonstrated not to be pathogenic or toxic in mammalian tests. In addition, all crops would be used for research purposes or destroyed so there would be no dietary exposure to humans.

Additionally, the recombinant bacterium was tested in the greenhouse to obtain initial data relating to the biology of interaction of the bacterium with plants and insects and to obtain preliminary data on the interaction of the bacterium with plants and insects and to obtain preliminary data on the interaction of the recombinant bacterium with plants and its efficacy in controlling insects. APHIS noted that it was normal for controlled field tests to be performed after greenhouse testing to confirm the efficacy data and to collect additional data on the complex interactions of organism and that such tests can only be validated in the environment using standard agricultural practices; such limited field testing is essential for the development of a potential agricultural product.

In its evaluation APHIS examined the proposed field plot design, field test protocol, and other factors necessary to identify the aspects of the environment that would potentially be affected. The agency also examined in detail the biology of the recipient/donor organisms involved in constructing the recombinant bacterium and attempted to identify the potential impacts to the environment inherent in those components and describe the ways in which the risk to the environment is limited, either by the nature of the organism or by specific safeguards that had been designed into the protocol.

For their studies CGI developed a number of methods to detect and monitor Cxc and Cxc/Btk organisms in soil, water, plants, and plant debris. Semi-quantitative population estimates of the organisms in extracted or expressed plant sap were performed by phase-contrast microscopy with a sensitivity of 1 x 10⁶ cells/ml of plant sap. A more sensitive assay used, quantitative isolation using a plating technique, had a detection limit of 2 x 10² CFU/g of fresh weight of colonized plant tissue. Spontaneous antibiotic resistant mutants of Cxc were developed for monitoring the organism's persistance in the environment. One Cxc strain was resistant to rifampicin and streptomycin, while a second strain was resistant to rifampicin, streptomycin, and spectinomycin. Using selective media, the detection limit for the doubly resistant strain in corn field runoff water was 1.2 x 10² CFU/ml. For the triply resistant strain the detection limits were 4.3 x 10¹ in river water and 3.4 x 10⁴ in field soil. Detection levels for the Cxc/Btk recombinant bacterium were similar to the spontaneous antibiotic resistant strains (2.7 x 10¹ CFU/ml, non-sterile water; 2.6 x 10⁴ CFU/g, non-sterile soil).

A radioimmunoassay was also developed for the detection of Cxc (no cross-reactivity was observed with any other coryneforms, except C. xyli subsp. xyli, or other plant-associated bacteria tested). The detection limit for the assay was 1 x 10⁶ cells/ml of extracted plant sap.

The data submitted by CGI demonstrated that bermudagrass colonized by the naturally occurring strain of Cxc may be stunted under growth-limiting conditions, such as light limitation and high humidity; however, the stunting was reversed when colonized plants were returned to normal growing conditions. In addition, CGI demonstrated that the yield of corn colonized by the naturally occurring strain of Cxc after artificial inoculation may be reduced in some cases by 2-17%, depending on the corn strain and geographic location. APHIS concluded that these observed effects on corn would not be a problem for the field test since Cxc occurs naturally in Maryland, and no reports of it affecting corn yield have surfaced. APHIS also concluded that the field test would not appreciably increase the total number of Cxc or Cxc/Btk organisms in the environment; nor would the engineered strain effectively compete with the parent organism in terms of growth rate.

Since there were relatively few literature references on Cxc, CGI generated extensive data on its environmental traits. These data included studies of the growth and survival of Cxc in soil, water, and plants, surveys of the natural occurrence of Cxc, and dissemination studies designed to determine if, and how, Cxc may be transmitted to other plants. (APHIS noted that environmental rate studies are inherently limited by the sensitivity of the methods used to detect the bacteria and that there are no detection methods that can reliably detect very low levels-approximately 10 bacteria/gram of soilof bacteria in environmental samples.) Under laboratory conditions, Cxc was found to colonize a wide range of plants when those plants were inoculated and then screened by the expressed sap method. Colonization was detected in 83 species of 26 families, and no colonization was detected in 52 species of 26 families. Some families included both host and nonhost species. These plants were inoculated with a relatively high level of Cxc (approximately 10⁶ CFUs/plant) through a puncture wound using a scalpel blade. Although many plants were shown to support growth of Cxc, no adverse effects were observed despite the presence of greater than 10⁶ cells of Cxc/ml of sap in the colonized plant.

In order to evaluate actual dissemination rates in the field, CGI sampled corn plants that were adjacent to corn plants experimentally colonized with Cxc. APHIS noted that the results were inconclusive (3 of 2264 sampled corn plants were colonized), since a relatively insensitive detection method was used, and the background rate of infection was not determined. The three colonized plants were found within a few meters of the inoculated test plants. No Cxc was found in 32 bermudagrass plants or 5 bindweed plants that were grown in close association with the intentionally infected corn plants.

In addition to the field surveys, CGI submitted studies designed to examine how Cxc could be transmitted in the soil. In greenhouse studies corn was planted in soil containing high levels of Cxc; no colonization of the corn was observed even after deliberate root wounding. Nor was Cxc detected in corn grown in field plots that previously harbored Cxc-infected corn plants tilled into the soil. CGI also submitted data demonstrating that Cxc can not be passed through corn seed to its progeny and thus this mode of transmission was not considered an important factor by APHIS in the field test since the seed would be destroyed. The potential for insect transmission of Cxc was tested in three species of insects (the green sharpshooter, Draeculacephala minerva, the red-headed sharpshooter, Carneocephala fulfida, and the armyworm, Pseudaletia unipuncta) with negative results.

The environmental fate studies consisted of laboratory studies, field surveys, and field transmission studies which APHIS noted was limited by the sensitivity of the detection methodology. Although some of the laboratory data on the growth and survival of Cxc in soil, water, and plants were limited by the use of relatively insensitive methods to screen for the presence of the bacteria, APHIS felt that the data generally supported the conclusions that Cxc does not survive well in soil and water, but grows well in plants. 

APHIS felt that it was not possible to conclude with confidence that the field survey data provided by CGI on the lack of plant colonization was valid, since the rate of low-level colonization could not be ascertained. In the studies, either the numbers of plants sampled were lower than the ratio of positive to negative plants in the corn-dissemination experiment or relatively insensitive screening detection methods were used. The corn transmission studies also used less sensitive detection methods. Nevertheless, in many cases, APHIS felt that these methods were useful in providing a general indication for the potential for colonization since CGI showed, using their most sensitive detection methods (10²-10³ CFUs/ml), that when young susceptible plants were inoculated with Cxc, they invariably supported growth of Cxc to levels detectable by the less sensitive screening methods developed by CGI, indicating that the plants function as an enrichment mechanism for infective doses of Cxc in the environment. By these methods, CGI was also able to detect colonization in the laboratory with inoculated plants used for the host range survey. APHIS noted that these considerations allow for use of these studies as a general indication of the potential for environmental colonization.

 The data indicated that Cxc can infect many plants after direct inoculation (an infective dose for corn was demonstrated at 100 CFUs/plant), but did not appear to naturally inhabit certain susceptible plants surveyed by CGI in the field. Cxc was shown to occur naturally only in bermudagrass, and the incidence of transmission from artificially colonized corn plants seems to be very low.

CGI demonstrated that there was no gross colonization of sugarcane from infected bermudagrass and no gross colonization of bermudagrass or bindweeds in fields of corn colonized by Cxc. APHIS concluded that Cxc is not likely to persist at levels high enough to lead to plant infection in water or soil, except at temperatures too low for plants to grow. This conclusion was supported by studies showing that corn was not infected by growing in soil infested with high levels of Cxc either in the laboratory or in the field. The data indicated that there may be an avenue for limited movement of Cxc to other plants, particularly by mechanical transfer (i.e. by cutting tools) or by transmission through seeds, but APHIS noted that the rate of movement was of a low magnitude and should be controlled with proper procedures. Thus, in the field test all tools used in the plots should be disinfested with a 10% solution of chlorine bleach. The seed produced in the plots should be destroyed or, if not, used in carefully controlled scientific experiments.

CGI submitted data demonstrating that the probability of gene transfer to another organism was extremely remote. No plasmids or prophage were found to be harbored by Cxc. No conjugation occurred in a filter mating experiment with C. flaccumfaciens subsp. cortii, the only related Corynebacterium known to have a conjugative system (Vidaver and Starr, 1981). APHIS noted that CGI truncated the transacting recombinase gene segment in their gene cassette to further reduce the possibility of gene transfer. The ICP gene was integrated into the chromosome of Cxc. In fact, in a 70 generation growth experiment, CGI demonstrated that the engineered genes were lost from the parental Cxc at a segregation loss frequency of 0.6%, which was equivalent to a calculated value of 8 x 10^-5/cell per generation. APHIS pointed out that this rate was almost 10⁴ greater than the natural mutation rate in bacteria, and it also is undirectional, i.e. it will result in irreversible loss of the gene as compared to mutational events which may be reversible, indicating that the ICP gene would eventually be eliminated from the population of bacteria that would develop from the recombinant bacterium introduced to the field.

APHIS additionally requested that CGI provide generation times and any supporting evidence for estimating gene loss time. The information provided showed that the calculated rate of ICP gene loss by the recombinant bacterium in corn plants was 7 x 10^-5/cell. The estimated frequency of revertant/corn plant was 0.25%. CGI also provided information that the parental Cxc was able to grow about 14% faster than the recombinant bacterium in vitro. Revertants likewise grew faster than the recombinant bacterium. Although the rate of ICP gene loss was not precisely known, the information that the presence of the gene decreases the growth rate of Cxc confirmed the conclusions that, in the absence of selective pressure, the recombinant bacterium would eventually be outcompeted by the faster-growing Cxc revertant bacteria that have lost the gene. Additionally, APHIS concluded that there should be no selective advantage to the presence of the gene since CGI demonstrated that Cxc/Btk could not be recovered from the European corn borer or its frass after being fed the recombinant bacterium-infected plants for 2 weeks. In APHIS's view, the target insect would not provide a reservoir for enrichment of the recombinant-bacterial population.

APHIS also addressed issues relevant to the effects on nontarget insect populations, other nontarget animals, nontarget plant communities, exposure of threatened and endangered organisms, and possible impact on the immediate physical environment and on human health from the data submitted. It was noted that CGI's Cxc/Btk strain had a relatively low order of toxicity to the target insect as compared to the parent B. thuringiensis subsp. kurstaki. In addition, the ICP gene would eventually be eliminated from the engineered population. Beneficial invertebrates, such as honey bees, would not likely be affected because the ICP is not toxic to honey bees. APHIS concluded, therefore, that the toxic hazard to nontarget insects would be minimal. APHIS also noted that, since the field plots would be surrounded with a chain-link fence, ingress to the plot by large vertebrate animals would be prevented. No factor unique to the field test was identified by APHIS that would have an effect on birds, wild animals, or vertebrate or invertebrate (except target lepidopteran insects) animals. APHIS further noted that, as described in data provided by CGI, at least 56 species of plants can be infected with Cxc when mechanically inoculated in the greenhouse with large doses of bacteria; however, with plants inoculated by CGI in their greenhouse studies, no signs of disease were detected.

APHIS concluded that dispersal of the recombinant organism from the test plot would be prevented or effectively limited by the field test design and protocols. Trap plants would be used to detect limited movement of numbers of bacteria sufficient to infect plants. Trap plants of both corn and the natural host, bermudagrass, would surround the test plots. The trap plants would be separated from the test plots by a strip of soil that is to be barren, but the plants would be located within the fence of the field plot. These trap plants would serve as a sensitive indicator not only of dispersal by known mechanisms, such as mechanical dispersal by phytophagous insects, and `of the even more remote possibility' of dispersal by windblown pollen.

Since the recipient organism is widely prevalent in Maryland and the recombinant bacterium was not altered to increase its pathogenicity to plants, APHIS concluded that the inserted gene would be eventually diluted or lost, and dispersal of the recombinant organism from the field test plot would be effectively limited with no risk to nontarget plants. Since the field tests would take place in the State of Maryland, while endangered lepidopteran species are in Washington, Oregon, California, and Florida, and since the dissemination from the test site and toxicity to insects would be minimal, APHIS also concluded that use of the recombinant bacterium under the conditions of the field test would not pose any risk to endangered/threatened species.

As to a possible impact on the immediate physical environment, APHIS noted upon examination of the CGI data that when levels of Cxc were inoculated into soil; the populations declined to undetectable levels (10⁴ CFUs/gm soil) after 1-28 days depending on the soil conditions at temperatures from 4 to 50^°C. CGI demonstrated that stable levels of Cxc could be recovered up to 28 days when kept in soil at 5°. Similar results were observed in water except that extended survival was seen at +5 and -5^°C. APHIS noted that although it is not possible to show that a microorganism is completely eliminated from the environment using this type of experiment, the data demonstrated that Cxc did not grow well in soil and water, but could survive at low temperatures. No potential, however, was seen for population levels of Cxc to increase in soil or water.

Because Cxc is a natural bacterial resident of plants and owing to its somewhat fastidious growth requirements, APHIS concluded that it was extremely unlikely that Cxc could infect mammals. The conclusion by APHIS of a lack of human health risk was further supported by the lack of pathogenicity or toxicity in CGI studies of the Cxc/Btk in laboratory mice. In these studies test mice were dosed by a number of exposure routes (oral, pulmonary, and intravenous) with a high titer (about 10⁸ bacteria) of the recombinant bacterium; no toxicity, infectivity, or pathogenicity was observed. Furthermore, in answer to an APHIS request for additional information, CGI presented data that the parental strain (C. xyli subsp. cynodontis MD69a) grows slowly at 34°C and does not grow at 36^°C or normal human body temperature. Additionally, APHIS pointed out there has been no confirmed records of human health hazards associated with exposure to B. thuringiensis var. kurstaki, which has been widely used for more than 20 years to control susceptible insects. Extensive pathogenicity/toxicity testing with B. thuringiensis in laboratory animals and in human volunteers has supported the lack of health risks associated with exposure to this bacterium and to the ICP. 

The two test sites in the United States were of similar plot design with only minor changes in the width of the barren and fallow zones and differences in the number of subplots to accommodate the size of each site. APHIS noted that the plots were designed to allow for efficient maintenance, ease of personnel movement between individual subplots, separation of specific, short-term studies from those studies of longer duration, and the detection of any dispersal of the recombinant bacterium outside the confines of the plot. It was noted that the plots consisted of (1) central corn plant populations arranged in distinct subplots, (2) a barren plant-free zone adjacent to and surrounding the subplots, (3) a `trap' plant zone consisting of corn and bermudagrass bordering the barren zone, (4) a black-plastic-covered earthern dike 2 feet wide and 18 inches high, constructed around the test site, just inside the security fence, (5) a standard 8 foot high, chainlink security fence, and (6) a low-maintenance fallow zone of naturally growing weeds (sampled and analyzed for the recombinant bacterium at intervals throughout the season to test for possible movement of the microbe from the test sites outside the security fences). The corn was irrigated as needed by trickle irrigation and a security guard was on location 24 hours/day. 

Conditions for the test included the requirement that plants and soil should be monitored in the tests by collection of the following data: (1) plant colonization by the recombinant bacterium 4 weeks after inoculation, (2) colonization by the recombinant bacterium of all plant parts monthly for 4 months, (3) natural and mechanical dispersal of the recombinant bacterium in the field after 60 days, (4) presence of the recombinant bacterium in runoff water, (5) presence of populations of the recombinant bacterium in the soil, (6) effect of the recombinant on crop yield; (7) effect of the recombinant bacterium on crop residue decomposition, (8) effect of the recombinant bacterium on vesicular-arbuscular mycorrhizae 3 and 6 weeks after planting, and (9) effects of the recombinant bacterium on populations of saprophytic gram-negative bacteria in the phylloplane. 

APHIS also felt that in the event the presence of corn flea beetles, leafhoppers, and adult southern corn rootworm were detected in association with inoculated test plants, any insects present should be analyzed for the presence of the recombinant bacterium. In all analyses that involved testing for potential dissemination of the recombinant bacterium (trap plants, indigenous plants in the fallow zone, corn residues in the field), CGI should use a detection method at least as sensitive as the plate-count method using plant homogenate (theoretical limit of detection of 2 x 10² CFUs/g). The weed species that grew in the fallow zone areas should be monitored for the presence of the recombinant bacterium to detect any potential spread from the test plot, and when surveying the potential for dispersal into naturally occurring weeds during the field test, all weed species tested should be selected from those weeds known to be susceptible to infection when artificially inoculated with the recombinant organism or parent organism. Also, prior to initiation of the field tests, a survey of indigenous plant species in the fallow zone of each test site should be carried out. All major plant species found should be tested for susceptibility to infection by artificial inoculation. APHIS stated that, during monitoring, the majority of samples from the fallow zone should be taken from plants known to be susceptible to infection.

 A number of other containment protocols were discussed by the reviewing organizations, in addition to the suggested containment provisions already mentioned, and included disinfection of field equipment (tractors, rototillers, sprayers, cultivators, hand tools, etc.). Machinery used in the field test should be disinfested at the edge of the test plot just inside the barren zone, and all organic debris should be thoroughly cleaned from surfaces to be disinfested (cutting blades, tires, etc.). All surfaces should be sprayed with a 10% solution of household bleach (final concentration 0.525% sodium hypochlorite) for a period of 10 minutes in order to make sure they were continuously wet. In addition, a surfactant should be used to ensure uniform wetting of the surfaces and all disinfested surfaces should be allowed to dry before removal from the field. Small tools and equipment used for inoculation and sampling should be decontaminated on site by continuous submersion in 10% solution of household bleach for a minimum of 10 minutes. 

For the transfer of the recombinant bacterium from CGI's laboratories in Hanover, Maryland, or the field laboratories, to the site of release the reviewing organizations suggested movement be carried out in secured, sealed containers. Additionally, investigators and visitors should sign in and sign out at the field test site. The sign-in/sign-out sheets should be maintained as part of the field study records. Footware should be disinfested before exiting the barren zone and all personnel involved in the field test experiments would be instructed in the environmental safety aspects and record-keeping requirements of the test.

APHIS also suggested that bermudagrass and corn trap plants be planted in the soil rather than in containers, as CGI originally specified in the application. The plants would thus be exposed to the soil and runoff water and would provide a more realistic environmental exposure. The bermudagrass should be tested and certified free of infection prior to planting. Each trap plant should be sampled monthly along with all plant parts of inoculated corn plants.

Inoculated corn plants should be surveyed weekly for the presence of flea beetles, leafhoppers, and the adult southern corn rootworm, and any collected insects should be analyzed for the presence of the recombinant bacterium. If the recombinant bacterium was detected in the insects surveyed from the infected corn plants or if the recombinant bacterium was detected in the corn trap plants, the test site should be treated with an insecticide that was effective against the insect found harboring the recombinant bacterium. APHIS had to be notified of the findings as soon as possible, but not later than 5 working days. Also, if the recombinant bacterium was detected in the trap plants (bermudagrass or corn), APHIS had to be informed as soon as possible, but not later than 5 working days. CGI, within 30 days, would take samples of the indigenous weeds in 200 one-foot square plots located in the fallow zone outside of the fence. This sampling would be conducted at least every 2 weeks following detection of the recombinant in trap plants and would be conducted as a replacement for, or in addition to, the monthly fallow zone sampling. If the recombinant bacterium was found in a sample from the fallow zone, APHIS had to be informed within 24 hours and the experiment had to be terminated.

If termination of the field test was required, a broad-spectrum herbicide should be used to kill all plant material in the test site and the fallow zone, all plant material should be collected and incinerated, the test site and the fallow zone should be fumigated with methyl bromide, and a layer of soil from elsewhere on the test farm should be spread over the test site and the fallow zone, and tilled in to provide an inoculum of naturally occurring organisms subsequent to the methyl bromide fumigation.

As specified by APHIS, a normal termination of the experiment could be accomplished by following the same procedure. Alternatively, CGI could monitor the test site in the following year (1989) by planting the site with bermudagrass and sampling the grass for the presence of the recombinant bacterium. If the recombinant bacterium was detected, then the supplemental monitoring could be terminated and the previously prescribed termination procedures should be followed. If no recombinant bacterium was detected, then fumigation and soil incorporation would not be necessary.

15.4 CONCLUSIONS

In a number of instances the potential to manipulate the genes of entomocidal bacteria has become a reality. Within the past ten years much has been learned about the genetics and molecular biology of these organisms and advances in gene manipulation and gene transfer technology have made possible the introduction and expression of foreign genes, including the B. thuringiensis and B. sphaericus entomocidal protein toxin, in transgenic organisms.

The location of the insecticidal crystal protein genes on transmissible plasmids permits construction of novel insecticidal combinations by transconjugation between diverse isolates. Through molecular cloning, gene transfer, recombinant-DNA manipulation, and site-directed mutagenesis more sophisticated biocides will undoubtedly be developed, and progress has already been made using some of these strategies. The nucleotide and deduced amino sequences and protoxin subdomains are now known for a wide variety of the B. thuringiensis subspecies that affect at least three orders of insects. Their comparisons have given some insight into the differences between their activities. The use of gene splicing methods and reconstruction of the various toxin domains should allow for greater insight and perhaps the creation of novel biocides based on B. thuringiensis (and B. sphaericus) that may exceed the present-day naturally occurring isolates in efficacy.

Products based on the ICP gene and developed from this technology have been appearing in increasing numbers. Controlled field releases are needed so that information on the efficacy and biology of the interaction of genetically engineered insecticidal bacteria with the environment can be obtained. In the case study summarized in this report, greenhouse tests were completed to obtain peliminary data on the interaction of the recombinant bacterium (Cxc/Btk) with the environment and the efficacy of the insecticidal organism to control insects. In an environmental assessment for the issuance of a permit by APHIS for Crop Genetics International (CGI) to conduct field tests with engineered Clavibacter xyli subsp. cynodontis harboring the insecticidal crystal protein gene of B. thuringiensis subsp. kurstaki, the APHIS stance was that it is normal for controlled field tests to be performed after greenhouse testing to confirm the efficacy data and to collect additional data on the complex interactions of organisms. The APHIS position indicated that tests can only be validated in the environment using standard agricultural practices and that such limited field testing is essential to develop a potential agricultural product. The stance was based on results of their review of the proposed field test and a subsequent finding that there was no risk of dissemination of a plant pest from the conduct of the test as described by CGI. The EPA also reviewed CGI's proposal (submitted as an application for an experimental use permit), as well as the local IBC; both of these reviewing groups were in concurrence with the APHIS findings.

Overall, the regulatory agencies have tended to emphasize risk reduction in their reviews of cases involving the release of genetically engineered organisms in the environment with limited attention being given to the potential benefits, although the benefit of field testing to the development of information needed for subsequent regulatory decision making on the materials is recognized. Conditions for field testing thus far seem to stress the importance of designing field-release tests to control and monitor the dissemination of genetically engineered organisms from the site of release. The regulatory agencies can apply special conditions when needed to limit the risks associated with potential movement of the test organism in a field-release experiment, and they generally require plans for mitigating any unexpected damage that might occur. They also have the authority to limit or terminate an experiment when deemed necessary. The uncertainties, of course, still remain about field releases of genetically engineered bacteria, but the experience with bacteria engineered to harbor insecticidal properties indicates that we can probably expect that planned introductions of these microbes can be carried out safely.

REFERENCES

Adang, M.J., Firoozabady, E., Klein, J., DeBoer, D., Sekar, V., Kemp, J.D., Murray, E., Rocheleau, T.A., Rashka, K., Staffeld, G., Stock, C., Sutton, D. and Merlo, D.J., (1987) Expression of a Bacillus thuringiensis insecticidal crystal protein gene in tobacco plants. In: Arntzen, C.J. and Ryan, C. (Eds.) Molecular Strategies for Crop Protection, Alan R. Liss, New York, pp. 345-53.

Anonymous (1988a) Ag bio update. In Wysiewicz, T.G. (Ed.) Bio Engineering News, Vol. 9, Port Angeles, Washington, pp. 1-4.

Anonymous (1988b) Biotechnology: managing the risks of field testing genetically-engineered organisms. In: GAO/RCED-88-27 Biotechnology: Risk Management, US GAO, Washington, DC, p. 108.

Aronson, A.I., Beckman, W. and Dunn, P. (1986) Bacillus thuringiensis and related insect pathogens. Microbiol. Rev. 50, 1-24.

BARC Biosafety Committee (1988) Field testing behavior and fate of Clavibacter xyli subsp. cynodontis. In: Permanent Records of the BARC Biosafety Committee Project RD 076, Beltsville Agriculture Research Center, Beltsville, Maryland. 

Biosafety Committee Project RD 076, Beltsville Agriculture Research Center, Beltsville, Maryland.

Barton, K.A., Whiteley, H.R. and Yang, N.S. (1987) Bacillus thuringiensis delta-endotoxin expressed in transgenic Nicotiana tabacum provides resistance to lepidopteran insects. Plant Physiol., 885, 1103-9.

Battisti, L., Green, B.D. and Thorne, C.B. (1985) Mating system for transfer of plasmids among Bacillus anthracis, Bacillus cereus and Bacillus thuringiensis. J. Bacteriol. 162, 543-50.

Baumann, P., Baumann, L., Bowditch, R.D. and Broadwell, A.H. (1987) Cloning of the gene for the larvicidal toxin of Bacillus sphaericus 2362: Evidence for a family of related sequences. J. Bacteriol. 169, 4061-7.

Calvin, D.D., Welch, S.M. and Poston, F.L. (1988) Evaluation of a management model for second-generation European corn borer (Lepidoptera:Pyralidae) for use in Kansas. J. Econ. Entomol. 81, 335-43.

Carlton, B.C. and Gonzalez, J.M. Jr (1985) The genetics and molecular biology of Bacillus thuringiensis. In: Dubrau, D. (Ed.) The Molecular Biology of the Bacilli, Vol. 2, Academic Press, New York, pp. 211-49.

Crop Genetics International (1987) EPA application for an experimental use permit to ship and use a pesticide for experimental purposes only (Transmittal document). Crop Genetics International Corporation, Hanover, Maryland.

Dean, D.H. (1984) Biochemical genetics of the bacterial insect-control agent Bacillus thuringiensis: basic principles and prospects for genetic engineering. Biotechnol. Gen. Engng Rev. 2, 341-63.

de Marsac, N.T., de la Torre, F. and Szulmajster, J. (1987) Expression of the larvicidal gene of Bacillus sphaericus 1593M in the cyanobacterium Anacystis nidulans R2. Mol. Gen. Genet. 209, 396-8.

Ellar, P.J., Thomas, W.E., Knowles, B.H., Ward, S., Todd, J., Drobniewski, F., Lewis, J., Sawyer, T., Last, D. and Nichols, C. (1985) Biochemistry, genetics, and mode of action of Bacillus thuringiensis delta-endotoxins. In: Hoch, J.A. and Setlow, P., (Eds.) Molecular Biology of Microbial Differentiation, American Society of Microbiology, Washington, DC, pp. 230-40.

Fischkoff, D.A., Bowdish, K.S., Perlak, F.J., Marrone, P.G., McCormick, S.M., Niedermeyer, J.G., Dean, D.A., Kusano-Kretzmer, K., Mayer, E.J., Rochester, E.E., Rogers, S.G. and Fraley, R.T. (1987) Insect tolerant transgenic tomato plants. Bio/Technology 5, 807-13.

Freedman, A.M. (1985) Rohm and Haas puts bug-battling gene into tobacco plants. Wall Street J., 24 January 1985, p. 18.

Ganesan, S., Kamdar, H., Jayaraman, K. and Szulmajster, J. (1983) Cloning and expression in Escherichia coli of a DNA fragment from Bacillus sphaericus coding for biocidal activity against mosquito larvae. Mol. Gen. Genet. 189, 181-3. 

Gonzalez, J.M., Brown, B.J. and Carlton, B.C. (1982) Transfer of Bacillus thuringiensis plasmids coding for delta-endotoxin among strains of B. thuringiensis and B. cereus, Proc. Natl Acad. Sci. USA. 79, 6951-5.

Graf, J.S. (1985) Mycogen gets EPA go ahead to test recombinant pesticide. Gen. Engng News 5, 1.

Herrnstadt, C., Gilroy, T.E., Sobieski, D.A., Bennett, B.D. and Gaertner, F.H. (1987) Nucleotide sequence and deduced amino acid sequence of a coleopteran-active delta-endotoxin gene from Bacillus thuringiensis subsp. san diego. Gene 57, 37-46.

Hindley, J. and Berry, C. (1987) Identification, cloning and sequence analysis of the Bacillus sphaericus 1593 41.9 kD larvicidal toxin gene. Mol. Microbiol. 1, 187-94.

Höfte, H., Seurinck, J., Van Houtnen, A. and Vaeck, M. (1987) Nucleotide sequence of a gene coding an insecticidal protein of Bacillus thuringiensis var. tenebrionis toxic against Coleoptera. Nucleic Acids Res. 15, 7183.

Klier, A., Bourgouin, C. and Rapoport, G. (1983) Mating between Bacillus subtilis and Bacillus thuringiensis and transfer of cloned crystal genes. Mol. Gen. Genet. 191, 257-62.

Kronstad, J.W. and Whiteley, H.R. (1986) Three classes of homologous Bacillus thuringiensis crystal protein genes. Gene 43, 29-40.

Obukowicz, M.G., Perlak, F.J., Kusano-Kretzmer, K., Mayer, E.J., Bolten, S.L. and Watrud, L.S. (1986a) Tn5-mediated integration of the delta-endotoxin gene from Bacillus thuringiensis into the chromosome of root-colonizing pseudomonads. J. Bacteriol. 168, 982-9.

Obukowicz, M.G., Perlak, F.J., Kusano-Kretzmer, K., Mayer, E.J. and Watrud, L.S. (1986b) Integration of the delta-endotoxin gene of Bacillus thuringiensis into the chromosome of root colonizing strains of pseudomonads using Tn5. Gene 45, 327-31.

Obukowicz, M.G., Perlak, F.J., Bolton, S.L., Kusano-Kretzmer, K., Mayer, E.J. and Watrud, L.S. (1987) IS50L as a non-self transposable vector used to integrate the Bacillus thuringiensis delta-endotoxin gene into the chromosome of root colonizing pseudonomads. Gene 51, 291-6.

Oeda, K., Oshie, K., Shimizu, M., Nakamura, K., Yamomoto, H., Nakayama, I. and Ohkawa, H. (1987) Nucleotide sequence of the insecticidal protein gene of Bacillus thuringiensis strain aizawai IPL7 and its high-level expression in Escherchia coli. Gene 53, 113-19.

Prefontaine, G., Fast, P., Lau, P.C.K., Hefford, M.A., Hanna, Z. and Brousseau, R. (1987) Use of oligonucleotide probes to study the relatedness of delta-endotoxin genes among Bacillus thuringiensis subspecies and strains. Appl. Environ. Microbiol. 53, 2808-14.

Sekar, V., Thompson, D.V., Maroney, M.J., Bookland, R.G. and Adang, M.J. (1987) Molecular cloning and characterization of the insecticidal crystal protein gene of Bacillus thuringiensis var. tenebrionis. Proc. Natl Acad. Sci. USA 84, 703040.

Vidaver, A.K. and Starr, M.P. (1981) Phytopathogenic coryneform and related bacteria. In: Star, M.P., Stolp, H., Truper, H., Balows, A. and Schlegel, H.G. (Eds.) The Prokaryotes: A Handbook on Habitats, Isolation, and Identification of Bacteria, Vol. 11, Springer-Verlag, Berlin, pp. 1879-87.

Watrud, L.S., Perlak, F.J., Tran, M.-T., Kusano, K., Mayer, E.J., Miller-Wideman, M.A., Obukowicz, M.G., Nelson, D.R., Kreitinger, J.P. and Kaufman, R.J. (1985) Cloning of the Bacillus thuringiensis subsp. kurstaki delta-endotoxin gene into Pseudomonas fluorescens: molecular biology and ecology of an engineered microbial pesticide. In: Halvorson, H.O., Pramer, D. and Rogul, M. (Eds.) Engineered Organisms in the Environment, American Society for Microbiology, Washington, DC, pp. 40-6.

Whiteley, H.R. and Schnepf, H.E. (1986) The molecular biology of parasporal crystal body formation in Bacillus thuringiensis. Ann. Rev. Microbiol. 40, 549-76. 

Whiteley, H.R., Schnepf, H.E., Kronstad, J.W. and Wong, H.C. (1984) Structural and regulatory analysis of a cloned Bacillus thuringiensis crystal protein gene. In: Ganesan, S. and Hoch, J.A. (Eds.) Genetics and Biotechnology of the Bacilli, Academic Press, New York, pp. 375-86.

Whiteley, H.R., Kronstad, J.W. and Schnepf, H.E. (1985) Structure and expression of cloned Bacillus thuringiensis toxin genes. In: Hoch, J.A. and Setlow, P. (Eds.) Molecular Biology of Microbial Differentiation, American Society for Microbiology, Washington, DC, pp. 225-9.

Whiteley, H.R., Schnepf, H.E., Tomczak, K. and Lara, J.C. (1987) Structure and regulation of the crystal protein gene of Bacillus thuringiensis. In: Maramorosch, K. (Ed.) Biotechnology in Invertebrate Pathology and Cell Culture, Academic Press, New York, pp. 13-27.