The differences between conventional Bacillus thuringiensis strains and transgenic insect resistant plants. Possible reasons for rapid resistance development and susceptibility of non-target organisms

by Beatrix Tappeser
Institute for Applied Ecology, Freiburg, Germany Prepared for the third meeting of the Open-ended Working Group on Biosafety, Okt 13-17 1998, Montreal, Canada


B. thuringiensis (Bt) was discovered by Ernst Berliner in 1911 when a consignment of flour moths sent in from Thuringia was found to be infected by some contagious pathogen. After some more properties of this Bacillus species had been identified, including its host specifity, it did not take long until first experiments were being carried out on its effectiveness in controlling the corn borer (1928-1931). However, it was not until after the Second World War, when first problems with synthetic pesticides turned up, that serious attempts were made to establish B. thuringiensis as a biological pesticide (Krieg, 1986).

In 1995 the market volume of Bt preparations was at an estimated 90 million US dollars, and 67 preparations were registered worldwide (Kumar et al., 1997). Bt preparations account for 80 - 90% of all biological pesticides. By contrast, their share in the whole insecticide market is no more than 1 - 2%, though there has been an upward tendency since the early nineties. Forecasts made in 1991 predicted that by the year 2000 Bt preparations would account for 5 - 10% of the world insecticide market (Bernhard and Utz, 1993).

The main target pests of Bt insecticides include various lepidopterous (butterfly), dipterous (flies and mosquitoes), and individual coleopterous (beatle) species. Some strains have also been found to kill off nematodes (Edward et al., 1988; Krieg and Franz, 1989). Conventional Bt preparations such as those registered in Germany but also worldwide are mostly derived from the highly potent strain Bacillus thurigiensis var. kurstaki HDI, which was isolated in the sixties (Dulmage, 1970, cited in Kumar et al., 1997).

Characterisation of Bt strains contained in conventional Bt preparations

Most Bt preparations available on the market, all of which are subject to individual licensing, contain spores with parasporal inclusion bodies composed of  d-endotoxins. To date 50 different d-endotoxin genes have been isolated which fall into sixteen subgroups (Crickmore et al., 1996). These genes can occur in different strains and in diverse combinations. Almost all Bt strains are able to form more than one type of crystalline inclusion body, and these in turn can be made up of several different d-endotoxin molecule species.

Bt strains are classified according to their membrane proteins and the endotoxins or toxin genes they contain. Bt d-endotoxins in turn are classified by the sequence homology of their genes and insect specifity.

Every Bt strain can have a variable number of plasmids responsible for the synthesis of different endotoxins. The HDI strain harbours at least five different plasmids. Plasmids can bear several, usually identical, toxin genes. Moreover, Bacillus thuringiensis has been found to bear different transposons which thus contribute to the great genetic variability of its toxin genes and consequently of the toxins themselves. This explains the great diversity of Bt strains. Bt strains can easily exchange their plasmids via a conjugation-like process, as has been demonstrated in the larval gut. In this way Bt strains can also exchange plasmids containing d-endotoxin genes and so express different activity patterns in different lepidopterous species. The spores and crystalline toxin molecules are inactivated quickly when exposed to UV-light.

Structure of toxin genes and proteins

Certain structures common to Bt toxin genes suggest a kind of "built-in" variability which gives Bt great flexibility in its action in different target insects. Just as antibodies contain conserved as well as variable domains, so are toxin genes and proteins made up of alternating conserved and variable regions. The N-terminal part of the toxin protein is responsible for its toxicity and specifity and contains five conserved regions. The C-terminal part is usually highly conserved and probably responsible for crystal formation.

Infection cycle

The first step of the infection cycle following the ingestion of Bt parasporal crystal bodies is the solubilisation of the crystalline proteins in the alkaline environment of the insect gut to proteins of 130 - 140 kDa size. The efficiency of crystal protein solubilisation depends on the enviromnent prevailing in the insect gut and the composition of the parasporal inclusion bodies.

In a second step the solubilised protein undergoes proteolytic cleavage, this giving rise to the actual toxin, a protein fragment of 60 - 65 kDa. The proteolytic process can consist of up to seven specific steps (Choma et al., 1990). The final toxic fragment of the most frequently encountered protein, Cry I A (c), is thought to span amino acids 29-608, counting from the Nterminus (Hofte and Whiteley, 1989). This activated protein diffuses through the peritrophic membrane, which is impermeable to the solubilised, still unprocessed 130 - 140 kDa protoxins, and so acts as a kind of molecular sieve. Then the toxin protein binds to specific receptors located in the insect gut. According to present knowledge this leads to the formation of pores and consequent destruction of ion gradients. These pores also permit the vegetative Bt cells germinating from the spores to migrate into the haemolymph and promote the intoxication process through the ensuing bacteraemia (Marrone and MacIntosh, 1993).

For a long time it was thought that the insecticidal action of Bacillus thuringiensis chiefly resides in the spores and inclusion bodies. It was not until 1995 that kurstakolin, an additional growth promoting factor which increases the toxicity of Bt preparations by 30%, was discovered in the supenatant of Bt cultures (Asano and Hori, 1995). This suggests that the insecticidal action of Bt must consist of highly complex interactions between the bacterium and its individual host insect species.

Host-pathogen interrelationships / development of resistance

On the pathogen side we have a great diversity of toxin genes which can occur in varying combinations within a strain and are capable of exchange between strains via conjugation-like processes. By virtue of their variable and conserved regions and occasional flanking transposon sequences Bt toxin genes are predestined for multiple transposition and recombination, suggesting that the great variability of Bt is also attributable to the individual structure of any given toxin gene. The insecticidal action of Bt is enhanced by a concurrent induction of bacteraemia following the binding of toxin proteins to and resultant formation of pores in the intestinal wall, as well as by the recently discovered growth factors.

Two prerequisites for infection on the host side are that the intestinal environment must permit efficient solubilisation of the crystalline inclusion bodies and that the host's own proteases cleave the solubilised protoxins such that the resulting active toxins are of the right size. This is essential if the proteins are to diffuse through the peritrophic membrane and reach their specific receptors in the intestinal wall. It is thus evident that insects have developed both specific and unspecific defences acting at different levels to protect them against the insecticidal action of Bt upon ingestion of parasporal inclusion bodies. Receptor specifity assumes a key role in this defence. The host's permissiveness of crystal solubilisation and its complement of proteases are not without influence, but their role in the evolution of resistance appears to be of minor importance. However, these factors do have a strong impact on endotoxin activity in susceptible hosts and possibly in non-target-hosts. The absence of the necessity to solubilize the endotoxin molecules and to cleave down the protein to a smaller size in insects feeding on transgenic plants may have an influence on the susceptibility of non-target organisms. Unfortunately there exists no systematic research to evaluate the differences between the unchanged Bt preparations and their transgenic counterparts.

To date there have only been few reports of initially susceptible insect species evolving resistance after intensive use of Bt preparations in the field. Some of the studies used purified toxins and insofar mimic the situation for transgenic plants but not for the conventional Bt preparations (cf. also Marrone and MacIntosh, 1993). So far only one crop pest, the diamondback moth has evolved resistance to Bt in open field populations (Tabashnik, 1994). More frequently it has happened in laboratory experiments. Recent evidence augmented the concerns that a rapid resistance evolution will take place. Gould et al. (1997) found out that the frequency of alleles for resistance is 1,5 X 10-3 in Heliothis virescens, a major cotton pest. This is at the upper level of any estimates used to calculate the time scale for resistance evolution. Heliothis virescens is very susceptible to Bt-preparations. Therefore the authors conclude that it will take about ten years for resistance becoming a serious problem in these polulations. This estimate takes into account that a 4% refuge plan is followed as it is proposed by the Environmental Protection Agency of the US.

However other pest species like cotton bollworm (Helicoverpazea) and the European corn borer (Ostrinia nubilalis) are less susceptible. The authors calculate that populations of these pests could become resistant in 3-4 years even with a 4% refuge given the same frequency of resistance alleles. This is a short time to render inefficient the most important biological insecticide, even too short to get back all the investments in developing the transgenic Bt-resistant plants. Cloning of two or more Bt toxins is supposed to extend the time period for resistance development because it is assumed that independent mutations are needed to counter each toxin. Tabashnik and co-workers (1997) showed that cross-resistance for four different Bt-toxins can be traced to one autosomal recessive gene which confers extremely high resistance to these toxins in one strain of the diamondback moth.

Some scientists think that the low incidence of evolved resistance in the field has to do with the presence of multiple toxin genes in the individual Bt strains. Studies by Dai and Gill showed that mosquitoes which were treated with Bt-var israeliensis for more than 20 generations and then exposed either to a sporal endotoxin preparation or a purified toxin preparation showed only a moderate degree of evolved resistance in the first case (threefold) relative to the 70-fold rate found in the second.

The fact that resistance development in the open field with conventional Bt-preparations is very rare up to now sheds light on the significance of the differences between transgenic insect resistant plants and the commercially available Bt-preparations.

It is surprising that the discussion on the possibility of resistance evolving in initially susceptible insects as a result of the use of transgenic plants has given so little attention to these aspects.

Insect resistant transgenic plants

As a rule insect resistant transgenic plants are engineered by means of single isolated toxin genes. The fact that the genetic code of plants differs slightly from that of bacteria makes it necessary to use synthetic genes whose nucleotide sequence is altered in such a way that it still encodes the desired bacterial amino acid sequence. However, it is not possible to use complete toxin genes in plants because these are not sufficiently soluble in plant cells (the protoxins are only soluble at pH greater 9.5, whereas the pH in plant cells is around 7.6). This problem is circumvented by using truncated genes which produce almost fully activated toxin molecules and reside in the plant cell in solubilised form.

The insect resistant maize engineered by Ciba (now Novartis) contains two copies of a truncated synthetic Cry I A (b) gene. This gene comprises the first 648 amino acids of a protoxin that is normally made up of 1155 amino acids. Because of its adaptation to the genetic code used by plants the nucleotide sequence of the synthetic gene is only 65% homologous to that of the native gene (referred to the truncated version). The synthetic gene is fused, firstly, to the phosphoenolpyruvate carboxylase promoter of maize, which is responsible for generating the protein in all green tissues of maize; and secondly, to a pollenspecific promoter of maize which enables the protein to be generated in pollen grains. The 648-residue protein probably has to undergo two or three more proteolytic steps in an alkaline environment before it becomes the 564-578-residue protein which Ciba (now Novartis) declares to be the fully active toxin. According to other sources, however, the active toxin in its most truncated version comprises amino acids 29-607 (Lereclus et al., 1993; Perlak et al., 1991). This gives compelling evidence that the active protein comprises 578 residues and that 28 residues are removed N-terminally and 31 C-terminally.

It follows that the insect-resistant plant contains a truncated version of a d-endotoxin gene which generates a shortened precursor protein that is synthesised throughout the vegetation cycle and resides in the plant in solubilised form. In view of its small size it is probable that the solubilised toxin molecule can diffuse directly through the peritrophic membrane without further cleavage steps.

There is no genetic background for varying toxin genes, there are no toxicity enhancers present, there are no possibilities for a bacteraemia to support the lethal work of the pathogen.

Concluding remarks

That means that the host-pathogen interaction has undergone important changes, changes that have the potential to promote resistance development, possibly by some orders of magnitude. Additionally, there are hints that the transgenic plants unfold their toxic potential to other insects than the pests in an unexpected way. Feeding of intoxicated larvae of the European corn borer to beneficial insects killed two out of three of these insects. Even feeding to nonsusceptible larvae of an other species killed those predators which fed on these larvae. These experiments were done by Swiss researchers employed at the national research station for agriculture and agricultural ecology. The study is submitted for publication to Environmental Entomology but not yet published (Facts 34/1997).

Effects to nontarget beneficial organisms were also shown for collembola and are mentioned in the pesticide fact sheet produced by EPA(1)and made available by the National Technical Information Service of the US. However, these effects were judged as not being disquieting.
(1) Published 1995

It is urgent to have a moratorium(2) for transgenic insect resistant plant in order to save one of the most valuable biological pesticides. This moratorium is also necessary to prevent genetic pollution via out-crossing. The changed toxin may have the potential to kill others than nontarget organisms which possibly will have far-reaching consequences in different environments. There is now more than evidence that insect-resistant transgenic plants have negative impacts on both sustainable agriculture and the environment. A science-based risk assessment should take into account its own data and not ignore them.  
(2) 31 groups of scientists, farmers and environmentalists filed legal action against USA EPA over its approvals, amongst them Greenpeace, the International Federation for Organic Agriculture, the Center for International Technology Assessment and the Institute for Agriculture and Trade Policy in Minneapolis


Asano S and Hori H (1995) Appl. Entomol. Zool. 30 (2), 369-374, zitiert nach Kumar PA, Sharma RP, Malik VS (1997) The Insecticidal Proteins of Bacillus thuringiensis. Advances in Applied Microbiology 42, 1-43

Bernhard K and Utz R (1993) Production of Bacillus thuringiensis Insecticides for Experimental and Commercial Uses. In: Entwistle PF, Cory JS, Bailey, MJ, Higgs S (Eds.) (1993) Bacillus thuringiensis, An Environmental Biopesticide: Theory and Practice. John Wiley & Sons, Chichester UK

Crickmore N, Zeigler DR, Feitelson J, Schnepf E, Lambert B, Lereclus D, Baum J, Dean DH (1996) Genes Microbiol. Rev., zitiert nach Kumar PA, Sharma RP, Malik VS (1997) The Insecticidal Proteins of Bacillus thuringiensis. Advances in Applied Microbiology 42, 1-43

Edwards DL, Payne J, Soares GG (1988) Novel isolates of Bacillus thuringiensis having activity against nematodes. European Patent Application, EP 0 303 426 A2

Environmental Protection Agency (1995) Pesticide fact sheet: 'Bacillus thuringiensis' CrylA(b) deltaendotoxin and the genetic material necessary for its production (plasmid vector pCIB4431) in corn. Washington, DC

Gould F, Anderson A, Jones A, Sumerford D, Heckel DG, Lopez J, Micinski S, Leonard R, Laster M (1997) Initial frequency of alleles for resistance to Bacillus thuringiensis toxins in field populations of Heliothis virescens. Proceedings of the National Academy of Sciences of the United States of America 94, 3519 - 3523

Hofte H, Whiteley HR (1989) Insecticidal Crystal Proteins of Bacillus thuringiensis. Microbiological Reviews 53, 242-255

Krieg A und Franz JM (1989) Lehrbuch der biologischen Schadlingsbekampfung. Verlag Paul Parey, Berlin und Hamburg

Krieg A, Huger AM (1986) Symposium in memoriam Dr. Ernst Berliner, Darmstadt 25.8.1986. Mitteilungen aus der Biologischen Bundesanstalt fur Land- und Forstwirtschaft, Berlin-Dahlem, Heft 233

Kumar PA, Sharma RP, Malik VS (1997) The Insecticidal Proteins of Bacillus thuringiensis. Advances in Applied Microbiology 42, 1-43

Lereclus D, Delecluse A, Lecadet MM (1993) Diversity of Bacillus thuringiensis_Toxins and Genes. In: Entwistle et al. (1993), a.a.O., S. 255 - 267

Marrone PG and MacIntosh SC (1993) Resistance to Bacillus thuringiensis and Resistance Management. In: Entwistle et al (1993), a.a.O., S. 221 - 235

Perlak FJ, Fuchs RL, Dean DA, McPherson SL, Fischhoff DA (1991) Modification of the coding sequence enhances plant expression of insect control protein genes. Proc. Natl. Acad. Sci. USA 88, 3324-3328

Tabashnik BE (1994) Evolution of resistance to Bacillus thuringiensis. Annual Review of Entomology 39, 47 -79

Tabashnik BE, Liu Y-B, Finson N, Masson L, Heckel DG (1997) One gene in diamondback moth confers resistance to four Bacillus thuringiensis toxins. Proceedings of the National Academy of Sciences of the USA 94, 1640 -1644

Visser B, Bosch D, Honee, G (1993) Domain-Function Studies of Bacillus thuringiensis Crystal Proteins: A Genetic Approach. In: Entwistle et al. (1993), a.a.O., S. 71 - 88