Well worth reading in full - even the summary at the end!

Prof Schubert is responding once again to Bradford et al who have been arguing that the regulation of GM food crops should be reduced or eliminated, based upon the assumption that the products of genetic engineering are no different than those produced by classical plant breeding.

Note also the reference to Prof Guy Cook's must-read book "Genetically Modified Language".

EXCERPTS: I, and hundreds before me, pointed out that [the claim that GM plants are no different from classically bred plants] is unambiguously not the case. I used specific references to show that many of their statements were misrepresentations of scientific fact. In their reply to my comments they used several new rhetorical techniques in addition to the standard ones such as taking statements out of context and misquoting sources. Of greatest concern is the new lexicon that has been evolving in the plant biotechnology industry over the last decade...

Perhaps the most curious aspect of all is that plant biotechnology is complaining about a regulatory system that was written by their lawyers (Eichenwald et al., 2001) and at least with respect to the FDA is voluntary and lacks safety testing requirements altogether (Gurian-Sherman, 2003; Freese & Schubert, 2004). Although they have what they asked for, they are still complaining about it.

David Schubert
Salk Institute
La Jolla, CA 92037

The following are some comments on the rebuttal (Bradford et al., 2005a) to my critique (Schubert, 2005) of a manuscript (Bradford et al., 2005b) that appeared in Nature Biotechnology. In their original article, Bradford and colleagues argue that the regulation of transgenic food crops should be reduced or eliminated, based upon the assumption that the products of genetic engineering (GE) are no different than those produced by classical plant breeding. I, and hundreds before me, pointed out that this is unambiguously not the case. I used specific references to show that many of their statements were misrepresentations of scientific fact. In their reply to my comments they used several new rhetorical techniques in addition to the standard ones such as taking statements out of context and misquoting sources. Of greatest concern is the new lexicon that has been evolving in the plant biotechnology industry over the last decade in order to deceive the less technically educated into believing that there should be no concern about GE food crops because, as they argue, the outcomes are identical to those obtained with standard breeding techniques. Since they cannot ignore the overwhelming evidence that GE is highly mutagenic, they are instead trying to equate GE with normal breeding by redefining the fundamental meaning of some relevant terminology. An excellent book entitled "Genetically Modified Language", written by a linguist, Guy Cook, shows how the plant biotechnology community is misusing language to promote themselves (Cook, 2005). As described in detail below, examples of "genetically modified language" are abundant in the rebuttal by Bradford et al. of my critique.

1. Lack of precision. The initial response of Bradford et al. in defense of the unambiguously high rate of mutagenesis in GE crops is a perfect example of how plant biotechnology is attempting to change the technical definitions of genetics for the purpose of self promotion. They state that "conventional breeding is based on essentially random induction or assembly of mutations", followed by "imprecise natural recombinations between genomes". Thus, they are equating recombination with mutagenesis, and so, by extension, GE with natural breeding. This is not only scientifically incorrect but exceptionally deceptive.

Recombination occurs with high fidelity between allelic genes. There is no mutagenesis involved in the standard recombination event, for if there were, there would be no such thing as a stable species of plant or animal. This section of the critique by Bradford et al. concludes by stating "changes accompanying GE may occur, but are irrelevant so long as the expected phenotype is produced". The problem here is that they redefine phenotype to suit their purposes. In general scientific usage phenotype refers to all traits, while these authors use 'phenotype' in both their original paper and their rebuttal to mean solely agricultural characteristics, ignoring other traits that might be caused by genotypic changes from GE. The tests used to assay unintended changes to phenotype are, to date, quite limited. The legitimate debate is whether these limited tests are adequate. Will an assay to detect changes in yield of peas detect an increase in rotenone or other harmful secondary metabolites?

2. Basic research vs. cultivar development. The discussion in this section is completely meaningless, for in my critique I was concerned about toxicological traits, not agronomic ones, and unless they can establish a causal link between plant height or yield and potentially toxic secondary metabolites, agronomic traits are not relevant to the health and safety issue.

3. Mutagenized cultivars. Since both the original Bradford paper and my critique deal only with US regulatory policies, I specifically stated that I was discussing food crops in the US. The manuscript by Ahloowalia et al. (Ahloowalia et al., 2004) lists all of the registered crops (non-food as well as food) in the world that have a mutagenized parent. The "2,275 varieties of 175 species" referred to by Bradford et al. include flowers and many other non-food crops, and the vast majority are not now and never were used commercially. As I stated in my critique of the food crops, the only one listed by Ahloowalia et al. as a commercial crop in the US is the sunflower. The major cultivars of the US crops of corn, soybeans and wheat are not derived by mutagenesis. The implication that I misrepresented the Ahloowalia article is therefore incorrect. Indeed, it would be of interest to many if Bradford et al. could list and document those vast numbers of crops in the US food supply that they claim are derived by mutagenesis.

4. Wide crosses. I agree that "genetic changes often accompany wide crosses". I don't doubt that genetic changes always occur during any breeding procedure. Indeed, that is the point of sexual reproduction. However, the question is whether or not those changes that do occur are the same as those caused by GE? First, Bradford et al. again try to equate recombination with mutagenesis which, as discussed above, is not correct. Knowing this group's propensity for "genetically modified language" I specifically pointed out the difference in my original critique. Second, their "large body of evidence" supporting the claim that wide crosses are mutagenic is rather paltry, and certainly does not justify all of the claims that they make for genomic modifications outside of changes in copy number and recombination, which are not mutations. For example, the cited paper by Madlung et al., 2005 used to support their claims of naturally occurring transposition in fact only shows that in Arabidopsis polyploids there is "transcriptional activity of several transposons although their transposition was limited" (Madlung et al., 2005). In other words, some transposon-dependent RNA was made, but it did not reverse transcribe and randomly insert into the chromosomal DNA to cause mutations (as occurs with GE manipulations). Both Madlung et al. (2005) and Liu & Wendel (2000) show that changes in DNA methylation at sites within or flanking the normally inactive transposons are responsible for their "limited" or "ephemeral" activation. While both papers show that transposons can transiently be transcribed, neither established that DNA products were made and incorporated into functional DNA, thereby possibly causing a mutation. Furthermore, the "silenced genes" in the cited manuscripts are in fact the transposons, and gene silencing is not a mutagenic event (half of the X chromosome complement in human females is silenced by methylation).

Again, aside from wheat, not a single one of the cited manuscripts showed that wide crosses produced mutations. In wheat allopolyploidization does cause the elimination of blocks of DNA and transient retrotransposition (Levy & Feldman, 2004). However, tetraploid wheat occurred about 500,000 years ago and hexaploid about 9500 years ago. Synthetic allopolyploid wheat has been made in the laboratory, but I am not aware of any commercial crops from this material.

5. Promoters. My comments have nothing to do with promoters, either viral or genomic, per se, but only with the fact that in GE plants they are used in synthetic DNA constructs to drive the expression of foreign genes in all plant tissues, and that this is by no stretch of fact or imagination a situation that occurs in nature.

Plant biologists are very defensive about this aspect of their technology, and as witnessed here they try to talk their way around it by presenting information unrelated to the expression of foreign genes in all tissues.

While I did not express any particular concern about the transfer of antibiotic resistance from GE plants to animals, it must be pointed out that contrary to the views expressed by plant biologists, it has clearly been shown that a transgene from GE soya can survive passage through the small intestine and can transfer its DNA to the microflora of the small intestine (Netherwood et al., 2004).

Although the gene was a fragment of the glyphosate resistance gene from soybeans, there is no reason why other genes could not also transfer. Therefore there is horizontal gene transfer from plant material to gut bacteria and if for some reason there is a selective advantage for those bacteria expressing the gene (for example, during a course of antibiotics), they could become the dominant population within the gut. Since plant DNA also can be taken up by and integrated into the cells lining the intestines and other tissues (Einspanier et al., 2001; Schubbert et al., 1994; Schubbert et al., 1997; Schubbert et al., 1998), the possible health consequences of this transfer cannot be ignored.

While I agree that antibiotic resistance may not be an issue for the common antibiotics like ampicillin and tetracycline, as pointed out by Bennett et al. (2004) "bacterial AR genes that are uncommon in bacterial pathogens, and for which any further spread would be undesirable, if not disastrous, should not be used as marker genes in GE plant development". Curiously, one of the papers cited by Bradford et al. is an attempted justification by a group sponsored by Monsanto to do exactly that - introduce kanamycin resistance into GE plants as a selectable marker (Flavell et al., 1992).

Transposition. Again Bradford et al. redefine scientific terminology to obscure the facts. In their statements, they explicitly equate the expression of mRNAs with the insertion of reverse transcribed DNA into genomic DNA (transposition). I state that there is no transposition, not that some plants (and animals) cannot occasionally transcribe some mRNA from these repetitive elements. It is possible that I have missed published data showing that transposition does occur in non GE food crops, but if this is the case, the appropriate reference should have been cited by Bradford et al. Thus they either do not understand the science or are purposely misrepresenting the data.

Screening. The statement that "humans have adapted to diverse plant chemistries" is curious in that it states exactly the opposite of the true situation and is one of my major concerns about GE. Human physiology did not evolve to fit that of plants and there certainly would be no selection against the ingestion of compounds with long term consequences such as carcinogens.

Instead, for the last 10,000 years, humans have selected and bred plants that did not make them sick and promoted their health. Bradford et al. contend the opposite. Since, as Roessner et al. (2001) clearly demonstrate, new chemicals not found in conventionally produced plants are indeed made by GE plants, it would be very naive to think that humans can "adapt" to all new plant metabolites. Humans are obviously not too good at adapting to rotenone or cyanide, both of which have been present in plants for thousands of years. Since GE can lead to the introduction of novel compounds, 10,000 years of experience with food safety is essentially disregarded by the promoters of GE.

The cited Ames & Gold paper (1997) has nothing to do with the normal consumption of plants and their metabolites. Instead it argues that animals fed almost any "pure" chemical at high enough doses to cause tissue damage will develop cancer due to the increased rate of cell division required for tissue repair, increasing the probability of cell transformation. Aberrant mitogenesis is a major cause of cancer in the developing world due to chronic infections and tissue lesions.

The argument that metabolic profiling would lead to chaos is ridiculous, for it would only have to be done with the few cultivars that are intended for production (the finalists in any given breeding program) and only needs to identify molecules that are toxic or novel. The real reason that the plant biotech companies do not want to do this or any other testing is because they fear potentially hazardous compounds will be detected.

With respect to the extensive quotation from the two co-authors on the Kuiper paper (2001) who supposedly changed their minds on the metabolic profiling issue, it should be pointed out that neither are the senior or corresponding authors on either paper and that they now work for a biotech company as opposed to the unbiased government health agency when they were co-authors on Kuiper's manuscript.

Finally, the comment about 3% of insertions leading to "visible phenotypes" says nothing about the invisible ones related to secondary metabolism, and the comment about the plant's ability to "buffer" itself against genetic changes is only minimally true and says nothing about what the plants are making in the way of compounds that have no visible phenotype, such as secondary metabolites.

Furthermore, plant defense compounds , which are of special concern because they are often also harmful to people, have been shown to be particularly susceptible to change (Schwab, 2003). This makes sense because they have to adapt to co-evolving pests, and this argues against effective buffering for classes of compounds that are of particular concern as toxins and allergens.

Unintended changes. I cite two independent papers by academic scientists showing that lignin levels are elevated in Bt corn and soybeans (Saxena and Stotzky, 2001; Gertz et al., 1999), while Bradford et al. cite one paper published in an agricultural trade journal funded by the Agricultural Biotechnology Stewardship Technical Council claiming that the original papers are incorrect (Jung and Shaeffer, 2004). Contrary data may have many explanations, such as subtle differences in methodologies or measuring somewhat different parameters. This is a common tactic that biotech promoters use frequently to counter any published data that is unfavorable to their industry.

Mutagenicity tests. The Ames test is a valid reflection of mutagenicity and potential carcinogenicity of compounds, and is required for approval for all drugs, cosmetics and chemicals that are released into the environment. It is simple (I have had a 7th grade student run the assays in my laboratory) and very cheap. It is used widely in many parts of the world to test plant products before giving them to humans (Ribnicky et al., 2004; Chen et al., 2003). The "high-dose test" is miscited by Bradford et al. and is not relevant to the issue of food safety, for only plant extracts need be tested. It is clear from the comments of Bradford et al. that they do not understand the Ames test or how it is used in other countries to screen plants and plant extracts.


The response of Bradford et al. to my critique of their article that argued for further reductions in the regulation of GE food crops is typical of GE promoters in that it is both misleading and does not correctly represent the facts. For example:

1. The biotech industry misuses language to redefine scientific terms in order to make the GE process sound similar to conventional plant breeding. Examples from Bradford, et al. include equating recombination with mutagenesis and calling the expression of mRNA from transposons transposition.

2. There is a lack of understanding of elementary biology when it is stated that humans have adapted to plant secondary metabolites rather than humans having selected non-toxic cultivars, as well as the belief that the Ames test is done in animals.

3. I focused my critique on food crops and the GE process, while Bradford et al. frequently cited work with flowers and other non-food crops, but did not state this in their text.

4. Papers were conveniently left out that showed the opposite of their claims. For example, the Netherwood paper that shows horizontal GE gene transfer between ingested plants and gut bacteria.

5. Their definition of phenotype in their original paper is extremely simplistic. It includes only basic agricultural properties. This is a convenient mechanism that allows them to ignore problems associated with more subtle and potentially more dangerous unintended effects.

6. Both Bradford et al. and I failed to point out that the unintended effect of a specific transgene may be directly correlated with transgene expression, not random mutagenesis as assumed (Roessner et al 2001; Gurian-Sherman, 2004).

7. Perhaps the most curious aspect of all is that plant biotechnology is complaining about a regulatory system that was written by their lawyers (Eichenwald et al., 2001) and at least with respect to the FDA is voluntary and lacks safety testing requirements altogether (Gurian-Sherman, 2003; Freese & Schubert, 2004). Although they have what they asked for, they are still complaining about it.


Ahloowalia, B. S., Maluszynski, M., and Nichterlein, K. (2004). Global impact of mutation-derived varieties. Euphytica 135, 187-204.

Ames, B. N., and Gold, L. S. (1997). Environmental pollution, pesticides, and the prevention of cancer: misconceptions. Faseb J 11, 1041-1052.

Bennett, P. M., Livesey, C. T., Nathwani, D., Reeves, D. S., Saunders, J. R., and Wise, R. (2004). An assessment of the risks associated with the use of antibiotic resistance genes in genetically modified plants: report of the Working Party of the British Society for Antimicrobial Chemotherapy. J Antimicrob Chemother 53, 418-431.

Bradford, K. J., Gutterson, N., Parrott, W., Van Deynze, A., and Strauss, S. H. (2005a). Reply to "Regulatory regimes for transgenic crops". Nat Biotechnol 23, 787-789.

Bradford, K. J., Van Deynze, A., Gutterson, N., Parrott, W., and Strauss, S. H. (2005b). Regulating transgenic crops sensibly: lessons from plant breeding, biotechnology and genomics. Nat Biotechnol 23, 439-444.

Chen, Z. L., Gu, H., Li, Y., Su, Y., Wu, P., Jiang, Z., Ming, X., Tian, J., Pan, N., and Qu, L. J. (2003). Safety assessment for genetically modified sweet pepper and tomato. Toxicology 188, 297-307.

Cook, G. W. D. (2005). Genetically Modified Language. The Discourse of Arguments for GE Crops and Food. (London, NY, Routledge Taylor & Francis Group).

Eichenwald, K., Kolata, G., and Petersen, M. (2001). Biotechnology food: from the lab to a debacle. In New York Times, pp. A1.

Einspanier, R., Klotz, A., Kraft, J., Aulrich, K., Poser, R., Schwagele, F., Jahreis, G., and Flachowsky, G. (2001). The fate of forage plant DNA in farm animals, a collaborative case-study investigating cattle and chicken fed recombinant plant material. Eur Food Res Technol 212, 129-134.

Flavell, R. B., Dart, E., Fuchs, R. L., and Fraley, R. T. (1992). Selectable marker genes: safe for plants? Biotechnology (N Y) 10, 141-144.

Freese, W., and Schubert, D. (2004). Safety testing of genetically engineered food. In Biotechnology and Genetic Engineering Reviews, S.E. Harding, ed. (Andover, Intercept Ltd.), pp. 299-325.

Gertz, J. M., Vencill, W. K., and Hill, N. S. (1999). Tolerance of transgenic soybean (Glycine max) to heat stress. Paper presented at: Proceedings of the 1999 Brighton Crop Protection Conference: Weeds (Farnham, UK, British Crop Protection Council).

Gurian-Sherman, D. (2003). Holes in the Biotech Safety Net: FFA Policy Does Not Assure the Safety of Genetically Engineered Foods. Center for Science in the Public Interest, January 2003.

Gurian-Sherman, D. (2004). A Look at the Unintended Effects of Genetically Engineering Food Plants Re: the National Academy of Sciences Report on Unintended Effects. Center for Food Safety.


Jung, H. G., and Shaeffer, C. C. (2004). Influence of Bt transgenes on cell wall lignification and digestibility of maize stover for silage. Crop Sci 44, 1781-1789.

Kuiper, H. A., Kleter, G. A., Noteborn, H. P., and Kok, E. J. (2001). Assessment of the food safety issues related to genetically modified foods. Plant J 27, 503-528.

Levy, A., and Feldman, M. (2004). Genetic and epigenetic reprogramming of the wheat genome upon allopolyploidization. Biol J Linnean Soc 82, 607-613.

Liu, B., and Wendel, J. F. (2000). Retrotransposon activation followed by rapid repression in introgressed rice plants. Genome 43, 874-880.

Madlung, A., Tyagi, A. P., Watson, B., Jiang, H., Kagochi, T., Doerge, R. W., Martienssen, R., and Comai, L. (2005). Genomic changes in synthetic Arabidopsis polyploids. Plant J 41, 221-230.

Netherwood, T., Martin-Orue, S. M., O'Donnell, A. G., Gockling, S., Graham, J., Mathers, J. C., and Gilbert, H. J. (2004). Assessing the survival of transgenic plant DNA in the human gastrointestinal tract. Nat Biotechnol 22, 204-209.

Ribnicky, D. M., Poulev, A., O'Neal, J., Wnorowski, G., Malek, D. E., Jager, R., and Raskin, I. (2004). Toxicological evaluation of the ethanolic extract of Artemisia dracunculus L. for use as a dietary supplement and in functional foods. Food Chem Toxicol 42, 585-598.

Roessner, U., Luedemann, A., Brust, D., Fiehn, O., Linke, T., Willmitzer, L., and Fernie, A. (2001). Metabolic profiling allows comprehensive phenotyping of genetically or environmentally modified plant systems. Plant Cell 13, 11-29.

Saxena, D., and Stotzky, G. (2001). Bt corn has a higher lignin content than non-Bt corn. Amer J Botany 88, 1704-1706.

Schubbert, R., Hohlweg, U., Renz, D., and Doerfler, W. (1998). On the fate of orally ingested foreign DNA in mice: chromosomal association and placental transmission to the fetus. Mol Gen Genet 259, 569-576.

Schubbert, R., Lettmann, C., and Doerfler, W. (1994). Ingested foreign (phage M13) DNA survives transiently in the gastrointestinal tract and enters the bloodstream of mice. Mol Gen Genet 242, 495-504.

Schubbert, R., Renz, D., Schmitz, B., and Doerfler, W. (1997). Foreign (M13) DNA ingested by mice reaches peripheral leukocytes, spleen, and liver via the intestinal wall mucosa and can be covalently linked to mouse DNA. Proc Natl Acad Sci U S A 94, 961-966.

Schubert, D. (2005). Regulatory regimes for transgenic crops. Nat Biotechnol 23, 785-787.

Schwab, W. (2003). Metabolome diversity: too few genes, too many metabolites? Phytochemistry 62, 837-849.