Horizontal Gene Transfer Happens - A Practical Exercise in How Scientific Evidence Must be Interpreted and Used in Accordance with the Precautionary Principle and Sound Science

by Mae-Wan Ho

At first, they said horizontal transfer of genes to unrelated species couldn't happen, then they said "just because it happens in the laboratory doesn't mean it happens in nature". Recently, Prof. Katz of Jena University found in field studies that GM genes may have transferred from GM pollen to bacteria and yeast in the gut of baby bees (The Observer, 28 May, 2000).

That study is not yet published.

But, researchers have earlier found evidence of horizontal gene transfer of GM genes to soil bacteria in the field where GM sugar beet was planted, and this has been reported in the scientific literature (1). Readers of ISIS News will note that there have already been several studies documenting the horizontal transfer of GM genes from GM plants to soil fungi and bacteria in the laboratory. (2).

In this article, I shall review the published study to show how the precautionary principle can be applied in practice to interpret and use scientific evidence responsibly and in accordance with sound science.

German geneticists Frank Gebhard and Kornelia Smalla began a series of experiments in 1993 to monitor field releases of GM rizomania-resistant sugar beet (Beta vulgaris) for persistence of the GM construct in the soil and for horizontal gene transfer.  They found that the GM construct has persisted in the soil for at least two years after the plants were grown and harvested, and different parts of the GM construct may have transferred to unknown soil bacteria.

The researchers are exemplary in documenting clearly their experimental material as well as the procedure, and I take pleasure in reporting their research in some detail. The GM sugar beet contained the following genes.

*       BNYYV cp (the coat protein of Beet Necrotic Yellow Vein Virus, with CaMV 35S promoter (from the cauliflower mosaic virus)  and 3'nos terminator (from soil bacterium Agrobacterium tumefaciens). A promoter is a gene switch required to turn the gene on, ie, to transcribe the gene; a terminator, in this context, is a genetic signal to ensure that the gene transcript will be translated into protein.

*       Marker genes nptII  (neomycin/kanamycin phosphotransferase (from Tn 5, a bacterial transposon) with terminator 3'ocs (from A. tumefaciens) and bar  (phosphinotricin acetyltransferase (from Streptomyces hygroscopicus, another soil bacterium) with terminator 3'g7 (source unspecified) both under the control of the bidirectional TR1/2 promoter (from A.  tumefaciens).

These two marker genes confer resistance, respectively, to the antibiotic kanamycin (Km) and the herbicide glufosinate ammonium.

PCR (Polymerase Chain Reaction) primers (short DNA sequences complementary to and hence specific for different parts of the construct) allowed amplification and detection of even trace amounts of GM construct. Bacteria in the soil samples were cultivated in media with, and without kanamycin, in order to detect the proportion that is kanamycin-resistant. Individual kanamycin resistant colonies were probed for the GM construct. To detect GM construct independently of cultivation, total soil DNA was extracted and amplified by PCR with the three different primer sets.

The GM construct or parts of it was found to have persisted for up to 2 years under field conditions and in soil microcosms with introduced GM plant DNA for up to six months. Let us look at the findings regarding horizontal gene transfer.

*       GM sugar beet litter introduced into the soil led to an increase in both the Km resistant and total bacterial populations. Most of the kanamycin resistant bacteria are those that already exist in the soil, as antibiotic resistance is widespread. Though the authors did not comment on it, the proportion of resistant bacteria did increase significantly between 1.5  and 2 years, suggesting that this increase may be due to the transfer of kanamycin resistance marker genes from the GM construct to soil bacteria. It takes time for litter to rot and the DNA contained to be released.

*       A total of 4000 isolates of Km resistant bacterial colonies were individually screened with a "dot blot" technique to identify sequences that bind to, or "hybridize with" GM -specific probes. This technique is more direct, but much less sensitive than PCR. "A few isolates giving weak hybridization signals ....were detected". These were checked with the PCR technique, but none gave PCR products, and hence the authors dismissed the results as false positives. There are obvious limitations to this experiment. First, 4000 is a small number of isolates, and most of them are probably from bacteria already carry pre-existing kanamycin resistance.

Second, the failure to obtain PCR products can be due to the fact that fragments of the GM constructs or rearranged versions of the GM construct have been transferred.  In order to rule out these possibilities, it is necessary to carry out more extensive molecular analyses, which were not done.

*       Construct-specific DNA was found in practically all soil samples 6 months after GM sugar beet litter was introduced into the soil, while no GM-specific DNA was present in the soil with young GM plants. GM-specific DNA persisted for up to 2 years in the field. This suggests that GM-DNA is released mainly after the plant litter has disintegrated.

*       When total bacteria from soil were isolated, treated with DNAase (enzyme which break down DNA) to remove free DNA, two out of seven samples were found to contain GM construct after 18 months. This again suggests that horizontal gene transfer has occurred. The authors were careful not to rule out the possibility that GM-DNA may simply have "adsorbed" onto the external surfaces of the bacteria.

*       Soil microcosm studies to which free DNA from the GM sugar-beet was added showed that the intensity of the signal for GM construct decreased during the first days and subsequently increased (strongest at 23 days).

This suggests that the GM-DNA may have been taken up by soil bacteria and become amplified by the multiplication of the bacteria; though the authors did not state this explicitly, nor offer any other explanation for the observation.

*       Soil samples from the microcosm experiments were plated and bacterial lawns grown for 4 days. These were harvested, treated with DNAse and the DNA released from the bacteria by boiling and freezing. PCR amplification with all three primer sets resulted in several positive signals, "which might indicate uptake of transgenic [GM] DNA by competent bacteria". But, "Because the isolates carrying the construct-specific DNA sequences were not accessible, an interpretation of the signals remains inconclusive."

The authors are scrupulously careful not to interpret the results as proof that horizontal gene transfer has taken place. The results, however, are prima facie evidence of horizontal gene transfer. The failure to isolate the bacteria which have taken up the GM construct is not surprising, as over 99 percent of soil bacteria are not isolatable by current culture techniques, and this is one major limitation to detecting horizontal gene transfer in the field. The authors further state, "The presence of bacterial genes, promoters, terminators, or origins of vegetative replication in transgenic plants will enhance the probability of stable integration of DNA stretches based on recombination events [should transgenic DNA be taken up by the bacteria]." (pp. 270-1).

I have reviewed the scientific paper in detail as a practical illustration of how scientific evidence is to be used responsibly, in accordance with the precautionary principle, which is also sound science. The precautionary principle states that where there is reasonable suspicion of harm, scientific uncertainty or lack of scientific consensus must not be used to postpone preventative action. Uncertainty is the hallmark of any active knowledge system, which is what science is, as opposed to religious fundamentalism. And this is ultimately why the precautionary principle must be part and parcel of sound science. The valid use of scientific evidence is to set precaution, and not to set permissive standards for scientists and corporations to use life and our life-support system as one vast laboratory, as has been the case for the past 50 years.

Gebhard and Smalla's paper does not provide positive proof of horizontal gene transfer that would stand up in a court of law, but it does provide reasonable suspicion that horizontal gene transfer has occurred, especially as it corroborates previous laboratory investigations demonstrating that horizontal gene transfer. There is already overwhelming evidence that horizontal gene transfer and recombination have created new bacterial and viral pathogens and spread drug and antibiotic resistance among the pathogens. GM constructs consist predominantly of bacterial and viral genetic material as well as antibiotic resistance marker genes. To persist in ignoring horizontal gene transfer in risk assessment not only violates the precautionary principle, it violates all the tenets of sound science and responsible governance.


1.      Gebhard, F. and Smalla, K. (1999). Monitoring field releases of genetically modified sugar beets for persistence of transgenic plant DNA and horizontal gene transfer. FEMS Microbiology Ecology 28, 261-272.

2.      For a more recent review, read "Horizontal Gene Transfer - Hidden Hazards of Genetic      Engineering" by Mae-Wan Ho, to be posted on ISIS website.