Who's Afraid of Horizontal Gene Transfer?
The Maize Gene War

6 March

Controversy erupted following claims by Berkeley scientists that transgenic maize has contaminated Mexican maize landraces. Horizontal gene transfer- the direct uptake and incorporation of foreign DNA into cells - is something pro-biotech scientists don't like to mention and top scientific journals like Nature are extremely reluctant to discuss. Dr. Mae-Wan Ho and Prof. Joe Cummins expose the weaknesses of the attack on the Berkeley researchers.

Ignacio Chapela and David Quist reported in Nature last November that transgenic maize genes were found in maize landraces growing in remote regions of Oaxaca, Mexico. The highlands of Oaxaca, Chiapas, and adjacent Guatemala are one of seven centers of genetic diversity for maize. To protect this diversity, an invaluable resource for crop breeders, the Mexican government declared a nation-wide moratorium in 1998 on planting transgenic maize.

Concerted attempts to discredit the research paper began almost immediately (see "Transgenic pollution by horizontal gene transfer?" Science in Society 13/14, February 2002), and has escalated since (see current issues of Nature and Science).

On 21 February, Mexican newspapers reported that two teams of government researchers have confirmed the findings. But a scathing editorial in the February issue of Transgenic Research written by editor Paul Christou charged that Chapela and Quist had presented no credible evidence to justify their claims.

To add to the confusion, Elena Alvarez-Buylla Roces, a biologist at the National Autonomous University of Mexico who appeared with Chapela at a press conference supposedly announcing the confirmation of his findings, insisted in a later e-mail to Science that the Mexican investigators still do not have definite answers.

Widely circulating anonymous e-mails have been accusing Chapela and Quist of conflicts of interest and other misdeeds. A notorious Fellow of the UK Royal Society, previously linked to a libel case against Greenpeace, posted a message to the pro-biotech website http://www.agbioworld.org/, suggesting that Berkeley asks Chapella to release his samples to be checked by others. "Refusal to do so should then be used to request Berkeley to relinquish Chapella's position".

Meanwhile, 144 civil-society groups have leapt to the Chapella's defence, in a joint statement issued on 19 February accusing the biotech industry of using intimidatory techniques to silence dissident scientists.

The molecular data under dispute are suggestive of horizontal gene transfer, and a lot of the confusion is created by a failure to acknowledge this possibility. The cauliflower mosaic virus (CaMV) 35S promoter was often found without downstream sequences, as would be expected from the recombination hotspot, or frequent breakpoint, known to be associated with the promoter.

One of us put this possibility to Chapela. He agreed, but indicated that Nature would never have published the paper if they had mentioned horizontal gene transfer. This was confirmed when a correspondence submitted to Nature (reproduced at the end of this article) was rejected.

The structural instability of transgenic DNA - its tendency to break and rejoin - is now undeniable. Reports of deletions, rearrangements and duplications are now legion (see Transgenic Instability, ISIS Reprints, ISIS Publications, March 2002). When the CaMV recombination hotspot was reported in 1999, we called for all GM crops containing the promoter to be withdrawn. The recombination hotspot further increases transgenic instability and enhances horizontal gene transfer and recombination.

Two years later, one of the scientists who first reported on the CaMV recombination hotspot recommended that it should be phased out, not on grounds of safety but on grounds that structural instability will compromise agronomic performance.

Imagine our surprise to find this same scientist now mounting the major attack on Chapela and Quist.

Paul Christou's editorial in the current issue of Transgenic Research (2002, 11, iii-v.), states, "Cross pollination and introgression would not produce those results." We quite agree. This was exactly what was stated in the correspondence rejected by Nature. But there our agreement ends.

"Sample contamination is the most likely explanation", he continues. This would be unlikely to produce such a diversity of sequences downstream of the CaMV 35S promoter that Chapela and Quist have found. Sample contamination is a very common and convenient way to discredit results one does not like.

So what is the main criticism? "Rather than rely on questionable PCR results, plants that were alleged to contain introgressed DNA should have been grown out and subject to more reliable confirming studies."

This criticism is irrelevant. The authors were looking for rare individual transformation events among the hundreds of kernels on each cob. PCR is the only method available for detecting rare events. To grow out all the kernels would require growing up hundreds of plants and then screening all of them individually. That would be quite a different project and it would take a lot more time and resources.

The inverse PCR results were "technically flawed" because only CaMV 35S promoter was analysed. Chapela and Quist were specifically looking for sequences linked to the CaMV 35S promoter, and they found a diversity of them, many not linked to other parts of the transgenic DNA in transgenic maize.

Christou, however, made no mention of the recombination hotspot of the CaMV 35S promoter reported by his group (Kohli et al, Plant J. 17, 591, 1999), nor did he mention that his group had recommended that the promoter should be phased out on that account.

Instead, he stated, "The fact that the authors have not been able to show the presence of intact inserts, which are more likely to be present than fragments of unknown origin, casts further doubt that the results observed come from a transgenic plant source." That would be true if simple cross-pollination from a stable transgenic line had occurred.

If horizontal gene transfer were involved, the transgenic DNA would most likely have been transferred to the plant cells as fragments. How might that have occurred? Insects with sharp mouthparts could have visited transgenic maize and landraces in succession. Or else composted transgenic maize might have been added to the soil in which the landraces were cultivated.

In the very next sentence, he ruled out recombination associated with the CaMV 35S recombination hotspot, which his own group discovered. "Recombination is not a satisfactory explanation either, since multiple generations of crossing have been done with all these constructs and they have been shown to be stable - or else they would not have made it through the regulator system."

Except that there is no evidence for the stability of any transgenic line. The required molecular data simply do not exist. We know, because we have challenged the companies to produce such data for years, when our regulators have failed to ask for them. Such event-specific molecular evidence is only now required in the new EC Directive.

More to the point, Monsanto's Roundup Ready Soya, after years of having been commercially grown, was the first and still the only commercial crop to be analysed in this manner last year. The results showed that the transgenic DNA was indeed scrambled, as was the host genome at the site of insertion, and more than 500 bps of unknown DNA was present as well. They were very different from the original data provided by Monsanto.

Christou's selective reporting is nothing new. He accused us of having misrepresented the stability of transgenic rice, and referred to a paper of his claiming to show that transgenic rice lines were stable. A detailed review of that paper revealed that, on the contrary, the vast majority of the lines were not stable (see "Questionable 'stability' at JIC" ISIS Report, February 2001, in Transgenic Instability, ISIS Reprints, ISIS Publications, March 2002).

We cannot tolerate the personal attacks meted out to all scientists who have ever reported findings that question the safety of GM. If Christou and others like him are dissatisfied with the lack of definitive experimental evidence, so are we. And it is up to them to convince us by performing the experiments and presenting them to public scrutiny. Meanwhile, there should be no further environmental releases of GM crops.

Correspondence to Nature by Mae-Wan Ho

(submitted 20 December 2001, rejected a month later)


Transgenic Pollution by Horizontal Gene Transfer?

The evidence that landraces growing in remote regions in Mexico have been contaminated by transgenic maize1 deserves further consideration.

Four of the six samples (cobs) tested positive for the CaMV 35S promoter used in all transgenic crops commercialised, while the blue maize of Cuzco Valley in Peru and seed samples from historic collection in Sierra Norte de Oaxaca tested negative.

Sequence analysis at the site of transgene insertion by inverse PCR yielded 1 to 4 DNA fragments differing in size in each sample. The sequences downstream of the CaMV 35S promoter were diverse. Two sequences were similar to synthetic constructs containing regions of the adh1 gene found in transgenic maize currently on the market, such as Novartis Bt11. Other sequences represented the criollo maize genome, including retrotransposon regions, whereas others showed no similarity to any GenBank sequence.

As a moratorium on planting transgenic maize has been in place in Mexico since 1998, it was suggested that the contamination might be due to "loose implementation of the moratorium", or to "introgression before 1998 followed by the survival of transgenes in the population".

However, simple cross-pollination cannot explain the fragmentary, diverse nature of the transgene contamination, which is a sign of horizontal gene transfer and recombination.

It is significant that all the contaminated samples had acquired the CaMV 35S promoter, with the rest of the transgenic construct either missing or recombined. This observation is consistent with our warning that CaMV 35S promoter has a recombination hotspot, and is hence expected to enhance horizontal gene transfer and recombination.2-4 We have demanded all transgenic crops with CaMV 35S promoter to be immediately withdrawn in 1999. Since then, researchers who have discovered the CaMV 35S recombination hotspot have recommended that the promoter should no longer be used5 but fell short of calling for existing crops containing it to be withdrawn.

1.. Quist, D. & Chapela, I.H. Nature 414, 541-543 (2001).
2.. Ho, M.W., Ryan, A. & Cummins, J. Microbial Ecology in Health and Disease 11, 194-197 (1999).
3.. Ho, M.W., Ryan, A. & Cummins, J. Microbial Ecology in Health and Disease 12, 6-11 (2000).
4.. Ho, M.W., Ryan, A. & Cummins, J. Microbial Ecology in Health and Disease 12,189 (2000).
5.. Christou, P et al, John Innes Centre & Sainsbury Laboratory Annual Report 1999/2000.