NATURE BIOTECHNOLOGY VOLUME 25 NUMBER 11 NOVEMBER 2007 1213

Open season?To the editor:In the September issue, your editorial entitled “A tragic GM ‘outing1’’ comments that “those who embrace GM crops must do it openly, as democratic society demands. Otherwise, activists will exploit secrecy to foment public mistrust, portraying themselves as heroes exposing covert GM planting operations.”

I waited for the other shoe to drop. Does not society in the shape of government and the police have a duty to protect and defend from attacks by vandals and other fanatics those going about their legitimate business—as Claude Lagorse was doing? Would those vandals not have attacked the farm if details had been widely published? Were those secret vandals after M. Lagorse’s secrecy or his crops?

You went on: “Ultimately, transparency and openness will make the continued destruction of property and intimidation of farmers difficult to justify.” In your view, does that mean that such destruction and intimidation is presently justified? Would you be happy if those intimidators ransacked the premises of Nature Biotechnology for publishing what they saw as ‘pro-GM material’? Should you therefore not be more open, inviting those very intimidators, qualified or not, to be part of your editorial process?

After all, they attack GM crops with little or no knowledge of agriculture—just as little as they doubtless have of biotech in general.

Vivian Moses

CropGen, P.O. Box 38589, London SW1A 1WE, UK e-mail: cropgen@f2s.com

1. Anonymous. Nat. Biotechnol. 25, 950 (2007).

Nature Biotechnology respondsNature Biotechnology unequivocally condemns the illegal actions of those who engage in intimidation and who vandalize legitimate cultivations of transgenic crops on private property. Those who are involved in such criminal acts should be prosecuted under the judicial system.

Our editorial sought to address the problem that round-the-clock police

protection and surveillance of transgenic crop plantings is impractical, given the resources and time involved. Given the difficulties faced by police and the rather mixed track record of the judicial system

in handing down stiff sentences to those found guilty of such offenses, how then should open democratic societies, such as France, respond to a small but significant minority who persist in this type of criminal activity?

Our answer is that any tacit public support that exists for such individuals should be marginalized to the extent that the actions of such people are

no longer tolerated. Make no mistake, this is a battle for the hearts and minds of the public, and biotech must clearly be on the side of the angels. In this regard, openness and transparency are key. Making the locations of trials of transgenic crops secret or even opaque merely plays into the hands of activists by making it appear that the government and the growers have something to hide.

Unfortunately, recent statements from the French environment minister, Jean-Louis Borloo, have suggested that his government may acquiesce to pressure for a moratorium of GM crops from José Bove and his acolytes. According to the French newspaper Le Monde, a freeze is reportedly planned on all transgenic crop trials1. In September, French representatives to the European Council of Ministers also abstained from voting on the import approval of three transgenic maize lines; the veto could hinder current negotiations on the extension of approval for MON810 corn—currently the only transgenic crop approved for cultivation in France.

As Nature Biotechnology went to press, a public consultation process about transgenic crops was underway in 15 French cities and on the internet. In addition, a working group on genetically modified organisms has been set up to discuss new legislation oriented towards transparency and the freedom of choice for farmers and for consumers and potentially the establishment of an independent national advisory body.

1. Jakubyszyn, C. & Kempf, H. Le Monde 20 September (2007) <http://www.lemonde.fr/web/article/0,1-0@2-3244,36-957270@51-951150,0.html>

GMO quantification in processed food and feedTo the editor:Reliable quantification of genetically modified organisms (GMOs) in food and feed is mandatory to fulfill European Union (EU) Regulations on the labeling of products containing GM ingredients over a 0.9% threshold—a threshold recommended to be defined as the ratio of genetically modified and unmodified haploid genomes1–4. A major challenge of these regulations is that they also require the labeling of highly processed products “containing, consisting or produced from GMOs,” where DNA is absent or heavily damaged and therefore

difficult to detect and quantify2.

Huge efforts are underway to develop and officially validate tools for GMO analysis. Polymerase chain reaction (PCR) and real-time PCR have become the methods of choice for GMO detection and quantification. It is possible to extract high-quality DNA for GMO analysis from raw materials like seeds or from Certified

Reference Materials (CRMs). Unfortunately, many processed food and feed products are often non-optimal sources of DNA5: food processing procedures often result in DNA fragments as small as (or even smaller than)

CORRESPONDENCE©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

ebio

tech

nolo

gy

1214 VOLUME 25 NUMBER 11 NOVEMBER 2007 NATURE BIOTECHNOLOGY

~200–1,000-bp fragments6. Moreover, in several refined products, such as glucose syrup, oil or soy lecithin, DNA is depleted to vanishingly small quantities. And, in several complex food products (e.g., sausages or chocolate), DNA may be embedded in substances that inhibit PCR reactions. Thus, overall, the more products are processed and complex, the poorer the quality of the extracted DNA.

Even though it is possible to extract and purify DNA from most process foods and feeds and analyze it by means of PCR, the key question is, how accurate are analytical results from these non-optimal DNA samples?

Where qualitative PCR analysis is performed, systems designed to boost sensitivity and specificity can enable detection of GMOs in highly degraded DNA samples. But what about quantitative analysis?

Most of the event-specific, real-time PCR protocols validated by the Community Reference Laboratory (CRL; Ispra, Italy) have been optimized on high-quality DNA samples extracted from sources such as leaves or seeds (http://gmo-crl.jrc.it/statusofdoss.htm)7; they have not been optimized for DNA samples extracted from sources, such as processed food or feed samples. Moreover, official CRL validation schemes do not include DNA samples simulating DNA degradation consequent to food processing. This is mostly due to the difficulty of reproducing all possible processing procedures used in industrial food manufacture. Consequently, good performance measured in validation studies cannot be assured when analyzing DNA extracted from processed food or feed samples.

Several factors influence the applicability and reliability of quantitative PCR analysis on processed samples. First, DNA degradation depletes amplifiable DNA sequences from the analytical sample. This influences the limit of detection (LOD) and the limit of quantification (LOQ) of a method, in that fewer intact target sequences are available as a consequence of DNA fragmentation8. This problem can usually be overcome by designing specific primer-probe systems targeting short marker sequences ~60–150 bp in length6.

Second, in real-time PCR, two measurements are performed, one targeting a GM event-specific marker sequence and a second targeting a species-specific wild-type reference-sequence, enabling detection of all genomes. The relative GMO content is

obtained by dividing the value of the GMO-specific measurement with the value of the wild-type reference sequence. Methods are usually developed using the convenient ‘modular approach’: that is, one universal primer-probe system (two primers/one probe) targeting a reference sequence from a particular species (e.g., maize, rice or soy) coupled with different GM event specific primer-probe systems (two primers/one probe). These assembled real-time PCR methods are validated by the CRL and, when they are found to perform well, become official methods in the EU.

Unfortunately, the species-specific and the event-specific amplicons often show significant differences in length and/or base composition of target sequences. This is not a problem as long as quantification is performed on DNA samples with degradation levels similar to those of the Certified Reference Materials (CRMs) used as calibrators. Eventual amplicon size and base composition differences can be considered negligible. But in the case of highly degraded DNA samples extracted from processed food and feed, the abundance of longer target sequences will most probably be underestimated as a consequence of the statistically higher probability that a nick will fall within the amplicon sequence9.

Third, DNA extraction from processed food products often represents a compromise between obtaining a high yield of DNA and keeping PCR inhibitors at low concentrations. Under such conditions, the choice of the most suitable DNA extraction methodology for a given matrix (i.e., processed food or feed product) becomes crucial. Several extraction methods have been developed or are under development, but it is not always possible to extract DNA with the required quality and purity, rendering real-time PCR quantification difficult. In addition, for example, Corbisier et al.6 have observed drifts in transgenic DNA content estimation when DNA from highly degraded GMO samples is purified using different extraction methodologies.

Fourth, the particle size homogeneity in a product can be a serious issue. In a recent study, Moreano et al.9 have shown how common food production processes can lead to GMO quantification distortions. Among other things, DNA extracted from larger-sized particles is under-represented with respect to DNA extracted from smaller sized particles. As a consequence, GMO content can be over- or underestimated as a result of different particle size distributions.

Finally, food products often contain different components derived from the same species, such as soybean flour, lecithin and oil. In practice, all components deriving from one species are considered together as one single ingredient3. Having one ingredient with components depicting different levels of DNA degradation and/or depletion mixed in variable unknown ratios renders a precise quantitative analysis virtually impossible. The true GMO content of such a product will necessarily be over- or underestimated.

All of these issues suggest that caution should be exercised when analyzing processed food and feed for GMO content. In normal practice the GMO labeling framework relies on the analysis of raw materials before processing and then their traceability throughout the production chain up to the final product. Even here, the analysis of raw materials must be extremely well tuned to avoid errors (for example, Macarthur et al.10 have reported the effect of heterogeneity of bulk lots of food and feed on the performance of PCR detection methods).

And yet, in this respect, the EU Regulations2 require that what is generically defined as ‘food’ or ‘feed’ can be analyzed regardless of the level of processing of the samples. Consequently, it is not excluded that national competent testing authorities could elect to analyze for GMO content processed products sampled directly from the shelf—a clearly flawed approach. As postmarketing analysis of processed food or feed samples is potentially error prone and risks mislabeling products, a sharply defined international approach for the analysis and certification of raw materials is thus the most prudent way forward for labeling food and feed.

Florian Weighardt

Via Milano 1095, 21027 Ispra (VA), Italy. e-mail: florian.weighardt@poste.it

1. Tsioumani, E. Rev. Eur. Comm. Int. Environ. Law 13, 279–288 (2004).

2. The European Parliament. Off. J. Eur. Commun. L 268, 1–23 (2003).

3. Weighardt, F. Nat. Biotechnol. 24, 23–25, (2006).4. The European Commission. Off. J. Eur. Commun. L

348, 18–26 (2004).5. Miraglia, M. et al. Food Chem. Toxicol. 42, 1157–

1180 (2004).6. Corbisier, P. et al. Anal. Bioanal. Chem. 383, 282–290

(2005).7. Trapmann, S. & Emons, H. Anal. Bioanal. Chem. 381,

72–74 (2005).8. Spiegelhalter, F., Lauter, F.R. & Russel, J.M. J. Food

Sci. 66, 634–640 (2001).9. Moreano, F., Busch, U. & Engel, K. H. J. Agric. Food

Chem. 53, 9971–9979 (2005).10. Macarthur, R. et al. Nat. Biotechnol. 25, 169–170

(2007).

CORRESPONDENCE©

2007

Nat

ure

Pub

lishi

ng G

roup

ht

tp://

ww

w.n

atur

e.co

m/n

atur

ebio

tech

nolo

gy