CORRESPONDENCE

http://biotech.nature.com • FEBRUARY 2002 • VOLUME 20 • nature biotechnology

tide. Our results broaden the hypothesis ofMourez et al. that the YWWL sequence maybind to the PA heptamer at the hydrophobicarea that interacts with the EF/LF factors(Fig. 1B,C), and identify the crucial site.

Using the technology devised for a much-publicized distributed-computing screen-saver project4, which now uses more than 1 million computers, we plan to initiate amassively distributed search among a virtuallibrary of 3.5 billion small molecules forcompounds that can bind to the PA heptamer–binding site identified by ouralgorithm2. Our approach may suggest newmolecules that can prevent the assembly ofthe anthrax toxin complex.

Meir Glick, Guy H. Grant,and W. Graham Richards, Department of

Chemistry, Central Chemistry Laboratory,University of Oxford, South Parks Road,

Oxford, OX1 3QH, UK(graham.richards@chem.ox.ac.uk)

1. Pannifer, A.D. et al. Nature 414, 229–233 (2001).2. Glick, M., Robinson, D.D., Grant, G.H., & Richards

W.G. J. Am. Chem. Soc., in press (2002).3. Petosa, C. et al. Nature 385, 833–838 (1997).4. http://www.chem.ox.ac.uk

the EU, so as to provide input to a workshoporganized by BEUC in September 20012.

To that end, we analyzed several notifica-tion dossiers for products derived from GMvarieties of maize and refined oil derivedfrom GM varieties of oilseed rape. Ouranalysis showed that compositional datasubmitted on the content of macro- andmicronutrients, vitamins, inherent planttoxins, and anti-nutrients lacked consisten-cy from case to case. Furthermore, thedesign of the GM crop field trial, the geo-graphical locations and seasons of plantingand harvesting, and the choice of controldiffered considerably from case to case.These case studies illustrated the lack of anoperational definition of the concept ofsubstantial equivalence in the EU.

As an alternative model, we have pro-posed that an operational definition forfood-safety assessment should include aminimum list of macro- and micronutri-ents, anti-nutrients, inherent plant toxins,secondary metabolites, and allergens to beanalyzed for each GM crop species and oftheir baseline concentrations in conven-tional varieties. It is also necessary to cre-ate detailed protocols to guide the designof field trials, to establish validated tech-niques that can reliably ascertain the con-tent of these compounds in plants, and touse common methods to analyze the datastatistically.

The participants at the BEUC workshopconcluded that “substantial equivalence”has been a controversial and misleadingterm for consumers that should not be usedin regulatory decision-making. They recog-nized that it remains crucial to systematical-ly detect unintended changes in the compo-sition of GM crops compared with anappropriate control, as such changes may beof toxicological, immunological, or nutri-tional concern.

The OECD, the European Commission,EU member states, and the Europeanbiotechnology industry association(EuropaBio) are now drafting “minimumlists” to underpin “substantial equivalence.”Although this is a good first step, there aresome discrepancies among these drafts. Forexample, in the case of GM maize, the Dutchauthorities have required additional dataabout five secondary metabolites that havenot been included in the list proposed byEuropaBio. Despite these differences, theefforts made by different bodies to stan-dardize these lists should be welcomed.

In July 2001, the European Commissionproposed to abandon the “light” notifica-tion procedure, which would mean that allGM food (components) would need toundergo a full authorization process. Theirsafety assessment obviously still requires a

systematic detection of unintended changesGM food (components) compared with anappropriate control. But whether it is usefulin that context to cling the term “substantialequivalence” is doubtful.

Piet Schenkelaars, SchenkelaarsBiotechnology Consultancy, Niels Bohrweg

11–13, 2333 CA Leiden, The Netherlands(pschenkelaars.sbcbiotech@planet.nl)

1. Millstone, E., Brunner, E., & Mayer, S. Nature 401,525–526 (1999).

2. Schenkelaars Biotechnology Consultancy. GM foodcrops and application of substantial equivalence inthe European Union. Commissioned byConsumentenbond and the Dutch Foundation‘Consument & Biotechnologie’ (SBC, Leiden, TheNetherlands, June 2001). The full report can bedownloaded at http://www.consubiotech.nl/

119

Rethinking substantial equivalence

To the editor:In 1993, the Organisation for EconomicCooperation and Development (OECD;Paris) introduced the concept of substantialequivalence as a guiding principle for safetyassessment of food containing a GM compo-nent. Since then, several meetings of expertssponsored by the OECD or jointly by theFood and Agricultural Organization (Rome)and World Health Organization (Geneva,Switzerland) have endorsed the use of thisconcept in regulatory decision-making.Under European Union (EU; Brussels)“Novel Food” Regulation 258/97, establish-ment of the “substantial equivalence” of aGM food (component) also is used to deter-mine whether a “light” notification proce-dure or a “heavy” authorization procedure isfollowed for a particular product.

In recent years, scientists concerned aboutthe potential health risks of GM foods havechallenged substantial equivalence as inade-quate for assessing food safety. In 1999, forexample, a commentary in Nature1 labeledthe concept “inherently anti-scientificbecause it was created to provide an excusefor not requiring biochemical and toxicolog-ical tests.”

Against this background, our consultancywas commissioned by the DutchConsumentbond and the European con-sumer organization Bureau Européen desUnions de Consommateurs (BEUC) to studythe application of substantial equivalence in

ES cell media

To the editor:The article by Xu et al. in the October issuedescribes the successful feeder-free propaga-tion of undifferentiated human embryonicstem (ES) cells (Nat. Biotechnol. 19, 971–974,2001). The authors reported, however, that inconditioned media from mouse embryoniccell lines and from human immortalized lines,the feeder-free human ES cells did not retainall the features of cells grown on feeder layers.

Did Xu et al. have the occasion to testhuman embryonic fibroblasts, which areavailable from cell repositories like theAmerican Type Culture Collection(Manassas, VA) and the European Collectionof Cell Cultures (Salisbury, UK), as feeders oras a source of conditioned medium? Theanswer might be of great help to the growingscientific community involved in the study ofhuman ES cells.

Massimo Battaglia, Institute of Neurobiologyand Molecular Medicine, Consiglio Nazionele

delle Ricerche, Rome, Italy(mbattaglia@mclink.it)

Chunhui Xu and Melissa Carpenter reply:We demonstrated that human embryonicstem (hES) cells can be maintained on appro-priate matrices and media conditioned byirradiated primary mouse feeders (MEF-CM) while maintaining all of the essentialfeatures of hES cells grown on feeders. WhenhES cells were grown in conditioned mediafrom a human telomerase immortalizedneonatal fibroblast line (BJ5ta) and embry-onic murine fibroblast lines (NHG 190 andSTO), fewer colonies with appropriate mor-phology were observed. Evaluation of addi-tional immortalized human cell lines is cur-rently in progress.Chunhui Xu and Melissa K. Carpenter, Geron

Corporation, Menlo Park, CA(mcarpenter@geron.com)

©20

02 N

atu

re P

ub

lish

ing

Gro

up

h

ttp

://b

iote

ch.n

atu

re.c

om