Over the next decade, plant scientists should also consider tackling the topic of molecular ecology, a discipline with difficult experimental set ups and lengthy periods of observation. The dynamics of plant evolution, diversifica- tion and adaptability need to be studied and understood. We all want a sustainable, environmentally friendly agriculture, a less polluting industry, and better protection of the remaining wildlife. To achieve this, we will have to rely on the implementation of new technologies driven by innovative plant science. With the increasing growth of the world’s population and the economic push for an agricul- ture that can produce not only food but also materials for industry, including biofuels, the remaining ‘pristine’ nature is threatened more than ever. 1360-1385/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tplants.2005.10.013 Plant biotechnology kicks off into the 21st century Richard A. Dixon Plant Biology Division, Samuel Roberts Noble Foundation, Ardmore, OK 73402-2180, USA In January 1996, I wrote a research news article in the first volume of Trends in Plant Science entitled ‘Plant Biotechnology shapes up for the 21st century’. It sum- marized key talks at an international symposium held at the University of Kentucky on engineering plants for commercial products and applications. Nearly ten years on, it is interesting to look back at how the earlier predictions have held up. This issue of Trends in Plant Science stands testament to the great strides made in plant biotechnol- ogy since 1996. Genomics (originally DNA and transcript based, but recently extended to integrate the proteome and metabolome) has revolutionized the speed of gene discovery for important plant traits. This has facilitated metabolic engineering of increasingly complex pathways to introduce health-beneficial compounds such as vitamins A and E (reviewed in this issue by Salim Al-Babili and Peter Beyer and Dean DellaPenna, respectively), polyunsaturated fatty acids, phytoestrogens and flavonoid antioxidants into plants. Advances in protein engineering have enabled production of several pharmaceutically active proteins, including vaccines, in plants (reviewed by Julian Ma et al.). However, there remains a disconnection between scientific discovery and commercialization. As Julian Ma et al. point out in their review, there are still no commercialized pharmaceutical products from plant biotechnology. The Golden Rice story exemplifies the complex interrelationships and tensions between basic science, commercialization, politics and sociology. The break- through of introducing a complete provitamin A pathway into the rice endosperm through the introduction of three transgenes spawned major debates on the efficacy of the product, its safety, and intellectual property issues for public–private partnerships. As is often the case with new technologies, governmental regulations lag behind the science, and public perception takes even longer to ‘come around’. The good news is that, even though Golden Rice has yet to be delivered to those most in need of it in developing countries, there is now a path laid out to this end, and the product can be improved to deliver additional vitamins (see Review by Dean DellaPenna) and micronutrients (see Review by Philip White and Martin Broadley). It was realized in 1996 that genomics would play a major role in driving plant biotechnology. With complete or ongoing genome projects in Arabidopsis, rice, poplar, Medicago, Lotus, tomato and maize, among others, the problem for the researcher now is not lack of data (essentially limited to EST and mapping information for Arabidopsis in 1996) but how best to interpret the wealth of publicly available gene sequence and expression information. This is a particular challenge in the area of secondary metabolism. Collectively, plants probably produce O200 000 natural products, many of which have value as pharmaceuticals or nutraceu- ticals, but these compounds are often genus or even species-specific. Too many compounds, too many genes! The key to rationalizing this apparent complex- ity is to understand the rules that govern structure– function relationships among the different protein families that feature repeatedly in creating the chemi- cal diversity of plants. Large gene families encoding such enzymes as cytochrome P450s, glycosyltrans- ferases, methyltransferases, acyltransferases and prenyltransferases make up a significant proportion of plant genomes. Some of these enzymes are catalyti- cally promiscuous and many are incorrectly annotated in plant genome databases. Combining protein struc- ture information (to develop predictive rules for functional annotation based on structure) with expression data related to substrate and product levels and localization (i.e. spatially resolved transcriptomics and metabolomics) will greatly aid gene functional annotation. Current bottlenecks to these approaches are throughput for structure determination by X-ray crystallography, and sensitivity for metabolomic analysis at the cellular level. As many of the articles in this issue attest, once genes are functionally characterized, plant metabolic engineering is moving apace, even in pathways such as terpenoid biosynthesis that once seemed chemically ‘difficult’ (reviewed by Harro Bouwmeester and colleagues). Gene delivery is still a major issue in plant biotechnology, particularly for complex pathway Corresponding author: Dixon, R.A. (radixon@noble.org). Editorial TRENDS in Plant Science Vol.10 No.12 December 2005 560 www.sciencedirect.com