Identification and modification of universal traits
The gene pool of crop plants has diverged over 150 million years, the diversity arising from variations in the basic building blocks of genes determining specific plant processes. A series of important traits in valuable crops like cereals, sugarcane, soybeans, cotton, potato, woody trees, horticultural and ornamental crops, can therefore be addressed from a general perspective using gene function analysis from model plants. For example, the gene for the LEAFY mutant phenotype is a single gene determining flowering initiation in Arabidopsis. But it can also be applied to manipulate the early flower initiation in plants as diverse as aspen trees and rice. Although such trait modification still needs to be fine-tuned, the first steps come from an understanding of the basic biology of the model plant.
The next step is to evaluate the trait in its genomic background, by identifying which other genes are required for its proper expression. The identification of a few key genes from Arabidopsis for generic traits like resistance to the abiotic stresses of cold, salinity and drought could therefore lead to their application in other crop plants that are more difficult to study because of their complex genomes and growth habits. Recently, the isolation of genes determining plant height from Arabidopsis led to the identification of ‘orthologous’ (similar and same function) ‘dwarf’ genes in rice and other cereals. Most surprisingly these were found to be the dwarf genes that were also introduced into modern varieties by conventional breeding in the Green Revolution. This is an example of the conservation of important traits throughout the plant kingdom, but also illustrates the potential of gene function discovery in model plants for application in a wide variety of crop plants.

Gene discovery race
Different approaches for discovering gene sequences and functions can be pursued, each with specific advantages and shortcomings:
• The expressed genes of a plant can be catalogued by sequencing expressed sequence tags (ESTs) or complementary DNA (cDNA) (see glossary). There are about 130,000 plant ESTs available in public databases from 19 plant species like Arabidopsis, rice, tomato, maize, soybean, cotton, and loblolly pine, which offers an efficient method for gene discovery in these plants. A comparison of EST databases from different plants, tissues and conditions reveals the diversity in coding sequences between plants. At the same time, however, it provides a global perspective of the similarities in genes for specific processes, such as ripening, or conditions, such as the induction of pathogens. A sequence similarity analysis using bioinformatics tools permits the assignment of probable gene function and the identification of genes similar between species. However, elucidating their exact function still requires experimental approaches.
• Systematic sequencing of the entire genome. It is evident that not all genes are transcribed in abundance so that they can be detected and represented in ESTs. It is likely that about 50 per cent of genes can be determined only by extensive, systematic sequencing of the entire genome.In the case of Arabidopsis, the complete sequence of the whole genome of the ecotype ‘Columbia’ will be determined by the year 2000. This information will reveal the rest of the genes, their structure and organization in the genome and will be made available through public databases.
• Shotgun sequencing. As an alternative, a strategy termed ‘Arabidopsis genome sampling’ was employed by Cereon Genomics (USA), a collaboration of the US companies Monsanto and Millenium Pharmaceuticals. They ‘shotgun-sequenced’ (see glossary) the genome of Arabidopsis (ecotype ‘Landsberg’) and additionally identified 10,000 novel ESTs. This rapidly-created proprietary sequence database will lead to the discovery of new genes even from gene families that have been most intensively characterized by public research. It brings the total of known unique Arabidopsis ESTs to 25,000, which is close to the expected number of genes. Most importantly, the sequencing of this second Arabidopsis genotype/ecotype directly yields to the detection of single nucleotide polymorphisms (SNPs). SNPs are point mutations in the DNA (see glossary) and can be employed as markers to accurately map known mutants, many of which are of commercial importance.

Commercially important crops
While Arabidopsis is the model for the dicot plant families, rice has been selected to be the model plant for the monocot families for several reasons:

•	Firstly, the genome of rice is relatively small, only three times larger than that of Arabidopsis.
•	Secondly, the genome organization in cereals is highly conservative: irrespective of the plant species, genes are lined up in the same order on the chromosomes. This phenomenon is called ‘synteny’. The identification of sequences and functional genes in rice will therefore help isolate the corresponding genes in the more complex cereal genomes of, for instance, maize and wheat, which are respectively five and 30 times longer.
•	Thirdly, rice feeds a quarter of the human population; its annual market value is considered to be US$ 45 billion in China alone.

It is therefore not surprising that deciphering the rice genome has attracted the interest of private companies. Earlier in 1999, Celera Genomics, a US genomics enterprise headed by Craig Venter, offered to sequence the whole genome of the rice plant in six weeks by using the shotgun sequencing technique. If contracted, Celera promised to do this job at a fraction of the cost entailed in public rice sequencing. At present, no interested parties from industry were offering the budget, which is still considerable. Yet Celera’s offer made scientists and politicians from all over the world sit up in alarm, concerned over issues of control of genetic information and freedom to use it. Venter had earlier revealed similar plans for sequencing the human genome (see also the article by Lehmann and Lorch). In both cases, one positive consequence of Celera’s competitive bid was that it pushed the efforts for public research. Led by Japan, the deadline of the International Rice Genome Sequencing Project (IRGSP) has been brought forward by almost four years to 2004. For this, Japan announced that it would inject extra funding; in 2000, the country’s total annual rice genome research budget is to reach US$ 67 million, a threefold increase from 1999. Of this, about US$ 27 million will be spent on sequencing alone. Furthermore, grants totalling US$ 12.3 million have been awarded to three US governmental agencies for the US component of the project. As DNA sequencing technology has become more accessible and cheaper, other industrialized countries such as France and Canada, as well as a number of Asian countries like China and Korea, have taken initiatives in systematically sequencing rice chromosomes.
Another crop that is targeted by the sequencing activities of private companies is maize. Sequencing maize completely would be as expensive as sequencing the human genome. However, while only about 35,000 ESTs of the maize genome are in public databases, more than 100,000 ESTs have already been sequenced by US companies like Pioneer HiBred and its new proprietor, DuPont. Similar proprietary EST databases exist for other commercial crops for which gene discovery is being exploited for product development. To balance the efforts of private companies, the US National Science Foundation (NSF) has launched a plant genome research initiative. This initiative has given grants to a number of projects to sequence and publicly release ESTs from maize (50,000 ESTs), soybean (30,000) and tomato (90,000), and to develop the infrastructure and technology for functional genomics in many plants.

Private-public partnerships for genomics
In the light of the genomics revolution, the emphasis of companies is changing from chemicals, pharmaceuticals and agrochemicals towards biotechnology and an integrated life science concept (see also the article by Bijman). One important aspect of this concept is strategic alliances between the life science giants and academic institutions all around the world. For instance, the University of Berkeley, California (USA) received US$ 50 million from Novartis for first rights to their genomic research. In the UK, the John Innes Centre and Sainsbury Laboratory are public research institutions that have established independent long-term research alliances with both DuPont and Zeneca (UK) in the area of plant genomics. Similar investment alliances in plant genomics have been formed between the Max-Planck-Institut für Züchtungsforschung (Germany) by a company consortium including AgrEvo (Germany), and between the Institute of Molecular Agrobiology (Singapore) and Rhône-Poulenc (France). These alliances of large public research organizations being funded by the private sector to carry out genomic research in a large-scale and efficient way will certainly channel discoveries into applications and advance biotechnology in general. On the other hand, as the infrastructure to carry out this research will not become publicly available, smaller research organizations will be left out.

Applications in the developing world
The increase that is needed in the world food supply will have to rely on increases in economically viable food production. This may be possible by reducing pre- and post-harvest losses due to pests and pathogens, stabilizing yields in poor soils, marginal and changing environments. Towards these ends, genetic improvements are possible with the use of modern genomics tools, assisting crop improvement through transgenic plants as well as marker-assisted breeding. The availability of the relevant knowledge and technology to the developing world will be the key factor in this development.
The mechanism of technology transfer should involve training and a build-up of biotechnological facilities in the developing countries as well as material transfer. Successful programmes in this regard are the rice and cassava programmes of the Rockefeller Foundation (USA), the programme of the Biotechnology Action Council of the United Nations Educational, Scientific and Cultural Organization (UNESCO) and the International Cooperation (INCO) programme of the European Union. Non-profit organizations like the International Service for the Acquisition of Agrobiotech Applications (ISAAA) are attempting to play a role in technology transfer from the public and private institutions of the industrialized world to developing countries. One of the favourable formulas is the donation of technology on a royalty-free basis for production aimed at internal markets of developing countries, with royalties only paid where export is possible. Promising examples are Monsanto’s donation of virus-resistant potatoes for the Mexican market, or the Rockefeller Foundation funded development of ‘golden rice’ (for Vitamin A) genotypes for rice breeding. Yet it remains to be seen in how far such attempts really meet the demands of the end-users in developing countries or if the transfer of technology is supply-driven.
Genomics is the logical evolution of the electronic age applying developments in informatics for developments in biology and it has a revolutionizing impact on agricultural research.
Andy Pereira

Department of Molecular Biology, DLO-Centre for Plant Breeding and Reproduction Research (CPRO-DLO), P.O. Box 16, 6700 AA, Wageningen, the Netherlands. Phone (+31) 317 47 70 01; Fax (+31) 317 41 80 94; E-mail A.Pereira@cpro.dlo.nl

Sources
Agris, C.H. (1999), "Patenting plants: What to claim." Nature Biotechnology 17, pp. 717-718.

Herrera-Estrella, L. (1999), "Transgenic plants for tropical regions: Some considerations about their development and their transfer to the small farmer." Proceedings of the National Academy of Science 96, pp. 5978-5981.

Kishore, G.M. and Shewmaker, C. (1999) Biotechnology: Enhancing human nutrition in developing and developed worlds. Proceedings of the National Academy of Science 96, pp. 5968-5972.

Nap, J.P. and Pereira, A. (1999), "From high throughput genomics to useful transgenic crops." Molecular Breeding 5, pp. 481-483.

Somerville, C. and Somerville, S. (1999), "Plant Functional Genomics." Science 285, pp. 380-383.