The increasing demand for decaffeinated coffee and tea has resulted from the occasional side effects of caffeine, which include palpitations, gastrointestinal disturbances, anxiety, tremor, increased blood pressure and insomnia1, 2. At present, the decaffeination process depends on supercritical fluid extraction with carbon dioxide to avoid introducing toxic residues from extraction solvents. However, this operation is expensive and, to discerning customers, flavours and aromas are lost.

There are some Coffea and Camellia species that produce low levels of caffeine, but these are not readily available for commercial use. Although it might be feasible to develop a breeding programme, it could take 20 years or so to establish and stabilize the desired traits3. The large-scale production of transgenic caffeine-deficient Camellia sinensis and Coffea arabica plants may thus be a more practical proposition4, but first some information is needed about the genes controlling key conversions in the biosynthesis of caffeine.

Caffeine is synthesized from purine nucleotides. The two final steps of the pathway, in which two methyl groups are added successively to 7-methylxanthine to produce theobromine and then caffeine, are catalysed by caffeine synthase, a bifunctional enzyme comprising two S-adenosylmethionine-dependent N-methyltransferase activities3. It has been hard to purify and isolate caffeine synthase and other enzymes of this pathway because they are extremely labile, but the amino-terminal sequence of caffeine synthase from young tea leaves has now been reported5.

We used the RACE (rapid amplification of complementary DNA ends) technique with degenerate gene-specific primers based on the amino-terminal sequence of caffeine synthase to obtain a 1.31-kilobase sequence of cDNA. The 5' untranslated sequence of the cDNA fragment was isolated by 5' RACE. The isolated cDNA, termed TCS1 (GenBank accession no. AB031280), consists of 1,438 base pairs and encodes a protein of 369 amino acids. The deduced amino-acid sequence of TCS1 shares a small amount of sequence similarity with other N-, S- and O-methyltransferases from plants and microorganisms, but considerably more with the salicylic acid O-methyltransferase6 (41.2%).

To determine whether our cDNA encoded an active caffeine synthase enzyme, we expressed TCS1 in Escherichia coli and incubated lysates of the bacterial cells with a variety of xanthine substrates in the presence of S-adenosylmethionine as methyl donor. We found that the substrate specificity of the recombinant enzyme was very similar to that of the native enzyme purified from young tea leaves (Table 1), with the recombinant enzyme mainly catalysing 3-N-methylation and 1-N-methylation of the purine ring of mono- and dimethylxanthines. There was no 7-N-methylation activity when xanthosine was the methyl acceptor. These results indicate that TCS1 encodes caffeine synthase.

The cloning of the caffeine synthase gene is an important advance towards the development of transgenic caffeine- deficient Camellia sinensis and Coffea arabica plants through antisense messenger RNA technology or by gene silencing. It is possible that the health benefits of tea7, 8, 10, 11, whose catechins and related polyphenols7-10 are thought to help protect against heart disease, may be enhanced without the potentially hypertensive effects of caffeine12, 13.

** Department of Biology, Faculty of Science, Ochanomizu University, Otsuka, Bunkyo-ku, Tokyo 112-8610, Japan
† Institute of Agricultural and Forest Engineering, University of Tsukuba, Ibaraki 305-8572 , Japan
‡ Plant Products and Human Nutrition Group, Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow G12 8QQ, UK

References

1. Chou, T. M. & Benowitz, N. L. Comp. Biochem. Physiol. 109C, 173-189 (1994).
2. Nurminen, M.-L., Niittynen, L., Korpela, R. & Vapaatalo, H. Eur. J. Clin. Nutr. 53, 831-839 (1999). Links
3. Ashihara, H. & Crozier, A. Adv. Bot. Res. 30, 117-205 (1999).
4. Mazzafera, P., Crozier, A. & Magalhes, A. C. Phytochemistry 30, 3913-3916 (1991).
5. Kato, M. et al. Plant Physiol. 120, 579-586 (1999). Links
6. Ross, J. R., Nam, K. H., D'Auria, J. C. & Pichersky, E. Arch. Biochem. Biophys. 367, 9-16 (1999). Links
7. Stensvold, I., Tverdal, A., Solvoll, K. & Foss, O. P. Prev. Med. 21, 546-553 (1992). Links
8. Hertog, M. G. L. et al. Lancet 342, 1007-1010 (1993). Links
9. Vinson, J. A. & Dabbagh, Y. F. FEBS Lett. 433, 44-46 (1998). Links
10. Leenen, R., Roodenburg, A. J. C., Tijburg, L. B. M. & Wiseman, S. A. Eur. J. Clin. Nutr. 54, 87-92 (2000). Links
11. Fujiki, H. J. Cancer Res. Clin. Oncol. 125, 589-597 (1999). Links
12. Lane, J. D. et al. Psychosom. Med. 60, 327-330 (1998). Links
13. James, J. E. Lancet 349, 279-281 (1997). Links

** NOTICE: In accordance with Title 17 U.S.C. Section 107, this material is distributed for research and educational purposes only. **