CLONING AND CHARACTERIZATION OF THE ARABIDOPSIS THALIANA LUPEOL SYNTHASE GENE JENNIFER B. R. HERRERA,$ BONNIE BARTEL,% WILLIAM K. WILSON% and SEIICHI P. T. MATSUDA*$% $Department of Chemistry, Rice University, 6100 S. Main St., Houston, TX 77005, U.S.A.; %Department of Biochemistry and Cell Biology, Rice University, 6100 S. Main St., Houston, TX 77005, U.S.A. (Received 7 January 1998; in revised form 13 May 1998) Key Word IndexÐArabidopsis thaliana; Brassicaceae; cloning; molecular evolution; triter- penes; lupeol. AbstractÐA 2274 bp Arabidopsis thaliana cDNA was isolated that encodes a protein 57% identical to cycloartenol synthase from the same organism. The expressed recombinant protein encodes lupeol synthase, which converts oxidosqualene to the triterpene lupeol as the major product. Lupeol synthase is a multi- functional enzyme that forms other triterpene alcohols, including b-amyrin, as minor products. Sequence analysis suggests that lupeol synthase diverged from cycloartenol synthase after plants diverged from fungi and animals. This evolutionary order is the reason that fungi and animals do not make lupeol. # 1998 Elsevier Science Ltd. All rights reserved INTRODUCTION Thousands of nonsteroidal triterpenes have been identi®ed and the ubiquity of these compounds among higher plants implies that they are essential. However, the function of the nonsteroidal triter- penes remains unclear. The cyclization, rearrange- ment and deprotonation reactions mandated by the biogenetic isoprene rule are well-established [1±4], but it is unknown how triterpene synthases mediate the speci®c bond changes, or how these enzymes evolved. Obtaining triterpene synthase genes is cru- cial to resolving these problems in terpene biosyn- thesis. If triterpene synthase genes from a model plant were known, transgenic plants with modi®ed triterpene production could be made and their phe- notypes studied. Correlating catalytic activity with sequences from multiple triterpene synthases would allow identi®cation of candidate catalytic residues. Phylogenetic trees constructed with orthologous and paralogous triterpene synthases would illumi- nate the evolutionary relationships between these enzymes. With these goals in mind, we have under- taken a program to characterize genes encoding tri- terpene synthases in a model organism, the plant Arabidopsis thaliana. RESULTS AND DISCUSSION Although A. thaliana has not been reported to produce nonsteroidal triterpenes, DNA sequence databases contain clues that this plant may cyclize oxidosqualene to at least one compound in addition to cycloartenol. A search for potential oxidosqua- lene cyclases in the GenBank database of randomly sequenced partial cDNA fragments uncovered a short A. thaliana sequence [5] (accession No. T22249) encoding a 39-residue peptide 61% identi- cal to a portion of A. thaliana cycloartenol synthase (CAS1) [6]. This fragment was sequenced as part of the Arabidopsis Genome Project [5] and although it was clearly related to cycloartenol synthase, it seemed too far diverged to represent a second iso- zyme that produced cycloartenol. A full-length cDNA corresponding to this fragment was sought by using the fragment to probe an A. thaliana cDNA plasmid library [7]. One hybridizing plasmid (pJR1) with an insert the same size (2.5 kbp) as A. thaliana CAS1 was analyzed in detail. The pJR1 insert was sequenced (the sequence has been deposited in the GenBank database, accession No. U49919) and a yeast expression plasmid was constructed as described in the experimental section. Recombinant yeast transformed with this plasmid acquired the ability to convert exogenous oxido- squalene to the pentacyclic triterpene alcohol Phytochemistry Vol. 49, No. 7, pp. 1905±1911, 1998 # 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0031-9422/98/$ - see front matter PII: S0031-9422(98)00366-5 *Author to whom correspondence should be addressed. 1905