Previous PageTable Of Contents

Elucidating biosynthesis of the rice allelochemical/phytoalexin momilactone B

Meimei Xu1, Sladjana Prisic1, P. Ross Wilderman1, Yinghua Jin2, Robert M. Coates2 and Reuben J. Peters1

1 Department of Biochemistry, Biophysics, & Molecular Biology, Iowa State University, Ames, IA 50011, USA
Department of Chemistry, University of Illinois, Urbana, IL 61801, USA


Rice (Oryza sativa) produces momilactone B as an allelopathic agent in its roots and as a phytoalexin in its leaves. Notably, momilactones are diterpenoids whose metabolism resembles that of the structurally related gibberellic acid (GA) phytohormones. In particular, biosynthesis of all such labdane-related diterpenoids is uniquely initiated by the consecutive action of two distinct classes of diterpene synthases/cyclases. However, rather than the ent-copalyl diphosphate (ent-CPP) and tetracyclic kaurene intermediates of GA biosynthesis, momilactones are produced via stereochemically distinct syn-CPP and tricylic pimaradiene intermediates. Using a functional genomics based approach we have identified and cloned a number of putative CPP (class II) and kaurene-like (class I) synthases from rice. Building on our biochemical studies of the GA biosynthetic enzymes from Arabidopsis, we have functionally expressed and characterized many of the rice labdane-related diterpene synthases; specifically including those required for momilactone metabolism. These two enzymatic genes not only catalyse the corresponding cyclization reactions, but also exhibit the expected root constitutive and leaf inducible expression pattern. Intriguingly, these two genes are also clustered together in the rice genome, suggesting evolutionary coupling through physical linkage and resulting co-segregation. Finally, we note that the expression patterns of many of the rice labdane-related diterpene synthases indicate that the corresponding natural products, which were previously identified as phytoalexins in rice leaves, are also constitutively secreted from roots into the rhizosphere, where their anti-microbial activity may be useful in root establishment and function.

Media summary

Identification of the terpene synthases/cyclases involved in biosynthesis of the rice allelochemical/antibiotic momilactone B, will enable metabolic engineering and biochemical/green synthesis.

Key Words

Terpene synthase, rice, phytoalexin, biosynthesis, natural products, labdane-related diterpene


Plants produce a huge number of natural products, small organic compounds that often have important ecological roles and mediate interactions between the producing plant and other organisms. Particularly notable among plant natural products are the labdane-related diterpenoids, with ~7,000 such compounds known (Buckingham 2002). The large numbers of this sub-family of natural product presumably arises from the absolute requirement for gibberellin (GA) phytohormone biosynthesis in all flowering plants. In fact, GAs are the seminal labdane-related diterpenoids, as it was study of their biosynthesis that demonstrated the intermediacy of labdadienyl/copalyl diphosphate (CPP) en route to the tetracyclic kaurene intermediate (Shechter and West 1969). Thus, duplication of the ubiquitous GA biosynthetic genes provides a ready source for derivation of alternative metabolic pathways. In particular, the committed step in labdane-related diterpenoid metabolism is catalysed by class II diterpene cyclases, which produce the characteristic bicyclic structure, such as that found in the ent-CPP intermediate of GA biosynthesis, from the acyclic universal diterpenoid precursor (E,E,E)-geranylgeranyl diphosphate (GGPP). Production of the final cyclized structure is then carried out a class I diterpene synthase, which typically produces a hydrocarbon olefin that generally must be further modified in order to exhibit biological activity (e.g. oxygenated by cytochromes P450).

Rice has been shown to produce a number of labdane-related diterpenoid beyond the universal GAs. These include momilactones A and B, oryzalexins A to F, oryzalexin S, and phytocassanes A to E, all of which act as phytoalexins in leaves (Cartwright et al. 1981; Akatsuka et al. 1985; Sekido et al. 1986; Kodama et al. 1992; Kato et al. 1993; Kato et al. 1994; Koga et al. 1995; Koga et al. 1997). In addition, momilactones A and B were originally identified as dormancy factors from rice seed husks (Kato et al. 1973). More recently, momilactone B has been found to be constitutively secreted from the roots of rice seedlings and act as an allelochemical (Kato-Noguchi et al. 2002; Kato-Noguchi and Ino 2003). Given the extensive genomic and expressed sequence information available for rice (Goff et al. 2002; Yu et al. 2002; Kikuchi et al. 2003), this plant has become a model system for investigating labdane-related diterpenoid biosynthesis. Accordingly, we (along with other groups) have been engaged in functional identification of the corresponding biosynthetic genes, particularly the relevant labdane-related diterpene synthases/cyclases.


Plant material

The rice plants (Oryza sativa L. subsp. japonica cv Nipponbare) used here have been previously described (Xu et al. 2004). Briefly, four week old greenhouse grown plants were dissected to obtain the roots, which were used directly, or leaves, which underwent either mock treatment or UV irradiation [this has previously been shown to induce phytoalexin biosynthesis (Kodama et al. 1988)]. These tissues were frozen in liquid N2 and total RNA extracted using Concert Plant Reagent (Invitrogen, Carlsbad, CA, USA) and stored at -80C.


Putative labdane-related diterpene synthases/cyclases were identified by BLAST queries of GenBank (, The Institute for Genomic Research (TIGR;, and the rice cDNA database at KOME ( with the nucleotide sequences of ent-CPP and kaurene synthases from Arabidopsis [Genbank entries U11034 and AF034774, respectively; (Sun and Kamiya 1994; Yamaguchi et al. 1998)]. Fragments corresponding to the rice genes were obtained via RT-PCR with gene-specific primers, cloned into pCR-Zero-Blunt (Invitrogen), and verified by complete sequencing. In some cases, full-length cDNA could be directly amplified in RT-PCR reactions using primers based on the predicted sequence. More often, it was necessary to use rapid amplification of cDNA ends (RACE) to obtain the correct 5' and/or 3' sequence. This then enabled amplification of the complete cDNA by RT-PCR. The full-length gene products were cloned by directional topoisomerization into pENTR/SD/D-TOPO (Invitrogen) and verified by complete sequencing.

Recombinant expression

The genes were transferred by directional recombination to the T7 based expression vectors pDEST14 and pDEST15 (Gateway system; Invitrogen) for expression either alone or fused to glutathione-S-transferase (GST), respectively. Recombinant expression was carried out with the BL21-derived C41 strain of E. coli, as previously described (Xu et al. 2004). Briefly, recombinant cells were grown to mid-log phase at 37C, then shifted to 16C for 1-2 hours prior to induction with 1 mM IPTG, and the expression allowed to proceed overnight (16-20 hours).

Biochemical analysis

Recombinant proteins were partially purified from clarified cell lysates either by batch binding to ceramic hydroxyapatite type II beads (Bio-Rad, Hercules, CA, USA) as previously described (Xu et al. 2004), or to glutathione agarose beads for the GST fusion proteins. Reactions were carried out in assay buffer (50 mM HEPES, pH 7.2, 100 mM KCl, 5 mM MgCl2, 5% glycerol, and 5 mM DTT) with approximately 50 M of substrate (GGPP, ent-, or syn-CPP) and 20 L of the enzymatic preparation in a total volume of 0.2 mL. After 4 hours at 30C reactions were extracted 3 times with 0.5 mL hexanes for analysis of class I enzymatic products. To analyse class II enzymatic products it was necessary to remove the diphosphate moiety by dephosphorylation with calf intestinal phosphorylase (New England Biolabs, Beverly, MA, USA) prior to extraction. The resulting organic extracts were then concentrated to 0.1 mL for gas chromatography-mass spectrometry (GC-MS) analysis using an Agilent (Palo Alto, CA, USA) 6890N GC instrument with 5973N mass selective detector. Samples (5 L) were injected onto an HP-1MS column in the splitless mode. After a 3 min. isothermal hold at 40C, the oven temperature was increased at 20C/min. to 300C, where it was held for an additional 4 min. MS data from 40 to 500 m/z was collected during the temperature ramp and high isothermal hold. Enzymatic products were identified by comparison to authentic standards, which were those previously described (Mohan et al. 1996).

Expression analysis

Transcriptional levels for each of the genes described herein was measured in response to UV-irradiation of detached leaves, as well as from untreated roots, as previously described (Xu et al. 2004). In particular, leaf material was collected, by freezing in liquid N2, at the indicated times after irradiation. This enabled RNA isolation and RT-PCR, using the afore-mentioned verified gene specific primers, to follow the relative levels of individual mRNA, or 18S rRNA (using QuantumRNA primers; Ambion, Austin, TX, USA).


Prompted by our interest in labdane-related diterpene synthases as model systems for biochemical analysis and metabolic engineering, we have undertaken identification of the catalytic activity of all such enzymatic genes from rice. Of particular interest were those involved in momilactone biosynthesis, due to both the intriguing range of bioactivities (i.e. momilactone B acts as both a phytoalexin and allelochemical) and novel stereochemical intermediate (i.e. syn- rather than the more common ent-CPP). Towards this end we searched the extensive sequence data available for rice to find putative labdane-related diterpene synthases/cyclases, as defined by homology to the ent-CPP and kaurene synthases from Arabidopsis. Clear evidence was found for three class II diterpene cyclases. Less clear was the number of class I diterpene synthases, we simply note that at least some sequence information was available for nine such genes, which is consistent with findings reported by other groups (Cho et al. 2004; Sakamoto et al. 2004). Based on this sequence data we were able to clone all three class II diterpene cyclases and three of the class I diterpene synthases.

Due to the plastidial origin of diterpenoid natural products, diterpene synthases/cyclases are encoded with N-terminal targeting sequences that are removed from the mature enzymes. Because of our interest in cyclization of labdane-related diterpenoids, we have previously developed protocols for truncation and recombinant expression of the Arabidopsis ent-CPP and kaurene synthases (unpublished results). Thus, we were able to functionally express all of the putative rice labdane-related diterpene synthases. Fortuitously, the first class II diterpene cyclase we cloned proved to produce the stereochemically novel syn-CPP intermediate required for momilactone biosynthesis (OsCPS4syn), and we demonstrated a root constitutive and leaf inducible expression pattern, as we have previously reported (Xu et al. 2004). With further work we were able to clone the two remaining class II diterpene cyclases, which both produce ent-CPP (OsCPS1ent and OsCPS2ent), as we have previously reported (Prisic et al. 2004). Similar results for all three rice class II diterpene cyclases also were reported by another group (Otomo et al. 2004b). Notably, previous work by (Sakamoto et al. 2004) had demonstrated that only one of the ent-CPP synthases is involved in GA biosynthesis (OsCPS1ent). Our results demonstrated that only transcription of the other (OsCPS2ent) was inducible in leaves, presumably for production of ent-CPP derived phytoalexins. In addition, OsCPS2ent is constitutively expressed in roots (Figure 1), indicating that the corresponding natural products may also be constitutively produced in rice roots. Finally, we note that the two secondary metabolism specific class II diterpene cyclases (OsCPS2ent and OsCPS4syn) are more closely related to each other (~53% amino acid identity), despite their stereochemically distinct products, than either is to the GA specific gene (OsCPS1ent). In fact, OsCPS1 is more closely related to the maize GA specific ent-CPP synthase An1 (Bensen et al. 1995), suggesting derivation of labdane-related diterpenoid secondary metabolism from GA biosynthesis prior to the early split between the separate lineages within the cereal/grass family (Poaceae) that resulted in modern rice and maize.

Figure 1. Comparison of OsCPS2ent mRNA expression levels in mature roots (R) or leaves (L). Note that transcript levels are dramatically increased by ~50-fold in leaves upon UV-irradiation, which has been shown to stimulate production of the corresponding phytoalexins (Kodama et al. 1988), as we have previously demonstrated (Prisic et al. 2004).

While the evolutionary relationship between the putative class I labdane-related diterpene synthases was not as clear, these also form a relatively closely related family (45-85% aa identity) that should provide an ideal system for examining structure-function relationships. Again, we were fortunate in that the first such enzyme we cloned (OsDTS2) proved stereospecific in its utilization of syn-CPP, rather than ent-CPP, and seems to be involved in momilactone biosynthesis, producing the corresponding pimara-7,15-diene olefin intermediate. We were again able to demonstrate the expected root constitutive and leaf inducible expression pattern, also as we have previously reported (Wilderman et al. 2004). Prior to our work, (Cho et al. 2004) had published functional identification of a ent-CPP specific cassadiene synthase that is also leaf inducible (OsDTC1). Thus, we noted that OsDTS2 and OsCPS4syn are clustered together on chromosome 4, while OsDTC1 and OsCPS2ent are also clustered together on chromosome 2, suggesting evolutionary coupling through physical linkage and resulting co-segregation of labdane-related diterpene synthases that act to produce and further process stereochemically common intermediates (i.e. syn- or ent-CPP). Accordingly, we have speculated that the additional class I labdane-related diterpene synthases also clustered with OsDTC1 and OsCPS2ent will also prove to be ent-CPP specific (Wilderman et al. 2004). Further work by other groups has also defined three other class I labdane-related diterpene synthases. In particular, the kaurene synthase (OsKS1) required for GA biosynthesis was defined by mutational phenotype (Sakamoto et al. 2004; Margis-Pinheiro et al. 2005), and the expected ent-sandaracopimaradiene (OsKS10) and syn-stemarene (OsDTC2) synthases by cloning and functional characterization (Nemoto et al. 2004; Otomo et al. 2004a). We already had clones corresponding to the cassadiene and stemarene synthases, and were also able to clone the sandaracopimaradiene synthase ourselves, while the kaurene synthase was obtained as a kind gift from N. Upadhyaya (CSIRO, Australia). We are in the process of functionally expressing and characterizing these ourselves. Nevertheless, from our own and others work, all of the expected labdane-related diterpene synthases have been cloned and functionally characterized (Figure 2). Notably, the expression patterns for many of these class I labdane-related diterpene synthases, as reported by Sakamoto et al. (2004) and Margis-Pinheiro et al. (2005), includes constitutively high levels of transcription in roots. This further strengthens our suggestion that many of the corresponding natural products will be constitutively produced and secreted from the roots.

Figure 2. Scheme for known cyclization reactions in rice labdane-related diterpenoid biosynthesis. All the corresponding enzymes have been cloned and functionally characterized, as indicated by the assigned gene names. Also indicated are the derived classes of natural products.


We have reported functional identification of the labdane-related diterpene synthases/cyclases that are involved in biosynthesis of the rice allelochemical/phytoalexin momilactone B. These enzymatic genes catalyse cyclization reactions to produce the corresponding syn-CPP and pimaradiene intermediates and also exhibit the expected root constitutive and leaf inducible expression pattern, as we have previously reported (Xu et al., 2004; Wilderman et al., 2004). Intriguingly, rice contains a number of other labdane-related diterpenoids that have been identified as phytoalexins in the leaves, but the expression of the corresponding synthases/cyclases indicates that these natural products may also be constitutively produced and secreted from the roots. We speculate here that the known anti-microbial activity of these compounds may offer a competitive advantage for root establishment, as well as potentially increasing nutrient availability.


Akatsuka T, Kodama O, Sekido H, Kono Y, Takeuchi S (1985). Novel Phytoalexins (Oryzalexins A, B, and C) Isolated from Rice Blast Leaves Infected with Pyricularia oryzae. Part I: Isolation, Characterization and Biological Activities of Oryzalexins. Agricultural and Biological Chemistry 49, 1689-1694.

Bensen RJ et al. (1995). Cloning and characterization of the maize An1 gene. Plant Cell 7, 75-84.

Buckingham J (ed) (2002) Dictionary of Natural Products (on-line web edition). In Chemical Databases Online. (Chapman & Hall/CRC Press).

Cartwright DW, Langcake P, Pryce RJ, Leworthy DP, Ride JP (1981). Isolation and characterization of two phytoalexins from rice as momilactones A and B. Phytochemistry 20, 535-537.

Cho E-M et al. (2004). Molecular cloning and characterization of a cDNA encoding ent-cassa-12,15-diene synthase, a putative diterpenoid phytoalexin biosynthetic enzyme, from suspension-cultured rice cells treated with a chitin elicitor. Plant Journal 37, 1-8.

Goff SA et al. (2002). A draft sequence of the rice genome (Oryza sativa L. ssp. japonica). Science 296, 92-100.

Kato H, Kodama O, Akatsuka T (1993). Oryzalexin E, a diterpene phytoalexin from UV-irradiated rice leaves. Phytochemistry 33, 79-81.

Kato H, Kodama O, Akatsuka T (1994). Oryzalexin F, a diterpene phytoalexin from UV-irradiated rice leaves. Phytochemistry 36, 299-301.

Kato T et al. (1973). Momilactones, growth inhibitors from rice, Oryza sativa L. Tetrahedron Letters 14, 3861-3864.

Kato-Noguchi H, Ino T (2003). Rice seedlings release momilactone B into the environment. Phytochemistry 63, 551-554.

Kato-Noguchi H, Ino T, Sata N, Yamamura S (2002). Isolation and identification of a potent allelopathic substance in rice root exudates. Physiologia Plantarum 115, 401-405.

Kikuchi S et al. (2003). Collection, Mapping, and Annotation of Over 28,000 cDNA Clones from japonica Rice. Science 301, 376-379.

Kodama O, Li WX, Tamogami S, Akatsuka T (1992). Oryzalexin-S, a novel stemarane-type diterpene rice phytoalexin. Bioscience, Biotechnology, and Biochemistry 56, 1002-1003.

Kodama O, Suzuki T, Miyakawa J, Akatsuka T (1988). Ultraviolet-induced Accumulation of Phytoalexins in Rice Leaves. Agricultural and Biological Chemistry 52, 2469-2473.

Koga J, Ogawa N, Yamauchi T, Kikuchi N, Ogasawara N, Shimura M (1997). Functional moiety for the antifungal activity of phytocassane E, a diterpene phytoalexin from rice. Phytochemistry 44, 249-253.

Koga J, Shimura M, Oshima K, Ogawa N, Yamauchi T, Ogasawara N (1995). Phytocassanes A, B, C, and D, Novel Diterpene Phytoalexins from Rice, Oryza sativa L. Tetrahedron 51, 7907-7918.

Margis-Pinheiro M, Zhou X-R, Zhu Q-H, Dennis ES, Upadhyaya NM (2005). Isolation and Characterization of a Ds-Tagged Rice (Oryza sativa L.) GA Responsive Dwarf Mutant Defective in An Early Step of the Gibberellin Biosynthesis Pathway. Plant Cell Reports 23, 819-833.

Mohan RS et al. (1996). Biosynthesis of cyclic diterpene hydrocarbons in rice cell suspensions: conversion of 9,10-syn-labda-8(17),13-dienyl diphosphate to 9-pimara-7,15-diene and stemar-13-ene. Archives of Biochemistry and Biophysics 330, 33-47.

Nemoto T et al. (2004). Stemar-13-ene synthase, a diterpene cyclase involved in the biosynthesis of the phytoalexin oryzalexin S in rice. FEBS Letters 571, 182-186.

Otomo K et al. (2004a). Diterpene Cyclases Responsible for the Biosynthesis of Phytoalexins, Momilactones A, B, and Oryzalexins A-F in Rice. Bioscience, Biotechnology, and Biochemistry 68, 2001-2006.

Otomo K et al. (2004b). Biological functions of ent- and syn-copalyl diphosphate synthases in rice: key enzymes for the branch point of gibberellin and phytoalexin biosynthesis. Plant Journal 39, 886-893.

Prisic S, Xu M, Wilderman PR, Peters RJ (2004). Rice contains disparate ent-copalyl diphosphate synthases with distinct metabolic functions. Plant Physiology 136, 4228-4236.

Sakamoto T et al. (2004). An Overview of Gibberellin Metabolism Enzyme Genes and Their Related Mutants in Rice. Plant Physiology 134, 1642-1653.

Sekido H et al. (1986). Oryzalexin D (3,7-Dihydroxy-(+)-sandaracopimaradiene), a New Phytoalexin Isolated from Blast-infected Rice Leaves. Journal of Pesticide Science 11, 369-372.

Shechter I, West CA (1969). Biosynthesis of gibberellins. IV. Biosynthesis of cyclic diterpenes from trans-geranylgeranyl pyrophosphate. Journal of Biological Chemistry 244, 3200-3209.

Sun T-P, Kamiya Y (1994). The Arabidopsis GA1 locus encodes the cyclase ent-kaurene synthetase A of gibberellin biosynthesis. Plant Cell 6, 1509-1518.

Wilderman PR, Xu M, Jin Y, Coates RM, Peters RJ (2004). Identification of syn-pimara-7,15-diene synthase reveals functional clustering of terpene synthases involved in rice phytoalexin/allelochemical biosynthesis. Plant Physiology 135, 2098-2105.

Xu M, Hillwig ML, Prisic S, Coates RM, Peters RJ (2004). Functional identification of rice syn-copalyl diphosphate synthase and its role in initiating biosynthesis of diterpenoid phytoalexin/allelopathic natural products. Plant Journal 39, 309-318.

Yamaguchi S, Sun T, Kawaide H, Kamiya Y (1998). The GA2 Locus of Arabidopsis thaliana Encodes ent-Kaurene Synthase of Gibberellin Biosynthesis. Plant Physiology 116, 1271-1278.

Yu J et al. (2002). A draft sequence of the rice genome (Oryza sativa L. ssp. indica). Science 296, 79-92.

Previous PageTop Of Page