Surrogate splicing for functional analysis of sesquiterpene synthase genes

Abstract

A method for the recovery of full-length cDNAs from predicted terpene synthase genes containing introns is described. The approach utilizes Agrobacterium-mediated transient expression coupled with a reverse transcription-polydeoxyribonucleotide chain reaction assay to facilitate expression cloning of processed transcripts. Subsequent expression of intronless cDNAs in a suitable prokaryotic host provides for direct functional testing of the encoded gene product. The method was optimized by examining the expression of an intron-containing beta-glucuronidase gene agroinfiltrated into petunia (Petunia hybrida) leaves, and its utility was demonstrated by defining the function of two previously uncharacterized terpene synthases. A tobacco (Nicotiana tabacum) terpene synthase-like gene containing six predicted introns was characterized as having 5-epi-aristolochene synthase activity, while an Arabidopsis (Arabidopsis thaliana) gene previously annotated as a terpene synthase was shown to possess a novel sesquiterpene synthase activity for alpha-barbatene, thujopsene, and beta-chamigrene biosynthesis.

Scheme I.

A cartoon depiction of the steps involved in generating a fully processed, expressible cDNA clone from a putative gene containing multiple introns via surrogate splicing in an agroinfiltration based system. Lane 1 of the ethidium bromide-stained agarose gel contains molecular size standard; lane 2 contains control PCR products from the intact gene (plus introns); and lane 3 contains PCR products representing a processed transcript derived from expression of the intron-containing gene in the agroinfiltration system.

Figure 1.

Transient expression of GUS activity and recovery of a fully processed GUS transcript after agroinfiltration of an intron-containing GUS gene into petunia leaves. A, Time course for transient expression of GUS enzyme activity in agroinfiltrated petunia leaves. Detached leaves were infiltrated with a suspension of A. tumefaciens carrying the CsVMV-GUSI construct. Samples of the infiltrated leaf tissue were collected at each time point and assayed for GUS enzyme activity. Each point represents the average of three samples. B, RT-PCR recovery of the processed GUS gene transcript. Total RNA was isolated from leaf discs 4 d postagroinfiltration and used for RT-PCR assays. Lanes 1 and 8, molecular size standards; lane 2, RT-PCR assay without template RNA or PCR primers added; lane 3, RT-PCR assay without forward PCR primer added; lane 4, RT-PCR assay without reverse PCR primer added; lane 5, RT-PCR assay without reverse transcriptase added; lane 6, complete RT-PCR assay; lane 7, positive size control PCR assay using the intron-containing GUS gene construct as a template.

Figure 2.

Surrogate splicing and functional characterization of a putative tobacco terpene synthase gene. A, A cartoon depiction of a putative tobacco terpene synthase gene, g110, which contains six introns and has sequence similarity to previously isolated terpene synthases. B, RT-PCR recovery of a processed transcript for this terpene synthase after infiltration of A. tumefaciens carrying the CsVMV-g110 construct into detached petunia leaves. Total RNA was isolated from leaf discs 4 d postinfiltration and used for RT-PCR assays. Lane 1, RT-PCR assay without template RNA or PCR primers added; lanes 2 and 9, molecular size standards; lane 3, RT-PCR assay without forward PCR primer added; lane 4, RT-PCR assay without reverse PCR primer added; lane 5, RT-PCR assay without reverse transcriptase added; lane 6, positive size control PCR assay using the tobacco EAS4 cDNA as the template with appropriate primers; lane 7, complete RT-PCR assay for g110 transcript; lane 8, positive size control PCR assay using the intron-containing CsVMV-110 gene construct as template. C, The RT110 cDNA generated by surrogate splicing was expressed in E. coli, the encoded protein purified, incubated with the substrate FPP, and the pentane extractable reaction product(s) analyzed by gas chromatography. D, Mass spectrum of the product peak (8.76 min) in C. E, Mass spectrum of an authentic 5-epi-aristolochene standard.

Figure 3.

Surrogate splicing and functional characterization of a putative Arabidopsis terpene synthase gene. A, A cartoon depiction of At5g44630, a terpene synthase gene predicted from the DNA sequence of the Arabidopsis genome to contain six introns (Aubourg et al., 2002). B, RT-PCR recovery of a processed transcript for this terpene synthase after infiltration of A. tumefaciens carrying the CsVMV-At5g44630 construct into detached petunia leaves. Total RNA was isolated from leaf discs collected 4 to 7 d postinfiltration and used for RT-PCR assays. Lane 1, complete RT-PCR assay for the At5g44630 transcript; lane 2, positive size control PCR assay using the intron-containing At5g44630 gene construct as template; and lane M, molecular size standards. C, The At5g44630 cDNA generated by surrogate splicing was expressed in E. coli, and extracts incubated with [³H]GPP, [³H]FPP, and [³H]GGPP to obtain a general measure of mono-, sesqui-, and diterpene synthase activity. D, Synthase activity partially purified from E. coli extracts by anion-exchange chromatography was incubated with FPP and the organic extractable products examined by GC-MS. A total ion chromatogram is shown. The mass spectra for reaction product peaks 1, 2, and 3 are shown relative to published spectra for (E) α-barbatene, (F) thujopsene, and (G) β-chamigrene.

Scheme II.

A chemical rationalization for the major sesquiterpene products generated by the Arabidopsis (+)-α-barbatene synthase. Allylic rearrangement of the diphosphate moiety of all-trans farnesyl diphosphate, (E, E)-FPP, and formation of nerolidyl diphosphate, (3R)-NPP, allows for rotation about the single 2,3 bond. Reionization of the diphosphate group in a cisoid conformation of NPP and presumably concerted anti,endo S_N′ cyclization generates the (3R)-bisabolyl cation. Anti-Markovnikov π-cyclization onto the 10, 11 double bond creates the 5-membered B ring and results in centering of the secondary carbocation at C10 of the B ring. A 1,4-hydride shift from the A ring repositions the reactive carbocation back onto the A ring, which creates a key branch point for the diversion of reaction intermediates to either α-barbatene or thujopsene/β-chamigrene biosynthesis. A methylene migration would generate the chamigrenyl intermediate, which could undergo either a direct proton abstraction to form (+)-β-chamigrene, or a ring expansion followed by a homoallyl-cyclopropylcarbinyl cyclization and proton elimination to yield (+)-thujopsene. The alternative branch pathway for the cuprenyl cation consists of sequential methyl migrations to generate a bazzanenyl intermediate. Instead of deprotonation at C12 as observed for trichodiene biosynthesis (Cane, 1990), another π-cyclization event generates the tricyclic barbatenyl cation that undergoes a final endocyclic deprotonation giving rise to (+)-α-barbatene. Precedent for these predictions is provided by the extensive characterization of trichodiene synthase (Cane, 1990; Rynkiewicz et al., 2002) and in vivo labeling studies of β-barbatene biosynthesis by liverwort cell cultures (Nabeta et al., 1998; Warmers and König, 2000). The absolute configurations of the three sesquiterpene products are assumed to be the same as those established for the same compounds detected in volatile emissions from Arabidopsis flowers (Chen et al. 2003).