Biochemical conservation and evolution of germacrene A oxidase in asteraceae

Abstract

Sesquiterpene lactones are characteristic natural products in Asteraceae, which constitutes approximately 8% of all plant species. Despite their physiological and pharmaceutical importance, the biochemistry and evolution of sesquiterpene lactones remain unexplored. Here we show that germacrene A oxidase (GAO), evolutionarily conserved in all major subfamilies of Asteraceae, catalyzes three consecutive oxidations of germacrene A to yield germacrene A acid. Furthermore, it is also capable of oxidizing non-natural substrate amorphadiene. Co-expression of lettuce GAO with germacrene synthase in engineered yeast synthesized aberrant products, costic acids and ilicic acid, in an acidic condition. However, cultivation in a neutral condition allowed the de novo synthesis of a single novel compound that was identified as germacrene A acid by gas and liquid chromatography and NMR analyses. To trace the evolutionary lineage of GAO in Asteraceae, homologous genes were further isolated from the representative species of three major subfamilies of Asteraceae (sunflower, chicory, and costus from Asteroideae, Cichorioideae, and Carduoideae, respectively) and also from the phylogenetically basal species, Barnadesia spinosa, from Barnadesioideae. The recombinant GAOs from these genes clearly showed germacrene A oxidase activities, suggesting that GAO activity is widely conserved in Asteraceae including the basal lineage. All GAOs could catalyze the three-step oxidation of non-natural substrate amorphadiene to artemisinic acid, whereas amorphadiene oxidase diverged from GAO displayed negligible activity for germacrene A oxidation. The observed amorphadiene oxidase activity in GAOs suggests that the catalytic plasticity is embedded in ancestral GAO enzymes that may contribute to the chemical and catalytic diversity in nature.

FIGURE 1.

Sesquiterpene lactone biosynthetic pathways in Asteraceae. Left, the proposed biosynthetic pathway of general sesquiterpene lactones in Asteraceae is shown. Right, the artemisinic acid biosynthetic pathway in A. annua is shown. FPP, farnesyl diphosphate.

FIGURE 2.

Analyses of the metabolites de novo synthesized from transgenic yeast. A, GC-MS chromatographs are shown for the sesquiterpenoids from yeast transformed with the indicated genes. Lines a and b are negative controls, and line c displays the metabolites unique to the yeast transformed with three genes (GAS, GAO, and CPR). Compounds 2 and 3 were not separated by DB-5 MS column but were clearly separated by the chiral column (Cyclodex-B column) as shown in the inset. B, the mass fragmentation patterns of compound 1 to 4 are given. C, proposed acid-induced rearrangements of germacrene A acid to α-, β-, and γ-costic acid and additional modification of costic acids to ilicic acid in yeast culture are shown. The speculated structure of peak 2 as α-costic acid was also given with a question mark. D, cope rearrangement of germacrene A acid to elemenic acid by heat is shown.

FIGURE 3.

GC-MS and LC-MS analyses of sesquiterpenoids from yeast expressing GAS, GAO, and CPR in buffered neutral culture. A, GC-MS chromatographs of the sesquiterpenoid products are shown. Inlet temperatures and chemical treatment are shown in front of the chromatograms. TCA, trichloroacetic acid; bracket, α-, β-, and γ-costic acids; asterisk, germacrene A acid; arrow, heat-induced rearranged product of germacrene A acid. B, shown are LC-MS ion traces from negative ions of m/z 171, 233, and 251. m/z 171 corresponds to the internal standard (IS), decanoic acid, with a retention time of 7.68 min; m/z 233 corresponds to germacrene A acid (*), with a retention time of 8.41 min and α-, β-, γ-costic acids (in the bracket) with a retention time of 9.02 min and 9.17 min (α- and β-costic acids co-migrate at 9.02 min); m/z 251 corresponds to ilicic acid (triangle), with retention time of 5.38 min. C, LC-MS chromatographs (negative m/z 233) of in vitro enzyme assay products are shown.

FIGURE 4.

Phylogenetic and sequence analyses of GAOs from various Asteraceae plants. A, the Asteraceae phylogeny simplified from the figure by Panero and Funk (9) is shown. Asterisks indicate the four subfamilies where GAOs were isolated, and the parentheses indicate specific species names. B, the phylogenetic tree of AMO/GAO in Asteraceae is shown. HPO and EAH used as outgroups are Hyoscyamus muticus premnaspirodiene oxygenase and tobacco 5-epi-aristolochene dihydroxylase, respectively. Bootstrap values were shown in percentage values from 1000 replicates. The bracket indicates the GAO clade that is clearly distinguished from BsGAO and AMO. C, alignment of deduced amino acids from GAOs and AMO is shown. Amino acid sequences were obtained from cDNAs deposited at the NCBI. AMO, amorphadiene oxidase from A. annua (DQ268763 or DQ315671); LsGAO, germacrene A oxidase from L. sativa (GU198171) or from C. intybus (Ci; GU256644), S. lappa (Sl; GU256645), H. annuus (Ha; GU256646), and B. spinosa (Bs; GU256647). Stars indicate the residues conserved in GAO but different in AMO. Circled marks indicate the residues unique in AMO but not conserved in GAO. The alignment is shaded to a 50% consensus. Dark and light shading indicate identical and similar residues, respectively.

FIGURE 5.

Biochemical analyses of GAOs from various Asteraceae plants. A, shown is LC-MS chromatography at selective (−)233 ion for germacrene A acid and (−)171 ion for the internal standard (IS), decanoic acid. The arrow is germacrene A acid, and the arrowhead is the internal standard. Yields of germacrene A acid from four independent transformants were given at the start of the chromatographs (mean ± S.D.). B, immunoblot analysis of recombinant GAOs is shown. FLAG antibodies were used to detect the epitope tags at the C termini of GAOs. Loaded microsome amounts are indicated. Ls, L. sativa; Ci, C. intybus; Sl, S. lappa; Bs, B. spinosa; Ha, H. annuus. C, LC-MS chromatography of B. spinosa extract at (−)233 ion was shown in the top line. Standards for germacrene A acid (GAA) and costic acids are shown in the second and third line, respectively.