Taxadiene synthase structure and evolution of modular architecture in terpene biosynthesis

Abstract

With more than 55,000 members identified so far in all forms of life, the family of terpene or terpenoid natural products represents the epitome of molecular biodiversity. A well-known and important member of this family is the polycyclic diterpenoid Taxol (paclitaxel), which promotes tubulin polymerization and shows remarkable efficacy in cancer chemotherapy. The first committed step of Taxol biosynthesis in the Pacific yew (Taxus brevifolia) is the cyclization of the linear isoprenoid substrate geranylgeranyl diphosphate (GGPP) to form taxa-4(5),11(12)diene, which is catalysed by taxadiene synthase. The full-length form of this diterpene cyclase contains 862 residues, but a roughly 80-residue amino-terminal transit sequence is cleaved on maturation in plastids. We now report the X-ray crystal structure of a truncation variant lacking the transit sequence and an additional 27 residues at the N terminus, hereafter designated TXS. Specifically, we have determined structures of TXS complexed with 13-aza-13,14-dihydrocopalyl diphosphate (1.82 Å resolution) and 2-fluorogeranylgeranyl diphosphate (2.25 Å resolution). The TXS structure reveals a modular assembly of three α-helical domains. The carboxy-terminal catalytic domain is a class I terpenoid cyclase, which binds and activates substrate GGPP with a three-metal ion cluster. The N-terminal domain and a third 'insertion' domain together adopt the fold of a vestigial class II terpenoid cyclase. A class II cyclase activates the isoprenoid substrate by protonation instead of ionization, and the TXS structure reveals a definitive connection between the two distinct cyclase classes in the evolution of terpenoid biosynthesis.

Figure 1. Proposed catalytic mechanism of taxadiene synthase

The cyclization of GGPP to form taxadiene is the first committed step of Taxol (paclitaxel) biosynthesis in yew species (OPP = diphosphate, Ph = phenyl, Ac = acetyl, Bz = benzyl). Taxadiene is converted to Taxol through a lengthy series of oxidation and acylation steps.

Figure 2. Structural relationships among terpenoid cyclases

The class I terpenoid cyclase fold of pentalenene synthase (PDB 1PS1) (blue, “α-domain”13) contains metal-binding motifs DDXXD and (N,D)DXX(S,T)XXXE (red and orange, respectively); in 5-epi-aristolochene synthase (PDB 1LZ9), this domain is linked to a smaller vestigial domain (green, “β-domain”13). A related domain is found in the class II terpenoid cyclase fold of squalene-hopene cyclase (PDB 1SQC), where it contains the general acid motif DXDD (brown) and a second domain (yellow, “γ-domain”13) inserted between the first and second helices; a hydrophobic plateau flanking helix 8 (gray stripes) enables membrane insertion. Taxadiene synthase (PDB 3P5R) contains both class I and class II terpenoid cyclase folds, but only the class I domain is catalytically active. The role of N-termini (purple) in class I plant cyclases is to "cap" the active site, as shown for 5-epi-aristolochene synthase.

Figure 3. Substrate analogue binding to TXS

a, Simulated annealing |F_o|-|F_c| omit map in which FGP and 3 Mg²⁺ ions are omitted from the structure factor calculation (contoured at 3.0σ); the side chains of metal ligands are indicated. b, Molecular recognition of the substrate diphosphate group in the TXS active site. For clarity, the isoprenoid moiety of FGP is truncated to one carbon (gray). Metal coordination and hydrogen bond interactions are indicated by green and black dashed lines, respectively. Atoms are color coded as follows: carbon = yellow, nitrogen = blue, oxygen = red, phosphorus = orange; Mg²⁺ ions and water molecules appear as purple or red spheres, respectively. A corresponding stereofigure is found in Figure S3.

Figure 4. Active site cavities of terpenoid synthases

a, Active site volumes are generally slightly larger than corresponding substrate and product volumes, perhaps to accommodate structural changes better during the cyclization cascade. Abbreviations are outlined in Table S2. b, Superposition of TXS (blue) and bornyl diphosphate synthase (green) guides the modeling of the J–K loop and the N-terminal segment of TXS (red) to define the enclosed active site cavity (solvent-accessible surface, magenta meshwork). c, One orientation of taxadiene (blue) fits in the active site cavity such that the H5β atom of the preceding taxen-4-yl cation would be oriented toward the diphosphate leaving group, suggesting that the PP_i anion could serve as the stereospecific base that terminates the cyclization cascade. Three Mg²⁺ ions and FGP are shown for reference; all protein atoms are omitted for clarity. A corresponding stereofigure is found in Figure S4.