A detailed view on 1,8-cineol biosynthesis by Streptomyces clavuligerus

The stereochemical course of the cyclisation reaction catalysed by the bacterial 1,8-cineol synthase from Streptomyces clavuligerus was investigated using stereospecifically deuterated substrates. In contrast to the well investigated plant enzyme from Salvia officinalis, the reaction proceeds via (S)-linalyl diphosphate and the (S)-terpinyl cation, while the final cyclisation reaction is in both cases a syn addition, as could be shown by incubation of (2-13C)geranyl diphosphate in deuterium oxide.


Introduction
Among all classes of natural products the climax of structural diversity and complexity is reached within the largest, the terpenoids. An estimated number of 75,000 different compounds are known from all kinds of organisms including plants [1], bacteria [2][3][4][5], fungi [6] and, as recently shown, even social amoebae [7]. These molecules are all made from only a handful of linear and achiral precursors such as geranyl diphosphate (GPP, monoterpenes), farnesyl diphosphate (FPP, sesquiterpenes) and geranylgeranyl diphosphate (GGPP, diterpenes). Terpene cyclases (type I) contain a trinuclear (Mg 2+ ) 3 cluster in their active site that is stabilised by binding to several highly conserved motifs including the aspartate-rich motif (DDXXD) and the NSE triad (ND(L,I,V)XSXXXE, modified in plants to a DTE triad: DD(L,I,V)XTXXXE) [8]. Their substrates bind with the diphosphate portion to the (Mg 2+ ) 3 cluster and via hydrogen bridges to a highly conserved arginine (diphosphate sensor) and a RY dimer [9]. The substrate is ionised by diphosphate abstraction and the resulting allyl cation undergoes a domino reaction via a series of cationic intermediates and a final deprotonation or attack of water to yield a terpene hydrocarbon or alcohol. This reaction cascade proceeds in a hydrophobic cavity from which water is excluded to enable carbocation chemistry in an aqueous environment. Furthermore, the hydrophobic cavity provides a template that arranges the substrate in a certain conformation to determine the formation of a specific product. Single residues such as phenylalanines are involved in the stabilisation of cationic intermediates, e.g., by cation-π interactions [8][9][10]. The overall process usually generates an en-antiomerically pure (poly)cyclic terpene with several stereogenic centres. A large variety of carbon skeletons is accessible, e g., more than 120 skeletons each representing various stereoisomers and constitutional isomers with different positioning of olefinic double bonds or alcohol functions are known just for sesquiterpenes [11]. The structural diversity of terpenoids can be further increased by the action of tailoring enzymes such as cytochrome P450 monooxygenases and acyl transferases [12,13]. Very few cases are known in which terpene cyclases generate an achiral product as exemplified by the monoterpene 1,8-cineol (eucalyptol, 1) and the sesquiterpenes germacrene B (2) and α-humulene (3) (Figure 1). A direct 1,6-cyclisation of the monoterpene precursor GPP to 1 is prevented by the topological constraints associated with the (2E) geometry which necessitates the isomerisation of GPP (4) to linalyl diphosphate (LPP, 5) followed by an anti,endo-S N 'cyclisation [14], but the stereochemical course of this reaction is not readily clear and may proceed via either enantiomer of the α-terpinyl cation (6, Scheme 1). Isotopic labelling experiments currently experience a revival [15] and are a very powerful method to follow the enzyme mechanisms of terpene cyclases [16][17][18][19][20][21][22][23][24] including the stereochemical courses of the cyclisation reactions (reviewed in [25]). While the stereochemical course of the GPP cyclisation to 1 has been investigated for the 1,8-cineol synthase from Salvia officinalis [26,27], it is unknown for the bacterial enzyme that was recently reported from Streptomyces clavuligerus [28]. Here we describe isotopic labelling experiments that gave insights into the cyclisation mechanism of the bacterial 1,8-cineol synthase.

Results
The absolute configuration of the intermediate terpinyl cation  Alpine borane reduction [29] (both enantiomers of this reagent are commercially available) of (1-2 H)geranial to (R)-and (S)-(1-2 H)geraniol that were obtained with high enantiomeric excess (>95% ee) as determined by Mosher ester analysis ( Figure S1, Supporting Information File 1). The alcohols were subsequently converted into the corresponding diphosphates using triethylammonium phosphate in trichloroacetonitrile [30,31]. The gene encoding the 1,8-cineol synthase [28] was cloned into the yeast-to-Escherichia coli shuttle vector pYE-Express by homologous recombination in yeast [32], followed by expression in E. coli BL21. The protein was purified by Ni 2+ -NTA affinity chromatography and used to convert both (R)-and (S)-(1-2 H)GPP into ( 2 H)-1 (in agreement with the findings described in reference [28], 1 is the only product from unlabelled GPP as was shown by GC-MS, Figure S2, Supporting Information File 1). The obtained products were analysed by HSQC spectroscopy (Figure 2). While for unlabelled 1 a 2:2 signal intensity is observed for the crosspeaks representing the Syn versus anti addition in the final ring closure The final cyclisation step from (S)-7 via 8 to 1 can in principle proceed either via a syn or an anti addition to the olefinic double bond, requiring a protonation of the original C-2 of GPP. To distinguish between these alternatives (2-13 C)GPP was synthesised from sulcatone (Scheme S2, Supporting Information File 1) and converted by the 1,8-cineol synthase in deuterium oxide. The obtained product was analysed by HSQC spectroscopy ( Figure 3A), showing that deuterium is taken up into the exo position at the 13 C-labelled C-2 of 1 (indicated by H'), while the endo position (H'') is occupied by the proton from the substrate, resulting in a strong crosspeak [22]. Furthermore, deuterium incorporation at C-2 was indicated by a strongly enhanced triplet in the 13 C NMR spectrum due to 13 C-2 H-spin coupling ( Figure S4, Supporting Information File 1) [11,20,23]. This finding is in agreement with a syn addition to the olefinic double bond of (S)-7 in the final cyclisation step. It is possible that the proton is directly transfered from the protonated hydroxy function in (S)-7 to C-2 ( Figure 3B). Alternatively, a deprotonation of (S)-7 to the hypothetical neutral intermediate α-terpineol followed by reprotonation at C-2 from the Si face can be assumed, but these two alternatives cannot be distinguished based on the labelling experiments described here.

Discussion
Plant and bacterial terpene cyclases show important structural differences [33]. While plant monoterpene synthases are composed of α and β domains and exhibit a quarternary α 2 β 2 structure [34,35], bacterial mono-and sesquiterpene cyclases are monodomain enzymes (α) [9,10,36]. Accordingly, also the 1,8-cineol synthases from Salvia officinalis and from Streptomyces clavuligerus are not related and have evolved independently. While the plant enzyme was shown to convert GPP via (R)-LPP ((R)-5) and the (R)-terpinyl cation ((R)-6) into 1 [27], the experiments described here revealed a different course for the bacterial enzyme via (S)-6. This finding is particularly interesting, because it reflects the frequent observation that the (chiral) products of bacterial terpene cyclases represent the opposite enantiomers as usually generated by plant enzymes [24,37,38]. Both enzymes from Salvia officinalis and from Streptomyces clavuligerus share the syn addition in the final cyclisation step which can be rationalised by a direct intramolecular proton transfer, circumventing the need of a low-energy neutral intermediate such as α-terpineol. However, in case of the sesquiterpene ethers corvol ethers A (19) and B (18) a reprotonation step was shown to proceed from the opposite face than the preceeding attack of water, thus excluding a direct proton transfer from oxygen to the neighbouring carbon (reactions from 12 to 14 in Scheme 2). Conclusively, reprotonation of the neutral intermediate 13 is possible in this case [21].

Experimental Cloning and homologous recombination
Cells of Streptomyces clavuligerus ATCC 27064 were obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ, Braunschweig, Germany). Cultivation was done in liquid 65 Gym medium (4 g yeast extract, 4 g glucose, 10 g malt extract, 1 L water, pH 7.2) and isolation of genomic DNA was performed using a standard protocol. The gene WP_003952918 encoding the 1,8-cineol synthase was amplified using forward primer (ATGCCCGCCGGCCAC-GAAGA) and reversed primer (TCACCAAGGGGTGGTG-GCCC). The isolated PCR product was elongated for homologous recombination with pYE-Express in yeast with a second set of primers (GGCAGCCATATGGCTAGCATGACTGGTG-GAATGCCCGCCGGCCACGAAGA and TCTCAGTGGTG-GTGGTGGTGGTGCTCGAGTTCACCAAGGGGTGGTGG CCC). This elongated product was transformed in Saccha-romyces cerevisiae FY834 together with linearised vector pYE-Express [32] (EcoRI and HindIII digestion) using the LiOAc/SS carrier DNA protocol [39]. Transformed cells were plated on SM-URA medium (20 g glucose, 1.7 g yeast nitrogen base, 5 g ammonium sulphate, 0.77 g nutritional supplement minus uracil, 24 g agar, 1 L water) and grown for 3 days at 28 °C. Plasmids were isolated using the kit Zymoprep Yeast Plasmid Miniprep II (Zymo Research, Irvine, USA), shuttled in E. coli BL21 by electroporation and confirmed by sequencing.

Incubation experiment with (2-13 C)GPP
Enzyme purification starting from an E. coli expression culture (0.5 L) was performed as described above. The last washing fraction was substituted with 2 H 2 O-based binding buffer and elution was done with 2 H 2 O-based elution buffer. The first elution fraction was incubated with (2-13 C)GPP (0.8 mg) for 16 h at 28 °C. The enzyme reaction was extracted with ( 2 H 6 )benzene (0.6 mL), dried with MgSO 4 and the extract was analysed directly by GC-MS and NMR.

GC-MS analysis
An Agilent 7890B gas chromatograph equipped with a HP5-MS silica column (30 m, 0.25 mm inner diameter, 0.50 μm film) connected to an Agilent 5977A inert mass selective detector was used to acquire GC-MS data. Instrumental settings were: (1) inlet pressure: 77.1 kPa, He: 23.3 mL/min, (2) transfer line: 250 °C, (3) electron energy: 70 eV. The GC was set to 50 °C starting temperature for 5 min, then increasing with 5 °C per minute to 320 °C and holding this temperature for another 5 min. The injection volume was 2 μL and the inlet was operating in split mode (10:1, 60 s valve time). Helium was used as the carrier gas at 1 mL/min. Retention indices were determined against a homologous series of n-alkanes (C 8 -C 40 ).
In a second flask, to a solution of (n-Bu 4 ) 3 HP 2 O 7 (97 mg, 0.11 mmol, 1.5 equiv) in dry CH 3 CN (1.0 mL) the crude product of the allyl bromide (1.0 equiv) was added and the reaction mixture was stirred for 2 h at room temperature and then concentrated under reduced pressure. The colorless oil was loaded onto an ion exchange column (DOWEX 50W-X8, NH 4 + form). Elution of the product was performed by addition of two column volumes of ion exchange buffer (0.03 M NH 4 HCO 3 in 2% iPrOH/H 2 O). Freeze drying yielded the product as a white solid (14.1 mg, 0.04 mmol, 55%).

Supporting Information
Synthesis schemes, Mosher ester analysis of (R)-and (S)-(1-2 H)GPP, gas chromatogram of the enzyme product of 1,8-cineol synthase, 13 C NMR of the enzyme product from (2-13 C)GPP in deuterium oxide buffer, and full NMR data of 1.

Supporting Information File 1
Additional material.