Pseudopterosin Biosynthesis: Aromatization of the Diterpene Cyclase Product, Elisabethatriene

Putative precursors in pseudopterosin biosynthesis, the hydrocarbons isoelisabethatriene (10) and erogorgiaene (11), have been identified from an extract of Pseudopterogorgia elisabethae collected in the Florida Keys. Biosynthetic experiments designed to test the utilization of these compounds in pseudopterosin production revealed that erogorgiaene is transformed to pseudopterosins A–D. Together with our previous data, it is now apparent that early steps in pseudopterosin biosynthesis involve the cyclization of geranylgeranyl diphosphate to elisabethatriene followed by the dehydrogenation and aromatization to erogorgiaene.


Introduction
The pseudopterosins are a class of diterpene glycosides isolated from the sea whip Pseudopterogorgia elisabethae [1][2][3]. As with many marine natural products, diverse congeners of the pseudopterosins are found in different locations and there are presently fifteen known pseudopterosin derivatives (A-O). All of the known pseudopterosins contain the amphilectane skeleton with a glycosidic linkage at either C-9 or C-10.
The identity of the sugar and the degree of acetylation account for the additional structural variation of this family of diterpenes. Pseudopterosins A-D (1)(2)(3)(4), from Sweetings Cay in the Bahamas, possess the amphilectane skeleton with an attached xylose sugar which is acetylated at different locations ( Figure 1).   The seco-pseudopterosins A-D are a related group of compounds (5)(6)(7)(8) belonging to the serrulatane class of diterpenes initially isolated from Pseudopterogorgia kallos in the Florida Keys [4]. More recently, novel seco-pseudopterosins were reported to co-occur with pseudopterosins in P. elisabethae [3]. The pseudopterosin and seco-pseudopterosin classes of diterpenes exhibit potent anti-inflammatory and analgesic activity [3,5]. The pseudopterosins are pharmacologically distinct from typical NSAIDs and they appear to act by a novel mechanism of action [6,7]. The commercial market for the pseudopterosins, presently as ingredients in a skin cream, indicates a need for the development of a sustainable supply of these compounds. Consequently, a continuing goal in our laboratory is to elucidate all steps in the biosynthetic pathway leading to the pseudopterosins.
We recently confirmed the identity of the diterpene cyclase product leading to the pseudopterosins as elisabethatriene (9) (Figure 2). This hydrocarbon, with the serrulatane skeleton, was isolated from extracts of P. elisabethae collected in the Florida Keys and in Sweetings Cay, Bahamas. The utilization of 9 in pseudopterosin biosynthesis was confirmed through biosynthetic experiments [8]. Erogorgiaene (11) was recently reported from a collection of P. elisabethae off Colombia [9]. Given the structure of the pseudopterosin class of diterpenes and our report of the transformation of 9 to 1-4, it seems reasonable to suggest that 11 is an intermediate in this biosynthetic pathway. Further, it seems plausible that an endocyclic isomer of 9 such as 10 could be an intermediate in the conversion of 9 to 11. This report describes the results of experiments directed at testing the hypothesis that isoelisabethatriene (10) and erogorgiaene (11) are early intermediates in pseudopterosin biosynthesis. Our approach was to identify these in our P. elisabethae extracts through the synthesis of standard samples of 10 and 11 from 9 and if present, test the compounds as metabolic intermediates.

Identification of Plausible Biosynthetic Intermediates
We have observed that the Bahamian populations (e.g. Sweetings Cay) of P. elisabethae contain much higher concentrations of pseudopterosins than P. elisabethae from the Florida Keys; the latter, however, exhibits a greater diversity of diterpene chemistry [10]. A preliminary survey of the non-polar fraction of an extract of P. elisabethae from Sweetings Cay indicated a lack of hydrocarbons that were structurally related to the pseudopterosins and therefore, not likely involved in pseudopterosin biosynthesis. Given the diverse diterpene chemistry of Floridian specimens of P. elisabethae, we examined these collections for the presence of hydrocarbons such as 10 and 11.
To aid in the search for erogorgiaene (11), we synthesized a standard sample of 11 from 9 using Pd/C under nitrogen [10] (Scheme 1). We utilized this sample of 11 as a standard to screen for the presence of this compound in a hexane extract. This analysis revealed the presence of compound 11 in 0.2% of the 100 g extract. NMR analysis of this HPLC fraction confirmed the purity and identity of erogorgiaene.

11
Pd/C H Similarly, a standard sample of 10 was synthesized from 9, as previously described [8], and this was used to monitor its presence in Floridian P. elisabethae extracts. We began the search for naturally occurring compound 10 in P. elisabethae by obtaining a crude organic extract and partially purifying the extract using a silica flash column. After eluting the silica column with 100% hexanes, analysis of this non-polar fraction by HPLC indicated the presence of a compound with the same retention time and UV absorption profile (λ max = 245) as synthetic 10. Subsequent NMR analysis confirmed the purity and identity of naturally occurring 10. Interestingly, neither 10 nor 11 were present in isolable quantities in extracts of Bahamian samples of P. elisabethae.
The identification of compound 10 in the Florida Keys' P. elisabethae represents the isolation of a novel marine metabolite which may be involved in pseudopterosin biosynthesis. Additionally, the natural occurrence of compounds 10 and 11 in the Florida Keys' P. elisabethae indicates that this specific population of the coral may be more useful for the identification of putative biosynthetic intermediates than the previously analyzed Bahamian populations. While we confirmed the presence of compounds 10 and 11 in P. elisabethae, the presence alone is, of course, not sufficient to confirm their involvement in pseudopterosin biosynthesis. This question was addressed by conducting various incubation experiments and using derivatizations to rigorously confirm radiochemical purity.

Purification of 3 H-elisabethatriene
Our approach for the biosynthetic experiments was to utilize a cell-free extract of P. elisabethae to produce 3 H-labeled 9 from [1-3 H]-geranylgeranyl diphosphate (GGPP) and then use this to test for the conversion of 9 to 10 and 9 to 11. Thus, a protein preparation of P. elisabethae was incubated with 50 µCi of [1-3 H]-GGPP. After extracting with hexanes, the organic fraction was partially purified through a small silica column and elisabethatriene (9) was rigorously purified by reversed phase HPLC and a portion subjected to scintillation counting. Compound 9 was generated in a radiochemical yield of 0.6% (300,000 dpm) and shown to be radiochemically pure as previously described [8]. This low yield is expected for a reactive intermediate and greater radioactivity was observed for more polar metabolites including the pseudopterosins.
Purified 3 H-labeled 9 (300,000 DPM) was reincubated with a cell-free extract of P. elisabethae (Florida Keys) for 1 hour. The quenched incubation mixture was partially purified by elution through a silica column with 100% hexanes and the resulting non-polar fraction subjected to repeated HPLC fractionation. The radioactivity of fractions corresponding to compounds 9-11 was measured using a scintillation counter (Table 1). Table 1. Recovered Radioactivity of Compounds 9-11 after Incubation of 3 H-labeled 9 with a Cell-free Extract of P. elisabethae (Florida Keys).

Compound
Radioactivity (DPM) 9 48,740 10 Background 11 2,060 The isolated 10 was not radioactive and consequently was deemed not to be involved in the pseudopterosin biosynthetic pathway. The lack of radioactivity in 10 could be explained by the fact that there could be two different diterpene cyclases in the Floridian populations of P. elisabethae yielding two different cyclase products. This is consistent with the wide variety of diterpenes present in the Florida samples of this gorgonian [10].
As described in Table 1, 9 was transformed to 11 with a radiochemical yield of 0.7% (2,060 dpm). Radiochemical purity was established for erogorgiaene by conversion of the recovered 11 to an epoxide and monitoring the specific activity of the substrate 11 and the epoxide product 12 (Scheme 2). Scheme 2. Synthetic derivatization of 11 to establish radiochemical purity. Conditions for this derivitization were established with non-radioactive erogorgiaene (11). The reaction of 11 with meta-chloroperoxybenzoic acid (m-CPBA) proceeded in excellent yield. Importantly, compound 12 eluted more than 10 min. earlier than 11 on reversed phase HPLC, thus ensuring a simple purification of the product. The HREIMS of 12 established a molecular formula of C 20 H 30 O. The 1 H-NMR spectrum of 12 was similar to that of 11. One variation between the NMR spectra of the two compounds was that the olefinic proton in 11 [δ5.27 (1 H, br t)] was replaced with a signal at δ2.57 (1 H, dd) for 12. Additionally, the two signals for the olefinic methyl groups in 11 [δ1.60 (3 H, br s) and 1.71 (3 H, br s)] were shifted upfield to δ1.14 (3 H, br d) and δ1.17 (3 H, br s) for 12.
Following the purification of radioactive 11 from the incubation with 3 H-labeled 9, 3 H-labeled 11 (4.27 x 10 6 dpm/mmol) was treated with m-CPBA in CHCl 3 and the peak corresponding to the epoxide (12) was purified by reversed phase HPLC. Compound 12 was found to have the same specific activity (4.94 x 10 6 dpm/mmol) as 11 (4.27 x 10 6 dpm/mmol). The lack of change of specific activity in this derivatization reaction establishes the radiochemical purity of compound 11 and thus confirms that erogorgiaene is derived from elisabethatriene (9).

Evaluation of 3 H-erogorgiaene in Pseudopterosin Biosynthesis
The data presented above demonstrated that elisabethatriene (9) is transformed to its aromatic derivative 11 in P. elisabethae. To test for the conversion of 11 to the pseudopterosins, and thus confirm the involvement of 11 in the pseudopterosin biosynthetic pathway, 3 H-labeled 11 (5,130 DPM) was incubated with a Bahamian P. elisabethae cell-free extract [8]. Radioactive 11 was obtained from the incubation of [1-3 H]-GGPP and also by oxidation of 9 [11]. Following incubation, pseudopterosins A-D (1-4) were rigorously purified by HPLC. Liquid scintillation counting of the pseudopterosins from the HPLC separation indicated transformation of 3 H-11 to 1-4. The overall radiochemical yield of the four pseudopterosins was 5.7% (290 DPM). The radiochemical yield of the pseudopterosins obtained in this experiment is an order of magnitude higher than that reported for the conversion of GGPP to 9 [8].

Conclusions
These data suggest that in pseudopterosin biosynthesis, erogorgiaene is produced from elisabethatriene by a dehydrogenation and spontaneous aromatization (Scheme 3). It is conceivable that isomers of elisabethatriene, other than 10, could be involved in this pathway. However, lack of such radioactive fractions in the HPLC analysis suggests that this is not likely.
The lack of transformation of elisabethatriene to the isomeric 10 suggests that the conversion of 9 to 11 involves a dehydrogenation which is then presumably followed by a spontaneous aromatization. In addition, utilization of the chemical variance of different populations of P. elisabethae for the isolation of biosynthetic intermediates and subsequent radiolabeling experiments proved to be invaluable for this investigation.

Cell-Free Extract Preparation
Flash frozen P. elisabethae (100 g) was homogenized with buffer [20 mM Tris-HCl (pH 7.7) containing 3 mM EDTA, 5 mM β-mercaptoethanol, and 5 mM MgCl 2 ] in a commercial blender. The buffer was added to the blender and was the coral blended in the presence of liquid nitrogen. After homogenizing, the cell-free extract was centrifuged to remove insoluble debris. Initially, centrifugation was conducted at 9,700 g for 10 min. The pellet was discarded and the supernatant was centrifuged again at 39,000 g for 20 min. The pellet was discarded and the supernatant centrifuged one last time at 39,000 for 30 min. A portion of the cell-free extract (40 mL) was incubated with radiolabelled compounds.

Identification of erogorgiaene (11) in P. elisabethae collected in the Florida Keys
Elisabethatriene (9) (0.5 mg) was reacted with a spatula tip of Pd/C in 500 µL of triethylene glycol dimethyl ether (triglyme), and the mixture refluxed under nitrogen for 3 hours. Following the reaction, the sample was filtered and the solvent evaporated under nitrogen. Analysis was then performed by reversed phase HPLC using a diode array detector (λ = 215 nm) with 100% methanol (2 mL/min.). To identify the compound in the Florida Keys' P. elisabethae extract, the hexane layer (~6 g) obtained from partitioning a crude methylene chloride/ethyl acetate extract with hexanes and methanol/water (9:1) was passed through a small silica column with 100% hexanes. The solvent was evaporated using a rotary evaporator and fractionated by HPLC using a diode array detector (λ = 215 nm) with 100% methanol (2 mL/min.). The peak corresponding to a retention time of 22 min. was subjected to NMR analysis.

Identification of compound 10 in P. elisabethae collected in the Florida Keys
A crude methylene chloride/methanol extract of P. elisabethae (Florida Keys) was partially purified through a silica column eluted with 100% hexanes. The sample was further purified using preparative TLC with hexanes as the eluent and the band corresponding to an R f of approximately 0.6 was isolated. Further fractionation was achieved by reversed phase HPLC using a refractive index detector and 100% methanol (2 mL/min.). The peak corresponding to a retention time of 27 min. was subjected to NMR analysis.

Purification of 3 H-labeled 9 from incubation of [1-3 H]-GGPP
Two separate reactions were conducted by incubating 20 µCi of [1-3 H]-GGPP with each of two 500 µL aliquots of a Sweetings Cay P. elisabethae cell-free extract for 1 hour at 29°C and 200 rpm. A Sweetings Cay P. elisabethae cell-free extract that had been partially purified by DEAE-cellulose anionexchange chromatography was also incubated with 10 µCi of [1-3 H]-GGPP for 1 hour at 29°C and 200 rpm. The samples were extracted with hexanes and passed through a small silica pasteur pipet column (5 cm). Elisabethatriene (300,000 DPM) was rigorously purified by reversed phase HPLC using a refractive index detector and 100% methanol (2 mL/min.). The radioactivity was measured using a liquid scintillation counter.

Incubation of Florida Keys' P. elisabethae Cell-free Extract with 3 H-9 and Purification of 3 H-11 and 3 H-10
3 H-Elisabethatriene (300,000 DPM) was collected from HPLC purification and the solvent was evaporated under a stream of N 2 . Following the addition of 1 mL of assay buffer and 0.05% Tween 20, the sample was sonicated for 10 min. and cell-free extract was added to a total volume of 40 mL. The sample was then incubated at 29°C and 200 rpm for 1 hour. The sample was lyophilized and partitioned between hexanes and methanol/H 2 0 (9:1). After partial purification through a silica column (7 cm) with 100% hexanes, the sample was purified using reversed phase HPLC with 100% methanol (2 mL/min.). Compounds 9-11 were collected in separate vials and the solvent evaporated under N 2 . Ten percent of each sample was then reinjected and fractions were collected before and after each peak. The solvent was evaporated from the fractions and they were subjected to liquid scintillation counting. For compound 10, another 40% was added to the sample and the liquid scintillation counting repeated. The remaining amount (90%) of 9 (48,740 DPM) and 11 (1,860 DPM) was then reinjected to purify further.
Derivitization of 11 to 12 meta-Chloroperoxybenzoic acid (m-CPBA, 76.2%, 3.5 mg) was dissolved in dry chloroform (5 mL). A portion of this mixture (1 mL, 4.6 µmol, 1.04 mg m-CPBA) was pipetted into a small separatory funnel and added to a stirred vial of 11 (4.6 µmol, 1.25 mg) over 10 minutes. The reaction was conducted at 0°C for 2.5 hours. The reaction mixture was concentrated under N 2 and purified by reversed phase HPLC with 100% methanol (2 mL/min.) to afford 12 (1.25 mg) as a colorless oil.

Synthesis and Purification of Radioactive 12
3 H-Labeled 9 (45.6 µg, 169 nmol, 1,420 DPM) was treated with m-CPBA (39.7 µg, 175 nmol) in dry CHCl 3 at 0°C for 2.5 hours as described previously. The reaction mixture was purified by reversed phase HPLC as described (methanol, 2 mL/min.) to afford 8.8 nmol (5.2%) of 12 and the radioactivity measured. Quantities of the compounds were obtained by integration of HPLC peaks.

Synthesis and Purification of Radioactive 11
3 H-9 (48,740 DPM) purified from the incubation with [1-3 H] GGPP was treated with H 5 PMo 10 V 2 O 40 -(26.5 mg) in tetraglyme (55 µL) and 1,2-dichloroethane at 70°C for 2 hours as described above. The sample was purified as described in the previous section and 5% was reinjected along with a small amount of "cold" erogorgiaene (11). Fractions were collected before and after the erogorgiaene peak and subjected to liquid scintillation counting. The other 95% was reinjected separately for purification.