Biocatalysis for terpene-based polymers

1.1 Extended substrate space

Prenylation of phenyls and indoles by aromatic PTs [16], [17] constitutes a highly interesting avenue to add terpene-based handles by biocatalysis. This is exemplified here by the biocatalytic addition of farnesyl pyrophosphate (FPP) to 3,5-dimethylorsellinic acid catalyzed by Trt2 from Aspergillus fumigatus [18], an aromatic PT. This transformation is formally an aromatic addition; one that is dependent on enzyme-catalyzed ionization of the pyrophosphate group of FPP. Aromatic PTs can display a different fold and active site architecture compared with that of the isoprenoid coupling enzymes shown in Scheme S1B.

Transfer of isoprene units is instrumental in assembling linear terpenes with other metabolites, such as indoles [19], [20]: a process of importance for bioactivity of some antibiotics [21]. Thus, Friedel–Crafts-like aromatic substitutions are known in biology. Promiscuous [20], [22] aromatic PTs are of particular biocatalytic interest toward the generation of fine chemical synthons and complex alkaloids [18], [22].

Generation of terpene-based alcohols can be imposed by carbocationic mechanisms of terpene synthases. Specifically, breakage of the labile allylic carbon-oxygen bond mediates the release of pyrophosphate, which is followed by water addition. This biotransformation is shown for trans,trans-nerolidol, which can be generated from FPP by sesquiterpene synthases [23].

Epoxidation of the terminal isoprene group by mono-oxygenases is instrumental for steroid biosynthesis. This biocatalytic process also reduces the activation barrier for subsequent cyclization steps, as protonation of an oxirane is facilitated compared with that of a terminal isoprene group.

Synthesis of alternative linear polyisoprenes through methylation of their backbone can remarkably bypass the well-known “isoprene rule” [24]. The biosynthesis of methyl-GPP (shown with methyl group in red), is founded on S-adenosylmethionine (SAM)-dependent methylation of 6, catalyzed by geranyl diphosphate methyltransferase [25].

1.2 Cyclization

Cyclizations of (functionalized) linear terpenoids by TCs significantly expand terpene diversity. Cyclization of prenylated terpenes, obtained from strategy 1 above, is key to incorporate aromatic rings into multicyclic natural products, which tailors their bioactivity [18]. We and others have demonstrated the high potential of using class II TCs to generate polyheterocyclic building blocks through cyclization of terpene-based alcohols [26], [27], [28]. Epoxidized linear terpenes significantly expand reaction space by adding an additional functional handle (i.e. hydroxyl group) in generated cyclic product, which is of basal importance in plant biology and cell signaling [4]. Recently, metabolic engineering and promiscuous TCs were used to generate unprecedented cyclic C₁₁ terpenes from methyl-GPP, exemplified here for 2-methyllimonene 9 [29]. The exciting possibility of interception of highly reactive carbocations generated during TC catalysis, by additional nucleophiles beyond water, have been demonstrated (corresponding to Nu=R-OH in Figure 1, top right) [30].

1.3 Product functionalization

Oxidation is one fundamental process to unlock the reactivity of cyclic building blocks for their controlled polymerization, e.g. via diols [31] and keto-groups shown for pinocarvone 18 [32]. Starting from cyclic building blocks generated by TCs above, and exemplified here by (–)-α-pinene (7), (–)-β-pinene (8), and 2-methyllimonene (9, Figure 1), P450 monooxygenases constitute potent biocatalysts to achieve oxidation of the carbocyclic ring(s) [33]. P450 monooxygenases are known to install oxirane groups (exemplified by limonene epoxide 10a), alcohols (represented by (1R,5S)-carveol 11) and even keto-groups (represented here by unsaturated carvone 12 and (–)-verbenone 17) [34], [35]. Optionally, the emerging family of peroxygenases, which rely on hydrogen peroxide to generate the oxoferryl-heme intermediate, are known to afford the generation of terpene-based alcohols and epoxides [36], [37]. Alcohol dehydrogenases can readily generate keto-functionalized terpenes [38], [39] (e.g. 12 and 17) from corresponding alcohols. Ene reductases [40] can afford subsequent reduction of the α,β-double bond in 12 to generate saturated carvone 13 [41]. Recently discovered [42] ene-reductases from natural product biosynthesis are expected to significantly advance the toolbox of terpene functionalization. Further oxidation of ketones into terpene-derived lactones by Baeyer–Villiger monooxygenases [43], [44] (BVMOs) is exemplified here for carvo-lactone 14a generated from 13 [45].

With respect to product diversity, prenylation of non-terpene-derived motifs in meroterpenoids is abundant in nature [46]. Halogenation by vanadium haloperoxidases [47] generate cyclic and halogenated (mero)terpenoids [48], shown here for the napyradiomycin family of natural products (15). Recently, non-heme-dependent iron dioxygenases have been in the spotlight for the generation of terpene-based diols (not shown) [49].

1.4 Polymerization

For linear terpenes, bio-based polyisoprene can be generated by pyrophosphate release from 4 catalyzed by isoprene synthase [23], followed by radical polymerization. Polymerization of decorated, multicyclic terpenoids, in particular obtained through enzymatic oxidation [37], constitutes a highly interesting strategy to generate bio-based materials with tailored backbone functionalization. This is shown here for limonene polycarbonate 10b, which can be manufactured from trans limonene oxide 10a [50]. We strongly believe that the introduction of the lactone functional group to cyclic building blocks will enable the design of new polymers. Recently, Messiha et al. investigated a semisynthetic approach using biocatalysis with renewable feedstocks as a venue to generate lactone monomers [51]. These monomers were polymerized using a mild metal-organic catalyst to result in the formation of previously unreported polymers [51]. For instance, carvone-lactone 14a could generate polyesters with pendant alkyl chains as shown here for the previously published biopolymer 14b′ [52] (analogous polyesters can be assembled from (–)-menthide [53]). Another example of the valorization of renewable resource-derived monomers is the polymerization of a lactone derived from verbenone 17 into the biopolymer 22, as recently demonstrated by us [54]. Poly(lactones) can be prepared by ring-opening polymerization (ROP) processes. ROP is a form of chain-growth polymerization in which the terminal end of the growing polymer will behave as a reactive moiety, enabling the addition of a cyclic monomer into the polymeric chain by ring opening. Three main mechanisms of ROP (anionic, cationic, and coordination-insertion) have been proposed based on the type of initiator/catalyst used [55], [56].

Anionic ring-opening polymerization (AROP) involves the addition of a small amount of a nucleophilic agent to attack the carbonyl carbon of the cyclic monomer and initiate its polymerization (Scheme 1A). This method can be efficiently controlled to produce high molecular weight polyesters. In contrast, cationic ring-opening polymerization (CROP) is a process that can be initiated by the addition of a trace amount of an electrophile to the cyclic monomer. In CROP, a positively charged intermediate will be formed and subsequently attacked by a cyclic monomer (Scheme 1B). In fact, not all cyclic esters undergo CROP. CROP depends on the ring size, to be more precise, on the ring strain. For instance, cyclic monomers with 4-, 6-, and 7-membered rings polymerize readily via CROP, unlike cyclic esters with smaller membered rings (or without ring strain); they will not be polymerized by CROP. On the other hand, coordination-insertion ring-opening polymerization (C-IROP) involve the coordination of the monomer to an active intermediate and then its insertion into the metal-oxygen bond by electron rearrangement (Scheme 1C). When two cyclic esters with comparable reactivity are mixed together in a C-IROP system, a block copolymer can be produced by sequential addition to this living system [57]. Last, but not least, ROP of lactones can also proceed through free radical polymerization [58]. This method is a powerful tool for enabling the introduction of various functional groups into the polymer by chain growth rather than step growth. By free radical polymerization, it is conceivable to design polymers with similar or inferior density than the monomers. This displays a remarkable importance for applications where it is required to keep a constant volume throughout the polymerization such as tooth fillings, coatings, and others [59].

Scheme 1:

Illustration of chain-growth in ROP processes. (A) Anionic, (B) cationic, and (C) coordination-insertion ROP. Reproduced from Stridsberg et al.

1.5 Potential

It is envisioned that bicyclic terpenoids, with the additional fused ring system shown in red (Figure 1), could be upgraded to novel terpene-based materials. Monoterpenes and their lactone derivatives exhibited a great interest due to their valuable characteristics. Lactone derivatives can be polymerized and used in value-added applications, this includes biodegradable composites of plastics [60], drug delivery vectors, and scaffolds in tissue engineering [61]. Lactone polymers are appropriate biomedical starting materials for copolymerization and blending to design medical devices with fascinating mechanical properties and degradation kinetics.

Norcamphor (19) is a ketone (C₇H₁₀O) readily available through synthesis [62] and with a structure closely related to that of bicyclic terpenes (Figure 1). In this part, we will shed some light on the chemical conversion of norcamphor (19) into its lactone derivatives (20 and 21) via Baeyer–Villiger (BV) reaction, and the feasibility to polymerize these lactones through ROP using methanesulfonic acid (MSA) as a catalyst. BV oxidation was reported as an efficient reaction to convert cyclohexanones to lactones [63], [64]. Likewise, BV oxidation of 19 has been reported [65]. The oxidation of 19 by BV using m-CPBA as an oxidizing agent can result in the formation of 20 and 21 lactones (Scheme S2). The characteristics and the chemical structures of the lactones were investigated by nuclear magnetic resonance (NMR) analysis (¹H, ¹³C, COSY, and HSQC NMR). By ¹H NMR, we were able to estimate the conversion (%) of 19 into its lactone derivatives (Figure S1). It is indicated that the conversion of 19 into either 20 or 21 is 100%, as evidenced by the total disappearance of proton H_a´ (2.68 ppm) after BV reaction. H_a´is shifted down-field to H_a (4.88 ppm) in case of 20; however, in the case of 21, H_e´1 (2.05 ppm) and H_e´2 (1.86 ppm) of 19 were also shifted down-field to H_ε1 (4.33 ppm) and H_ε2 (4.12 ppm), respectively. Furthermore, based on the ¹H NMR data, the chemical BV oxidation of 19 resulted in the formation of 20 (93.4%) and a trace amount of 21 (6.6%).

The products were separated by column chromatography and we were able to retain 96.1% pure 20 lactone form (Figure S2). The spectral data (¹H and HSQC) confirmed that the retained lactone is 20 (Figure S2). The value of chemical shift of H_a (single proton) in the ¹H NMR spectra of 20 is 4.88 ppm, which reveals that this proton is attached to a carbon linked with an alkoxy oxygen atom. This provides a clear indication that the oxygen atom is inserted between carbon C_a and the carbonyl functionality.

Furthermore, the enzymatic transformation of 19 into the normal lactone 20 was performed with the Escherichia coli cell lysate containing the cyclohexanone monooxygenase from Acinetobacter calcoaceticus (CHMO_Acineto) containing amino acid variations for increased stability (CHMO_Acineto_QM) [43]. The E. coli cell lysate containing CHMO_Acineto_QM was also analyzed in terms of activity on nicotinamide adenine dinucleotide phosphate (NADPH), using both the natural substrate cyclohexanone and norcamphor 19, which showed a volumetric activity of 1.318±0.049 U mL⁻¹ and 1.088±0.090 U mL⁻¹, respectively. These data corroborate the high activity of the biocatalyst toward 19 (assuming similar uncoupling for the two substrates), in line with previous reports [65]. The high industrial potential of applying oxidoreductases in biosynthesis at larger scale has been discussed recently [37]. The biotransformation in the presence of FAD, cofactor regeneration system containing glucose dehydrogenase (GDH) and glucose, with NADPH and catalase for the uncoupling reaction, allowed 100% conversion of 19 into 20, with a 100% presence of the normal lactone under our working conditions. The results were analyzed by GC/FID analysis (Figure S5). The upscale of the reaction was performed in a 2 L baffled flask containing 200 mL of the reaction described above. After 24 h, 100% conversion was identified by GC/FID, responding to 50 mg product formed. To sum up, our results showed an extraordinary chemo-selective BV oxidation of 96% and 100% toward the formation of the isomer 20, using chemical and enzymatic catalysis, respectively. In the case of chemical transformation, it is attributed to the mechanism of action of BV reaction as discussed by Corma et al. [66].

1.6 Polymerization of lactone 20

Herein, we report the polymerization of lactone 20 using the ROP method through a cationic mechanism as shown in Scheme 1B. Benzyl alcohol (BnOH) was used as an initiator and MSA as a highly efficient catalyst for the ROP of lactones [67]. We have recently shown that this catalysts is suitable to generate novel biopolymers by ROP from bicyclic terpene-derived lactones [54]. The polymerization of 20 is illustrated in Scheme 2. Figure 2 shows the crude ¹H NMR spectra of polymer 20 after 24 h of reaction. Herein, we were able to confirm the polymerization of 20 for the first time, as the proton H_am at 4.88 ppm (H_a in the monomeric form lactone) is shifted down-field to H_ap (H_a after polymerization) at 5.18 ppm. The polymer conversion (PC%) representing the percentage of lactone 20 monomer that undergoes polymerization, and the degree of polymerization (DP) representing the number average length of lactone 20, indicated a PC of 75.23% and an average DP of 19.2. Based on the average DP, the molecular weight (Mn) of the synthesized polymer was estimated to be 2.23×10³ g mol⁻¹.

Scheme 2:

Illustration of ROP of lactone 20 using BnOH as an initiator and MSA as a catalyst.

Figure 2:

¹H NMR spectra of lactone 20 polymer (crude) after 24 h of reaction.

The molecular weight was further assessed by size exclusion chromatography (SEC) (Figure 3). In agreement with the NMR results, SEC data showed that the produced polymer has the following molecular weight: Mn 1.38×10³ g mol⁻¹ and Mw 2.16×10³ g mol⁻¹ with an intermediate polydispersity of Ð 1.56. The SEC results showed a minor difference from the NMR data. In fact, molecular weight measured by SEC was confounded because SEC was most sensitive to the hydrodynamic volume of the polymer; ROP added weight to the molecule but did not necessarily increase the hydrodynamic volume in proportion to the increase in the length of the polymer. Our data in combination revealed that ROP is an efficient method to polymerize bicyclic terpene-based lactones, which could serve as green substituents to other aliphatic polyesters.

Figure 3:

Size exclusion chromatography of lactone 20 polymer (crude) after 24 h of reaction.

3.1 Materials

Norcamphor 98% (CAS 497-38-1) was provided by Sigma-Aldrich (St. Louis, MO, USA). Kanamycin sulfate, Isopropyl β-d-1-thiogalactopyranoside, nicotinamide adenine dinucleotide phosphate, and flavine adenine dinucleotide were purchased from Sigma-Aldrich (St. Louis, MO, USA). Methanesulfonic acid (MSA, CAS 75-75-2) was provided by Sigma-Aldrich (St. Louis, MO, USA). Benzyl alcohol anhydrous, 99.8% (CAS 100-51-6) from Sigma-Aldrich (St. Louis, MO, USA). NMR spectra were recorded using Bruker Avance III 400 MHz spectrometer (Bruker, Billerica, MA, USA). SEC is performed using A TOSOH EcoSEC HLC-8320GPC system (Tokyo, Japan) equipped with an EcoSEC RI detector and three PSS PFG 5 µm columns (microguard, 100 Å, and 300 Å). The GC/FID and GC/MS analysis were performed on a GCMS-QP2010 Ultra (Shimadzu, Kyoto, Japan) equipped with an AOC-20i autoinjector (Shimadzu, Kyoto, Japan). Rxi-5ms capillary columns (30 m×250 μm×0.25 μm, Restek, Bellefonte, PA, USA).

3.2 Bacterial strain, plasmids, and media

The in-house available gene of the BVMO (cyclohexanone monooxygenase) from A. calcoaceticus, previously cloned in the plasmid pET28a(+) vector containing an N-terminal His₆-tag, was transformed into E. coli BL21(DE3) for expression. The cells were grown in 2×YT medium (16 g L⁻¹ tryptone, 10 g L⁻¹ yeast extract, 5 g L⁻¹ sodium chloride) containing 40 μg mL⁻¹ kanamycin sulfate (Sigma, St. Louis, MO, USA). Cell density at OD₆₀₀ was determined using a plate reader SpectraMax i3x (Molecular Devices, San José, CA, USA).

3.3 Recombinant expression and preparation of cell lysate

Freshly transformed E. coli BL21(DE3) cells were inoculated in 1 mL 2× YT medium containing 40 μg mL⁻¹ kanamycin sulfate in 96-deep well plates, and were cultivated overnight at 37 °C and 200 rpm. The overnight culture was used to inoculate 3 mL of fresh medium to OD₆₀₀=0.1 and incubated at 37 °C and 200 rpm until an OD₆₀₀=0.6–0.8 was reached. Induction was carried out for 20 h at 25 °C and 180 rpm using 0.05 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) (Sigma-Aldrich, USA). Cells were harvested by centrifugation at 2276×g for 15 min at 4 °C on a Thermo Scientific Sorvall ST16R centrifuge coupled to a M20 rotor (Waltham, MA, USA). Cell lysis was performed with B-PER Complete Bacterial Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, MA, USA) using 5 mL g⁻¹ to resuspend the cell pellet. The solution was incubated at room temperature for 15 min under mixing (750 rpm) to lyse the cells. The cell debris were removed by centrifugation at 2276×g for 30 min at 4 °C.

For high-scale conversion, a preculture of 30 mL 2× YT medium containing 40 μ g mL⁻¹ kanamycin sulfate was incubated overnight at 37 °C and 200 rpm. The overnight culture was diluted in 200 mL of fresh medium to OD₆₀₀=0.1 and incubated at 37 °C and 200 rpm until an OD₆₀₀=0.6–0.8 was reached. Induction and cell harvesting were performed as described above. Tris-buffer pH 8.5 and 50 mM (5 mL) was used to resuspend the wet cell pellet (1 g). The suspension was sonicated for 1 min with 1 s pulse and 2 s pause (total time 3 min), 60% duty cycle under ice cooling with Misonix sonifier cell disruptor ultrasonic S-4000 probe (MIsonix Inc., Farmingdale, NY, USA). Cellular debris were removed by centrifugation at 40,000×g and 4 °C for 20 min.

3.4 Protein analysis

The protein concentration was measured by Bradford using the Bio-Rad protein assay kit (Bio-Rad, Hercules, CA, USA) with bovine serum albumin as the protein standard. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed according to Laemmli [68], using a 4% stacking and 15% separating gels, purchased from Bio-Rad. SeeBlue Plus2 prestained protein standard was purchased from Thermo Fisher Scientific and used as a molecular weight marker. Proteins were stained using InstantBlue (Expedeon, Heidelberg, Germany) .

3.5 BVMO activity by NADPH measurements

The measurement of the activity of the E. coli cell lysate containing CHMO_Acineto_QM was performed spectrophotometrically, using the Spark TECAN plate reader (Männedorf, Switzerland). A total of 20 μL of cell lysate was mixed with 200 μL reaction mixture (50 mM Tris-HCl pH 8.5, 0.6 mM cyclohexanone or norcamphor (60 mM in ethanol), 0.25 mM NADPH. The consumption of reduced NADPH was determined at 340 nm for 120 s in 96-well plates (using an extinction coefficient of ε_NADPH=6.4 mM⁻¹ cm⁻¹).

3.6 Chemical synthesis of lactones

A total of 4.06 g of 77% pure m-CPBA was dissolved in dichloromethane (DCM), dried over magnesium sulfate, filtered, and then concentrated under nitrogen flow. Once the m-CPBA (18.13 mmol, 2 eq) is pure, 1.8 mL of DCM and 1 g of norcamphor (0.9 mmol, 1 eq) were added to a three-neck flask to obtain a final concentration of 0.5 M. The reaction was stirred under reflux for 18 h at room temperature under nitrogen atmosphere. The organics were washed twice with sodium bisulfite and twice with sodium bicarbonate to remove excess of m-CPBA, and then dried over magnesium sulfate. The product was then purified on column using (10% EtOAc in heptane over 2CV then up to 30% EtOAc in heptane over 8CV). DCM was dried over aluminum oxide before use. Final yield was 53.2% with 532 mg of lactone recovered.

3.7 Polymerization of lactone 20

The ROP method was used to polymerize lactone 20. Briefly, 128 mg of lactone 20 (1.02 mmol, 1 eq), 1.06 μL of benzyl alcohol (0.01 eq), and 1.32 μL of MSA (0.02 eq) were poured into a 2 mL vial containing 0.5 mL toluene and mixed at 75 °C under nitrogen flow. The reaction was stopped after 24 h and MSA activity was quenched by adding 1.8 μL pyridine. The crude specimens were used for analysis.

PC (%) and DP were estimated by ¹H NMR based on the peak intensity of the corresponding signals according to the following equation:

PC (%)=Lactone 20(polymerized)Lactone 20(total)×100

=I(Hap)I(Hap+Ham )×100

DP=Lactone 20 (polymerized)BnOH(terminal)=2×I(Hap)I(Hδ)

where (I) is the integral (area) of the selected proton and Hδ are the two protons of the carbon linked to alkoxy oxygen atom in BnOH after polymerization.

3.8 Nuclear magnetic resonance

NMR spectra (¹H, ¹³C, homonuclear correlation spectroscopy (¹H-¹H COSY), and heteronuclear single-quantum correlation (HSQC)) were recorded using Bruker Avance III 400 MHz spectrometer (Billerica, MA, USA), equipped with a 5 mm multinuclear broad band probe (BBFO+) and z-gradient coil. The samples (~20 mg) were dissolved in d₁-chloroform (~1 mL) and the spectra were recorded at room temperature and with 16 scans for 1D ¹H, 256 scans for 1D ¹³C, and 2 scans for 2D ¹H-¹H COSY and 2D ¹H-¹³C HSQC. Chemical shifts (in ppm) were expressed relative to the resonance of chloroform (δ=7.26 ppm).

3.9 Size exclusion chromatography

The molecular weight of the synthesized lactone 20-based polymer was determined by SEC analysis. A TOSOH EcoSEC HLC-8320GPC system (Minato City, Tokyo, Japan) was used equipped with an EcoSEC RI detector and three PSS PFG 5 μm columns (microguard, 100 Å, and 300 Å). Polyethylene glycol standards were used for calibration, and toluene was used as internal standard. DMF was used to dissolve 20 crude.