Pinene-Based Oxidative Synthetic Toolbox for Scalable Polyester Synthesis

Generation of renewable polymers is a long-standing goal toward reaching a more sustainable society, but building blocks in biomass can be incompatible with desired polymerization type, hampering the full implementation potential of biomaterials. Herein, we show how conceptually simple oxidative transformations can be used to unlock the inherent reactivity of terpene synthons in generating polyesters by two different mechanisms starting from the same α-pinene substrate. In the first pathway, α-pinene was oxidized into the bicyclic verbanone-based lactone and subsequently polymerized into star-shaped polymers via ring-opening polymerization, resulting in a biobased semicrystalline polyester with tunable glass transition and melting temperatures. In a second pathway, polyesters were synthesized via polycondensation, utilizing the diol 1-(1′-hydroxyethyl)-3-(2′-hydroxy-ethyl)-2,2-dimethylcyclobutane (HHDC) synthesized by oxidative cleavage of the double bond of α-pinene, together with unsaturated biobased diesters such as dimethyl maleate (DMM) and dimethyl itaconate (DMI). The resulting families of terpene-based polyesters were thereafter successfully cross-linked by either transetherification, utilizing the terminal hydroxyl groups of the synthesized verbanone-based materials, or by UV irradiation, utilizing the unsaturation provided by the DMM or DMI moieties within the HHDC-based copolymers. This work highlights the potential to apply an oxidative toolbox to valorize inert terpene metabolites enabling generation of biosourced polyesters and coatings thereof by complementary mechanisms.

. Schematic structure of the FDCA-containing copolymers. The ratio between and FDCA and DMI and DMM was set to 1:1 respectively.

Polymerization of the abnormal verbanone lactone
To obtain a polymer based on the abnormal verbanone lactone, a small amount of ethanol was added to the mixture of the normal and abnormal lactone and the temperature was set to 50 °C to polymerize the more unstable abnormal lactone out of the mixture (Figure S2a: 1 H-NMR of the mixture obtained from distillation). The polymerization was conducted for about 6 hours. The polymer could then be precipitated to separate the polymer based on the abnormal lactone from the mostly unreacted normal verbanone lactone. The polymer was then precipitated a second time to remove residual monomers and characterized by 1 H-NMR and SEC. The 1 H-NMR of the polymer is shown in Figure S2b. To calculate the theoretical molecular weights, Mn, the ratio between the epsilon protons of the repeating unit (abnormal lactone: d= 4.0 ppm and 3.6 ppm; normal lactone: 4.6) and the ethanol protons (d = 4.15 ppm, CH2) were used, resulting in an average Mn of 3100 g/mol. The 1 H-NMR further showed partial polymerization of the normal lactone. However, the isolated polymers contained more than 80% of the abnormal lactone, showcasing the potential to separate the two lactones by polymerizing the abnormal lactone out of the distillate. To analyse the end groups of the synthesized material, the polymers were reacted with trifluoroacetic anhydride in DCM. The 1 H-NMR of the resulting end-capped polymers is shown in Figure S2c and reveals that the normal lactone was the dominant end group. It should be noted that the molecular weight in those samples occurs higher due to several washing steps that were conducted to remove the triethylamine and pyridine. The molecular weight distribution was further measured by size exclusion chromatography ( Figure S3). The obtained molecular weight Mn of the polymer was around 2.4 kg/mol, which is in good agreement to the molecular weight detected by NMR.

Condition
To investigate the thermal properties and the influence of the stereochemistry on the resulting material properties, DSC analysis was conducted on the polymer. The profiles of the cooling and the second heating cycle at different heating rates are shown in Figure S4. The polymer of the abnormal Lactone exhibited semi-crystallinity with a Tm around 150°C and a glass transition temperature around 50°C. The cold crystallization peak in the thermograms with 10 and 20°C/min in the heating curve at around 100°C is a result of the slow crystallization behaviour of the polymer most probably related to its bulky chemical structure. To visualize this effect, films were coated on a glass substrate. The glass substrate was then heated to 170°C for 10 min and rapidly cooled down by placing the glass substrate in the freezer. Afterwards the films were heated to 70°C for 30 min to allow the material to crystallize ( Figure S5).

Polyol4640
To assess the amount of ethoxy groups per molecule, polyol 4640 was reacted with trifluoroacetic anhydride in the presence of pyridine. The reaction was conducted in DCM at room temperature overnight. To stop the reaction, DCM was added, and the solution was washed twice with brine. The downfield shift of terminal methylene protons (signal around 4.5 ppm) was then used to calculate the ratio of terminal/nonterminal protons. Based on the assumption of complete functionalization of all hydroxyls, the amount of ethoxy groups was determined to be l+m+n+o=5, which is in good agreement with the nominal molecular weight reported by the manufacturer of around 360 g/mol.     Figure S11. FTIR of pVaL_2k, characteristic melamine peaks are marked. Note that for the names of experiments given in the box, pVL is synonymous to pVaL (scheme 1 in main paper). Figure S12. IR of pVaL_8k, characteristic melamine peaks are marked. Note that for the names of experiments given in the box, pVL is synonymous to pVaL (scheme 1 in main paper). Figure S13. HHDC-based films applied on glass substrate. Figure S14. FTIR of pHHDC-DMI, the C=C-signal at ~1640 cm -1 is highlighted. The post-cure signal decrease indicates that the majority of double bonds has been consumed in the curing reaction. Figure S15. FTIR of pHHDC-DMM-FDCA, the C=C-signal at ~1640 cm -1 is highlighted. Figure S16. FTIR of pHHDC-DMI-FDCA, the C=C-signal at ~1640 cm -1 is highlighted. Figure S17. HHDC based coatings on paper.

Cobb-test and contact angle analysis
Figure S18. Verbanone-based paper coatings. Note that for the names of experiments shown, pVL is synonymous to pVaL (scheme 1 in main paper).    Table 2. Pictures were taken at the end of each run. Figure S22. Contact angle measurements of HHDC based materials. Measurements were conducted for 10 seconds and CA were continuously measured. Average contact angles are reported in Table 2. Pictures were taken at the end of each run. Note that for the names of experiments shown, pVL is synonymous to pVaL (scheme 1 in main paper). Figure S23. Overview over the performed contact angle analyses. Note that for the names of experiments shown, pVL is synonymous to pVaL (scheme 1 in main paper).

Synthesis of sobrerol lactone
Building up on our study on the use of sobrerol as an intermediate building block for the synthesis of functional polymethacrylates and with the aim to explore new paths towards bio-based polyesters, we further investigated the versatility of sobrerol as an intermediate building block. Our strategy was to transform sobrerol into its lactone form, leaving the tertiary alcohol group unreacted during the polymerization and ready for post-functionalization.

Chemical Synthesis of the sobrerol lactone.
The sobrerol lactone (SobL) was synthesized in a three-step reaction. In the first step, trans-sobrerol was oxidized using manganese oxide, yielding in carvone hydrate in 80% yield. The double bond in ketone was then quantitatively reduced using hydrogenation chemistry ( Fig. S24a and b). Afterwards, 8-hydroxy-p-menthan-2-one was oxidized using the same principle as described for verbanone 2 resulting in lactone in 70% yield after flash chromatography, respectively (Fig. S24c). Surprisingly, the produced lactone was not the 7-membered SobL but the respective five membered rearrangement product. The rearrangement was quantitative and occurred in both the chemical and the enzymatic transformation. The 5-membered lactone was identified by the too low shift for the alpha proton of the 7-membered lactone (3.6 ppm), as well as the two peaks originating from the endocyclic methylene groups of the 5-membered lactone (1.45 and 1.3 ppm), which show up as one peak in previous intermediates. To further proof the existence of the 5-membered lactone, the tertiary alcohol was reacted with trifluoroacetic anhydride. The resulting shift of the proton neighbouring the tertiary alcohol and the shift of the exocyclic methylene group clearly identified the 5 membered isomers ( Figure S25). To prevent this isomerization from happening, one method could be to protect the tertiary alcohol prior to the Baeyer-Villiger reaction and deprotect again after polymerization. This could potentially lead to a built-in trigger for degradation of the polymer. However, it should be considered that once polymerized, the ring strain of the lactone cannot act as driving force for the rearrangement, which potentially prevents the polymer units from degrading even post deprotection. The 5-membered lactone was attempted to polymerize but results are premature and will therefore be part of future publications. Figure S25. 1 H-NMR of SobL (up) and the crude reaction product of the reaction with Trifluoroacetic anhydride. Arrows used to highlight relevant shifts.

Enzymatic synthesis of sobrerol lactone
Expression and extraction of XenB One E. coli colony harboring XenB in plasmid pGAS-XenB-C-His6 was picked to prepare overnight culture in 2YT media with 100 μg/ml ampicillin. The next day, a second culture was prepared from the overnight culture with a start OD600 of 0.1. Once the OD600 reached 0.6, 0.1 mM IPTG (final concentration) was added to induce the expression of XenB that was set for 20 hours at 25 °C 180 rpm. Harvesting of the cells, extraction of XenB, confirmation of expression, and protein concentration assessment was performed in the same way as described for CHMOActineto in Materials and Methods.

Biotransformation
The biocatalytic conversion of unsaturated ketone derivative of sobrerol 9, to lactone was carried out in closed 20 ml glass vials, at 30°C  Scheme S1. Biotransformation of unsaturated ketone derivative of sobrerol to lactone. XenB reduces the unsaturated ketone, CHMOAcineto (CHMO_QM) performs Baeyer-Villiger oxidation of the saturated ketone, transforming it into the seven-member ring lactone which spontaneously rearranges to its five-member ring structure. Figure S26. GC-MS chromatogram of extract of biotransformation from unsaturated ketone derivative of sobrerol to fivemember ring lactone. At the bottom are the chemically synthesized standards; saturated ketone (8.8 min), unsaturated ketone (9.5 min), and five-member ring lactone (11.4 min). The other peaks were also visible in empty cell control.