Solvent-Driven Selectivity on the One-Step Catalytic Synthesis of Manoyl Oxide Based on a Novel and Sustainable “Zeolite Catalyst–Solvent” System

Application of a novel “zeolite catalyst–solvent” system for the sustainable one-step synthesis of the terpenoid manoyl oxide, the potential precursor of forskolin and ambrox. Manoyl oxide high-yield and large-scale production over a zeolite catalyst has been infeasible so far, while this system results in 90% yields at 135 °C and atmospheric pressure. Substrate-controlled methodology is used to achieve selectivity. Solvent-driven catalysis is shown, as the activation energy barrier decreases in the presence of appropriate solvents, being 62.7 and 93.46 kJmol−1 for a glyme-type solvent and dodecane, respectively. Finally, catalyst acidity is key parameter for the process.


Introduction
Cyclic ethers are fundamental heterocycles of bioactive compounds, natural products, etc. [1][2][3]. Therefore, the use of cyclodehydration reactions is of significant importance in the chemical industry.
The cyclic ether manoyl oxide (MO) is a compound that belongs to labdane diterpenes and is known to exhibit a number of important properties. MO is a compound of general formula C 20 H 34 O and a mixture of two C13 epimers; 13R-MO (compound 1 in Scheme 1) and 13S-MO (compound 2 in Scheme 1) [4][5][6].
Both C13 MO epimers are valuable compounds as they have been proven to be an antibacterial agents against Staphylococcus aureus, Pseudomonas aeruginosae and Klebsiella pneumonia [7], as well as to have antimicrobial activity against Borrelia burgdorferi [8] and against Gram-positive bacteria (B. cereus, Str.faecalis, St. aureus, St. epidermidis, and B. subtilis) [9].
Furthermore, 13R-MO has been reported as a potential precursor of two valuable compounds forskolin [6,10,11] and ambrox [12][13][14] (Scheme 2). These two compounds have received much attention for their broad range of applications [13][14][15][16][17][18][19][20][21]. In particular, the limited availability of pure forskolin is currently hindering its full utilization; therefore, a new and sustainable strategy is needed for forskolin production. Recently, the entire biosynthetic pathway leading to forskolin was elucidated. A key process of this pathway include the biosynthesis of 13R-MO from GGPP (geranylgeranyl diphosphate), as 13R-MO consists the diterpene backbone of forskolin [10,22,23]. Then 13R-MO is further bioconverted to forskolin [24,25]. Recent work has improved the efficiency towards forskolin biosynthesis in yeast from 13R-MO [25]. However, the optimization of the first step of the process i.e. the synthesis of 13R-MO, still remains a challenge. Maybe a solution to this challenge would be to replace the total biosynthetic pathway for the synthesis of Forskolin, by a process that combines a chemical pathway for the synthesis of 13R-MO, followed by a biosynthetic one for the synthesis of Forskolin from 13R-MO.
A number of synthetic procedures for the production of MO have been reported. Most of them start from sclareol (compound 3 in Scheme 1), which is a diterpene found in Clary Sage (Salvia sclarea). However, the already developed processes lead to relatively low yields of MO or comprise multistep procedures [26,27].
Prior art includes also patents on the production of MO or cycloether derivatives by the cyclodehydration of diols, such as US10208326B2 [28] and US9469622B2 [29], respectively. Nevertheless, in US10208326B2 [28] a biosynthetic method of 13R-MO is proposed where long preparation/ processing times are involved, while US9469622B2 [29] has the limitation that either high temperature water (HTW) conditions should be followed, or at least one of the hydroxyl groups of the diol must be non-tertiary and therefore production of MO from sclareol is not possible according to the means of this patent, or long reaction times are required (30 h). In addition, in the process described by Alvarez-Manzaneda EJ, Chaboun R., Alvarez E. et al. cerium (IV) ammonium nitrate (CAN) is utilized for the synthesis of MO from sclareol, but reaction times up to 10 h and stoichiometric amounts of CAN are required [30]. Previous work also includes processes were 90% yield in MO is achieved, by using trifluoroacetic acid in dioxane solution [31]. The drawbacks of such process include the employment of corrosive reagents as trifluoroacetic acid, while dioxane is classified from ECHA (Europpean Chemicals Agency) as a substance suspected of causing cancer. Finally, recent research includes the production of a 13R-MO yeast cell factory from glucose, but such process requires metabolic engineering strategies and long fermentation times [32].
Recently in EP3409663A1 [33] we have presented a new "zeolite catalyst-solvent", which can transform the diol sclareol to the cyclic ether MO with excellent selectivity in a one-step process over a mordenite catalyst. This novel system lead to yields in MO higher than 85%.
Solid catalysts have the advantage of being easily separated from the reaction mixture. However, a challenge of solid catalysts is that sometimes their various active sites can accelerate side reactions leading to undesired by-products. One solution to this problem could be the use of a solvent-controlled methodology to obtain high selectivity. In particular, fine chemicals synthesis over heterogeneous catalysts sometimes involves the use of solvents that can strongly influence the catalytic performance [34][35][36][37]. In order to find the most appropriate solvent, it is essential to examine extensively the relationship between the chemical nature of the solvents and their interaction with the catalyst and the substrate [38][39][40].
Herein, we provide a step forward towards EP3409663A1 [33] and propose that solvent nature is a key parameter for the process, as a substrate-controlled methodology is used to achieve selectivity.
In particular, a zeolite acid catalyst (TECHNOSA-H2) prepared by acid activation of natural mordenite from volcanic soils [41,42] was combined with various solvents. The efficiency of the corresponding "zeolite catalyst-solvent" systems for the dehydration of sclareol to MO was thoroughly investigated. The solvents examined were of different boiling points and polarities, in order to not only identify the most appropriate solvent from a yield standpoint, but also understand the mechanistics of the aforementioned system.
Considering the sustainability of the process, reaction conditions were optimized to develop a cost-effective and low energy-consuming method for providing high yield of MO by an easily applicable in industrial scale one-step process. In addition, in frame of utilizing renewable raw materials for chemical synthesis, the employment of acid treated natural mordenite as catalyst offers a low-cost and renewable catalyst alternative for the studied reaction.
Finally, the effects of catalyst's surface area and acidity were also inspected by testing two additional catalyst samples suitably modified to have different surface areas and acidities.

Catalyst Preparation and Characterization
Natural mordenite was stirred with distilled water for 2 h in a solid mass to liquid volume ratio 1 g/3 mL and the undesirable materials were removed by filtering. The slurry was dried for 2 h at 90 °C and then it was grinded to become fine-grained. This material was treated with 2 M acidic solutions of HCl, and CH 3 COOH (all purchased by Fisher scientific) at 70 °C for 4 h and mass solid to solution volume 1 g/20 mL. Full details on the acid activation process have been reported previously [41,42]. The symbols TECHNOSA-H2, and TECHNOSA-A2 stand, respectively, for the catalysts resulted by activation of natural mordenite with 2 M hydrochloric and acetic acid aqueous solutions. TECHNOSA-H2C catalyst was prepared by air-calcination of TECHNOSA-H2 catalyst at 500 °C for 2 h. Full characterization of TECHNOSA-H2, TECHNOSA-A2 and TECHNOSA-H2C with N 2 -physisorption, XRD, ATR-FTIR, SEM-EDS, TEM, Microelectrophoresis and Equilibrium pH measurements have been reported elsewhere [42,43]

Catalytic Conversion of Sclareol
One gram sclareol (Mw = 308.5 g/mol), 3 g of the selected glyme solvent and catalyst (0.9% mass ratio of sclareol) were combined in a round bottom flask equipped with a condenser. The flask was added in a preheated oil bath at the determined temperature and the mixture was stirred vigorously at a predetermined time. At the end of the reaction, the mixture was cooled at 25 °C, hexane was added, and the mixture was filtered to remove the catalyst. The filtrate was double washed with water and dried over anhydrous sodium sulfate. Finally, hexane removal was performed by condensation. The isolated product gave 0.84 g MO (MW = 290.5 g/ mol). In other tests, the reaction was also performed in a solvent-free system. The same procedure as above was followed except from the initial addition of the solvent.

Analysis
The reaction mixture was analyzed by gas chromatography-mass spectrometry (Shimadzu GCMS-QP2010 Ultra equipped with capillary column: 20 m × 0.18 mm × 0.18 μm (using He as carrier gas), in addition to gas chromatography (AGILENT Technologies 6890 N with FID detector, using a Varian capillary column CP Sil 5 CB 25 m × 0.25 mm × 0.25 μm and He as the carrier gas). The identification of all compounds was performed using the mass spectra libraries of VIORYL S.A. and where necessary confirmation with analytical standards. All activity evaluation experiments were repeated at least three times and the standard error was found lower than ± 3%.

Solvent-Driven Selectivity and Mechanism
Although, yields up to 50% to MO can be obtained in a solvent-free system (Table 1), specific solvents were found to strongly influence the reaction results. For the investigation of solvent applicability in the catalytic conversion of sclareol to MO, several polar and non-polar solvents have been tested at the same reaction conditions (reaction time = 10 min, T = 66 °C), as shown in Table 1. The reaction temperature of 66 °C was chosen, as it constitutes the lowest boiling point temperature of the tested solvents.
In the tests with dimethylformamide, no reaction took place as the solvent blocked the reactivity of the acid sites of the catalyst. Therefore, basic polar solvents are considered as inappropriate solvents for the investigated "zeolite catalyst-solvent" system.
Inspection of Table 1 shows that the higher selectivities in MO were achieved when sclareol is initially dissolved in a glyme-type solvent i.e. diglyme, triglyme, tetraglyme. Glymes, i.e. glycol diethers, are saturated polyethers containing no other functional groups. They are aprotic solvents with high chemical stability [47]. Glymes have a variety of industrial applications and their ability to form hydrogen bonds with the reactants/intermediates of several chemical processes has been extensively studied [47,[49][50][51].
Taking into account that: (i) sclareol dehydration is highly accelerated over our acid catalyst; (ii) a glyme-type solvent, Scheme 3 Mechanistic proposal for the catalytic dehydration of sclareol over TECHNOSA-H2 catalyst. Pathway (1): cyclization of sclareol 3 to produce manoyl oxide 1 + 2. Pathway (2): partial dehydration of sclareol 3 to produce manool 4, labd-7,14-dien-13-ol 5, abienol 6 and labd-7,13-dien-15-ol 7, while further dehydration of these alcohols leads to labd-7,12,14-triene 8 and sclarene 9 which can interact with sclareol via hydrogen bonds formation, increases the MO selectivity, while the non-hydrogen bond forming solvents don't; and (iii) the variety of the partial dehydration products of Scheme 3 obtained over the TECHNOSA-H2 catalyst indicating that water elimination starts either from OH group being in the ring of sclareol or from the other one being in the side chain, we can propose the catalytic mechanism illustrated in Scheme 4. According to this mechanism when "TECHNOSA-H2-Glyme" system is applied (Scheme 4), formation of hydrogen bonds between the glyme solvent and OH groups of sclareol takes place. The formation of these hydrogen bonds changes the configuration of sclareol in order to minimize the distance between its hydroxyl groups. Then an acidic proton from the catalyst surface attacks to the C 13 oxygen of sclareol. As follows, the protonated hydroxyl group is substituted by the C 8 oxygen and leads to the production of MO along with the formation of one water molecule. The respective procedure follows when the acidic proton attacks to the C 8 hydroxyl group of sclareol. Based on the reaction results, it is seen that there is not any preference in the attack of the acidic proton of TECHNOSA-H2, but it randomly attacks C 8 or C 13 oxygen of sclareol. This is proved by the ratio of the C 8 and C 13 alcohols that are each time produced which differs from test to test.
From the process described in Scheme 4 the final product is a mixture of the C13 isomers of MO. This is in line with previous research [26] where it is shown that the initial C13 configuration of sclareol does not limit the synthesis of 13R-or 13S-MO. Scheme 4 Proposed mechanism using the "TECHNOSA-H2-Glyme" system for the conversion of sclareol (3) to manoyl oxide (1 + 2) Fig. 1 Conversion of sclareol and yield of manoyl oxide (MO) obtained over TECHNOSA-H2 catalyst, using triglyme as solvent at various reaction times (temperature = 135 °C, catalyst weight = 0.9% weight of sclareol)   (Fig. 2) or a non-polar solvent (Fig. 3). Thus, one can make a direct comparison of the effects of the aforementioned types of solvents on the title reaction. The solvents that were chosen for this comparison were triglyme and dodecane, since they both have high boiling point values and they represent polar and non-polar solvents, respectively. The results obtained confirmed the successful cooperation of a glyme-type solvent with the TECHNOSA-H2 catalyst in the whole reaction temperature range studied. Taking into account that the highest MO yield was achieved at ~ 135 °C one can conclude that a glyme-type solvent (like triglyme) with boiling point higher than 135 °C is a good choice for the studied reaction.

Activation Energy Calculation
Using the results presented in Fig. 2 and Fig. 3 and assuming first order kinetics up to 135 °C for triglyme and up to 145 °C for dodecane, we calculated the reaction constant (k) values at various reaction temperatures using Eq. (1) where x and t stand for yield in MO and reaction time, respectively. By the insertion of x and t for various reaction temperatures to the above equation, the energetics of the reaction pathway which leads only to MO and not to the total amount of the reaction products, can be calculated (Table S1). Introducing these values in an Arrhenius plot, we obtained a very good linear correlation as shown in Supplementary material (Fig.S1), which confirms the aforementioned first order assumption. Using the slope of the straight line obtained for the tests with triglyme as solvent, we calculated an activation energy value, E a = 62.70 ± 0.1 kJ mol −1 . On the other hand, using the slope of the straight line obtained for the tests with dodecane as solvent the activation energy value was calculated to be E a = 93.46 ± 0.16 kJ mol −1 .
From the above results, we can safely assume that indeed the use of glyme-type solvents changes the energetics of the reaction network for the conversion of sclareol, to favour the formation of the desired product. The activation energy barrier for sclareol dehydration to MO (Scheme 3: reaction pathway 1) is decreased in the presence of the glyme-type solvent (62.70 kJmol −1 ), whereas the same barrier is higher when dodecane is used as solvent (93.46 kJmol −1 ). The critical role of glyme-type solvents in the mechanistic pathways and the yield of the reaction to the desired product is therefore confirmed.
Finally, additional experiments have been performed in air-free conditions, using nitrogen as the inert gas. It was shown that such conditions do not alter the product distribution.

Investigation of the Effect of BET Surface Area and Acidity of the Catalyst
In order to investigate the influence of the surface area and the acidity of the catalyst to the reaction results, two more catalysts were prepared and studied. These are TECHNOSA-H2C (prepared via air-calcination of TECH-NOSA-H2 at 500 °C for 2 h) and TECHNOSA-A2 (natural mordenite activated with acetic acid). The properties of the catalysts tested are presented in Table 2.
The value S BET concerns all the pores of the catalyst including micropores (≤ 2 nm), mesopores (2-50 nm) and macropores (larger than 50 nm) and the external surface as well [52]. The equilibrium pH of a dense "solid/water" suspension is an easily determined parameter and indicates the acidity of the solid surface [53]. Sclareol conversion and MO yield over the catalysts TECHNOSA-H2, TECHNOSA-H2C and TECHNOSA-A2 are shown in Fig. 4.
Inspection of Fig. 4 and Table 2 reveals that the acidity of the catalyst is the key factor determining the catalytic performance. Indeed, the highest sclareol conversion and MO yield is obtained over the TECHNOSA-H2 catalyst which has the  highest acidity (lowest equilibrium pH). The aforementioned kinetic parameters are dramatically decreased over the TECHNOSA-H2C catalyst, which has the lowest acidity among the catalysts studied although it exhibits similar S BET value with its non-calcined counterpart ( Table 2). On the other hand, the TECHNOSA-A2 catalyst with considerably lower S BET value but somewhat higher acidity than the TECHNOSA-H2C catalyst exhibits slightly higher conversion of sclareol and yield of MO than those obtained over the latter catalyst.

Conclusions
According to the results presented the water elimination from sclareol is accelerated over the strong acid catalyst (TECHNOSA-H2) studied in this work. However, in absence of a solvent or in the presence of an inappropriate solvent this elimination is random leading in variety of dehydration products.
The mechanism of the catalytic dehydration reaction of sclareol was investigated and it was shown that sclareol can be dehydrated leading to either the formation of a cyclic ether which is a mixture of MO C13 epimers or the formation of several alcohols. Further dehydration of the produced alcohols leads to the synthesis of totally dehydrated compounds which are labd-7,12,14-triene and sclarene.
To control the product distribution towards the desired compound, a solvent-driven selectivity of the natural mordenite acid catalyst is proposed. The combination of TECH-NOSA-H2 catalyst with a glyme type solvent increases the yield of the cyclic ether MO. The key to success for this controllable system is the use of the solvent polarity as an effective strategy to regulate the linkage between sclareol and the acid catalyst and modify the inter-molecular distances.
The effect of the solvent was proved by comparing the product distribution when a glyme-type solvent is used, towards other polar and non-polar solvents. Finally, we can safely assume that indeed the use of glymes as solvents alters the energetics of the reaction network, since the activation energy barrier for sclareol dehydration to afford MO is decreased in the presence of a glyme solvent (62.70 kJmol −1 ), whereas the same barrier is higher using dodecane as solvent (93.46 kJmol −1 ).
Finally, the influence of the BET surface and the acidity of the catalyst was also studied. It has been shown that acidity is a key parameter of catalytic performance in the present reaction.
In contrary with prior art on the synthesis of MO where long reaction times, multistep processes or toxic reagents are involved, our new "TECHNOSA-H2-Glyme" system provides a sustainable alternative to the already mentioned processes. This new system offers high specificity substrate dehydration in a one-step process, where short reaction times (t = 10 min), low temperatures (T = 135 °C) and atmospheric pressure conditions are only involved, resulting to yields in MO of up to 90%.
Our work provides a step forward towards the sustainable synthesis of 13R-MO as a precursor of Ambrox and a promising option for application in the synthesis of Forskolin where the total biosynthetic pathway for Forskolin synthesis can be replaced by a process which combines a chemical catalytic step for the production of 13R-MO, followed by a biosynthetic pathway for the synthesis of Forskolin from 13R-MO.

Declarations
Conflict of interest Ms. D. Makarouni, Prof. C. Kordulis and Prof. V. Dourtoglou declare that there is no conflict of interest. All co-authors have seen and agree with the contents of the manuscript and there is no financial interest to report.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.