An Improved Scalable Synthesis of α- and β-Amyrin

The synthesis of α- and β-amyrin was accomplished starting from easily accessible starting materials, oleanolic, and ursolic acid. The procedures allow the preparation of β-amyrin in an exceptionally short scalable manner via selective iodation and reduction. For α-amyrin, a different synthetic approach had to be chosen providing access to α-amyrin in medium-to-large scale.


Introduction
Pentacyclic triterpenoids of the amyrin family ( Figure 1) represent a class of the most abundant secondary metabolites in plants sharing plant protective properties [1,2]. Among their high value as antimicrobial [3] and antifungal [4] agents in plants, they share pharmacologically relevant properties, and have proven anti-inflammatory [5][6][7], anti-nociceptive [8], and gastro-protective [9] activities, and inhibit HIV transcriptase-1 [10]. In addition, studies show an increasing biological activity by derivatization. As a consequence, this compound class still receives interest, and offers possible candidates for enhanced bioactivity [11].

Introduction
Pentacyclic triterpenoids of the amyrin family ( Figure 1) represent a class of the most abundant secondary metabolites in plants sharing plant protective properties [1,2]. Among their high value as antimicrobial [3] and antifungal [4] agents in plants, they share pharmacologically relevant properties, and have proven anti-inflammatory [5][6][7], anti-nociceptive [8], and gastro-protective [9] activities, and inhibit HIV transcriptase-1 [10]. In addition, studies show an increasing biological activity by derivatization. As a consequence, this compound class still receives interest, and offers possible candidates for enhanced bioactivity [11]. In spite of belonging to the most abundant secondary metabolites, the problem for accessing larger quantities of α-amyrin (1) and β-amyrin (2) persists [12]. Amyrins represent a precursor for more complex triterpenoids like boswellic acid (found in frankincense) or maslinic acid (found in olives), but their concentration in plant tissue is rather low compared to other triterpenoids, such as oleanolic acid (OA) and ursolic acid (UA). Furthermore, their purification is challenging, especially in large scale. Over the past decades, there have been several studies on purification and identification of sources for these compounds. However, it is still an ongoing task to gain access to αand β-amyrin [6,13]. The first partial synthesis of β-amyrin was done by Barton in 1968 [14], and the first total synthesis of β-amyrin was accomplished by E.J. Corey in 1993 [15]. Different synthetic approaches also targeted the β-amyrin and triterpenoids of the β-amyrin family [16][17][18]. Until now, In spite of belonging to the most abundant secondary metabolites, the problem for accessing larger quantities of α-amyrin (1) and β-amyrin (2) persists [12]. Amyrins represent a precursor for more complex triterpenoids like boswellic acid (found in frankincense) or maslinic acid (found in olives), but their concentration in plant tissue is rather low compared to other triterpenoids, such as oleanolic acid (OA) and ursolic acid (UA). Furthermore, their purification is challenging, especially in large scale. Over the past decades, there have been several studies on purification and identification of sources for these compounds. However, it is still an ongoing task to gain access to αand β-amyrin [6,13]. The first partial synthesis of β-amyrin was done by Barton in 1968 [14], and the first total synthesis of there has been no publication describing the total synthesis of α-amyrin. A recent partial synthesis, starting from easily accessible starting materials-oleanolic acid (OA) and ursolic acid (UA)-made α-amyrin, β-amyrin, and lupeol accessible [19].
Our aim was to improve the synthetic route to gain access to larger quantities of α-and β-amyrin to enable further derivatization and research.

Unsuccessful Approaches
The first approaches of a direct reduction of oleanolic and ursolic acid by reaction with tris(pentafluorophenyl)borane and alkylsilanes did not afford the desired reaction product [20,21]. Likewise, an Arndt-Eistert reaction followed by decarboxylation was not successful, because we were not able to obtain the desired products.

High Yield Preparation of β-Amyrin
A straightforward approach, previously already used in the reduction of decahydronapthalenes, led to good results ( Figure 2) [22]. Thus, starting from oleanolic acid (OA) its reduction with lithium aluminum hydride afforded the corresponding alcohol 3 in high yield. Therefore, the next step was carried out without any further purification. In an Appel-like reaction, a selective substitution of the primary hydroxy group with iodine proceeded in short reaction times and good isolated yields. Longer reaction times, however, led to the formation of unwanted byproducts. After purification of compound 4, a reduction with zinc powder gave β-amyrin, and again, no further purification was necessary. Therefore, over three steps, we were able to convert oleanolic acid to β-amyrin in an 81% overall yield. This represents a good improvement to the previously reported 42% [19]. Unfortunately, this approach did not work with ursolic acid; due to its limited reactivity on C-28, the substitution reaction with iodine at C-28 did not proceed well.

Preparation of α-and β-Amyrin
Therefore, we decided to introduce a different leaving group than iodine. Unfortunately, a protection of the hydroxy group at C-3 was necessary ( Figure 3). For oleanolic acid methansulfonic esters (12) represented a suitable leaving group and resulted in good yields in the reduction to follow, to yield the corresponding alkane 14. However, it was not possible to carry out the same protocol with ursolic acid as a starting material. It was already known that the shift of a methyl group in the α-amyrin type triterpenoids reduces the reactivity at the carboxylic group of ursolic acid compared to the β-amyrin type oleanolic acid. This holds true for the corresponding alcohol 9 as well. It was not possible to reduce the mesylate with lithium aluminum hydride to alkane 13. Different leaving groups, such as iodine or tosylates, could not be introduced. With trifluoromethanesulfonic anhydride, however, it was possible to generate triflate 11. This triflate (stable at −20 °C for a couple of days in the dark) could be reduced with lithium aluminum hydride to yield the TBDMS-protected

Preparation of αand β-Amyrin
Therefore, we decided to introduce a different leaving group than iodine. Unfortunately, a protection of the hydroxy group at C-3 was necessary ( Figure 3). For oleanolic acid methansulfonic esters (12) represented a suitable leaving group and resulted in good yields in the reduction to follow, to yield the corresponding alkane 14. However, it was not possible to carry out the same protocol with ursolic acid as a starting material. It was already known that the shift of a methyl group in the α-amyrin type triterpenoids reduces the reactivity at the carboxylic group of ursolic acid compared to the β-amyrin type oleanolic acid. This holds true for the corresponding alcohol 9 as well. It was not possible to reduce the mesylate with lithium aluminum hydride to alkane 13. Different leaving groups, such as iodine or tosylates, could not be introduced. With trifluoromethanesulfonic anhydride, however, it was possible to generate triflate 11. This triflate (stable at −20 • C for a couple of days in the dark) could be reduced with lithium aluminum hydride to yield the TBDMS-protected α-amyrin-ether 13. This approach yielded α-amyrin in 64% overall yield, while β-amyrin was formed in around 69% overall yield. The latter yield is a significant improvement to known procedures for the synthesis of β-amyrin [19]. Upon scaling up of the reactions by a factor of 5, the overall yield drops slightly (62 and 66%, respectively).
Molecules 2018, 23, x FOR PEER REVIEW 3 of 11 α-amyrin-ether 13. This approach yielded α-amyrin in 64% overall yield, while β-amyrin was formed in around 69% overall yield. The latter yield is a significant improvement to known procedures for the synthesis of β-amyrin [19]. Upon scaling up of the reactions by a factor of 5, the overall yield drops slightly (62 and 66%, respectively).

α-Amyrin
Procedure 2: Compound 4 (4 g, 7.24 mmol) was dissolved in acetic acid (1 L) at 40 • C. Zinc (2.5 g, 38.2 mmol) was added, and the reaction was stirred for 4 h at room temperature. The zinc was filtered off, and the filtrate was concentrated under reduced pressure at 40 • C. As soon as β-amyrin began to crystalize, the product was precipitated by adding water (1 L). The precipitate was filtered off and washed thoroughly with water. The white crystalline solid was dried to furnish 2 as a white crystalline solid (   (3β) Methyl 3-hydroxyurs-12-en-28-oate (5) Ursolic acid (20.0 g, 0.044 mol) was dissolved in DMF (250 mL) and potassium carbonate (30.0 g, 0.220 mol) was added. The mixture was stirred for 30 min, and methyl iodide (5 mL, 0.080 mol) was added. After stirring for 12 h at room temperature, the product was precipitated from the mixture by adding water (1 L). The solid was filtered off and washed with 2 M hydrochloric acid (2 × 20 mL), water (2 × 20 mL), and the resulting product was dried in a desiccator yielding 5 as a colorless crystal solid (19. ν = 3446br, 2946m,  2926m, 2870w, 1724w, 1636m, 1456w, 1384w, 1232w, 1200w, 1166w, 1144w, 1112w, 1092w, 1044m, 1032m (3β) Ethyl 3-hydroxyolean-12-en-28-oate (6) Oleanolic acid (10.0 g, 0.022 mol) was dissolved in DMF (200 mL) and potassium carbonate (15.0 g, 0.110 mol) was added. The mixture was stirred for 30 min, and ethyl iodide (3.5 mL, 0.044 mol) was added. After stirring for 12 h at room temperature, the product was precipitated by adding water (800 mL). The product was filtered off and washed with 2 M hydrochloric acid (2 × 20 mL), water (2 × 20 mL); drying in a desiccator gave 6 as a colorless crystalline solid ( ν= 3446br, 2946m, 2868w, 1722m, 1636w, 1462w,  1386w, 1364w, 1260w, 1178m, 1162m, 1124w, 1094w, 1064w, 1036m   Compound 5 (19.0 g, 0.040 mol) was dissolved in dry DMF (300 mL), and imidazole (7.0 g, 0.103 mol) and TBDMSCl (6.8 g, 0.044 mol) were added. The reaction mixture was stirred for 24 h at 50 • C. The product was precipitated by adding water; it was filtered off and washed with 2 M hydrochloric acid (2 × 20 mL) and water (2 × 50 mL). The resulting solid was dried in a desiccator to yield 7 as a colorless crystalline solid (  Compound 6 (10.0 g, 0.021 mol) was dissolved in DMF (100 mL), imidazole (3.5 g, 0.520 mol) and TBDMSCl (3.5 g, 0.045 mol) were added. The reaction mixture was stirred for 24 h at 50 • C. The product was precipitated by adding water, which was filtrated off and washed with 2 M hydrochloric acid (2 × 20 mL) and water (2 × 50 mL), and dried in a desiccator to yield 8 as a colorless crystalline solid (11.83 g, 94%) [30] (9) A solution of 7 (20 g, 0.03 mol) dissolved in dry THF (150 mL) was added to a solution of lithium aluminum hydride (5 g, 0.13 mol) in dry THF (300 mL) at 0 • C. The reaction mixture was allowed to warm to room temperature and finally heated for 3 h under reflux. Surplus lithium aluminum hydride was deactivated by adding THF/water (1:1) at 0 • C. An aqueous solution of sodium hydroxide (20 mL) was added, and the mixture was stirred for 15 min. The precipitate was filtered off and washed with diethyl ether (2 × 100 mL). The aqueous phase was extracted with diethyl ether (2 × 50 mL) and the combined organic extracts were washed with brine and dried over magnesium sulfate. After filtration, the solvent was removed under reduced pressure to yield crude 9. Further purification with column chromatography afforded (silica gel, hexane/EtOAc 9:1) 9 as a colorless crystalline solid (18.  A solution of 8 (10.0 g, 0.017 mol) in dry THF (100 mL) was added slowly to a solution of lithium aluminum hydride (3.0 g, 0.080 mol) in dry THF (150 mL) at 0 • C. The solution was heated under reflux until TLC showed the reaction to be complete. The mixture was cooled to 0 • C and surplus of LiAlH 4 was deactivated by adding THF/water (1:1). An aqueous solution (2 M) of sodium hydroxide (10 mL) was added, and the mixture was stirred for 15 min. The solid was filtered off, washed with diethyl ether (2 × 50 mL), and the organic phase was washed with brine (25 mL). The aqueous phase was washed with diethyl ether (2 × 50 mL). The combined organic phases were dried over magnesium sulfate