Practical and scalable synthesis of beclomethasone dipropionate

Steroidal advance intermediate 3 (Fig. 1) was selected as the starting material to synthesize 1 due to its low cost and wide availability. The first step was the regioselective opening of C₉-C₁₁ epoxide to obtain 2. We investigated the previously reported procedures using 12 M concentrated hydrochloric acid (HCl), which produced C₁₆-C₁₇ -unsaturated 4 as the major product instead of product 2 (Scheme 1) [22]. Moreover, the reaction mixture progressively turned pink indicating the formation of the eliminated by-products 4 and 5. This was confirmed by LCMS and observed even at a 1 g scale, which resulted in low and inconsistent yields.

To reduce the formation of the eliminated by-products 4 and 5, the reaction was monitored at short time intervals and optimized using dilute HCl (1–5 M) (Table 1). Initially, when 1 M and 2 M HCL solutions were used, the reactions showed low conversion even after 300 min (Table 1, entries 1, 2, 3, and 4) This could be due to the low acidic concentrations, i.e., the S_N1 mechanism to open up the epoxide was not feasible. When the concentration was increased to 3 M HCl solution, complete consumption of the starting material was observed rapidly (Table 1, entry 5). However, the yield was not optimal due to the formation of the mixture of eliminated products. Prolonging the reaction only produced more eliminated products, which resulted in a low isolated yield of 2 (Table 1, entry 6). Therefore, we decided to cool the reaction and introduce the HCl and then let the mixture return to room temperature (Table 1, entry 7). This resulted in the best yields and time profiles for the conversion of 3 to 2. Further, an increase in the concentration of HCl made more of 4 and 5, which resulted in low yields (Table 1, entries 8 and 9).

Typically, when the reactions in Table 1 was completed, the addition of the sodium bicarbonate (NaHCO₃) solution directly to the reaction mixture made 2 precipitate from the reaction mixture. Filtering of the solid was followed by washing it with deionized water, which was followed by washing by a cold solvent yielded a white precipitate of 2 in high purity. However, the filtrate consisted of significant amounts of 2 resulting in the overall reaction yield being erroneous. Due to the poor solubility of 2, chromatography was not an option and further optimization was needed to improve the yield of this reaction. In practical considerations, chloroform should be avoided in scale-up reactions due to its high cost, low availability, and the safety and environmental concerns pertaining to it [23]. Therefore, screening for a solvent was explored. Thus, we confined our investigations only to certain criteria in choosing a new solvent including the following: (i) the conversion to 2 should be as high as observed in chloroform (Table 1, entry 7), (ii) it should produce minimum amounts of byproducts 4 and 5, (iii) it must readily precipitate 2 by adding NaHCO₃ solution directly to the reaction mixture. To find out the suitable solvent that provides a conversion as high as that observed in chloroform, some commonly available solvents, such as toluene, dichloromethane (DCM), THF, ethyl acetate, and diethyl ether, were explored. To our delight, it was observed that when switching the solvent from chloroform to DCM (Table 1, entry 12) the by-product formation was minimized and the precipitate formation of 2 was elevated by the addition of NaHCO₃ solution directly to the reaction mixture. The resulting precipitate of 2 was pure enough to be used in the next reactions without any further purification. This optimized procedure was successful to produce high yields (92–97 %) and was consistent up to 1 kg scale-up reactions. This outcome is very important from the point of view of industrial-scale production of 2 as it eliminates the practical difficulties of using chloroform and the cost of additional purification steps from the total production costs. To date, this is the highest yielding process reported to obtain 2.

With high purity 2 in hand, next, we attempted the propionylation of C₂₁ and C₁₇ hydroxyl groups (Scheme 2). According to the most recent method reported, by simply adding propionyl chloride or propionyl anhydride to 2 can only propionylate C₂₁ and C₁₁ hydroxyl groups and not C₁₇ hydroxyl group [12]. Even under harsh conditions, the C₁₇ hydroxyl group did not propionylate and the reaction halted at C₂₁, C₁₁ dipropionate 7 (Scheme 2).

It has been reported that Trifluoroacetic acid and p-toulenesulfonic acid (pTSA) can be used to propionate tertiary alcohols in steroidal systems [19]. Therefore, we investigated whether these acid pair conditions can propionate the C₁₇-hydroxyl of 7. Compound 7 was subjected to the acid pair conditions which gave C₁₁, C₁₇, and C₂₁ tripropionated product 6 in moderate yield (Scheme 3). However, selective hydrolysis of the C₁₁ ester to yield 1 was impossible since C₁₇ and C₂₁ esters was also hydrolyzing (Scheme 3).

Since 3 has the C₁₁-hydroxyl already protected with the epoxide, we envisioned that the propionylation of this would lead to 1 without going through the hydrolysis of the tripropionated by-product 6. Therefore, we first reacted 3 with propionic anhydride to get the C₂₁-hydroxyl group propionylated to yield 8 (Scheme 4). Thereafter, 8 was treated with acid pair conditions to yield 9. However, this reaction produced many by-products with only trace amounts of the desired product 8. This could be due to the strongly acidic conditions forcing the epoxide to open which leads to these by-products. As our goal was to develop a short, simple, and robust method to synthesize 1 without going through a protection-deprotection sequence, we concluded that moving forward with this approach was not fruitful.

Given that all the previous attempts failed to selectively dipropionate C₁₇ and C₂₁ hydroxyl groups, we decided to use the 1,3 dihydroxyl uniqueness of substrate 2 to get the desired product, i.e., product 1. Therefore, we treated 2 with triethyl orthopropionate (TEOP) to obtain 10 which was converted in situ via 6-exo-tet favorable cyclization to 10 (Scheme 5) [24], [25], [26]. Notably, 10 was isolated via crystallization and subjected to 3 M HCl conditions for 3 h. To our delight, this substate-controlled reaction to obtain C₁₇ mono propionylated product 11 produced a quantitative yield (Scheme 5).

A detailed reaction mechanism of the cyclocondensation of 2 by TEOP followed by the hydrolysis to yield 11 has been outlined in Scheme 6. We believed that the transformation from 2 to 11 could be due to the water molecule attacking the C₂₁ carbon of the oxocarbenium ion (13) (Scheme 6).

Inspired by the proposed mechanism (Scheme 6), we attempted to develop a one-pot reaction to convert 2 directly to 11 by skipping the isolation of 10. Therefore, 2 was treated with TEOP and this reaction was closely monitored via TLC. After the reaction was completed, 6 M HCl was directly added to the reaction, which turned the solution to light yellow color (Fig. 2). After completion of the reaction, the reaction mixture was neutralized with NaHCO₃ solution and worked up to get 88% yield of compound 11 (Scheme 5). The purity of 11 after the workup was high enough to proceed to the next step without further purifying it. This protocol produced 11 in high yield and was consistent with 1 kg scale-up reactions (Fig. 2). With 11 in hand, next, we attempted to selectively propionylate the C₂₁-hydroxyl group under mild conditions. We treated 11 in less reactive propionic anhydride conditions to get the much-anticipated 1, and its yield was 96 %. In all these transformations, we did not observe any chiral erosion of the intermediates.

The overall yield of 1 starting with the yield from 3 was 82 %. To our knowledge, this is the highest yield reported for the synthesis of 1. We believe that this procedure will provide a more practical alternative to the existing methods to synthesize 1 in kilo-scale as all the reactions were optimized and tested from 1 g to 1 kg scale to repeatedly give consistent yields. For the fine-chemical industries, the development of straightforward methods to prepare highly valuable synthetic intermediates from commercial and easily synthesizable chemicals is of enormous interest. Our improved methodology and detailed experimental data would, no doubt, be beneficial to the pharmaceutical and other fine chemical industries.