Bismuth(III) reagents in steroid and terpene chemistry.

Steroid and terpene chemistry still have a great impact on medicinal chemistry. Therefore, the development of new reactions or "greener" processes in this field is a contemporaneous issue. In this review, the use of bismuth(III) salts, as "ecofriendly" reagents/catalysts, on new chemical processes involving steroids and terpenes as substrates will be focused. Special attention will be given to some mechanistic considerations concerning selected reactions.

1. Introduction

Steroids [1] and terpenes [2] constitute a large and structurally diverse family of natural products and are considered important scaffolds for the synthesis of molecules of pharmaceutical interest. Different contributions ranging from natural product isolation and characterization to the synthesis of new compounds and their biological evaluation are reported every day. These investigations are justified by the well known biological properties of steroids and terpenes that make them useful molecules in pharmacy and medicine.

The literature abounds with classical chemical reactions employed in steroid and terpene chemistry. Despite their wide applicability, these transformations often suffer from disadvantages such as handling of toxic, sensitive and/or expensive reagents, difficult work-up procedures, low yields and poor selectivities. Moreover, the lack of catalytic methods to perform a great number of reactions is evident.

The growing relevance of green and sustainable chemistry and the application of its guiding principles to the development of new reactions and chemical processes is changing the face of chemistry [3,4,5,6,7]. Several strategies have been developed for leading to more efficient, sustainable and environmentally friendly chemical processes and products. Among those strategies, catalysis and the design of new processes that avoid the use of toxic reagents have been the subject of intense research. In this context, bismuth(III) salts are suitable reagents for the design of “ecofriendly catalysts”, and a large number of methods involving the use of Bi(III) compounds in organic synthesis have been reported over the last decades [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22]. Two review papers focusing on the recent advances of bismuth(III) salts usage in organic chemistry, with special emphasis on their application to the synthesis of compounds of pharmaceutical interest have quite recently been published [23,24].

In this review, the use of bismuth(III) salts as reagents and/or catalyst in reactions involving steroids and terpenes as substrates is presented. The review is organized by reaction type with emphasis on the following topics: oxidation reactions, formation and removal of common protecting groups, ring-opening of epoxides by nucleophiles, rearrangement reactions and miscellaneous reactions. Particular attention is given to works that highlight what could be called the hidden “Bi” behavior of bismuth. In fact, increasing evidence supports that the catalytic activity of bismuth(III) salts and its hydrates cannot be attributed to single Lewis acid activation of the substrates by bismuth cation, only. Indeed, the observed catalytic activities is thought to be attributable to a subtle balance between each bismuth species of Brønsted and/or Lewis acidic activation [25].

2. Oxidation Reactions

2.1. Oxidation of α-hydroxyketones by Bi2O3

The use of Bi2O3, in refluxing acetic acid was found to be an efficient and selective oxidant for the oxidation of α-hydroxyketones to the corresponding diones. According to Rigby, the true oxidant is presumably Bi(OAc)3, formed in situ under the reaction conditions, and is subsequently reduced to elemental bismuth (dark powder) [26,27]. This classical and versatile process has been extensively applied in the preparation of various steroids and terpenes, due, mainly, to its chemical selectivity. For example, this method has been useful in the synthesis of 11-hydroxy-12-keto- Δ9(11)-steroids with cholane [28,29], pregnane [30] and androstane [31] backbones. The oxidation of the α-hydroxyketone moiety present in ring A of several steroids with Bi2O3/AcOH was reported as an efficient way to obtain 2-hydroxy-3-keto-Δ1-steroids [32,33,34]. A different reactivity was observed in the Bi2O3/AcOHoxidation of the 20,21-ketol group of a 21-chloromethyl-pregnane derivative resulted in the formation of a dehydrohalogenated 20,21-diketo product, in 52% yield () [35].

Bi2O3/AcOH oxidation of the 20,21-ketol group of a 21-chloromethylpregnane derivative.

The Bi2O3/AcOH system has been applied to the preparation of an intermediate in the synthesis of cortisone starting from hecogenin, since under this reaction conditions the labile spirostan side chain remained intact () [28,29].

Bi2O3/AcOH oxidation of the 11-keto-12β-hydroxy moiety of a hecogenin derivative.

The triterpene alkaloid cevine, bearing a hemiacetal function, was converted into a 5-membered ring A hydroxy-δ-lactone product by treatment with the Bi2O3/AcOH system () [36]. Similarly, the conversion of the related triterpenes veracevine and cevagenine into similar hydroxy-δ-lactone derivatives was also reported by Kupchan and Lavie [36].

Bi2O3/AcOH oxidation of cevine.

Other examples of the application of this Bi2O3/AcOH system include the final step in the synthesis of diosphenolene [37], the preparation of α-diketo intermediate during the total synthesis of the racemate of aphidicolin [38] and the oxidation the α-hydroxyketone functionality present in ring A of highly oxygenated tetracyclic triterpenes, such as cucurbitacin D [39] and cucurbitacin B [40]. More recently, during the total synthesis of bruceantin, the Bi2O3/AcOH system was applied to the oxidation of an α-hydroxy ketone intermediate to afford the corresponding diosphenol, in 72% yield () [41].

Bi2O3/AcOH oxidation of an intermediate for the synthesis of bruceantin.

2.2. Allylic oxidation

The allylic oxidation of Δ5-steroids [42,43] using several homogeneous or heterogeneous bismuth catalysts in combination with t-BuOOH has been recently reported [44,45]. BiCl3 was found to be the best catalyst and several Δ5-steroids were converted into the corresponding Δ5-7-oxosteroids in good to high yields (). The BiCl3/t-BuOOH system proved to be very selective for this reaction, since significant epoxidation of the double bond, secondary hydroxyl group oxidation () or cleavage of the diosgenin side chain were not observed.

Bismuth (III) salt-catalyzed allylic oxidation of Δ5-steroids.

Moreover, a white insoluble solid was formed during the BiCl3-catayzed reactions. This solid was recovered by filtration and identified by X-ray diffraction (XRD) analysis as BiOCl. Thus, BiCl3 can be recovered at the end of the reaction and reused as BiOCl, which is also active under these reaction conditions, or reconverted again into BiCl3.

Under similar reaction conditions, the allylic oxidation of (+)-valencene was reported and the corresponding α,β-unsaturated ketone, nootkatone, was obtained in 35% yield () [44,45].

Bismuth (III) salt-catalyzed allylic oxidation of (+)-valencene.

2.3. Oxidation of alcohols

Considering the excellent chemoselectivity observed in the allylic oxidation of dehydroepiandrosterone (), it was interesting to evaluate the selective allylic alcohol oxidation in the presence of a secondary saturated hydroxyl group using the BiCl3/t-BuOOH system. This study was performed using androst-4-ene-3β,17β-diol and it was observed that after complete consumption of this substrate, the important steroid hormone testosterone could be obtained, in 45% yield, after flash column chromatography () [46].

Selective allylic alcohol oxidation of androst-4-ene-3β,17β-diol by the BiCl3/t-BuOOH system.

In another study, the oxidation of the allylic alcohol moiety of carveol has been effected with montmorillonite impregnated with Bi(NO3)3·5H2O affording carvone, a naturally occurring monoterpene () [47].

Oxidation of carveol with montmorillonite impregnated with Bi(NO3)3·5H2O.

3. Formation of Common Protecting Groups

3.1. Acylation of alcohols

The conversion of cholesterol and cholesterol formate into the corresponding 3β-acetoxy derivative has been described by Reese and co-workers using stoichiometric amounts of Bi(OAc)3 () [48].

More recently, acylation procedures using bismuth(III) salts as catalysts have been reported, and the acetylation of cholesterol was performed using Bi(OTf)3·xH2O in the presence of acetic anhydride () [49,50]. Remarkably, all the three hydroxyl groups of cholic acid methyl ester were converted to the corresponding acetoxy groups in excellent yield, regardless of the sterically hindered 12α-hydroxy group [49,50].

Different procedures for the acylation of cholesterol and menthol using bismuth(III) salts.

The Ac2O/Bi(OTf)3·xH2O system has been applied to the acetylation of several monooxygenated monoterpenes, such as geraniol [49,50], borneol [49,50], linalool [49,50] and menthol () [50,51,52,53]. BiCl3 and Bi(CF3COO)3 were also reported as efficient catalysts for the acetylation of menthol [52,53], however higher temperatures and/or reaction times were needed to achieve yields identical to those obtained with Bi(OTf)3·xH2O.

A very efficient process for the acetylation of cholesterol () [54] and menthol (), among several other alcohols, using in situ generated BiCl3 from the procatalyst BiOCl and acetyl chloride was also described.

The pivaloylation of cholesterol (), borneol and menthol () was accomplished by Orita and co-workers using Bi(OTf)3·xH2O (3 mol%) in the presence of either pivaloic anhydride or pivaloyl chloride [50].

Menthol was quantitatively converted into the corresponding formate and benzoate derivatives after reaction with ethyl formate [53] and benzoic anhydride [52], respectively, in the presence of BiCl3, Bi(TFA)3 or Bi(OTf)3·xH2O (). Bismuth(III) salts catalyzed the direct conversion of the THP ether of menthol to the corresponding formate, acetate and benzoate derivatives by reaction with appropriate acylation reagents (). The reactions proceeded at reflux with ethyl formate or acetic acid [55], whereas the use of acetic or benzoic anhydrides, at room temperature, was enough to achieve high yields with short reaction times () [56].

Bismuth(III) salt-catalyzed transformations of the THP ether of (–)-menthol to the corresponding formate, acetate and benzoate derivatives.

An interesting transesterification of betulin with ethyl acetate has been recently reported, using Bi(OTf)3·xH2O as catalyst. Betulin-28-yl acetate was obtained as major reaction product along with allobetulin, resulting from the Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement () [57].

Transesterification of betulin with ethyl acetate catalyzed by Bi(OTf)3·xH2O.

3.2. Formation of ethers

The tetrahydropyranylation of the secondary hydroxyl groups of cholesterol (), geraniol and (–)-menthol and () and the tertiary alcohol of linalool has been accomplished by reaction with 3,4-dihydro-2H-pyran (DHP) in the presence of 0.1 mol% of Bi(OTf)3·xH2O [58]. More recently, Bi(NO3)3·5H2O was reported to be a more efficient catalyst, affording the tetrahydropyranyl (THP) ethers of menthol () and geraniol in high yield, in shorter reaction times [59].

Bismuth(III) salt-catalyzed tetrahydropyranylation of cholesterol and menthol.

Keramane and co-workers reported the benzylation of (S)-(–)-menthol with racemic 1-phenylethanol in the presence of catalytic amounts of BiBr3 [60]. The reaction was performed using the toxic CCl4 as solvent and the corresponding ether was obtained as an equimolar mixture of diastereomers in 90% yield (), and thus the reaction was shown to occur specifically with retention of configuration.

BiBr3-catalyzed benzylation of (S)-(–)-menthol.

Very recently, a mild and efficient solvent-free method for the silylation of alcohols and phenols with hexamethyldisilazane (HDMS) and Bi(OTf)3·xH2O as catalyst was reported. This process was successfully applied to geraniol, an acid sensitive alcohol with two double bonds, and the corresponding trimethylsilyl ether (TMS) was prepared in 95% yield, after just 5 min. of reaction [61].

4. Removal of Common Protecting Groups

4.1. Deprotection of O,O-acetals

The deprotection of citral dimethylacetal has been performed in the presence of 0.1 mol % of Bi(OTf)3·xH2O using THF/H2O (4:1 v/v) as solvent () [62]. More recently, the same reaction has also been reported with BiI3 (1 mol%) in H2O () [63].

Bismuth(III) salt-catalyzed deprotection of citral dimethylacetal.

4.2. Deprotection of ether groups

The high yielding depyranylation reactions of geraniol, menthol and linalool THP ether derivatives have been reported using 1 mol% Bi(OTf)3·xH2O in DMF-MeOH (9:1 v/v) [58] or BiCl3 (3 mol%), Bi(TFA)3 (5 mol%) and Bi(OTf)3·xH2O (1 mol%) in MeOH () [64].

Bismuth(III) salt-catalyzed depyranylation reaction of (–)-menthol THP ether.

The deprotection of the tert-butyldimethylsilyl ether of (–)-menthol was accomplished using the BiCl3/NaI system in acetonitrile, at room temperature, in 80% yield, without loss of the original configuration () [65].

Deprotection of the tert-butyldimethylsilyl ether of (–)-menthol by the BiCl3/NaI system.

4.3. Deprotection of oximes to carbonyl compounds

A catalytic procedure using 10 mol% of BiCl3, in THF, under microwave irradiation, was reported for the regeneration of carbonyl groups from their oximes. The methodology was applied to 5α-cholestan-3-one oxime and the corresponding 3-keto-steroid was obtained in 80% yield () [66].

BiCl3-catalyzed conversion of 5α-cholestan-3-one oxime to the parent 3-keto-steroid.

5. Nucleophilic Ring-opening of Epoxides

5.1. Ritter reaction of epoxides

The Ritter reaction of 5α,6α- and 5β,6β-epoxysteroids with nitriles in the presence of bismuth(III) salts was reported to give the corresponding vic-acylamino-hydroxy products, in high yields () [67,68,69]. This process was stereo- and regioselective as the result of the trans-diaxial ring-opening of the epoxysteroids. In the presence of several other functional groups, such as hydroxyl, ketone or ester, the reaction occurred selectively at the epoxide group. Interestingly, a quite recent study using DFT calculation showed that the higher affinity of Bi for Br can prevent side reactions [70] and makes BiBr3 a better Lewis acid catalyst, which is in accordance with the better results achieved with BiBr3 in this Ritter reaction, when compared with BiCl3.

During the study of the bismuth(III)-promoted Ritter reaction of epoxides, the conversion of (–)-caryophyllene oxide into a clovan-9-ol derivative bearing a 2β-acetamide group at ring A was reported () [67]. Despite the low yield (33%), this reaction is an alternative approach for the synthesis of clovane-type compounds with nitrogen atoms directly attached to C-2.

BiBr3-catalyzed Ritter reaction of epoxysteroids.

BiBr3-catalyzed Ritter reaction of (–)-caryophyllene oxide.

5.2. Ring opening of epoxides with bismuth(III) salt

The Ritter reaction of epoxysteroids was found to occur under catalytic and stoichiometric conditions with BiBr3 [67] However, when stoichiometric amounts of BiCl3 or Bi(NO3)3·5H2O were used to promote the Ritter reaction in acetonitrile, a competitive side product was detected, in both cases. These products were found to be the result of the epoxide ring-opening by BiCl3 and Bi(NO3)3·5H2O, respectively [71]. Thus, using stoichiometric amounts of bismuth(III) salts, such as BiCl3, BiBr3 or Bi(NO3)3·5H2O, and, simply, by changing the solvent to 1,4-dioxane, halohydrins and β-hydroxy-nitrates could be obtained in high yields by ring-opening of 5α,6α-, 5β,6β- and 2α,3α-epoxysteroids () [71,72,73].

The reactions were also stereo-, regio- and chemoselective. Notably, the ring opening of a 5α,6α;16α,17α-diepoxysteroid proved to be highly specific for the 5α,6α-epoxide group () [71,73]. In order to understand the nature of the nucleophilic species, the reaction was performed in the presence of 2,6-di-t-butylpyridine (2,6-DTBP), a known proton scavenger which only binds to protons and is unable to coordinate to metal ions due to the bulky t-butyl groups [74]. Thus, in these reaction conditions, both BiCl3 and Bi(NO3)3·5H2O were able to afford the corresponding β-hydroxy substituted products, in a shorter reaction time than when the reaction was performed without 2,6-DTBP [71]. Taking into account the mechanism proposed by Keramane and co-workers for the chlorination of alcohols with BiCl3 [75] and a recent DFT work that studied the Lewis acidity of BiX3 (X = Cl, Br and I) towards alcohols [70], it has been suggested that the ring opening reaction is mediated by the Lewis acidity of bismuth towards the epoxide (. As suggested by Keramane et al., a new bismuth species should be formed (“BiCl2OH”) [75]. In the ring opening of epoxides, one molecule of H2O, participates by furnishing a H+, present in the final hydroxyl group, whereas the OH– will act as the third ligand of the newly formed bismuth species. Stabilization of Bi–O complex or the final “BiCl2OH” by 2,6-DTBP may help to explain the higher reaction rate observed in the reactions performed in the presence of this organic base.

Ring opening of epoxysteroids with BiX3 (X = Cl, Br and ONO2).

Possible mechanism for the ring opening of epoxysteroids with BiX3 (X = Cl, Br and ONO2).

5.3. Trans-diaxial hydroxylation of Δ5-steroids

Taking into account a recent report for the epoxidation of Δ5-steroids by the use of MMPP [76] and the previous knowledge of bismuth-catalyzed ring-opening reactions of epoxysteroids [67,71], Carvalho et al. described a fast and efficient high-yielding sequential approach for the preparation of 5α,6β-dihydroxysteroids using Δ5-steroids as raw materials [77]. This new synthetic protocol involves two steps: (i) formation of the epoxide from Δ5-steroids, using MMPP as oxidant; and (ii) trans-diaxial ring opening of the diastereomeric mixture of 5α,6α- and 65,6β-epoxides with Bi(OTf)3·xH2O in acetone. A catalytic ring-opening of the epoxide was achieved when a filtration was performed to remove the insoluble salts formed in the first step () [77]. With this methodology, a library of 5α,6β-dihydroxy steroids has been easily obtained, and further research has provided important data regarding the cytotoxic profile of these compounds [78].

Trans-diaxial hydroxylation of Δ5-steroids by sequential combination of MMPP epoxidation and Bi(OTf)3·xH2O-catalyzed ring opening of epoxides.

In this work, some considerations were done regarding the true catalytic species in the ring-opening step. Although the authors demonstrated the importance of Brønsted acid species, either derived from Bi(OTf)3·xH2O or from the oxidant, it has been suggested that Bi(OTf)3·xH2O also participates as Lewis acid in the catalytic cycle, since reactivity has been found, even when the reaction is performed in the presence of stoichiometric amounts of 2,6-DTPB [77].

6. Rearrangement Reactions

6.1. Ferrier rearrangement

The Ferrier rearrangement is a well known methodology for the preparation of alkyl and aryl 2,3-unsaturated-O-glycosides by reaction of glycals with alcohols catalyzed by Lewis acids.

Bismuth(III) salt-catalyzed Ferrier rearrangement.

The reaction of cholesterol with 3,4,5-tri-O-acetyl-D-glucal in the presence of BiCl3 or Bi(OTf)3·xH2O afforded the corresponding 2,3-unsaturated-O-glycoside, in good yield [79,80]. The use of Bi(OTf)3·xH2O or Bi(OTf)3·xH2O/SiO2, proved to be more stereoselective, affording the α-anomer almost exclusively () [80].

6.2. Epoxyolefin cyclizations

Epoxyolefin cyclizations are an important area of research since the discovery that these reactions are involved in biosynthetic pathways of terpenes. The cyclization of geraniolene oxide has been investigated in the presence of several metal triflates, including Bi(OTf)3·xH2O. The reaction product ratio was found to be mainly influenced by the choice of the solvent and substrate concentration rather than by the choice of the metal triflate. Cyclization products were preferentially formed in CH2Cl2 and under high dilution conditions whereas acyclic compounds were mostly obtained in ethereal solvents () [81].

Bi(OTf)3·xH2O-catalyzed cyclization of geraniolene oxide.

Later, Smith and co-workers evaluated the use of stoichiometric amounts of BiCl3 and BiOClO4·xH2O in the Lewis acid mediated cyclization reactions of 6,7-epoxygeranyl pivalate ester also. Despite the fact that good activity was observed, poor selectivity for the desired bicyclic ether was found [82].

6.3. Westphalen and “Backbone” rearrangements

Bismuth(III) salts were described as catalysts for the Westphalen and “backbone” rearrangements of 5β,6β-epoxysteroids [83,84]. The reactions were found to be particularly sensitive to changes on the solvent, temperature, stereochemistry of the starting epoxysteroids and their substituents at C-17. Thus, the Bi(OTf)3·xH2O-catalyzed reaction of 5β,6β-epoxysteroids from the cholestane, androstane and pregnane series, in 1,4-dioxane, at 80 ºC, afforded the corresponding 5β-methyl-Δ9(10)-19-norsteroids, in moderate yields (). On the other hand, using a solvent with high dielectric constant, such as CH3NO2, the rearrangement is extended to rings C and D of the steroid core and the 5β,14β-dimethyl-Δ13(17)-18,19-dinorsteroids from the cholestane and pregnane series were isolated in good yields () [83,84].

Bi(OTf)3·xH2O-catalyzed Westphalen and “backbone” rearrangements of 5β,6β-epoxysteroids.

With the androstane derivative, due to the lack of a 17α-H, a full “backbone” rearrangement was not expected and, in addition to the 5β-methyl-Δ9(10)-norsteroid derivative, the 5β-methyl-Δ8(14)-norsteroid was also formed, in Bi(OTf)3·xH2O/CH3NO2, and isolated in 33% yield () [83]. An interesting feature of the “backbone” rearranged compounds is their abnormal A/B ring trans-fused (5β,10α) steroid structure that was confirmed by X-ray crystallography [83,84]. These rearrangements were found to occur specifically with 5β,6β-epoxides, and not with their corresponding 5α,6α-diastereomers. Data obtained from X-ray crystallography and quantum chemical calculation on the equilibrium geometry of the free molecule [84,85] revealed that important steric factors, such as the change in the torsion angle of the steroid nucleus, may contribute to explain the difference in reactivity observed between 5β,6β- and 5α,6α-epoxysteroids, which do not react under the same reaction conditions, although similar electronic effects are present in these diasteriomeric epoxides. Thus the 5β,6β-stereochemistry is essential for the development of 1,3-syn-diaxial interactions with the angular 10β-methyl group, that are relieved during the rearrangement. In another related study, it has been shown that the “backbone” rearrangement of the 5β,6β-epoxide group occurs preferencially to the Wagner-Meerwein rearrangement of the 16α,17α-epoxy-20-oxo moiety () [86].

Bi(OTf)3·xH2O-catalyzed “backbone” rearrangement of 5β,6β;16α,17α-diepoxy-20-oxopregnan-3β-yl acetate.

6.4. Wagner-Meerwein rearrangements

The application of Westphalen and “backbone” rearrangement reaction conditions [83] to 16α,17α-epoxy-20-oxosteroids has also been studied [86]. At room temperature, with 0.05 eq. of catalyst, low conversion of 16α,17α-epoxy-20-oxopregn-5-en-3β-yl acetate was observed. However, on increasing the temperature to 50 ºC, 16α- and 16β-hydroxy-17β-methyl-Δ13-18-nor pregnane derivatives were formed in near similar amounts from 16α,17α-epoxy-20-oxosteroids () [86].

Several experiments, using HOTf or La(OTf)3 as catalysts instead of Bi(OTf)3·xH2O or with Bi(OTf)3·xH2O in presence of proton scavengers (2,6-DTPB and BiPh3), demonstrated that this rearrangement is mediated by Brønsted acid species generated in situ from the catalyst Bi(OTf)3·xH2O. It was also observed that the 16α-hydroxy-17β-methyl-Δ13-18-nor pregnane derivative undergoes epimerization under these reaction conditions.

Thus, it was suggested that an in situ generated Brønsted acid species from Bi(OTf)3·xH2O catalyzes the ring opening of the 16α,17α-epoxide, creating a tertiary carbocation at C17 followed by stereoselective 1,2-migration of the C18-methyl group to the 17β-position. Due to the 18-CH3 → 17-CH3 shift, a carbocation centered at C13 is formed, and further 14α-H elimination originates the Δ13-double bond. 16β-Epimers can be formed due to an acid-catalyzed reverse-aldol equilibrium involving the 16-hydroxy-20-keto function of the rearranged steroid, under the reaction conditions employed, which is responsible for the epimerization at C16. () [86].

Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement of 16α,17α-epoxy-20-oxosteroids.

Possible mechanism for the formation of 16α- and 16β-hydroxy-17β-methyl-Δ13-18-norsteroids.

Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement in the presence of acylation reagents.

In order to increase the stereoselectivity of the rearrangement, the reactions were studied in the presence of Ac2O. In fact, it was observed that 16α-acetoxy rearranged derivatives were selectively prepared. To enlarge the scope of this Bi(OTf)3·xH2O-catalyzed rearrangement, other acylation agents were used, and several 17β-methyl-Δ13-18-norsteroids bearing different acyl groups at positions C3, C16 and C21 were selectively prepared () [86].

Wagner-Meerwein type rearrangements have also been widely reported in terpene chemistry [87]. One of such well-known transformations involves the acid catalyzed ring E rearrangement of betulin and important related triterpene derivatives, to 18α-oleanane compounds bearing a 19β,28-epoxide or 28,19β-lactone ring, and is commonly referred to “betulin-allobetulin rearrangement”. Following the previous knowledge that bismuth(III) salts are able to catalyze Wagner-Meerwein type rearrangements, their advantageous use as catalysts for the Wagner-Meerwein rearrangement of lupanes has recently been reported [57]. Thus, using CH2Cl2 as solvent and with Bi(OTf)3·xH2O catalysis (0.05 to 0.5 eq.), several lupane derivatives undergo a Wagner-Meerwein rearrangement with expansion of ring E and formation of an additional O-containing ring. By using this process [with 0.05 eq. of Bi(OTf)3·xH2O], several interesting 18α-oleananes could be efficiently obtained from lupanes by a single Wagner-Meerwein rearrangement () [57,88].

Some mechanistic studies were performed, in a similar fashion to what has been done during the study of the Wagner-Meerwein rearrangement of 16α,17α-epoxy-20-oxosteroids [86], and it was demonstrated that these reactions are catalyzed by Brønsted acid species generated in situ from the hydrolysis of Bi(OTf)3·xH2O. Therefore, the reaction mechanism can be postulated as follow. Under this reaction conditions, Bi(OTf)3·xH2O undergo a reversible hydrolysis leading to the formation of bismuth-containing cationic species [17] and thus the formation of an in situ Brønsted acid species. The hydrolysis of Bi(OTf)3·xH2O has been shown to lead to the quantitative formation of triflic acid, in water [89], however experimental data under this reaction conditions do not allowed to conclude it, unquivocally. Anyway, the Brønsted acid species formed in situ catalysis the formation of successive carbocations that are eventually trapped by intramolecular nuclophylic attach by the C28-hydroxyl group () [57].

Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement of betulin and betulinic acid derivatives.

Under more vigorous conditions (0.2-0.5 eq. of Bi(OTf)3·xH2O), the dehydration of the 3β-hydroxyl group occurred and the resulting compounds underwent a double Wagner-Meerwein rearrangement, originating A-neo-18α-oleanene compounds in high yields (). This process was also applied the sesquiterpene (–)-caryophyllene oxide, to afford clov-2-en-9α-ol by a “caryophyllene-clovane rearrangement” ().

Possible mechanism for the Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement of betulin to allobetulin.

Bi(OTf)3·xH2O-promoted dehydration and double Wagner-Meerwein rearrangement of betulin (R1 = H2) and betulinic acid (R1 = O).

Bi(OTf)3·xH2O-catalyzed rearrangement of (–)-caryophyllene oxide.

6.5. Oxidative rearrangement of tertiary allylic alcohols

Recently, a new method for the oxidative rearrangement of tertiary allylic alcohols to the corresponding transposed carbonyl derivatives using TEMPO in combination with PhIO and Bi(OTf)3·xH2O was described. However, when applied to the terpenic derivative 2-methylcarveol, this procedure only led to 19% yield of 3-methylcarvone ( [90,91].

Bi(OTf)3·xH2O-catalyzed oxidative rearrangement of 2-methylcarveol.

7. Miscellaneous Reactions

7.1. Diels-Alder cycloaddition

(+)-Menthol was used as an inexpensive chiral auxiliary in the course of the development of a large-scale production process for the preparation of the enantiomerically pure (2R,3R)-bicyclo[2.2.1]hept-5-ene-2,3-dicarboxylic acid, an intermediate in the synthesis of the adenosine A1 antagonist BG9719 (CVT-124). In this context, he asymmetric Diels-Alder reaction of (+)-dimenthyl fumarate and cyclopentadiene has been extensively investigated in order to determine which conditions lead to the best yield and diastereomeric excess. Among the Lewis acids tested, BiCl3 promoted reaction rendered 100% conversion, but poor diastereomeric excess () [92].

BiCl3-promoted Diels-Alder reaction of (+)-dimenthyl fumarate and cyclopentadiene.

7.2. Ene reaction

The BiCl3-catalyzed reaction of (–)-β-pinene with an equimolar amount of chloral in CH2Cl2/Et2O gave the corresponding H-ene adduct, in 49% yield [ratio (11R)/(11S)-diastereomers = 64:36] () [93].

BiCl3-catalyzed ene reaction of (–)-β-pinene with chloral.

A well known example of an intramolecular ene reaction is the cyclization of citronellal to (–)-isopulegol, an important intermediate in the industrial production of (–)-menthol (Takasago process). This reaction has been reported to be catalyzed by BiCl3 (2–5 mol%) [93] and Bi(OTf)3·xH2O (0.1 mol%) () [94]. Along with the desired product, neoisopulegol was also formed in low amounts (<30%) in both processes.

Bi(OTf)3·xH2O-catalyzed cyclization of citronellal.

7.3. Aromatic nitration reactions using Bi(NO3)3·5H2O

Montmorillonite impregnated with bismuth nitrate was found to be an efficient reagent for the nitration of estrone affording the 2-nitro and 4-nitro derivatives as a 1:1 mixture, in 94% yield [95,96].

Nitration of estrone using Bi(NO3)3·5H2O.

More recently, the same group reported further developments of the initial methodology using various metal nitrates, including Bi(NO3)3·5H2O, and screened several solid supports and reaction conditions. The best results with Bi(NO3)3·5H2O were accomplished using Florisil or molecular sieves, in refluxing benzene, in a Dean Stark apparatus, or alumina, under dry conditions. With these reaction conditions, a 5:1 ratio of the 2-nitro and the 2,4-dinitro derivatives was obtained () [97].

7.4. Lactonization of Δ12-oleananes

During the study of the Bi(OTf)3·xH2O-catalyzed Wagner-Meerwein rearrangement of terpenes, it was found that this reagent is able to promote the 28,13β-lactonization of oleanonic acid () [57]. Interestingly, under bismuth catalysis, a carbocation is formed not only at C13 but also at C18, as shown by the observed inversion of the configuration of the C18-stereocenter, which has been confirmed by X-ray data [98].

Bi(OTf)3·xH2O-promoted lactonization of oleanonic acid.

7.5. Cleavage of the C17-side chain of corticosteroids

A quite recent report showed that Bi(OTf)3·xH2O is an efficient catalyst for the conversion of corticosteroids into highly functionalized 17-ketosteroids, by cleavage of the C17-dihydroxyacetone side chain [99]. This reaction was typically performed using large amounts of oxidative or basic reactants [100,101,102]. The use of other bismuth(III) salts and other solvents has been tested, however the best results have been obtained with 0.05 eq. of Bi(OTf)3·xH2O, in 1,4-dioxane, at 80 °C. Under this reaction conditions, non-steroidal by-products have been observed. Their presence probably arises from the acid-catalyzed polymerization of the α-hydroxy ketone moiety derived from the cleavage of the corticosteroid side chain. Thus, purification by flash chromatography was needed to obtain pure 17-ketosteroids. Using this method several highly functionalized 17-ketosteroids have been prepared ().

The process proved to be very chemoselective, since functionalities of the starting corticosteroids, such as Δ4-3-keto, Δ1,4-3-keto, 11β-hydroxyl, and 9β,11β-epoxide, remained intact [99]. As an important drug class in the treatment of a variety of clinical situations [103], corticosteroids are readily available compounds. In fact, their industrial production reaches several tonnes per year [104], making this new process an useful source of substrates to obtain other steroids bearing a variety of chemical functions in rings A, B, C and D.

Bi(OTf)3·xH2O-catalyzed cleavage of the C17-dihydroxyacetone side chain of corticosteroids.

It is interesting to note that when the reaction was performed using HOTf as catalyst, only 11% of the 11β-hydroxy-17-ketosteroid was obtained along with 21-(1,3-dioxalan-2-yl)-11β-hydroxypregn-4-ene-3,20-dione, in 45% yield, as the major reaction product (). Moreover, using the Bi(OTf)3·xH2O/2,6-DTBP system, the desired product was obtained in good yield, after a longer reaction time (). Therefore, although Brønsted acid catalysis was observed, it seems the Bi(OTf)3·xH2O-catalyzed cleavage of the C17-dihydroxyacetone side chain of corticosteroids is mediated by Lewis acid catalysis [99]. In the absence of proton scavenger, Brønsted acid-assisted Lewis acid catalysis [25] is observed, as suggested by the higher reaction rate.

Cleavage of the C17-dihydroxyacetone side chain of corticosteroids under Brønsted acid and Lewis acid catalysis.

4. Conclusions

In the past five years, several new bismuth-based processes have been reported in the field of steroid and terpene chemistry, which added new “ecofriendly” tools for the synthesis of valorous molecules.

Efforts have been made to tentatively identify the nature of the true catalytic species. Many literature reports attributed the reactivity to the Lewis acid nature of the bismuth(III) salts, however others explain the observed catalytic reaction as result of a Brønsted acid species formed in situ from bismuth(III) salt. In addition, Brønsted acid-assisted Lewis acid catalysis and Lewis acid-assisted Brønsted acid catalysis have also been found in selected transformations, involving bismuth(III) salts. This chemical behavior is responsible for many interesting reactivities of bismuth(III) salts towards steroids and terpenes. In fact, these catalysts were able to promote transformations that only have been earlier performed under extremely hard reaction conditions, such as strong mineral acids or powerful oxidant or basic reactants.

Besides the great advances that have been obtained so far, more research in the development of new bismuth-based processes using steroid and terpenes of pharmaceutical interest is required. Moreover, incorporation of some of these procedures, in the design of large scale synthetic approaches for the preparation of selected steroid and terpenes, in a near future, will be expected.