Allobetulin and Its Derivatives: Synthesis and Biological Activity

This review covers the chemistry of allobetulin analogs, including their formation by rearrangement from betulin derivatives, their further derivatisation, their fusion with heterocyclic rings, and any further rearrangements of allobetulin compounds including ring opening, ring contraction and ring expansion reactions. In the last part, the most important biological activities of allobetulin derivatives are listed. One hundred and fifteen references are cited and the relevant literature is covered, starting in 1922 up to the end of 2010.


Introduction
Triterpenes and triterpenoids are numerous and widely distributed in Nature. Biosynthetically, they are derived from squalene. Earlier studies have focused on the isolation and structural elucidation of the compounds, and there is still a lot of ongoing research in this area that has been regularly reviewed by Connolly and Hill [1]. During recent years, several interesting biological properties were found for this class of compounds, which in combination with their low toxicities lead to an increased research OPEN ACCESS effort [2,3]. More particularly, the oleanane group displays a number of significant pharmacological activities. Allobetulin (2) and its derivatives, obtained from the readily available lupane betulin (1), form a part of the oleanane group.
In this review, we summarize the chemistry of allobetulin analogs including: (1) their formation by rearrangement from betulin derivatives, (2) their further derivatisation, (3) their fusion with heterocyclic rings, and (4) the further rearrangements of allobetulin including ring opening, ring contraction and ring expansion reactions. In the final part (5), the most important biological activities of the allobetulin derivatives mentioned in sections 1-4 are listed.
There are also a number of allobetulin derivatives that are isolated from plant extracts. For a recent example see [4]. These will not be treated in this review. We also did not cover the chemistry of the ring contracted or seco-derivatives of allobetulin, other than their formation from allobetulin derivatives.

Betulin-Allobetulin Rearrangement
In 1922, Schulze and Pieroh reported that when betulin (1) was heated in formic acid, an unexpected formate ester product resulted, that gave an isomeric product after saponification that was named allobetulin (2) (Scheme 1) [5]. At that time, very little was known about the structure of (allo)betulin due to the lack of adequate characterisation techniques, but the authors were able to conclude that the obtained product was a monoalcohol, containing an ether function and an otherwise strongly rearranged structure as compared to the dialcohol betulin (1). Dischendorfer et al. determined the correct molecular bruto formula of 2 not much later [6]. In the following years several authors carried out similar rearrangements and prepared derivatives of allobetulin (2), but breakthroughs regarding its structure came only after the work of Davy [7] who oxidized the acetate of allobetulin to the corresponding 28-oxo derivative, and then saponified it to the alcohol and oxidized this compound to oxyallobetulone (3). The latter was identical to a product ("ketone-lactone-A") derived by rearrangement of betulonic acid. Only recently was an X-ray structure of allobetulin (2) reported [8]. Various acidic conditions have been applied for this transformation, which is now known to belong to the class of Wagner-Meerwein rearrangements. Hydrobromic acid in chloroform [9], sulfuric acid in acetic acid [10], concentrated hydrochloric acid in ethanol [11,12] and even the highly toxic dimethyl sulfate [13] have been used for the transformation of 1 to 2 in moderate to good yields. The yield can be substantially improved by using acid reagents adsorbed on solid supports. Li et al. used "solid acids" such as sulfuric acid or tosic acid on silica, Montmorillonite K10 and KSF, bleaching clays and kaolinite to obtain allobetulin and its derivatives in close to quantitative yield [14]. Pichette et al. have used ferric nitrate or ferric chloride absorbed on silica gel or alumina to convert betulin (1) into allobetulin (2) in excellent yield. Longer reaction times lead to the formation of allobetulone (4) or Aring contracted products, respectively [15]. Ferric chloride hydrate itself (not supported) was also used for a larger scale reaction (approx. 5 g, 92% yield) [16]. Trifluoroacetic acid [17] or bismuth triflate (via triflic acid liberated by hydrolysis) [18] also give excellent results for this transformation. Russian researchers, including patent literature, mention the use of diluted sulfuric acid [19] and orthophosphoric acid [20] to combine the process of extraction of 1 from birch bark and rearrangement to 2. This rearrangement can in fact be seen as an interesting undergraduate laboratory experiment [21].
Simple derivatives of betulin, such as betulone, 3-acetylbetulin, and betulinic acid have been transformed by the above methods to the corresponding allobetulin analogs allobetulone (4), 3acetoxyallobetulin (5), and 28-oxoallobetulin (6). Betulinic acid is slower to rearrange in comparison to other betulin analogs and may give substantial amounts of side products. 28-Oxoallobetulin (6) may be prepared more effectively in two steps by rearrangement of the 3-acetylated betulinic acid, followed by hydrolysis [22]. As mentioned earlier, rearrangement of betulonic acid or its methyl ester [23] affords triterpene 3, which can be reduced back to 6 ( Figure 1) [24].

28-oxoallobetulin (6)
Another example is the preparation of 3-amino-28-oxoallobetulin (7) after attempted trifluoroacetic acid deprotection of the corresponding Boc-protected betulinic acid derivative [25]. Treatment of betulin (1) with bromine was reported to give a good yield of the dibromoallobetulin (8) [26]. The structure of rearrangement product 8 was proven by X-ray crystallography. However, this good yield is difficult to reproduce so an efficient procedure towards this interesting product is still lacking. Pradhan  In the recent work of Czuk et al., allobetulin homologues 12 were prepared in almost quantitative yields by trifluoroacetic acid induced rearrangement of secondary alcohols 11 that were prepared from 3-acetylated betulinic aldehyde by aldol condensation reactions (Scheme 3) [31].  The naturally occurring 23-hydroxybetulin (13, obtained from the bark of Sorbus aucuparia L.) was transformed to the diformate 14a (R = CHO) by an adaptation of the Schulze-Pieroh procedure [11]. Removal of the formate lead to 23-hydroxyallobetulin (14b, R = H). Oxidation of the latter with Jones reagent lead to formation of the norketone 15, after decarboxylation of the intermediate ketoacid. The latter compound was used as a means to functionalize the B-ring, and 19β,28-epoxy-18α-olean-5-ene derivatives such as the interesting unsaturated allobetulone analog 16 were obtained after a bromination, dehydrobromination and methylation sequence ( Figure 3) [32].

Further Rearrangements of Allobetulin, Including Ring Contractions and Ring Expansions
Often, the rearrangement of betulin (1) to allobetulin (2) is accompanied with the formation of the dehydrated, isomeric "apoallobetulins". The latter have a variety of structures and can also be obtained from isolated allobetulin (2) by treatment with different acidic reagents. The structure of the δallobetulin 51 obtained by treatment of allobetulin (2) with PCl 5 or phosphorous pentoxide at 0 °C [5,6,15] was shown later by ozonolysis to have an exocyclic double bond [89]. The so-called αapoallobetulin 52 has an endocyclic double bond and is formed on treatment of betulin (1) with Fuller's earth [6,89]. More recently, different solid acids such as Montmorrilonite K10 have successfully transformed 1, 2 or even the δ-isomer 51 to mixtures of 52 and the "rearranged αapoallobetulin" [47] 53. In general, the amount of 53 in the mixture increases at higher temperatures. 28-Oxo derivatives of 52 and 53 are formed accordingly from betulinic acid or 28-oxoallobetulin (6) [15]. Silica-or alumina supported FeCl 3 hydrate gave a similar mixture (55:45 ratio) of 52 and 53 on extended reaction of betulin (1), via allobetulin (2) [16]. The reaction of allobetulin (2) with PCl 5 has been reinvestigated and was shown to lead directly to 52 at slightly higher temperatures (5-10 °C). At −10-0 °C, the expected 51 was formed [47]. The highest yields and selectivities of apoallobetulin isomers were obtained on treatment of betulin (1) with bismuth triflate. The relative amount of catalyst is important. Thus, heating 1 with 20 mol% catalyst for 40 h at reflux in dichloromethane gave 98% yield of 52. On the other hand, heating of 1 or 52 for 8-15 h in the same solvent with 50 mol% bismuth triflate gave the isomer 53 almost quantitatively (96-98% yield) (Figure 13) [19]. Treatment of allobetulin (2) with acid chlorides in high boiling solvents leads to rearranged and ring opened diacylated products 54a, that can be saponified to the so-called "heterobetulin" 54b, which has an ursane framework [9,47,90,91]. "Alloheterobetulin" 55 is a ring closed isomer of the latter which can be obtained after treatment of 54b with toluenesulfonic acid [92]. A remarkable rearrangement/O,C-diacylation was recently reported to occur (55% yield of 56) when allobetulin (2) was treated with acetic anhydride and a few drops of perchloric acid ( Figure 14) [93]. The Baeyer-Villiger oxidation of allobetulone (4) was investigated by different groups under different circumstances [40,94,95]. With MCPBA in dichloromethane, the main product (83%) is the ring-expanded lactone 57a. Other peracids (performic, peracetic) give similar results. However, reaction of 4 with MCPBA in the presence of acid (acetic + sulfuric) leads to the formation of a nor-lactone 57b. The 3,4-seco derivatives 58a,b were obtained in good yield either from 57a by alkaline hydrolysis or directly from 4, carrying out the oxidation in methanol with a trace amount of sulfuric acid [95]. Larger amounts of acid (0.15%) lead to the formation of the 2α-hydroxyallobetulone 28c in good yield (86%) (Figure 15) [65]. The Beckmann rearrangement of allobetulin oxime 17a, induced by TsCl/pyridine or phosphoryl chloride, gave rise to the formation of a lactam 59a (major product) and a 3,4-seco-triterpene nitrile 60 (minor product). The lactam 59a could be transformed into the nitrile 60 on extended heating [96,97]. Upon Schmidt reaction of methyl betulonate or Beckmann rearrangement (POCl 3 ) of its oxime, the 28-oxo derivatives 59b and 60b were formed after two consecutive rearrangements [98]. Other 2,3seco-derivatives were prepared via Beckmann fragmentation of allobetulin derivatives ( Figure 16) [33,45,99,100]. The dibromoallobetulin 29 underwent a quasi-Favorskii rearrangement on treatment with base, leading to the ring contracted product 61a. Oxidative decarboxylation of the latter with lead tetraacetate gave the norketone 62 [45,56,96,101]. The latter is an interesting starting material that was used in many follow-up reactions that will not be discussed here. Benzilic acid rearrangement of diketone 20a gives the same hydroxyacid 61a. The photochemical Wolff rearrangement of diazo compound 21c gave the ring contracted carboxylic acid 61b (Figure 17) [45]. Dischendorfer reported oxidation of the A-ring of 28-oxoallobetulone 3 to "allobetulinic acid" which formed a cyclic anhydride 63 [102,103]. This seco-derivative 63 was recently used to prepare spirocyclic derivatives 64 after treatment with benzylamines and oxalyl chloride [104]. Recently, the diacid analog of 63 was prepared by ozonolysis of the Claisen ester condensation product of 3 (i.e. the 3-oxo analog of 27) [105]. This procedure has some similarity with earlier work by Ruzicka, who used chromic acid to prepare the diacid from hydroxymethyleneallobetulone (27, R = H) or directly from allobetulin (2) (Scheme 7) [106].  Another obvious position for ring cleavage is the lactone bridge of 28-oxoallobetulin derivatives. This bridge is quite stable towards saponification, but LiAlH 4 reduction of the 3β-acetoxy derivative of 5 [108] or 28-oxoallobetulone (3) [25,30,109] gives a germanicanetriol derivative 67 which was further transformed to different germanicanes by selective acylation, oxidation and dehydration reactions [108,109]. The lactone ring of the protected 28-oxoallobetulone 22b was reductive cleaved with LiAlH 4 and after deprotection, 28-acetylation, dehydration with POCl 3 , saponification, and stepwise oxidation, moronic acid (68) and the reduced morolic acid 69 were obtained ( Figure 19) [41]. In general, allobetulin and its derivatives are important starting materials for the synthesis of rare germanicanes and olean-18(19)-ene triterpenoids.

Biological Properties of Allobetulin Analogs
The biological properties of betulin, betulinic acid and its derivatives are well known [2] and often activity studies of allobetulin derivatives are found back in the literature together with or in comparison to their betulin isomers. A wide spectrum of biological properties have been reported, including antiviral, antifeedant, immunotropic, antibacterial, antifungal, and anti-inflammatory activities, cytotoxicity and inhibition of glycogen phosphorylase activities.

Antiviral properties
In 1995, it was found that allobetulin (2) itself showed moderate inhibitory activity against the influenza B virus [111]. It was claimed in the patent literature that different derivatives of allobetulin, including 2 and its 3-O-acylated and phosphorylated derivatives, 4, 30a, 32, exhibited significant antiviral activity and could be used to treat herpes virus (HSV-herpes simplex virus) infection [35]. Also in 2002, compound 3 was shown in cell culture to inhibit influenza A growth while being inactive against HSV and the enterovirus ECHO-6 [112]. Somewhat later, allobetulin derivatives 3, 6, and different O-acylated oximes 39 were tested against several viruses such as HSV, influenza and ECHO-6 [24,37]. In fact, the non-acetylated oxime 17a had the largest effect against influenza virus A, while being only moderately active against enterovirus ECHO-6 and inactive with respect to HSV. The Nacetylated oximes 39 had a moderate activity towards HSV, but were inactive against the other viruses [37]. It was confirmed that 28-oxoallobetulone (3) strongly inhibited the influenza virus, but did not influence HSV reproduction [24]. Rearranged product 56 showed only moderate inhibition of the Papilloma virus [93].

Antifeedant properties
In 1990, Lugemwa et al. reported high antifeedant activity against the bollworm larvae, Heliothis zea, for the glycoside derivative of 30a. Simple allobetulin derivatives such as 2, 20a and 30a itself were not active. The antifeedant property was selective and the glycoside did not display high activity against either the Colorado potato beetle (Leptinotarsa decemlineata) or the fall armyworm (Spodoptera frugiperda) [40].

Immunotropic activities
Different 2-substituted allobetulone derivatives, including the formyl analog 27 (R = H) and different condensation products with amines 41 were screened [53]. In fact, compounds 27 and 41 (R = i Pr) had the most promising activity combined with low toxicity. In later work it was shown that these compounds had high biological activity on chronic administration and that their immunosuppressive activity was the result of toxic effect on the lymphocytes [113].

Anti-inflammatory and anti-ulcer properties
Biological tests on mice with the carrageenan and formalin edema models showed that acylated derivatives of allobetulin (2) possessed anti-inflammatory activity comparable to ortophen (diclofenac) [71,115]. Moderate antiulcer activity of 3 and 3-O-acylated allobetulin derivatives were observed in mice [112,115].

Conclusions
Allobetulin and its analogues are easily accessible starting from the corresponding betulin derivatives. Although a large structural variety of allobetulin analogs is already available by functionalisation, ring fusion to the A ring, further rearrangements, ring contractions, ring expansions, and ring cleavages, there is still much chemical space unexplored. Further investigations are certainly worthwile because of the interesting bioactivities.