3 4 -Anhydrovinblastine

Two deoxygenations of naturally occurring compounds have been reported. Anhydrovinblastine, an important intermediate for the anticancer drug navelbine, was prepared by Doris and coworkers from leurosine, an abundant alkaloid from the Madagascan periwinkle Catharantus roseus, in 70% yield by using Cp2TiCl (Scheme 5) [46]. 

The isolation of Catharine (10), C45H54N40,o, mp 271-275°C, an on-colytically inactive alkaloid, has been reported from several Catharan-thus species C. roseus (29-31), C. ovalis (32), and C. longifolius (33). The structure of Catharine (10) has been elucidated by X-ray crystallography (89-91) of its acetone solvate. Catharine (10) can be obtained by mild oxidation of either leurosine (11) (30) or anhydrovinblastine (8) (92-93). In view of the ease of this oxidation, Catharine (10) may be considered as an artifact of the isolation process. 

Additionally, there had, for many years, been a vast synthetic effort underway aimed at the synthesis of the two monomeric units, where it was anticipated that the two units could be joined to form the vinblastine-type bisindole alkaloids. Coincidentally, as it transpired, 20 years of effort in the areas of synthesis and biosynthesis converged, at almost the same time, on the compound 3, 4 -anhydrovinblastine (8). 

Goodbody and co-workers (7/9) have examined the production of alkaloids in root and shoot cultures induced from seedlings of C. roseus. The pattern of alkaloids in the root cultures was similar to that of the roots from intact plants. Thus ajmalicine (39) and catharanthine (4) were produced, but no vindoline (3), a major leaf alkaloid, and no bisindole alkaloids. Similarly, the pattern of the alkaloid content of the shoot cultures was like that of the leaves of the intact plant, showing the presence of vindoline (3), catharanthine (4), and ajmalicine (39), with 3 predominating. A search for the bisindole alkaloids in the cultures indicated the presence of anhydrovinblastine (8) and leurosine (11) in the shoot cultures (2.6 and 0.3 xg/g fresh weight, respectively), but no vinblastine (1) or vincristine (2). 

The enzyme-catalyzed formation of anhydrovinblastine (8) from catharanthine (4) and vindoline (3) was first examined by Kutney and co-workers (170,219) using a cell-free preparation. [ao f- H]Catharanthine (4) and [acety/- C]vindoline (3) were incubated for 3-8 hr, both separately and jointly with a preparation from C. roseus, which led to the isolation of labeled anhydrovinblastine (8) and leurosine (11) incorporations were of the order of 0.54 and 0.36%, respectively. On this basis, anhydrovinblastine (8) was proposed as the key biosynthetic intermediate en route to vinblastine (1) and vincristine (2). 

More recently, Kutney and co-workers (220) have investigated whether the same dihydropyridinium intermediate 109 is involved in the enzymatic conversion of catharanthine (4) and vindoline (3) to anhydrovinblastine (8) as is involved in the chemical conversion. Use of a cell-free preparation from a 5-day culture of the AC3 cell line gave 18% of the bisindole alkaloids leurosine (11), Catharine (10), vinamidine (25), and hydroxy-vinamidine (110), with 10 predominating. When the incubations were carried out for only 5-10 min, the dihydropyridinium intermediate was detected followed by conversion to the other bisindole alkaloids, with FAD and MnClj required as cofactors. Clearly a multienzyme complex is present in the supernatant, but further purification led to substantial loss of enzymatic activity. The chemical formation of anhydrovinblastine (3) is carried out with catharanthine A-oxide (107), but when this compound was used in the enzyme preparation described, no condensation with vindoline (3) occurred to give bisindole alkaloids. This has led Kutney and co-workers to suggest (220) that the A-oxide 108 is not an intermediate in the biosynthetic pathway, but rather that a 7-hydroperoxyindolenine... 

FMN. Similar conversion yields for each of the isozymes were in the range 34-50%. Paralleling these data was the observation that horseradish peroxidase was also capable of converting vindoline (3) and catharanthine (4) to anhydrovinblastine (8) with the correct C-18 stereochemistry (222). 

The conversion of anhydrovinblastine (8) to vinblastine (1) has been examined by several different groups, using intact plants, cell suspension systems, and cell-free preparations. From the studies discussed above it was clear that 3, 4 -anhydrovinblastine (8) was probably the initially formed intermediate in the condensation of vindoline (3) and catharanthine (4) prior to vinblastine (1). Kutney and co-workers have reported (225,226) on the biotransformation of 3, 4 -anhydrovinblastine (8) using cell suspension cultures of the 916 cell line from C. roseus a line which did not produce the normal spectrum of indole alkaloids. After 24 hr the major alkaloid products were leurosine (11) and Catharine (10) in 31 and 9% yields, respectively, with about 40% of the starting alkaloid consumed. 

When 3, 4 -[ao - H]anhydrovinblastine (8) was incubated with a cell-free preparation at pH 6.3 for 50 hr, leurosine (11) and Catharine (10) were labeled to the extent of 8.15 and 15.15%, respectively, and vinblastine (1) was labeled to 1.84% (776,227). Approximately the same level of incorporation was obtained by Scott s group, using 3, 4 -[21 - H]anhydrovinblas-tine (8) and isolating vinblastine (1) from cell-free extracts of C. roseus (228). Scott s failure (87) to observe incorporation of the same precursor into vinblastine (1) in whole plants was explained by the established (82) instability of anhydrovinblastine (8). 

The question of the possible artifactual nature of leurosine (11) was examined by Kutney and co-workers (229) using a cell-free preparation at pH 6.3. After a 3-hr incubation, a 22% yield of leurosine (11) was formed from anhydrovinblastine (8), and when anhydrovinblastine (8) was incubated with horseradish peroxidase in the presence of HjOj, leurosine (11) was formed in 65% yield. 

To examine whether a dihydropyridinium intermediate is involved in the conversion of anhydrovinblastine (8) to vinblastine (1), Scott et al. fed methyl- C]anhydrovinblastine (8) to a cell-free suspension... 

The extremely low yield of vincristine (2) from intact plants has made pursuit of its biosynthesis a very challenging problem, which at this point in time remains unsolved. Kutney et al. have used both anhydrovinblastine (8) (227) and catharanthine N-oxide (107) (233) as precursors to vincristine (2) in a cell-free preparation, but incorporation levels were extremely low. Therefore, the question of whether vinblastine (1) is an in vivo, as well as an in vitro, precursor remains to be answered. Several possibilities exist for the overall oxidation of vinblastine (1) to vincristine (2), including a direct oxidation of the A-methyl group or oxidative loss of the N-methyl group followed by N-formylation. 

A coupling of vindoline (3) with catharanthine A-oxide (45), mediated by trifluoroacetic anhydride, and subsequent reduction with sodium boro-hydride delivered anhydrovinblastine (42) in up to 40% yield, accompanied by minor amounts of the undesired C-16 (R) isomer 24 as well as some 17-deacetylanhydrovinblastine (46) (Scheme 13) (38-41). The struc-... 

Anhydrovinblastine (42) was produced from 15-(3-acetoxy-20-a-ethyldi-hydrocatharanthine (65) in that study [63,64-, an identical observation has been reported by Langlois et al. (65,66)], while the Kutney group reported the formation of 15-a-acetoxy-20-deoxyleurosidine (66) for this reaction 67,68). In all cases the major products of the reactions, 67-69, were again... 

The relatively facile formation of anhydrovinblastine (42) by the modified Polonovski reaction, and the poor yields experienced on coupling of other catharanthine derivatives, made anhydrovinblastine an attractive precursor for the preparation of additional binary alkaloids, including vinblastine itself (50,57). Hydrogenation to 20 -deoxyleurosidine (61a), formation of its Af -oxide (78), and reaction with trifluoroacetic anhydride led to the enamine 59 (Scheme 23). Oxidation of this enamine with thal-... 

It is interesting to note that these experiments strongly suggested that anhydrovinblastine is the biogenetic precursor of the VLB-type alkaloids in the plant, a proposal recently substantiated (47,48,82-88). 

Coupling of other vindoline derivatives with ring D or E modified oxidation levels (92-96) to catharanthine N-oxide provided new binary products for biological evaluation 39, 95-97). The two diastereomeric C-16 -C-14 PARE anhydrovinblastines 42 and 97 were obtained in a 46 54 ratio (50% yield) from racemic catharanthine (98), and the corresponding 20 -desethyl compounds 98 and 99 were generated at - 20°C in a 1 1 ratio (16% yield each), and at -76°C in lower yields, together with the corre-... 

The stereochemistry of hydrogenation of anhydrovinblastine was initially misassigned. See also ref. 70-74. 

The chemical reactivity of N-6 (or N ), directed entirely by the basicity of this atom, is controlled by the nature and stereochemistry of the substituents at C-4 (vide supra). Oxidation of N-6 occurs under mild conditions in several naturally occurring bisindole alkaloids. Thus, treatment of a dichloromethane solution of leurosine (4) with m-chloroperben-zoic acid at -20°C for 4 hr gives the N -oxide (15) in greater than 90% yield after preparative reversed-phase chromatography (46). Leurosine A/ -oxide has also been isolated from Catharanthus roseus and should therefore be considered a naturally occurring bisindole (50). The analogous conversion of vinblastine (1) to its A/ -oxide (16) proceeds under similar conditions but requires longer exposure to the peraeid (24 hr) (5/) 3, 4 -anhydrovinblastine is converted to its N -oxide (17) in 10 min at 0°C... 

Two dimeric Vinca alkaloids, vinorelbine and anhydrovinblastine, have also been investigated using the CHCl3-HF-SbF5 system.550-552 In both cases, the corresponding products difluorinated in the ethyl side chain (140) were isolated in modest yields [Eq. (5.206)]. The mechanistic pathway suggested that the transformation includes steps already depicted in Schemes 5.57 and 5.58. 

The product from the reaction of anhydrovinblastine Nfc-oxide 1 with trifluoroacetic anhydride in methylene chloride was treated, after evaporation of the excess of the reagent, with a mixture of water and THF. This gave a mixture of products, only one of which could be obtained pure (27%), namely 5 -noranhydrovinblastine 2. 

The diols (15 / )- and (15 S)-hydroxyleurosidine, and the four 15 -hydroxy-compounds derived from deoxyleurosidine and deoxyvinblastine, have been prepared by standard procedures from anhydrovinblastine for pharmacological evaluation.122... 

---