Radomir N.
Saičić
a,
Radomir N.
Saičić
*a and
Radomir
Matović
b
aFaculty of Chemistry, University of Belgrade, Studentski trg 16, P.O. Box 158, YU-11001, Belgrade, Yugoslavia. E-mail: rsaicic@chem.bg.ac.yu.
bICTM, Center for Chemistry, Njegoševa 12, 11000, Belgrade, Yugoslavia
First published on 27th January 2000
A semisynthesis of 7-deoxypaclitaxel 4 is described, starting from taxine 6 – the most abundant naturally occurring taxane diterpene fraction. A key step in this transformation is a tandem reaction: stereoselective osmylation of cinnamic ester 14/intramolecularly assisted methanolysis of 16, which gives the key intermediate 5, along with the optically pure ester 17 – a precursor for the synthesis of the paclitaxel side-chain. In this way, the cinnamoyltaxicine 9 is converted into 7-deoxybaccatin III 25 in 11 steps, and in 15% overall yield.
® (Paclitaxel) 1 (Scheme 1), isolated from the bark of the Pacific yew (Taxus brevifolia),1 has attracted intense interest from the scientific community, owing to its great potential in the successful treatment of many types of cancer, unusual mode of antimitotic action, and complex molecular architecture.2 Although six total syntheses of Paclitaxel have been reported,3 the structural complexity of the compound makes it elusive for an economical total synthesis. For its commercial production, which was once based solely on extraction from the bark of Taxus brevifolia, which destroys the tree, and arguably threatened the very slow growing yew population, there is now available a useful alternative, which is semisynthesis from 10-O-deacetylbaccatin III (10-DAB) 3, which is in turn isolated, most notably, from the renewable leaves of the European yew, Taxus baccata.4 This method, more environmentally friendly, gave rise to a novel, non-natural taxoid Taxotere
® (Docetaxel) 2, which is comparable to, if not better than, Paclitaxel in terms of therapeutic indications and efficiency.5 However, 10-DAB is not the most abundant secondary metabolite of the yew tree: its variable content in the needles of Taxus baccata ranges from 0.01 to 1 g per kg of leaves.6 In addition, the isolation of pure 10-DAB involves somewhat tedious purification by multiple-column chromatography. In contrast, a mixture of alkaloids collectively referred to as ‘Taxine’
6 can be obtained by a simple extraction procedure in yields of 7–12 g kg−1,7 the two major constituents of this fraction (≈35%) being taxine B 6a and isotaxine B 6b.8 Therefore, the development of a procedure that would allow for the efficient use of this starting material for the preparation of biologically active paclitaxel analogues would be of considerable interest.
![]() | ||
| Scheme 1 Paclitaxel, Docetaxel, and their common precursor 10-DAB. | ||
Structure–activity relationship (SAR) studies on paclitaxel derivatives have shown that a hydroxy group at position 7 is not essential for biological activity. In fact, reports on 7-deoxypaclitaxel 4, obtained by radical deoxygenation of paclitaxel, or of a baccatin derivative, showed this compound to be of comparable, or even superior, cytotoxic activity with respect to paclitaxel.9 Its structural congruence with taxine B stimulated several groups to investigate the possibility of a semisynthetic approach to 7-deoxypaclitaxel 4, starting from taxine 6.10 As the preparative separation of taxine alkaloids is difficult, the crude taxine mixture was hydrolysed in order to obtain a well defined starting material, i.e. tetraol 7. However, the differentiation between hydroxy groups in tetraol 7 required extensive use of protective groups, leading to long synthetic sequences, and lowering considerably the overall yield.11 Recently, we reported an efficient procedure for the conversion of taxine into 7-deoxypaclitaxel, which in principle could give access not only to the title compound, but also to other derivatives suitable for further SAR studies.12 In this paper we present the results of our study giving full experimental details, together with some observations on the reactivity of the intermediates.
![]() | ||
| Scheme 2 Retrosynthetic analysis of 7-deoxypaclitaxel. | ||
We first re-examined the possibility of exploiting the favourable arrangement of functional groups in the major constituent of taxine, i.e. taxine B 6a (Scheme 3). Quaternization of crude taxine, followed by DBU-induced elimination under anhydrous conditions, afforded a mixture of 9- and 10-O-acetyl-5-O-cinnamoyltaxicine I (8 and 9),13 which could be separated by flash chromatography on a SiO2 column, and were isolated in yields of 1 and 1.7 g kg−1 of needles, respectively. Some time ago, it was reported that 6a and 6b interconvert on storage on TLC silica plates;7c,13 this finding prompted us to examine the possibility of preparative isomerization of 8 into 9. After some experimentation, we found that 8 can be isomerized into the required 10-O-acetyl derivative 9 by treatment with methanolic KOAc (43% yield at 70% conversion, without optimization), to afford a total of 2 g of 10-O-acetyl-5-O-cinnamoyltaxicine I 9 per 1 kg of dry leaves. Later on, we found it advantageous to perform the separation of regiomeric acetates at the level of carbonates 10 and 11, where 11 could be isomerized into 10 on treatment with methanolic KOAc. Alternatively, the mixture of carbonates 10 and 11 can be oxidized with Jones reagent to ketones 14 and 15, which could be easily separated by dry-flash chromatography. We hoped that 11 could be converted into 14 by the following sequence of reactions: a) protection of 11 as the O-ethoxyethyl derivative 12; b) hydrolysis of C-9 acetate in 12 to give alcohol 13, and c) direct oxidation of 13 into the required ketone 14 [eqn. (1)]. These attempts failed, however, as the C-9 acetate in 12 proved resistant towards hydrolysis even under conditions where the cyclic carbonate was cleaved.
![]() | ||
| Scheme 3 Transformation of taxine into suitable precursor for 7-deoxypaclitaxel synthesis. (A) previous work (ref. 10); (B) this work. Reagents and conditions: (i) MeI, THF, rt, 5 h; (ii) K2CO3, H2O, EtOH, rt, 3 h; (iii) NaOMe, MeOH, 0 °C, 16 h; (iv) column chromatography on SiO2; (v) DBU, CHCl3, rt, 1.5 h; (vi) 10% KOAc in MeOH, rt, 3 h. | ||
| (1) |
With the suitable starting compound 9 in hand, the synthesis of 7-deoxybaccatin 25 proceeded as displayed in Scheme 4. Treatment of 9 with phosgene, followed by hydrolytic work-up, furnished the cyclic carbonate 10 (90%), which was oxidized to diketone 14 (Dess–Martin periodinane: 80%; PCC: 75%; Jones reagent: 50%), thus establishing the final functionalization of the ‘upper’ part of the molecule. Elaboration of the oxetane ring was envisaged to proceed via triol 5. For that purpose a method for selective removal of the cinnamoyl chain was needed, as the simultaneous cleavage of the C-10 acetate (before, or after, the transformation of the C4–C20 double bond into a vicinal diol) would create a system with two secondary hydroxy groups that might be difficult to differentiate, and which would certainly require additional protective steps. To this end, allylic cinnamate 14 was oxidized with OsO4–NMO, as it was anticipated that intramolecular hydrogen bonding in 16 should activate the dihydroxyphenylpropionate ester towards hydrolysis under very mild conditions; in this way, both the desired functional transformation of the taxane core and the activation of the ‘side-chain’ would be accomplished in a single step. Although some concern existed about the stereochemical outcome of the osmylation, to our pleasure the reaction proceeded with complete stereoselectivity, and in quantitative yield. TLC monitoring of the reaction mixture indicated that the reaction proceeded via two intermediates of close polarity (major: less polar; minor: more polar), which, during the course of the reaction, converged to a single product 16. When the hydroxylation was interrupted before completion, these intermediates could be isolated and identified as diols 26 and 27 (Scheme 5). However, this bifurcate reactivity does not compromise diastereoselectivity, as a single isomer 16 is obtained. Whether intramolecular assistance is included in the reaction mechanism or not remains unclear.
![]() | ||
| Scheme 4 Reagents and conditions: (i) COCl2 (20 equiv.), CH2Cl2, 0 °C, 20 min; then Et2O, H2O, imidazole (cat.), 0 °C, 30 min; (ii) Dess–Martin periodinane (2 equiv.), CH2Cl2, CF3CO2H (cat.), rt, 12 h; (iii) OsO4 (cat.), NMO, THF, H2O, rt, 4 h; (iv) 10% KOAc in MeOH, reflux, 30 min; (v) TBDMSCl, Et3N, DMAP (cat.), CH2Cl2, rt, 24 h; (vi) MsCl, Py, 0 °C to rt, 24 h; (vii) 7% HF in CH3CN, rt, 7 h; (viii) iPr2NEt (7 equiv.), toluene, reflux, 39 h; (ix) Ac2O (7 equiv.), DMAP (14 equiv.), CH2Cl2, rt, 4 h; (x) PhLi (10 equiv.), THF, −78 °C, 0.5 h; then Ac2O, DMAP, CH2Cl2, rt, 1 h; (xi) NaBH4 (excess), MeOH, rt, 3 h. | ||
![]() | ||
| Scheme 5 Reagents and conditions: (i) OsO4 (cat.), NMO, THF, H2O, rt; (ii) KOAc, 18-cr-6 (cat.), CH2Cl2, rt, 15 h. | ||
The conversion of 14 to 16 brought about the expected modification in the side-chain ester reactivity, as indicated by its proclivity towards spontaneous migration of the ester side-chain from O-5 to O-20 on storage in solution (MeOH, THF
) at room temperature; on treatment with KOAc in CH2Cl2, in the presence of a catalytic amount of 18-crown-6, tetraol 16 was converted into 28 quantitatively. Submitting 16 to K2CO3 or NaHCO3 in aq. methanol at 0 °C induced very rapid hydrolysis of the dihydroxypropanoate ester, but hydrolysis of the C-10 acetate also occurred under these conditions. Eventually, refluxing of 16 with methanolic KOAc afforded the desired triol 5 in 94% yield.†
Stereochemical identity of 5 was verified by an NOE-difference experiment on its C-20 acetate 29 (in 1H NMR spectrum of 5 some signals overlapped) which confirmed the proposed (i.e., correct) configuration at C-4 (Fig. 1). Optically pure (2S,3R)-(−)-(methyl 2,3-dihydroxy-3-phenylpropanoate) 17 was isolated as the side product of this reaction, and further transformed into the paclitaxel side-chain 18 according to the previously published procedure.14 Silylation of the primary alcohol in 5 and mesylation of 5-OH were accomplished in the usual way (80% over two steps), but TBAF- induced deprotection of 20 gave 21 in only 40% yield, indicating its instability under basic conditions, and possible complications in the cyclization step. A higher yield of 21 was obtained when the deprotection was carried out under acidic conditions (HF–CH3CN, 80%). Tetrabutylammonium acetate in refluxing butanone (‘standard’ reagent for the construction of the oxetane ring on structurally similar systems)
10a caused the rapid decomposition of 21, as did many other reagents.‡ After considerable experimentation it was found that the oxetane-ring closure could be accomplished with Hunig’s base in refluxing toluene (82%). Acetylation of tertiary alcohol 22 afforded 23 in moderate yield (66%). Installation of the requisite 1-hydroxy, 2-benzoate functionalities, and re-acetylation of the secondary alcohol at C-10 were performed following a known protocol,15 furnishing 24 in 77% yield. Finally, reduction of enone 24 with NaBH4 gave rise to 7-deoxybaccatin III 25 (83% at 85% conversion),9b,c,10d thus completing the synthesis of this compound from 10-O-acetyl-5-O-cinnamoyltaxicine I 9 in 15% overall yield (unoptimized).
![]() | ||
| Fig. 1 NOE correlations in 29. | ||
7-Deoxybaccatin is a direct precursor of paclitaxel analogues, as exemplified by its conversion into 7-deoxypaclitaxel (Scheme 6): esterification of 25 with acid 30
16 (30 3 equiv., DCC 3 equiv., DMAP 0.5 equiv., THF, rt, 1 h), followed by acidic hydrolysis (5% p-TsOH in MeOH, rt, 1 h) afforded the title compound 4
9 in 74% yield.
We believe that the described chemistry offers an efficient pathway for the preparation of new paclitaxel derivatives, and points to the naturally abundant taxane diterpene fraction – taxine – as a valuable starting material for further semisynthetic studies.
Taxine 6 was obtained according to the previously reported procedure,7c in yields of 5–10 g kg−1 of dry leaves.
1H and 13C NMR spectra for compounds 8 and 9 were identical to those previously reported.13
17 (4.23 g) in CH2Cl2 (210 ml) was added a solution of 10 (1.6 g) in CH2Cl2 (50 ml), followed by a catalytic amount of trifluoroacetic acid (10 μl). The reaction mixture was stirred for 12 h at rt, then saturated aq. NaHSO3 (200 ml) was added, and stirring was continued for 45 min. The organic phase was washed successively with aq. NaHSO3 and brine, dried over anhydrous MgSO4, evaporated to dryness and purified by dry-flash chromatography (eluent: petroleum spirit–acetone 85∶15) to give ketone 14 (1.27 g, 80%) as a white foam, mp 192–193 °C (Calc. for C32H34O9: C, 68.31; H, 6.09. Found: C, 67.99; H, 5.84%); [α]20D +74.3 (c 1, CHCl3), IR (KBr) νmax 2928, 1821, 1752, 1713, 1687, 1638, 1229, 1204, 1164, 1020; 1H NMR δ 7.73 (m, 2H), 7.68 (d, J 16, 1H), 7.46 (m, 3H), 6.71 (s, 1H), 6.32 (d, J 16.1, 1H), 5.62 (d, J 1.4, 1H), 5.43 (br t, 1H), 5.40 (br s, 1H), 4.43 (d, J 5.5, 1H), 3.82 (d, J 5.5, 1H), 3.0 (s, 2H), 2.33 (s, 3H), 2.26 (s, 3H), 2.0 (m, 2H), 1.55 (m, 2H), 1.38 (s, 3H), 1.29 (s, 3H), 1.28 (s, 3H); 13C NMR δ 202.94 (C), 196.05 (C), 169.23 (C), 165.72 (C), 152.07 (C), 148.86 (C), 146.42 (CH), 141.85 (C), 138.72 (C), 134.15 (C), 130.62 (CH), 128.96 (2 × CH), 128.38 (2 × CH), 118.69 (CH2), 116.91 (CH), 88.6 (C), 80.99 (CH), 76.55 (CH), 76.19 (CH), 57.03 (C), 44.54 (CH), 41.31 (C), 40.51 (CH2), 31.54 (CH3), 31.01 (CH2), 27.33 (CH2), 20.67 (CH3), 18.32 (CH3), 15.28 (CH3), 14.75 (CH3), MS/CIisobutane 563 (M + 1).
18 (0.16 ml) dropwise. After the addition was complete the reaction mixture was stirred for 15 min at 0 °C then allowed to reach rt, and was stirred for an additional 30 min. Aq. NaHSO3 was added, the mixture was extracted with CH2Cl2, and the extract was dried over anhydrous MgSO4 and evaporated to dryness. Purification by dry-flash chromatography (eluent benzene–EtOAc 975∶25) afforded compound 14 (100 mg, 50%) physical data as above.
Physical data for 5: white foam, mp >240 °C (decomp.) (Calc. for C23H30O10·H2O: C, 56.97; H, 6.60. Found: C, 57.28; H, 6.57%); [α]20D −108 (c 1, CHCl3), IR (KBr) νmax 3468, 2957, 1794, 1747, 1707, 1681, 1230, 1210, 1022; 1H NMR (in CDCl3 + D2O) δ 6.66 (s, 1H), 4.36 (d, J 4.63, 1H), 4.0 (d, J 10.5, 1H), 3.95 (d, J 19.37, 1H), 3.72 (br s, 1H), 3.57 (d, J 10.52, 1H), 3.52 (d, J 4.64, 1H), 2.84 (d, J 19.37, 1H), 2.23 (s, 3H), 2.22 (s, 3H), 2.2–1.6 (m, 3H), 1.35 (s, 3H), 1.25 (m, 1H), 1.22 (2 × s superimposed, 6H); 13C NMR δ 203.52 (C), 197.73 (C), 169.26 (C), 153.11 (C), 148.21 (C), 143.09 (C), 89.08 (C), 82.27 (CH), 75.74 (CH), 73.97 (CH), 70.0 (CH), 62.18 (CH2), 56.26 (C), 42.06 (CH), 41.58 (C), 40.52 (CH2), 31.1 (CH3), 28.99 (CH2), 24.16 (CH2), 20.7 (CH3), 18.57 (CH3), 16.67 (CH3), 14.28 (CH3), MS/CIisobutane 467 (M + 1).
Spectral data for 19: IR (film) νmax 3475, 3023, 2931, 1752, 1711, 1692, 1606, 1259, 1226, 1154, 1125, 1026; 1H NMR δ 6.66 (s, 1H), 4.31 (d, J 4.7, 1H), 3.98 (d, J 19.4, 1H), 3.97 (d, J 9.5, 1H), 3.96 (s, 1H, OH), 3.64 (br s, 1H), 3.53 (d, J 9.5, 1H), 3.52 (d, J 4.7, 1H), 2.83 (d, J 19.4, 1H), 2.78 (s, 1H, OH), 2.23 (s, 3H), 2.22 (s, 3H), 2.2–1.65 (m, 4H), 1.34 (s, 3H), 1.21 (s, 6H), 0.9 (s, 9H), 0.14 (s, 3H), 0.11 (s, 3H); 13C NMR δ 203.52 (C), 197.58 (C), 169.11 (C), 152.43 (C), 147.97 (C), 142.91 (C), 88.36 (C), 81.84 (CH), 75.65 (CH), 73.87 (C), 69.61 (CH), 62.98 (CH2), 56.12 (C), 41.52 (CH), 41.48 (C), 40.46 (CH2), 30.99 (CH3), 28.94 (CH2), 25.64 (CH3), 23.98 (CH2), 20.58 (CH3), 18.43 (CH3), 18.05 (C), 16.59 (CH3), 14.17 (CH3), −5.64 (CH3), MS/CIisobutane 581 (M + 1).
Compound 23 was obtained as a white solid, mp 275 °C (decomp.); [α]25D +17.1 (Calc. for C25H30O10: C, 61.22; H, 6.16. Found: C, 61.04; H, 6.00%); IR (KBr) νmax 3019, 2992, 2956, 1822, 1747, 1731, 1710, 1684, 1232, 1206, 1026, 772; 1H NMR δ 6.54 (s, 1H), 4.95 (d, J 8.74, 1H), 4.64 (d, J 8.91, 1H), 4.51 (d, J 5.46, 1H), 4.46 (dd, J1 9.1, J2 0.73, 1H), 3.47 (d, J 5.28, 1H), 2.96 (d, J 20, 1H), 2.84 (d, J 20, 1H), 2.25 (s, 3H), 2.1 (s, 3H), 2.08 (s, 3H), 2.08–1.95 (m, 2H), 1.77 (s, 3H), 1.75–1.65 (m, 2H), 1.34 (s, 3H), 1.25 (s, 3H); 13C NMR δ 203.41 (C), 195.58 (C), 170.38 (C), 169.1 (C), 152.09 (C), 149.98 (C), 142.89 (C), 88.23 (C), 83.27 (CH), 80.77 (CH), 79.9 (C), 76.1 (CH), 75.65 (CH2), 55.32 (C), 42.25 (CH), 41.37 (C), 39.8 (CH2), 31.92 (CH2), 31.25 (CH3), 27.0 (CH2), 21.55 (CH3), 20.67 (CH3), 18.52 (CH3), 14.47 (CH3), 14.12 (CH3), MS/CIisobutane 491 (M + 1).
Spectral data for compound 24 were identical with those previously reported.10d
Physical data for compound 25 were identical with those previously reported.9b,c,10d
16 (137 mg) in THF (5 ml) was added DCC (71 mg). After 5 min a solution of 7-deoxybaccatin III 25 (65 mg) in THF (2 ml) was added, followed by DMAP (7 mg), and stirring was continued for 1 h, when TLC monitoring indicated completion of the reaction. The resulting suspension was filtered through a cotton plug, diluted with EtOAc, washed successively with aq. NaHCO3 and brine, dried over anhydrous MgSO4, and evaporated to dryness. The product was roughly purified by dry-flash chromatography (eluent benzene–EtOAc 85∶15), then was dissolved in methanol (10 ml), toluene-p-sulfonic acid (0.5 g) was added, and the mixture was stirred 2 h at rt, diluted with EtOAc, washed successively with aq. NaHCO3 and brine, dried over anhydrous MgSO4 and evaporated to dryness, and the residue was purified by dry-flash chromatography (eluent n-hexane–acetone 3∶1 → 3∶2). A second purification by flash chromatography (eluent CH2Cl2–methanol 95∶5) afforded 7-deoxypaclitaxel 4 (70 mg, 74% as a white powder, 1H NMR spectrum was identical with that previously reported.913C NMR spectrum of this compound has not been reported: δ 206.18 (C), 172.71 (C), 170.27 (C), 169.69 (C), 167.12 (C), 167.07 (C), 140.25 (C), 137.99 (C), 133.64 (CH), 133.60 (CH), 133.50 (C), 131.93 (CH), 130.22 (CH), 129.27 (C), 128.98 (CH), 128.71 (C), 128.65 (CH), 128.29 (CH), 127.05 (CH), 126.96 (CH), 84.50 (CH), 82.04 (C), 78.93 (C), 76.62 (CH2), 75.67 (CH), 74.07 (CH), 73.20 (CH), 72.19 (CH), 54.93 (CH), 52.75 (C), 44.99 (CH), 42.84 (C), 35.84 (CH2), 35.09 (CH2), 26.95 (CH2), 26.20 (CH3), 22.64 (CH3), 21.40 (CH3), 20.78 (CH3), 14.68 (CH3), 14.37 (CH3).
Compound 26: white foam, 1H NMR δ 7.5–7.3 (m, 5H), 6.65 (s, 1H), 5.64 (d, J 1.41, 1H), 5.34 (s, 1H), 5.25 (br s, 1H), 4.88 (d, J 3.94, 1H), 4.41 (d, J 5.67, 1H), 4.22 (d, J 3.93, 1H), 3.53 (d, J 5.52, 1H), 2.98 (d, J 19.67, 1H), 2.87 (d, J 20.4, 1H), 2.26 (s, 3H), 2.25 (s, 3H), 2.12–1.85 (m, 2H), 1.6–1.4 (m, 2H), 1.36 (s, 3H), 1.26 (s, 3H), 1.24 (s, 3H); 13C NMR δ 202.77 (C), 196.76 (C), 171.89 (C), 169.25 (C), 151.99 (C), 148.95 (C), 141.92 (C), 138.81 (C), 137.83 (C), 128.32 (CH), 128.27 (CH), 126.39 (CH), 120.04 (CH2), 88.56 (C), 81.0 (CH), 78.97 (CH), 76.04 (CH), 75.18 (CH), 75.09 (CH), 56.85 (C), 44.88 (CH), 41.30 (C), 40.60 (CH2), 31.61 (CH3), 31.03 (CH2), 27.04 (CH2), 20.70 (CH3), 18.34 (CH3), 15.41 (CH3), 14.97 (CH3), MS/CIisobutane 597 (M + 1).
R: Science and Applications, ed. M. Suffness, CRC Press, Boca Raton, 1995,
Search PubMedand references cited therein;;
(c) K. C. Nicolaou, W.-M. Dai and R. K. Guy, Angew. Chem., Int. Ed. Engl., 1994, 33, 15 CrossRef
Angew. Chem., 1994, 106, 38.
Search PubMed.
)..Footnotes |
| † In this reaction yields varied between 60 and 94%. It is interesting to compare these reaction conditions with those needed for the hydrolysis of structurally related cinnamic esters, which require 20 M NaOH at reflux, for many days.10a,b For a study on selective hydrolysis of the paclitaxel side-chain, with the retention of C-10 acetate, see ref. 19. |
‡ Other reagents tried include: TBAF–THF, KF–18-crown-6–THF, KOAc–MeOH, NaH–THF, DBU and pyridine. Attempts to effect the cyclization via the corresponding triflate 3c gave a complex mixture from which the oxetane was isolated in 4% yield. |
| This journal is © The Royal Society of Chemistry 2000 |