286
J. Org. Chem. 2002, 67, 286-289
Revision of Absolute Configuration of Enantiomeric (Methylenecyclopropyl)carbinols Obtained from (R)-(-)- and (S)-(+)-Epichlorohydrin and Methylenetriphenylphosphorane. Implications for Reaction Mechanism and Improved Synthesis of Antiviral Methylenecyclopropane Analogues of Nucleosides
Scheme 1
Scheme 2
Xinchao Chen and Jiri Zemlicka* Department of Chemistry, Developmental Therapeutics Program, Barbara Ann Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, Michigan 48201-1379
[email protected] Received May 21, 2001
Abstract: Absolute configurations of enantiomeric methylenecyclopropanecarbinols obtained by reaction of (R)- and (S)-epichlorohydrin 5 with methylenetriphenylphosphorane or resolution of the corresponding oxaphospholane 6 via a salt with L-(+)-tartaric acid and subsequent Wittig transformation with formaldehyde were revised. The (-)-oxaphospholane 6 has the S,S and (-)-(methylenecyclopropyl)carbinol (4) the R configuration. The configurations of (+)-6 and (+)-4 are then R,R and S, respectively. These assignments are in accord with an initial attack of phosphorane at the oxirane ring of epichlorohydrin. An improved preparation of key enantiomeric intermediates (R)-1a and (S)-1a, important for synthesis of antiviral purine methylenecyclopropane analogues of nucleosides, is also described.
Methylenecyclopropane analogues of purine nucleosides 1 are potent antiviral agents of broad selectivity.1
During investigation of structure-activity relationships in this series of analogues, we have studied the enantioselectivity2,3 of antiviral effects of R- and S-enantiomers (R)- and (S)-1. For synthesis of both enantiomeric series, the known4 diastereoisomeric amides of (R)-phenylglycinol (R,R)-2 and (S,R)-3 were converted3 by a seven-step procedure to key intermediates (R)-(-)-1a and (S)-(+)1a, respectively (Scheme 1). Although this approach provided sufficient amounts of enantiomeric analogues (1) Zemlicka, J. In Recent Advances in Nucleosides: Chemistry and Chemotherapy; Chu, C. K., Ed.; Elsevier Science: Amsterdam, The Netherlands, in press. (2) Qiu, Y.-L.; Hempel, A.; Camerman, N.; Camerman, A.; Geiser, F.; Ptak, R. G.; Breitenbach, J. M.; Kira, T.; Li, L.; Gullen, E.; Cheng, Y.-C.; Kern, E. R.; Drach, J. C.; Zemlicka, J. J. Med. Chem. 1998, 41, 5257-5264. (3) Qiu, Y.-L.; Geiser, F.; Kira, T.; Gullen, E.; Cheng, Y.-C.; Ptak, R. G.; Breitenbach, J. M.; Drach, J. C.; Hartline, C. B.; Kern, E. R.; Zemlicka, J. Antiviral Chem. Chemother. 2000, 11, 191-202. (4) Lai, M.-t.; Liu, L.-d.; Liu, H.-w. J. Am. Chem. Soc. 1991, 113, 7388-7397.
for initial biological investigations, it was poorly suitable for a large-scale synthesis. Thus, up to 15 chromatographic separations on silica gel columns were necessary to provide diastereoisomers 2 and 3 in desirable optical purity (95%). For these reasons we sought an alternate approach that would limit the number of synthetic steps and avoid tedious chromatographic separation in an early stage of synthesis. We anticipated that readily available (R)- and (S)(methylenecyclopropyl)carbinols5 (4) could serve as starting materials for the synthesis of enantiomeric analogues 1 (Scheme 2). Racemic epichlorohydrin (5) was first alkylated with methylenetriphenylphosphorane to give oxaphospholane 6, which was resolved by crystallization of its salt with L-(+)-tartaric acid. Treatment of the diastereoisomeric (-)- and (+)-tartarates with a base furnished the (S)-(+)- and (R)-(-)-oxaphospholanes6 6. Reaction of (S)-(+)-6 with paraformaldehyde gave (R)(+)-(methylenecyclopropyl)carbinol (4), whereas enantiomer (R)-(-)-6 provided (S)-(-)-4. Absolute configurations were based on a transformation7 of (R)-(-)- and (S)-(+)epichlorohydrin 5 to (R)-(-)- and (S)-(+)-oxaphospholanes (6) and then (S)-(-)- and (R)-(+)-(methylenecyclopropyl)carbinols (4). Following this protocol5 (Scheme 2) we obtained both enantiomeric methylenecyclopropylcarbinols 4 on a 50-g scale, which were first converted to the corresponding acetates 7 (Scheme 3). Addition of elements of bromine via pyridinium tribromide gave two pairs of diastereoisomeric dibromocyclopropanes 8, which were previously obtained by a different procedure.3 Diastereoisomers 8 derived from (-)-4 were used in an alkylation-elimination procedure3 with 2-amino-6-chloropurine to give the (5) Le Corre, M.; Hercouet, A.; Bessieres, B. J. Org. Chem. 1994, 59, 5483-5484. (6) Only single R and S designations were used5, 7 for compounds 6 with two centers of chirality. The absolute configurations in Scheme 2 are those of refs 5 and 7. (7) Okuma, K.; Tanaka, Y.; Yoshihara, K.; Ezaki, A.; Koda, G.; Ohta, H.; Hara, K.; Kashimura, S. J. Org. Chem. 1993, 58, 5915-5917.
10.1021/jo010511j CCC: $22.00 © 2002 American Chemical Society Published on Web 12/13/2001
Notes
J. Org. Chem., Vol. 67, No. 1, 2002 287 Scheme 3
Scheme 5
Scheme 4
Scheme 6
Z- and E-isomeric mixture of methylenecyclopropane intermediates. Deacetylation afforded, after chromatographic separation, the (Z)-(-)-isomer 1a and the corresponding E-isomer. The (-)-1a was identical (optical rotation and chiral HPLC) with the previously prepared3 (R)-(-)-enantiomer 1a. However, the (-)-(methylenecyclopropyl)carbinol (4), which was previously assigned the S configuration,5,7 should give the S-enantiomer 1a. In a similar fashion, diastereoisomeric dibromides 5 derived from (+)-(methylenecyclopropyl)carbinol (4), presumably of an R configuration, provided the (+)-enantiomer 1a, which was identical with the (S)-(+)-enantiomer 1a. It should be stated that the absolute configuration of analogues 1 was determined by X-ray diffraction of (R)(-)-synadenol (1b) obtained by deamination of racemic compound with adenosine deaminase and by correlation2 of this product with material synthesized from diastereoisomer (R,R)-2 (Scheme 1). To clarify this discrepancy, we repeated the procedure used7 for establishing the absolute configuration of carbinol (-)-4 (Scheme 4). The latter was converted to mesylate (-)-9, which was then transformed to nitrile (-)-10. The optical rotations of both compounds agreed with the described values.7 In contrast, nitrile (-)-10 was converted to methylenecyclopropaneacetic acid (11) with an optical rotation opposite ([R]20D +8.7°) to the reported value7 ([R]D -8.1°). The latter optical rotation was used to determine the R configuration of 11sand hence the S configuration of (-)-4sby comparison with the known values4,8 of (R)-(-)-11 and (S)-(+)-11. It is then clear that the reported5,7 absolute configurations of (S)-(-)- and (R)-(+)-(methylenecyclopropyl)carbinol 4 must be revised. Thus, the (-)-enantiomer 4 has an R configuration and (+)-4 is then assigned an S configuration. These findings are in agreement with the fact that (S)-(+)-methylenecyclopropane analogue 1a was obtained from (S)-(+)-carbinol 4 (Scheme 3), whereas (R)(-)-4 led to (R)-(-)-1a, confirming thus the previous2,3 assignment of absolute configuration of analogues 1. It
is also important to note that, independently, an S configuration was assigned to carbinol (+)-4 obtained by a different procedure from (S)-glycidol tosylate.9 These results have implications for the mechanism of alkylation of epichlorohydrin (5) with methylenetriphenylphosphorane (Scheme 5). An attack of the phosphorane at the halogen-carrying carbon of (R)-5 gives rise to the (S)-triphenylphosphonium salt 12, which then undergoes an intramolecular alkylation via carbanion (S)-13 to the corresponding (R,R)-oxaphospholane 6 or its betaine form (R,R)-14. A Wittig reaction with formaldehyde then leads to (S)-(methylenecyclopropane)carbinol (4). In contrast, an attack at the oxirane ring of (R)-5 results in formation of intermediate (R)-15 (Scheme 6). In the next step, a new oxirane ring is closed, compound (R)12. An intramolecular attack of the phosphorane portion at the oxirane moiety via carbanion (R)-13 will then furnish oxaphospholane (S,S)-6 or the corresponding
(8) Baldwin, J. E.; Ostrander, R. L.; Simon, C. D.; Widdison, W. C. J. Am. Chem. Soc. 1990, 112, 2021-2022.
(9) Baldwin, J. E.; Adlington, R. M.; Bebbington, D.; Russell, A. T. Tetrahedron 1994, 50, 12015-12028.
288
J. Org. Chem., Vol. 67, No. 1, 2002
betaine (S,S)-14 in a fashion similar to that described in Scheme 5 (compound (S)-12). Reaction with formaldehyde gives then carbinol (R)-4. Support for the first mechanism (Scheme 5) was based on isolation7 of triphenylphosphonium salt (S)-12 (Cl ) I), a presumed product of the initial attack of phosphorane at the chloromethyl group of (R)-5. However, this intermediate is enantiomeric with (R)-12 (Cl ) I), an intermediate in the sequence described in Scheme 6. The configurational assignment of (S)-12 (Cl ) I) relied on conversion to (presumed) (R,R)-oxaphospholane 6 and (S)-(methylenecyclopropyl)carbinol (4). In contrast, our findings indicate that only the mechanism described in Scheme 6 is compatible with the formation of R-configured (methylenecyclopropyl)carbinol (4) from (R)-epichlorohydrin (5) or (S,S)-oxaphospholane 6 as a starting material. A similar mechanism (initial attack at an oxirane ring) was observed in alkylation of (S)-glycidol tosylate with phenyl trimethylsilylethyl sulfone.9 Interestingly, a preferential opening of the oxirane ring of (racemic) epichlorohydrin (5) (intermediate 15) was observed10 in the reaction with methylenetriphenylphosphorane in toluene catalyzed by BuLi. The following Wittig transformation with benzaldehyde gave the corresponding benzylidenecyclobutanol in 62% yield. The different outcome (formation of cyclobutane vs cyclopropane5 ring) is probably a result of the stronger base used (BuLi vs tBuOK). The described approach (Scheme 3) significantly simplifies the synthesis of key enantiomeric methylenecyclopropane intermediates (R)-(-)-1a and (S)-(+)-1a.
Notes
Resolution of (()-Oxaphospholane 6. The reaction of racemic epichlorohydrin 5 with methylenetriphenylphosphorane was performed as described.5 Methyltriphenylphosphonium bromide was used for generation of the phosphorane instead of iodide to give oxaphospholane 6. The resolution5 of a salt of 6 with L-(+)-tartaric acid (1600 g, 3.32 mol) was also modified. The latter was recrystallized three times from EtOH to give the (-)-tartarate of 6 as an ethanolate (554 g, 63%): mp 112-116 °C; [R]20D -60.4° (c 1, MeOH), lit. -71.92° (anhydrous compound); 1H NMR (DMSO-d6) δ 0.97-1.05 (m, 1H), 1.03 (t, 3H, CH3 of EtOH), 1.71-1.78 (m, 1H), 2.12-2.25 (m, 2H), 2.78-2.85 (m, 1H), 3.27 (dd, 1H), 3.2 (q, 2H, CH2 of EtOH), 3.77 (s, 2H), 5.27 (bs, 5H), 7.70-7.74 (m, 6H), 7.84-7.91 (m, 9H); 13C NMR 6.2 (d, J ) 86.1 Hz), 10.5, 19.2 (CH3 of EtOH), 24.6, 56.7 (CH2 of EtOH), 59.2, 77.8, 120.8 (d, J ) 88.3 Hz), 130.6 (d, J ) 12.6 Hz), 134.6 (d, J ) 10.4 Hz), 135.5, 175.0; 31P NMR 25.3. The product from the mother liquor after the first crystallization was recrystallized four times from MeCN to afford the (+)-tartarate of 6 (495 g, 62%): mp 183.5-185 °C; [R]20D 82.0° (c 1, MeOH), lit.5 71.42°. The 1H, 13C, and 31P NMR spectra corresponded to those of (-)-tartarate 6. (R,R)-(+)-Oxaphospholane 6. NaOH (2 M, 1620 mL, 3.24 mol) was added with stirring to a suspension of (-)-tartarate 6 (ethanolate, 554 g, 1.046 mol) in CH2Cl2 (3600 mL) at room temperature. The stirring was continued for 1.5 h, the organic phase was separated, dried (Na2SO4), and evaporated to give product (R,R)-(+)-6 (308.5 g, 90%): mp 128.5-130 °C; [R]20D 131.6° (c 1, CHCl3), lit.5 129.5°; 1H NMR (CDCl3) δ 1.19-1.25 (m, 1H), 1.36 (ddd), 1.39-1.43 (m, 1H), 1.48 (ddd), 1.75-1.80 (m, 1H), 3.04 (dd, 1H), 3.60 (d, 1H), 7.21-7.28 (m, 9H), 7.477.50 (m, 6H); 13C NMR 7.3 (d, J ) 4.6 Hz), 12.0 (d, J ) 139 Hz), 15.0, 58.1, 127.6, 127.7, 127.9, 131.2 (d, J ) 8.4 Hz); 31P NMR -56.4.
(S,S)-(-)-Oxaphospholane 6. The procedure for (R,R)-(+)-6 was used with (+)-tartarate 6 (495 g, 1.03 mol) to give (S,S)-()-6 (301.5 g, 88.5%): mp 125-127 °C; [R]20D -137.8° (c 1, CHCl3), lit.5 -131.9°. The 1H, 13C, and 31P NMR spectra were identical with those of (R,R)-(+)-6. (S)-(+)-(Methylenecyclopropyl)carbinol (4). A mixture of (R,R)-(+)-oxaphospholane 6 (308.5 g, 0.929 mol) and paraformaldehyde (55.30 g, 1.84 mol) was heated in sulfolane (215 mL) at 100-110 °C for 1 h. After cooling, crude carbinol was distilled in vacuo (bp 50-85 °C/20 Torr, 66.26 g). Redistillation gave pure (S)-(+)-carbinol 4: bp 59-66 °C/20 Torr, 54.28 g (69.5%); [R]20D 7.4° (c 2, CHCl3), lit., as (R)-(+)-enantiomer,5,7 6.75° (c 1) and 5.1° (c 2.2); [R]20D 42.7° (c 1.8, Et2O), lit.9 [R]20D 47.8° (c, 0.95, Et2O). The 1H and 13C NMR spectra were identical with those of (()-4. (R)-(-)-(Methylenecyclopropyl)carbinol (4). The procedure described for (S)-(+)-carbinol 4 was performed with (S,S)(-)-oxaphospholane 6 (301.5 g, 0.91 mol) to give (R)-(-)-4 (55.1 g, 72%): [R]20D -6.9° (c 2, CHCl3), lit., as (S)-(-)-enantiomer,5,7 -6.98° (c 1) and -5.1° (c 2); [R]20D -42.7° (c 0.95, Et2O). The 1H and 13C NMR spectra were identical with those of (()-4. (S)-(+)-(Methylenecyclopropyl)carbinol Acetate (7). The procedure reported11 for (()-7 was modified as follows. Acetic anhydride (13.2 g, 130 mmol) was added dropwise within 30 min to a solution of (S)-(+)-(methylenecyclopropyl)carbinol 4 (10.35 g, 123 mmol) and pyridine (11.7 g, 148 mmol) in CH2Cl2 (40 mL) with stirring at room temperature. The stirring was continued for 3 h and the mixture was worked up as described11 for (()-7. Distillation afforded (S)-(+)-7 (bp 68-70 °C/20 Torr, 13.16 g, 85%): [R]20D 22.9° (c 2.0, CHCl3). The 1H NMR spectrum was identical with that of (()-7. (R)-(-)-(Methylenecyclopropyl)carbinol Acetate (7). The procedure for the (S)-(+)-enantiomer 7 was followed on a 0.32 mol scale of (R)-(-)-4 to give (R)-(-)-7 (86%): [R]20D -24.2° (c 2, CHCl3). The 1H NMR spectrum was identical with that of (()-7. (1R)-Acetoxymethyl-(2R)-bromo-2-bromomethylcyclopropane and (1R)-Acetoxymethyl-(2S)-bromo-2-bromomethylcyclopropane [(1R,2R)- and (1R,2S)-8]. Pyridine· HBr3 (2.42 g, 7.56 mmol) was added to a solution of (S)-(+)-7 (0.635 g, 5.04 mmol) in CH2Cl2 (20 mL) at -10 °C with stirring. The stirring was continued at 0 °C for 3 h. Ether (60 mL) was added, the solids were filtered off, and the organic phase was washed with a saturated aqueous solution of Na2S2O3 followed by NaHCO3. After drying (Na2SO4) ether was evaporated to give product (1R,2R)- + (1R,2S)-8 as a syrup (1.274 g, 88%) that was identical (1H NMR, except isomeric composition) with the compound prepared by a different method.3 (1S)-Acetoxymethyl-(2R)-bromo-2-bromomethylcyclopropane and (1S)-Acetoxymethyl-(2S)-bromo-2-bromomethylcyclopropane [(1S,2R)- and (1S,2S)-8]. The abovementioned procedure was performed with (R)-(-)-7 to give diastereoisomers (1S,2R)- + (1S,2S)-8, whose 1H NMR spectrum corresponded to that of the previously obtained compound.2 (Z,S)-(+)-2-Amino-6-chloro-9-[(2-hydroxymethyl)cyclopropylidene]methylpurine (1a). The described procedure3 was modified. A mixture of diastereoisomers (1R,2R)- and (1R,2S)-8 (4.634 g, 16.2 mmol), Cs2CO3 (31.6 g, 97 mmol), and 2-amino-6-chloropurine (3.30 g, 19.4 mmol) in DMF (200 mL) was stirred at 70-80 °C under N2 for 40 h. The solids were filtered off, and the solvent was evaporated to give isomeric mixture of acetates (Z/E ) 1.5/1, 3.04 g, 64.5%) after chromatography on a silica gel column (2-5% MeOH in CH2Cl2). The 1H and 13C NMR spectra corresponded to authentic material.3 Acetates (2.945 g, 10.05 mmol) and K2CO3 (1.385 g, 10 mmol) were stirred in CH2Cl2/MeOH/H2O ) 5/2/1 (195 mL) at room temperature for 4.5 h. Acetic acid was added (0.75 g, 12.1 mmol), the solvents were evaporated, and Z/E-isomers were separated by chromatography on a silica gel column using 5-7% MeOH in CH2Cl2 to give Z-isomer (S)-(+)-1a (1.349 g, 50%): mp 185186 °C, lit.3 188-191 °C; [R]20D 73.7° (c 0.35, DMF), lit.3 [R]25D 74.8° (c 0.32, DMF); identical by chiral HPLC with an authentic
(10) Okuma, K.; Tsubakihara, K.; Tanaka, Y.; Koda, G.; Ohta, H. Tetrahedron Lett. 1995, 36, 5591-5594.
(11) Guan, H.-P.; Ksebati, M. B.; Cheng, Y.-C.; Drach, J. C.; Kern, E. R.; Zemlicka, J. J. Org. Chem. 2000, 65, 1280-1290.
Experimental Section
Notes sample.3 The (E,S)-(+)-isomer (879 mg, 35%) was also identical with the previously obtained compound.3 (Z,R)-(-)-2-Amino-6-chloro-9-[(2-hydroxymethyl)cyclopropylidene]methylpurine (1a). A mixture of intermediary acetates was prepared from diastereoisomers (1S,2R)- and (1S,2S)-8 as reported.3 Deacetylation was perfomed as described above for enantiomer (S)-(+)-1a. The Z-isomer (R)-(-)-1a: mp 188-190 °C, lit.3 185-187 °C; [R]20D -75.6° (c 0.65, DMF), lit.3 [R]25D -76.1° (c 0.36, DMF); identical by chiral HPLC with an authentic sample.3 (1R)-(-)-(Methylenecyclopropyl)carbinol mesylate (9) was prepared as described:4,7 [R]20D -35.9° (c 2.0, CHCl3), lit.7 [R]D -33.6° (c 2.0, CHCl3). (1S)-(-)-Methylenecyclopropaneacetonitrile (10). The literature procedures4,7 were modified as follows. A mixture of mesylate 9 (1.501 g, 9.26 mmol) and KCN (905 mg, 13.9 mmol) in DMF (15 mL) was heated at 60 °C for 16 h. After cooling, water was added and the solution was extracted with ether (3 × 15 mL). The organic phase was washed with water and brine, whereupon it was dried (Na2SO4). Ether was removed by evaporation at atmospheric pressure and the residue was distilled to give nitrile 10 (bp 65-70 °C/25 Torr, 495 mg, 58%): [R]20D -22.9° (c 1.4, CHCl3), lit.7 [R]D -27.3° (c 1.5, CHCl3). The 1H and 13C NMR spectra corresponded to the reported data.4 (1S)-(+)-Methylenecyclopropaneacetic Acid (11). The described protocol4,7 was modified as follows. DIBALH in CH2Cl2
J. Org. Chem., Vol. 67, No. 1, 2002 289 (1 M, 3.2 mL, 3.20 mmol) was added dropwise to a solution of nitrile 10 (275 mg, 3.0 mmol) in CH2Cl2 (15 mL) at -78 °C with stirring under N2. The stirring was continued for 3 h, whereupon the reaction was quenched with a saturated solution of NH4Cl (5 mL). The mixture was stirred for 10 min at room temperature and it was acidified with 2 M HCl. The organic layer was separated and the aqueous portion was extracted with ether. The combined organic phase was dried (Na2SO4) and evaporated to give crude (1S)-methylenecyclopropaneacetaldehyde (194 mg, 65%) as an oil. Oxidation of this product (185 mg, 1.9 mmol) in acetone (10 mL) with Jones’s reagent and subsequent workup were perfomed as described.4 Chromatography on a silica gel column using 10% ether in CH2Cl2 gave acid 11 as an oil (173 mg, 80%): [R]20D 8.7° (c 1.5, CHCl3), lit. [R]D 8.6° (c 1.6, CHCl3),4 [R]D 9.0° (c 0.56, CDCl3),8 [R]D -8.1° (c 0.5, CHCl3) for (1R)-(-)enantiomer.7 The 1H and 13C NMR corresponded to the reported spectra.4
Acknowledgment. We thank Dr. Yongchun Hou and Muhammad Umair Alam for a large-scale synthesis of racemic 6 and its salt with L-(+)-tartaric acid. This research was supported by a grant AI46390 from the National Institute of Allergy and Infectious Diseases, NIH, Bethesda, MD. JO010511J
