Stereoselective synthesis of chiral furan amino acid analogues of D- and L-serine from D-sugars

LETTER 1327 Stereoselective Synthesis of Chiral Furan Amino Acid Analogues of D- and L-Serine from D-Sugars Synthesis of Chiral Furan Amino Acids from D -Sugars Lidia Molina, Antonio J. Moreno-Vargas,* Ana T. Carmona, Inmaculada Robina* Department of Organic Chemistry, Faculty of Chemistry, University of Seville, P. O. Box 553, 41071 Seville, Spain Fax +34(954)624960; E-mail: [email protected]; E-mail: robi[email protected] Received 31 January 2006 SYNLETT 2006, No. 9, pp 1327–1330 01.06.2006 Advanced online publication: 22.05.2006 DOI: 10.1055/s-2006-941570; Art ID: G04606ST © Georg Thieme Verlag Stuttgart · New York Abstract: The synthesis of chiral furan amino acid analogues of D- and L-serine is reported. The developed methodology starting from D-xylose affords the corresponding amino acid derivative analogue of D-serine enantiomerically pure. Starting from D-arabinose, the corresponding analogue of L-serine was isolated in 92.3% enantiomeric purity. The analogue of D-serine was transformed into a stable Fmoc activated derivative ready to be incorporated into peptides. Key words: furan amino acids, cyclic sulfites, serine analogues, polyhydroxyalkyl furans, a-furfuryl amines The rational design of new peptide-based drugs requires compounds with improved stability towards proteolytic cleavage on physiological systems, and a lessening of the peptide’s intrinsic flexibility in order to obtain the effective conformation required for receptor binding. These requirements have been achieved by insertion of rigid non-peptide moieties into the appropriate site of the peptide backbone. This is a common approach to restrict the conformational degrees of freedom or stabilization of secondary structures that favors the binding to receptors.1 In this context, many structurally rigid amino acids have been designed. Among them, heteroaromatic amino acids have recently attracted much attention,2 and have been assembled into peptidomimetics for the purpose of drug discovery.3 As part of our continuing interest in the synthesis and biological applications of chiral furan amino acids,4 we present the synthesis of constrained analogues of D- and L- serine (1 and ent-1, Scheme 1) as novel scaffolds for the synthesis of non-proteinogenic peptides. Given that complex biochemical processes involve molecular recognition based not only on polar but also on hydrophobic interactions, the structure of our targets is of interest because they contain an alkyl furan moiety which increases the rigidity and hydrophobicity.5 In addition, constrained serine analogues incorporated into peptides have been reported to clarify the conformational role of the hydroxymethyl group6 and several serine analogues have been reported with this goal.7 The stereoselective introduction of an amino function into the a-position of a furan moiety has been widely studied8 due to their interest as building blocks for the synthesis of azasugar derivatives9 such as piperidine alkaloids, indolizidines and quinolizidines through the aza-Achma- towicz rearrangement.10 Here we report a novel stereoselective route for the synthesis of a-furfuryl amines and their transformation into D- and L-serine analogues. This methodology implies a nucleophilic displacement by azido anions on the intermediate sulfites that were obtained starting from D-xylose and D-arabinose, respectively (Scheme 1). Additionally, our target compounds are analogues of b-hydroxy-a-ami- no acids, which are constituents of many biologically active natural products and medicinally important compounds.11 Scheme 1 Thus, the reaction of D-xylose and D-arabinose with benzyl acetoacetate in the presence of ZnCl2 as catalyst12 afforded trihydroxypropyl furans 2 and 3, respectively, in moderate to good yield (Scheme 2). The trihydroxypropyl derivatives were obtained with good stereoselectivities which implies almost no epimerization at C-1¢ of the polyolic chain (S/R= 36 for 2 and R/S= 19 for 3). Recently, it has been reported that the reaction of aldopentoses and other sugars including disaccharides with b-dicarbon- yl compounds can be performed using CeCl3 as catalyst.13 In our hands, the condensation of D-xylose and D-arabi- nose with benzyl acetoacetate and CeCl3 (25%), provoked the total epimerization at C-1¢ after six hours of reaction, which is in contradiction with the results reported in the literature.13 Regioselective protection of the primary alcohol function in 2 afforded silyl derivative 4 after separation of the mi- nor epimer by flash chromatography (Scheme 3). For the introduction of an azide function at C-1¢ of the polyolic O COOP1 Me P2O O O S D -xylose D -arabinos e O COOH Me H2N OH O COOH Me H 2N OH 1 ent-1 O O COOP1 Me P2O O O S O Downloaded by: Universidad de Sevilla. Copyrighted material. 1328 L. Molina et al. LETTER Synlett 2006, No. 9, 1327–1330 © Thieme Stuttgart · New York chain of a tetrahydroxybutylfuran derivative, we have recently reported14 on the regioselective chlorination at C-1¢ followed by nucleophilic displacement with an azido anion. However, this procedure is poorly stereoselective giving a mixture of epimers at C-1¢ in a 2:1 ratio. We present in this letter a methodology that improves this functionalization in order to obtain a-furfuryl amino derivatives with high stereoselectivity and good yield. Reaction of 4 with thionyl chloride and triethylamine afforded the stable cyclic sulfite 5 in 92% yield, which could be even purified by column chromatography. Although nucleophilic displacements are commonly performed on key sulfate intermediates, the activated benzylic position C-1¢ in sulfite 5 is thought to be reactive enough for nucleophilic ring-opening. Treatment of 5 with NaN3/DMF at 60 °C gave a mixture of azido derivatives 6,15,16 epimers at C- 1¢, in a ratio R/S= 4, indicating that SN2 and SN1-like mechanisms participated in the displacement. In order to avoid epimerization, we decided to use the mixture TMSN3/TBAF with increased nucleophilicity and solubility of the azide anion in organic solvents. Thus, the reaction of 5 with TMSN3/TBAF in THF at room temperature gave the azido derivative 717 as unique compound in 77% yield with the concomitant removal of the silyl protecting group. This fact indicates that the reaction occurred through an SN2 mechanism allowing the total stereoselective introduction of the azide function. Oxidative cleavage of diol 7 followed by reduction of the aldehyde function with NaBH4 afforded alcohol 8 in 83% overall yield. In order to confirm that no epimerization at C-1¢ occurred in the reduction, Mosher’s ester of compound 8 was prepared. 1H NMR and 13C NMR experiments showed that only diastereomer 918 was present and no signals for the epimerized derivative could be found. Finally, hydrogenation of 8 using Pd/C (10%) as catalyst afforded the furan amino acid 1 in good yield. In a similar way, the synthesis of ent-1 was carried out starting from trihydroxypropyl furan 3 (C-1¢, R/S= 19) that was obtained from D-arabinose in 69% yield (Scheme 4). tert-Butyldiphenylsilyl protection followed by sulfite formation afforded 10 (C-1¢, R/S= 23) after purification by column chromatography (Scheme 4). Displacement reaction of 10 using the mixture TMSN3/TBAF in THF at room temperature gave the azido derivative 11 (57% after 48 h of reaction). The displacement–deprotection step for compound 10 proved to be slower than for its epimer 5. Besides, partial epimerization was detected (ratio on C-1¢, S/R=13=11/7 measured by 1H NMR in the crude mixture) which is presumably due to the participation of an SN1-like mechanism. Attempts to improve the stereoselectivity of the displacement using DMF as solvent or lower temperatures were not successful. Oxidative cleavage of 11 followed by reduction with NaBH4 afforded alcohol ent-8, that was reduced to give the amino acid derivative ent-1 in good-to-moderate yield (Scheme 4). The enantiomeric purity (92.3%) of ent-8 was determined by formation of the corresponding Mosher’s ester. S cheme 2 D -xylose OBn O O OBn O O D -arabinose +S S/R = 36 R/S = 19 ZnCl 2 , EtOH reflux, 30 min 67% 2 +ZnCl 2 , EtOH reflux, 30 min 69% O COOB n Me HO OH OH R 3 O COOB n Me HO OH OH S cheme 3 1) TBDPSCl, imidazole DMF, 0 °C 2) flash chromatography 60% SOCl2 Et3N, CH2Cl2 92% TMSN3, TBAF THF, r.t. 77% 2 1) NaIO4, MeOH 2) NaBH4 83% 9 Mosher's acid chloride pyridine, r.t. quant. R/S = 4 44% H2, Pd/C MeOH 74% 4O COOBn Me PO OH OH O COOBn Me P O O O S 56 O COOB n Me PO N3 OH O COOBn Me H O N3 OH 7 R O COOBn Me N 3 OH 8 O COOH Me H2N OH 1 O COOBn Me N3 O O Ph F3C OCH3 NaN3 DMF, 60 °C O P = tert-butyldiphenylsilyl Downloaded by: Universidad de Sevilla. Copyrighted material. LETTER Synthesis of Chiral Furan Amino Acids from D-Sugars 1329 Synlett 2006, No. 9, 1327–1330 © Thieme Stuttgart · New York The absolute configuration of C-1¢ in 1 and ent-1 was confirmed in the corresponding precursors 7 and 11 by transformation into their corresponding oxazolidine-2-thione derivatives 13 and 14 (Scheme 5). Appropriate NOEs confirmed the proposed structures. Amino acid derivative 1 is a suitable scaffold for its incorporation into peptides through the Fmoc solid-phase methodology. For this purpose, Fmoc derivative 15 was prepared starting from 1 following Koole’s methodology19 (Scheme 6). Moreover, reaction of 15 with PyBOP/DIEA afforded the activated ester 1620 that could be isolated by chromatography column. In summary, the synthesis of new chiral furan amino acid analogues of D- and L-serine is described starting from D- aldopentoses. In the case of the D-serine analogue 1 the method proved to be very efficient as the final product was obtained enantiomerically pure. For the L-serine analogue ent-1, the method provides the target compound with 92.3% of enantiomeric purity. Compound 1 was easily transformed into the corresponding Fmoc-activated derivative ready to be incorporated into a peptide or peptidomimetic following the Fmoc strategy for solid-phase peptide synthesis. Due to the aromatic character of the furan-3-carboxylic acid, compound 16 is stable which makes it an attractive building block for solid phase peptide synthesis as compared to most OBt esters21 of Fmocamino acids which are not stable enough and need to be generated in situ. Work is in progress to prepare peptidomimetics with conformational bias by solid phase synthesis using the new scaffold, and to extend this methodology to other constrained hetaryl amino acids. Acknowledgment We thank the Ministerio de Educación y Ciencia of Spain (CTQ2004-00649/BQU), the European Commission (FP6, TRIoH, LSHG-CT-2003-503480) and the Junta de Andalucía (FQM-345) for financial support. References and Notes (1) (a) Recent Advances in Peptidomimetics; Tetrahedron Symposia-in-Print, No. 83; Aubé, J., Ed.; Tetrahedron; 2000, 56, 50. (b) Robinson, J. A. Synlett 2000, 429. (c) Smith, A. B. III; Favor, D. A.; Sprengeler, P. A.; Guzman, M. C.; Carrol, P. J.; Furst, G. T.; Hirschmann, R. Bioorg. Med. Chem. 1999, 7, 9. (d) Hanessian, S.; McNaughton-Smith, G.; Lombart, H.-G.; Lubell, W. D. Tetrahedron 1997, 53, 12789. (e) Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1244. (f) Kinshenbaum, K.; Zuckerman, R. N.; Dill, K. A. Curr. Opin. Struct. Biol. 1999, 9, 530. (g) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173. (2) (a) Singh, R. K.; Sinha, N.; Jain, S.; Salman, M.; Nazqui, F.; Anand, N. Tetrahedron 2005, 61, 8868. (b) You, S.-L.; Razavi, H.; Kelly, J. W. Angew. Chem. Int. Ed. 2003, 42, 83. (c) You, S.-L.; Kelly, J. W. J. Org. Chem. 2003, 68, 9506. (d) You, S.-L.; Kelly, J. W. Chem. Eur. J. 2004, 10, 71. (e) Davidson, B. S. Chem. Rev. 1993, 93, 1771. (f) Wipf, P. Chem. Rev. 1993, 95, 2115. (g) Phillips, A. J.; Uto, Y.; Wipf, P.; Reno, M. J.; Williams, D. R. Org. Lett. 2000, 2, 1165. (3) (a) Bertran, A.; Pattendem, G. Heterocycles 2002, 58, 521. (b) Wang, W.; Nan, F. J. Org. Chem. 2003, 68, 1636. (c) You, S.-L. i.; Deechongkit, S.; Kelly, J. W. Org. Lett. 2004, 6, 2627. (d) Pattenden, G.; Thompson, T. Tetrahedron Lett. 2002, 43, 2459. (e) Deeley, J.; Pattenden, G. Chem. Commun. 2005, 797. S cheme 4 10 3 O COOBn Me P O O O S S 11 O COOB n Me HO N3 OH (92.3% enantiomeric purity) Mosher's acid chloride pyridine quant. H2, Pd/C MeOH ent-8ent-1 1) NaIO4, MeOH 2) NaBH4 80% 71% O COOBn Me N3 OH O COO H Me H2N OH 12 O COOBn Me N3 O O Ph F3C OCH3 O R/S = 23 57% S/R = 13 TMSN3/TBAF THF, r.t. R P = tert-butyldiphenylsilyl S cheme 5 7 ONH S T BDPSO O COOBn Me HH 1. TBDPSCl/Imidazole CH2Cl2, 0 °C 2. a) Ph3P/CH2Cl2; b) H2O 3 (Im)2C=S, CH2Cl2 11 13 14 71% 1' 2' 3' H NOEs ONH S TBDPSO O COOB n M e 1' 2' 3' H H NOE 44 79% (arrows indicate NOE enhancements) S cheme 6 1 PyBOP DIEA, DMF 78% (overall) 1. TMSCl/Py 2. FmocCl 3. H2O 16 Bt = N NN O COOH Me F mocHN OH 15 O COOB t Me FmocHN OH Downloaded by: Universidad de Sevilla. Copyrighted material. 1330 L. Molina et al. LETTER Synlett 2006, No. 9, 1327–1330 © Thieme Stuttgart · New York (4) (a) Moreno-Vargas, A. J.; Fernández-Bolaños, J. G.; Fuentes, J.; Robina, I. Tetrahedron Lett. 2001, 42, 1283. (b) Moreno-Vargas, A. J.; Jiménez-Barbero, J.; Robina, I. J. Org. Chem. 2003, 68, 4138. (5) (a) Silverman, D. N.; Lindskog, S. Acc. Chem. Res. 1988, 21, 30. (b) Greener, B.; Rose, J. Chem. Commun. 1999, 2361. (c) Lehn, J. M. Supramolecular Chemistry: Concepts and Perspectives; VCH: Weinheim, 1995. (6) Cleland, W. W.; Kreevoy, M. M. Science 1994, 264, 1887. (7) (a) Clerici, F.; Gelmi, M. L.; Gambini, A. J. Org. Chem. 1999, 64, 5764. (b) Clerici, F.; Gelmi, M. L.; Gambini, A. J. Org. Chem. 2000, 65, 6138. (c) Clerici, F.; Gelmi, M. L.; Gambini, A.; Nava, D. Tetrahedron 2001, 57, 6429. (d) Avenoza, A.; Busto, J. H.; Canal, N.; Peregrina, J. M. J. Org. Chem. 2005, 70, 330. (e) Avenoza, A.; Busto, J. H.; Canal, N.; Peregrina, J. M.; Pérez-Fernández, M. Org. Lett. 2005, 7, 3597. (8) (a) Sun, X.-L.; Kai, T.; Takayanagi, H.; Furuhata, K. Synlett 1999, 1399. (b) Bushey, M. L.; Haukaas, M. H.; O’Doherty, G. A. J. Org. Chem. 1999, 64, 2984. (c) Saaby, S.; Bayón, P.; Aburel, P. S.; Jørgensen, K. A. J. Org. Chem. 2002, 67, 4352. (d) Miyata, O.; Iba, R.; Hashimoto, J.; Naito, T. Org. Biomol. Chem. 2003, 1, 772. (9) For a review on a-furfuryl amine derivatives, see: Zhou, W.-S.; Lu, Z.-H.; Xu, Y.-M.; Liao, L.-X.; Wang, Z.-M. Tetrahedron 1999, 55, 11959. (10) (a) Ciufolini, M. A.; Wood, C. W. Tetrahedron Lett. 1986, 27, 5085. (b) Ciufolini, M. A.; Hermann, C. Y. W.; Dong, Q.; Shimizu, T.; Swaminathan, S.; Xi, N. Synlett 1998, 105. (11) (a) McMillan, J. B.; Molinski, T. F. Org. Lett. 2002, 4, 1883. (b) Willis, M. C.; Cutting, G. A.; Piccio, V. J.-D.; Durbin, M. J.; John, M. P. Angew. Chem. Int. Ed. 2005, 44, 1543. (c) Amino Acids, Peptides and Proteins, Vol.1-28; Special Periodical Reports, Chemistry Society: London, 1968-1995. (12) For similar condensations between aldohexoses and ketohexoses with b-dicarbonyl compounds see: García- González, F. Adv. Carbohydr. Chem. 1956, 11, 97. (13) Misra, A. K.; Agnihotri, G. Carbohydr. Res. 2004, 339, 1381. (14) (a) Moreno-Vargas, A. J.; Robina, I.; Demange, R.; Vogel, P. Helv. Chim. Acta 2003, 86, 1894. (b) Robina, I.; Moreno- Vargas, A. J.; Fernández-Bolaños, J. G.; Fuentes, J.; Demange, R.; Vogel, P. Bioorg. Med. Chem. Lett. 2001, 11, 2555. (15) (a) For the same reaction on a similar but unstable cyclic sulfite giving only one diasteroisomer, see ref. 8d. (b) Seki, M.; Mori, K. Eur. J. Org. Chem. 1999, 2965. (16) For reviews on cyclic sulfite reactivity, see: (a) Lohray, B. B. Synthesis 1992, 1035. (b) Byun, H.-S.; He, L.; Bittman, R. Tetrahedron 2000, 56, 7051. (17) Selected data for compound 7: [a]D20 +99 (c 0.98, CH2Cl2). 1H NMR (300 MHz, CDCl3, 298 K): d = 7.40–7.28 (m, 5 H, H-arom.), 6.75 (s, 1 H, H-4), 5.29 (s, 2 H, CH2Ph), 4.56 (d, 1 H, J1¢,2¢= 7.3 Hz, H-1¢), 4.02 (m, 1 H, H-2¢), 3.76 (dd, 1 H, J3¢a,3¢b= 11.5 Hz, J3¢a,2¢= 3.4 Hz, H-3¢a), 3.72 (dd, 1 H, J3¢b,2¢= 5.2 Hz, H-3¢b), 2.60 (s, 3 H, CH3) ppm. 13C NMR (75.4 MHz, CDCl3, 298 K): d = 163.5 (CO), 160.4, 147.5 (C- 2, C-5), 135.9, 128.6, 128.3, 128.2 (6 C-arom.), 114.2 (C-3), 110.9 (C-4), 71.9 (C-2¢), 66.2 (CH2Ph), 62.9 (C-3¢), 59.9 (C- 1¢), 14.3 (CH3) ppm. HRMS (CI): m/z calcd for C16H18N3O5 + H+: 332.1246; found: 332.1236. (18) Selected data for compound 9: [a]D20 +25 (c 2.62, CH2Cl2). 1H NMR (300 MHz, CDCl3, 298 K): d = 7.42–7.36 (m, 10 H, H-arom.), 6.67 (s, 1 H, H-4), 5.30 (s, 2 H, CH2Ph), 4.75 (t, 1 H, J1¢,2¢a=J1¢,2¢b= 6.5 Hz, H-1¢), 4.56 (d, 2 H, H-2¢a and H-2¢b), 3.55 (q, 3 H, JCH,F = 1.2 Hz, CH3O), 2.58 (s, 3 H, CH3) ppm. 13C NMR (75.4 MHz, CDCl3, 298 K): d = 166.2 (CO), 163.1 (CO), 160.5, 146.0 (C-2, C-5), 131.8–127.2 (12 C, C-arom.), 123.1 (q, 1 C, 1JC,F =288, CF 3), 114.2 (C-3), 110.5 (C-4), 84.7 [q, 1 C, 2JC,F = 28.2 Hz, (C(OMe)(CF3)Ph], 66.2 (CH2Ph), 65.0 (C-2¢), 56.5 (C-1¢), 55.5 (OCH3), 13.9 (CH3) ppm. (19) Koole, L. H.; Moody, H. M.; Broeders, N. L. H. L.; Quaedflieg, P. J. L. M.; Kuijpers, W. H. A.; van Genderen, M. H. P.; Coenen, A. J. J. M.; van der Wal, S.; Buck, H. M. J. Org. Chem. 1989, 54, 1657. (20) Experimental Procedure for the Preparation of Fmoc- Activated Amino Acid Derivative 16 from 1. To a stirred mixture of 1 (26 mg, 0.138 mmol) in dry pyridine (2 mL) at 0 °C, TMSCl (54 mL, 0.414 mmol) was dropped and the reaction mixture stirred for 45 min at r.t. Then the reaction mixture was cooled to 0 °C, 9-fluorenyl- methoxycarbonyl chloride (46 mg, 0.18 mmol) was added and the mixture stirred for 1.5 h at r.t. Afterwards, H2O (0.1 mL) was added, the mixture stirred for 1 h at r.t., and then evaporated to give 15 that was used in the next step without any purification. Crude 15 was dissolved in DMF, then DIEA (53 mL, 0.3 mmol) and PyBOP (88 mg, 0.168 mmol) were added. The mixture was stirred for 1 h at r.t., then the solution was evaporated in vacuo. The resulting residue was purified by column chromatography (EtOAc–PE, 1:1) to give 16 (57 mg, 0.108 mmol, 78%) as a white solid. [a]D20 +38 (c 0.84, CH2Cl2). 1H NMR (300 MHz, CDCl3, 298 K): d = 7.57–7.27 (m, 12 H, H-arom.), 6.74 (s, 1 H, H-4), 5.56 (d, 1 H, JNH,1¢= 7.5 Hz, NHFmoc) 4.90 (br s, 1 H, H-1¢), 4.51 (d, 2 H, J=6.5 Hz, CH 2 of Fmoc), 4.22 (t, 1 H, CH of Fmoc), 3.94 (br m, 2 H, H-2¢a and H-2¢b), 2.62 (CH3) ppm. 13C NMR (75.4 MHz, CDCl3, 298 K): d = 163.4 (CO), 159.5 (C-2), 156.1 (CO of Fmoc), 152.3 (C-5), 143.7, 143.4, 141.4, 128.8, 127.8, 127.1, 124.9, 120.4, 120.0, 107.3 (18 C, C- arom.), 108.8 (C-3), 108.4 (C-4), 67.0 (CH2 of Fmoc), 63.3 (C-2¢), 50.9 (C-1¢), 47.2 (CH of Fmoc), 14.2 (CH3). (21) The OBt esters of protected amino acids are not isolable and must be generated in situ. Only less reactive OPfp esters are commercially available, see: Novabiochem,® 2004/5 catalogue. Downloaded by: Universidad de Sevilla. Copyrighted material.