Intramolecular Radical Cyclizations of Dehydroamino Acid Derivatives

디하이드로 아미노산 유도체의 라디칼에 의한 분자 내 고리형성반응에 관한 연구

INTRODUCTION

Free radicals are reactive intermediates of considerable importance in organic chemistry. Over the past years, detailed studies of the reactivity, selectivity, and stability of many types of organic radicals have been reported.1 Recently, stereochemical control in cyclic and acyclic radical reactions has shown great promise, and a general understanding of this problem is becoming clear.2 The most popular method for this purpose is attaching a chiral auxiliary near the reaction center,3 and the incorporation of a Lewis acid for chelation control was published. 4 Finding a general rule and making a prediction possible are still important issues. Steric hindrance is considered a main factor again.2a The preferred conformations of intermediates based on steric hindrance are postulated to explain many experimental results, but still there are exceptions.

Fig. 1.

Ring construction by intramolecular free radical cyclization has proved to be particularly useful. The regio and stereochemistry of cyclization has been widely studied and a general understanding and prediction of stereoselectivity becomes possible.5 Steric effects have proved to be the most important factors controlling the stereochemistry.2e,f This methodology has been extended to various synthetic processes.2d

If there is a protein that contains dehydroaminoacids like 1 or 2 (Fig. 1) in its polypeptide chain, intramolecular radical cyclization of these would generate a new stereocenter at the α-carbon. A stereospecific structural change might be able to trigger a specific alteration in the tertiary structure of the protein. Thus, such an internal trigger which might induce structural changes can provide a bioactive peptide with a certain activation or deactivation signal which can turn on or off specific enzymatic transformations. Our initial purpose in studying the intramolecular radical cyclization of dehydroamino acid dipeptides was: a) to establish the reaction conditions necessary to effect the transformation, including the compatibility with a fully functionalized protein; b) to determine the stereoselectivity (if any) in solution or in the solid phase for small optically pure peptides.

Not many examples of intramolecular cyclization of dehydroamino acids have been reported. This is a little surprising because many examples of the amenability of various derivatives of side chain functionalized amino acids by radical methods have been demonstrated by several groups.6 Intermolecular radical additions to dehydroamino acid derivatives were reported by D. Crich and Davis.7

 

RESULT AND DISCUSSION

Intramolecular cyclization to cyclopentylamino acid dipeptide was examined. As a radical precursor, iodinated dehydroamino acid dipeptide 10Z was chosen (Scheme 1).

Phenyl thiocarbamate 12Z was considered, too. However, this precursor has been reported to have a limitation. Tin hydride meditated free radical generation via xanthate esters and related thiocarbamates has been developed by Barton and McCombie, and it’s often the method of choice for deoxygenation of secondary alcohols in synthesis.8 Relatively few examples of use of this radical precursor for intramolecular cyclization were reported, and most of them were with secondary radicals.9

The synthesis of the dehydroamino acid dipeptide is shown in Schemes 1. Compound 310 was easily converted into aldehyde 4 with NBS and 2,6-lutidine in acetone/water (9:1) solvent. Because 4 was not stable enough to be separated with silica gel, the removal of excess 2,6-lutidine was carried out under vacuum for 5 hours, and the remaining crude oil was used for next step without purification. The Erlenmeyer azlactone synthesis gave 5Z/E in good yield (84%) and the Z:E ratio was 5:1. To the Z and E mixture of azlactones was added S-leucine methyl ester (6) to yield dipeptie 7Z/E. The TBS protecting group of 7Z/E was taken off with tetrabutyl ammonium fluoride. Direct conversion of hydroxyl group in 8Z/E to bromide with trioctylphosphine and carbon tetrabromide was not satisfactory. The reaction proceeded too slowly with low yield. However, mesylation of 8Z/E followed by substitution reaction with NaI in DMF produced the target radical precursor 10Z/E in 87% yield.

For the final cyclization, the traditional method, tributyltin hydride and AIBN (azoisobutyronitrile) with heating was tried. The removal of oxygen by bubbling dry argon through the solvent for 30 minutes before adding Bu3SnH was necessary. After heating at 70 ℃ for 12 hours under the Ar atmosphere, high yield conversions into 11 were performed starting from pure Z isomer of 10. However, no diastereoselectivity was found.

Schemes 1.

Recently, several new techniques have been developed for asymmetric radical reactions. The most common method is lowering the temperature with a different radical initiator instead of AlBN and heating. One of them is triethylborane that has proved to be an effective radical initiator in the presence of trace amounts of oxygen.11 Several examples of use of the Et3B have been reported in the reduction of alkyl, alkenyl, and aryl halides.12 This methodology is attractive for the following reasons: 1) It is mild enough not to perturb other functional groups (carbonyl, ether, hydroxyl). 2) The reaction temperature can be lowered to -78 ℃. Especially, K. Oshima reported interesting reactions of α-halo ketones and aldehydes with Ph3SnH and Et3B at room temperature.13 The important two roles of Et3B in this reaction were both initiation of radical reaction to generate α-carbonyl radical and trapping this radical as boron enolate. He also extended this to the three component coupling reaction of alkyl iodide, α, β-unsaturated ketone, and aldehyde.

Et3B methodology was attractive for our purpose of enhancing diastereoselectivity by lowering the reaction temperature. Additionally, if boron enolate was forming as an intermediate, an intramolecular chelation of boron enolate could exist, and it might have a positive effect to improve the diastereoselectivity.

Pure Z isomer of 10 was dissolved in toluene and O2 was removed by bubbling dry argon through the reaction mixture. Et3B (1.3 equiv.) in THF and Bu3SnH (1.3 equiv.) was added, and the reaction was stirred at -78 ℃, -50 ℃, -20 ℃, and 0 ℃. At -78 ℃ and -50 ℃, no product formation was found, and starting material was recovered. It might be because of the low solubility of the reagents at low temperature. Actually, a white precipitate in the reaction flask was found during the reaction. At -20 ℃, reaction was finished in 5 hours, and a 2:1 mixture of two diastereomers (11) was obtained in 74% yield. The amounts of added Et3B was essential. With less than 1 equivalent, reaction was not complete, but with more than 1.5 equivalents, side products were formed. This result suggests that Et3B is not just a radical initiator, but also participates in forming an intermediate, which seems to be a boron enolate.

Diastereoselectivity is induced by the asymmetric H-abstraction of boron enloate which might depend marginally on the stereochemistry of adjacent amino acid side chain (R). The exact mechanism is still unclear and studies to figure out the absolute stereochemistry of major diastereomer of 11 are going on. At 0 ℃, reaction was completed within 30 minutes, and slightly lower diastereoselectivity (3:2 ratio of two diastereomers) with similar yield was obtained.

Our dipeptide was designed using adjacent amino acid groups as the chiral auxiliary. However, for this purpose, O=C-N and C-C=O bond rotations in the peptide bond should be considered carefully. For the best stereoselectivity, the orientation of these two bonds must be fixed. In the field of stereoselective alkylation of enolates and enamines, several techniques have been developed to restrict these two bond rotations. Especially, for O=C-N bond fixing, the C2 symmetry strategy14, chelation15, H-bonding, dipole-dipole control16,and steric control techniques have been reported. In the radical chemistry field, some chiral amide auxiliaries - dimethylpyrrolidine4a-c, Oppolzer camphorsultam2a, and oxazolidines4d - have been applied successfully for this purpose. In our dipeptide, although the orientation of C-C=O bond rotation might be fixed by forming boron enolate, the steric hindrance from chiral auxiliary in adjacent amino acid was proved to have little effect. However, If we do same radical cyclization of dehydroamino acid included in a longer peptide chain, we carefully expect better diastereoselectivity, because the tertiary structure of the long peptide chain might be rigid enough not to allow the free rotation of individual peptide bonds, and the adjacent amino acids will be used as effective chiral auxiliaries.

For better diastereocontrol, intramolecular cyclization of a dehydroamino acid 21Z (Scheme 2) which has a chiral carbon next to double bond was tried. In this case, two stereocenters are generated by asymmetric intramolecular cycloaddition and subsequent H-abstraction. Different from the previous dipeptide, the stereocontrol on cycloaddition was designed to be performed by a chiral group attached next to the double bond, and the newly formed chiral cyclopentyl moiety controlled the direction of H-abstraction of α-carbon.

Schemes 2.

The radical precursor 21Z was easily synthesized from 20Z.17 Mesyl compound 20Z was converted into 21Z by refluxing with NaI in DMF for 8 hours. Fot the cyclization, 21Z was dissolved in dry toluene, and O2 was removed before adding Bu3SnH and AIBN. After heating at 70 ℃ for 8 hours, cyclized product 22 was separated by column chromatography with silica gel. The yield was 90%, but 22 was inseparable mixtures of four diastereomers in 9:3:3:1 ratio. Deprotection of three PMB (p-methoxybenzyl) groups of 22 with TMSI (tetramethylsilyl iodide) gave 23 which was separated as mixtures of two diastereomers (3:1). Because of the failure of purification, there were difficulties in assigning the configurations of two new forming stereocenters in the major isomer. However, the configuration of the β-carbon in the major isomer could be barely figured out by an NOE experiment. Its result is in Fig. 2. Transannular NOE between H-1 and H-3, and H-4 is evidence supporting H-1, H-3 and H-4 have a syn configuration. Therefore, the configuration of the β-carbon on the major diastereomer of 23 could be assigned as R.

Low diastereoselectivity can be enhanced when the reaction is carried out at lower temperature with irradiation or using the Et3B initiator, and we hope to report its result near future.

Fig. 2.

 

EXPERIMENT

4-[5-(t-Butyldimethylsilyloxy)-pentylidene]-2-phenyl-5(4H)-oxazolone 5Z/E. To 12 g (0.037 mol) of 5-t-butyldimethylsilyloxy-1,1-dithioethylpentanal (3) in 200 mL of acetone/water mixture (10:1) was added 17.3 mL (0.149 mol, 4 equiv.) of 2,6-lutidine, and 13.2 g (0.074 mol, 2 equiv.) of N-bromosuccinimide (NBS) was titrated into the reaction mixture with stirring at rt (room temperature) until the yellow color remained. After stirring for 15 min, the reaction mixture was transferred into a 1M sodium sulfite aqueous solution. Extractions were carried out with hexane/CH2Cl2 (4:1) three times. The organic extracts were combined, dried over anhydrous MgSO4, filtered, and concentrated at reduced pressure. The remaining clear oil was dried by toluene azeotropic distillation, and excess 2,6-lutidine was also removed by evaporation. Without further purification, crude aldehyde was dissolved in 100 mL dry toluene and transferred into a mixture of 10 g (0.0555 mol, 1.5 equiv.) of hippuric acid, 13.7 mL (0.233 mol, 4 equiv.) of acetic anhydride, and 5 g (0.0112 mol, 1.2 equiv.) of Pb(OAc)4 in 200 mL dry toluene under nitrogen atmosphere. The reaction mixture was heated for 12 hr at 80 ℃, and the color became yellow. Cooling the reaction flask in ice bath, 50 mL of 0.1 M NaHSO4 aqueous solution was added slowly, and an extraction was carried out with three portions of toluene. The organic layers were combined, dried over anhydrous MgSO4, filtered, and concentrated at reduced pressure. The remaining yellow oil was purified by column chromatography with florisil to yield 11.3 g (84%) of 5Z and 5E mixture (Z:E=5:1). For analytical purpose, 5Z was isolated.

5Z: IR (neat, cm−1) 2928, 2856, 1810, 1654, 1602, 1580, 1522, 1472, 1324, 1293, 1179, 1091, 836, 777, 710; 1H NMR (360 MHz, CDCl3) δ 8.07 (d, J =7.17 Hz, 2H, ArH), 7.56 (m, 1H, ArH), 7.42 (m, 2H, ArH), 6.69 (dd, J=7.98, 7.98 Hz, 1H, vinyl-H), 3.65 (dd, J=5.97, 5.97 Hz, 2H, CH2OTBS), 2.69 (ddd, J=7.63, 7.63, 7.63 Hz, 2H, allylic-CH2), 1.53 (m, 4H, CH2), 0.90 (s, 9H, (CH3)3C), 0.05 (s, 6H, Si(CH3)2); 13C NMR (90 MHz, CDCl3) δ 166.0, 162.7, 139.6, 136.3, 133.1, 128.8, 128.1, 125.7, 62.6, 32.3, 28.5, 25.9, 24.9, 18.3, -5.4.

N-(1S-methoxycarbonyl-3-methylbutyl)-7-(t-butyldimethylsilyloxy)-2-phenacylamino-2-heptenamide 7Z/E. To 10 g of L-leucine methyl ester hydrochloride (6) in 300 mL of dry CH2Cl2 was added NH3 gas by bubbling into the reaction mixture for 2 hr. The reaction was stirred for 6 hr at rt. A white precipitate of NH4Cl was removed by filtering, and the CH2Cl2 was removed at reduced pressure. The remaining clear oil was distilled in vacuum (25 mmHg) to yield leucine methylester as a clear oil. 1.5 g (4.18 mmol) of 5Z/E (Z:E=5:1) and 1.1 g (8.36 mmol, 2 equiv.) of leucine methyl ester were dissolved in 10 mL dry THF. After refluxing 12 hr, the solvent was removed at reduced pressure, and the remaining yellow oil was purified by column chromatography with silica gel to yield 1.89 g (90%) of a 7Z/E mixture (5:1). For analytical purpose, a pure sample of 7Z was obtained.

7Z: IR (neat, cm−1) 3258, 3045, 2954, 2857, 1751, 1636, 1558, 1522, 1486, 1472, 1435, 1361, 1255, 1198, 1157, 1101, 836, 775; 1H NMR (360 MHz, CDCl3) δ 8.08 (s, 1H, NH), 7.86 (d, J=7.35 Hz, 2H, ArH), 7.49 (m, 1H, ArH), 7.40 (m, 2H, ArH), 6.83 (d, J=8.06 Hz, 1H, NH), 6.39 (dd, J=7.15, 7.15 Hz, 1H, vinyl-H), 4.62 (m, 1H, NCHCO), 3.64 (s, 3H, OCH3), 3.54 (dd, J=5.68, 5.68 Hz, 2H, CH2OTBS), 2.15 (m, 2H, allylic-CH2), 1.62 (m, 4H, CH2), 1.48 (m, 3H, CH2, CH), 0.88 (m, 6H, CH3), 0.83 (s, 9H, (CH3)3C), 0.01 (s, 6H, (CH3)2Si); 13C NMR (90 MHz, CDCl3) δ 173.7, 173.6, 165.1, 133.7, 133.4, 131.9, 128.7, 128.5, 127.5, 62.6, 52.1, 51.0, 41.6, 36.1, 32.4, 28.1, 25.8, 24.7, 22.7, 21.8, 18.2, -5.4.

N-(1S-methoxycarbonyl-3-methylbutyl)-7-hydroxy-2-phenacylamino-2-hepenamide 8Z/E. To 1.50 g (2.97 mmol) of 7Z/E (Z:E=5:1) dissolved in 100 mL of dry THF, was added 3.56 mL (3.56 mmol, 1.2 equiv.) of tetrabutylammonium fluoride (1M solution in THF). After stirring 2 hr, the reaction mixture was transferred into 50 mL saturated aqueous NH4Cl solution. Extraction was carried out with three portions of ethylacetate, and the combined organic layers were dried over anhydrous MgSO4, filtered, and concentrated at reduced pressure. The remaining oil was purified by column chromatography over silica gel to yield 1.07 g (92%) of 8Z/E (Z:E=5:1).

8Z: IR (neat, cm−1) 3288, 2954, 2869, 1736, 1648, 1522, 1482, 1438, 1369, 1276, 1207, 1158, 1058, 709; 1H NMR (360 MHz, CDCl3) δ 8.20 (s, 1H, NH), 7.88 (m, 2H, ArH), 7.53 (m, 1H, ArH), 7.45 (m, 2H, ArH), 6.77 (d, J=8.21 Hz, 1H, NH), 6.41 (dd, J=7.36, 7.36 Hz, 1H, vinyl-H), 4.65 (m, 1H, NCHCO), 3.65 (s, 3H, OCH3), 3.61 (m, 2H, CH2OH), 2.22 (m, 2H, allylic-H2), 1.62 (m, 7H, CH2, CH), 0.92 (m, 6H, CH3); 13C NMR (90 MHz, CDCl3) δ 173.7, 165.0, 133.5, 133.2, 132.1, 128.9, 128.6, 128.4, 127.5, 62.4, 52.3, 51.1, 41.5, 31.7, 27.9, 24.8, 24.7, 22.8, 21.9.

N-(1S-methoxycarbonyl-3-methylbutyl)-7-(methanesulfonyloxy)-2-phenacylamino-2-heptenamide 9Z/E. To 1.03 g (2.64 mmol) of 8Z/E (Z:E=5:1) dissolved in 50 mL of dry CH2Cl2 were added 0.63 mL (7.92 mmol, 3 equiv.) of pyridine and 0.61 L (7.92 mmol, 3 equiv.) of methanesulfonyl chloride, and the reaction was stirred for 12 hr at rt. The reaction mixture was transferred into 20 mL of a saturated aqueous NH4Cl solution and extracted with CH2Cl2 three times. The combined organic layers were dried over anhydrous MgSO4 and filtered. The solvent was removed at reduced pressure, and the remaining oil was separated by column chromatography with silica gel to yield 1.10 g (89 %) of 9Z/E (Z:E=5:1). For analytical purpose, a pure sample of 9Z was obtained.

9Z: IR (neat, cm−1) 3270, 2956, 2870, 1744, 1647, 1522, 1479, 1438, 1351, 1277, 1200, 1173, 975, 937, 832, 710; 1H NMR (360 MHz, CDCl3) 8.25 (s, 1H, NH), 7.88 (m, 2H, ArH), 7.50 (m, 1H, ArH), 7.41 (m, 2H, ArH), 7.00 (d, J=8.05 Hz, 1H, NH), 6.40 (dd, J=7.20, 7.20 Hz, 1H, vinyl-H), 4.60 (m, 1H, NCHCO), 4.17 (dd, J=6.21, 6.21 Hz, 2H, CH2OMs), 3.63 (s, 3H, OCH3), 2.95 (s, 3H, OCH3), 2.20 (ddd, J=7.35, 7.35, 7.35 Hz, 2H, allylic-CH2), 1.70 (m, 4H, CH2), 1.58 (m, 3H, CH2, CH), 0.91 (m, 6H, CH3); 13C NMR (90 MHz, CDCl3) δ 173.7, 173.6, 164.9, 133.3, 132.8, 132.0, 129.2, 128.5, 127.5, 69.6, 52.2, 51.1, 41.1, 37.2, 28.6, 27.5, 24.7, 24.0, 22.7, 21.8.

N-(1S-methoxycarbonyl-3-methylbutyl)-7-iodo-2-phenylacylamino-2-heptenamide 10Z/E. To 1.05 g (2.24 mmol) of 9Z/E (Z:E=5:1) dissolved in 50 mL dry DMF was added 0.50 g (3.36 mmol, 1.5 equiv.) of NaI, and the reaction was heated at 70 ℃ for 12 hr under nitrogen atmosphere. The DMF solvent was concentrated at reduced pressure, and the remaining yellow solid was purified by column chromatography with silica gel to yield 0.97 g (87%) of 10Z/E (Z:E=5:1). Approximately 200 mg of the 10Z/E mixture was repurified by preparative TLC and 110 mg of pure 10Z was isolated.

10Z: IR (neat, cm−1) 3055, 2956, 2867, 1749, 1634, 1556, 1524, 1458, 1434, 1367, 1296, 1201, 1159, 1028, 984, 692; 1H NMR (360 MHz, CDCl3) δ 8.04 (s, 1H, NH), 7.86 (m, 2H, ArH), 7.52 (m, 1H, ArH), 7.44 (m, 2H, ArH), 6.74 (d, J=8.07 Hz, 1H, NH), 6.36 (dd, J=7.12, 7.12 Hz, vinyl-H), 4.65 (m, 1H, NCHCO), 3.68 (s, 3H, OCH3), 3.15 (dd, J =6.91, 6.91 Hz, 2H, CH2I), 2.21 (ddd, J=7.49, 7.49, 7.49 Hz, 2H, allylic-CH2), 1.79 (m, 2H, CH2), 1.61 (m, 5H, CH2, CH), 0.93 (d, J=5.99 Hz, 3H, CH3), 0.91 (d, J=5.99 Hz, 3H, CH3); 13C NMR (90 MHz, CDCl3) δ 173.6, 164.9, 133.4, 132.1, 128.9, 128.4, 127.5, 52.3, 52.1, 41.4, 32.9, 29.1, 27.4, 24.8, 22.8, 21.9, 6.5.

(N-phenylacyl-2-cyclopentylglycyl)-(S)-leucine methyl ester 11. A) Bu3SnH/ AIBN method: To 60 mg (0.120 mmol) of 10Z dissolved in 50 mL of dry toluene was added a catalytic amount of azoisobutyronitrile (AlBN). Dry argon was bubbled through the reaction mixture for 30 min, then 41.9 μL (0.144 mmol, 1.2 equiv.) of Bu3SnH was injected via syringe and the reaction was heated at 75 ℃ under an argon atmosphere for 12 hr. The reaction mixture was transferred into a saturated aqueous NH4Cl solution and extracted with toluene three times. The combined organic extracts were dried over anhydrous MgSO4, filtered, and concentrated at reduced pressure. The remaining yellow oil was purified by column chromatography with silica gel to yield 39.5 mg (88%) of 11 (1:1 mixture of two diastereomers).

B) Et3B/Bu3SnH method: To 20 mg (0.040 mmol) of 10Z in dry toluene degassed with dry argon was added 52.0 μL (1.3 equiv.) of Et3B 1.0 M solution in THF and 15.1 μL (0.052 mmol, 1.3 equiv.) of Bu3SnH at -20 ℃. After stirring 5 hours, small water was added, and extractions were carried out with three portions of toluene. Combined organic layers was dried over anhydrous MgSO4, filtered, and concentrated. The resultant oil was purified by column chromatography with silica gel to yield 11.0 mg (74%) of 11 (2:1 mixture of two diastereomers).

11 (a mixture of two diastereomers): IR (neat, cm−1) 3068, 2955, 2870, 1750, 1633, 1579, 1539, 1490, 1436, 1386, 1328, 1259, 1203, 1171, 1152, 1027, 692; 1H NMR (360 MHz, CDCl3) δ 7.79 (m, 4H, ArH), 7.45 (m, 6H, ArH), 6.86 (dd, J=6.00, 6.00 Hz, 2H, NH), 6.76 (d, J=8.20 Hz, 1H, NH), 6.67 (d, J=7.93 Hz, 1H, NH), 4.59 (m, 4H, NHCHCO), 3.73 (s, 3H, OCH3), 3.66 (s, 3H, OCH3), 2.39 (m, 2H, CH), 1.79 (m, 4H, CH2), 1.60 (m, 10H, CH2), 1.39 (m, 8H, CH2, CH), 0.94 (d, J=6.05 Hz, 3H, CH3), 0.92 (d, J=5.85 Hz, 3H, CH3), 0.89 (d, J=6.16 Hz, 3H, CH3), 0.87 (d, J=5.96 Hz, 3H, CH3); 13C NMR (90 MHz, CDCl3) δ 173.0, 172.9, 171.7, 167.5, 167.4, 134.0, 133.9, 131.7, 128.5, 128.4, 127.1, 127.0, 77.4, 77.0, 76.6, 57.1, 52.2, 50.8, 42.7, 42.5, 41.2, 41.0, 29.5, 29.4, 28.9, 28.8, 26.8, 26.6, 25.3, 25.1, 25.0, 24.9, 24.7, 22.8, 21.7, 21.7; HRMS (FAB) [MH]+ calculated for C21H30O4N2 375.2284, found 375.2285.

Methyl (Z)-7-iodo-4R,5S,6S-tris(p-methoxybenzyloxy)-2-trifluoroacetyl-2-heptenoate 21Z. To 195.5 mg (0.25 mmol) of 20Z in DMF (15 mL) was added 74.9 mg (0.50 mmol, 2 equiv.) of NaI. The reaction mixture was stirred at 70 ℃ for 12 hr, and the solvent was removed at reduced pressure. The remaining oil and excess NaI were separated by column chromatography with silica gel to yield 140 mg (71%) of 21Z: IR (neat, cm−1) 3309, 3001, 2954, 2837, 1735, 1655, 1612, 1586, 1515, 1465, 1437, 1303, 1250, 1210, 1174, 1034, 910, 822, 760; 1H NMR (360 MHz, CDCl3) δ 8.68 (s, 1H, NH), 7.20 (m, 4H, ArH), 7.10 (m, 1H, ArH), 6.82 (m, 6H, ArH), 6.43 (d, J=6.72 Hz, 1H, vinyl-H), 4.71 (d, J=10.48 Hz, 1H, OCH2Ar), 4.50 (d, J=10.48 Hz, 2H, OCH2Ar), 4.43 (d, J=11.17 Hz, 1H, OCH2Ar), 4.37 (dd, J=6.80, 2.53 Hz, 1H, CHOPMB), 4.23 (d, J=10.72 Hz, 1H, OCH2Ar), 4.19 (d, J=11.24 Hz, 1H, OCH2Ar), 3.83 (s, 3H, ArOCH3), 3.80 (s, 3H, ArOCH3), 3.79 (s, 3H, ArOCH3), 3.76 (s, 3H, COCH3), 3.70 (dd, J=6.58, 2.54 Hz, 1H, CHOPMB), 3.56 (dd, J=4.13, 4.13 Hz, 2H, CH2I), 3.50 (m, 1H, CHOPMB); 13C NMR (90 MHz, CDCl3) δ 162.9, 159.7, 159.5, 159.4, 155.0 (ddd, J=37.8, 37.8, 37.8 Hz, CF3CO), 130.7, 130.2, 129.8, 129.7,129.2, 128.8, 128.7, 128.2, 113.9, 113.8, 81.5, 76.1, 75.2, 74.4, 71.3, 71.2, 55.2, 55.1, 52.8, 8.8; HRMS (FAB) [MNa]+ calculated for C34H37F3INO9 810.1364, found 810.1367.

Methyl N-trifluoroacetyl-2-[2R,3R,4R-tris(p-methoxybenzyloxy)-cyclopentyl]-glycinate 22. To 100 mg (0.13 mmol) of 21Z dissolved in dry toluene was added a catalytic amount (5 mg) of AlBN. Dry argon was bubbled through the reaction mixture for 30 min, then 45.5 μL (0.17 mmol, 1.3 equiv.) of Bu3SnH was added via syringe and the reaction mixture was heated to 75 ℃ for 12 hr. Toluene was removed at reduced pressure, and the remaining oil was purified by column chromatography with silica gel to yield 77 mg (90%) of 23 as an inseparable mixture of four diastereomers (9:3:3:1).

Major diastereomer. (in a mixture of diastereomers): IR (neat, cm−1), 3268, 3074, 3000, 2952, 2909, 2837, 1724, 1612, 1586, 1550, 1515, 1464, 1442, 1361, 1302, 1250, 1210, 1174, 1111, 1034, 821, 728; 1H NMR (360 MHz, CDCl3) δ 8.12 (d, J =5.56 Hz, 1H, NH), 7.25 (m, 4H, ArH), 7.14 (m, 2H, ArH), 6.88 (m, 6H, ArH), 4.39 (dd, 1H, NCHCO), 3.95 (m, 2H, CHOPMB), 3.86 (m, 1H, CHOPMB), 3.85 (s, 3H, ArOCH3), 3.82 (s, 3H, ArOCH3), 3.81 (s, 3H, ArOCH3), 3.64 (s, 3H, COOCH3), 2.58 (m, 1H, CH), 2.13 (m, 1H, CH2), 1.70 (m, 1H, CH2); 13C NMR (90 MHz, CDCl3) δ 169.9, 159.4, 159.3, 159.2, 157.8 (ddd, J=37.5, 37.5, 37.5 Hz, CF3CO), 129.7, 129.6, 129.5, 129.4, 129.3, 115.6 (ddd, J=285.9, 285.9, 285.9 Hz, CF3), 113.8, 113.7, 113.6, 82.6, 81.9, 75.8, 72.0, 71.9, 71.2, 55.2, 54.8, 52.4, 41.7, 29.3; HMRS (FAB) [MH]+ calculated for C34H38F3NO9 660.2420, found 660.2426.

Methyl N-trifluoroacetyl-2-(2R,3R,4R-trihydroxycyclopentyl)-glycinate 23. To 55 mg (0.083 mmol) of 22 (mixture of four diastereomers) in 5 mL dry CH2Cl2 at 0 ℃ was added 63.2 μL (0.44 mmol, 5 equiv.) of TMSI portionwise for 2 hr under nitrogen atmosphere. Stirring was continued until all starting material disappeared at 0 ℃, and the aqueous CH3OH was added to quench excess TMSI. After stirring 1 hr, the solvent was removed at reduced pressure. The remaining oil was transferred into a saturated aqueous NH4Cl solution, and extraction was carried out with three portions of EtOAc. The combined organic extracts were dried over anhydrous MgSO4, filtered, and concentrated at reduced pressure. The remaining dark brown oil was purified by column chromatography with silica gel to yield 21.5 mg (86%) of an inseparable mixture (9:3:3:1) of diastereomers 23. Small amount of 23 were repurified with preparative TLC, but still a mixture of two diastereomers (3:1) was obtained.

Major diastereomer. (in a mixture of two diastereomers): IR (neat, cm−1) 3342, 3084, 2927, 1716, 1565, 1440, 1220, 1182, 1069, 1037, 975, 878, 729; 1H NMR (360 MHz, CD3COCD3) δ 9.08 (brd, 1H, NH), 4.28 (dd, J=6.00, 6.00 Hz, 1H, NCHCO), 4.11 (m, 1H, CHOH), 4.00 (dd, J=4.70, 10.19 Hz, 1H, CHOH), 3.77 (dd, J=4.50, 9.43 Hz, 1H, CHOH), 3.69 (s, 3H, OCH3), 2.45 (m,1H, CH), 2.25 (m, 1H, CH2), 1.56 (m, 1H, CH2); 13C NMR (90 MHz, CD3CN) δ 170.9, 158.2 (ddd, J=36.8, 36.8, 36.8 Hz, CF3CO), 115.0 (ddd, J=285.1, 285.1, 285.1 Hz, CF3), 79.9, 77.3, 72.0, 56.3, 52.8, 43.7, 32.1; HRMS (FAB) [MNH4]+ calculated for C10H14F3NO6 319.1117, found 319.1115.

 

CONCLUSION

Various intramolecular radical cyclizations of dehydroamino acid derivatives were studied. The 5-exo radical cyclization of dehydroamino acid derivative in which adjacent amino acid was used as a chiral auxiliary was performed with high yield but with poor stereoselectivity. In low temperature reaction, enhanced stereo selectivity was found, but it was still unsatisfactory. However, a 5-exo radical cyclization of a dehydroamino acid in which chiral carbons were located around the double bond gave promising results. Two new stereocenters were generatedat the same time, and even at high temperature reaction condition, affordable stereoselectivity was found. It is expected that the changing reaction condition at low temperature with Et3B initiator can improved the stereoselectivity. We believe intramolecular radical cyclization of dehydroamino acids must be a useful methodology for the synthesis of amino acids that have a cyclic side chain.