Catalytic Asymmetric Addition of Diorganozinc Reagents to N

Catalytic asymmetric addition of diorganozinc SPECIAL FEATURE reagents to N-phosphinoylalkylimines

Alexandre Coˆ te´ , Alessandro A. Boezio, and Andre´ B. Charette*

Department of Chemistry, University of Montreal, P.O. Box 6128, Station Downtown, Montreal, QC, Canada H3C 3J7

Edited by Jack Halpern, University of Chicago, Chicago, IL, and approved February 4, 2004 (received for review October 31, 2003)

The synthesis of ␣-chiral amines bearing two alkyl groups has been hampered by the accessibility and stability of the alkylimine precursor. Herein, we report an efficient strategy to generate the alkyl-substituted imine in situ that is compatible with the Me- DuPHOS monoxide⅐Cu(I) catalyzed addition of diorganozinc reagents. The sulfinic acid adduct of the imine is readily prepared by mixing diphenylphosphinic amide, the aldehyde, and sulfinic acid. The sulfinic acid adduct is generally isolated by filtration. The addition of diorganozinc reagents in the presence of Me-DuPHOS monoxide⅐Cu(I) and the in situ-generated imines affords the cor- Fig. 1. Bioactive ␣-chiral amines. responding ␣-chiral amines in high yields and enantiomeric excesses. sensitive imines from stable precursors has been a strategy that he synthesis of ␣-chiral amines using the catalytic asymmetric has been quite successful in a number of cases. Typically, a stable Taddition of diorganozinc reagents has produced very exciting imine adduct is used as a precursor and is converted to the imine results in recent years (1–3). This very important subunit is in situ (Scheme 2). The method involves the use of a leaving

␣ CHEMISTRY commonly found in many pharmaceuticals and other biologically group (LG) on the -carbon of the N-protected amine. Several important compounds. Some specific examples include Flomax leaving groups including benzotriazolates (29), sulfinates (30), (antihypertensive) and Mexitil (antiarrhythmics) (Fig. 1) (4). and succinimidates (31) have been used for the in situ prepara- The most important methodologies developed to prepare tion of N-acylimines. ␣-chiral amines in enantiomerically pure form rely on either In this article, we describe the feasibility of the in situ synthesis chemical resolution or on the use of readily available chiral of N-phosphinoylimines and we show that these precursors are also suitable reaction partners for the preparation of ␣-chiral synthons as building blocks. More recently, the development of ⅐ efficient chiral auxiliaries has led to the extensive use of chiral amines in high yields and ee in the (R,R)-BozPHOS Cu catalyzed imines as precursors to ␣-chiral amines by alkylation chemistry addition of diorganozinc reagents (Scheme 3). (5–9) or hydrogenation (10). Experimental Methods Catalytic asymmetric nucleophilic addition reactions of orga- nometallic reagents to imines have been reported with N- General. All nonaqueous reactions were run under an inert tosylimines (11, 12), N-arylimines (13–15), and N-acylimines (16, atmosphere (nitrogen or argon) with rigid exclusion of moisture 17) as amine precursors. The cleavage of the protecting group from reagents and glassware by using standard techniques for leads to the ␣-chiral amine; however, in the former two cases, manipulating air-sensitive compounds. All glassware was stored in the oven and͞or was flame-dried before use under an inert harsh conditions are usually necessary [N-tosyl, SmI2; N-aryl, ϩ atmosphere of gas. Anhydrous solvents were obtained either by ceric ammonium nitrate or PhI(OAc)2; N-formyl, H3O , heat]. Furthermore, the reactions usually require excess of the dior- filtration through drying columns (ether, toluene) on a Glass- ͞ ganozinc reagent, and in a number of cases, the addition of Contour system (Irvine, CA) or by distillation over sodium dimethylzinc is unsuccessful or affords to the amine in much benzophenone (toluene). Analytical TLC was performed on lower yields and enantiomeric excesses (ee). Recently, we re- precoated, glass-backed silica gel (Merck 60 F254). Visualization ported that the copper-catalyzed addition of diorganozinc re- of the developed chromatogram was performed by UV absor- agents to N-phosphinoylarylimines proceeds in high yields and bance, aqueous cerium molybdate, ethanolic phosphomolybdic enantioselectivities in the presence of a catalytic amount of the acid, iodine, or aqueous potassium permanganate. Flash column novel Me-DuPHOS monoxide [(R,R)-BozPHOS] chiral ligand chromatography was performed by using 230–400 mesh silica (Scheme 1) (18, 19). In addition to being readily available, this [EM Science or Silicycle (Quebec City, QC, Canada)] of the hemilabile bidentate ligand offers superb catalytic activity, indicated solvent system according to standard technique. Melt- broad substrate generality, and mild reaction conditions. Fur- ing points were obtained on a Buchi (Flawil, Switzerland) thermore, the increase in imine electrophilicity imparted by the melting-point apparatus and are uncorrected. NMR spectra (1H, N-phosphinoyl protecting group, combined with its ease of 13C, 31P) were recorded on Bruker AV 300, AMX 300, AV 400, cleavage under mildly acidic conditions, makes this method very or ARX 400 spectrometers. Chemical shifts for 1H NMR spectra attractive to prepare ␣-chiral amines. are recorded in ppm with the solvent resonance as the internal The one major limitation of this methodology is our inability standard (chloroform, ␦ 7.27 ppm, or DMSO, ␦ 2.50 ppm). Data to prepare alkylaldehyde-derived N-phosphinoylimines bearing are reported as follows: chemical shift, multiplicity (s, singlet; ␣-enolizable protons in reasonable yields. This has been a d, doublet; t, triplet; q, quartet; qn, quintet; sept, septuplet; limitation not only in diorganozinc addition chemistry (20–25), but also in catalytic asymmetric nitro-Mannich (26), Mannich (27), and Strecker (28) processes that use similar electrophilic This paper was submitted directly (Track II) to the PNAS ofﬁce. precursors. This inherent limitation is also present in the prep- Abbreviation: ee, enantiomeric excess. aration of N-acylimines, and several strategies have been devel- *To whom correspondence should be addressed. E-mail: [email protected]. oped to circumvent this problem. The in situ generation of © 2004 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0307096101 PNAS ͉ April 13, 2004 ͉ vol. 101 ͉ no. 15 ͉ 5405–5410 Downloaded by guest on September 29, 2021 Scheme 3.

1 h at room temperature, diethylzinc (205 ␮l, 2.0 mmol, 2.5 eq) was added, and the resulting dark-brown suspension was stirred for 20 min and cooled to the desired temperature. A cold suspension of sulfinyl adduct (0.80 mmol, 1.0 eq) in toluene (3 ml ϩ 2 ml to rinse ϩ 1 ml to rinse) was transferred (Teflon cannula; o.d., 1͞8th of an inch; i.d., 1͞16th of an inch) to the catalyst solution. After stirring, the reaction was quenched with saturated aqueous NH4Cl and the aqueous layer was extracted ϫ with CH2Cl2 (3 40 ml). The combined organic layers were dried over Na2SO4. Concentration and purification by silica gel column chromatography (100% EtOAc) gave the pure protected amine. Scheme 1. p-Toluenesulfinic Acid. p-Toluenesulfinic acid was formed by dis- solving its hydrated sodium salt in a minimum of hot 10% vol͞vol m, multiplet; br, broad), coupling constant in Hz, and integra- HCl (the resulting pH must be lower than 3) and crystallization tion. Chemical shifts for 13C NMR spectra are recorded in ppm at 4°C. Further filtration and drying under vacuum led to white from tetramethylsilane by using the central peak of deutero- crystals. chloroform (77.00 ppm) or deuterated DMSO (39.52 ppm) as the internal standard. All spectra were obtained with complete General Procedure for the Synthesis of Sulfinyl Adducts. To a sus- proton decoupling. Optical rotations were determined with a pension of P,P-diphenylphosphinic amide (1.00 g, 4.6 mmol, 1 Perkin–Elmer 341 polarimeter at 589 or 546 nm. Data are eq) and sulfinic acid (1.08 g, 6.9 mmol, 1.5 eq) in anhydrous ␣ temp ͞ reported as follows: [ ]␭ , concentration (c in g 100 ml), and diethylether (40 ml, ACS grade) was added the freshly distilled solvent. High-resolution mass spectra were performed by the aldehyde (6.9 mmol, 1.5 eq) at room temperature. The mixture Centre re´gional de spectroscopie de masse of the University of was stirred for 15 h, during which a white precipitate was slowly Montreal. Combustion analyses were performed by the Elemen- formed. On completion of the reaction, the solution was filtered, tal Analysis laboratory of the University of Montreal. and the white solid was washed with anhydrous diethylether (15 Analytical HPLC was performed on a Hewlett–Packard an- ml, ACS grade) and dried under vacuum. alytical instrument (model 1100) equipped with a diode array UV detector. Data for determination of the enantiomeric excess N-[(4-methylphenyl)sulfonyl](phenyl)methyl]-P,P-diphenylphosphinic is reported as follows: column type, eluent, flow rate, and amide (3). The general procedure was followed. The crude retention time (tr). compound (71% yield) was used without purification for the next step: (white powder) mp 122–124°C (dec.) HRMS (FAB) m͞z Reagents. ϩ Cu(OTf)2 was purchased from Strem Chemicals (New- calc. for C H NO PS [M ϩH] : 462.1293 found: 462.1305. buryport, MA). All starting materials were purchased from 26 25 3 Aldrich or Alfa Aesar (Ward Hill, MA). Unless otherwise stated, P,P-diphenyl-N-[(1S)-1-phenylpropyl]phosphinic amide. The product commercial reagents were used without purification. Diethyl- 2 was obtained following the general procedure (specific con- zinc was purchased neat from Akzo Nobel and used without ditions: 0°C for 24 h). Yield: 87%. The ee (97%) was determined purification. by HPLC analysis [Chiralpak AD, 80:20 hexanes:i-PrOH, 1.0 Racemic samples for HPLC analysis were prepared by ml͞min: (R)-2 t ϭ 9.6 min, (S)-2 t ϭ 13.0 min]. The S addition of Et Zn͞CuCN in toluene to the sulfinyl adducts. r r 2 configuration was assigned by comparison of the optical rotation Ligand (R,R)-BozPHOS was prepared according to literature procedure (18). of the deprotected amine hydrochloride with the literature (32). N-[1-methyl-(toluene-4-sulfonyl)-methyl]-P,P-diphenylphosphinic amide. General Procedure for Diethylzinc Addition on Sulfinyl Adducts. The general procedure was followed. The crude compound 9 (R,R)-BozPHOS (13 mg, 0.040 mmol, 0.050 eq) and Cu(OTf)2 (91% yield) was used without purification for the next step: (13 mg, 0.036 mmol, 0.045 eq) were dissolved in toluene (3 ml). 1 The resulting dark-green heterogeneous solution was stirred for (white powder) mp 120–121°C (dec.); H NMR (DMSO-d6, 400 MHz) ␦ 7.68–7.63 (m, 4H), 7.54–7.47 (m, 4H), 7.37–7.29 (m, 6H), 6.37 (t, J ϭ 10.6 Hz, 1H), 4.46 (m, 1H), 2.41 (s, 3H), 1.48 ϭ 13 ␦ (d, J 6.8 Hz, 3H); C NMR (DMSO-d6, 75 MHz) 145.2, 135.0 ϭ ϭ (d, JC-P 68.4 Hz), 134.1, 133.3 (d, JC-P 72.7 Hz), 132.6 (d, JC-P ϭ ϭ ϭ 2.7 Hz), 132.3 (d, JC-P 3.0 Hz), 132.2 (d, JC-P 10.2 Hz), ϭ ϭ 132.0 (d, JC-P 10.2 Hz), 130.4, 130.2, 129.2 (d, JC-P 12.5 Hz), ϭ ϭ 31 129.0 (d, JC-P 12.6 Hz), 69.0, 21.9, 16.6 (d, JC-P 1.9 Hz); P ␦ ͞ NMR (DMSO-d6, 162 MHz) 26.0; HRMS (APCI) m z [M- ϩ SO2Tol] calc.: 244.1 found: 244.2; Anal. calc. for C21H22NO3PS: Scheme 2. C 63.14, H 5.55, N 3.51 found: C 63.10, H 5.70, N 3.60.

5406 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0307096101 Coˆ te´ et al. Downloaded by guest on September 29, 2021 1 ␦ ϭ N-[(1R)-1-methylpropyl]-P,P-diphenylphosphinic amide. The product CHCl3); H NMR (300 MHz, CDCl3) 7.93 (dd, J 8.0, 1.6 Hz, 16 ϭ was obtained by following the general procedure (specific 2H), 7.89 (dd, J 8.1, 1.6 Hz, 2H), 7.53–7.38 (m, 6H), 3.04 (m, SPECIAL FEATURE conditions: Ϫ60°C for 48 h). Yield 97%, ee (90%) was deter- 1H), 2.64 (br s, 1H), 1.75 (sept, J ϭ 6.6 Hz, 1H), 1.66–147 (m, mined by HPLC analysis [Chiralpak AD-H, 80:20 hexanes:i- 2H), 1.47–1.23 (m, 2H), 0.90 (t, J ϭ 7.4 Hz, 3H), 0.78 (t, J ϭ 6.9 ͞ ϭ ϭ 13 ␦ ϭ PrOH, 1.0 ml min: (S)-16 tr 7.8 min, (R)-16 tr 8.7 min]. mp Hz, 6H); C NMR (75 MHz, CHCl3) 134.3 (d, JC-P 20.0 Hz), ␣ 20 Ϫ ϭ ϭ 128–130°C; Rf 0.25 (100% AcOEt); [ ]D 18.7° (c 1.27, CHCl3); 133.7 (d, JC-P 9.3 Hz), 132.6 (d, JC-P 9.3 Hz), 132.1 (d, JC-P 1 ␦ ϭ ϭ ϭ H NMR (300 MHz, CDCl3) 7.94–7.83 (m, 4H), 7.48–7.36 (m, 1.4 Hz), 128.9 (d, JC-P 1.1 Hz), 128.7 (d, JC-P 1.2 Hz), 51.2 ϭ ϭ ϭ 6H), 3.11 (m, 1H), 2.73 (br s, 1H), 1.59 (m, 1H), 1.46 (m, 1H), (d, JC-P 2.0 Hz), 46.5 (d, JC-P 5.9 Hz), 30.2 (d, JC-P 4.0 Hz), ϭ ϭ 13 31 ␦ 1.19 (d, J 6.4 Hz, 3H), 0.88 (t, J 7.4 Hz, 3H); C NMR (75 25.1, 23.2, 23.0, 9.7; P NMR (121 MHz, CDCl3) 22.5; LRMS ␦ ϭ ϭ ͞ ϩ ϩ MHz, CHCl3) 133.9 (d, JC-P 8.6 Hz), 132.1 (d, JC-P 8.6 Hz), (APCI) m z [M H] calc.: 316.4 found: 316.1; Anal. calc. for ϭ ϭ 132.0 (d, JC-P 4.6 Hz), 131.9 (d, JC-P 4.6 Hz), 131.5, 128.4, C19H26NOP: C 72.36, H 8.31, N 4.44 found: C 72.10, H 8.54, ϭ ϭ 128.2, 48.8 (d, JC-P 1.7 Hz), 32.2 (d, JC-P 6.1 Hz), 23.1 (d, JC-P N 4.46. ϭ 31 ␦ 4.4 Hz), 10.2; P NMR (162 MHz, CDCl3) 23.1; LRMS (APCI) m͞z [M ϩ H]ϩ calc.: 274.2 found: 274.1; Anal. calc. for N-{cyclopentyl[(4-methylphenyl)sulfonyl]methyl}-P,P-diphenylphos- C16H20NOP: C 70.31, H 7.38, N 5.12 found: C 70.32, H 7.44, phinic amide. The general procedure was followed. The crude N 5.15. compound 4 (95% yield) was used without purification for the next step: (white powder) mp 111–113°C (dec.); 1H NMR ϭ N-{2-methyl-1-[(4-methylphenyl)sulfonyl]propyl}-P,P-diphenylphosphinic (DMSO-d6, 300 MHz) 7.77–7.65 (dd, J 11.3, 7.7 Hz, 2H), 7.56 amide. The general procedure was followed. The crude com- (d, J ϭ 7.7 Hz, 2H), 7.50–7.40 (m, 4H), 7.38–7.29 (m, 4H), 7.18 pound 8 (88% yield) was used without purification for the next (d, J ϭ 7.7 Hz, 2H), 6.31 (t, J ϭ 13.1 Hz, 1H), 4.66 (dt, J ϭ 11.5, 1 step: (white powder) mp 116–117°C (dec.); H NMR (DMSO-d6, 5.4 Hz, 1H), 2.31 (s, 3H), 1.72–1.55 (m, 2H), 1.50–1.15 (m, 7H); ϭ 13 ␦ ϭ 400 MHz) 7.81–7.69 (m, 2H), 7.58 (d, J 8.3 Hz, 2H), 7.54–7.43 C NMR (DMSO-d6, 75 MHz) 144.8, 136.1 (d, JC-P 59.9 Hz), ϭ ϭ ϭ ϭ (m, 4H), 7.38–7.25 (m, 4H), 7.22 (d, J 8.3 Hz, 2H), 6.25 (t, J 135.4, 134.4 (d, JC-P 62.0 Hz), 132.7 (d, JC-P 2.5 Hz), 132.3 ϭ ϭ ϭ ϭ 13.0 Hz, 1H), 4.56 (dt, J 11.8, 2.9 Hz, 1H), 2.49 (m, 1H), 2.35 (d, JC-P 2.4 Hz), 131.8 (d, JC-P 11.3 Hz), 131.6 (d, JC-P 10.4 ϭ ϭ ϭ ϭ (s, 3H), 1.03 (d, J 6.8 Hz, 3H), 0.90 (d, J 7.0 Hz, 3H); LRMS Hz), 130.2, 129.6, 129.0 (d, JC-P 12.5 Hz), 128.7 (d, JC-P 12.8 ͞ ϩ 31 (APCI) m z [M-SO2Tol] calc.: 272.1 found: 272.1; Anal. calc. Hz), 75.8, 30.4, 28.7, 25.7, 25.5, 22.0; P NMR (DMSO-d6, 162 for C H NO PS: C 64.62, H 6.13, N 3.28 found: C 64.61, H 6.32, MHz) ␦ 25.0; HRMS (FAB) m͞z calc. for C H NO PS [M ϩ

23 26 3 25 29 3 CHEMISTRY N 3.39. H]ϩ: 454.1606 found: 454.1601.

N-[(1S)-1-ethyl-2-methylpropyl]-P,P-diphenylphosphinic amide. The N-[(1S)-1-cyclopentylpropyl]-P,P-diphenylphosphinic amide. The product 15 was obtained by following the general procedure product 12 was obtained by following the general procedure (specific conditions: room temperature for 24 h). Yield 86%, ee (specific conditions: 0°C for 24 h). Yield 92%, ee (95%) was (96%) was determined by HPLC analysis [Chiralpak AD-H, determined by HPLC analysis [Chiralpak AD-H, 80:20 hex- ͞ ϭ ͞ ϭ ϭ 90:10 hexanes:i-PrOH, 1.0 ml min: (S)-15 tr 13.1 min, (R)-15 anes:i-PrOH, 1.0 ml min: (S)-12 tr 7.3 min, (R)-12 tr 9.2 ϭ ␣ 20 ϩ tr 14.7 min]. mp 135–136°C; Rf 0.43 (100% AcOEt); [ ]D 7.7° min]. The S configuration is tentatively assigned based on 1 ␦ (c 1.04, CHCl3); H NMR (300 MHz, CDCl3) 7.92 (m, 4H), 7.45 compounds 3, 5, 8, 9 and 10.mp145–146°C; Rf 0.45 (100% ␣ 20 ϩ 1 (m, 6H), 2.82 (br s, 1H), 2.70 (br s, 1H), 1.88 (m, 1H), 1.52 (m, AcOEt); [ ]D 8.8° (c 1.03, CHCl3); H NMR (300 MHz, 13 ␦ ␦ ϭ 2H), 0.89 (m, 9H); C NMR (75 MHz, CHCl3) 134.3 (d, JC-P CDCl3) 7.90 (dt, J 11.7, 5.9 Hz, 4H), 7.46 (m, 6H), 2.92 (br ϭ ϭ ϭ 10.3 Hz), 132.7 (d, JC-P 3.4 Hz), 132.6 (d, JC-P 3.3 Hz), s, 1H), 2.72 (br s, 1H), 1.96 (m, 1H), 1.82 (m, 1H), 1.58 (m, 7H), ϭ ϭ ϭ 13 132.2 (d, JC-P 9.4 Hz), 132.1 (d, JC-P 1.0 Hz), 132.1 (d, JC-P 1.27 (m, 2H), 0.91 (t, J 7.4 Hz, 3H); C NMR (75 MHz, ϭ ϭ ϭ ␦ ϭ ϭ 1.0 Hz), 128.9 (d, JC-P 2.9 Hz), 128.7 (d, JC-P 2.9 Hz), 58.6 CHCl3) 134.4 (d, JC-P 33.8 Hz), 132.7 (d, JC-P 6.2 Hz), 132.7 ϭ ϭ ϭ ϭ ϭ ϭ (d, JC-P 2.2 Hz), 32.1 (d, JC-P 8.7 Hz), 26.7 (d, JC-P 14.0 (d, JC-P 32.5 Hz), 132.6 (d, JC-P 6.2 Hz), 132.1 (d, JC-P 2.3 31 ␦ ϭ ϭ Hz), 19.0, 18.2, 10.9; P NMR (121 MHz, CDCl3) 23.4; HRMS Hz), 132.0 (d, JC-P 2.3 Hz), 128.9, 128.7, 56.9 (d, JC-P 2.3 Hz), ͞ ϭ ϭ ϭ (MAB) m z calc. for C18H24NOP [M]: 301.1587 found 301.1596. 45.0 (d, JC-P 6.3 Hz), 29.8 (d, JC-P 8.8 Hz), 28.6 (d, JC-P ϭ 31 3.2 Hz), 25.8 (d, JC-P 14.0 Hz), 9.5; P NMR (121 MHz, ␦ ͞ ϩ ϩ N-{3-methyl-1-[(4-methylphenyl)sulfonyl]butyl}-P,P-diphenylphos- CDCl3) 22.5; LRMS (APCI) m z [M H] calc.: 328.2 found: phinic amide. The general procedure was followed. The crude 328.2; Anal. calc. for C20H26NOP: C 73.37, H 8.00, N 4.28 found: compound 10 (84% yield) was used without purification for the C 73.37, H 8.14, N 3.93. next step: (white powder) mp 112–114°C (dec.); 1H NMR (400 ␦ ϭ MHz, DMSO-d6) 7.72–7.67 (m, 2H), 7.68 (d, J 8.2 Hz, 2H), N-{1-[(4-methylphenyl)sulfonyl]-3-phenylpropyl}-P,P-diphenylphos- 7.53–7.40 (m, 8H), 7.28 (d, J ϭ 8.3 Hz, 2H), 6.32 (dd, J ϭ 13.7, phinic amide. The general procedure was followed. The crude 11.0 Hz, 1H), 4.49 (ddd, J ϭ 23.0, 11.1, 3.8 Hz, 1H), 2.36 (s, 3H), compound 5 (97% yield) was used without purification for the 1.79–1.57 (m, 3H), 0.79 (d, J ϭ 6.2 Hz, 3H), 0.64 (d, J ϭ 6.0 Hz, next step: (white powder) mp 118–120°C (dec.); 1H NMR 13 ϭ ␦ ϭ 3H); C NMR (100 MHz, CHCl3) 135.6 (d, JC-P 90.9 Hz), (DMSO-d6, 300 MHz) 7.78 (m, 2H), 7.49 (m, 10H), 7.28 (d, J ϭ ϭ ϭ ϭ 134.5, 134.4 (d, JC-P 94.3 Hz), 132.8 (d, JC-P 14.6 Hz), 132.0 8.2 Hz, 2H), 7.18 (m, 3H), 7.0 (d, J 6.9 Hz, 2H), 6.45 (t, J ϭ ϭ ϭ (d, JC-P 10.4 Hz), 130.5 (d, JC-P 2.8 Hz), 130.1, 129.2 (d, JC-P 11.6 Hz, 1H), 4.39 (tdd, J 21.2, 21.2, 2.1 Hz, 1H), 2.65 (m, 1H), ϭ ϭ ϭ ␮ 13 11.8 Hz), 129.0 (d, JC-P 12.5 Hz), 72.2 (d, JC-P 10.4 Hz), 2.48 (m, 1H), 2.35 (s, 3H), 2.27 (m, 1H), 1.93 ( , 1H); C NMR 31 ␦ 38.6, 24.6, 24.3, 22.1, 21.4; P NMR (162 MHz, CDCl3) 26.0; (DMSO-d6, 75 MHz) 145.3, 141.4, 136.0, 135.1, 134.3, 133.4, ͞ ϩ ϭ ϭ LRMS (APCI) m z [M-SO2Tol] calc.: 286.1 found: 286.1; 132.6 (J 2.4 Hz), 132.4 (d, J 2.4 Hz), 132.3, 132.2, 132.1, Anal. calc. for C24H28NO3PS: C 65.29, H 6.39, N 3.17 found: C 132.0, 130.5, 130.1, 129.4, 129.3, 129.2, 129.2, 129.1, 129.0, 128.9, 31 ␦ 65.18, H 6.36, N 3.20. 127.0, 72.9, 31.9, 31.7, 22.0; P NMR (DMSO-d6, 121 MHz) ͞ ϩ 24.7. LRMS (APCI) m z [M-SO2Tol] calc.: 334.1 found: 334.1; N-[(1S)-1-ethyl-3-methylbutyl]-P,P-diphenylphosphinic amide. The Anal. calc. for C28H28NO3PS: C 68.69, H 5.76, N 2.86 found: C product 17 was obtained by following the general procedure 68.56, H 5.94, N 2.93. (specific conditions: Ϫ20°C for 16 h). Yield 97%, ee (96%) was determined by HPLC analysis [Chiralpak AD-H, 80:20 hex- N-[(1S)-1-ethyl-3-phenylpropyl]-P,P-diphenylphosphinic amide. The ͞ ϭ ϭ anes:i-PrOH, 1.0 ml min: (S)-17 tr 7.2 min, (R)-17 tr 11.0 product 11 was obtained by following the general procedure ␣ 20 ϩ Ϫ min]. mp 103–104°C; Rf 0.49 (100% AcOEt); [ ]D 16.0° (c 1.05, (specific conditions: 20°C for 16 h). Yield 98%, ee (96%) was

Coˆ te´ et al. PNAS ͉ April 13, 2004 ͉ vol. 101 ͉ no. 15 ͉ 5407 Downloaded by guest on September 29, 2021 determined by HPLC analysis [Chiralpak AD-H, 80:20 hex- Table 1. Effect of additives in the catalytic asymmetric synthesis ͞ ϭ ϭ anes:i-PrOH, 1.0 ml min: (R)-11 tr 9.2 min, (S)-11 tr 12.8 of amines ␣ 20 Ϫ min]. mp 138–139°C; Rf 0.54 (100% AcOEt); [ ]D 20.3° (c 1 ␦ ϭ 1.32, CHCl3); H NMR (300 MHz, CDCl3) 7.91 (dt, J 11.6, 5.9 Hz, 4H), 7.52–7.39 (m, 6H), 7.28–7.21 (m, 2H), 7.20–7.13 (m, 3H), 3.09 (br s, 1H), 2.85–2.56 (m, 3H), 1.85 (q, J ϭ 7.4 Hz, 2H), 1.64 (qn, J ϭ 6.9 Hz, 2H), 0.93 (t, J ϭ 7.4 Hz, 3H); 13C NMR (75 ␦ ϭ ϭ MHz, CHCl3) 142.4, 134.3 (d, JC-P 2.1 Hz), 133.6 (d, JC-P ϭ ϭ 6.0 Hz), 132.6 (d, JC-P 5.9 Hz), 132.1 (d, JC-P 2.2 Hz), 128.9 Entry Additive (LG-ZnEt) Yield*, % ee, % ϭ ϭ (d, JC-P 1.1 Hz), 128.8 (d, JC-P 0.7 Hz), 126.2, 53.0 (d, JC-P ϭ ϭ ϭ 1 None Ͼ95 98 1.5 Hz), 38.3 (d, JC-P 5.2 Hz), 32.4, 29.9 (d, JC-P 4.5 Hz), 31 ␦ ͞ 10.1; P NMR (121 MHz, CDCl3) 22.9; HRMS (MAB) m z calc. for C23H26NOP [M] calc.: 363.1752 found: 363.1748. 244–95 78–94

N-{1-[(4-methylphenyl)sulfonyl]heptyl}-P,P-diphenylphosphinic amide. The general procedure was followed. The crude compound 7 (87% yield) was used without purification for the next step: Ͼ 1 3 95 68 (white powder) mp 117–118°C (dec.); H NMR (DMSO-d6, 400 MHz) 7.71 (dd, J ϭ 11.9, 7.37 Hz, 2H), 7.59 (d, J ϭ 8.0 Hz, 2H), 7.54–7.35 (m, 8H), 7.27 (d, J ϭ 7.9 Hz, 2H), 6.32 (t, J ϭ 12.3 Hz, 1H), 4.47 (dd, J ϭ 11.0, 2.2 Hz, 1H), 2.34 (s, 3H), 1.95–1.86 (m, 1H), 1.73–1.62 (m, 1H), 1.32–1.20 (m, 1H), 1.15–0.85 (m, 7H), 4 Ͼ95 97 ϭ 13 ␦ 0.75 (t, J 7.2 Hz, 3H); C NMR (DMSO-d6, 100 MHz) 145.0, ϭ ϭ 135.6 (d, JC-P 72.2 Hz), 134.4, 134.4 (d, JC-P 75.9 Hz), 132.3 ϭ ϭ ϭ (d, JC-P 3.3 Hz), 131.9 (d, JC-P 10.0 Hz), 131.8 (d, JC-P 10.3 57195 ϭ ϭ Hz), 130.3, 130.0, 129.1 (d, JC-P 12.5 Hz), 128.9 (d, JC-P 12.7 31 Hz), 72.9, 31.7, 28.5, 25.5, 22.7, 22.0, 14.7; P NMR (DMSO-d6, ␦ ͞ ϩ 6 Ͼ95 96 162 MHz) 26.0; LRMS (APCI) m z [M-SO2Tol] calc.: 314.2 found: 314.1; Anal. calc. for C26H32NO3PS: C 66.50, H 6.87, 7† None Ͻ5N͞A N 2.98 found: C 66.43, H 6.87, N 2.92. *Yields determined by 31P NMR with an internal standard. † N-[(1S)-1-ethylheptyl]-P,P-diphenylphosphinic amide. The product Et2Zn replaced by EtZnOTf (4 eq). 14 was obtained by following the general procedure (specific conditions: Ϫ20°C for 16 h). Yield 98%, ee (95%) was deter- compounds 3, 5, 8, 9 and 10.mp151–152°C; Rf 0.43 (100% mined by HPLC analysis [Chiralpak AD-H, 90:10 hexanes:i- ␣ 20 Ϫ 1 ͞ ϭ ϭ AcOEt); [ ]D 5.7° (c 0.99, CHCl3); H NMR (400 MHz, PrOH, 1.0 ml min: (S)-14 tr 10.3 min, (R)-14 tr 14.7 min]. ␦ The S configuration is tentatively assigned based on the com- CDCl3) 7.92 (m, 4H), 7.47 (m, 6H), 2.80 (br s, 1H), 2.72 (br s, pounds 3, 5, 8, 9 and 10.mp110–111°C; R 0.51 (100% AcOEt); 1H), 1,80–140 (m, 8H), 1.29–1.07 (m, 4H), 1.03–0.92 (m, 1H), f ϭ 13 ␦ ␣ 20 Ϫ 1 ␦ 0.90 (t, J 7.4 Hz, 3H); C NMR (75 MHz, CHCl3) 134.3 (d, [ ]D 5.4° (c 1.06, CHCl3); H NMR (300 MHz, CDCl3) 7.93 ϭ ϭ ϭ ϭ ϭ JC-P 23.7 Hz), 132.7 (d, JC-P 20.5 Hz), 132.7 (d, JC-P 1.8 (d, J 7.6 Hz, 2H), 7.90 (d, J 7.1 Hz, 2H), 7.52–7.35 (m, 6H), ϭ ϭ Hz), 132.2 (d, JC-P 20.5 Hz), 132.0 (d, JC-P 2.5 Hz), 128.9 (d, 3.01 (br s, 1H), 2.68 (br s, 1H), 1.62–1.41 (m, 4H), 1.41–1.11 (m, ϭ ϭ ϭ 13 ␦ JC-P 2.1 Hz), 128.7 (d, JC-P 2.1 Hz), 58.1 (d, JC-P 2.3 Hz), 8H), 0.93–0.77 (m, 6H); C NMR (75 MHz, CHCl3) 134.4 (d, ϭ ϭ ϭ ϭ ϭ 42.2 (d, JC-P 5.1 Hz), 29.7, 28.8, 27.0 (d, JC-P 3.0 Hz), 26.8 JC-P 9.2 Hz), 133.6 (d, JC-P 16.5 Hz), 132.6 (d, JC-P 1.9 Hz), ϭ ϭ (d, J ϭ 3.7 Hz), 10.6; 31P NMR (121 MHz, CDCl ) ␦ 22.3; 132.1 (d, JC-P 1.9 Hz), 128.9, 128.7, 60.8, 53.1 (d, JC-P 2.0 Hz), C-P 3 ϭ ϭ HRMS (MAB) m͞z calc. for C H NOP [M]: 341.1909 found: 36.6 (d, JC-P 5.3 Hz), 32.2, 29.7 (d, JC-P 5.5 Hz), 25.9, 23.0, 21 28 31 ␦ 14.5, 10.0; P NMR (121 MHz, CDCl3) 22.8; LRMS (APCI) 341.1905. m͞z [M ϩ H]ϩ calc.: 344.4 found: 344.2; Anal. calc. for Results C21H30NOP: C 73.44, H 8.80, N 4.08 found: C 73.11, H 9.08, N 3.84. One important aspect of the in situ generation of N- phosphinoylimines from stable precursors is the concomitant N-{cyclohexyl[(4-methylphenyl)sulfonyl]methyl}-P,P-diphenylphosphinic formation of a stoichiometric amount of an EtZnLG species amide. The general procedure was followed. The crude com- (Scheme 3). Because the basic nature of this by-product may pound 6 (72% yield) was used without purification for the next seriously hamper the subsequent catalytic asymmetric nucleo- 1 step: (white powder) mp 113–115°C (dec.); H NMR (DMSO-d6, philic addition step, we initially elected to test a standard 300 MHz) 7.71 (dd, J ϭ 11.4, 7.8 Hz, 2H), 7.56 (d, J ϭ 8.2 Hz, catalytic asymmetric addition on a preformed and stable 2H), 7.52–7.41 (m, 4H), 7.35–7.24 (m, 4H), 7.20 (d, J ϭ 7.8 Hz, N-phosphinoylimine in the presence of several additives (Table 2H), 6.28 (t, J ϭ 12.9 Hz, 1H), 4.49 (dt, J ϭ 11.9, 2.7, 1H), 2.34 1). These reactions were then compared to the standard one (s, 3H), 2.18 (dt, J ϭ 10.5, 2.1 Hz, 1H), 1.90 (d, J ϭ 11.7 Hz, 1H), in the absence of additives to identify which additives had the 1.52 (dt, J ϭ 35.9, 10.8, 4H), 1.35–0.95 (m, 4H), 0.95–0.77 (m, minimal detrimental effect on the efficiency and level of 31 ␦ ͞ 1H); P NMR (DMSO-d6, 121 MHz) 25.0; HRMS m z (FAB) enantioselection for the reaction. The data collected in Table ϩ ϩ calc. for C26H31NO3PS [M H] : 468.1762 found: 468.1767. 1 illustrate that the benzotriazoyl group (entry 2) resulted in irreproducible yields depending on the reaction time (longer N-[(1S)-1-cyclohexylpropyl]-P,P-diphenylphosphinic amide. The times led to higher yields and ee of the desired product: 78% product 13 was obtained by following the general procedure ee after 6 h and 94% ee after 20 h). Conversely, the succin- (specific conditions: 0°C for 24 h). Yield 89%, ee (96%) was imidate group (entries 3) allowed for an excellent yield but determined by HPLC analysis [Chiralpak AD-H, 90:10 hex- eroded ee when compared to the standard reaction run in the ͞ ϭ ϭ anes:i-PrOH, 1.0 ml min: (S)-13 tr 10.4 min, (R)-13 tr 14.7 absence of the additive (entry 1). The ethylzinc p- min]. The S configuration is tentatively assigned based on the toluenesulfinyl was fully compatible with the reagents and

5408 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0307096101 Coˆ te´ et al. Downloaded by guest on September 29, 2021 Table 3. Optimization of the catalytic asymmetric addition SPECIAL FEATURE

Scheme 4. mol% Entry Cu salt, mol% (R,R)-BozPHOS Yield, % ee, %

catalytic species, and high yield and enantiocontrol was ob- 1 Cu(OTf)2 (10) 5 97 94 served in the presence of a stoichiometric amount of this 2 Cu(OTf)2 (6) 5 93 97 additive (entry 4). It is noteworthy that high enantiocontrol 3 Cu(OTf)2 (4.5) 5 98 96 was also achieved in the presence of ethylzinc phenoxide (entry 4 Cu(SO2Tol)2 (6) 5 90 82 5) or ethylzinc methoxide (entry 6). To further corroborate 5 Cu(OTf)2 (6) 3 90 98 these findings, the benzotriazoyl adduct 1 of the imine derived 6 CuOTf (10) 5 83 93 from benzaldehyde was prepared and submitted to the reaction conditions (Scheme 4). Although a low conversion to the desired amine 2 was observed after 6 h (43% conversion and between the copper salt and (R,R)-BozPHOS, as well as the 38% isolated yield), the ee of the product was excellent, optimal copper salt. The Cu͞ligand stoichiometry is important perhaps indicating that the benzotriazoyl group may be a very because it defines the amount of ethylzinc triflate present in the good choice for the in situ preparation of imines. However, the mixture. Although ethylzinc triflate is not sufficiently nucleo- data presented in Table 1 led us to focus initially on the philic to add to the imine (Table 1, entry 7), we believe its role p-toluenesulfinyl adduct of the imine because it appeared to be as a Lewis acid to activate the imine is crucial in this reaction. the most promising precursor for the in situ preparation of ␣ We also elected to use 2.5 eq of diethylzinc because 1 eq is dialkyl-substituted -chiral amines. consumed to form the imine, leaving 1.5 eq for the nucleophilic The synthesis of the sulfinic acid adduct of N-phosphinoylimi- addition process. The first three entries illustrate the effect of the CHEMISTRY nes could not be accomplished by using standard procedures that stoichiometry between (R,R)-BozPHOS and copper(II) triflate involve acidic conditions and heat. However, we found that they on the efficiency of the reaction. It appears that a 1:1 ratio of Cu could be prepared simply by stirring an aldehyde, diphenylphos- and the chiral ligand is optimal for this reaction. Excellent ee of phinic amide, and p-toluenesulfinic acid in ether (Table 2). The the product was observed when the amount of the chiral ligand mixture is stirred for several hours (16–36 h) during which the resulting sulfinic acid adduct precipitates and is isolated by was lowered to 3 mol%; however, the yield was slightly lower filtration in high yield. The reaction proceeds well with several (Table 1, entry 5). A similar reduction in yield was observed if aryl- and alkyl-substituted aldehydes. More importantly, the copper(I) triflate was used as the catalyst precursor (Table 1, reaction involving aldehydes containing enolizable protons pro- entry 6). Because a stoichiometric amount of the sulfinyl coun- ceeded smoothly and without any significant self-condensation terion is formed in the reaction, we were intrigued by the of the aldehyde. Typically, the solid that is recovered on filtration possibility of using copper p-toluenesulfinyl as the catalyst is sufficiently pure to be used directly in the catalytic asymmetric precursor. Unfortunately, much lower ee resulted with this alkylation reaction with diorganozinc reagents. One limitation of copper salt (Table 1, entry 4). this approach is that ␣,␤-unsaturated aldehydes are not com- With the optimal conditions in hand, the copper catalyzed patible with this procedure as the sulfinic moiety is prone to addition reactions of diethylzinc to sulfinic acid adducts 3-10 undergo conjugate addition reaction with this substrate. were tested (Table 4). We were quite pleased to find that the The next step was to test several reaction conditions and other additions proceeded smoothly in all of the cases and that, more parameters to identify any peculiar aspect of copper catalyzed importantly, excellent enantiocontrol was observed. Further- addition of diorganozinc reagents to the sulfinic acid adduct 5 more, the reaction was shown to be highly effective with (Table 3). Based on our previous work on this reaction, the first ␣-branched imines (Table 4, entries 2, 4, and 6). In all cases, the important variable to be tested was the optimal stoichiometry

Table 4. Catalytic asymmetric synthesis of amines Table 2. Synthesis of the sulﬁnic adduct of N-phosphinoylimines

Entry R1 Yield, % ee, %, (er) Entry R1 Yield, % 1Ph87(2) 97 (98.5:1.5)

1Ph71(3) 2 c-C5H11 92 (12) 95 (97.5:2.5) 2 c-C5H11 95 (4) 3 PhCH2CH2 98 (11) 96 (98:2) 3 PhCH2CH2 97 (5) 4 c-C6H13 89 (13) 96 (98:2) 4 c-C6H13 72 (6) 5C6H13 98 (14) 95 (97.5:2.5) 5C6H13 87 (7) 6 i-Pr 86 (15) 96 (98:2) 6 i-Pr 88 (8) 7Me97(16) 90 (95:5) 7Me 91(9) 8 i-Bu 97 (17) 96 (98:2) 8 i-Bu 84 (10) er, Enantiomeric ratio.

Coˆ te´ et al. PNAS ͉ April 13, 2004 ͉ vol. 101 ͉ no. 15 ͉ 5409 Downloaded by guest on September 29, 2021 N-phosphinoyl protecting group could be cleaved under mild advantages of this process are high yields and enantioselectivi- conditions to liberate the amine (HCl, MeOH). ties, as well as the mild conditions for the deprotection of the N-protecting group. Conclusion In conclusion, we have shown that Me-DuPHOS monoxide This work was supported by the Natural Sciences and Engineering Research Council (NSERC), Merck Frosst Canada, Boehringer In- [(R,R)-BozPHOS] is a very effective ligand in the copper- gelheim (Canada), and the University of Montreal. A.A.B. is grateful to catalyzed addition of diethylzinc to N-phosphinoylalkylimines NSERC (PGF B) and Fonds pour la Formation de Chercheurs et l’Aide prepared in situ from their p-toluenesulfinyl adducts. The major `a la Recherche (B2) for postgraduate fellowships.

1. Enders, D. & Reinhold, U. (1997) Tetrahedron Asymmetry 8, 1895–1946. 17. Dahmen, S. & Bra¨se, S. (2002) J. Am. Chem. Soc. 124, 5940–5941. 2. Bloch, R. (1998) Chem. Rev. 98, 1407–1438. 18. Boezio, A. A. & Charette, A. B. (2003) J. Am. Chem. Soc. 125, 1692–1693. 3. Kobayashi, S. & Ishitani, H. (1999) Chem. Rev. 99, 1069–1094. 19. Boezio, A. A., Pytkowicz, J., Coˆte´, A. & Charette, A. B. (2003) J. Am. Chem. 4. Kleemann, A., Engel, J., Kutscher, B. & Reichert, D. (2001) Pharmaceutical Soc. 125, 14260–14261. Substances: Syntheses, Patents, Applications (Thieme, Stuttgart). 20. Wipf, P., Kendall, C. & Stephenson, C. R. J. (2003) J. Am. Chem. Soc. 125, 5. Ellman, J. A., Owens, T. D. & Tang, T. P. (2002) Acc. Chem. Res. 35, 984–995. 761–768. 6. Liu, G. C., Cogan, D. A. & Ellman, J. A. (1997) J. Am. Chem. Soc. 119, 21. Shi, M. & Wang, C.-J. (2003) Adv. Synth. Cat. 345, 971–973. 9913–9914. 22. Jimeno, C., Reddy, K. S., Sola, L., Moyano, A., Pericas, M. A. & Riera, A. 7. Enders, D., Schubert, H. & Nu¨bling, C. (1986) Angew. Chem. Int. Ed. Engl. 25, (2000) Org. Lett. 2, 3157–3159. 1109–1110. 23. Sato, I., Kodaka, R. & Soai, K. (2001) J. Chem. Soc. Perkin Trans. 1, 2912–2914. 8. Denmark, S. E., Weber, T. & Piotroswski, D. W. (1987) J. Am. Chem. Soc. 109, 24. Pinho, P. & Andersson, P. G. (2001) Tetrahedron 57, 1615–1618. 2224–2225. 25. Zhang, H. L., Zhang, X. M., Gong, L. Z., Mi, A. Q., Cui, X., Jiang, Y. Z., Choi, 9. Sibi, M. P. & Asano, Y. (2001) J. Am. Chem. Soc. 123, 9708–9709. M. C. K. & Chan, A. S. C. (2002) Org. Lett. 4, 1399–1402. 10. Overberger, C. G., Marullo, N. P. & Hiskey, R. C. (1961) J. Am. Chem. Soc. 26. Yamada, K., Harwood, S. J., Groger, H. & Shibasaki, M. (1999) Angew. Chem. 83, 1374–1378. Int. Ed. Engl. 38, 3504–3506. 11. Fujihara, H., Nagai, K. & Tomioka, K. (2000) J. Am. Chem. Soc. 122, 12055–12056. 27. Matsunaga, S., Kumagai, N., Harada, S. & Shibasaki, M. (2003) J. Am. Chem. 12. Hayashi, T. & Ishigedani, M. (2000) J. Am. Chem. Soc. 122, 976–977. Soc. 125, 4712–4713. 13. Porter, J. R., Traverse, J. F., Hoveyda, A. H. & Snapper, M. L. (2001) J. Am. 28. Masumoto, S., Usuda, H., Suzuki, M., Kanai, M. & Shibasaki, M. (2003) J. Am. Chem. Soc. 123, 984–985. Chem. Soc. 125, 5634–5635. 14. Porter, J. R., Traverse, J. F., Hoveyda, A. H. & Snapper, M. L. (2001) J. Am. 29. Katritzky, A. R. & Harris, P. A. (1992) Tetrahedron Asymmetry 3, 437–442. Chem. Soc. 123, 10409–10410. 30. Mecozzi, T. & Petrini, M. (1999) J. Org. Chem. 64, 8970–8972. 15. Wei, C. & Li, C.-J. (2002) J. Am. Chem. Soc. 124, 5638–5639. 31. Kohn, H., Sawhney, K. A., Robertson, D. W. & Leander, J. D. (1994) J. Pharm. 16. Hermanns, N., Dahmen, S., Bolm, C. & Bra¨se, S. (2002) Angew. Chem. Int. Ed. Sci. 83, 689–691. Engl. 41, 3692–3694. 32. Wu, M.-J. & Pridgen, L. N. (1991) J. Org. Chem. 56, 1340–1344.

5410 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0307096101 Coˆ te´ et al. Downloaded by guest on September 29, 2021