Edgars Sūna Chiral 1,2,3,4-Tetrahydroisoquinolines :埃德加sūna手性1,2,3,4 -四氢异喹啉类化合物
Chapter B
Synthesis of Chiral Proton Donors via Catalytic
Asymmetric Transfer Hydrogenation
1. Introduction to the catalytic enantioselective reduction of cyclic imines.
The most direct route to chiral isoquinolines is the resolution of racemates
1 Although chiral acids used for using diastereomeric salts crystallization technique.
amines resolution can be quantitatively recovered, the process is highly substrate dependent and success often relies on chemist’s fortune (see Chapter A). Both
2asymmetric synthesis of optically active isoquinolines and diastereoselective reduction
3of dihydroisoquinolines usually employs stoichiometric amount of chiral building
blocks, auxiliaries or chiral reagents. Since chiral auxiliaries and reducing agents can not be recovered, overall asymmetric synthesis process usually is expensive. Catalytic
enantioselective reduction is an important alternative to these techniques, because a large quantity of the chiral compound can be produced using a small amount of a chiral catalyst.
In contrary to catalytic asymmetric carbonyl group reduction, corresponding
3reaction for imines is much more less developed. The subsequent literature review
covers the most important examples of catalytic enantioselective reduction of cyclic imines. Particular attention will be paid to 1-aryl substituted cyclic imines as well as 3,4-dihydroisoquinolines because the main purpose is to find the most suitable and efficient method for asymmetric synthesis of chiral tetrahydroisoquinolines with an aniline subunit.
1.1. Rhodium and iridium catalyzed asymmetric hydrogenation and
hydrosilylation.
Rhodium (I) chloride modified with various chiral bidentate phosphorous ligands was the first transition metal catalyst applied for asymmetric C=N bond hydrogenation. In contrast to high enantioselectivities observed in hydrogenation of various acyclic substrates (up to 95% ee for acyclic imines and 97% ee for
45hydrazones), only very limited success has been achieved in the case of cyclic imines.
While reduction of dihydroisoquinoline B1 with in situ prepared Rh(I)-DIOP catalyst
B4 afforded amine “optically pure or very nearly so after recrystallization of the
5a5bhydrochloride salt”, hydrogenation of imine B2 was completely non-selective.
Isoquinoline B3 in the presence of modified Rh(I) catalyst B5 also yielded racemic
5c product.
27
MeOMeOFigure B1.NNNMeOOH
B2OMeClPB3RhB1PSolv
OMeOPPhSolv = ROH 2B4: P P = DIOP:PPhOO2P2 POPPh2B5: P P = CYCPHOS:OMe2PPh2 (R,R)-MOD-DIOP
The replacement of Rh (I) with the corresponding Ir (I) catalyst completely
changed the reduction course. Hydrogenation of imine B2 with Ir(I) analogue of
5b catalyst B4 afforded optically enriched amine with 66% ee.Enantiocontrol was even higher employing modified ligand - MOD-DIOP (see Figure B1): Table B1. Comparison of catalytic asymmetric imine B2 hydrogenation in the presence of Rh(I) and Ir(I) catalysts.
Entry Metal Ligand Conversion ee
(%) (%)
1 Rh(I) DIOP 95 0
2 Ir(I) DIOP 100 66
3 Rh(I) MOD-DIOP 60 0
4 Ir(I) MOD-DIOP 100 81
Somewhat more promising method for reduction of cyclic imines is rhodium (I)
66acatalyzed asymmetric hydrosilylation procedure. In 1975 Kagan obtained several enantiomerically enriched tetrahydroisoquinolines B3, B6-B7, while Brunner and
6bWiegrebe reported hydrosilylation of various 2-phenyl-3,4-dihydropyrrole derivatives
B8-B11:
28
Scheme B1.
CFXN3N*1. PhSiH / catalyst B422HNOX2. (CFCO)O32R
RRB3, B6-B7
B8-B11
Pyrrolines were separated from the unreacted starting imines B8-B11 via
distillation of in situ prepared N-trifluoroacetamides. Reductions were run in toluene, however the best ee is achieved in the absence of solvent (entry 4, Table B2): Table B2. Asymmetric hydrosilylation of various cyclic imines employing PhSiH and 22
2 mol% of in situ generated [RhCl]-DIOP B4.
Entry Imine R X Conversiee
on (%) (%)
1 CHPh H 78 B6 23 2
2 CH OCH 93 B3 6 33
3 CHPh OCH 98 B7 39 23
4 H - 84 B8 64
5 2-OCH - 85 B9 31 3
6 3,4,5-(OCH) - 81 B10 31 33
7 4-Br - 82 B11 60
Drop in optical induction for MeO substituted substrates (entries 5-6 vs. entries 4 and 7, Table B2) was attributed to intermolecular coordination of MeO-groups to Rh. Rh(I)-Phephos catalyzed hydrosilylation procedure was also applied for the reduction of cyclic structure B12 yielding enantiomer of the antidepressant “Pyrazidole” with
6c73% enantioselectivity:
HCHC33
Rh(I)-(S)-Phephos*NNN
SiH22NNHPh273% eeB12PPh69% yield(S)-Phephos
In general, Rh(I)-catalyzed asymmetric reductions afford cyclic amines with moderate enantioselectivities, lower than observed using other transition metal
29
catalysts (see also Table B1). Evidently, this is the reason why further development of asymmetric catalytic reduction methods has been based on transition metals other that Rh, such as Ir, Ru and Ti.
789 Ir(I) and cationic Ir(I) catalysts have been widely used Chiral neutral Ir(III),
in asymmetric hydrogenation of various imines. Ir(III) catalyst exists as stable and
7easy-handled dimer that was successfully introduced and explored by Osborn. In the
reaction mixture dimeric species equilibrate with monomers which was proposed to be the active catalyst:
XHScheme B2.XPP DimericIrIr** Ir(III) catalystXPPHX
18 e complex
P P = chiral ligands
*
X = halogenHActive Ir(III)P catalystIr2*XPNX
16 e complexB2
HNHH
PNIr*Ir(III) / imine XPH2X complexHN18 e complexPIr*XPX
In contrary, neutral as well as cationic iridium (I) catalysts usually are prepared in situ from commercially available chloro(1,5-cyclooctadiene)iridium (I) or chloro-(norbornadienyl)iridium (I) dimers and an appropriate chiral diphosphine ligand:
Scheme B3.H
PORIrCatalytic cycle*2OR*P(see Scheme B2)P PH
ROHB15Clneutral Ir(I) catalystIrIr
Cl*P P+-+-[Ir(COD)Cl]Cl2ClOP4PCHClAnion 2222Ir*orIr* or-PexchangeP THFPF 6 B14B13Cationic Ir(I) catalyst
30
Cationic Ir(I) complex B13 forms if catalyst preparation is carried out in
Cl or THF. Chlorine replacement by non-nucleophilic coordinating anion such as CH22
--ClO or PF affords cationic Ir(I) catalysts B14, which are stable enough to be 46
isolated and purified by recrystallization. In the presence of alcohol, however,
is proposed. Catalysts B15 usually are prepared in formation of neutral complex B15
situ within minutes before hydrogenation is carried out and utilized without isolation.
Iridium catalysts of different oxidation states (Ir(I) and Ir(III)) as well as neutral and cationic Ir(I) species have been used for reduction of structurally distinct cyclic imines, so it is difficult to compare the reactivity and selectivity of catalysts. Fortunately, there are several common substrates for all catalysts - imines B2 and B16
and rough selectivity comparison can be made.
Figure B2.CH3++-ClO-4PF6ONCHHC3HC3N3PhPhPPhPPN2IrIrNPhPh
B2B16B17B18
CyP2PhP2HPPhOPPh2PPh22PAr2NPPhPPhO2PAr2H2PPhOO2(R,R)-DIOP
(S,S)-BDPP(R,R)-BICP(2S,4S)-BCPM: (R)-BINAP: Ar = Ph(SKEWPHOS)
(R)-Tol-BINAP: Ar = p-CHCH364
31
Table B3. Asymmetric hydrogenation of cyclic imines B2 and B16 with different
iridium catalysts (see Figure B2 for ligands structure):
a Ligand Imine Pressure Yield Lit. Entry Catalystee
(atm) (%) ref. (%)
1 Ir (III) BDPP 40 99 7a B2 80
2 Ir (III) DIOP 28 99 7a B2 51
b3 Ir (I) DIOP 100 99 5b B2 66
b4 Ir (I) BCPM 100 32 8a B2 28
c5 Ir (I) BCPM 100 99 8a B2 66
d6 Ir (I) BICP 68 99 8b B2 78
7 80 99 9a Cationic Ir (I) B17 B2 8
d8 Ir (I) BICP 68 99 8b B16 65
9 Ir (I) Tol-BINAP 60 99 8c B16 23
10 Ir (I) Tol-BINAP 60 46 8c B16 89
11 100 - 9b Cationic Ir (I) B18 B16 64
(a) 0.2 Mol% Ir(III), 1mol% Ir(I) and 0.2 mol% catalyst B17. (b) Hydrogenation was +-carried in the presence of 2mol% BuNI as additive. (c) Additive: 2 mol% BiI. (d) 43-Hydrogenation in the presence of 4mol% phthalimide as additive.
Although DIOP-modified Ir (I) catalyst is more selective than the corresponding Ir (III) species (entry 3 vs. 2), simple DIOP ligand replacement by BDPP (entry 1) makes Ir (III) catalyst enantioselectivity similar to the best Ir (I) example (entry 6; see also Table B1, entry 4). Cationic catalyst B17 is significantly less
selective than the neutral one (entry 7 vs. 6), however deactivation of the catalyst B17
during hydrogenation is reported. In contrary, cationic Ir (I) catalyst B18 is
comparable to Ir(I)-BICP (entries 8 and 11). Thus, Table B3 shows that there is no decrease in selectivity comparing in situ prepared catalysts with isolated neutral Ir (III) and cationic ones. Obviously, iridium oxidation state and type of catalyst complex is a minor issue compared to chiral ligand structure that has to be optimized for each particular substrate. Additives, solvents and hydrogenation temperature are additional
+-variables of great importance. For example, replacement of BuNI for BiI as additive 43
results in more than twofold enantioselectivity improvement (from 28% to 66% ee,
00entries 4-5), while hydrogenation temperature lowering by 50 C (to -30) resulted in
32
further increase in enantioselectivity to 91% ee. Finally, hydrogenation selectivity is sensitive also to the solvent used. Thus, hydrogenation of cyclic imine B16 in methanol
yields corresponding amine with 23% ee, while in benzene 89% ee was observed (entry 9 and 10).
Because hydrogenation success depends on many variables that have to be carefully adjusted for each particular substrate, further literature analysis will be
10 focused mainly on asymmetric transfer hydrogenation of various isoquinolines.
Table B4. Effects of additives, solvents and reduction temperature on Ir(I)-BCPM
acatalyzed asymmetric hydrogenation of 3,4-dihydroisoquinolines.
MeOMeOMeOMeO
NNNNMeOMeOMeOMeO( )MeOnOPhCH2B3
MeOMeOB22
OMeB19: n=1B21
B20: n=2
Entry T Conv. ee Imine Additive Solvent Lit. 0(%) (%) (C) ref.
20 90 18 1 none PhH-MeOH 10a B3
- 92 12 2 BiI PhH-MeOH 10a B3 3
1,8-naphthalimide - 66 3 3 PhH-MeOH 10a B3
- 96 44 4 phthalimide PhH-MeOH 10a B3
20 95 41 5 phthalimide THF 10a B3
- 94 70 6 phthalimide CHCl 10a B3 22
20 94 79 7 phthalimide toluene 10a B3
2-5 95 85-93 8 phthalimide toluene 10a,b B3
5 84 88 9 F-phthalimide toluene 10b B19 4
2 75 87 10 phthalimide toluene 10b B20
b5 50 31 11 phthalimide PhH-MeOH 10b B21
ctoluene-MeOH 2-5 85 86 12 F-phthalimide 10c B22 4
(a) Under 100 atm hydrogen pressure and with 1 mol% of the catalyst. (b) 0.5 Mol% Ir(I)-BINAP catalyst was used. (c) In the presence of 2 mol% Ir(I)-BINAP.
33
+- and BuNI is added as a co-If no additive is present or an iodide such as BiI34
catalyst, enantioselectivities are lower than 20% ee (entries 1-2). Six-membered imides are ineffective, while five-membered imides improve ee’s (entry 3 vs. 4). Clear solvents effect is observed: in less polar solvents higher selectivities are obtained (entries 4-7). The lowering of hydrogenation temperature further improves ee’s (entry 8 vs. 7).
Optimized conditions (entry 8) were applied for hydrogenation of various isoquinolines B19-B22 affording chiral alkaloids or their precursors in reasonable optical purity (86-88% ee).
10a,bEnantiocontrol achieved with Ir-catalysts is relatively high (up to 93 % ee).
Moreover, Ir-catalyst system tolerates presence of various functional groups such as
7bnitro-group, ketones, esters and nitriles. On the other hand, disadvantages of Ir-
catalyzed procedure (relatively high hydrogen pressure (28-100 atm), sensitivity to temperature and solvents used) combined with need for empirical adjustment of ligand structure and additives for each particular substrate make it less attractive compared to alternative Ti and Ru catalyst systems.
1.2. Hydrogenation and hydrosilylation catalyzed by chiral titanocene catalysts.
Buchwald achieved excellent enantioselectivities and chemical yields in
1112asymmetric hydrogenation and hydrosilylation of various cyclic and acyclic imines
employing a chiral ansa-titanocene catalyst. Based on extensive mechanistic studies, ansa-titanocene (III) hydride B23 was proposed to be the active hydrogenation
catalyst. Because catalyst B23 is highly reactive and air-sensitive it is generated in situ
from air-stable chiral precatalysts B24-B26. Treatment of chiral ansa-titanocene
dichloride B24 and 1,1’-binaphth-2,2’-diolate B25 with n-BuLi and phenylsilane under
hydrogen atmosphere generates green colored solution of active catalyst B23. In the
presence of 5 mol% catalyst reduction of various cyclic imines B28-B30 proceeds with
excellent enantiocontrol (95-99% ee) and high yields.
34
Scheme B4.
TiFFTiXX1. 10 mol% n-BuLiPhSiH32. 15 mol% PhSiH3
B26 B24: X = Cl
O B25:XX=OTiH
RL
HNHRLB23RRSNN
RRSN
H2RLRLTiTiHNHRNNRNRSRS
B271,2-insertion
( )nMeONN
NMeO
NCH3B8: n = 1
B28: n = 2B3B29: n = 3
B30
The first step of the proposed catalytic cycle is reaction of titanium (III)
11chydride B23 with an imine via 1,2-insertion reaction to form titanium amide B27.
Reduction enantioselectivity is controlled at the stage of amide B27 formation and depends on sterical interaction between catalyst ligands and imine substituents,
11cparticularly at imine nitrogen (R) (Scheme B4). According to the proposed model N
syn-imine should give (S)-isomer of amine while anti-imine should afford opposite product enantiomer (R)-isomer. Experimental observations supported stereochemistry
predicted by models (Table B5, entries 1-2, anti-imine vs. entries 5-6, syn-imine). The second step in catalytic cycle is the hydrogenolysis of amide B27 via ,-bond metathesis process to form amine enantiomer and regenerate the titanium hydride B23.
35
11a-c (entries 1-7) and Table B5. Catalytic asymmetric hydrogenation
12ahydrosilylation (entries 8-9) of cyclic imines using chiral titanocene catalyst.
0Entry Imine Catalyst Pressure T ( C) Yield ee (%)
mol% (atm) (%) (config)
1 5 34 21 86 99 (R) B8
2 5 5.4 65 83 99 (R) B8
3 5 34 65 78 98 B28
4 5 34 45 71 98 B29
5 5 136 65 82 98 (S) B3
6 5 5.4 65 79 95 (S) B3
7 5 5.4 65 72 99 B31
8 0.1 - 35 96 98 B8
9 2 - r.t. 64 98 B30
Ansa-titanocene catalyzed hydrogenation was applied also for kinetic
13resolution of various 2,3, 2,4 and 2,5-disubstituted pyrrolines. The best result was
obtained in reduction of rac-2,5-diphenylpyrroline B32. The reaction was allowed to
proceed to 50% conversion and enantioselectivity measured for reduction product - cis-2,5-pyrrolidine B33 was 99% ee (34% yield). Unreacted enantiomer of starting material B32 (99% optically pure) was recovered in 37% yield.
B25 / 5.5 atm H 2
PhPh1. 2 eq n-BuLiPhPh+PhPhNNN2. 2.5 eq PhSiHH3
(S)-B32cis-B33rac-B32
99% ee99% eePhPh
PhPhNN
B34B35
Modest result, however, was achieved for 2,3-diphenylpyrroline B34 (75% ee
for unreacted starting material), while 2,4-diphenylpyrroline B35 showed relatively
poor selectivity (49% ee for unreacted B35).
Although various cyclic imines are reduced with excellent enantiocontrol, relatively high hydrogen pressure (5.4-136 atm) and highly demanding (air and
36
moisture-free) reduction conditions combined with necessity of the catalyst preactivation are important drawbacks of Buchwald’s hydrogenation procedure.
Even more important disadvantage is poor catalyst compatibility with various
11b While N-benzyl pyrrolylimine B31 is reduced with 99% ee and in functional groups.
72% chemical yield (Table B5, entry 7), free pyrrole B36 is reported to destruct
catalyst, but N-lithio and N-TMS derivatives (B37 and B38, resp.) failed to react at all.
Pyridyl substituted imine B39 also failed to react, but reduction of 2-furyl-2-pyrroline B40 could not be forced to completion even under harsh conditions, possibly due to catalyst inhibition by binding of the amine to the metal in a bidentate fashion:
Figure B3.
RNNNONN
B40B39CHCH33
B31: R = CHPhCH23PhNNB36: R = H
B37: R = LiXClB38: R = SiMe3B41: X = BrB43B42: X = CF3
It was also found that an imine containing an aromatic bromide B41 deactivates
the catalyst and ca. 5% debrominated product was detected. Besides, imine B42
containing trifluoromethyl group was found to destroy catalyst. On the other hand, titanium catalyst tolerates presence of oxygenated functional groups, such as acetals, silyl ethers and alcohols (alcohols are silylated in situ).
More convenient precatalyst B26 activation as well as simpler experimental
12aprocedure was achieved using hydrosilylation procedure. Ansa-difluoro titanocene
B26 was in situ converted to the active catalyst B23 by reaction with PhSiH (Si-F 3
bond formation is proposed to be driving force for this reaction). Hydrosilylation proceeded at room temperature in argon atmosphere with lower catalyst loading (0.1-2 mol%) and cyclic chiral imines B8 and B30 were reduced with 98-99% ee and in
substantially higher chemical yields (96-97%; silylamines were never isolated due to their lability) than in hydrogenation experiments (77-86%). Chiral substituted pyrrolidines were obtained with the same absolute configuration as in the case of titanium-catalyzed hydrogenation. Hydrosilylation is proposed to proceed by a
37
catalytic cycle similar to that for hydrogenation (Scheme B4). It was also found that reaction tolerates presence of an aromatic chloride B43.
was Recently, polymethylhydrosilane (PMHS) in the presence of i-BuNH2
employed as more convenient and inexpensive hydride source for the reduction of
12bacyclic imines.
Finally, hydrosilylation procedure was successfully employed in asymmetric
14 total synthesis of piperidine alkaloids (S)-Coniine and (2R,6R)-trans-Solenopsin A.
Conclusion. Titanocene catalyzed asymmetric hydrogenation and
hydrosilylation affords cyclic amines with excellent enantioselectivities (97-99% ee) and in high yields. However, due to low titanium catalyst compatibility with various functional groups at current level of development it can not be employed for synthesis of chiral anilino-isoquinolines.
1.3. Ruthenium(II) catalyzed asymmetric reduction of cyclic imines.
Although the first highly selective ruthenium catalyzed imine B44 asymmetric
15hydrogenation was reported by Oppolzer in 1990, (see Scheme B5), this remained
the only attempt to employ Ru(II) catalysts for reduction of cyclic imines until 1996 when Noyori applied Ru(II)-catalyzed asymmetric transfer hydrogenation protocol to
16reduction of various cyclic imines B3, B19-B21 and B45-B47:
Scheme B5.Cl[(R)-(+)-BINAP]EtNRu2423
NHNOppolzer:4 atm HSS2OOOOCHCl/ EtOH22 B4499% ee after recrystalliz.
72% chem. yield
RSOArn2Noyori:NRuNHClMeO2MeO
NNNHN*MeOMeOHHCOOH - NEt 5:2R3RR
B46: R = CHB3: R = CH33B47: R = PhB19: R = 3,4-(MeO)CHCH2632B20: R = 3,4-(MeO)CH(CH)26322: R = 3,4-(MeO)CHB21263B45: R = Ph
38
Due to excellent optical yields, operational simplicity and functional groups compatibility and selectivity Noyori procedure is regarded as breakthrough in
development of methods for catalytic asymmetric isoquinolines reduction. The catalytic method is particularly useful for transfer hydrogenation of cyclic imines with ee values ranging from 90% to 97% ee:
Table B6. Asymmetric Transfer Hydrogenation of Imines by Chiral Ru(II) complexes B48-B50:
RSOArn62B48: ,(arene) = p-cymene; Ar = 4-CHCH 364N6B49: ,(arene) = p-cymene; Ar = 2,3,4-(CH)CH 3364Ru6NHB50: ,(arene) = benzene; Ar = 1-naphthylCl2
Entry Imine Catalyst S/C Time Yield ee
(h) (%) (%)
1 200 3 99 B3 B48 95
2 1000 12 97 B3 B48 94
3 200 7 90 B19 B49 95
4 200 12 99 B20 B49 92
5 100 12 99 B21 B50 84
6 200 8 99 B45 B50 84
7 200 5 86 B46 B48 97
8 1000 12 89 B46 B48 93
9 200 5 83 B47 B48 96
Reduction of 1-methyl-3,4-dihydroisoquinoline B3 proceeds with
enantioselectivity comparable to that obtained using ansa-titanocene catalyst. The
experiment simplicity, however, makes this method more attractive than Buchwald’s
procedure (compare Table B6, entries 1-2 and Table B5, entry 4). Substituted 1-aryl (entries 5-6, Table B6) and 1-arylalkyl-3,4-dihydroisoquinolines (entries 3-4) are
10breduced in higher optical and chemical yields compared also to Ir(I)-catalyst (Table
B4, entries 9-11). Asymmetric reduction is successfully extended to the synthesis of optically active indoles from the corresponding imines B46-B47 (entries 7-9).
Transfer hydrogenation is reported to proceed smoothly in various aprotic polar solvents such as DMF, DMSO, MeCN and CHCl, using inexpensive, stable and 22
39
easy-handled formic acid-triethylamine 5:2 azeotropic mixture as a hydrogen source.
6(,-arene)] and N-sulfonylated 1,2-Generally, Ru catalyst is prepared from [RuCl222
diphenylethylenediamine, however, the same result can be obtained using catalyst formed in situ. The C=N/C=O chemoselectivity is superior to that observed in the stoichiometric reduction using NaCNBH and imine B3 can be reduced even in 3
acetone. A competitive experiments show that imine B3 is >1000 times more reactive
than structurally related acetophenone B51 or ,-methylstirene B52.
MeOMeO
NOMeOMeO
B51B3B52
6 The hydrogenation rate and enantioselectivity is influenced by Ru catalyst (,-
arene and arylsulfonyl group in catalyst ligand) as well as by the solvent and reduction
16ctemperature used. The general sense of asymmetric induction with Ru(II)-catalyst B48-B50 system is illustrated in Figure B4. In the stereodetermining hydrogen-transfer
2step, Ru catalyst discriminates between the enantiofaces at the sp nitrogen atom of the
3imine, generating a stereogenic sp carbon.
"H"2Figure B4.
(S,S)
3R*L3H.R.NNN
RuC12RR2RH
1R(R,R)
""H2
Hydride transfer from active catalyst - Ru-hydride species to an imine requires out-of-
16bplane interaction between the Ru-H moiety and C=N bond. Noyori suggests that N-
H linkage in Ru catalyst B48-B50 can stabilize a transition state through hydrogen bonding with imine nitrogen (see Figure B4).
Recently, chiral Ru-(oxazolinylferrocenyl)phosphine catalyst B53 has been
17employed in asymmetric hydrosilylation of 2-aryl-3,4-dihydropyrrole B8.
40
PPh3PhClPhPRuClPhNFe
NNHO
1 mol%B53H
PhSiH22
B8
Corresponding (S)-amine was isolated with 88% ee and in 60% yield. Analogous Rh(I) catalyst showed considerably lower selectivity (34% ee) (see also Table B2, entry 4).
Conclusion. Ru(II) catalyzed transfer hydrogenation of cyclic imines affords the highest enantioselectivities for almost all substrates tested. In combination with operational simplicity it is the method of choice for the synthesis of chiral isoquinolines containing an aniline subunit.
181.4. Asymmetric hydrogenation of cyclic enamides.
Being closely related to olefins hydrogenation, asymmetric reduction of cyclic enamides is one of the most efficient tool for highly enantioselective synthesis of 1-
1919a,balkyl and 1-arylalkyl substituted isoquinolines. Initially, chiral Rh(I)-complexes
such as B55 were employed affording N-acetyl-1-methyl-1,2,3,4-
tetrahydroisoquinoline with reasonable enantioselectivities (81% ee in the best case): Scheme B6.
+PhPhOMePOMeRhPPh Ph-MeOMeOBFMeO4AcCl or AcO2B55NNCH3CHNNEt or Py3MeOMeO3*1 atm H, S/C = 500MeO2OO
B381% eeB5499% conversion
19c,d The method became synthetically important after Noyori had shownthat various N-
acylated 1-alkylidene and 1-benzylidene isoquinolines can be hydrogenated in the presence of Ru(II)-BINAP catalyst with excellent enantioselectivities (up to 100% ee, see Table B7).
41
MeOMeOXRRHNNNMeOMeOXOOOMeOOMe
MeOOMeOMe
(Z)-B56(E)-B56(Z)-B57: X = CH 3(Z)-B58: X = (CH) 24
zArAr+XPhOHPPhOPNRu*X-OORhPClO4OPArArPhPhz OMe
(R)-B62: Z=CH; Ar=Ph(S)-B613B59: X = CH 3; Ar=4-CHCH(R)-B63: Z=CF3364B60: X = (CH) 24
Table B7. Asymmetric hydrogenation of 2-acyl-1-alkylidene-1,2,3,4-tetrahydroisoquinolines B54, B56-B58.
Pressure Catalyst Entry Substrate S/C T Yield ee 0(atm) (C) (%) (%)
1 100 4 30 95 (Z)-B56 (R=CH) (S)-B61 75 3
2 50-200 4 - 100 B54 (S)-B62 96
3 200 1 30 100 (Z)-B56 (R=H) (R)-B62 >99.5
4 200 100 30 100 (Z)-B56 (R=H) (R)-B62 96
5 200 4 0 0 (Z)-B56 (R=H) (R)-B62 -
6 200 100 60 98 (Z)-B56 (R=H) (R)-B62 91
7 200 4 24 100 (Z)-B56 (R=CH) (R)-B62 >99.5 3
8 200 4 24 <3 (E)-B56 (R=CH) (R)-B62 - 3
9 200 4 24 10 (Z)-B56 (R=CF) (R)-B62 - 3
10 200 4 24 100 (Z)-B56 (R=tBu) (R)-B62 50
11 50-200 100 - 99 (Z)-B57 (R)-B63 96
12 50-200 100 - 98 (Z)-B58 (R)-B63 98
The comparison of catalysts clearly shows that Ru-complex is superior to cationic Rh(I)-catalyst (entry 7 vs. 1 and entry 2 vs. Scheme B6). Extensive studies revealed that N-acyl group is crucial for the reaction because it acts as a binding tether
to the catalytic metal center. Z-Olefin geometry is important for high reactivity and
19denantioselectivity. E-olefin could not be reduced under standard conditions (entry 8
42
vs. 7). Hydrogenation occurs regioselectively at the enamide part leaving tetra-substituted olefinic linkage intact (entry 11-12, product amines B59-B60). Both N-
formyl and N-acetyl amides can be used (entries 3 and 7), but the strongly electron-
CO-group and bulky pivaloyl group decreases reactivity and/or withdrawing CF3
enantioselectivity (entries 9-10 vs. 3 and 7). The reaction usually is run under 1-4 atm
0of hydrogen at 30C. Increased temperature and pressure results in lower
0enantioselectivity (entry 4 and 6 vs. 3), while diminishing the temperature to 0 C
causes inhibition of the reaction (entry 5 vs. 3).
Despite the excellent enantioselectivities achieved for a variety of isoquinolines, hydrogenation can not be employed for reduction of 1-aryl substituted substrates (formation of the corresponding enamide would require non-aromatic, unstable quinone-type structure B64):
MeOMeORRNNMeOOMeOO
B64
Asymmetric hydrogenation has been successfully employed for large-scale
20synthesis of piperazine-2-carboxamide B66, building block of Merck HIV protease
?inhibitor Indinavir.
43
OHNOHHNNNScheme B7.ONHtBuO
,Indinavir
OORR22-*+ )]SbF[Rh(NBD)(P-P6 a: R = NHtBu, R= BnO, R = t-BuONN1 244H / ClCHCHCl222b: R = OMe, R= Me, R = t-BuO1 211RRorc: R = NHtBu, R= H, R = t-BuONN1 2-+[(P-P*)Rh(MeOH)] TfO OOd: R = NHtBu,R= t-BuO, R = PhO21 2OOR1RH / MeOH12B65B66
HMeP(i-Bu)2PPh2
Fe*FeP-P : i-BuTRAP =P-P* : [2,2]-Phanephos =PPh2
(i-Bu)P2MeH
The enantioselectivity of reduction was found to depend on electronic character of olefinic double bond, which can be regulated by N-protecting groups,
4particularly at nitrogen in 4th position. N-Boc group was found to be the best. Careful
optimization of reduction conditions for each particular substrate combined with extensive screening of diphosphine ligands resulted in highly enantioselective large-scale synthesis of chiral piperazine B66.
Table B8. Optimized conditions for asymmetric hydrogenation of cyclic enamide B65
in the presence of Rh(I) catalysts.
Entry Enamide Catalyst T Yield Pressure Lit. Ligand ee 0*mol% (C) (%) ref. (atm) (P-P) (%)
1 BINAP 2 40 70 96 16a B65a 99
2 [2,2]-Phanephos - -40 1.5 >99 16b B65b 86
3 BINAP 7 40 100 >99 16c B65c 97
4 i-BuTRAP 1 50 1 85 16d B65d 96
The hydrogenation of enamides has been recently extended also to the synthesis
21aof cyclic amino acids B67 and B68 with various ring size.
44
Scheme B8.
[Rh(COD)-(R,R)-(Et-DuPHOS)]OTf
Size Chem(%) ee(%) 6 mol%( )nCOMeN( )a: n=0 84 0 2nCOMeN2H (6 atm) / r.t. / MeOHCOPhb: n=1 95 27 22PhCO2c: n=2 96 95 EtB67d: n=3 96 97 B68Pe: n=4 97 95 EtEt-DuPHOS:Etf: n=8 96 91 Pg: n=11 86 86
Et
Excellent enantioselectivities were achieved in the case of medium B68c-e and
large ring size B68f-g while 5- and 6-membered cyclic enamides B68a-b were
obtained either racemic or with low ee’s. In contrary, Comins was able to reduce
21benamide B67b with reasonable 80%ee employing [(S)-BINAP]Ru catalyst.
1.5. Catalytic enantioselective imines reduction by borane.
Despite to the very few reports on successful borane-mediated catalytic asymmetric reduction of imines, potentially it is an alternative to transition metals catalyzed hydrogenation, especially because there is no need of high-pressure equipment. Various chiral boranes have been used as catalysts and chiral Lewis acid complexation to nitrogen lone pair allowed to discriminate imine enantiofaces by a non-chiral hydride source, usually borane (BH-THF, BH-SMe etc.). 332
Enantioselective catalytic reduction has been applied mainly for acyclic imines and only recently Kang reported enantioselective reduction of cyclic imines - 3,4-
dihydroisoquinolines in the presence of 20 mol% chiral thiazazincolidine catalyst
22B70:
Scheme B9.MeOZnNMeOEtSMeO+MeMeONPhB70-RNHZnNMeONEtBoraneMeOHSRRMePh
B71B3: R = CH3CH3B19: R = 3,4-(MeO)CHCH2632HC3BDMPB: B21: R = 3,4-(MeO)CHZnO263BH N2B73:R = 3,4,5-(MeO)CHCHEt3622CHCHS33MePhB72
45
The reduction is assumed to proceed via formation of enantioface-selective
complex B71 between chiral Lewis acid B70 or B72 and isoquinoline nitrogen lone
pair and the complex geometry controls the reaction stereoselectivity. Borane-THF is the best reducing agent for all substrates (entries 1-3, Table B9), except of 1-aryl-3,4-dihydroisoquinoline B21, where BDMPB is superior (entry 5 vs. 4):
Table B9 Reduction of 3,4-dihydroisoquinolines with 20 mol% thiazazincolidine
0C: catalyst in toluene at -5
Entry Imine Catalyst Borane Yield (%) ee (%)
1 BH-THF 65 B3 B70 86 3
2 BH-THF 83 B19 B70 78 3
3 BH-THF 25 B73 B70 76 3
4 BH-THF 92 B21 B70 24 3
5 BDMPB 69 B21 B72 56
46
1.6. Summary.
Literature analysis was carried out to reveal the most suitable method for catalytic asymmetric synthesis of chiral isoquinolines with aniline subunit. The following conclusions can be made:
1. Rh(I) and Ir(I) catalyzed asymmetric hydrogenation and hydrosilylation procedure requires a careful optimization of catalyst, additives and reduction conditions and in the best case affords chiral isoquinolines with lower enantioselectivity than the corresponding titanocene and Ru(II) catalyzed processes.
-2. Despite to excellent enantiocontrol, several limitations in the case of ansa
titanocene catalyst, namely, relatively high hydrogen pressure (5.4-136 atm), highly demanding (air and moisture-free) reduction conditions combined with the necessity to activate titanocene catalyst prior to use is an important drawback of Buchwald’s
hydrogenation procedure. Moreover, ansa-titanocene catalyst shows low tolerance
toward various functional groups, such as unprotected imine N-H and aryl halogens.
3. Asymmetric hydrogenation of 3,4-dihydroisoquinoline-derived enamides is questionable in the case of 1-aryl substituted substrate since it requires formation of potentially unstable quinone-type structure.
4. Ru(II) catalyzed transfer hydrogenation of various 3,4-dihydroisoquinolines proceeds with excellent enantioselectivity and chemical yield. Operational simplicity (reduction can be preformed in open reaction vessel with in situ prepared catalyst)
makes Noyori procedure the method of choice for the synthesis of various chiral 1-aryl substituted isoquinolines.
47