Stereoselective hydrogen bromide-promoted hydrogenation of an α-hydroxyoxime

Article abstract of DOI:10.1016/S0040-4039(00)01404-0

Hydrogen bromide has been demonstrated to provide optimal cis-selectivity in the reduction of 4-chromanone α-hydroxyoxime (25:1 cis/trans) in 94% yield. This reaction is pivotal in the synthesis of cis-aminochromanol. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01404-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 8021–8025
Stereoselective hydrogen bromide-promoted hydrogenation
of an a-hydroxyoxime
Ian W. Davies,* Mark Taylor, Jean-Franc¸ois Marcoux, Louis Matty, Jimmy Wu,
David Hughes and Paul J. Reider
Department of Process Research, Merck & Co., Inc., PO Box 2000, Rahway, NJ 07065, USA
Received 13 July 2000; accepted 7 August 2000
Abstract
Hydrogen bromide has been demonstrated to provide optimal cis-selectivity in the reduction of
4-chromanone a-hydroxyoxime (25:1 cis/trans) in 94% yield. This reaction is pivotal in the synthesis of
cis-aminochromanol. © 2000 Elsevier Science Ltd. All rights reserved.
In the synthesis of the HIV protease inhibitor Crixivan®,1 an expedient way of producing
cis-aminoindanol and related aminoalcohols was established using a Jacobsen epoxidation/
Ritter-type reaction sequence.2 These aminoalcohols have been used successfully in a ‘conforma-
tional toolbox’ of oxazoline ligands.3 Recently, we have begun to explore the use of aminoalco-
hols that are inaccessible via the Jacobsen/Ritter sequence, e.g. cis-aminochromanol 1.4 The
racemic cis-aminochromanol has previously been prepared by reaction of chromene with INCO5
and hydrolysis of the oxazolidinone or borane reduction of the O-benzyloxime.6
One of the most ‘atom-economical’ means of preparing an amine is hydrogenation of an
oxime where the only by-product is water.7 The hydrogenation of indane-derived keto- or
hydroxy-oximes has been known for almost four decades.8 The cis-aminoindanols are formed in
acidic media9 whereas the trans isomers are formed predominantly in neutral or alkaline
solution. Cis-aminochromanol 1 might also be accessible via reduction of the hydroxyoxime 3
under acid conditions, although the outcome in this case would be less clear-cut than in the
indane series due to the plasticity of the chromane ring system. In this paper we describe the
* Corresponding author. E-mail: ian davies1@merck.com
–
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01404-0

8022
discovery of a stereoselective hydrogen bromide promoted hydrogenation and our studies
directed at understanding the origin of the stereoselectivity. For our initial studies in the racemic
series, hydroxyoxime 3 was prepared as a crystalline solid in 85% overall yield from chromanone
by Moriarty oxidation10 and in situ oxime formation. Palladium black has often been used for
oxime reductions but we have found the outcomes to be dependent on the catalyst vendor. High
catalyst loadings were required and this coupled with the lack of availability at large scale drove
our change to Pd/C as catalyst.11 Hydrogenations using a variety of acids (1.0 equiv.) were run
at 40 psig, 12 hours, 10°C, 0.3 M in methanol and with 3 mol% catalyst (Table 1).
Table 1
Entry
Acid
1/2 cis/transa
Assay yielda (%)
1
2
3
4
5
1.3:1
2.2:1
7:1
96
79
95
94
0
HF
HCl
HBr
HI
23:1
6
7
8
9
10
11
12
13
14
H2SO4
HNO3
HPF6
CF3SO3H
CH3SO3H
B(OH)3
CF3CO2H
CH3CO2H
(CF3)3COH
1.9:1
2.1:1
2.0:1
3.2:1
4.7:1
1.4:1
1.7:1
1.5:1
6.3:1
80
92
89
90
86
90
92
87
52
a Based on HPLC assay.13
From this panel of experiments, it is clear that HBr provides optimal selectivity. Due to the
levelling effect, there is no acid stronger than MeOH2+ (pKa=−2 relative to water) in methanol.
Hence, bromide ion (entry 4) and not acid strength is implicated as having a role in the
hydrogenation selectivity.12 HI (entry 5) led to precipitation of a palladium mirror which was
not a competent catalyst. In order to determine if HBr was also causing desorption of the
palladium, the reaction solution was analyzed but <1 ppm Pd was detected by ICP-MS.
Table 2
Entry
Halide (1.0 equiv.)
1/2 cis/transa
Assay yielda (%)
1
2
3
4
5
6
7
48% HBr
HBr (acetic acid)
Lutidine·HBr
23:1
23:1
94
93
74
88
89
90
87
3.8:1
1.9:1
8.1:1
7.9:1
2.6:1
Bu4NBr
Bu4NBr/HBr (1:1)
Bu4NCl/HBr (1:1)
Bu4NBr/(CF3)3COH (1:1)
a Based on HPLC assay.

8023
The role of bromide ion was examined in our next panel of experiments (Table 2). It seems
clear that an acid and a bromide ion are required for high selectivity. Entries 5 and 6, both using
0.5 equiv. HBr gave the same selectivity irrespective of which supplemental halide was added to
the reaction medium.
Solvent effects were also briefly examined (Table 3).14 The use of higher levels of water
significantly reduced the selectivity, presumably due to increased solvation of the bromide ion
and attenuation of the system acidity (pKa=−1.7). Having clearly defined the need for both
bromide ion and proton we investigated the effect of stoichiometry (Table 4). Optimal selectivity
was achieved using 1.0 equiv. of HBr (entry 5). From a practical sense, the use of higher levels
of HBr leads to increased solvolysis to give 4 if hydrogenation is not begun promptly after the
addition of the acid.
Table 3
Entry
Solvent
1/2 cis/transa
Assay yielda (%)
1
2
3
MeOH
1:1 MeOH/H2O
1:2 MeOH/H2O
23:1
12:1
11:1
94
92
89
a Based on HPLC assay.
Table 4
Entry
Equiv. 48% HBr
1/2 cis/transa
Assay yielda (%)
1
2
3
4
5
6
7
1.3:1
5.4:1
12:1
17:1
23:1
23:1
23:1
96
94
92
96
94
93
91
0.25
0.50
0.75
1.00
1.25
1.50
a Based on HPLC assay.
The hydroxylamine 515 was isolated at incomplete conversion. In the hydrogenation of oxime
with 94% ee, the aminoalcohol is isolated with 94% ee,16 i.e. the first reduction involves the CꢀN
bond rather than the enamine tautomer 6. NMR experiments indicated protonation of oxygen
rather than at nitrogen. Whilst we have no clear explanation for the role of HBr, it might be
speculated that it serves to reorganise the catalyst surface to provide an optimal site for selective
hydrogenation.

8024
With the optimal conditions selected, the hydrogenation of the racemic oxime was run at
multi-kilo scale at 0–5°C (cis/trans 25:1, 94% assay yield). After filtration of the catalyst, ion
exchange followed by concentration and addition of (S)-mandelic acid gave the (S,S,S) salt
(>99% ee) free of the trans isomer in 40% overall yield from the racemic oxime.17
In summary, we have developed a highly selective synthesis of S,S-aminochromanol in 32%
overall yield from chromanone. The synthesis involves a pivotal cis-selective hydrogenation of
the hydroxyoxime either as a racemate or single enantiomer form. We are currently exploring
the generality of the Pd/CꢁHBr catalyst system.
Acknowledgements
We would like to thank Barry Trost and Barry Sharpless for helpful discussions.
References
1. Vacca, J. P.; Dorsey, B. D.; Schleif, W. A.; Levin, R. B.; McDaniel, S. L.; Darke, P. L.; Zugay, J; Quintero, J.
C.; Blahy, O. M.; Roth, E.; Sardana, V. V.; Schlabach, A. J.; Graham, P. I.; Condra, J. H.; Gotlib, L.; Holloway,
M. K.; Lin, J.; Chen, I.-W.; Vastag, K.; Ostovic, D.; Anderson, P. S.; Emini, E. A.; Huff, J. R. Proc. Natl. Acad.
Sci. USA 1994, 91, 4096.
2. Davies, I. W.; Reider, P. J. Chem. Ind. 1996, 412.
3. Davies, I. W.; Gerena, L.; Cai, D.-W.; Larsen, R. D.; Verhoeven, T. R.; Reider, P. J. Tetrahedron Lett. 1997, 38,
1145.
4. Lee, N. H.; Muci, A. R.; Jacobsen, E. N. Tetrahedron Lett. 1991, 32, 6533.
5. Julian, D. R.; Matusiak, Z. S. J. Heterocycl. Chem. 1975, 12, 1179.
6. Ghosh, A. K.; McKee, S. P.; Sanders, W. M. Tetrahedron Lett. 1991, 32, 711.
7. Trost, B. M. Science 1991, 254, 1471. Trost, B. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 259.
8. Rosen, W. E.; Green, M. J. J. Org.Chem. 1963, 28, 2797. Rimek, H.-J.; Yurraphat, T.; Zymalkowski, F. Liebigs
Ann. Chem. 1969, 725, 116. Huebener, C. F.; Donoghue, E. M.; Novak, C. J.; Dorfman, L.; Wenkert, E. J. Org.
Chem. 1970, 35, 1149.
9. Kajiro, H.; Mitaamura, S.; Mori, A.; Hiyama, T. Synlett 1998, 51. Kajiro, H.; Mitaamura, S.; Mori, A.; Hiyama,
T. Bull. Chem. Soc. Jpn. 1999, 72, 1093.
1
10. Mp 122–123°C. Anal. C, 59.94; H, 5.17; N, 7.60. C9H9NO3 requires C, 60.33; H, 5.06; N, 7.82; H NMR (400
MHz DMSO-d6) Major isomer: l 9.70–9.93 (br, 1H), 8.35 (dd, 1H, J=7.8, 1.6), 7.86 (dt, 1H, J=7.8,1.7), 7.54
(dt, 1H, J=11.7, 1.1), 7.49 (dd, 1H, J=8.2, 1.1), 5.68 (t, 1H, J=2.2), 4.91 (dd, 1H, J=12.4, 2.3), 4.59 (dd, 1H,
J=12.4, 2.1), 4.13–4.38 (br, 1H). Selected minor isomer peaks: l 9.13 (dd, 1H, J=8.1, 1.7), 7.91 (dt, 1H, J=7.9,
1.7), 4.94 (dd, 1H, J=9.8, 2.6), 4.76 (dd, 1H, J=12.8, 2.8). Moriarty, R. M.; Hu, H.; Gupta, S. C. Tetrahedron
Lett. 1981, 232, 1283. Moriarty, R. M.; Prakash, O.; Thachet, C. T. Synth. Commun. 1984, 14, 1373.
11. Rh/C gave 1.5:1 cis selectivity and no reaction in the presence of 1.0 equiv. HBr (18 h, 40 psig).
12. Bromide is believed to inhibit the non-selective hydrogenation sites in the Raney–Ni tartaric acid asymmetric
reduction of methylacetoacetate. Harada, T.; Yamamoto, M.; Onaka, S.; Imaida, M.; Ozaki, H.; Tai, A.; Izumi,
Y. Bull. Chem. Soc. Jpn. 1981, 54, 2323.
,
13. YMC ODS AQ (4.6×250 mm, S-5 micron 120 A). Eluent CH3CN, 5 mM K2HPO4 at pH 8, 35°C.
14. Other polar solvents, e.g. DMF, DMPU gave 2:1 selectivity.
1
15. Mp 92–95°C; LC/MS m/z 181; H NMR (400 MHz, DMSO-d6) l 7.25 (1H, dd, J=8, 1), 7.10 (1H, dt, J=8, 1),
6.83 (1H, dt, J=8, 1), 6.69 (1H, dd, J=8, 1), 5.15 (1H, br.s), 4.95 (1H, br.s), 4.55 (1H, d, 3.5), 3.96–3.99 (2H,
m), 3.80–3.92 (2H, m).
16. Hydroxychromanone starting material had 94% ee by SFC assay—Chiralpak AD, 300 Bar; 4–40% MeOH at
2%/min; 1.5 mL/min; 35°C; 215 nm.

8025
17. To a solution of oxime 3 (2.50 kg, 14 mol) in methanol (49 L) at 0°C was charged 48% aqueous HBr (1.9 L).
10% Pd/C (1.9 kg, 62% water wet) was charged and the mixture was hydrogenated in a stirred autoclave at
0–5°C, 40 psig for 12 h (cis/trans 25:1, 90% assay yield of cis isomer). The mixture was filtered through solka floc.
The batch was eluted through Dowex 1x2 on the base-cycle using methanol. The solution of racemic amine was
solvent switched to ethanol and S-mandelic acid (2.1 kg, 14 mol) was added in ethanol. The mixture was cooled
and the mandelate salt was isolated by filtration. The batch was dried to give 1.8 kg of (S,S,S)-mandelate salt
(>99% ee, 40% overall yield). To a slurry of the salt (1.7 kg) in isopropylacetate (IPAC) was added aq.
ethanolamine (6 L). The layers were separated and the aqueous layer was extracted with IPAC (3×). The IPAC
extracts were concentrated to ꢀ8 L and crystallised by the addition of n-heptane (8 L). (S,S)-aminochromanol
1 was isolated by filtration and dried to give a colorless solid (0.80 kg, 93%). The rotation for (S,S)-aminochro-
manol (1% in MeOH, 405 nm) was −177.9, DSC onset 110°C, peak 111°C. (R,R)-aminochromanol was prepared
similarly, rotation (1% in MeOH, 405 nm) −179.8, DSC onset 110°C, peak 112°C. Ee was determined by HPLC
using a crownpak CR(+) column using HClO4.
.