Enantioselective kinetic resolution of trans-cycloalkane-1,2-diols

Article abstract of DOI:10.1002/anie.200800641

Finally! The title resolution is achieved with a nonnatural, partially rigid, lipophilic tetrapeptide at low catalyst loadings without additional base or cosolvents. The transition-state model (ball-and-stick model in the scheme; C gray, N blue, O red) emphasizes the interplay between hydrogen-bonding and hydrophobic interactions. (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200800641

Communications
DOI: 10.1002/anie.200800641
Organocatalysis
Enantioselective Kinetic Resolution of trans-Cycloalkane-1,2-diols**
Christian E. Müller, Lukas Wanka, Kevin Jewell, and Peter R. Schreiner*
The kinetic resolution[1] of chiral trans-cycloalkane-1,2-diols
(1) presents a formidable challenge[2] and a rare case where
chemical methods are superior to enzymatic approaches.
Taking the kinetic resolution of trans-cyclohexane-1,2-diol
(1a, n = 2) through monoacylation as the delineating test
reaction, it was shown that various Pseudomonas lipases
display both low activities (reaction times typically within the
range of days) and low selectivities.[3] Purely chemical trans-
formations utilizing benzoyl transfer in CuII-catalyzed reac-
tions with C2-symmetric bisoxazoline ligands give good ee
values for the monobenzoylated product (around 80%) and
good conversions (37–46%; selectivity factor s = 14–22)
within hours;[4,5] Diol 1a is resolved with up to 66% ee.[4]
Hence, the availability of a practical chemical method for the
catalytic enantioselective kinetic resolution of diols 1 is
desirable.
resolution of monoprotected diols and amino alcohols,[9] as
well as of polyols through acyl transfer.[8] The tetrapeptide
catalysts such as 3 utilized in some of these reactions are, at
least in our hands, much less effective for the resolution of
(Æ)-1a (see Figure 1 and the Supporting Information).
Toniolo et al.improved the efficacy of 3 also through the
introduction of lipophilic a-methylvaline.[10]
Figure 1. Catalyst screening (2) for the enantioselective acylation test
reaction. Reactions were run at low conversions (<10%) to determine
maximum activity. Enantiomeric ratios (e.r.) are given for 4a. The most
efficient catalyst, 2i, is outlined with a dotted line.
Herein we present an approach that utilizes the novel
lipophilic chiral tetrapeptide platform 2 (Boc = tert-butoxy-
carbonyl, AGly = g-aminoadamantanecarboxylic acid; AGly in
our shorthand notation emphasizes the relationship to the a-
amino acid glycine[12]), which is equipped with a nucleophilic
Our strategy for developing a practical method for the
resolution of (Æ )-1 was the preparation and utilization of
more lipophilic and somewhat structurally less flexible
oligopeptides; we hoped that diminished catalyst self-associ-
ation (dimerization or folding) would lead to low catalyst
loadings and would allow the use of nonpolar organic
solvents.It has been shown recently that organic solvents of
low polarity play a key role in the effective regeneration of
the catalyst such that they even allow the omission of auxiliary
base in 4-dimethylaminopyridine(DMAP)-catalyzed acyla-
tion reactions.[11] Not having to use an additional base
simplifies the workup and product purification.[9]
Our approach does not follow established design princi-
ples for oligopeptide catalysts which emphasize the formation
of catalytically important secondary structures (indicated by
the internal hydrogen bonds in 3).[6,8] Our concept culminated
in tetrapeptide catalysts of type 2 incorporating a rigid
nonnatural g-aminoadamantanecarboxylic acid as well as
several hydrophobic natural as well as nonnatural amino
acids.
N-p-methylhistidine moiety for enantioselective acyl trans-
[7,8]
fer.[6] Miller et al.have been highly successful
in the
[*] Dipl.-Chem. C. E. Müller, Dr. L. Wanka, K. Jewell,
Prof. Dr. P. R. Schreiner
Institute of Organic Chemistry
Justus-Liebig University
Heinrich-Buff-Ring 58, 35392 Giessen (Germany)
Fax: (+49)-641-9934309
E-mail: prs@org.chemie.uni-giessen.de
Homepage: http://www.chemie.uni-giessen.de
[**] This work was supported by the Deutsche Forschungsgemeinschaft
(SPP1179). We thank Prof. Scott J. Miller (Yale University) for
providing us his protocol for the synthesis of N-p-methylhistidine
and Dr. E. Röcker for competent analytical support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200800641.
6180
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6180 –6183

Angewandte
Chemie
Table 1: Enantioselective kinetic resolution of trans-cycloalkane-1,2-diols with oligopeptide catalyst 2i in
toluene.
Oligopeptides 2 with a varia-
tion only in R can be prepared in a
straightforward manner by auto-
mated solid-phase peptide-cou-
pling methods (see the Supporting
Information for details), and we
prepared and screened a large
variety of structures; the coupling
of the novel Fmoc-protected bulky
adamantyl g-amino acid proved
unproblematic despite its bulki-
ness.[12]
Substrate
Cat. [mol%] T [8C] t [h] Conv. [%][a] ee [%] (+)-1 ee [%] (À)-4 s[a]
2
À20
4
57
>99
75
>50
(Æ)-1a (0.4 mmol)[b]
(Æ)-1a (1 mmol)[c]
1
1
0
0
5
4
39[d]
37[e]
>99
>99
78
78
>50
>50
Millerꢀs peptide catalyst 3 (with
variation in using l- and d-proline)
was prepared for comparison by
published protocols.[9] Catalyst 3b
(d-pro) was shown to be highly
effective for the kinetic resolution
of trans-1,2-acetamidocyclohexa-
nol, which provides additional
amide hydrogen-bond interactions
with the catalyst.[6] The b-hairpin
structure of 3b, which was inferred
by NMR experiments, was held
responsible for its high selectivity
(s = 28).In contrast, the kinetic
2
À20
9
63
85
49
8[f]
2
À20
5
57
>99
77
>50
(Æ)-1c (0.4 mmol)[b]
(Æ)-1c (1 mmol)[c]
1
1
0
0
5
5
39[d]
41[e]
97
>99
80
79
37
>50
2
À20
6
55
>99
83
>50
(Æ)-1d (0.4 mmol)[b]
(Æ)-1d (1 mmol)[c]
1
1
0
0
5
5
40[d]
44[e]
>99
>99
82
85
>50
>50
[a] Conversions and s factors determined following the procedure of Kagan and Fiaud;[14] s factors above
50 are not reliable; since s factors do not vary with catalyst concentration (1–10 mol%), the approximate
formula is valid up to 50.[1,15] [b] Preparative experiment on a 0.43 mmol scale. [c] Preparative experiment
on a 1.0 mmol scale. [d] Yield of isolated product for 1 in preparative experiments on a 0.43 mmol scale.
[e] Yield of isolated product for 1 in preparative experiments on a 1.0 mmol scale. [f] Dichloromethane
added for solubility.
resolution
of
monoacylated
(Æ)-4a proved rather difficult (s =
1.4).[13] Our experiments with 3a
and 3b confirm these findings and
also show that the kinetic resolu-
tion of (Æ )-1a is practically inef-
fective (Figure 1).
The amino acid backbone presented in the twelve
tetrapeptides 2a–2l, with only a variation in R, turned out
to be generally effective for the enantioselective acylation of
(Æ )-1a (n = 2), which we utilized as our test reaction
(Figure 1).More lipophilic R groups give better results, with
R = methylenecyclohexane (2i, Cha as building block) being
the most effective.The optimal conditions for our acylation
protocol with nonpolar toluene as the only solvent and 1–2
mol% catalyst loading are also applicable to other racemic
trans-1,2-diols (Table 1).The selectivities for these reactions
are generally very good; trans-cyclopentane-1,2-diol (1b) is
the exception, partially because addition of some CH2Cl2 is
necessary to overcome the poor solubility of 1b in toluene.As
anticipated, the selectivities indeed largely depend on solvent
polarity.For comparison, we also examined the acylation of
1a in acetonitrile, CH2Cl2, and trifluoromethylbenzene under
otherwise identical conditions.The reaction times were much
longer to achieve appreciable conversions and the s factors
were significantly lower: 2.4 (CH3CN, 5.1% conversion, 48 h),
9.6 (CH2Cl2, 23.5% conversion, 24 h), and 8.9 (PhCF3, 23.2%
conversion, 4 h) (for details see the Supporting Information).
The omission of auxiliary base is possible because acetic
acid (pKa = 4.74) equilibrates with the methylimidazolium ion
(pKa = 7.3)[16] such that an appreciable amount of unproto-
nated catalyst is always available; the pKa values are likely to
be more similar in organic solvents owing to the higher
polarizability of methylimidazole.As a consequence, other
anhydrides delivering even weaker acids could show even
higher selectivities.Indeed, using isobutyric anhydride under
otherwise identical conditions for the resolution of (Æ )-1, we
found marginally better selectivity, however, at a much longer
reaction time for comparable conversion.This is apparently
because of higher steric demand; pivalic anhydride, with even
greater steric demand, is practically unreactive.
We also determined the absolute stereochemistry of
enantiomerically pure (+)-1a obtained with 2i as 1S,2S by
comparison with literature data (for details see the Support-
ing Information).[3] The (À) enantiomer can be prepared
readily by changing the stereochemistry at the histidine
moiety and the R position in tetrapeptide 2.This is
remarkable as this implies that the stereochemistry is
determined by the homo configuration of the amino acid
defining R and the catalytically active His moiety.We
followed up on this finding and searched for indications of
secondary-structure formation of 2i by NMR polarization-
transfer experiments, but these did not provide clear indica-
tions of specific intramolecular interactions.Hence, there
must be a structure-forming element at the stage of the
complexed acylium ion.As NMR studies of such ions are
hampered by various issues, we resorted to a molecular
dynamics search for low-lying conformations of the catalyst/
acylium ion adduct utilizing the Merck Molecular Force Field
Angew. Chem. Int. Ed. 2008, 47, 6180 –6183
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6181

Communications
(MMFF);[17] as we used a nonpolar solvent, this comparison
of the Fmoc protecting group as described above, the peptide was
elongated using Fmoc-AGly-OH (0.250 g, 0.6 mmol), HBTU, HOBt,
and DIPEA in the same stoichiometric ratio as given above.After
washing and subsequent cleavage of the Fmoc protecting group as
described above, the peptide was elongated using another double
coupling strategy (2 h shaking per coupling step) Boc-l-(p-Me)-His-
OH (0.121 g, 0.45 mmol), HBTU (0.228 g, 0.6 mmol), HOBt (0.092 g,
0.6 mmol), and DIPEA (0.155 g, 204.1 mL, 1.2 mmol) per coupling
step (1.5:2:2:4 equiv, respectively). After washing (five times each
with DMF, dichloromethane, and diethyl ether), 2i was cleaved from
the resin by shaking two times for two days with methanol, triethyl-
amine, and THF (9:1:1, v/v).The resin was filtered off and washed
several times with THF.The collected solutions were concentrated
and the residue was purified by HPLC (eluent: tert-butyl methyl
ether(TBME)/CH3OH 85:15, 6 mLminÀ1; UV detector l = 254 nm,
should be qualitatively valid.Irrespective of the starting
geometry, the most favorable conformers always placed the
cyclohexyl group in 2i in close proximity to the imidazole/
acylium ion adduct (Figure 2).
Emax = 2.56; refractometer; column l = 250 mm, d = 8 mm, LiChro-
sorb Diol (7 mm, Merck); retention time (2i) = 10.43 min). The
peptide was characterized by ESI-MS, HR-ESI-MS, NMR, IR, and
EA.
1H NMR (600 MHz, CDCl3): d = 7.35 [s, 1H, CH-imidazole
(His)], 7.24–7.15 [m, 3H, HAr (Phe)], 7.05–7.01 [m, 2H, HAr (Phe)],
6.79 [s, 1H, CH-imidazole (His)], 6.44 [d, J = 7.8 Hz, 1H, NH (Phe)],
5.91 [d, J = 7.9 Hz, 1H, NH (Cha)], 5.68 [s, 1H, NH (AGly)], 5.09 [d,
J = 8.3 Hz, 1H, NH (His)], 4.78–4.70 [m, 1H, Ha (Phe)], 4.41–4.30 [m,
1H, Ha (Cha)], 4.13–4.03 [m, 1H, Ha (His)], 3.64 (s, 3H, OCH3), 3.54
(s, 3H, NCH3), 3.09–2.98 [m, 2H, Hb (Phe)], 2.98–2.88 [m, 2H, Hb
(His)], 2.13 (m, 2H, adamantane), 1.93–1.80 (m, 6H, adamantane +
Cha), 1.71–1.51 (m, 12H, adamantane + Cha), 1.40–1.36 (m, 1H,
Figure 2. Model for the enantioselective acylation of trans-cycloalkane-
1,2-diols in the “pocket” of the acylated catalyst. Hydrogen atoms on
the catalyst are omitted for clarity. C gray, N blue, O red.
=
The two geometrically near C O groups are likely to
provide the hydrogen-bonding contacts needed for chiral
recognition of the diols.Our finding that more hydrophobic R
groups provide higher ee values could also be rationalized by
the additional hydrophobic interactions with the substrate.
The model also emphasizes that the AGly building block
provides the scaffold that holds the catalytic site and the
centers governing recognition and stereochemistry in place.
This model will provide the basis for further catalyst develop-
ment.
We have identified a tetrapeptide incorporating natural
and unnatural amino acids capable of stereoselective acyl-
group transfer onto trans-cycloalkane-1,2-diols.The kinetic
resolutions presented here provide exceptionally high selec-
tivities that are made possible through the interplay of an
unnatural cage g-amino acid that provides some rigidity and a
lipophilic amino acid in the chain allowing for hydrophobic
interactions in our proposed transition-state model.The lack
of secondary structure in the free catalyst implies that the
factors determining the stereochemistry are developed in the
charged acylium ion complex with the peptide catalyst and
the subsequent stereodifferentiating interactions of this com-
plex with the substrate.
Cha), 1.37 [s, 9H, C(CH3)3], 1.23–1.00 (m, 4H, Cha), 0.92–0.69 ppm
13
=
(m, 2H, Cha). C NMR (150 MHz, CDCl3): d = 176.3 (C O), 171.9
=
=
=
=
(C O), 171.6 (C O), 169.7 (C O), 155.4 (C O), 138.3, 135.7, 129.2,
128.6, 128.2, 127.2, 127.2, 80.5, 54.4, 53.2, 52.3, 50.7, 42.5, 42.1, 40.3,
40.3, 39.5, 38.2, 38.0, 37.8, 35.1, 34.2, 33.5, 32.7, 31.5, 29.1, 29.1, 28.3,
26.8, 26.3, 26.1, 26.1 ppm; IR (KBr): n˜ = 3427, 2921, 2853, 2912, 1746,
1661, 1510, 1518, 1450, 1366, 1280, 1249, 1169 cmÀ1.MS: a) ESI: m/z =
761.5 [M+H]+ (calcd m/z = 761.5), m/z = 783.4 [M+Na]+ (calcd m/z =
783.4), m/z = 1521.3 [2M+H]+ (calcd m/z = 1521.9), m/z = 1543.3
[2M+Na]+ (calcd m/z = 1543.9); b) HR-ESI: m/z = 761.45963
[M+H]+ (calcd m/z = 761.45963). Anal. calcd for C32H43N5O6: C
66.29, H 7.95, N 11.04; found C 64.45, H 7.75, N 10.33.
Example illustrating the general procedure for the preparative
kinetic resolution of the cyclic diols: Catalyst 2i (3.3 mg, 0.0043 mmol,
1 mol%) and diol (Æ )-1a (50 mg, 0.43 mmol) were dissolved in
80 mL of anhydrous toluene to produce a clear solution.The reaction
mixture was cooled to 08C, and acetic anhydride (0.215 mL,
2.28 mmol, 5.3 equiv), which was cooled to 08C, was added and
allowed to stir for 4.5 h at 08C.The reaction mixture was quenched
with 10 mL methanol and filtered using 37 g silica gel suspended with
EtOAc to remove the catalyst and acetic acid (the silica gel was
washed with EtOAc).After the filtration, the solvents were removed
under reduced pressure.The crude product was then applied directly
to a silica gel column.Eluting with EtOAc afforded 339.mg
(0.214 mmol, 50.0%) of monoacetate 4a (Rf = 0.47) and 19.4 mg
(0.167 mmol, 38.8%) of diol 1a (Rf = 0.20). The products were then
directly characterized by chiral GC analysis and NMR.
Experimental Section
Tetramer 2i was synthesized on solid support using commercially
available Wang polystyrene resin end-capped and preloaded with
Fmoc-protected l-phenylalanine (0.405 g, 0.74 mmolgÀ1, 0.3 mmol).
Fmoc cleavage was performed by shaking the resin twice in 25%
piperidine in DMF (v/v).The resin was washed five times each with
DMF, dichloromethane, and DMF.Chain elongation with Fmoc- l-
Cha-OH was performed by a double coupling procedure (1 h shaking
per coupling step) using Fmoc-l-Cha-OH (0.237 g, 0.6 mmol), O-
(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophos-
phate (HBTU; 0.228 g, 0.6 mmol), 1-hydroxy-1H-benzotriazole
monohydrate (HOBt·H2O (0.092 g, 0.6 mmol), and diisopropylethyl-
amine (DIPEA; 0.155 g, 204.1 mL, 1.2 mmol) per coupling step
(2:2:2:4 equiv, respectively).After washing and subsequent cleavage
Received: February 8, 2008
Revised: April 14, 2008
Published online: July 10, 2008
Keywords: acylation · alcohols · kinetic resolution ·
.
organocatalysis · peptides
[1] E.Vedejs, M.Jure, Angew. Chem. 2005, 117, 4040; Angew. Chem.
Int. Ed. 2005, 44, 3974.
6182
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 6180 –6183

Angewandte
Chemie
[2] P.Somfai, Angew. Chem. 1997, 109, 2849; Angew. Chem. Int. Ed.
Engl. 1997, 36, 2731.
[11] A.Sakakura, K.Kawajiri, T.Ohkubo, Y.Kosugi, K.Ishihara,
J.
Am. Chem. Soc. 2007, 129, 14775.
[3] K.Naemura, R.Fukuda, M.Murata, M.Konishi, K.Hirose, Y.
Tobe, Tetrahedron: Asymmetry 1995, 6, 2385.
[4] C.Mazet, S.Roseblade, V.Kohler, A.Pfaltz, Org. Lett. 2006, 8,
1879.
[12] L.Wanka, C.Cabrele, M.Vanejews, P.R.Schreiner,
Eur. J. Org.
Chem. 2007, 1474.
[13] S.J.Miller, G.T.Copeland, N.Papaioannou, T.E.Horstmann,
E.M.Ruel, J. Am. Chem. Soc. 1998, 120, 1629.
[5] Y.Matsumura, T.Maki, S.Murakami, O.Onomura,
J. Am.
[14] H.B.Kagan, J.C.Fiaud, Top. Stereochem. 1988, 18, 249.
Chem. Soc. 2003, 125, 2052; A.Gissibl, M.G.Finn, O.Reiser,
Org. Lett. 2005, 7, 2325.
[15] J.M.Goodman, A-.K.Köhler, S.C.M. Alderton,
Lett. 1999, 40, 8715; H.F.T.Klare, M.Oestreich, Angew. Chem.
2007, 119, 9496; Angew. Chem. Int. Ed. 2007, 46, 9335.
Tetrahedron
[6] E.A.C.Davie, S.M.Mennen, Y.J.Xu, S.J.Miller,
2007, 107, 5759.
Chem. Rev.
´
[16] J.J.Klicic , R.A.Friesner, S.-Y.Liu, W.C.Guida,
J. Phys. Chem.
[7] E.R.Jarvo, S.J.Miller, Tetrahedron 2002, 58, 2481.
[8] J.T.Blank, S.J.Miller, Biopolymers 2006, 84, 38.
[9] E.R.Jarvo, G.T.Copeland, N.Papaioannou, P.J.Bonitatebus,
S.J.Miller, J. Am. Chem. Soc. 1999, 121, 11638.
[10] F.Formaggio, A.Barazza, A.Bertocco, C.Toniolo, QB. .
Broxterman, B.Kaptein, E.Brasola, P.Pengo, L.Pasquato, P.
Scrimin, J. Org. Chem. 2004, 69, 3849.
A 2002, 106, 1327.
[17] T.A. Halgren, in Encyclopedia of Computational Chemistry,
Vol. 2 (Eds.: P.von R. Schleyer, N.L. Allinger, T. Clark, J.
Gasteiger, P.A.Kollman, H.F.Schaefer, P.R.Schreiner), Wiley,
Chichester, 1998, p.1033.
Angew. Chem. Int. Ed. 2008, 47, 6180 –6183
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6183