Steroidal guanidinium receptors for the enantioselective recognition of N-acyl α-amino acids

Article abstract of DOI:10.1039/a808245f

Guanidinium cations 4 and 5 extract N-acetyl α-amino acids into CHCl₃ from an aqueous medium with enantiomeric excesses of up to 80%.

Full text of DOI:10.1039/a808245f

Steroidal guanidinium receptors for the enantioselective recognition of N-acyl
a-amino acids
Anthony P. Davis* and Laurence J. Lawless
Department of Chemistry, Trinity College, Dublin 2, Ireland. E-mail: adavis@tcd.ie
Received (in Liverpool, UK) 23rd October 1998, Accepted 9th November 1998
Guanidinium cations 4 and 5 extract N-acetyl a-amino acids
O
into CHCl3 from an aqueous medium with enantiomeric
R
N
H
N
excesses of up to 80%.
O
MeO
O
O
H
N
H
O
Enantioselective recognition is of continuing interest in supra-
molecular chemistry,1 especially where relevant to the large-
scale resolution of racemates. Enantioselective phase transfer is
particularly attractive, raising the possibility of ‘catalytic’
resolutions based on the transport of substrates through
otherwise impermeable barriers.2 Carboxylates/carboxylic
acids are suitable targets for this approach, because of their
ability to partition between aqueous and organic phases. Amino
acids, and their N-acyl derivatives, are attractive substrates
because of their biological significance and practical im-
portance.3
Herein we describe a new strategy for the enantioselective
extraction of chiral carboxylates from aqueous into organic
media, and its realisation in the form of receptors which show
significant enantiodiscrimination in the case of N-acetyl a-
amino acids. In common with previous work on carbohydrate4
and anion recognition,5 our design exploits cholic acid 1 as a
starting material. The secondary hydroxy groups in 1 can be
independently modified to give differentiated, co-directed
substituents as in 2.6 Groups A–C are spaced so as to allow, in
O
N
O–
Ph
N+
N
H
H
H
H
N
N+
H
R′
Ph
3•–O2CR′
4
O
O
MeO
O
O
H
O
N
Ph
Cl
N
N
H
H
N
H
N+
H
Cl
5
dependent, allowing the determination of enantioselectivities,
as well as extraction efficiencies, by simple integration. The
results are shown in Table 1. Extraction efficiencies were
moderate to good for substrates with non-polar side-chains,
although neither receptor was effective with the polar aspar-
agine derivative. Receptor 4 proved remarkably consistent in its
12
O
7
HO
OH
3
OH
1
OH
O
OH
OMe
O
OH
O
i,ii
OMe
MeO
C
OH
N3
B
H
6
A
2
OH
HN
H
a typical case, cooperative effects on a substrate with minimum
interference from intramolecular interactions. The design is
suggestive of ‘three-point contact’, as required for the classical
model of enantioselective recognition.7
HN
S
iii–vi
O
OH
For carboxylate extraction, it is useful that one of groups A–C
should form a specific, electroneutral complex with the anionic
centre. To serve this purpose we have chosen the guanidinium
unit 3, capable of binding carboxylates as shown.8 Of the
possible variations on our general theme, we report the
synthesis and properties of two initial examples; the bis-
phenylcarbamate 4 and its asymmetrically-substituted relative
5. Receptor 4 was accessible from 3a-azide 69 as shown in
Scheme 1, while receptor 5 was prepared via a longer sequence
with alcohol 7 as the penultimate intermediate.
BocNH
OMe
vii,vi
4•Cl–
OH
HN
H
+
NH
HN
Cl–
As anticipated, solutions of 4·Cl2 and 5·Cl2 in CHCl3 were
capable of extracting carboxylic acids from neutral or basic
aqueous solutions, presumably through exchange of carbox-
ylate for chloride. In the case of N-acetyl a-amino acids the 1H
NMR spectra of the complexed substrates were enantiomer-
Scheme 1 Reagents and conditions: i, Zn dust, AcOH, room temp., 24 h; ii,
SCN(CH2)3NHBoc, Pri2NEt, dry CH2Cl2, room temp., 72 h; iii, MeI,
MeOH, reflux; iv, TFA, CH2Cl2, room temp.; v, Pri2NEt, MeOH, room
temp., 24 h; vi, aq. NaOH then aq. HCl; vii, PhNCO, conc. aq. HCl (cat.),
CH2ClCH2Cl, reflux, 72 h.
Chem. Commun., 1999, 9–10
9

Table 1 Extractions by 4·Cl2 and 5·Cl2 of racemic N-acetyl a-amino acids
intermolecular NOE was observed from the a-CH in -8 to the
L
a
from aqueous buffer (pH 7.4) into CHCl3
ortho protons of 5-NHPh (consistent with Fig. 1, allowing for
some rotational freedom about the N–Ph bond‡). A similar
MCMM search† on 5 + -8 yielded a higher-energy structure in
Receptor 4
Receptor 5
D
which the acetyl O···HN interaction is missing, the 12-carba-
moyl NH forming a fourth (apparently strained) hydrogen bond
to the carboxylate.
Extraction Enantio-
Extraction Enantio-
efficiency selectivity efficiency selectivity
c
c
(L:D)
Substrate
(mol%)b
(
L
:
D)
(mol%)b
Viewed as forerunners of an extended family of receptors, 4
and 5 show encouraging levels of enantioselectivity. Many
variants are within easy reach, a majority with much greater
differentiation between the three substituents. We hope to report
examples with improved performance in the foreseeable
future.
Financial support for this work was provided by Forbairt, the
Irish Science and Technology Agency, Schering Plough
(Avondale) Ltd., and the EU Training and Mobility for
Researchers programme. We are grateful to Peter Ashton
(University of Birmingham) for mass spectra, Dr John O’Brien
for non-routine NMR experiments, and Freedom Chemical
Diamalt GmbH for generous gifts of cholic acid.
N-Ac-alanine
N-Ac-phenylalanine
N-Ac-tryptophan
N-Ac-valine
52
87
83
7:1
7:1
7:1
76
93
92
89
82
93
74
0
6:1
9:1
6:1
9:1
5:2
9:1
4:1
—
71
7:1
d
d
N-Ac-tert-leucine
N-Ac-methionine
N-Ac-proline
N-Ac-asparagine
a
d
d
d
d
0
—
Solutions of receptor in CHCl3 (6 m
M
, 1 ml), and substrate in
phosphate buffer (7–8 m
M, 5 ml), were stirred vigorously for 2 h. The
organic phases were isolated, dried by passage through hydrophobic filter
paper, then evaporated. The residues were dissolved in CDCl3 (0.6 ml) and
analysed by 1H NMR spectroscopy. b Determined by 1H NMR integration
of substrate a-CH and NH vs. 7/12b-H of receptor. Determined by 1H
c
NMR integration of a-CH and NH signals for enantiomers of substrates.
Assignments confirmed through control experiments with enantiopure
substrates. d Not determined.
Notes and references
† Calculations employed MacroModel V5.5 (ref. 10), the Amber* force
field, CHCl3 GB/SA solvation, and 5000 steps of MCMM. Six and three
separate searches were conducted for the L and D substrates respectively, all
from widely differing starting geometries and all yielding essentially similar
final structures.
‡ Rotation about N–Ph allows an ortho proton to make van der Waals
contact with the substrate a-CH.
ability to differentiate between enantiomers, irrespective of
side-chain bulk. Receptor 5 showed generally higher extraction
abilities, possibly due to the greater acidity of the di-
chlorophenylcarbamoyl NH, and was more sensitive to side-
chain structure. Perhaps surprisingly, the substrate with the
most sterically hindered asymmetric centre (N-Ac-tert-leucine)
gave the lowest selectivity.
1 T. H. Webb and C. S. Wilcox, Chem. Soc Rev., 1993, 22, 383.
2 M. Newcomb, J. L. Toner, R. C. Hegelson and D. J. Cram, J. Am. Chem.
Soc., 1979, 111, 6294.
1H NMR spectroscopy and molecular modelling combined to
suggest plausible models for the binding geometries. A Monte
Carlo Molecular Mechanics (MCMM) search† on the complex
3 Enantioselective extraction/transport of amino acids or N-acyl amino
acids : (a) J. L. Sessler and A. Andrievsky, Chem. Eur. J., 1998, 4, 159;
(b) N. Voyer and B. Guerin, Chem. Commun., 1997, 2329; (c) J. Y.
Zheng, K. Konishi and T. Aida, Tetrahedron, 1997, 53, 9115; (d) K.
Konishi, K. Yahara, H. Toshishige, T. Aida and S. Inoue, J. Am. Chem.
Soc., 1994, 116, 1337; (e) A. Metzger, K. Gloe, H. Stephan and F. P.
Schmidtchen, J. Org. Chem., 1996, 61, 2051; (f) H. Tsukube, J. Uenishi,
T. Kanatani, H. Itoh and O. Yonemitsu, Chem. Commun., 1996, 477; (g)
G. J. Pernia, J. D. Kilburn and M. Rowley, J. Chem. Soc., Chem.
Commun., 1995, 305; (h) A. Galán, D. Andreu, A. M. Echavarren, P.
Prados and J. de Mendoza, J. Am. Chem. Soc., 1992, 114, 1511; (i) J. de
Mendoza and F. Gago, in Computational approaches in supramolecular
chemistry, ed. G. Wipff, Kluwer Academic Publishers, 1994, p. 79. Ref.
3(a) serves as a leading reference to amino acid recognition in
general.
4 A. P. Davis, Chem. Soc. Rev., 1993, 22, 243; A. P. Davis, R. P. Bonar-
Law and J. K. M. Sanders, in Comprehensive Supramolecular
Chemistry, ed. Y. Murakami, Pergamon, Oxford, 1996, vol. 4
(Supramolecular Reactivity and Transport: Bioorganic Systems), p.
257.
5 A. P. Davis, J. F. Gilmer and J. J. Perry, Angew. Chem., Int. Ed. Engl.,
1996, 35, 1312; A. P. Davis, J. J. Perry and R. P. Williams, J. Am. Chem.
Soc., 1997, 119, 1793.
between 5 and N-acetyl- -valinate 8 yielded the configuration
L
shown in Fig. 1. The carboxylate accepts H-bonds from the
7-carbamoyl and two guanidinium NH groups, while the acetyl
oxygen is bound to the 12-carbamoyl NH. In support of this
structure, the receptor carbamate and 2 of the 3 guanidinium NH
signals moved downfield on complex formation, while a weak
O
Ph
O
OMe
N
H
O
O
H
–O
N
O
OH
HN
H
+
NH
Cl–
7
8
HN
6 Podand-type receptors derived from the bile acids have been reported by
other groups. See for example: R. Boyce, G. Li, H. P. Nestler, T.
Suenaga and W. C. Still, J. Am. Chem. Soc., 1994, 116, 7955; Y. A.
Cheng, T. Suenaga and W. C. Still, J. Am. Chem. Soc., 1996, 118, 1813;
L. J. D’Souza and U. Maitra, J. Org. Chem., 1996, 61, 9494. However,
none of these systems possesses three differentiated a-substituents as
implied for 2.
7 C. Dalgliesh, J. Chem. Soc., 1952, 3940.
8 The analogous five-membered ring has been widely used in carboxylate
and phosphate receptors. For examples, see: E. Fan, S. A. V. Arman, S.
Kincaid and A. D. Hamilton, J. Am. Chem. Soc., 1993, 115, 369; M. S.
Muche and M. W. Göbel, Angew. Chem., Int. Ed. Engl., 1996, 35, 2126;
A. Metzger, V. M. Lynch and E. V. Anslyn, Angew. Chem., Int. Ed.
Engl., 1997, 36, 862. The six-membered 3 was chosen for the present
work because of its greater stability and lipophilicity, and because it was
expected to hold a substrate closer to the a-face of the steroid.
9 For a two-step synthesis of 6 from methyl cholate, see: A. P. Davis, S.
Dresen and L. J. Lawless, Tetrahedron Lett., 1997, 38, 4305.
10 F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, C. Caufield,
G. Chang, T. Hendrickson and W. C. Still, J. Comput. Chem., 1990, 11,
440.
Fig. 1 Structure of 5 +
L-8 (black) derived from computer-based molecular
modelling. Intermolecular hydrogen bonds are shown as broken lines.
Communication 8/08245F
10
Chem. Commun., 1999, 9–10