Table 1 Enantioselective discrimination for receptor 2 with different
guests in CDCl at 20 ЊC.
3
a
Guest
Krel
Ethoxycarbonyl--proline
57.0
16.0
15.0
8.4
7.6
4.0
Cbz--phenylglycine
Cbz--phenylalanine
Ethoxycarbonyl--alanine
Ethoxycarbonyl--leucine
Boc--leucine
a
Krel refers to the relative association constants between the guest and
the two enantiomers of receptor 2.
resulted in two different bands (R = 0.72, R = 0.23) for each of
f
f
Fig. 1 X-ray crystal structure of receptor 1 showing intermolecular
hydrogen bonds, which are indicated by dashed lines: N(4) ؒ ؒ ؒ O(m) =
host 2 enantiomers. From these bands, receptors were obtained
as complexes. The free hosts were released by washing an ethyl
acetate solution with 4% aqueous sodium carbonate.
2
.96 Å; O(6) ؒ ؒ ؒ O(m) = 2.67 Å. One of the methanol molecules has
been omitted for clarity.
Since association constants for receptor 2 were too large for
an accurate direct measurement, we used a competitive
7
method. A relative association constant of 5.8 was established
4
4
Ϫ1
against a known receptor with Kass = 1.4 × 10 M when
ethoxycarbonyl--leucine was the guest. Therefore the weak
receptor 2 complex should show K = 8.4 × 10 M while Kass
6.4 × 10 M can be calculated for the strong one with the
4
Ϫ1
ass
5
Ϫ1
=
previous guest.
NOE experiments on the strong complex were specially
revealing since they supported a possible geometry for this
associate. Crossed effects between host and guest place the
phenylenediamine Ho proton close to the α carbon CH in
the guest (3.5%) and the H aromatic proton in the host in the
7
proximity of the methyl of the leucine isobutyl group (2.5%)
(
Fig. 2). From these data, we propose the configuration shown
in Fig. 2 for the strong complex (host 2–-leucine derivative) as
6S, 5aR, 11aS/). The proximity between the leucine CH α
(
carbon and the phenylenediamine group in the host provides
an explanation for the chiral recognition, since in the weak
complex the leucine α proton exchanges position with the
amino acid side chain, and the bulky isobutyl group then
collides with the host phenylenediamine unit.
Titration of guests revealed the importance of the amino
acid side chain and the carbamoyl substituent. Benzyloxy-
carbonyl derivatives provide the best substrates, with chiral
recognitions of up to 15 (with Cbz-phenylglycine), while steric
hindrance from the tert-butyl group probably yields Boc deriv-
atives with small association constants and poor enantio-
selectivities. The combination of the ethoxycarbonyl group
with the rigidity of the proline afforded specially good
discrimination up to 57.
Fig. 2 Proposed structure for the complex formed by receptor 2
(
6S, 5aR, 11aS) and ethoxycarbonyl--leucine.
receptor nor chiral discrimination; the presence alone of a
chiral environment does not seem to be sufficient to achieve
discrimination between the guest enantiomers; further inter-
actions between the amino acid α carbon substituents and the
receptor are necessary to secure reasonable stability differences
between both diastereomeric complexes. CPK models show
that, in receptor 2, a new H-bond between the phenylene-
diamine NH in the receptor and the carbonyl group of a
carbamoyl amino acid derivative could fix the free rotation of
the guest, increasing the difference in association constants of
both enantiomeric guests (Fig. 2). Several groups were tested to
activate the phenylenediamine NH (acetate, mesylate and tri-
flate), triflamide showing the best results, probably due to the
high acidity of its NH.
We thank the “Dirección General de Investigación Científica
y Técnica” (DGICYT Grant PB 98-0275) for its support of this
work. The MEC is acknowledged for three fellowships (A. I. O.,
L. S., F. M. M.).
Notes and references
Transformation of receptor 1 into receptor 2§ (Scheme 1) can
be accomplished in high yield by treating 1 with the lithium salt
of m-phenylenediamine in THF, followed by reaction with tri-
† Electronic supplementary information (ESI) available: binding data.
See http://www.rsc.org/suppdata/p2/b2/b203054c/
ϩ
‡
Selected physical data for receptor 1: FABMS: 57 (M ), 100%; 603,
1
7
60%; 385, 45%; 129, 45%; 483, 40%; 605, 30%; 91, 20%. H NMR
fluoromethanesulfonic anhydride. Competitive titrations were
(
(
(
400 MHz, CDCl ): 8.47 (d, J = 2 Hz, 1H), 7.72 (d, J = 8 Hz, 1H), 7.33
d, J = 2 Hz, 1H), 7.26 (t, J = 8 Hz, 1H), 7.18 (t, J = 8 Hz, 1H), 7.10
d, J = 2 Hz, 1H), 7.03 (d, J = 8 Hz, 1H), 6.79 (d, J = 2 Hz, 1H), 5.33 (d,
3
carried out by adding small amounts of the enantiomeric pure
guests (amino acid derivatives, Table 1) to a deuterochloroform
1
solution of the racemic host, yielding splitting of the H-NMR
J = 9 Hz, 1H), 4.28–4.20 (m, 2H), 3.05–2.65 (m, 4H), 2.60 (dd, J = 3 Hz,
J = 18 Hz, 1H), 1.45 (s, 9H), 1.34 (s, 9H), 1.20 (t, J = 7 Hz, 3H). mp 136–
138 ЊC.
host 2 signals. Graphic representation of the chemical shifts of
these protons against each other, and the use of a home made
curve fitting program, provide the chiral discrimination, and
disclosed significant enantioselectivity for receptor 2.
Crystal data for receptor 1: C H ClN O × 2CH OH, M = 666.25,
3
5
39
2
5
3
monoclinic, space group P2 /n (nЊ 14), a = 16.012(1) Å, b = 12.401(1) Å,
1
3
c = 19.494(1) Å, α = λ = 90Њ, β = 111.98(1)Њ, V = 3589.5(4) Å , Z = 4, D
These promising results suggested that it might be possible to
resolve a racemic mixture of host 2, using its supramolecular
c
Ϫ3
Ϫ1
=
1.235 Mg m , µ(Cu-Kα) = 1.345 mm , F(000) = 1424. Data (6113
3
total reflections and 2091 observed reflections [I > 2 σ(I )]) were meas-
ured on a Seifert 3003 SC rotating anode diffractometer with (Cu-Kα)
radiation (graphite monochromator) using 2θ–ω scans at 293 K.
The crystallographic data for the structure reported in this paper
are deposited at the Cambridge Crystallographic Data Centre as
properties. Ethoxycarbonyl--leucine was selected as the guest
due to its lack of UV absorption at 250 nm. Preparative TLCs
were impregnated with a 2% amino acid solution in chloroform,
2
dried and loaded with host 2 (50 mg); elution with CH Cl2
J. Chem. Soc., Perkin Trans. 2, 2002, 1050–1052
1051