Bowl-shaped C(3)-symmetric receptor with concave phosphine oxide with a remarkable selectivity for asparagine derivatives.

Article abstract of DOI:10.1021/ol034168b

A bowl-shaped C(3)-symmetric receptor (1a) that has a phosphine oxide functionality in the interior of a "molecular bowl" shows remarkable selectivity for Asn derivatives. [structure: see text]

Full text of DOI:10.1021/ol034168b

ORGANIC
LETTERS
2003
Vol. 5, No. 9
1431-1433
Bowl-Shaped C -Symmetric Receptor
3
with Concave Phosphine Oxide with a
Remarkable Selectivity for Asparagine
Derivatives
Kwan Hee Lee,† Dong Hoon Lee,† Sungu Hwang,† One Sun Lee,†
,†,‡
Doo Soo Chung,† and Jong-In Hong*
School of Chemistry, College of Natural Sciences, Seoul National UniVersity,
Seoul 151-747, Korea, and Center for Molecular Design and Synthesis, KAIST,
Daejon 305-701, Korea
jihong@plaza.snu.ac.kr
Received January 30, 2003
ABSTRACT
A bowl-shaped C -symmetric receptor (1a) that has a phophine oxide functionality in the interior of a “molecular bowl” shows remarkable
3
selectivity for Asn derivatives.
Construction of host molecules possessing a rigidly defined
cavity with a concave functionality is of great interest.1,2
Incorporation of an inwardly pointing functionality into the
cavity of a bowl-shaped receptor3 is reminiscent of the active
site of enzymes. If a functional group is embedded in an
appropriately sized cavity-shaped molecular host with a rigid
framework, the cavity will function as a reaction site or
binding site with unique properties.
A previous C3-symmetric receptor synthesized in our group
had a hydrophobic binding cavity with a potential preference
for binding lipophilic residues.4 We expected that introduc-
tion of a H-bonding polar functionality within the binding
cavity would alter the binding affinity and selectivity toward
H-bonding guests compared to hosts without a polar group
within the cavity.5 To introduce a concave functionality into
(3) (a) Kemp, D. S.; McNamara, P. E. J. Org. Chem. 1985, 50, 5834.
(b) Hong, J.-I.; Namgoong, S. K.; Bernardi, A.; Still, W. C. J. Am. Chem.
Soc. 1991, 113, 5111. (c) Yoon, S. S.; Still, W. C. J. Am. Chem. Soc. 1993,
115, 832. (d) Still, W. C. Acc. Chem. Res. 1996, 29, 155-163. (e) Ballester,
P.; Shivanyuk, A.; Far, A. R.; Rebek, J., Jr. J. Am. Chem. Soc. 2002, 124,
14014-14016. (f) Grynszpan, F.; Aleksiuk, O.; Biali, S. E. J. Chem. Soc.,
Chem. Commun. 1993, 13. (g) Shevchenko, I.; Zhang, H.; Lattman, M.
Inorg. Chem. 1995, 34, 5405.
† Seoul National University.
‡ Center for Molecular Design and Synthesis.
(1) (a) Goto, K.; Tokitoh, N.; Okazaki, R. Angew. Chem., Int. Ed. Engl.
1995, 34, 1124. (b) Goto, K.; Tokitoh, N.; Okazaki, R. Angew. Chem., Int.
Ed. Engl. 1995, 34, 1124. (c) Lu¨ning, U. Top. Curr. Chem. 1995, 175, 57.
(2) (a) Chavez, F. A.; Que, L., Jr.; Tolman, W. B. Chem. Commun. 2001,
111. (b) Kennan, A. J.; Whitlock, H. W. J. Am. Chem. Soc. 1996, 118,
3027. (c) Sheridan, R. E.; Whitlock, H. W. J. Am. Chem. Soc. 1986, 108,
7120.
(4) Kim, T. W.; Hong, J.-I. Bull. Korean Chem. Soc. 1995, 16, 781.
(5) Carrasco, M. R.; Still, W. C. Chem. Biol. 1995, 2, 205.
10.1021/ol034168b CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/02/2003

the cavity, we designed a C3-symmetric receptor (1a,b) with
a phosphine oxide moiety at the bottom of the bowl. A CPK
model of the designed receptor indicates that PdO is directed
either inside or outside the cavity. Composed of a binding
cavity with a H-bonding functionality, and H-bond donor
and acceptor functionalities on the periphery of the sur-
rounding wall of the bowl-shaped host, 1a (or 1b) is expected
to show enantioselectivity and residue selectivity in the
binding of amino acid derivatives and small peptides. Herein
we report a remarkable residue selectivity for Asn alkyl
amide against Gln, Glu, and Asp derivatives with quite
similar H-bond donor/acceptor geometries in polar media.
major product 1a has the PdO moiety within the cavity.
Furthermore, 1a showed a less polar character on the TLC
(Rf ) 0.4, CH2Cl2/MeOH, 10:1, v/v, compared to Rf ) 0.3
of 1b), indicating that the PdO inside the cavity of 1a is
more shielded from the solvent environment. The lowest-
energy structure of the macrotricyclic compound also reveals
that PdO is pointing into the cavity.8
NMR titration experiments of 1a (or 1b) with various
N-dodecylamide amino acid derivatives were performed in
CDCl3/CD3OD (10:1, v/v) solution (Table 1). With Boc-
Table 1. Binding Constants of 1a and 1b with Various
Ammonium Guestsa
Scheme 1
guestb
H
Ka (M-1
)
esc
1
2
3
4
5
6
7
8
9
D,L-Val-NHRd
D,L-Phe-NHR
D,L-Ser-NHR
N-Boc-D-Val-NHR
D,L-Thr-NHR
D,L-Asn-NHR
D,L-Asn-(â-NHMe)-NHR 1a 3100 (D), 2000 (L)
D,L-Asn-NHR
D,L-Asp-NHR
1a 400 (D), 80 (L)
1a 1000 (D), 170 (L)
1a 1800 (D), 1500 (L)
1a nce
1a 2250 (D), 1050 (L)
1a 45000 (D), 12000 (L) 79:21
83:17
85:15
55:45
68:32
61:39
55:45
76:24
60:40
88:12
1b 2050 (D), 1650 (L)
1a 5000 (D), 1600 (L)
1a 3000 (D), 2000 (L)
1a 5200 (D), 700 (L)
10 D,L-Gln-NHR
11 D,L-Glu-NHR
The syntheses of 1a and 1b began with trialkylation of tris-
(chloromethyl)phosphine oxide6 with dimethyl 5-mercapto-
isophthalate. An intermolecular macrolactamization between
hexakis(pentafluorophenyl ester) and (1R,2R)-diamino-
cyclohexane provided 1a and 1b in 47 and 5% yields,
respectively.
a Measured by 1H NMR titration in CDCl3/CD3OD (10:1, v/v) at 25 °C.
b Guests were used as their trifluoroacetate salts. c Es (enantioselectivity)
) Ka(D)/Ka(L). d R ) docecyl. e No complexation detected.
protected, N-alkylamide amino acid derivatives, no com-
plexation was detected (entry 4), which means that the
ammonium group is crucial for the binding through the
charged H-bonding interaction. N-Alkylamide amino acid
derivatives with a lipophilic side chain show weaker binding
affinity compared to those with a hydrophilic side chain
(entries 1 and 2 vs 3 and 5-11). In particular, Asp, Asn,
Glu, and Gln derivatives with stronger side chain H-bond
donors exhibited better binding affinity compared to those
with a weaker H-bond donor in the side chain (entries 6-11).
It turns out that the D-isomer always binds preferentially.
Compound 1b shows more than a 20-fold decrease in binding
affinity to D-Asn-NHR, which clearly indicates that out-
wardly directed PdO of 1b cannot be involved in H-bond
interaction with the guest. Job analysis for the complex
between 1a and D-Asn-NHR confirmed a 1:1 stoichiometry.
31P resonance of 1a moves downfield upon addition of Asn-
NHR. The saturation binding curve from the 31P NMR
titration of 1a with Asn-NHR implies that PdO in the cavity
of 1a interacts with one of the H-bond donors of the guest.
The highest affinity was found for Asn derivatives (entry
6). In comparison, Gln derivatives with the same H-bond
donor and acceptor geometry except for an additional
methylene on the side chain showed far reduced binding
affinity to 1a (entry 10). Surprisingly, Asp derivatives with
a stronger H-bond donor (carboxylic acid group) in place of
the amide group of Asn showed a dramatic decrease in
binding constants (entry 9). This is remarkable in that subtle
31P NMR spectroscopy was used in order to show the
direction of phosphine oxide at the bottom of receptors.
Compounds 1a and 1b were shown to have PdO inside the
cavity and outside the cavity, respectively, which was
elucidated through the changes of the complexation-induced
+
31P chemical shift between 1a (or 1b) and NH4 , Ph2SnCl2,
+
+
and tBuNH3 . Ph2SnCl2 and tBuNH3 are expected to act
as bulky Lewis acids7 and thus form 1:1 complex with
phosphine oxide outside the cavity. Addition of excess
tBuNH3+ to a ∼10:1 mixture of 1a and 1b resulted in a large
downfield shift (∆δ ) 5.5 ppm) of 31P resonance for 1b
and a small downfield shift (∆δ ) 0.5 ppm) for 1a. Addition
of excess Ph2SnCl2 to 1a in 10:1 (v/v) CDCl3/CD3OD caused
a small downfield shift of the 31P NMR signal. Successive
+
addition of an NH4 solution to the mixture of 1a and
Ph2SnCl2 gave rise to a large downfield shift (∆δ ) 4.2 ppm)
of the 31P NMR signal. Therefore, the result implies that the
(6) Hoffmann, A. J. Am. Chem. Soc. 1921, 43, 1684, 2995.
(7) (a) Whitlock, B. J.; Whitlock, H. W. J. Am. Chem. Soc. 1990, 112,
3910-3915. (b) Mullins, P. Can. J. Chem. 1971, 49, 2719.
(8) (a) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440. (b) Missing force field parameters for a phosphine
oxide group in the original MacroModel AMBER* parameter set were
generated by performing quantum mechanical calculations (ab initio HF/
6-31G**) for model molecules (Supporting Information). (c) Conformational
search for the complex between 1a and D- or L-Asn-NHPr was performed
by keeping a constrained H-bond between â-amide protons of the Asn guest
and PdO during the simulation (Supporting Information).
1432
Org. Lett., Vol. 5, No. 9, 2003

changes in H-bond functionalities and lengths of the side
chain of the guest caused dramatic effects on the binding
affinity. We were particularly interested in which part of the
-37.4 cal mol-1 degree-1; D-Asn-NHR, ∆H° ) -9.6 kcal/
mol and ∆S° ) -19.7 cal mol-1 degree-1). The fact that
binding to D-Asn-NHR is enthalpically less favorable and
entropically more favorable compared to L-Asn-NHR sug-
gests that the structural changes in 1a and D-Asn-NHR are
less pronounced in the process of complexation in compari-
son with those in 1a and L-Asn-NHR.
We need to answer why Asp derivatives with a stronger
H-bond donor in the side chain show much weaker binding
affinity compared to Asn guests. The energy-minimized
structures of D-Asn-NHPr+CF3COO- and D-Asp-NHPr+CF3-
COO- revealed exactly the same H-bonding patterns and
H-bond donor/acceptor geometries (Figure 2).8 The only
1
Asn guest points into the cavity to interact with PdO. H
NMR titration of the Asn guest with 1a in 10% CD3OH in
CDCl3 revealed that the primary amide proton resonances
on the side chain of Asn moved upfield (L-Asn, ∆δ of the
syn amide proton ) -0.73 ppm and ∆δ of the anti amide
proton ) -0.51 ppm; D-Asn, ∆δ of the syn amide proton
) -0.47 ppm and ∆δ of the anti amide proton ) -0.43
ppm, upon addition of 3 equiv of 1a to the guest), which
suggests that the primary amide group is located near the
shielding faces of three aromatics (Figure 1).
Figure 2. Lowest-energy structures for D-Asp-NHPr+CF3COO-
(left) and D-Asn-NHPr+CF3COO- (right). H-bonds are shown as
dotted lines. Dotted circle in yellow indicates the â-anti amide
proton participating in the H-bond with PdO of 1a.
Figure 1. 1H NMR spectra of (a) 1a, (b) complex upon addition
of 3 equiv of 1a to L-Asn-NHR, and (c) L-Asn-NHR in 10% CD3OH
in CDCl3 (Ha, R-amide proton; Hb, â-syn amide proton; Hc, â-anti
amide proton; Hd, chiral methine proton). R ) dodecyl.
difference comes from the side chain geometry. As shown
in the energy-minimized structures of the complexes, the anti
amide proton in the side chain of Asn participates in a
H-bond with PdO (Supporting Information). However, in
the case of Asp derivatives, a â carboxylic acid proton should
be syn to the carboxyl oxygen and therefore hardly involved
in the H-bond with PdO. We believe this is the main reason
for the reduced affinity of the Asp guest to 1a.
In conclusion, we have developed a bowl-shaped C3-
symmetric receptor (1a) that has a phophine oxide function-
ality in the interior of a molecular bowl. Remarkable
selectivity for Asn derivatives turns out to be due to the
participation of PdO in cooperative H-bonding with the
guest inside the cavity of 1a and subtle differences in the
intermolecular H-bond mode within the cavity.
Since the primary amide proton resonances experience an
upfield shift, we believe that it participates in an H-bond
with PdO inside the cavity. Indeed, Asn-(â-NHMe)-NHR
shows much weaker affinity, indicating the absence of an
H-bonding interaction of the â-amide proton of the guest
with PdO of 1a (entry 7). Other amide protons of 1a and
the Asn guest display downfield shifts, indicating inter-
molecular H-bonding interaction outside the cavity upon
complexation. Additional support comes from intermolecular
NOE experiments that indicate contacts between the primary
amide protons of Asn and the three aromatic protons of 1a.
The sensitivity of binding to steric and electronic effects of
the side chain of the guest is compatible with a complex in
which H-bond donors in the side chain are buried within
the binding cavity to interact with PdO.
Although we do not know the exact structural origin of
the chiral discrimination between 1a/D-Asn-NHR and 1a/L-
Asn-NHR, van’t Hoff plots between 1a and D- or L-Asn-
NHR in CD3OD/CDCl3 (1:4, v/v) provided thermodynamic
parameters controlling the enantioselective complexation
process: the complexation process is driven by favorable
enthalpy changes with a negative contribution by entropy
changes (L-Asn-NHR, ∆H° ) -14.8 kcal/mol and ∆S° )
Acknowledgment. Support of this work by the KOSEF
(Grant R02-2002-000-00137-0) is gratefully acknowledged.
Supporting Information Available: Synthetic, spectral,
NMR titration, Job plot, van’t Hoff plot, and computational
modeling details. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL034168B
Org. Lett., Vol. 5, No. 9, 2003
1433