Brief Articles
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 2 491
In the case of shorter or longer side chains, such as
4-guanidinomethylproline or 4-(3′-guanidino-n-propyl)-
proline, their molecular volumes did not overlap well
with that of arginine (data not shown).
The size of the intersection region of all molecular
volumes (295 Å3) was approximately one-third of that
of arginine (969 Å3). For arginine, 36 conformations
were included inside the intersection region while 97
conformations were searched. At least one of those
conformations would bind to the receptor. The intersec-
tion region contains mainly the equatorial configuration
of the guanidinoethyl groups in 4S- and 4R-GEPro (1,
2) and arginine with an extended side chain. Since those
conformations were calculated to be relatively stable,
the guanidino group in each peptide (18-20) takes the
same arrangement easily.
Con clu sion
This report demonstrates the use of the arginine-
proline fused amino acids to determine the conforma-
tional space where the guanidino group interacts with
the receptor. The backbones of the arginine analogues
fit the turn because of the bent backbone of proline. The
systematic conformational search showed that the side
chains of 4R- and 4S-GEPro are restricted compared to
that of arginine, and the conformational spaces acces-
sible to the guanidino group differed depending on the
chirality of the C4 atom as shown in Figure 2. Since the
biological activity of the analogues (19, 20) was as potent
as that of miniANP (18), it can be concluded that the
receptor-bound conformation of the guanidinoethyl group
would be one of the conformations included in the
intersection region of each conformationally accessible
space.
F igu r e 2. Conformations accepted by the systematic confor-
mational search and molecular volumes of the guanidino
groups of Arg (A and E), 4S-GEPro (B and F), and 4R-GEPro
(C and G) and the intersection region of each volume (D and
H). Parts D and H contain all conformations. (A-D) Front view
of the turn. (E-H) Top view. The red, blue, green, and magenta
areas represent Arg, 4S-GEPro, 4R-GEPro, and the intersec-
tion region, respectively. The conformations are aligned using
backbone atoms. The side chains of Met and Asp are omitted
for clarity.
Analogous proline analogues can also be designed and
synthesized readily. Therefore, the method we proposed
here can be used widely and may shed light on the
structural analysis of ligands containing a turn.
Exp er im en ta l Section
should possess almost the same conformational proper-
ties as for the backbone.
NMR Mea su r em en ts a n d Str u ctu r e Ca lcu la tion s. 2D
DQF-COSY, TOCSY, and NOESY experiments were carried
out on 2.0 mM samples dissolved in 90% H2O/10% D2O using
a Bruker DMX 750 spectrometer. Proton signals were assigned
by the conventional assignment strategy. NOE volumes were
converted into upper-bound distance restraints classified as
strong, medium, weak, or very weak, corresponding to 2.7, 3.7,
5.0, or 6.0 Å, respectively. The lower bound distance restraints
System a tic Con for m a tion a l Sea r ch . The NMR
results above suggest that peptides 18-20 share a
similar backbone structure. Since the analogues 19 and
20 have similar biological activity as well, it seems likely
that their guanidino groups interact with the ANP
receptor in a similar conformational space. The confor-
mational space accessible to the guanidino groups of 4S-
and 4R-GEPro (1, 2) and arginine were mapped out
systematically and compared. It should be noted that
the receptor-bound conformation of residues 6-9 could
be different from the typical type I â-turn used in the
conformational search. However, our previous study
indicated that miniANP (18) has a turnlike conforma-
tion on binding12 and proline has a strong propensity
to form a â-turn as shown by the solution structures.
Therefore, the difference, if any, would be only to a
slight degree.
3
were set to 1.8 Å. J HN,Ha coupling constants were measured
from a 1D spectrum and the DQF-COSY spectrum and were
converted into dihedral angle restraints. An amide proton
temperature coefficient (dsHN/dT), whose value greater than
-4.6 ppb/K is a good indicator of a hydrogen bond, was
analyzed by measuring 1D NMR spectra at 15, 25, 35, 45, and
55 °C. Structure calculations were carried out using torsion-
angle molecular dynamics with the program XPLOR-NIH. A
total of 100 structures were generated, and the 10 lowest
energy structures were submitted to subsequent analysis.
System a tic Con for m a tion a l Sea r ch . Conformational
spaces accessible to the guanidinoethyl group at a turn were
searched for systematically with the program Search Compare
(Accelrys, San Diego, CA). To simplify the search, only a type
I â-turn (Gly-Pro-Met-Asp) was used. First, (4S)- and (4R)-
methylproline were constructed and the proline ring conforma-
tions were searched, during which the N1-C2 bond was
cleaved, allowing other bonds to rotate freely. While C3-C4
and C4-C5 bonds were rotated 360° every 10°, C2-C3 and C5-
N1 bonds were automatically rotated by the algorithm. The
Figure 2 shows the conformations accepted by the
search together with molecular volumes. Since the
potential energy of the conformations in the binding site
could not be estimated, all conformations were used in
the analysis. The volume of arginine can encompass
almost the total volumes of 4S- and 4R-GEPro (1, 2).