Oxidative induction of β-turn conformations in cyclic peptidomimetics: Conformational analyses as indicators of configuration

Article abstract of DOI:10.1021/ja0171411

Solid-phase syntheses of the cyclic peptidomimetics 1-4 were performed. Sulfide 1 and sulfoxide 3 did not tend to exist in β-turn conformations in solution, but the sulfone 2 and the epimeric sulfoxide 4 did. The assignment of configuration of the sulfoxi

Full text of DOI:10.1021/ja0171411

Published on Web 07/11/2002
Oxidative Induction of â-Turn Conformations in Cyclic Peptidomimetics:
Conformational Analyses As Indicators of Configuration
Luyong Jiang and Kevin Burgess*
Department of Chemistry, Texas A&M UniVersity, PO Box 30012, College Station, Texas 77842-3012
Received September 22, 2001
Table 1. Key NMR Parameters for Compounds 1a-4a
Oxidation of methionine to methionine oxide can perturb peptide
conformations. For instance, this type of transformation may
establish an intramolecular hydrogen bond to a main-chain amide-
NH that disrupts helicity,1 or decrease hydrophobicities in am-
phiphilic helices causing them to revert to â-sheet conformations.2
Reported here is a case where oxidation of a sulfide causes a cyclic
peptidomimetic to adopt a â-turn conformation. Moreover, it
emerged that occurrence of the â-turn conformation is so closely
correlated to only one of the two sulfoxide epimers that confor-
mational studies can be used to predict absolute configurations at
sulfur.
compound
parameter
1a
fast
6.37
medium
3a
2a
4a
ideal type-I
NHi+2 H/D exchangeb
NHi +3Tc (-ppb/K)a
NHi +3 H/D exchangeb
3J GluR,NH (Hz)a
fast
1.26
slow
5.0
7.5
W
slow
0.60
slow
4.5
8.5
M
slow
1.70
slow
5.5
9.0
M
slow
<3.0
slow
4.0
9.0
M
8.5
8.0
M
M
S
3J LysR,NH (Hz)a
NH(Glu)-NH(Lys)a,c
NH(Lys)-NH(hCys)a,c
NH(Glu)-H(Aryl)a,c
M
S
M
W
M
none
M
n/a
a In DMSO-d6. b In CD3OD; rates are classified as “fast”, “medium”, or
“slow” as compared with other NH in the same molecule. c Classified as
relatively “weak”, “medium”, or “strong” cross-peaks.
located in quenched molecular dynamics studies (QMD, Supporting
Information) is consistent with the NMR data (Table 1), but it does
not closely resemble any identifiable secondary structure. It is
difficult to identify the reasons for this difference, but they are
probably related to the larger van der Waals radius of sulfur
compared with those of nitrogen and oxygen, and the fact that S-C
bonds are longer than N-C and O-C.6
The failure of 1a to rest in â-turn conformations motivated us
to explore oxidized derivatives of this thioether. Initially, the
corresponding sulfoxides (3a and 4a) were avoided because of the
difficulties that would be encountered determining the relative
configuration of the stereocenter at sulfur. Consequently, the sulfone
2a was prepared next.
NMR, CD, and QMD analyses of compound 2a led to a
surprising and welcome conclusion: this compound has a preference
for type-I â-turn conformational states. The temperature coefficient
for the NHi+3 was low, indicative of H-bonding or solvent shielding7
or both and that same NH exchanged slowly with CD3OD; the CRH
to NH coupling constants were very near to the ideal values
expected for type-I turns, and the close contacts observed from
ROESY studies were also consistent. A representative low-energy
conformation simulated from QMD studies (i.e., without applying
spectroscopic constraints) is shown in Figure 1.
Why should the sulfone 2a have a preference for a â-turn
conformation if the thioether 1a does not? A clue for the origin of
the conformational switch to â-turn conformations upon oxidation
was found in the 1D NMR studies. Those experiments revealed
that, unexpectedly, there were two NH protons that were relatively
slow to exchange when the molecule was placed in CD3OD. It was
the NHi+2 residue that had these unexpected spectroscopic char-
acteristics, and this led us to propose that the â-turn in compound
2a was stabilized by an NHi+2 to OS hydrogen bond. Consistent
with this, one of the NHi+2 to OSO distances in a representative
low-energy conformer was only 2.38 Å (Figure 1).
This project arose from a modest set of objectives. Previous work
had shown that cyclic amine and ether analogues of 1 (where the
sulfur atom is substituted with NH or O) tend to adopt â-turn
conformations in DMSO.3 These conformations were desired for
preparations of libraries of molecules that might mimic or disrupt
certain protein-protein interactions.4 To expand the series, we
decided to make the cyclic thioether 1a and study its preferred
conformation in solution.
Thioether 1a was prepared via methods analogous to those
previously reported,5 and its conformation in solution was studied
by a combination of experiments as already outlined for the cyclic
amine and ether analogues.4 The syntheses gave good purities and
yields of the desired products. However, unlike its NH- and
O-analogues, the preferred conformations of thioether 1a do not
include â-turn structures. The minimum energy conformation
Only one of the diastereotopic sulfone oxygens participates in
the unusual transannular H-bond postulated for compound 2a.
* To whom correspondence should be addressed. Email: burgess@tamu.edu.
9
9028
J. AM. CHEM. SOC. 2002, 124, 9028-9029
10.1021/ja0171411 CCC: $22.00 © 2002 American Chemical Society

C O M M U N I C A T I O N S
Figure 2. CD spectra for compounds 1a-4a.
The conformational biases simulated by QMD are consistent with
â-turn conformations for the (S)-sulfoxides, and not the (R)-isomers.
These observations do not provide incontrovertible assignments of
the absolute configurations at the sulfoxide S-atoms; this would
almost certainly require crystallographic analyses, and we were
unable to form suitable crystals from these peptidomimetics.
Nevertheless, the weight of evidence presented in favor of the
proposed assignments is substantial.
This work demonstrates that unusual transannular H-bonds for
peptidomimetics containing sulfone or sulfoxide functionalities can
profoundly effect their preferred conformational states in solution.
It is also highlights a rare situation in which conformational analyses
can be used to infer stereochemical configurations.
Figure 1. Simulated low-energy conformers for compounds 1a-4a.
Therefore it seemed logical that only one of the corresponding
sulfoxides would display a similar interaction.
Oxidation of the thioether 1a with sodium periodate gave a 1:2
mixture of sulfoxides that were separable by preparative HPLC.
There was no immediate basis for a stereochemical assignment of
the sulfur configuration, but QMD simulations of these two
diastereomers (without any spectroscopic constraints), indicated that
only the (S)-isomer 4a could adopt a type-I â-turn conformation
with an extra transannular H-bond, whereas â-turn conformers did
not feature prominently for the (R)-sulfoxide 3a. Spectroscopic
analyses showed that the major stereoisomer had a preferred â-turn
conformer (Table 1); hence, it appeared likely that this compound
had the (S)-sulfoxide configuration, that is, it was compound 4a.
Consistent with this assignment, the spectroscopic studies gave data
for the minor stereoisomer that matched the preferred conformation
simulated for sulfoxide 3a, whereas the virtual preferred â-turn
conformation for 4a matched the spectroscopic data for the major
isomer nearly perfectly. Moreover, exchange of the NHi+2 proton
in CD3OD was slow for compound 4a relative to that for 3a.
Glu-Lys side chains are polar and capable of H-bonding, and
thus it was important to explore whether the trends outlined for
compounds 1-4 were unique to this particular substitution pattern.
Consequently, a similar approach was used to study the compounds
1b-d to 4b-d. The NMR and CD data accumulated show that
the sulfones 2 and one of the sulfoxide epimers 4, adopt â-turn
conformations, whereas the sulfides 1 and the sulfoxides 3 have
no similar conformational biases. The trends in the NMR data that
are implied in Table 1 are much clearer when the whole data set
for all the compounds 1-4 are plotted together (Table S1,
Supporting Information).
Acknowledgment. Support for this work was provided by
Procter and Gamble, The NIH (CA 82642), and by The Robert A.
Welch Foundation. We thank Dr. S. Srivastava at Bristol Myers
Squibb for supplying us with a sample of bis-N-trifluoroacetyl
homocystine, Ms. M. Pattarawarapan and Mr. S. Reyes for MS
analyses, and Mr. C.-. H. Park for help with the QMD studies.
Supporting Information Available: Details of the syntheses,
spectroscopic data for characterization and conformational studies,
results from the QMD calculations (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
References
(1) Watson, A. A.; Fairlie, D. P.; Craik, D. J. Biochem. 1998, 37, 12700.
(2) Schenck, H. L.; Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1996,
118, 12487. Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1993, 115,
12609.
(3) Feng, Y.; Wang, Z.; Jin, S.; Burgess, K. Chem.sEur. J. 1999, 5, 3273.
(4) Burgess, K. Acc. Chem. Res. 2001, 34, 826-835.
(5) Feng, Y.; Burgess, K. Chem.sEur. J. 1999, 5, 3261.
(6) N 0.74 Å, O 0.66 Å, S 1.04 Å from Inorganic Chemistry, 2nd ed.; Shriver,
D. F.; Atkins, P.; Langford, C. H. Freeeman: New York, 1994.
Circular dichroism (CD) studies support the assertion that unusual
transannular H-bonds stabilize the â-turn conformations in com-
pounds 2a and 4a, and also provide further evidence for their
stereochemical assignments (Figure 2). Compounds 1a, 2a, and 4a
give an ellipticity minimum at approximately 205 nm characteristic
of a type-I â-turn,8 whereas compound 3a does not.
(7) Kopple, K. D.; Ohnishi, M.; Go, A. J. Am. Chem. Soc. 1969, 91, 4264,
Ohnishi, M.; Urry, D. W. Biochem. Biophys. Res. Commun. 1969, 36,
194.
(8) Perczel, A.; Hollosi, M. In Circular Dichroism and the Conformational
Analysis of Biomolecules; Fasman, G. D., Ed.; Plenum Press: New York
and London, 1996; p 362.
JA0171411
9
J. AM. CHEM. SOC. VOL. 124, NO. 31, 2002 9029