β-turn and β-hairpin mimicry with tetrasubstituted alkenes

Article abstract of DOI:10.1021/ja9824526

Synthesis and conformational analysis are reported for molecules containing the trans-5-amino-3,4-dimethylpent-3-enoic acid residue (ADPA, 1). This amino acid is a glycylglycine mimic, in which the central amide group is replaced with an E-tetrasubstituted alkene. It was anticipated that this isosteric replacement would promote specific local (β-turn) and nonlocal (β-hairpin) conformational preferences. Previous work has shown that the most common β-turn conformations (type I and type II) are not strong inducers of β-hairpin formation, while the rare 'mirror image' β-turns (type I' and type II') promote β-hairpin formation. We therefore sought an achiral unit with a strong turn-forming propensity, since the lack of stereogenic centers within such a unit would eliminate the energetic distinction between common and 'mirror image' turn conformations. In, the ADPA unit, avoidance of allylic strain was expected to preorganize the backbone for adoption of folded conformations. A combination of NMR and IR data for di-, tri-, and tetrapeptide analogues containing the ADPA residue reveal that β-turn- and β-hairpin-like folding is promoted in methylene chloride solution.

Full text of DOI:10.1021/ja9824526

1806
J. Am. Chem. Soc. 1999, 121, 1806-1816
â-Turn and â-Hairpin Mimicry with Tetrasubstituted Alkenes
Robb R. Gardner, Gui-Bai Liang, and Samuel H. Gellman*
Contribution from the Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706
ReceiVed July 13, 1998. ReVised Manuscript ReceiVed December 30, 1998
Abstract: Synthesis and conformational analysis are reported for molecules containing the trans-5-amino-
3
,4-dimethylpent-3-enoic acid residue (ADPA, 1). This amino acid is a glycylglycine mimic, in which the
central amide group is replaced with an E-tetrasubstituted alkene. It was anticipated that this isosteric replacement
would promote specific local (â-turn) and nonlocal (â-hairpin) conformational preferences. Previous work has
shown that the most common â-turn conformations (type I and type II) are not strong inducers of â-hairpin
formation, while the rare “mirror image” â-turns (type I′ and type II′) promote â-hairpin formation. We therefore
sought an achiral unit with a strong turn-forming propensity, since the lack of stereogenic centers within such
a unit would eliminate the energetic distinction between common and “mirror image” turn conformations. In,
the ADPA unit, avoidance of allylic strain was expected to preorganize the backbone for adoption of folded
conformations. A combination of NMR and IR data for di-, tri-, and tetrapeptide analogues containing the
ADPA residue reveal that â-turn- and â-hairpin-like folding is promoted in methylene chloride solution.
Introduction
A â-hairpin (two strands connected by a short loop) can be
viewed as the smallest increment of antiparallel â-sheet second-
ary structure. The most common loop contains only two
There is great interest in peptide analogues that have some
or all of the backbone replaced by non-R-amino acid segments.
6
1
residues, and in this case, the two loop residues are also the
Molecular chimerae of this type may display improved phar-
macological properties relative to peptides, e.g., resistance to
peptidase degradation, and passive diffusion across biological
membranes. A second source of interest is the prospect of
diminishing the conformational flexibility that characterizes
short R-amino acid oligomers. Backbone rigidification can
enhance pharmacological behavior because biological activity
usually requires that a peptide or peptide mimic adopt a specific
conformation. Peptide mimetic units that provide conformational
control can also serve as tools in basic biostructural research,
e.g., for construction of model systems that probe the origins
of protein folding preferences.2
two central residues of a â-turn (â-turns are usually defined in
terms of four residues ). Surveys of â-hairpins in crystalline
proteins led Thornton and co-workers to recognize that two rare
7
classes of â-turn, type I′ and type II′, often form two-residue
7
a
hairpin loops. In contast, type I and type II â-turns, by far the
most common among â-turns in crystalline proteins, are seldom
involved in two-residue â-hairpin loops. (â-Turn “types” are
defined by the φ and ψ torsion angles of the central two residues;
the prime (′) indicates that all four torsion angles are mirror
8
images of the torsion angles in the unprimed version. ) This
discovery has prompted several groups to try to stabilize
â-hairpin conformations in short peptides by employing loop
sequences that are predisposed toward type I′ or type II′ â-turn
Here we describe the development of a dipeptide mimic that
promotes â-hairpin-like folding when coupled to R-amino acid
9
conformations. Our approach has involved use of D-residues
3
residues. The dipeptide mimic is centered on a tetrasubstituted
9a-h
to induce loop formation.
We have shown that D-proline at
alkene; this design builds upon the widespread use of alkene
units as peptidase-resistant isosteric replacements for backbone
the first of the two loop positions strongly stabilizes â-hairpin
9a,b
conformations of tetrapeptides in organic solvents. We have
4
amide groups. We extend beyond these precedents by taking
subsequently demonstrated that D-proline exerts a similar effect
advantage of allylic strain (actually, the avoidance of allylic
strain) to induce U-shaped conformations,5 which in turn
promote â-hairpin-like folding. Wipf et al. have recently reported
related studies.4
9c-e
in 12-16 residue peptides in aqueous solution.
(4) (a) Hann, M. M.; Sammes, P. G.; Kennewell, P. D.; Taylor, J. B. J.
Chem. Soc., Chem. Commun. 1980, 234. (b) Cox, M. T.; Heaton, D. W.;
Horbury, J. J. Chem. Soc., Chem. Commun. 1980, 799. (c) Cox, M. T.;
Gormley, J. J.; Hayward, C. F.; Petter, N. N. J. Chem. Soc., Chem. Commun.
1980, 800. (d) Wipf, P.; Fritch, P. C. J. Org. Chem. 1994, 59, 4875. (e)
Wipf, P.; Henninger, T. C.; Geib, S. J. J. Org. Chem. 1998, 63, 6088.
(5) General discussions of allylic strain: Johnson, F. Chem. ReV. 1968,
68, 375. Berg, U.; Sandstr o¨ m, J. AdV. Phys. Org. Chem. 1989, 25, 1.
Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1992, 31, 1124. For the
use of acyclic stereochemical control to preorganize the backbone of a
saturated alkane hairpin mimic, see: Schopfer, U.; Stahl, M.; Brandl, T.;
Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1997, 36, 1745.
(6) Gellman, S. H. Curr. Opin. Chem. Biol. 1998, 2, 717.
(7) (a) Sibanda, B. L.; Thornton, J. M. Nature 1985, 316, 170. (b) Wilmot,
C. M.; Thornton, J. M. J. Mol. Biol. 1988, 203, 221. (c) Sibanda, B. L.;
Thornton, J. M. J. Mol. Biol. 1993, 229, 428. (d) Mattos, C.; Petsko, G. A.;
Karplus, M. J. Mol. Biol. 1994, 238, 733. (e) Gunasekaran, K.; Ramakrish-
nan, C.; Balaram, P. Protein Eng. 1997, 10, 1131.
e
(1) For leading references, see: (a) Ball, J. B.; Alewood, P. F. J. Mol.
Recog. 1990, 3, 55. (b) Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl.
1
1
993, 32, 1244. (c) Gillespie, P.; Cicariello, J.; Olsen, G. L. Biopolymers
997, 43, 191. (d) Hanessian, S.; McNaughton-Smith, G.; Lombart, H.-G.;
Lubell, W. D. Tetrahedron 1997, 53, 12789.
2) For leading references, see: (a) Kemp, D. S.; Oslick, S. L.; Allen,
(
T. J. J. Am. Chem. Soc. 1996, 118, 4249. (b) Kemp, D. S.; Bowen, B. R.;
Muendel, C. C. J. Org. Chem. 1990, 55, 4650. (c) Brandmeier, V.; Sauer,
W. H. B.; Feigel, M. HelV. Chim. Acta 1994, 77, 70. (d) Nesloney, C. L.;
Kelly, J. W. J. Am. Chem. Soc. 1996, 118, 5836. (e) Nowick, J. S.; Insaf,
S. J. Am. Chem. Soc. 1997, 119, 10903. (f) Schopfer, U.; Stahl, M.; Brandl,
T.; Hoffmann, R. W. Angew. Chem., Int. Ed. Engl. 1997, 36, 1745. (g)
Schneider, J. P.; Kelly, J. W. Chem. ReV. 1995, 95, 2169. (h) Nowick, J.
S.; Smith, E. M.; Pairish, M. Chem. Soc. ReV. 1996, 401.
(3) Preliminary report: Gardner, R. R.; Liang, G.-B.; Gellman, S. H. J.
(8) Rose, G. D.; Gierasch, L. M.; Smith, J. A. AdV. Protein Chem. 1985,
37, 1 and references therein.
Am. Chem. Soc. 1995, 117, 3280.
1
0.1021/ja9824526 CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/19/1999

â-Turn and â-Hairpin Mimicry
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1807
Scheme 1a
3
2
4
a
Key: (a) 2-Lithio-1,3-dithiane (1 equiv). (b) NaN
3
, Bu
4
NBr. (c) Ph
3
P, THF (wet). (d) Di-tert-butyl dicarbonate. (e) MeSO
3
H, MeOH. (f)
Isobutyryl chloride, Et
3
N. (g) Bis(trifluoromethyl)acetoxyiodobenzene, MeOH. (h) 0.5 M HCl, THF/H
2
O. (i) NaClO , 2-methyl-2-butene. (j) DCC,
2
N-hydroxysuccinimide, isopropylamine.
Nonhydrolyzable â-hairpin-promoting dipeptide mimics may
be useful to medicinal chemists because compact â-hairpins,
with two-residue loops, have been shown to play key roles in
protein-protein and protein-RNA recognition. Dipeptide
mimics of this type may allow construction of metabolically
stable antagonists for natural complexation processes. â-Hairpin
inducers can also be used to elucidate the sources of â-sheet
folding preferences in proteins, if these inducers allow one to
stabilize small increments of â-sheet secondary structure in
solution.2
virtue that substantial conformational preorganization is antici-
pated in the absence of asymmetric centers, which simplifies
the synthetic challenge. In addition, the distinction between
“mirror image” turn conformations (e.g., type I vs type I′
â-turns) becomes irrelevant in an achiral dipeptide mimic like
ADPA.
1
0
11
The synthesis of diamide 2, an end-capped derivative of
glycylglycine mimic 1, began with (E)-1,4-dibromo-2,3-di-
1
5a
methyl-2-butene (Scheme 1). Treatment of this dibromide
with 1 equiv of 2-lithio-1,3-dithiane followed by NaN3 gave a
mixture of compounds that included the desired unsymmetrically
functionalized molecule. NMR analysis of this mixture, how-
ever, revealed that the double-bond configuration was scrambled
during the NaN3 reaction. This mixture was subjected to azide-
reducing conditions, followed by tBoc protection of the resulting
amino group. At this point, intermediate 3 and its Z-isomer were
readily separated and purified via chromatography. The tBoc
group was then removed from each alkene, independently, and
the amino group was acylated with isobutyryl chloride. These
dithianes were converted to the corresponding dimethyl acetals,
Results and Discussion
Design and Synthesis of the Dipeptide Mimic. Trans-
disubstituted alkene units have been widely employed as
isosteric replacements for the backbone amide groups of peptides
because alkenes are inert to peptidases and the trans configu-
ration mimics the conformational preference of a secondary
amide.4 This substitution does not, however, diminish the
12
conformational flexibility that is characteristic of peptides. We
chose to deviate from the typical alkene isostere approach by
employing a tetrasubstituted alkene unit because we anticipated
4
and its Z-isomer. The alkene configurations of the isomeric
acetals were established by both NOE data (summarized in the
Experimental Section) and IR data (4 showed no sign of
intramolecular N-H- -OCH3 hydrogen bonding, while the
isomer displayed intramolecular hydrogen bonding, in dilute
CH2Cl2 solution). Conversion of 4 to diamide 2 involved
standard methods.
that avoidance of allylic strain would limit conformational
freedom in a way that favors â-hairpin folding.1
3,14
We selected
trans-5-amino-3,4-dimethylpent-3-enoic acid (ADPA, 1), which
Intermediate 3 was used to prepare a wide variety of other
molecules containing the ADPA unit, including tri- and tet-
rapeptide analogues, e.g., 11. These syntheses involved standard
methods for the most part; details are provided in the Supporting
Information. Two of the molecules required for our studies, 16
and 17 (analogues of 11), had an ester group in place of one
secondary amide group. For 16, synthesis involved use of the
R-hydroxy analogue of L-leucine, which is readily available from
mimics glycylglycine, as our initial target. This target has the
(
9) Use of D-Pro-Xxx segments to promote â-hairpin formation: (a)
Haque, T. S.; Little, J. C.; Gellman, S. H. J. Am. Chem. Soc. 1994, 116,
4
1
1
1
1
105. (b) Haque, T. S.; Little, J. C.; Gellman, S. H. J. Am. Chem. Soc.
996, 118, 6975. (c) Haque, T. S.; Gellman, S. H. J. Am. Chem. Soc. 1997,
19, 2303. (d) Stanger, H. E.; Gellman, S. H. J. Am. Chem. Soc. 1998,
20, 4236. (e) Schenck, H. L.; Gellman, S. H. J. Am. Chem. Soc. 1997,
20, 44869. (f) Awasthi, S. K.; Raghothama, S.; Balaram, P. Biochem.
(12) Replacement of an ethylene unit with an alkene unit leads to little
2
3
Biophys. Res. Commun. 1995, 216, 375. (g) Karle, I. L.; Awasthi, S. K.;
Balaram, P. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 8189. (h) Ragothama,
S. R.; Awasthi, S. K.; Balaram, P. J. Chem. Soc., Perkin Trans. 2 1998,
rigidification because of the low barrier to rotation about sp -sp bonds:
Page, M. I.; Jencks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1678.
Expansion of this concept: Mammen, M.; Shakhnovich, E. I.; Whitesides,
G. M. J. Org. Chem. 1998, 63, 3168.
1
37. Use of the L-Asn-Gly segment to promote â-hairpin formation: (i)
Ram ´ı rez-Alvarado, M.; Blanco, F. J.; Serrano, L. Nature Struct. Biol. 1996,
(13) We have previously shown that tri- and tetrasubstituted trans-alkene
units provide greater backbone preorganization than a disubstituted trans-
alkene unit, as manifested in intramolecular hydrogen bond formation: (a)
Liang, G.-B.; Dado, G. P.; Gellman, S. H. J. Am. Chem. Soc. 1991, 113,
3994. (b) Liang, G.-B.; Desper, J. M.; Gellman, S. H. J. Am. Chem. Soc.
1993, 115, 925.
(14) Computational prediction that a trans-trisubstituted alkene-based
amide isostere will more strongly induce chain folding than a trans-
disubstituted alkene: Deschrijver, P.; Tourw e´ , D. FEBS Lett. 1982, 146,
353. Recent experimental efforts to verify this prediction were unsuccess-
ful: Devadder, S.; Verheyden, P.; Jaspers, H. C. M.; Van Binst, G.; Tourw e´ ,
D. Tetrahedron Lett. 1996, 37, 703.
3, 604. (j) de Alba, E.; Jimenez, M. A.; Rico, M. J. Am. Chem. Soc. 1997,
119, 175. (k) Maynard, A. J.; Searle, M. S. J. Chem. Soc., Chem. Commun.
1997, 1297. (l) Sharman, G. J.; Searle, M. S. J. Chem. Soc., Chem. Commun.
1997, 1955. (m) Griffiths-Jones, S. R.; Maynard, A. J.; Sharman, G. J.;
Searle, M. S. J. Chem. Soc., Chem. Commun. 1998, 789. (n) Maynard, A.
J.; Sharman, G. J.; Searle, M. S. J. Am. Chem. Soc. 1998, 120, 1996. (o)
Sharman, G. J.; Searle, M. S. J. Am. Chem. Soc. 1998, 120, 5291. (p) We
have recently shown that D-Pro-Gly is a stronger hairpin promoter than
L-Asn-Gly (ref 9d).
(
10) Strynadka, N. C. J.; Jensen, S. E.; Alzari, P. M.; James, M. N. G.
Nat. Struct. Biol. 1996, 3, 290.
11) (a) Puglisi, J. D.; Chen, L.; Blanchard, S.; Frankel, A. D. Science
(
(15) (a) Murray, R. W.; Agarwal, S. K. J. Org. Chem. 1985, 50, 4698.
(b) Gordon, B.; Blumenthal, M.; Mera, A. E.; Kumpf, R. J. J. Org. Chem.
1985, 50, 1540.
1
995, 270, 1200. (b) Ye, X.; Kumar, R. A.; Patel, D. J. Chem. Biol. 1995,
2, 827. (c) Varani, G. Acc. Chem. Res. 1997, 30, 189 and references therein.

1808 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Gardner et al.
Scheme 2a
a
t
2
Key: (a) KO Bu, n-BuLi; MeI. (b) Br .
extended conformation very similar to that seen for 2 (Figure
), with all hydrogen bonding intermolecular. This observation
Figure 1. Conformation of 2 in the solid state.
1
is contradictory to our hypothesis but consistent with the results
presented below that show the folding of 2 and 5 to be very
similar in solution.
To evaluate our hypothesis that avoidance of allylic strain
would provide useful conformational preorganization to a
peptide chimera containing dipeptide mimic 1, we required
analogues containing the glycylglycine mimic trans-5-amino-
pent-3-enoic acid (APA, 6). Scheme 3 outlines preparation of
Figure 2. Influence of avoided allylic strain on the position of
substituents R and R relative to the alkene plane.
1 2
L-leucine. Synthesis of 17 was accomplished by treating (E)-
,4-dibromo-2,3-dimethyl-2-butene with 1 equiv of 2-lithio-1,3-
1
dithiane followed by CsOAc (rather than NaN3) and then
proceeding as described in the Supporting Information.
The alkene configuration of 2 was confirmed by X-ray
crystallography (Figure 1). Interestingly, the molecule crystal-
lized in an extended conformation, even though it was designed
to adopt a folded conformation in solution (this folded confor-
mation is highly populated in methylene chloride solution, as
demonstrated below). For 2 and related molecules, avoidance
of allylic strain is expected to keep the carboxyl and amino
functional groups out of the alkene plane, but these groups can
either reside on the same side of the alkene plane (desired;
expected to promote â-hairpin folding) or on the opposite side
diamide 8, which is an analogue of 2. Intermediate 7 was useful
for preparing other derivatives that are discussed below.
Solvent Choice. All molecules were examined in methylene
chloride solution. The di-, tri-, and tetrapeptide analogues we
prepared are not expected to fold extensively in a highly polar
solvent, like water, because there is insufficient noncovalent
driving force for adoption of a compact conformation. Indeed,
conventional peptides comprised of two to four residues do not
generally display significant populations of folded conformers
in aqueous solution. By working in a relatively nonpolar solvent,
we allow intramolecular hydrogen bonds to serve as a driving
force for folding. Hydrogen bond formation also provides a
sensitive method for detecting folding conformations and
identifying the folding pattern under these conditions.
(undesired; precludes â-hairpin folding). These two possibilities
are illustrated in Figure 2. In the solid state, the undesired
arrangement is displayed by 2 (Figure 1) and by four other
molecules derived from 1 (including 17).17 (The observation
that 2 adopts an extended conformation in the solid state shows
how profoundly intermolecular interactions in a crystal lattice
can perturb the intrinsic conformational preferences of flexible
8
a,b,13,19
20
Extensive reports from our lab
and others demon-
strate that formation of one or two amide-amide hydrogen
bonds provides a significant but not overwhelming inducement
for folding in solvents such as methylene chloride and chloro-
form. This situation is ideal for examining the intrinsic folding
propensity of the covalent structure that links the hydrogen-
bonding sites, particularly if the analysis involves comparing
molecules that could form identical numbers and types of
hydrogen bonds (e.g., comparison of dipeptide mimics 2, 5, and
16
molecules. ) Wipf et al. have recently reported crystal structures
of two optically active alanylalanine dipeptide mimics related
to 2 in which the alkene is trisubstituted, and a â-turn-like
conformation is observed in both cases.4e
It seemed possible that greater conformational preorganization
might be achieved by replacing the vinyl methyl groups of
amino acid 1 with larger groups, since avoidance of 1,5-syn-
pentane-like interactions could induce the diagonal substituents
8
). The intrinsic folding propensity is determined by several
factors, including torsional strain and other nonbonded repul-
sions that develop as the backbone folds back upon itself, and
the conformational entropy loss that accompanies folding. Since
solvation forces are expected to exert relatively little influence
on the intrinsic folding propensity, the insights gained from
analysis in organic solvents should apply to aqueous solution
as well. Indeed, we have established such a correlation in our
analysis of the stereochemical requirements for formation of
(
i.e., the carbonyl and nitrogen functional groups) to residue
on the same side of the alkene. This type of conformation has
1
8
been observed for tetrabenzylethylene in the solid state. We
tested this hypothesis by examining diamide 5, the diethyl
analogue of 2. Synthesis of 5 proceeded from (E)-2,3-dibro-
momethyl-3-hexene via the route described above for prepara-
tion of 2 from (E)-1,4-dibromo-2,3-dimethyl-2-butane. (E)-2,3-
Dibromomethyl-3-hexene was prepared by converting 2,3-
9
a-e
1
5b
â-hairpins with two-residue loops.
We initially used tet-
dimethylbutadiene to 2,3-diethylbutadiene,
followed by
1
5a
rapeptides and related compounds, examined in methylene
chloride, to demonstrate that placing D-proline at the first of
the two loop residues strongly promotes â-hairpin folding,
bromination (Scheme 2). The alkene configuration of 5 was
verified crystallographically.17 In the solid state, 5 displays an
9
a,b
(16) For previous examples of the effects of crystal lattices on the
relative to L-proline at the same position. Subsequently, we
conformations of flexible, polar molecules, see: (a) Dado, G. P.; Desper,
J. M.; Holmgren, S. K.; Rito, C. J.; Gellman, S. H. J. Am. Chem. Soc.
(19) (a) Gellman, S. H.; Dado, G. P.; Liang, G.-B.; Adams, B. R. J. Am.
Chem. Soc. 1991, 113, 1164. (b) Liang, G.-B.; Desper, J. M.; Gellman, S.
H. J. Am. Chem. Soc. 1993, 115, 925.
(20) For leading references, see: (a) Tsang, K. Y.; Diaz, H.; Graciani,
N.; Kelly, J. W. J. Am. Chem. Soc. 1994, 116, 3988. (b) Nowick, J. S.;
Abdi, M.; Bellamo, K. A.; Love, J. A.; Martinez, E. J.; Noronha, G.; Smith,
E. M.; Ziller, J. W. J. Am. Chem. Soc. 1995, 117, 89. (c) Gung, B. W.;
Zhu, Z. J. Org. Chem. 1996, 61, 6482.
1
992, 114, 4834. (b) Gellman, S. H.; Powell, D. R.; Desper, J. M.
Tetrahedron Lett. 1992, 33, 1963. (c) Kessler, H.; Zimmermann, G.; F o¨ rster,
H.; Engel, J.; Oepen, G.; Sheldrick, W. S. Angew. Chem., Int. Ed. Engl.
1
981, 20, 1053.
17) Gardner, R. R., Ph.D. Thesis, University of WisconsinsMadison,
997.
18) Andersen, L.; Berg, U.; Pettersson, I. J. Org. Chem. 1985, 50, 493.
(
1
(

â-Turn and â-Hairpin Mimicry
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1809
Scheme 3a
7
8
a
Key: (a) (Ethanol, H
SO
2 4
2 3 2
(cat.). (b) (PhO) P(O)N , tert-butanol, reflux. (c) NaOH, ethanol, H O. (d) DCC, N-hydroxysuccinimide, isopropylamine.
3
(e) 4 N HCl, dioxane. (f) Isobutyryl chloride, Et N.
-
1
proposed to be are solvent-exposed (3436 cm ) and an
-
1
intramolecular N-H- -π hydrogen-bonded form (3410 cm ),
2
2
-1
involving the alkene as acceptor. The ca. 17 cm difference
between non-hydrogen-bonded N-terminal and C-terminal N-H
stretch bands is consistent with prior observations that a
branched alkyl group on nitrogen causes the N-H bands of
-1
secondary amides to move 15-20 cm to lower energy, relative
to N-H bands of secondary amides with linear alkyl append-
2
1
ages.
Our conclusion that 2 experiences extensive 10-membered
ring but no eight-membered ring hydrogen bonding is supported
by the data for reference compound 9, in which only the eight-
membered ring hydrogen bond is possible. This molecule
-
1
displays a single N-H band, at 3440 cm , in the region for
solvent-exposed N-H. The difference between the positions of
this band and the band assigned to the analogous solvent-
2 2
Figure 3. Amide proton NMR chemical shifts for 2 in CD Cl at room
temperature, as a function of the logarithm of concentration. These
data suggest that aggregation occurs above 1 mM.
-
1
exposed N-H in 2 (ca. 13 cm ) may indicate that there is a
different torsional preference about the (Od)CN-CC(dC) bond
2
3
between these two molecules. This difference also indicates
that great caution must be exercised in trying to interpret
showed that analogous effects are manifested with 12-16
-1
residue peptides in aqueous solution.9
e-c
variations of <20 cm in N-H stretch band positions between
different molecules.
Conformational Analysis of Dipeptide Mimics. Study of
hydrogen-bond-driven folding processes in nonpolar solvents
must begin with control experiments to determine the lowest
concentration at which intermolecular hydrogen bonding occurs.
Figure 3 shows how the amide proton chemical shifts of 2 vary
as a function of the logarithm of concentration in CD2Cl2. These
data suggest that intermolecular hydrogen bonding occurs only
above 1 mM. The NMR and IR data discussed below were
obtained at 1 mM and should reflect effects of only intramo-
lecular hydrogen bonding, unless otherwise mentioned.
Comparison of IR data for 2 with the data for 8, the analogue
containing a disubstituted alkene, confirms our prediction that
the tetrasubstituted alkene strongly preorganizes the backbone
for â-turn-like folding. Dipeptide 8 experiences much less
intramolecular hydrogen bonding than does 2. Further, the data
for reference compound 10 indicate that intramolecular hydrogen
bonding in the disubstituted alkene series arises in part or
entirely from eight-membered ring formation. These results
demonstrate that avoidance of allylic strain provides useful
control of conformational preferences.
Comparison of IR data for 2 with the data for 5, the analogue
containing ethyl groups rather than methyl groups on the
tetrasubstituted alkene, contradicts our hypothesis that avoidance
of syn-pentane-like interactions would lead to enhanced con-
formational control. Diamides 2 and 5 display nearly identical
extents of internal hydrogen bonding. It is possible that the
desired level of conformational control could be achieved with
substituents larger than ethyl on the alkene.
Figure 4 shows N-H stretch region IR data for dipeptide
mimics 2, 5, and 8 and for reference compounds 9 and 10, each
1
mM in CH2Cl2 at room temperature. There are at least four
bands for 2 in this region, as shown by the mathematical
1
3,19,21
decomposition of the spectral data. Extensive precedent
indicates that the bands at 3453 and 3436 cm-1 arise from N-H
groups exposed to solvent, the band at 3330 cm arises from
-
1
N-H involved in an intramolecular amide-amide hydrogen
-
1
bond, and the band at 3410 cm arises from N-H involved in
-
1
a weaker hydrogen bond. We assign the band at 3453 cm to
the N-terminal N-H, which appears to be free of intramolecular
hydrogen bonding. The other three bands are assigned to the
C-terminal N-H. The most highly populated state of this N-H
involves a 10-membered ring CdO- -H-N hydrogen bond (the
desired â-turn-like folding pattern); the two less populated states
Conformational Analysis of Tetrapeptide Mimics. A large
propensity to adopt a U-shape does not ensure that a molecular
fragment will serve as a successful â-hairpin promoter.9 We
therefore evaluated the â-hairpin-promoting properties of the
ADPA and APA glycylglycine mimics by placing them at the
center of tetrapeptide analogues. Tetrapeptides can fold to
a-d
(
21) (a) Boussard, G.; Marraud, M. Biopolymers 1979, 18, 1297. (b)
(22) (a) Gallo, E. A.; Gellman, S. H. Tetrahedron Lett. 1992, 33, 7485.
(b) Ref. 13b.
Maxfield, F. R.; Leach, S. J.; Stimson, E. R.; Powers, S. P.; Scheraga, H.
A. Biopolymers 1979, 18, 2507.
(23) Aaron, H. S. Top. Stereochem. 1980, 11, 1.

1810 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Gardner et al.
Figure 4. N-H stretch region FT-IR data for 1 mM samples of 2, 5, and 8-10 in CH
of pure CH Cl . For 2, the bands implied by curve fitting analysis are also shown.
2
Cl
2
at room temperature, after subtraction of the spectrum
2
2
Figure 5. Amide proton NMR chemical shifts for 11 and 12 in CD
2
Cl
2
at 20 °C, as a function of the logarithm of concentration.
minimal â-hairpins, in which the central two residues form the
loop and each terminal residue constitutes a strand.9 In such
systems, two internal amide-amide hydrogen bonds are formed.
The inner hydrogen bond (10-membered ring) corresponds to
â-turn formation, and the outer hydrogen bond (14-membered
ring) is a key criterion for â-hairpin formation.
Molecule 11 is a tetrapeptide analogue in which the ADPA
glycylglycine mimic replaces the central two residues; if 11
adopted a â-hairpin-like folding pattern, the valine and leucine
residues would constitute the strands. Tetrapeptide analogue 12
is based on the APA disubstituted alkene mimic. Figure 5 shows
the variation of amide proton chemical shifts of 11 and 12 as a
function of the logarithm of concentration. These data indicate
that 12 is more prone to aggregation than is 11. This difference
provides indirect evidence that 11 is intramolecularly hydrogen
bonded to a greater extent, and therefore less available for
intermolecular hydrogen bonding, than is 12.
proton chemical shifts at low concentrations, along with
analogous data from reference compounds 13-15 (Figure 6).
Equilibration among internally hydrogen-bonded and non-
hydrogen-bonded states of small molecules is generally rapid
on the NMR time scale. An amide proton involved in such
equilibration gives rise to a single resonance, which is a
population-weighted average of the contributing hydrogen-
bonded and non-hydrogen-bonded states. Hydrogen-bonded
forms are generally downfield of non-hydrogen-bonded forms.
Because of differences in spectroscopic time scales, hydrogen-
bonding information from NMR chemical shifts is complemen-
tary to information provided by IR data. Equilibration between
hydrogen-bonded and non-hydrogen-bonded states is slow on
the IR time scale; therefore, hydrogen-bonded and non-
hydrogen-bonded states of a given N-H group give rise to
distinct bands in the IR spectrum (cf. Figure 4).
a,b
The data in Figure 6 indicate extensive folding of tetrapeptide
mimic 11 in CD2Cl2 solution. Each of the two NH chemical
shifts observed for reference compound 13 is nearly identical
Insight on the internal hydrogen-bonding patterns experienced
by 11 and 12 in methylene chloride is available from the amide

â-Turn and â-Hairpin Mimicry
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1811
that the chemical shifts reported for 12 in Figure 6 are
downfield-shifted relative to the true “monomeric” values. The
differences between 11 and 12 in Figure 6 support our
conclusion that the tetrasubstituted alkene unit more effectively
promotes â-hairpin-like folding than does the disubstituted
alkene unit.
Comparison of N-H stretch region IR data for tetrapeptide
mimics 11 and 12 (Figure 8; 1 mM in CH2Cl2) indicates that
there is a greater extent of intramolecular hydrogen bonding in
1
1 than in 12, which is consistent with a greater folding
-
1
propensity for 11. Bands above 3400 cm arise largely from
solvent-exposed N-H groups, while bands below 3400 cm
-
1
arise from hydrogen-bonded N-H groups. Although each
tetrapeptide mimic displays a complex spectrum in this region,
the larger amount of CdO- -H-N hydrogen bonding in 11 is
-1
readily apparent from the dominant band at 3326 cm . IR data
for reference compounds 13-15 reveal no bands below 3420
-
1
cm ; thus, the N-H groups in these molecules are solvent
exposed or engaged in “C5 interactions”. (The term C5 desig-
nates CdO- -H-N interactions within a single residue; whether
this five-membered ring interaction constitutes a weak hydrogen
bond has been a subject of debate because of the necessarily
poor geometry (O- -H-N angle < 100°).) The absence of
conventional CdO- -H-N hydrogen bonding within 13-15
indicated by the IR data is consistent with the conclusions drawn
above from the NMR data in Figure 6, i.e., that internal
hydrogen bonding in 11 results from folding patterns I plus II
and/or III.
Figure 6. Structures of 11-15, with NH chemical shifts measured at
1
2 2
mM in CD Cl .
1
5
We used N-labeling to gain further information from N-H
to the analogous NH chemical shift for reference compound
4, which shows that the carbonyl of the dipeptide mimetic unit
15
stretch IR spectroscopy of 11. The stretch band of a N-H
1
-1
unit is expected to occur ca. 12 cm lower than the stretch
does not form hydrogen bonds to the N-terminal residue. Thus,
neither 8- nor 11-membered ring hydrogen bonding is expected
in tetrapeptide mimic 11. (The absence of eight-membered ring
hydrogen bonding in 13 is consistent with our findings for 9,
discussed above.) Relative to the appropriate NH of reference
compounds 13 and 14, the N-terminal NH of 11 is shifted
downfield by 1.2 ppm, which suggests that this NH is involved
in considerable intramolecular hydrogen bonding. These results
imply that the acceptor is the leucine CdO (14-membered ring
hydrogen bond), and this end-to-end interaction suggests that
the desired â-hairpin-like folding pattern (I, Figure 7) is
significantly populated.
One surprising feature of the chemical shift data in Figure 6
is the 0.8 ppm downfield shift of the central NH of 11 relative
to the analogous NHs of 13 and 14. In the â-hairpin-like folding
pattern (I, Figure 7), this central NH is exposed to solvent, but
the downfield shift in 11 relative to 13 and 14 suggests that
this NH participates in intramolecular hydrogen bonding with
the C-terminal (leucine) carbonyl (11-membered ring hydrogen
bond). Since this hydrogen bond cannot occur in the â-hairpin-
like conformation, we propose that 11 equilibrates between the
â-hairpin-like folding pattern and at least one other folded
conformation. This additional folded state could involve only
the 11-membered ring hydrogen bond to the C-terminal carbonyl
14
25
band of the analogous N-H unit. We have previously shown
15
that site-specific incorporation of N into molecules with
multiple secondary amide groups can facilitate assignment of
26
N-H stretch bands. Figure 9 shows N-H stretch region IR
data for 11 without labels, and for two labeled versions, one
15
15
15
with N at the leucine N-H (11- N-Leu) and one with N at
15
the valine NH (11- N-Val). Also shown in Figure 9 is a
deconvolution of each spectrum into five component bands.
Assignments of the five N-H stretch bands identified for 11
can be made, based on the effects of labeling two of the three
15
secondary amide groups with N. We present these assignments
for 11 by considering each NH group individually.
The valine NH contributes to the 3421 and 3288 cm bands
-1
-1
identified for 11 because the former shifts to 3415 cm and
-1
15
the latter to 3280 cm in 11- N-Val, while the other three
bands are not significantly shifted in this compound. The band
-1
at 3288 cm arises from a strong amide-amide hydrogen bond;
we assign this band to valine NH involved in the 14-membered
ring hydrogen bond of â-hairpin-like conformation I (Figure
-1
7
). The band at 3421 cm is assigned to valine NH involved
in a C5 interaction, which is consistent with conformation II,
but not with the bifurcated interaction of conformation III.
(24) (a) Avignon, M.; Huong, P. V.; Lascombe, J.; Marraud, M.; Neel,
J. Biopolymers 1969, 8, 69. (b) Burgess, A. W.; Scheraga, H. A. Biopolymers
(II, Figure 7), or simultaneous 11- and 14-membered ring
1
973, 12, 2177.
25) For a localized A-B stretch, the band position can be estimated
from the equation
hydrogen bonds to this carbonyl (III, Figure 7). An N-H- -π
interaction could occur between the leucine NH and the alkene
in the additional folded state (shown in both II and III, Figure
(
ν ) (2πc)- [k(M + M )/M M ]
1
1/2
A
B
A
B
where c is the speed of light, k is the force constant of the A-B bond, MA
is the mass of atom A, and MB is the mass of atom B. (Silverstein, R. M.;
Bassler, G. C.; Morrill, T. C. Spectrometric Identification of Organic
Compounds, 5th ed.; John Wiley & Sons: New York, 1991; p 93). This
7
).
NH chemical shift data for 12 at 1 mM in CD2Cl2 (Figure 6)
suggest that intramolecular hydrogen bonding is diminished,
relative to 11, because all NH resonances of 12 are upfield of
the analogous NH resonances of 11. The data in Figure 5 show
that 12 experiences some aggregation at 1 mM, which means
1
5
-1
calculation predicts a localized N-H stretch to be ca. 12 cm lower in
energy than a 14N-H stretch.
(
26) (a) Gallo, E. A.; Gellman, S. H. J. Am. Chem. Soc. 1993, 115, 9774.
(b) Gardner, R. R.; Gellman, S. H. J. Am. Chem. Soc. 1995, 117, 10411.

1812 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Gardner et al.
Figure 7. Alternative folding patterns for tetrapeptide mimic 11 (see text for details).
Figure 8. N-H stretch region FT-IR data for 1 mM samples of 11-15 in CH
CH Cl
2 2
Cl at room temperature, after subtraction of the spectrum of pure
2
2
.
Figure 9. N-H stretch region FT-IR data for 1 mM samples of 11 and two 15N-labeled versions in CH
of the spectrum of pure CH Cl . The bands implied by curve fitting analysis are also shown.
2 2
Cl at room temperature, after subtraction
2
2
The leucine NH contributes to the 3387 and 3331 cm-1 bands
to leucine NH involved with the alkene in an N-H- -π hydrogen
-
1
-1
identified for 11 because the former shifts to 3383 cm and
bond. This band is 23 cm lower than the band assigned to
-
1
15
the latter to 3325 cm in 11- N-Leu, while the other three
bands are not significantly shifted. The large band at 3325 cm-
represents NH involved in an amide-amide hydrogen bond,
and we assign this band to the 10-membered ring hydrogen bond
of â-hairpin-like conformation I; as discussed below, we suspect
that there is also another hydrogen-bonded NH contribution to
the N-H- -π interaction in 2; we have no explanation for this
difference. The proposed N-H- -π interaction in 11 is consistent
with conformation II or III.
1
The internal NH of 11 (i.e., the NH of the ADPA glycyl-
1
5
glycine mimic) was not N-labeled because of the synthetic
-
1
difficulty of this task. Nevertheless, the band at 3437 cm can
be assigned to this internal NH, since this band does not shift
-
1
-1
the large 3325 cm band. The band at 3387 cm is assigned

â-Turn and â-Hairpin Mimicry
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1813
2
Figure 10. Summary of nonadjacent NOEs detected for 11 in CD -
Cl .
2
in either 15N-labeled molecule. This band position indicates a
solvent-exposed NH (cf. IR data for reference compound 9), as
expected for â-hairpin-like folding pattern I (Figure 7). Since
Figure 11. Structures of 16 and 17, with NH chemical shifts measured
at 1 mM in CD Cl .
2 2
-
1
the 3437 cm band is relatively small, it is likely that the
internal NH of 11 contributes to at least one other band in the
spectrum. We propose that the internal NH contributes to the
-
1
large band at 3331 cm , which would imply that this NH is
involved in a strong amide-amide hydrogen bond, as required
for conformation II or III.
The 15N-labeling/deconvolution studies provide support for
our conclusion that 11 adopts â-hairpin-like folding pattern I.
These studies also support our deduction that 11 adopts a second
folded conformation, most likely II, in which the leucine NH
is involved in an N-H- -π interaction with the alkene, the valine
NH is exposed to solvent or engaged in a C5 interaction, and
the internal NH forms an 11-membered ring hydrogen bond
with the leucine carbonyl.
NOESY27 studies were carried out with tetrapeptide mimic
11 to gain additional understanding of the folded conformations
adopted by this molecule in methylene chloride solution. The
long-range NOEs observed in a 1 mM solution are summarized
in Figure 10. None of the conformations shown in Figure 7,
alone, can account for all of these long-range NOEs; therefore,
the NOESY data provide further evidence of multiple folded
conformations. The NOE between valine NH and leucine NH
can be explained only by the â-hairpin-like folding pattern (I,
Figure 7). The NOEs between leucine NH and the two alkene
methyl groups, and between the terminal methyl groups, are
consistent with the â-hairpin-like folding pattern but could also
be explained in other ways: the leucine NH- -alkene methyl
NOEs could arise from conformation II or III, and the NOE
between the terminal methyl groups could arise from III (but
not II). The NOEs between valine NH and the internal NH,
and between the internal NH and leucine NH, cannot arise from
â-hairpin-like conformation I; the former is consistent with
conformation III, and the former, with II or III.
Figure 12. N-H stretch region FT-IR data for 1 mM samples of 16
and 17 in CH Cl at room temperature, after subtraction of the spectrum
of pure CH Cl
2
2
2
2
.
In 16, the leucine NH is replaced by an ester oxygen; this
molecule allows us to test the favorability of proposed folding
patterns II and III in Figure 7. A â-hairpin-like folding pattern,
analogous to I, should be strongly disfavored in 16, since the
ester placement precludes formation of the “inner” 10-membered
ring hydrogen bond. On the other hand, the favorability of the
folding pattern analogous to II should not be compromised in
Further information on the folding behavior of 11 was
obtained by examination of two analogues, 16 and 17, in which
one secondary amide group is replaced by an ester group. As
we have previously pointed out, this replacement is conservative
in the sense that the secondary amide and ester groups have
similar conformational preferences.26 However, ester and sec-
ondary amide groups differ profoundly in their hydrogen-
bonding properties, since the ester cannot donate a hydrogen
16 relative to 11. Although the N-H- -π hydrogen bond shown
in Figure 7 for II and III cannot form in 16, it is unlikely that
this weak interaction would be a dominant contributor to
conformational stability.
2
8
bond and the ester is a much weaker acceptor than the amide.
Figure 11 summarizes amide proton chemical shifts of 16 and
7 (1 mM in CD2Cl2), for comparison with Figure 5. Figure 12
shows N-H stretch region IR data for 16 and 17.
The behavior of 16 indicates that folding patterns such as II,
and perhaps III, are reasonable conformations for this class of
tetrapeptide analogues, but that neither II nor III can be the
major conformation adopted by 11. Comparison of the amide
proton chemical shifts for 16 (Figure 11) and reference
compound 12 (Figure 6) suggests that there is significant 11-
membered ring hydrogen bonding in 16, between the C-terminal
1
(
27) Macura, S.; Ernst, R. R. Mol. Phys. 1980, 41, 95.
(28) (a) Arnett, E. M.; Mitchell, E. J.; Murty, T. S. S. R. J. Am. Chem.
Soc. 1974, 96, 3875. (b) Spencer, J. N.; Garrett, R. C.; Moyer, F. J.; Merkle,
J. E.; Powell, C. R.; Tran, M. T.; Berger, S. K. Can. J. Chem. 1980, 58,
1
372.

1814 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Gardner et al.
amide carbonyl and the NH of the dipeptide mimic because
this δNH in 16 is 0.6 ppm downfield relative to 13. In contrast,
the valine NH resonances are less than 0.2 ppm apart in 16 and
except for hexane, which was purified by distillation. Anhydrous
reaction conditions were maintained under a slightly positive nitrogen
atmosphere in oven-dried glassware. Silica gel chromatography was
performed using 230-400 mesh silica purchased from EM Science.
Routine H and C NMR spectra were obtained on a Bruker AM-300
spectrometer at 300.132 and 75.033 MHz frequencies, respectively.
1
3, which suggests that a folding pattern like II is preferable to
a folding pattern like III. The N-H stretch region IR data for
6 (Figure 12) confirm that this molecule experiences significant
1
13
1
NMR spectra were referenced to tetramethylsilane (TMS) (0 ppm) or
-
1
intramolecular hydrogen bonding (3323 cm ); however, most
1
to the isotopic impurity peak of CDCl
3
(7.26 and 77.0 ppm for H and
of the N-H groups of 16 are solvent-exposed or engaged in a
13
C, respectively). Routine infrared spectra were obtained using a
-
1
C5 interaction because the 3427 cm band is larger than the
Nicolet 740 FT-infrared spectrometer. High-resolution electron impact
ionization mass spectroscopy was performed using a Kratos MS-25
spectrometer.
-
1
3
323 cm band (hydrogen-bonded bands have larger extinction
29
coefficients than non-hydrogen-bonded bands ). Comparison
of the IR data for 16 with the IR data for 11 (Figure 8) shows
that there is considerably less intramolecular hydrogen bonding
in 16 than in 11. Thus, there must be another major intramo-
lecularly hydrogen-bonding pattern available to 11, presumably
â-hairpin-like conformation I.
IR Studies. High-quality infrared spectra (128 scans) were obtained
-
1
at 2 cm resolution using a 1 mm CaF
740 FT-infrared spectrometer. All spectra were obtained in 1 mM
solutions in anhydrous CH Cl at 297 K. All compounds that were
2
solution cell and a Nicolet
2
2
1
studied were judged to be >95% pure by examination of 500 MHz H
NMR spectra. All compounds were dried in vacuo at elevated
In 17, the NH of the dipeptide mimic unit is replaced with
an ester oxygen. In this case, it is impossible to have the 11-
membered ring hydrogen bond central to folding patterns II
and III (Figure 7), but a â-hairpin-like folding pattern, analogous
to I can form, with the ester serving as the acceptor for the
2 5 2 2 2
temperatures in the presence of P O . CH Cl was distilled from CaH
and stored over activated 4 Å molecular sieves. All sample preparations
were performed in a nitrogen atmosphere.
1
1
H NMR Studies. H NMR spectra for the concentration-dependence
studies were obtained using a Bruker AM-500 spectrometer. Com-
pounds were dried in vacuo at elevated temperatures in the presence
of P O . CD Cl was dried by being stored over activated 4 Å molecular
2 5 2 2
“
inner” 10-membered ring hydrogen bond. Both NMR and IR
data suggest that there is considerable â-hairpin-like folding of
1
7, even though the ester-amide hydrogen bond is weaker than
sieves for a minimum of 4 days. All sample preparations were
performed in a nitrogen atmosphere. Solutions for temperature-
26,28
the corresponding amide-amide hydrogen bond of 11.
The
valine δNH is quite far downfield (Figure 11; compare
analogous δNH of 13 and 11, in Figure 6), which is consistent
with substantial 14-membered ring hydrogen bonding. The
leucine δNH of 17 is also downfield shifted, relative to reference
dependence studies were 1 mM in CD
2 2
Cl and were degassed and
placed in sealed NMR tubes. All spectra were referenced relative to
the isotopic impurity peak in CD
at all temperatures.
2 2
Cl , which was assumed to be 5.320
1
26,28
Samples for two-dimensional H NMR experiments were prepared
as described above. All two-dimensional NMR spectra were obtained
on a Varian Unity-500 spectrometer at 298 K. Proton signals were
assigned from TOCSY30 and NOESY spectra. Two-dimensional
spectra were obtained using standard Varian pulse sequences, and the
data were processed with Varian V-NMR version 4.3 software. TOCSY
compound 15; since esters are weaker acceptors than amides,
the maximum downfield shift for leucine δNH in 16 is probably
smaller than the fully hydrogen bonded limit of the leucine δNH
in 11. IR data (Figure 12) show that 17 is intramolecularly
hydrogen bonded to a large extent. Only a relatively small band
27
-
1
(3427 cm ) is observed in the solvent-exposed/C5 region. The
spectra were recorded with 2048 data points in t
1
, 256 data points in
-
1
large band at 3375 cm is assigned to the ester-amide
hydrogen bond (10-membered ring) of the â-hairpin-like folding
pattern, and the shoulder at 3330 cm is assigned to the amide-
amide hydrogen bond (14-membered ring).
t , 16 scans per t increment, sweep width ) 4000 Hz, relaxation delay
2
2
) 600 ms, and a mixing time of 60 ms. NOESY spectra were recorded
with 2048 data points in t , 512 data points in t , 128 scans per t
increment, sweep width ) 4000 Hz, relaxation delay ) 2500 ms, and
a mixing time of 900 ms. (This mixing time was well below the T
-1
1
2
2
1
Conclusion. Our analysis of di-, tri-, and tetrapeptide
analogues containing the ADPA glycylglycine mimic demon-
strates that avoidance of allylic strain within a tetrasubstituted
alkene unit provides significant conformational preorganization.
Use of the tetrasubstituted (E)-alkene favors adoption of folded
conformations more strongly than does the more conventional
APA disubstituted alkene “amide isostere.” The desired â-hairpin-
like folding pattern is adopted when the ADPA glycylglycine
mimic is placed at the center of a tetrapeptide analogue (11);
however, at least one other folding pattern is also populated in
this case, in methylene chloride solution. Our analysis shows
how a combination of NMR and IR techniques, along with
careful choice of reference compounds, can elucidate a con-
formational equilibrium involving multiple intramolecularly
hydrogen bonded states. These results suggest that the ADPA
glycylglycine mimic may be useful for pharmaceutical and
molecular design applications.
measured for the amide NH resonances and most other resonances in
the molecule.)
Synthesis. Preparative routes for diamides 2 and 5 are provided
below. Additional synthetic protocols may be found in the Supporting
Information.
N-(tert-Butoxycarbonyl)-2,3-dimethyl-4-(2,6-dithianyl)-(E)-2-bute-
namine (3). A solution of 1,3-dithiane (7.08 g, 59 mmol) in anhydrous
tetrahydrofuran (THF) (130 mL) was cooled to -30 °C under N . A
2
solution of n-butyllithium (22.5 mL, 2.6 M in hexane, 59 mmol) was
added dropwise over 40 min. After being stirred for an additional 80
min, the solution was added dropwise (via a cannula) over 2 h to a
solution of 1,4-dibromo-2,3-dimethyl-3-(E)-butene (14.4 g, 59.5 mmol)
in anhydrous THF (100 mL) at -78 °C. The resulting yellow solution
was stirred for 6 h while being warmed to room temperature. The
reaction mixture was quenched with water (50 mL) and then concen-
trated to a total volume of 100 mL. The organic layer was separated,
and the aqueous layer was extracted with Et
combined organic layers were concentrated in vacuo, and the semisolid
residue was taken up in water (50 mL). This mixture was treated with
2
O (2 × 60 mL). The
Experimental Section
3 4
NaN (4.36 g, 67 mmol) and Bu NBr (1.9 g, 5.9 mmol), and the two-
phase mixture was stirred at 80 °C for 10 h. The reaction mixture
was cooled, and the organic phase was separated. The aqueous layer
was extracted with Et O (50 mL), and the combined organic layers
2
were concentrated to yield a brown viscous liquid. This material was
taken up in THF (170 mL) and water (30 mL) and then was treated
General Procedure. All melting points are uncorrected. Reagents
employed were either commercially available or prepared according
to a known procedure. Anhydrous CH
2 2
Cl and t-BuOH were obtained
by distillation from CaH . All other solvents used were reagent grade
2
(
29) Pimentel, G. C.; McClellan, A. L. The Hydrogen Bond; Freeman:
San Francisco, CA, 1960.
(30) Bax, A.; Davis, D. G. J. Magn. Reson. 1985, 65, 355.

â-Turn and â-Hairpin Mimicry
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1815
ylene groups of 4 itself were not observed.
with triphenylphoshine (23.2 g, 88.5 mmol) and stirred at room
temperature for 30 h. The THF was removed in vacuo, and the resulting
3
material was mixed with CHCl . This solution was extracted with 1 N
HCl (4 × 50 mL). The combined acid extracts were carefully treated
with 10 N NaOH until the pH of the solution was greater than 10,
which caused formation of a dark layer at the top of the mixture. The
two-phase mixture was extracted with Et
combined organic layers were dried with Na
give the desired amine (4.32 g) as a mixture of isomers. H NMR
CDCl ),
.8-2.3 (m, 2H, 2 × CH ), 2.53 (d,
J ) 8.30 Hz, 2H, CH ), 3.27 (s, 2H,
NCH
2
O (5 × 50 mL), and the
2
SO and concentrated to
4
1
(
1
3
/TMS, ppm): 1.50 (broad, 2H, NH), 1.7-1.8 (m, 6H, 4 × CH
3
2
), 2.48 (d, J ) 7.90 Hz, 2H, CH
2
2
), 2.8-2.9 (m, 4H, 4 × SCH
2
2
2
), 3.28 (s, 2H, NCH ), 4.18 (t, J ) 7.90 Hz, 1H, SCHS), 4.22 (t,
J ) 8.30 Hz, 1H, SCHS). The E/Z isomer ratio was ∼1:1. A solution
of the crude isomeric mixture of amines (1.0 g, 4.6 mmol) in dioxane
(10 mL) and water (5 mL) was treated with potassium carbonate (0.73
g, 5.3 mmol) and di-tert-butyl dicarbonate (1.1 g, 5.06 mmol). This
mixture was stirred at 0 °C for 3 h. The dioxane and water were
removed in vacuo, yielding a dark oil that was purified by silica gel
chromatography eluting with 15% (v/v) EtOAc in hexanes. The second
Isobutyryl-ADPA-NHiPr (2). A solution of 4 (97 mg, 0.4 mmol)
in THF (1.4 mL) was treated with 1 N HCl (1.4 mL, 1.4 mmol). After
isomer to elute was the desired E-isomer (3) as a white solid (0.74 g,
1
1
9
7.5%). Mp: 52.0-55.0 °C. H NMR (CDCl
3
/TMS, ppm): 1.45 (s,
), 1.8-2.2 (m, 1H, CH ), 2.48
), 2.83 (m, 4H, CH ), 3.76 (d, J ) 5.55 Hz,
), 4.20 (t, J ) 7.67 Hz, 1H, SCHS), 4.45 (broad, 1H, NH).
C NMR (CDCl , ppm): 17.4, 18.2 (CdC-CH ), 25.8 (C-CH -C),
8.3 (C(CH ), 40.1, 43.5 (CH ), 46.4 (S-CH -S),
), 30.6 (2 × SCH
4.2 (C-CH ), 128.5, 129.2 (CdC), 156.0 (CdO). IR (neat, cm ):
358 (m), 2975 (m), 2929 (m), 2914 (m), 2901 (m), 1705 (s), 1507
317.1483).
N-(Isobutyryl)-5,5-dimethoxy-2,3-dimethyl-(E)-2-pentenamine (4).
stirring for 30 min, the reaction mixture was extracted with CHCl
3
(5
and
then concentrated to yield the aldehyde intermediate (76 mg). H NMR
CDCl /TMS, ppm): 1.17 (d, J ) 6.81 Hz, 6H, CH
), 1.69, 1.77 (2 ×
m, 6H, CdCCH ), 2.36 (septet, J ) 6.88 Hz, 1H, CH), 3.19 (s, 2H,
CH ), 3.96 (d, J ) 3.55 Hz, 2H, CH ), 5.40 (broad, 1H, NH), 9.62 (t,
J ) 2.01 Hz, 1H, HCO). A solution of the crude aldehyde (76 mg,
.38 mmol) in t-BuOH (2 mL) and phosphate buffer (2 mL, 1 N, pH
4) was treated with 2-methyl-2-butene (0.1 mL, 0.94 mmol) and
NaClO (36 mg, 0.40 mmol) and was stirred at room temperature for
0 min. The reaction mixture was extracted with CHCl
(10 × 5 mL),
and the combined organic layers were dried with Na SO and
H, CH
3
), 1.75, 1.77 (2 × m, 6H, CH
3
2
×
5 mL). The combined organic layers were dried with Na SO
2
1
4
(d, J ) 7.66 Hz, 2H, CH
2
2
2
H, NCH
2
(
3
2
13
3
3
2
3
2
5
3
)
3 3
2
2
2
-
1
2
2
3
0
)
(
2 2
m). MS (EI, m/z): 317.1499 (calcd for C15H27NO S
2
A solution of 3 (0.26 g, 0.82 mmol) in anhydrous MeOH (22 mL) was
treated with methanesulfonic acid (0.81 mL, 12.5 mmol), and the
3
3
2
4
mixture was stirred under N
aqueous NaHCO (1 N) and then extracted with CH
The combined organic layers were dried with MgSO
2
for 18 h. The reaction was quenched with
concentrated to yield the carboxylic acid intermediate (74 mg) as a
1
3
2
Cl
and concentrated
/TMS, ppm):
), 1.8-2.2 (m, 2H, CH ), 2.48
), 2.8-2.9 (m, 4H, CH ), 3.27 (s, 2H, NCH ),
.22 (t, J ) 7.65 Hz, 1H, SCHS). This material (0.17 g, ∼0.76 mmol)
at 0 °C and then
treated with triethylamine (0.22 mL, 1.52 mmol) and isobutyryl chloride
0.1 mL, 0.95 mmol). A white solid precipitated upon addition of the
2
(5 × 10 mL).
yellow oil. H NMR (CDCl /TMS, ppm): 1.16 (d, J ) 6.89 Hz, 6H,
3
4
CH ), 1.17, 1.83 (2 × m, 6H, dCCH ), 2.37 (septet, J ) 6.85 Hz, 1H,
3
3
1
to give 0.17 g (93%) of a yellow oil. H NMR (CDCl
.34 (broad, 2H, NH), 1.78 (m, 6H, CH
d, J ) 7.66 Hz, 2H, CH
3
2 2
CH), 3.12 (s, 2H, CH ), 3.92 (d, J ) 5.58 Hz, 2H, NCH ), 5.52 (broad,
1
3
2
1H, NH). A solution of the crude acid (84 mg, ∼0.4 mmol) in anhydrous
THF (5.2 mL) was treated with N-hydroxysuccinimide (61 mg, 0.53
mmol) and dicyclohexylcarbodiimide (DCC) (109 mg, 0.53 mmol), and
(
2
2
2
4
was dissolved in anhydrous THF (10 mL) under N
2
2
the mixture was stirred for 2 h under N . Isopropylamine (0.101 mL,
1.2 mmol) was added, and the reaction was stirred for an additional 12
h. The reaction mixture was suction-filtered, and the filtrate was
concentrated. The residue was purified by silica gel chromatography
(
isobutyryl chloride. After 30 min, the reaction mixture was suction
filtered and the filtrate was concentrated. This gave a tan solid (0.20
g) that was dissolved in anhydrous MeOH (5 mL) and then treated
with small portions of bis(trifluoroacetoxy)iodobenzene (PIFA) (0.32
g, 0.76 mmol). All solids dissolved quickly, and a slight exotherm
eluting with EtOAc, yielding 2 as a white solid (23.9 mg, 28%). Mp
1
(
(
6
recrystallized from EtOAc and hexanes): 199.0-200.0 °C. H NMR
CDCl
.56 Hz, 6H, CH
J ) 6.88 Hz, 1H, CH), 2.99 (s, 2H, CH
NCH ), 4.08 (m, 1H, NCH), 5.69 (broad, 1H, NH), 6.08 (broad, 1H,
NH). 13C NMR (CDCl
CH ), 41.2 (NCH), 42.9 (CH
3
/TMS, ppm): 1.16 (d, J ) 6.88 Hz, 6H, CH
), 1.64, 1.85 (2 × m, 6H, CdC-CH
), 3.85 (d, J ) 5.39 Hz, 2H,
3
), 1.13 (d, J )
3
3
), 2.37 (septet,
2
resulted upon PIFA addition. After 30 min of stirring under N
reaction was quenched with 1 N NaHCO and then extracted with Et
SO
4
2
, the
2
3
2
O
3
, ppm): 16.3 (CH
3
), 19.4 (CH
3
), 22.5, 19.7 (4
(
3 × 10 mL). The combined organic layers were dried with Na
2
×
3
2
), 43.0 (CH
2
), 126.9, 129.6 (CdC),
and then concentrated. The crude product was purified by silica gel
-
1
1
69.0, 172.4 (CdO). IR (1 mM in CH
m), 1667 (s), 1513 (s), 1467 (m), 1381 (w), 1368 (w). MS (EI, m/z):
54.2009 (calcd for C14 254.1994).
N-(Isobutyryl)-5,5-dimethoxy-2,3-diethyl-(E)-2-pentenamine. A
slurry of potassium tert-butoxide (31.6 g, 258 mmol) and n-butyllithium
98.7 mL, 2.6 M in hexane, 258 mmol) in anhydrous pentane (100
mL) was treated dropwise with a solution of 2,3-dimethyl-1,3-butadiene
15 mL, 130 mmol) in anhydrous pentane (50 mL) over 20 min. The
slurry was stirred at room temperature for 30 min under N . The stirring
was stopped and the resulting orange salt allowed to settle. The pentane
was removed over a 2 h period by blowing dry N through the flask.
2 2
Cl , cm ): 3453 (m), 3330
chromatography, eluting with 15% (v/v) EtOAc in hexanes, yielding 4
as a white solid (0.13 g, 64%). Mp (recrystallized from EtOAc and
(
2
26 2 2
H N O
1
hexanes): 60.0-61.5 °C. H NMR (CDCl
3
/TMS, ppm): 1.16 (d, J )
), 2.37 (septet, J )
), 3.34 (s, 6H, OCH ),
), 4.43 (t, J ) 5.54 Hz, 1H, OCHO),
.31 (broad, 1H, NH). 13C NMR (CDCl
, ppm): 17.1 (CH ), 18.9 (CH ),
9.7 (2 × CH ), 35.7 (CH), 38.3 (CH ), 42.3 (CH N), 53.3 (OCH ),
03.2 (OCHO), 128.6 (CdC), 129.6 (CdC), 176.0 (CdO). IR (1 mM
6
.90 Hz, 6H, CH
.90 Hz, 1H, CH), 2.39 (d, J ) 5.68 Hz, 2H, CH
.89 (d, J ) 5.40 Hz, 2H, NCH
3
), 1.71, 1.78 (2 × m, 6H, C-CH
3
6
2
3
(
3
2
5
3
3
3
(
1
3
2
2
3
2
1
-
1
in CH
2
Cl
2
, cm ): 3442 (w), 1668 (s), 1511 (s), 1447 (m), 1368 (m).
243.1844).
2
MS (EI, m/z): 243.1834 (calcd for C13H25NO
3
Once the pentane was removed, anhydrous THF (100 mL) was added
and the resulting slurry was cooled to -78 °C. The cooled slurry was
cannulated into a -78 °C solution of MeI (16.1 mL, 258 mmol) in
anhydrous THF (100 mL) over a 1 h period. The reaction mixture was
stirred and allowed to warm to room temperature over 30 min. The
reaction mixture was mixed with pentane (250 mL) and then was
washed with water (8 × 100 mL) to remove THF. Pentane in the
NOE Study of Compound 4. An NOE experiment was performed
on 4 and its Z-isomer using a Bruker AM-500 NMR spectrometer. The
automated program NOEDIFF.AU was used for the NOE experiment
(
0
using the parameters: D3 ) 0.1, S3 ) 35L, D1 ) 15, D2 ) 2, RD )
. PW ) 10.5, DE ) 113.8, NS ) 8, NE ) 8). The results for the
Z-isomer of 4 are summarized below. NOEs between the vinylmeth-

1816 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Gardner et al.
resulting yellow liquid was removed by Kugelrohr distillation. The
crude material contained significant amounts of pentane and was carried
on to the next step without further purification. The yellow oil was
(v/v) EtOAc in hexane, to yield 0.18 g of the desired product as a
1
white solid. Mp: 52.0-53.5 °C. H NMR (CDCl
3
/TMS, ppm): 0.97,
), 1.15 (d, J ) 6.89 Hz, 6H,
), 2.09, 2.14 (2 × q, J ) 7.52, 7.54 Hz, 4H, CH ), 2.37 (septet, J
) 6.90 Hz, 1H, CH), 2.39 (d, J ) 5.73 Hz, 2H), 3.34 (s, 6H, CH ),
3.88 (d, J ) 5.73 Hz, 2H, NCH ), 4.39 (t, J ) 5.55 Hz, 1H, OCHO),
5.36 (broad, 1H, NH). 13C NMR (CDCl
, ppm): 13.3, 13.8 (CH ), 19.6
(CH ), 24.0 (CH ), 25.2 (CH ), 34.9 (CH ), 35.6 (CH), 39.4 (NCH ),
53.4 (OCH ), 133.7 (CdC), 135.3 (CdC), 176.7 (CdO); IR (1 mM in
CH Cl
, cm-1): 3442 (w), 1712 (w), 1668 (s), 1509 (m), 1460 (w),
0.98 (2 × t, J ) 7.41, 7.48 Hz, 6H, CH
3
dissolved in CHCl
bromine (0.55 mL, 11 mmol) in CHCl
3
at -25 °C and then was treated with a solution of
(10 mL) over a 2 h period.
CH
3
2
3
3
The mixture was stirred an additional 2 h while warming to room
temperature. The reaction mixture was concentrated, and the resulting
brown material was purified by silica gel chromatography eluting with
2
3
3
3
2
2
2
2
hexane, yielding 3.76 g of an isomeric mixture (3:1, E:Z) of 1,4-
3
1
dibromo-2,3-diethyl-2-butene as a white solid. H NMR (CDCl
3
/TMS,
), 2.23, 2.34 (2 × q, J
), 4.03, 4.08 (2 × s, 4H, CH Br). MS (EI, m/z):
69.9436 (calcd for C 269.9439).
A solution of 1,3-dithiane (2.12 g, 17.6 mmol) in anhydrous THF
40 mL) was cooled to -30 °C under N . A solution of n-butyllithium
12.0 mL, 1.6 M in hexane, 19.2 mmol) was added dropwise (via
2
2
ppm): 1.06, 1.11 (2 × t, J ) 7.59 Hz, 6H, CH
3
1382 (m), 1370 (w). MS (EI, m/z): 271.2144 (calcd for C15H29NO
3
271.2147).
N-Isopropyl-3,4-diethyl-5-(isobutyramido)-(E)-3-pentenamide (5).
A solution of N-((E)-2,3-diethyl-5,5-dimethoxy-2-pentenyl)-2-methyl-
propanamide (0.17 g, 0.62 mmol) in THF (2.15 mL) was treated with
1 N HCl (2.15 mL, 2.15 mmol). After being stirred for 3 h, the reaction
)
2
7.59 Hz, 4H, CH
2
2
8
H14Br
2
(
(
2
cannula) over 40 min. After stirring for an additional 80 min, the
solution was added dropwise over 2 h to a solution of 1,4-dibromo-
mixture was extracted with CHCl
layers were dried with Na SO
aldehyde intermediate as a yellow oil (0.13 g). H NMR (CDCl
ppm): 0.98 (t, J ) 7.52 Hz, 6H, CH ), 1.15 (d, J ) 6.87 Hz, 6H,
CH ), 2.38 (septet, J
), 2.05, 2.16 (2 × q, J ) 7.56, 7.51 Hz, 4H, CH
) 6.89 Hz, 1H, CH), 3.20 (d, J ) 2.08 Hz, 2H, CH ), 3.94 (d, J )
5.25 Hz, 2H, CH ), 5.68 (broad, 1H, NH), 9.61 (t, J ) 2.01 Hz, 1H,
3
(5 × 5 mL). The combined organic
2
4
and were concentrated to yield the
1
2
,3-diethyl-3-(E)-butene (3.76 g, 13.9 mmol) in anhydrous THF (30
3
/TMS,
mL) at -78 °C. The resulting yellow solution was stirred for 6 h while
being warmed to room temperature. The reaction mixture was quenched
with water and then was concentrated to a total volume of 30 mL. The
organic layer was separated off, and the aqueous layer was extracted
3
3
2
2
2
with Et
in vacuo, and the semisolid residue was taken up in water (35 mL).
This mixture was treated with NaN (1.50 g, 23 mmol) and Bu NBr
0.50 g, 1.6 mmol), and the two-phase mixture was stirred at 80 °C for
2
O (2 × 50 mL). The combined organic layers were concentrated
HCO). A solution of the aldehyde (0.13 g, 0.58 mmol) in t-BuOH (5
mL) and phosphate buffer (5 mL, 1 N, pH ) 4) was treated with
2-methyl-2-butene (0.3 mL, 2.82 mmol), and the mixture was stirred
3
4
(
for 30 min. The reaction mixture was extracted with CHCl
mL), and the combined organic layers were dried with Na
Concentration of the organic layers yielded the desired carboxylic acid
3
(10 × 10
10 h. The reaction mixture was cooled, the organic phase was separated,
2
SO .
4
and the aqueous layer was extracted with Et
2
O (3 × 40 mL). The
1
combined organic layers were concentrated to a brown oil. This material
was dissolved in THF (20 mL) and water (5 mL) and with triphen-
ylphoshine (5.24 g, 20 mmol). The reaction mixture was stirred at room
temperature for 30 h. The THF was removed in vacuo, and the resulting
intermediate as a yellow semisolid (0.13 g). H NMR (CDCl
ppm): 0.98 (t, J ) 7.50 Hz, 6H, CH ), 1.15 (d, J ) 6.88 Hz, 6H,
CH ), 2.38 (septet, J
), 2.06, 2.18 (2 × q, J ) 7.53, 7.53 Hz, 4H, CH
) 6.87 Hz, 1H, CH), 3.13 (s, 2H, CH ), 3.90 (d, J ) 5.00 Hz, 2H,
CH ), 5.68 (broad, 1H, NH), 6.95 (broad, 1H, COOH). A solution of
the acid (136 mg, ∼0.57 mmol) in anhydrous CH Cl (5.2 mL) was
treated with N-hydroxysuccinimide (89 mg, 0.77 mmol) and DCC (159
mg, 0.77 mmol), and the mixture was stirred for 2 h under N
3
/TMS,
3
3
2
2
material dissolved in CHCl
6 × 30 mL). The combined acidic extracts were made basic with 10
N NaOH, which resulted in the formation of a dark layer at the top of
the mixture. The two-phase mixture was extracted with Et
O (4 × 50
mL), and the combined organic layers were dried with Na SO to yield
.3 g of amine. The isolated material was a 1:1 mixture of E- and
3
. This solution was extracted with 1 N HCl
2
(
2
2
2
2
.
2
4
Isopropylamine (0.15 mL, 1.7 mmol) was added, and the reaction was
stirred for an additional 12 h. The reaction mixture was filtered, and
the filtrate was concentrated. The resulting residue was purified by silica
gel chromatography eluting with 50% (v/v) EtOAc in hexane, yielding
1
Z-isomers and was taken to the next reaction without further purifica-
tion. A stirred solution of the amine (1.3 g, ∼5.3 mmol) in anhydrous
THF (30 mL) was cooled to 0 °C under N
2
and then was treated with
5 (76 mg, 44%) as a white solid. Mp (recrystallized from EtOAc and
1
triethylamine (TEA) (1.12 mL, 8 mmol) and isobutyryl chloride (0.75
mL, 8 mmol). Copious amounts of a white solid precipitated upon
addition of the isobutyryl chloride. After 30 min, the reaction mixture
was vacuum filtered and the filtrate was concentrated, yielding 1.33 g
of crude N-(2,3-dimethyl-4-(2,6-dithianyl)-2-pentenyl-2-methylpro-
panamide. This material was purified by silica gel chromatography,
hexane): 159.0-161.0 °C. H NMR (CDCl
3
/TMS, ppm): 0.97 (t, J )
), 1.16 (d, 6.95 Hz,
), 2.03, 2.23 (2 × q, J ) 7.57, 7.53 Hz, 4H, CH ), 2.35 (septet,
J ) 6.87 Hz, 1H, CH), 3.01 (s, 2H, CH ), 3.88 (d, J ) 5.27 Hz, 2H,
NCH ), 4.07 (m, 1H, NCH), 5.56 (broad, 1H, NH), 6.00 (broad, 1H,
NH). 13C NMR (CDCl
, ppm): 13.1 (CH ), 13.4 (CH ), 22.5, 23.2 (4
× CH ), 23.2, 26.3 (CH ), 35.6 (CH), 39.7, 39.9 (CH
7.49 Hz, 6H, CH
6H, CH
3
), 1.13 (d, J ) 6.63 Hz, 6H, CH
3
2
2
2
2
3
3
3
eluting with 25% (v/v) EtOAc in hexane, yielding a tan solid (0.28 g).
3
2
2
), 41.2 (NCH),
1
-1
H NMR (CDCl
.15 (d, J ) 6.90 Hz, 6H, CH
q, J ) 7.45, 7.47 Hz, 4H, CH
.52 (d, J ) 7.78 Hz, 2H), 2.8-2.9 (m, 4H, SCH
Hz, 2H, NCH ), 4.19 (t, J ) 7.53 Hz, 1H, SCHS), 5.21 (broad, 1H,
NH). IR (neat, cm ): 3281(s), 1640 (s), 1560 (m), 1472 (m), 1458
m), 1430 (w). MS (EI, m/z): 315.1661 (calcd for C16
15.1690).
A solution N-(2,3-dimethyl-4-(2,6-dithianyl)-2-pentenyl 2-methyl-
3
/TMS, ppm): 0.99 (2 × t, J ) 7.48 Hz, 6H, CH
), 1.8-2.0 (m, 2H, CH ), 2.12, 2.13 (2
), 2.34 (septet, J ) 6.90 Hz, 1H, CH),
), 3.88 (d, J ) 6.52
3
),
2 2
130.9, 135.6 (CdC), 169.6, 177.2 (CdO). IR (1 mM in CH Cl , cm ):
1
3
2
3450 (w), 3439 (w), 3412 (w), 3332 (w), 1711 (w), 1669 (s), 1650
(m), 1605 (w). MS (EI, m/z): 282.2295 (calcd for C16H N O
30 2 2
282.2209).
×
2
2
2
2
-
1
Acknowledgment. This research was supported by the
National Science Foundation (CHE-9622953). NMR spectrom-
eters were purchased through Grants NSF CHE-8813550 and
NIH 1 S10 RR04981.
(
3
H29NOS
2
propanamide (0.21 g, 0.67 mmol) in anhydrous MeOH (10 mL) was
treated with small portions of PIFA (0.43 g, 1.01 mmol). All solids
dissolved quickly, and a slight exotherm resulted upon PIFA addition.
Supporting Information Available: Text describing syn-
thetic protocols and characterization for molecules not discussed
in the Experimental Section (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
After 30 min of stirring under N
NaHCO and then was extracted with Et
organic layers were dried with Na SO
2
, the reaction was quenched with 1 N
3
2
O (3 × 10 mL). The combined
2
4
and concentrated. The crude
product was purified by silica gel chromatography, eluting with 50%
JA9824526