Stapling of a 310-helix with click chemistry

Article abstract of DOI:10.1021/jo101670a

Short peptides are important as lead compounds and molecular probes in drug discovery and chemical biology, but their well-known drawbacks, such as high conformational flexibility, protease lability, poor bioavailability and short half-lives in vivo, have prevented their potential from being fully realized. Side chain-to-side chain cyclization, e.g., by ring-closing olefin metathesis, known as stapling, is one approach to increase the biological activity of short peptides that has shown promise when applied to 3₁₀-and α-helical peptides. However, atomic resolution structural information on the effect of side chain-to-side chain cyclization in 3₁₀-helical peptides is scarce, and reported data suggest that there is significant potential for improvement of existing methodologies. Here, we report a novel stapling methodology for 3₁₀-helical peptides using the copper(I)-catalyzed azide-alkyne cycloaddition (CuAAC) reaction in a model aminoisobutyric acid (Aib) rich peptide and examine the structural effect of side chain-to-side chain cyclization by NMR, X-ray diffraction, linear IR and femtosecond 2D IR spectroscopy. Our data show that the resulting cyclic peptide represents a more ideal 3₁₀-helix than its acyclic precursor and other stapled 3₁₀-helical peptides reported to date. Side chain-to-side chain stapling by CuAAC should prove useful when applied to 3 ₁₀-helical peptides and protein segments of interest in biomedicine.(Figure Presented)

Full text of DOI:10.1021/jo101670a

pubs.acs.org/joc
Stapling of a 310-Helix with Click Chemistry
,†
Øyvind Jacobsen,* Hiroaki Maekawa, Nien-Hui Ge, Carl Henrik Gorbitz,
‡
‡
§
€
Pa
˚
l Rongved,† Ole Petter Ottersen,z Mahmood Amiry-Moghaddam,z and Jo Klaveness†
†School of Pharmacy, University of Oslo, P.O. Box 1068 Blindern, 0316 Oslo, Norway, ‡Department of
Chemistry, University of California at Irvine, Irvine, California 92697-2025, United States, §Department of
Chemistry, University of Oslo, P.O. Box 1033 Blindern, 0315 Oslo, Norway, and Centre for Molecular
Biology and Neuroscience, University of Oslo, P.O. Box 1105 Blindern, 0317 Oslo, Norway
z
oyvind.jacobsen@farmasi.uio.no
Received August 25, 2010
Short peptides are important as lead compounds and molecular probes in drug discovery and
chemical biology, but their well-known drawbacks, such as high conformational flexibility, protease
lability, poor bioavailability and short half-lives in vivo, have prevented their potential from being
fully realized. Side chain-to-side chain cyclization, e.g., by ring-closing olefin metathesis, known as
stapling, is one approach to increase the biological activity of short peptides that has shown promise
when applied to 310- and R-helical peptides. However, atomic resolution structural information on
the effect of side chain-to-side chain cyclization in 310-helical peptides is scarce, and reported data
suggest that there is significant potential for improvement of existing methodologies. Here, we report
a novel stapling methodology for 310-helical peptides using the copper(I)-catalyzed azide-alkyne
cycloaddition (CuAAC) reaction in a model aminoisobutyric acid (Aib) rich peptide and examine the
structural effect of side chain-to-side chain cyclization by NMR, X-ray diffraction, linear IR and
femtosecond 2D IR spectroscopy. Our data show that the resulting cyclic peptide represents a more
ideal 310-helix than its acyclic precursor and other stapled 310-helical peptides reported to date. Side
chain-to-side chain stapling by CuAAC should prove useful when applied to 310-helical peptides and
protein segments of interest in biomedicine.
Introduction
this concept has been successfully applied to R-helical pep-
tides and protein segments. Examples include peptides with
intramolecular H-bond surrogates1 and so-called stapled
The concept of introducing conformational constraints in
peptides that stabilize their biologically active secondary
structure has attracted a lot of interest as a way to improve
the pharmacological properties of peptides. In particular,
(1) Vernall, A. J.; Cassidy, P.; Alewood, P. F. Angew. Chem., Int. Ed.
2009, 48, 5675.
1228 J. Org. Chem. 2011, 76, 1228–1238
Published on Web 01/28/2011
DOI: 10.1021/jo101670a
r
2011 American Chemical Society

Jacobsen et al.
JOCArticle
peptides, the latter deriving helix stabilization from side chain-
to-side chain hydrophobic interactions,2 salt bridges,3 disulfide
bridges,4 lactams,5 and metathesis derived hydrocarbon
bridges.6-8 Significantly, hydrocarbon stapling of R-helical
peptides has afforded several compounds that could have
clinical potential, e.g., against cancer.9-11 The structural
and pharmacological effects of ring-closing metathesis in pep-
tides have recently been reviewed.12 Recently, hydrocarbon
stapling has also been successfully applied to 314-helical
β-peptides,13 extending its range of applicability beyond
R-peptides.
The 310-helix, which is defined by intramolecular ifiþ3
H-bonds, is an important structural element in proteins,
peptide antibiotics known as peptaibols,14 and many bio-
logical recognition processes, as well as a postulated inter-
mediate structure in protein folding.15
Over the past decade the predominant water channel in the
mammalian brain, aquaporin-4 (AQP4), has emerged as an
important target for treatment of brain edema after stroke or
trauma.16-19 As part of an ongoing project to design selec-
tive inhibitors of AQP4 we have been interested in side chain-
to-side chain bridges that allow some stabilization of the
310-helical conformation of the Pro138-Gly144 segment of
human AQP4,20 which has been postulated to mediate adhesive
interactions between two AQP4 tetramers.21-23
ferrocenedicarboxylic acid Lys diamides,25 photoinduced
1,3-dipolar cycloaddition,26 metathesis derived hydrocarbon
bridges,20,27,28 and a p-phenylenediacetic acid bridge29 between
two R,R-disubstituted 4-aminopiperidine-4-carboxylic acid
(Api) residueshavebeenreported.However,onlytwostudies28,29
have provided atomic resolution detail of the effect of cycliza-
tion on helix regularity, i.e., on backbone dihedral angles and
H-bond lengths.
In the first X-ray crystallographic study28 of the effect of
side chain-to-side chain cyclization in a 310-helical peptide it
was observed that the backbone is distorted by an ifiþ3
metathesis derived olefinic bridge, resulting in the breakage
of one intramolecular H-bond, thus disrupting the 310-helix.
The p-phenylenediacetic acid bridge on the other hand
appears to afford a highly regular Api/Aib based 310-helix.29
However, R,R-disubstituted amino acids such as Aib and
N-acylated Api are generally hydrophobic and tend to distort
the dihedral angles of neighboring monosubstituted, protei-
nogenic residues away from ideality.24,28,30 In the context
of a helical peptide primarily consisting of proteinogenic
amino acids, monosubstituted residues are expected to be
better tolerated. Hence, new methodology for cross-linking
of monosubstituted residues, which does not significantly
distort the regularity of the 310-helix, is highly desirable. If, at
the same time, the cross-linking results in a more hydrophilic
bridge, thus increasing the aqueous solubility of the stapled
peptide, such a methodology could potentially have broad
utility. In particular, the resulting stapled peptides should be
useful for the study and modulation of biologically impor-
tant recognition processes involving 310-helical peptides and
protein segments.
Examples of ifiþ3 and ifiþ4 side chain-to-side chain cross-
linking in 310-helical peptides by Glu-Lys lactam formation,24
(2) Munoz, V.; Blanco, F. J.; Serrano, L. Nat. Struct. Biol. 1995, 2, 380.
(3) Scholtz, J. M.; Qian, H.; Robbins, V. H.; Baldwin, R. L. Biochemistry
1993, 32, 9668 and references therein.
(4) Jackson, D. Y.; King, D. S.; Chmielewski, J.; Singh, S.; Schultz, P. G.
J. Am. Chem. Soc. 1991, 113, 9391.
(5) Phelan, J. C.; Skelton, N. J.; Braisted, A. C.; McDowell, R. S. J. Am.
Chem. Soc. 1997, 119, 455 and references therein.
(6) Blackwell, H. E.; Grubbs, R. H. Angew. Chem., Int. Ed. 1998, 37, 3281.
(7) Schafmeister, C. E.; Po, J.; Verdine, G. L. J. Am. Chem. Soc. 2000,
122, 5891.
There has been an explosion of interest in click chemistry31
in recent years, exemplified by the highly popular copper(I)-
catalyzed azide-alkyne cycloaddition (CuAAC) reaction.32-35
This reaction has been successfully applied to ifiþ4 side
chain-to-side chain cyclization in an R-helical peptide36,37
and ifiþ3 cyclization in peptoids (peptides composed of
N-substituted glycines).38 The high functional group tolerance
of the CuAAC reaction, the very large dipole moment
(8) Young-Woo, K.; Verdine, G. L. Bioorg. Med. Chem. Lett. 2009, 19,
2533.
(9) Walensky, L. D.; Kung, A. L.; Escher, I.; Malia, T. J.; Barbuto, S.; Wright,
R. D.; Wagner, G.; Verdine, G. L.; Korsmeyer, S. J. Science 2004, 305, 1466.
(10) Moellering, R. E.; Cornejo, M.; Davis, T. N.; Del Bianco, C.; Aster,
J. C.; Blacklow, S. C.; Kung, A. L.; Gilliland, D. G.; Verdine, G. L.; Bradner,
J. E. Nature 2009, 462, 182.
(11) Gavathiotis, E.; Suzuki, M.; Davis, M. L.; Pitter, K.; Bird, G. H.;
Katz, S. G.; Tu, H.-C.; Kim, H.; Cheng, E. H.-Y.; Tjandra, N.; Walensky,
L. D. Nature 2008, 455, 1076.
(12) Jacobsen, Ø.; Klaveness, J.; Rongved, P. Molecules 2010, 15, 6638.
(13) Ebert, M.-O.; Gardiner, J.; Ballet, S.; Abell, A. D.; Seebach, D. Helv.
Chim. Acta 2009, 92, 2643.
(25) Curran, T. P.; Handy, E. L. J. Organomet. Chem. 2009, 694,
902.
(26) Madden, M. M.; Rivera Vera, C. I.; Song, W.; Lin, Q. Chem.
Commun. 2009, No. 37, 5588.
(27) Blackwell, H. E.; Sadowsky, J. D.; Howard, R. J.; Sampson, J. N.;
Chao, J. A.; Steinmetz, W. E.; O’Leary, D. J.; Grubbs, R. H. J. Org. Chem.
2001, 66, 5291.
(28) Boal, A. K.; Guryanov, I.; Moretto, A.; Crisma, M.; Lanni, E. L.;
Toniolo, C.; Grubbs, R. H.; O’Leary, D. J. J. Am. Chem. Soc. 2007, 129,
6986.
(29) Ousaka, N.; Sato, T.; Kuroda, R. J. Am. Chem. Soc. 2008, 130, 463.
(30) Bavoso, A.; Benedetti, E.; Di Blasio, B.; Pavone, V.; Pedone, C.;
Toniolo, C.; Bonora, G. M. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 1988.
(31) Kolb, H. C.; Finn, M. G.; Sharpless, K. B. Angew. Chem., Int. Ed.
2001, 40, 2004.
(32) Tornøe, C. W.; Meldal, M. Peptides 2001, Proceedings of the Amer-
ican Peptide Symposium; American Peptide Society and Kluwer Academic
Publishers: San Diego, 2001; pp 263-264.
(33) Tornøe, C. W.; Christensen, C.; Meldal, M. J. Org. Chem. 2002, 67,
3057.
(34) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, B. K.
Angew. Chem., Int. Ed. 2002, 41, 2596.
(35) Meldal, M.; Tornøe, C. W. Chem. Rev. 2008, 108, 2952.
(36) Cantel, S.; Le Chevalier Isaad, A.; Scrima, M.; Levy, J. J.; DiMarchi,
R. D.; Rovero, P.; Halperin, J. A.; D’Ursi, A. M.; Papini, A. M.; Chorev, M.
J. Org. Chem. 2008, 73, 5663.
(37) Scrima, M.; Le Chevalier-Isaad, A.; Rovero, P.; Papini, A. M.;
Chorev, M.; D’Ursi, A. M. Eur. J. Org. Chem. 2010, No. 3, 446.
(38) Holub, J. M.; Jang, H.; Kirshenbaum, K. Org. Lett. 2007, 9, 3275.
(14) Peptaibiotics: Fungal Peptides Containing R-Dialkyl R-Amino Acids;
€
Toniolo, C., Bruckner, H., Eds.; Wiley-VCH: New York, 2009.
(15) Millhauser, G. L. Biochemistry 1995, 34, 3873.
(16) Manley, G. T.; Fujimura, M.; Ma, T.; Noshita, N.; Fliz, F.; Bollen,
A. W.; Chan, P.; Verkman, A. S. Nat. Med. (N. Y.) 2000, 6, 159.
(17) Vajda, Z.; Pedersen, M.; Fuchtbauer, E.-M.; Wertz, K.; Stodkilde-
Jorgensen, H.; Sulyok, E.; Doczi, T.; Neely, J. D.; Agre, P.; Frokiaer, J.;
Nielsen, S. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 13131.
(18) Amiry-Moghaddam, M. R.; Otsuka, T.; Hurn, P. D.; Traystman,
R. J.; Haug, F. M.; Stanley, S. C.; Adams, M. E.; Neely, J. D.; Agre, P.;
Ottersen, O. P.; Bhardwaj, A. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 2106.
(19) Amiry-Moghaddam, M.; Ottersen, O. P. Nat. Rev. Neurosci. 2003, 4,
991.
(20) Jacobsen, Ø.; Klaveness, J.; Ottersen, O. P.; Amiry-Moghaddam,
M. R.; Rongved, P. Org. Biomol. Chem. 2009, 7, 1599.
(21) Tani, K.; Mitsuma, T.; Hiroaki, Y.; Kamegawa, A.; Nishikawa, K.;
Tanimura, Y.; Fujiyoshi, Y. J. Mol. Biol. 2009, 389, 694.
(22) Engel, A.; Fujiyoshi, Y. Curr. Opin. Struct. Biol. 2008, 18, 229.
(23) Hiroaki, Y.; Tani, K.; Kamegawa, A.; Gyobu, N.; Nishikawa, K.;
Suzuki, H.; Walz, T.; Sasaki, S.; Mitsuoka, K.; Kimura, K.; Mizoguchi, A.;
Fujiyoshi, Y. J. Mol. Biol. 2006, 355, 628.
(24) Schievano, E.; Pagano, K.; Mammi, S.; Peggion, E. Biopolymers
2005, 80, 294.
J. Org. Chem. Vol. 76, No. 5, 2011 1229

Jacobsen et al.
JOCArticle
SCHEME 1. Final Steps of the Synthesis of the Acyclic Peptide 1
Results and Discussion
Peptide Synthesis. The building block NR-Boc-δ-azido-
L-norvaline42 was synthesized in 63% yield from NR-Boc-
L-ornithineusingarecently developed shelf-stable and crystalline
diazo transfer reagent, imidazole-1-sulfonyl azide hydro-
chloride.43 N-Boc-O-propynyl-L-serine44 was synthesized
in 77% yield by a variation of Sugano’s method for synthesis
of N-Boc-O-benzyl-L-serine.45 The octapeptide 1 was assembled
by a segment condensation strategy using standard solution
phase peptide coupling chemistry employing EDC/HOBt in
DMF or CH2Cl2 (Supporting Information). The final steps
involved deprotection of pentapeptide 3 with TFA/CH2Cl2
(1:1(v/v)) and coupling of the resulting trifluoroacetate 4
with Boc-Aib-Aib-Aib-OH28 to afford the octapeptide 1 in
43% yield (Scheme 1).
A recent investigation of Aib oligopeptides by 2D IR
spectroscopy revealed that the onset of 310-helical structure
appears to occur already at the pentapeptide level in CDCl3.46
This, together with our previous success in cyclizing an olefinic
pentapeptide by ring-closing metathesis20 in CH2Cl2 sug-
gested that cyclization by CuAAC might be possible at the
pentapeptide level in CH2Cl2. At high dilution (∼0.15 mM) 3
was cyclized to 5 in 83% yield (Scheme 2) in the presence of
SCHEME 2. Final Steps of the Synthesis of the Cyclic Peptide 2
0.31equiv of the organic-soluble copper(I) complex CuI P(OEt)3,
3
which was synthesized according to a literature procedure.47,48
Dimerization and cyclodimerization are competing processes
and have resulted in relatively low yields of cyclic monomer
in several instances of intramolecular CuAACs, even at high
dilution.35,49,50 The relatively high yield in this case suggests
a high degree of substrate preorganization in CH2Cl2.
Deprotection of 5 with TFA/CH2Cl2 (1:1(v/v)), yielding
the trifluoroacetate 6, followed by segment condensation
with Boc-Aib-Aib-Aib-OH28 afforded octapeptide 2 in
73% yield over two steps (Scheme 2). This synthesis
strategy for the synthesis of 2 has the advantage over
direct cyclization of 1 that it allows the most challenging
synthetic step, namely, the macrocyclization, to be carried
out at the earliest possible stage. Later structural studies
(below) demonstrated that 1 is predominantly 310-helical
in CH2Cl2 solution, suggesting that cyclization would
have been successful also at the octapeptide level. How-
ever, the yield obtained on cyclization of 3 was considered
satisfactory and rendered the alternative strategy super-
fluous and untested in this particular study.
(∼5 D)39 and the relatively high resistance to metabolic
degradation40,41 of the 1,2,3-triazole moiety make 310-helical
peptides with a side chain-to-side chain triazole bridge highly
interesting objects of study.
In this paper we report the installation of an ifiþ3
constraint by side chain-to-side chain CuAAC between two
monosubstituted residues in the context of a 310-helical Aib
rich peptide and examine in detail the effect of cyclization on
helix regularity by X-ray crystallography, NMR, linear IR
and 2D IR spectroscopy. To allow a direct comparison with the
results for the ifiþ3 hydrocarbon bridge, two octapeptides
1 (Scheme 1) and 2 (Scheme 2) with the reactive/cross-linked
residues in the same Aib rich context as the olefinic peptides
of Boal et al.28 were chosen as synthetic targets.
To the best of our knowledge, this study represents the first
X-ray structural investigation of a (R- or 310-) helical peptide
after stapling by CuAAC or with a triazole derived confor-
mational constraint and the first 2D IR structural investigation
of a 310-helical peptide with a conformational constraint
installed. Given the widespread interest in the CuAAC reaction,
it is important to note that this study also affords what appears
to be the first crystal structure of a bifunctional azide-alkyne
compound.
X-ray Crystallography. Peptide 1 (C41H69N11O12, Mw =
908.05) (Scheme 1) crystallized as colorless, plate-shaped
crystals in space group P212121, with unit cell parameters a=
˚
16.239(12), b=18.236(14), c=18.655(14) A, R=90.00ꢀ, β=
90.00ꢀ, γ=90.00ꢀ (orthorhombic crystal system) and Z=4.
The X-ray structure was refined to a final R-factor of 0.068 for
(42) Le Chevalier Isaad, A.; Barbetti, F.; Rovero, P.; D’Ursi, A. M.;
Chelli, M.; Chorev, M.; Papini, A. M. Eur. J. Org. Chem. 2008, No. 31, 5308.
(43) Goddard-Borger, E. D.; Stick, R. V. Org. Lett. 2007, 9, 3797.
(44) ten Brink, H. T.; Rijkers, D. T. S.; Liskamp, R. M. J. J. Org. Chem.
2006, 71, 1817.
(45) Sugano, H.; Miyoshi, M. J. Org. Chem. 1976, 41, 2352.
(46) Maekawa, H.; Formaggio, F.; Toniolo, C.; Ge, N.-H. J. Am. Chem.
Soc. 2008, 130, 6556.
(47) Langille, N. F.; Jamison, T. F. Org. Lett. 2006, 8, 3761.
(48) Nishizawa, Y. Bull. Chem. Soc. Jpn. 1961, 34, 1170.
(49) Jagasia, R.; Holub, J. M.; Bollinger, M.; Kirshenbaum, M.; Finn,
M. G. J. Org. Chem. 2009, 74, 2964.
(39) Bourne, Y.; Kolb, H. C.; Radic, Z.; Sharpless, K. B.; Taylor, P.;
Marchot, P. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 1449.
(40) Whiting, M.; Muldoon, J.; Lin, Y. C.; Silverman, S. M.; Lindstrom,
W.; Olson, A. J.; Kolb, H. C.; Finn, M. G.; Sharpless, K. B.; Elder, J. H.;
Fokin, V. V. Angew. Chem., Int. Ed. 2006, 45, 1435.
(41) Tornøe, C. W.; Sanderson, S. J.; Mottram, J. C.; Coombs, G. H.;
Meldal, M. J. Comb. Chem. 2004, 6, 312.
(50) Angell, Y.; Burgess, K. J. Org. Chem. 2005, 70, 9595.
1230 J. Org. Chem. Vol. 76, No. 5, 2011

Jacobsen et al.
JOCArticle
FIGURE 3. Thermal ellipsoid plot of the X-ray crystal structure of
the cyclic peptide 2 at the 50% probability level with intramolecular
H-bonds indicated. Apolar hydrogens have been omitted for clarity.
FIGURE 1. Thermal ellipsoid plot of the X-ray crystal structure of
the acyclic peptide 1 at the 50% probability level with intramole-
cular H-bonds indicated. Apolar hydrogens have been omitted for
clarity. The major backbone conformation at Aib6 is shown (see
Supporting Information for details).
β-turn52,53 together with residue 6 in its minor conformation
(-72.56ꢀ, -7.39ꢀ). In contrast to the acyclic olefinic analogue,
where residue 8 is in the RR conformation,28 the C-terminal
residue in 1 adopts a left-handed polyproline II (PIIL) con-
formation with (j, ψ) = (-57.83ꢀ, 159.12ꢀ). This fits well,
however, into the empirical pattern found in a recent survey
of nonhelical conformations of Aib residues in peptides. In a
database of 143 crystal structures of Aib-containing helices
with >3 residues with a C-terminal Aib, 86.5% of the C-
terminal residues adopted the opposite helix sense than the
rest of the molecule and 20.3% of these fell within the PIIL
region.54
The intramolecular H-bond lengths (dCdO HN) vary
˚
3 3 3
between 2.115 A (Aib1fAib4) and 2.341 A (Aib3fAib6),
˚
˚
with mean 2.193 A and standard deviation 0.088 A. These
˚
values are very similar to the ones found for the acyclic
olefinic analogue (2.210 A and 0.115 A).28 For both of these
acyclic peptides the same pattern of H-bond length variations is
observed. For both acyclic peptides, the longest H-bonds are
between pairs of Aib residues on the N-terminal (residues 3
and 6) and on the C-terminal (residues 5 and 8) sides of the
monosubstituted residues, respectively.
˚
˚
FIGURE 2. Partial Ramachandran plot for residues 1-7 of 1 and 2.
Both conformations of Aib6 in the crystal structure of 1 are indi-
cated (Res. 6: major conformation; Res. 60: minor conformation).
See Supporting Information for details and complete Ramachandran
plots.
The carbonyl group of Aib7 forms an intermolecular H-bond
to the carbamate NH of the Boc group, but there is no inter-
molecular peptide-peptide H-bond to the carbonyl of the methyl
ester as is often seen in structures of Aib rich peptides.54
The cyclic peptide2 (C41H69N11O12, Mw=908.05) (Scheme 2)
crystallized as colorless, plate-shaped crystals in space group
C2, with unit cell parameters a=36.417(12), b=13.382(5),
data obtained for a very small crystal (0.100 mmꢀ0.010 mmꢀ
0.009 mm). The peptide forms a fully developed right handed
310-helix with all possible intramolecular ifiþ3 H-bonds
present, including between the tert-butoxycarbonyl (Boc)
group and the amide NH of Aib3 (Figure 1).
With the exceptions of residues 4, 6, 7, and 8 the confor-
mations of all remaining residues fall into the 310-helical
region of (j, ψ)-space, with the mean absolute deviations
from the ideal (i.e., average observed in peptides) 310-helical
angles of (-57ꢀ, -30ꢀ)51 being 3.68ꢀ and 4.48ꢀ, respectively
(Figure 2).
However, the dihedral angles of the two chiral, monosub-
stituted residues 4 and 7 deviate significantly from the ideal
310-helical angles, with [(|Δj|, |Δψ|) = (13.38ꢀ, 17.90ꢀ) and
(36.55ꢀ, 30.96ꢀ)] respectively. Similarly large deviations from
ideality, albeit slightly smaller for residue 4, were observed
for the acyclic olefinic peptide of Boal et al.28 Here, residue 7,
for which (j, ψ)7 = (-93.55ꢀ, 0.96ꢀ), forms part of a type-I
˚
c = 11.873(4) A, R = 90.00ꢀ, β = 102.360(4)ꢀ, γ = 90.00ꢀ
(monoclinic crystal system) and Z = 4. The X-ray structure
was refined to a final R-factor of 0.039, which is unusually
low for a molecule of this size.
Like the acyclic peptide 1 the cyclic octapeptide 2 forms
a fully developed right handed 310-helix with all possible
ifiþ3 intramolecular H-bonds present (Figure 3).
Whereas significant deviations from an ideal 310-helix with
respect to individual dihedral angles were observed in the
crystal structure of peptide 1, the structure of peptide 2 re-
presents a strikingly ideal 310-helix from residue 1-7. The
average (j, ψ)-angles are (-54.96ꢀ, -30.17ꢀ), deviating a
mere 2.04ꢀ and 0.17ꢀ from ideality, making peptide 2 the
most perfect cross-linked 310-helix to date (Table 1).
Importantly, the triazole bridge appears to strongly enforce
a 310-helical conformation for residues 4, 6, and 7, effectively
(51) Toniolo, C.; Benedetti, E. Trends Biochem. Sci. 1991, 16, 350.
(52) Chou, P. Y.; Fasman, G. D. J. Mol. Biol. 1977, 115, 135.
(53) Venkatachalam, C. M. Biopolymers 1968, 6, 1425.
(54) Aravinda, S.; Shamala, N.; Balaram, P. Chem. Biodivers. 2008, 5,
1238.
J. Org. Chem. Vol. 76, No. 5, 2011 1231

Jacobsen et al.
JOCArticle
TABLE 1. Mean (u, ψ)-Angles with Standard Deviations
for the Cross-linked 310-Helical Peptides for Which Crystallographic
Data Have Been Reporteda
TABLE 2. Mean O H Distances with Standard Deviations for the
3 3 3
Intramolecular H-bonds in the Crystal Structures of 1 and 2a
compound
O
H
σ
O
Hb
σb
3 3 3
3 3 3
compound
j
σj
ψ
σψ
1
2
2.193
2.118
2.087
2.210
2.384
2.335
0.088
0.075
0.084
0.115
0.760
0.195
2.164
2.102
2.058
0.056
0.071
0.051
ideal 310-helix
1 (major)
1 (minor)
-57
-30
-58.75
-64.22
-54.96
-62.32
-68.35
-59.99
-54.20
-55.52
-54.94
18.68
15.61
6.79
15.23
24.04
14.94
4.35
-27.93
-22.38
-30.17
-24.29
-16.94
-27.52
-28.97
-26.78
-27.73
16.34
15.63
5.87
10.27
24.50
15.59
8.07
2 (scaled)
acyclic olefinic28
cyclic olefinic28
cyclic hydrogen28
2
2.075
2.257
0.357
0.034
acyclic olefinic28
cyclic olefinic28
cyclic hydrogen28
acyclic api29
cyclic api (mol 1)29
cyclic api (mol 2)29
aPositional parameters for amide H atoms were refined for 2 only. To
facilitate comparison with 1, scaled values are included for 2 after
˚
normalization of all N-H bonds to 0.880 A (i.e., H atoms are moved
2.61
3.83
6.11
7.20
along the covalent bond vectors so as to make the N-H distances equal
b
˚
to 0.880 A, the fixed N-H distance used in the refinement of 1). The
aThe dihedral angles for residue 8 have been omitted for all peptides.
longest H-bond, i.e., between residues 4 and 7 for 2 and between 3 and 6
for 1 and the hydrocarbon stapled peptides, has been omitted.
Boc group in capped peptides or with solvent.54 Interest-
ingly, in the structure of 2 the dihedral angles of (46.80ꢀ,
49.93ꢀ) allow two peptide molecules to contact each other in
a tail-to-tail fashion forming two bifurcated nonclassical
C-H OdC H-bonds between the triazole hydrogen and
3 3 3
the carbonyl groups of residues 7 and 8 (Figure 4). This
underlines a potential added advantage of a triazole in a helix
stabilizing bridge, namely, its ability to make useful contacts
to peptides/proteins. The triazole is approximately coplanar
with the bifurcated H-bond. Interestingly, both of the Api
peptides,29 which also are highly 310-helical at residue 7, have
C-terminal residues with conformations very similar to that
of the C-terminal residue of 2.
The overall similarity to an ideal 310-helix is also reflected
in significantly shorter intramolecular H-bonds compared to
1 and the hydrocarbon stapled analogues (Table 2). The data
for 1 refer to the structure with the alkyne side chain in its
major orientation and the backbone in its major conforma-
tion (see Supporting Information for details).
FIGURE 4. The triazole proton of one molecule forms a bifurcated
nonclassical H-bond with residues 7 and 8 in another molecule.
Only one of the two bifurcated H-bonds between two peptide mole-
cules is shown. The remainder of the molecules has been omitted for
clarity.
removing these as outliers in the Ramachandran plot (Figure 2).
The deviations from ideality for these residues in peptide 2
are (|Δj|, |Δψ|) = (2.20ꢀ, 4.88ꢀ), (0.51ꢀ, 0.78ꢀ), and (10.93ꢀ,
9.25ꢀ), respectively, dramatically improved relative to 1. This
is opposite to the trend observed in the structures of the cyclic
olefinic peptide and its hydrogenated analogue, where cycli-
zation appeared to cause larger or unchanged deviations,
with (j, ψ)4=(-96.57ꢀ, 19.84ꢀ) and (-66.64ꢀ, -22.33ꢀ) and
(j, ψ)7 = (-108.91ꢀ, 13.69ꢀ) and (-90.06ꢀ, 3.05ꢀ), respec-
tively.28 The sizes of the macrocycles and the rigid elements
(triazole and carbon-carbon double bond) are important
parameters that could be partially responsible for the differ-
ent effects of the two staples on the backbone conformation.
The macrocycle in 2 is slightly larger (19 atoms along the
shortest path) than in the corresponding olefinic peptide
(18 atoms). In addition, the rigid element in the CuAAC-
derived bridge is also slightly larger than a carbon-carbon
double bond.
The residues of the p-phenylene diacetic acid cross-linked
Api/Aib peptide generally have close to ideal dihedral angles
from residue 1 through to 7, but residue 4 (next to the first
Api residue) has a slightly distorted ψ-angle (-22.69ꢀ), and
the deviation from the ideal ψ-angle for residue 6 (-15.26ꢀ,
|Δψ|=14.74ꢀ) is larger than any j/ψ-deviation from ideality
for the first 7 residues of 2.29
The C-terminal residue of 2 adopts an RL conformation, in
other words, the opposite helix sense as the rest of the molecule.
This is statistically the most common conformation for a
C-terminal Aib in Aib rich helices with >3 residues and is
often due to head-to-tail intermolecular interactions with the
As expected the longest H-bond observed in the structure
˚
˚
of 2 is between residues 4 and 7 (2.232 A, Δ2-1 =þ0.103 A),
whose conformations change the most as they are pulled in
toward more ideal 310-helical dihedral angles. However, all
the remaining 5 intramolecular H-bonds are shorter in 2
than in 1. The largest improvements are seen for BocfAib3
˚
˚
(Δ2-1 = -0.172 A), Aib3fAib6 (Δ2-1 = -0.193 A) and
˚
Aib5fAib8 (Δ2-1 = -0.184 A). Interestingly, the Aib3f
Aib6 H-bond is the one stretched the most in the cyclic
hydrocarbon stapled peptides relative to their acyclic precursor
˚
˚
(Δcyclic-acyclic=þ0.386 A and Δcyclic-acyclic=þ1.585 A) and
28
˚
is in fact broken in the cyclic olefinic peptide (dO3 3 3 H=3.927A).
NMR Spectroscopy. While the X-ray crystallographic
studies provided atomic resolution detail of the conformations
of 1 and 2 in the solid state, these may not be very repre-
sentative of their solution phase structures. A careful inves-
tigation of the solution phase structures of 1 and 2 in the
polar aprotic solvent CD2Cl2, including a detailed comparison
with the corresponding crystal structures, was therefore
undertaken using various NMR spectroscopic techniques.
The 1H NMR spectra of 1 and 2 were found to be concen-
tration dependent (Figure 5). In particular, the NH(Aib1)
(Δδ = 0.37 ppm) and NH(Aib2) (Δδ = 0.20 ppm) reso-
nances display remarkable downfield shifts when moving
from a relatively dilute solution (9.9 mM) to a more con-
centrated solution (51 mM). The chemical shifts of the
remaining protons, with the notable exception of one of
the methyl groups, are much less concentration dependent
1232 J. Org. Chem. Vol. 76, No. 5, 2011

Jacobsen et al.
JOCArticle
not discernible, possibly due to overlap with the corresponding
diagonal peaks) and the only possible CrH(i)fNH(iþ2) ROE
(Supporting Information), which are indicative of 310- or
R-helical peptides.55-57 However, the presence of the long-
range CRH(i)fNH(iþ3) ROE also typical of 310-helical
peptides could not be established with certainty as the CRH-
(azidonorVal) peak is overlapping with the CRH(propSer)-
CH2O peak, which has an ROE to NH(propSer) (Support-
ing Information).
Gratifyingly, ROEs characteristic of a 310-helical structure
was also found in the 2D ROESY spectrum of 2, including
all possible short-range NH(i)fNH(iþ1), medium range
CRH(i)fNH(iþ2), and long-range CRH(i)fNH(iþ3) ROEs
(Figure 6).
A commonly used upper distance constraint for a weak
ROE is on the order of 5.0 A.58 In order to determine more
˚
precisely the extent to which the solution structures of 1 and 2
mimic their corresponding crystal structures, all proton pairs
˚
closer than 5.0 A in the crystal structures, which should give
FIGURE 5. Overlay of the NH region of the 1H NMR spectra of 2
in dilute (9.9 mM, 296 K, 400 MHz, black) and concentrated solution
(51 mM, 298 K, 300 MHz, red). The NH(Aib1) and NH(Aib2)
resonances are significantly downfield shifted in the more concen-
trated solution.
rise to ROEs if the crystal structure is preserved in solution,
were identified. In the case of 1 the distance measurements
were performed on the structure with the alkyne side chain
and backbone at Aib6 in their major conformations. A careful
analysis of the experimental 2D ROESY spectra recorded in
dilute CD2Cl2 solution for the presence or absence of these
ROEs was undertaken. The results are summarized in Figure 7.
For 1 the general correlation between the predictions based
on the crystal structure and the experimental solution phase
spectrum is quite good, with 18 (45%) predicted ROEs present
and 11 (27.5%) absent. However, the presence/absence of
an additional 11 (27.5%) ROEs could not unambiguously be
decided because of overlap with other peaks. Importantly,
the correlation appears to be even better for the cyclic peptide 2,
with 27 (63%) predicted ROEs present in the experimental
spectrum, only 3 (7%) undecidable, and 13 (30%) absent.
For both compounds it should be added that a large fraction
of the correlations that were not found experimentally
than δNH(Aib1) and δNH(Aib2). Remarkably, the concentration
dependence of the H NMR spectra, the selectivity of the
1
effect on δNH(Aib1) and δNH(Aib2) and the magnitude and sign
of the chemical shift changes may all be explained by invok-
ing (concentration-independent) predominantly 310-helical
backbone structures for 1 and 2 in CD2Cl2. It should be
noted that NH(Aib1) and NH(Aib2) are the only amide NHs
which do not participate in intramolecular H-bonding if the
peptides adopt 310-helical structures (Figures 1 and 3) and
the only NHs which thus may engage in intermolecular
H-bonding without compromising the overall 310-helical struc-
tures of the peptides. We believe that the large and selective
downfield shifts of δNH(Aib1) and δNH(Aib2) could be explained
by hypothesizing that NH(Aib1) and NH(Aib2) form inter-
molecular H-bonds in concentrated CD2Cl2 solution, while the
other NHs remain intramolecularly H-bonded. This would
imply the presence of a monomer-dimer equilibrium, which
could explain the concentration dependence of the 1H NMR
spectra.
˚
corresponds to distances between 4.0 and 5.0 A in the crystal
structure and are expected to be weak and potentially diffi-
cult to distinguish from background noise in dilute solutions.
An examination of the individual backbone-to-backbone
ROEs reveals the presence of all possible NH(i)fNH(iþ1)
ROEs, but none of the predicted NH(i)fNH(iþ2) ROEs. How-
ever, the latter finding should not be considered as strong evi-
dence against a 310-helical structure, as the NH(i)fNH(iþ2)
1
The concentration dependence of the H NMR spectra
prompted the question of whether the structures in dilute
solutions or in concentrated solutions are more similar to the
crystal structures. The fact that NH(Aib1) and NH(Aib2) are
not H-bonded to other peptide molecules in the crystal
structures suggests that the structures in dilute solution are
likely to be most similar to the solid state structures. It should
be added that this concentration regime is also of greater
biomedical relevance. The cyclic peptide 2 does form a dimer
in the crystal, but this does not involve H-bonding to NH-
(Aib1) or NH(Aib2) (Figure 4).
At any rate, caution is required when comparing solution
phase structures in CD2Cl2 with the crystal structures, given
the fact that the crystals were grown by slow evaporation of
EtOAc (1) and (effectively) iPrOH solutions (2) rather than
CD2Cl2 solutions.
˚
distance is approximately 4.5 A in a 310-helix, giving rise to a
weak ROE at best, but can be much shorter in nonhelical
peptides, as can be appreciated from a molecular model.
When it comes to the predicted CRHfNH type ROEs the
accordance between theory and experiment is very good for
both peptides.
Some of the side chain-to-backbone ROEs, i.e., side chain-
to-NH and side chain-to-CRH ROEs, are quite informative,
as their presence/absence appear to be particularly sensitive
to small changes in backbone structure away from the
(55) Rai, R.; Aravinda, S.; Kanagarajadurai, K.; Raghothama, S.; Shamala,
N.; Balaram, P. J. Am. Chem. Soc. 2006, 128, 7916.
(56) Williams, D. H.; Fleming, I. Spectroscopic Methods in Organic
Chemistry, 5th ed.; McGraw-Hill: New York, 1995; pp 140.
To extract more detailed structural information 2D ROESY
spectra were recorded in dilute CD2Cl2 solution (11 mM,
296 K, Figure 6 and Supporting Information). The 2D
ROESY spectrum of 1 demonstrated the presence of 5 out
of 7 possible NH(i)fNH(iþ1) ROEs (the remaining two were
€
(57) Wuthrich, K. NMR of Proteins and Nucleic Acids; John Wiley &
Sons: New York, 1986.
(58) Ma, M. T.; Hoang, H. N.; Scully, C. C. G.; Appleton, T. G.; Fairlie,
D. P. J. Am. Chem. Soc. 2009, 131, 4505.
J. Org. Chem. Vol. 76, No. 5, 2011 1233

Jacobsen et al.
JOCArticle
FIGURE 6. Partial 2D ROESY spectrum of 2 displaying the only possible medium range CRH(i)fNH(iþ2) (black circle) and long-range
CRH(i)fNH(iþ3) ROE (red circle). Note that not all ROEs referred to in Figure 7 are visible at the chosen spectrum intensity.
FIGURE 7. ROEs observed in the 2D ROESY spectra of 1 and 2 in dilute CD2Cl2 solution and comparison of the solution structures of 1 and 2
in CD2Cl2 with their respective crystal structures. Color code: Green signifies an ROE that should be observed if the crystal structure is
preserved in solution and is found experimentally. Yellow signifies an ROE that should be observed if the crystal structure is preserved in
solution, but cannot unambiguously be identified experimentally because of overlap with other ROEs, artifact peaks (e.g., TOCSY peaks), or
diagonal peaks under the given experimental conditions. Red signifies an ROE that should be observed if the crystal structure is preserved in
solution, but is not found in the experimental spectra or could not be distinguished from background noise. White fields correspond to distances
˚
greater than 5.0 A in the crystal structure. Blue signifies the presence of a correlation in the experimental spectrum that was not predicted by the
crystal structure. For each correlation the left-hand columns refer to peptide 1 and the right-hand columns to peptide 2. Note that CCH is the alkyne
proton in 1 and the triazole proton in 2. All methyl groups and the Boc group have been omitted.
1234 J. Org. Chem. Vol. 76, No. 5, 2011

Jacobsen et al.
JOCArticle
310-helical crystal structures. This is especially the case for the
CδH2(i)fNH(i) interactions in the acyclic peptide 1, and to a
lesser degree the CβH2(i)fNH(i) interactions. In the 310-
helical crystal structure the torsional angle Cβ-CR-N-H is
small and positions the NH proton and the side chain on the
same side of the backbone with close to minimal distance
δ
˚
separation. The distance C H2(4)fNH(4) is 4.20 A in the
crystal structure, making this a quite sensitive test for local
helical structure. The absence of an observable CδH2-
(4)fNH(4) correlation for 1 could indicate that the peptide
backbone enjoys substantial conformational freedom, atleast
locally. The CδH2(4)fNH(4) ROE is present in the 2D
ROESY spectrum of 2, for which the corresponding crystal
˚
structure distance is 4.87 A. This is an important difference
between the two compounds. The CδH2(7)fNH(7) correlation
˚
(3.56 A in the crystal structure) is, however, present in 1. The
δ
˚
shorter distance compared with C H2(4)fNH(4) (4.20 A)
1
FIGURE 8. Methyl resonances in the H NMR spectra of 1 (red)
and 2 (blue) (c1 = 10 mM, c2 = 9.9 mM, T1 = 295 K, T2 = 296 K).
The greater dispersion of the resonances for the cyclic peptide may
be taken as evidence for a more structured peptide in solution (less
conformational averaging).
suggests that it is less sensitive to backbone changes away
from the crystal structure than the CδH2(4)fNH(4) ROE. In
contrast to 1, the CδH2(7)fNH(7) is absent from the 2D
ROESY spectrum of the cyclic peptide 2, but this does not
necessarily imply a large deviation from the crystal structure,
˚
as the crystal structure distance is 4.69 A, placing it close to
the detection limit already in a perfect 310-helix.
TABLE 3. Chemical Shifts of NH Resonances in Dilute (c1 = 10 mM,
c2 = 9.9 mM, T1 = 295 K, T2 = 296 K) CD2Cl2 Solution
residue
δacyclic
δcyclic
Δδ
There are no observable differences between 1 and 2 regard-
ing the presence/absence of CβH2fNH type ROEs. The
presence of CβH2(i)fNH(i) interactions in both peptides
should be noted in support of a predominantly 310-helical
backbone structure. However, larger backbone conforma-
tional changes are required to completely abolish these than
is the case for the CδH2(i)fNH(i) ROEs discussed earlier.
The absence of predicted NH(i)fCβH2(iþ1) and presence
of CβH2(i)fNH(iþ1) correlations in both peptides are less
informative.
Aib1
Aib2
Aib3
azidonorVal
Aib5
Aib6
5.21
6.69
7.92
7.80
7.84
7.29
7.55
7.56
5.24
6.72
7.91
7.77
7.88
7.00
7.40
7.53
þ0.03
þ0.03
-0.01
-0.03
þ0.04
-0.29
-0.15
-0.03
propSer
Aib8
a smaller degree of conformational mobility/averaging on
the NMR time scale in the former case and thus provides
evidence for a more stable 310-helix.
The absence of CRH(4)fCβH2(7) interactions in the cyclic
˚
peptide 2 is a bit surprising (distance 4.21 A in the crystal).
For the acyclic peptide 1 the presence or absence of this ROE
cannot be decided.
The chemical shifts of the NH resonances of the cyclic pep-
tide 2 are generally very similar to those of 1 (Table 3) and
demonstrate that the H-bond lengths and angles of the two
peptides most likely are relatively similar in CD2Cl2 solution.
The only exceptions are NH(Aib6) and NH(propSer), which
are shifted 0.29 and 0.15 ppm upfield for 2. The 1D and 2D
NMR data taken together demonstrate that ifiþ3 side chain-
to-side chain macrocyclization by CuAAC results in a close-
to-ideal and conformationally stable 310-helical peptide.
Measurement and Simulation of 2D IR Spectra. Conven-
tional linear IR response of the amide-I mode is widely used
to obtain structural information of polypeptides.59 Going
beyond 1D, 2D IR spectroscopy measures nonlinear response
of the mode, which has higher sensitivity to the underlying
biomolecular structure.60-62 In this study we measured FT
IR and 2D IR spectra of 1 and 2 in CH2Cl2 to obtain insights
into the backbone conformation of these peptides in a polar
aprotic solvent. Also, the spectral profiles were simulated
based on crystal structures established by the X-ray diffraction
analysis. Details of our 2D IR spectrometer, data collection,
and calculation protocol have been described previously.46,63,64
Side chain-to-side chain correlations may also be informa-
tive, especially for the acyclic peptide, for which the side
chains of residues 4 and 7 are likely to be too far apart to give
rise to observable ROEs in many nonhelical conformations.
Most of these are absent in both peptides, including the
expected CβγH2(4)fCβH2(7) and CδH2(4)fCβH2(7) ROEs
for 2 and the CβγH2(4)fCδH2(7) ROE for 1.
Finally, a very weak ROE that was not predicted on the
basis of the X-ray structure was observed between CδH2(4)
and CδH2(7) for 1. In general, the absence of many extra peaks
not predicted by the crystal structure provides further evidence
that the solution structure closely mimics the crystal structure.
However, the presence of the unexpected CδH2(4)fCδH2(7)
ROE may nevertheless be taken as evidence in favor of a
310-helical structure, because it means that the side chains of
residues 4 and 7 are on the same side of the molecule, which is
what one would expect for a 310-helical peptide.
As a final note, the 2D ROESY spectra display ROEs
characteristic of 310-helices independently of concentration.
Hence, the possible presence of a monomer-dimer equilib-
rium is not expected to constitute a major complication for
the structural analysis of 1 and 2 in solution.
(59) Krimm, S.; Bandekar, J. Adv. Protein Chem. 1986, 38, 181.
(60) Zhuang, W.; Hayashi, T.; Mukamel, S. Angew. Chem., Int. Ed. 2009,
48, 3750.
(61) Kim, Y. S.; Hochstrasser, R. M. J. Phys. Chem. B 2009, 113, 8231.
(62) Hunt, N. T. Chem. Soc. Rev. 2009, 38, 1837.
The signal dispersion of the methyl resonances is signifi-
cantly greater for 2 than for 1 (Figure 8). This indicates
J. Org. Chem. Vol. 76, No. 5, 2011 1235

Jacobsen et al.
JOCArticle
FIGURE 9. (a, b) Measured (black solid) and simulated (red dashed) linear IR spectra for 1 and 2. (c, d) 2D IR cross-peak patterns in CH2Cl2
observed under the double-crossed polarization configuration. (e, f) Simulated cross-peak patterns based on the peptide backbone confor-
mations of 1 and 2 in the crystal state. Shown on the right are the measured and simulated absorptive and the real parts of the rephasing and
nonrephasing 2D IR spectra under the perpendicular polarization configuration.
The specifics relevant to the data discussed below are given in
the Supporting Information.
nodal lines between the 0-1 and 1-2 transitions are parallel
and perpendicular to the diagonal, respectively. These 2D IR
spectra look almost indistinguishable between 1 and 2.
The amide-I 2D IR cross-peak pattern obtained under the
double-crossed polarization can much more sensitively dis-
tinguish subtle structural differences, for example, between
310- and R-helices.63,64 In general, the experimental and simu-
lated 2D profile exhibits a doublet pattern for the former and
a multiple-peak pattern for the latter.64,65 Figure 9c and d
presents the absolute magnitude cross-peak patterns of 1 and 2,
respectively. A doublet clearly shows up in the amide-I region.
Weak cross-peaks between the amide-I and the urethane
CdO modes are also observed. The doublet pattern of 1 is
similar to that of 2, but some subtle differences can also be
noticed. For 2, the line shape and relative intensity of the
lower diagonal peak to the upper peak (0.68) are very close to
those observed for other 310-helical peptides with Aib and
(RMe)Val residues.63,64 For 1, the two peaks in the doublet
are more elongated along the diagonal, merged into each
other at a lower frequency, and their intensity ratio is 0.45.
This result suggests that 1 and 2 are both 310-helical in the
solution but their structures are slightly different. It is con-
ceivable that 1 may be more disordered than 2 because it
lacks the side chain-to-side chain CuAAC constraint. Our
previous theoretical study shows that the rephasing cross-
peak pattern differs from a doublet with the appearance
The top panels in Figure 9a and b show the measured FT
IR spectra (black solid) of 1 and 2, respectively, in CH2Cl2
(∼6 mM, a thickness of 180 μm). The spectra were normal-
ized by the peak absorbance of the methyl ester CdO band at
1738 cm-1 after subtracting the solvent spectrum. We assign
the small band at 1701 cm-1 to the urethane CdO of the
N-terminus Boc group and the broad band at 1662 cm-1 to
the amide-I modes. The line shapes of the amide-I bands of
the two peptides are slightly different, more rounded at the
peak for 2 but pointy for 1. The full-width-half-maximum are
27.7 and 30.8 cm-1, respectively. It is not straightforward,
however, to infer from the linear spectra whether 1 and 2 in
CH2Cl2 maintain the same 310-helical conformation as ob-
served in the crystal states, or if their structures have changed
to other conformations, such as R-helix and PIIL
.
The absorptive and the real parts of rephasing and non-
rephasing 2D IR spectra of 1 and 2 measured under the ÆY, Y,
Z, Zæ polarization configuration are shown on the right in
Figure 9. The three positive peaks along the diagonal line of
the absorptive spectrum correspond to the 0-1 transitions,
whereas the three negative peaks are the 1-2 transitions,
which are anharmonically shifted from the 0-1 in the ωt
direction. For the rephasing and nonrephasing spectra, the
(63) Maekawa, H.; Toniolo, C.; Broxterman, Q. B.; Ge, N.-H. J. Phys.
Chem. B 2007, 111, 3222.
(64) Maekawa, H.; Toniolo, C.; Moretto, A.; Broxterman, Q. B.; Ge,
N.-H. J. Phys. Chem. B 2006, 110, 5834.
(65) Sengupta, N.; Maekawa, H.; Zhuang, W.; Toniolo, C.; Mukamel, S.;
Tobias, D. J.; Ge, N.-H. J. Phys. Chem. B 2009, 113, 12037.
1236 J. Org. Chem. Vol. 76, No. 5, 2011

Jacobsen et al.
JOCArticle
of extra features as the peptide conformation increasingly
deviates from an ideal 310-helix.65 The absence of extra features
suggests that the structural difference between 1 and 2 is small
in CH2Cl2 despite the quite different dihedral angles of the
residues 4, 6, and 7 observed in the crystal state.
chain-to-side chain cross-linking methodology may have signi-
ficant utility applied to peptides and peptidomimetics of
interest in chemical biology and biomedicine, in particular
to synthetic analogues of the Pro138-Gly144 segment of
human AQP4.20
To further address these points, we performed model
calculations to examine how similar or different the 2D IR
spectral patterns would be if the crystal structures are pre-
served in solution. Figure 9e and f presents the simulated 2D
IR cross-peak pattern using an ensemble of peptide structures
with the backbone dihedral angles in Gaussian distributions
centered at the crystal structures of 1 and 2, respectively. The
simulated 2D IR pattern of 2 shows a clear amide-I doublet
along with some cross-peaks between the amide-I and the
capping CdOs. The linear IR spectrum (Figure 9b, red dashed)
and 2D absorptive and the real parts of the rephasing and
nonrephasing spectra (Figure 9, bottom panels on the right)
were also calculated using the same parameters, and the
agreement with the experimental spectrum is quite good.
This suggests that the solution structure of 2 is close to that of
an ideal 310-helix. In contrast, the crystal structure of 1 gave
rise to quite different linear and 2D IR spectra from those of 2.
The calculated linear spectrum of the major backbone con-
former consists of two overlapping, broad amide-I bands with
the lower frequency band having a stronger peak intensity.
The linear spectrum of the minor backbone conformer also
exhibits two bands with an opposite trend in the peak intensity.
The population weighted spectrum (Figure 9a) shows hints
of overlapping features. The cross-peak pattern in Figure 9e
exhibits a more spread doublet with additional shoulders
than that characteristic of an ideal 310-helix. Note that the
simulated cross-peak pattern is still different from the ideal
R-helix conformation [(φ, ψ) = (-63ꢀ, -42ꢀ)] we obtained
previously.63,64 Also, the calculated 2D absorptive spectrum
of 1 is more elongated along the diagonal than that of 2. The
experimental and simulation results indicate that the differ-
ent backbone conformations of 1 and 2 in the crystal state are
no longer preserved in CH2Cl2.
Experimental Section
Full experimental details for the crystallization of 1 and 2,
acquisition of NMR, FT IR and 2D IR data, model calculations
of linear and 2D IR spectra and syntheses of intermediates may
be found in the available Supporting Information. The crystal
structures of 1 and 2 have been deposited at the Cambridge
Crystallographic Data Centre (accession codes CCDC 770131
and CCDC 770132).
N-tert-Butoxycarbonyl r,r-Dimethylglycyl r,r-Dimethylglycyl
r,r-Dimethylglycyl δ-azido-L-norvalyl r,r-Dimethylglycyl r,r-
Dimethylglycyl O-Propynyl-L-seryl r,r-Dimethylglycyl Methyl
Ester 1. δ-Azido-L-norvalyl R,R-dimethylglycyl R,R-dimethylglycyl
O-propynyl-L-seryl R,R-dimethylglycyl methyl ester trifluoroacetate
4 (0.551 g, 0.827 mmol) and N-tert-butoxycarbonyl R,R-dimethyl-
glycyl R,R-dimethylglycyl R,R-dimethylglycine28 (0.309 g, 0.827
mmol) (see Supporting Information for details on preparation)
were suspended in CH2Cl2 (5 mL), and a solution of N,N-diiso-
propylethylamine (0.107 g, 0.828 mmol) in CH2Cl2 (7 mL) added.
HOBt hydrate (0.127 g, 0.829 mmol) and then EDC hydrochloride
(0.174 g, 0.908 mmol) were added together with more CH2Cl2
(5 mL) at room temperature. The reaction mixture was stirred for
45 h at room temperature before being diluted with CH2Cl2 (65 mL).
The solution was washed with 5% (w/w) citric acid monohydrate
solution (3ꢀ35 mL), 7.5% (w/w) K2CO3 solution (3ꢀ35 mL), and
saturated brine (35 mL). The solution was dried with anhydrous
MgSO4 and the solvent was evaporated affording a white solid
(0.595 g). The solid (0.573 g) was purified by flash column chro-
matography (eluent CH2Cl2/acetone (3:1)) affording the title com-
pound as a white solid (0.323 g, 43%): 1H NMR (400 MHz,
CD2Cl2, 10 mM, 295 K) δ 7.92 (s, 1H, NH(Aib3)), 7.84 (s, 1H,
NH(Aib5)), 7.80(d,J=5.1 Hz, 1H, NH(azidonorVal)), 7.56(s,1H,
NH(Aib8)), 7.55 (d, J = 8.0 Hz, 1H, NH(propSer)), 7.29 (s, 1H,
NH(Aib6)), 6.69 (s, 1H, NH(Aib2)), 5.21 (s, 1H, NH(Aib1)), 4.44
(td, J=8.3, 3.5 Hz, 1H, CRH(propSer)), 4.23 (dd, J=15.6, 2.4 Hz,
1H, CHHCCH), 4.18 (dd, J = 16.0, 2.4 Hz, 1H, CHHCCH),
3.97-3.80 (m, 3H, CRH(azidonorVal)/CH2O), 3.65 (s, 3H, OCH3),
3.33 (t, J = 6.8 Hz, 2H, CH2N3), 2.48 (t, J = 2.3 Hz, 1H, CCH),
2.02-1.62 (m, 4H, CH2CH2CH2N3), 1.61-1.36 (m, 45H, CH3/
(CH3)3); 13C NMR (75 MHz, CD2Cl2, 8.1 mM, 298 K) δ 177.2,
176.1, 175.5, 175.5, 175.4, 175.2, 173.9, 170.0, 156.2, 82.1, 80.7, 74.6,
70.0, 58.8, 57.5, 57.4, 57.4, 57.4, 57.1, 57.0, 56.3, 54.7, 52.4, 51.8,
28.5, 28.5, 28.0, 28.0, 27.7, 27.5, 27.4, 26.5, 25.5, 25.2, 23.8, 23.5,
23.3, 23.3, 23.1; HRMS (m/z) [M þ Na]þ calcd for C41H69N11O12-
Na, 930.5024, found 930.5017. Anal. Calcd for C41H69N11O12: C,
54.23; H, 7.66; N, 16.97. Found: C, 54.2; H, 7.6; N, 16.6.
The results from our linear and 2D IR experiments and
simulations are consistent with the findings from 1D and 2D
NMR measurements. Both peptides remain 310-helical in solu-
tion. Comparing to 2, the solution structure of the acyclic
peptide 1 is more flexible and/or disordered. It also exhibits
larger deviations from the crystal structure. The difference in
their behavior can be attributed to the effect of the side chain-
to-side chain constraint on peptide conformation.
Methyl 2-Methyl-2-((5S,14S)-8,8,11,11-tetramethyl-5-(2,2,6,6,9,
9,12,12-octamethyl-4,7,10-trioxo-3-oxa-5,8,11-triazatridecanamido)-
6,9,12-trioxo-16-oxa-1,7,10,13,19,20-hexaazabicyclo[16.2.1] henicosa-
18(21),19-dienecarboxamido)propanoate 2. (5S,14S)-14-(1-Methoxy-
2-methyl-1-oxopropan-2-ylcarbamoyl)-8,8,11,11-tetramethyl-6,9,
12-trioxo-16-oxa-1,7,10,13,19,20-hexaazabicyclo[16.2.1]henicosa-
18(21),19-dien-5-aminium 2,2,2-trifluoroacetate 6 (0.332 g, 0.498
mmol) and N-tert-butoxycarbonyl R,R-dimethylglycyl R,R-dimethyl-
glycyl R,R-dimethylglycine28 (0.186 g, 0.498 mmol) (see Supporting
Information for details on preparation) were suspended in CH2Cl2
(3 mL) and a solution of N,N-diisopropylethylamine (0.065 g, 0.50
mmol) in CH2Cl2 (4 mL) was added. HOBt hydrate (0.076 g, 0.50
mmol) and then EDC hydrochloride (0.105 g, 0.548 mmol) were
added at room temperature together with additional CH2Cl2 (3mL).
The reaction mixture was stirred for 44 h at room temperature before
being diluted with CH2Cl2 (40 mL). The solution was washed with
Conclusion
In summary, the feasibility of side chain-to-side chain cross-
linking by CuAAC in a 310-helical Aib rich peptide has been
demonstrated. An attractive feature of the cyclic product 2 is
its significantly higher aqueous solubility (>1 mM) com-
pared to 1. 2D IR and 2D ROESY experiments confirmed
that the cyclic peptide 2 retains a 310-helical structure in the
polar aprotic solvent CD2Cl2. The first X-ray crystallo-
graphic investigation of a helical peptide with a triazole
derived cross-link has revealed that 2 is the most perfect cross-
linked 310-helical peptide so far studied in the crystal state,
with mean (j, ψ)-angles deviating less than 2ꢀ from ideality.
The closeness to ideality of the conformational angles
in the solid state strongly suggests that the CuAAC side
J. Org. Chem. Vol. 76, No. 5, 2011 1237

Jacobsen et al.
JOCArticle
5% (w/w) citric acid monohydrate solution (3ꢀ20 mL), 7.5% (w/w)
K2CO3 solution (3ꢀ20 mL), and saturated brine (20 mL). The solu-
tion was dried with anhydrous MgSO4, and the solvent was evapo-
rated affording the title compound as a white solid (0.331 g, 73%):
1H NMR (400 MHz, CD2Cl2, 9.9 mM, 296 K) δ 7.91 (s, 1H,
NH(Aib3)), 7.88 (s, 1H, NH(Aib5)), 7.77 (d, J = 6.0 Hz, 1H,
NH(azidonorVal)), 7.73 (s, 1H, C2HN3), 7.53 (s, 1H, NH(Aib8)),
7.40 (d, J=8.4 Hz, 1H, NH(propSer)), 7.00 (s, 1H, NH(Aib6)), 6.72
(s, 1H, NH(Aib2)), 5.24 (s, 1H, NH(Aib1)), 4.82 (d, J=13.0 Hz, 1H,
OCHHC2HN3), 4.55 (td, J=8.4, 2.8 Hz, 1H, CRH(propSer)), 4.51
(d, J = 13.2 Hz, 1H, OCHHC2HN3), 4.38 (t, J = 5.7 Hz, 2H,
CH2N3C2H), 3.88 (t, J=8.5 Hz, 1H, CRHCHHO), 3.72 (dd, J=9.0,
2.9 Hz, 1H, CRHCHHO), 3.65 (s, 3H, OCH3), 3.23 (ddd, J=11.7,
5.9, 3.8 Hz, 1H, CRH(azidonorVal)), 2.31-2.17 (m, 1H, CH2-
CHHCH2), 2.07-1.93 (m, 2H, CRHCHHCH2CH2/CH2CHH-
CH2), 1.79-1.68 (m, 1H, CRHCHHCH2CH2), 1.64 (s, 3H, CH3),
1.58-1.35 (m, 42H, CH3/(CH3)3); 13C NMR (75 MHz, CD2Cl2, 51
mM, 298 K) δ 177.4, 175.9, 175.8, 175.6, 175.6, 175.5, 173.7, 170.0,
156.4, 145.3, 125.2, 81.9, 70.0, 65.3, 57.6, 57.5, 57.4, 57.3, 57.1, 56.4,
55.2, 54.4, 52.5, 49.7, 28.6, 28.3, 27.9, 27.9, 27.3, 27.2, 27.0, 25.4, 25.3,
24.8, 23.7, 23.4, 23.3, 23.2, 23.0; HRMS (m/z) [M þ H]þ calcd
for C41H70N11O12, 908.5205, found 908.5194. Anal. Calcd for
C41H69N11O12: C, 54.23; H, 7.66; N, 16.97. Found: C, 54.0; H, 7.6;
N, 16.5.
Acknowledgment. Ø.J., C.H.G., P.R., O.P.O., M.A.-M.,
and J.K. gratefully acknowledge Inven2 AS, the Technology
Transfer Office of the University of Oslo, and the Norwegian
Research Council for financial support through a verifica-
tion grant and professor Frode Rise (University of Oslo) and
€
€
Goran Schomer (Bruker Biospin Scandinavia) for NMR
technical support. H.M. and N.-H.G. gratefully acknowledge
grant support from the US National Science Foundation
(CHE-0802913).
Supporting Information Available: Synthesis schemes, full
experimental details for syntheses, compound characterization data,
crystallization conditions, crystallographic data in CIF format,
experimental details for acquisition of X-ray, NMR, linear IR and
2D IR data, model calculations of linear and 2D IR spectra,
Ramachandran plots, and partial 2D ROESY spectra. This material
is available free of charge via the Internet at http://pubs.acs.org.
1238 J. Org. Chem. Vol. 76, No. 5, 2011