A general method for constraining short peptides to an α-helical conformation

Article abstract of DOI:10.1021/ja9611654

A method for constraining short peptides (<20 residues) of arbitrary sequence to an α-helical conformation (~100% helical in H₂O at 25°C) is presented. Glutamine residues at positions i and i + 7 of the peptides were tethered with an alkanediyl

Full text of DOI:10.1021/ja9611654

VOLUME 119, NUMBER 3
J ANUARY 22, 1997
© Copyright 1997 by the
American Chemical Society
A General Method for Constraining Short Peptides to an
R-Helical Conformation
J. Christopher Phelan,*,† Nicholas J. Skelton,‡ Andrew C. Braisted,‡ and
Robert S. McDowell†
Contribution from the Department of Bioorganic Chemistry and Department of
Protein Engineering, Genentech, Inc., 460 Point San Bruno BouleVard,
South San Francisco, California 94080
ReceiVed April 8, 1996. ReVised Manuscript ReceiVed September 16, 1996X
Abstract: A method for constraining short peptides (<20 residues) of arbitrary sequence to an R-helical conformation
(∼100% helical in H2O at 25 °C) is presented. Glutamine residues at positions i and i + 7 of the peptides were
tethered with an alkanediyl chain between the side chain nitrogen atoms. Peptides containing this tether were readily
synthesized on the solid phase by amide formation between an R,ω-diaminoalkane and the side chain carboxylates
of glutamate residues. The resulting cyclic peptides were studied by NMR and CD and were found to adopt an
R-helical conformation in aqueous solution. The R-helix was thermally stable to g40 °C. Corresponding untethered
control peptides with N-methylglutamine at the i and i + 7 positions lacked helicity under the same conditions.
Analogous peptides were also prepared for comparison using the thiolysine cross-linking method described previously
[Jackson, D. Y.; King, D. S.; Chmielewski, J.; Singh, S.; Schultz, P. G. J. Am. Chem. Soc. 1991, 113, 9391-9392].
Introduction
A variety of methods for stabilizing R-helical peptides have
been described previously. Such peptides have been stabilized
by modification of the solvent (e.g., with trifluoroethanol) and
by dimerization at hydrophobic interfaces.3,4 Short R-helical
peptides have been stabilized from the end(s) by stabilization
of the intrinsic helix dipole5 and by incorporation of naturally-
The R-helix is the most prevalent secondary structure element
observed in proteins. However, investigations of the chemical
and physical properties of isolated R-helices have been hampered
by the inability of most linear peptides to sustain an R-helical
conformation in aqueous solution.1,2 A general method allowing
the study of isolated R-helices would constitute a useful tool
for experimental investigation of theories on the formation and
stability of R-helices, accurate measurements of the energetics
of binding interactions involving R-helices, and “epitope reduc-
tion” of R-helical protein domains involved in biologically
important interactions.
(3) Ho, S. P.; DeGrado, W. F. J. Am. Chem. Soc. 1987, 109, 6751-
6758.
(4) Fezoui, Y.; Weaver, D. L.; Osterhout, J. J. Proc. Natl. Acad. Sci.
U.S.A. 1994, 91, 3675-3697.
(5) Shoemaker, K. R.; Kim, P. S.; York, E. J.; Stewart, J. M.; Baldwin,
R. L. Nature 1987, 326, 563-567.
(6) Lyu, P. C.; Wemmer, D. E.; Zhou, H. X.; Pinker, R. J.; Kallenbach,
N. R. Biochemistry 1993, 32, 421-425.
* To whom correspondence should be addressed at Arris Pharmaceutical,
385 Oyster Point Blvd., South San Francisco, CA 94080.
† Department of Bioorganic Chemistry.
(7) Forood, B.; Feliciano, E. J.; Nambiar, K. P. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 838-842.
(8) Zhou, H. X.; Lyu, P. C.; Wemmer, D. E.; Kallenbach, N. R. J. Am.
Chem. Soc. 1994, 116, 1139-1140.
‡ Department of Protein Engineering.
X Abstract published in AdVance ACS Abstracts, December 15, 1996.
(1) Marqusee, S.; Robbins, V. M.; Baldwin, R. L. Proc. Natl. Acad. Sci.
U.S.A. 1989, 86, 5286-5290.
(9) Kemp, D. S. TIBTech 1990, 249-255.
(10) Kemp, D. S.; Curran, T. P.; Boyd, J. G.; Allen, T. J. J. Org. Chem.
1991, 56, 6683-6697.
(2) Finkelstein, A. V.; Badretdinov, A. Y.; Btitsyn, O. B. Proteins:
Struct., Funct., Genet. 1991, 10, 287-299.
(11) Kemp, D. S.; Curran, T. P.; Davis, W. M.; Boyd, J. G.; Muendel,
C. J. Org. Chem. 1991, 56, 6672-6682.
S0002-7863(96)01165-1 CCC: $14.00 © 1997 American Chemical Society

456 J. Am. Chem. Soc., Vol. 119, No. 3, 1997
Phelan et al.
occurring capping motifs6-8 as well as organic templates.9-12
Noncovalent side chain constraints that have been used for
R-helix stabilization include hydrophobic interactions,13 salt
bridges,14,15 and metal ion chelation by both natural16-18 and
unnatural19 amino acids. Finally, R-helices have been stabilized
by covalent side chain tethers: side chain to side chain
lactamization between residues i and i + 3, i and i + 4, or i
and i + 7,20-25 and disulfide bonds between residues i and i +
426 or i and i + 7.27
that D-amino acids tend to destabilize helical peptides by
approximately 1 kcal mol-1 33
The method also requires a
.
multistep synthesis of D- and L-thiolysine. Hence, we sought
to develop an alternative i to i + 7 tethering scheme that could
be implemented using commercially available L-amino acids.
We established the following criteria for a method that would
be generally useful for our applications: (1) The method must
allow the presentation of arbitrary amino acid sequences. (2)
It must maintain helicity approaching 100% in H2O at room
temperature in short peptides (less than 20 residues). (3) It must
be synthetically straightforward and compatible with solid-phase
peptide synthesis chemistry. (4) It must allow variation of
solvent (e.g., osmolarity and pH) and temperature. We report
here a new method for constraining peptides to an R-helical
conformation with an alkanediyl tether between amide side
chains at residues i and i + 7. The initial design of the tether
is described, along with the synthesis and structural character-
ization by NMR and CD of several short tethered R-helical
peptides. Structural studies of analogous nontethered peptides
and peptides constrained using the thiolysine “lock” are also
presented for comparison.
Earlier work on the bee venom peptide apamin28 as a scaffold
for presentation of R-helical peptide sequences29,30 served as a
starting point for our design of an amide-based tether. Apamin
presents a C-terminal helix that is stabilized by two disulfide
bonds to a structured N-terminal loop. The utility of apamin
as a general scaffold for helix display is limited by the fact that
the N-terminus of the helix is “capped” and cannot be extended.
Synthetic apamin peptides in which either of the cysteines is
replaced by a pair of alanines show a marked decrease in helicity
as evaluated by circular dichroism31 and proton NMR;32
comparison of the peptide lacking the Cys3-Cys15 pair with
that lacking the Cys1-Cys11 pair suggests that an i to i + 7
tether is more effective than an i to i + 3 tether for inducing
helicity.
Results and Discussion
We evaluated the i to i + 7 tethering scheme recently
described by Jackson et al.27 Although peptides with this
disulfide tether appear helical by CD, NMR studies suggest a
loss of helicity surrounding the D-amino acid in the i position
(see the Supporting Information), consistent with the observation
Design Considerations. As a tether we chose an alkanediyl
chain between the side chain nitrogen atoms of glutamine
residues at positions i and i + 7. This structure was compatible
with our chemical criteria and was expected to provide a
minimum of strain and steric hindrance. Straightforward
peptide-compatible chemistries with which to construct the
desired tether include amide and disulfide bonds. Disulfides,
however, introduce an unwanted 90° twist into the linkage. A
representative set of protein crystal structures from the Brookhaven
Database34 was searched for all occurrences of glutamine in an
R-helical context (with φ ) -60° ( 30° and ψ ) -45° (
30°). The resulting data set was used to determine the side
chain rotamer distributions of naturally occurring helical
glutamine residues. In general, amino acid residues in an
R-helical context have ø1 ≈ -60°, a conformation suitable for
the i + 7 position of a side chain linker. Glutamine has a
relatively high population (14.6%) of the ø1 ) 180° rotamer,
representing a suitable conformation for the i position that points
the side chain toward the C-terminal end of the helix. Rotamer
combinations were identified that minimized the Nꢀ2-Nꢀ2
distance between the i and i + 7 side chains in a model helical
peptide; the optimized distances ranged from 5.3 to 7 Å.
Model building suggested that a 4-methylene “bridge” could
optimally link these two glutamine side chains without incurring
unfavorable torsional interactions. Models of 3-, 4-, and
5-methylene-bridged helical peptides were constructed using
distance geometry methods35 followed by energy minimization.
All residues except the linked glutamines were alanine. The
conformational stabilities of helical peptides were assessed using
1 ns of unconstrained molecular dynamics at 298 K following
an initial 100 ps equilibration period during which harmonic
restraints (25 kcal mol-1 Å-1) were applied to maintain helicity.
As a control, a polyalanine helix was calculated for 1 ns in the
presence of identical restraints.
(12) Lieberman, M.; Tabet, M.; Sasaki, T. In Pept.: Chem. Biol., Proc.
Am. Pept. Symp., 12th; Smith, J. A., Rivier, J. E., Eds.; ESCOM: Leiden;
1991; pp 332-334.
(13) Albert, J. S.; Hamilton, A. D. J. Am. Chem. Soc. 1995, 34, 984-
990.
(14) Scholtz, J. M.; Zian, H.; Robbins, V. H.; Baldwin, R. L. Biochemistry
1993, 32, 9668-9676.
(15) Bierzynski, A.; Kim, P. S.; Baldwin, R. L. Proc. Natl. Acad. Sci.
U.S.A. 1982, 79, 2470-2474.
(16) Todd, R. J.; Van Dam, M. E.; Casimiro, D.; Haymore, B. L.; Arnold,
F. H. Proteins: Struct., Funct., Genet. 1991, 10, 156-161.
(17) Ghadiri, M. R.; Choi, C. J. Am. Chem. Soc., 1990, 112, 1630-
1632.
(18) Ghadiri, M. R.; Fernholz, A. K. J. Am. Chem. Soc. 1990, 112, 9633-
9635.
(19) Ruan, F.; Chen, Y.; Hopkins, P. B. J. Am. Chem. Soc. 1990, 112,
9403-9404.
(20) Chorev, M.; Roubini, E.; McKee, R. L.; Gibbons, S. W.; Goldman,
M. E.; Caulfield, M.P.; Rosenblatt, M. Biochemistry 1991, 30, 5968-5974.
(21) Osapay, G.; Taylor, J. W. J. Am. Chem. Soc. 1990, 112, 6046-
6051.
(22) Osapay, G.; Taylor, J. W. J. Am. Chem. Soc. 1990, 114, 6966-
6973.
(23) Bracken, C.; Gulya´s, J.; Taylor, J. W.; Baum, J. J. Am. Chem. Soc.
1994, 116, 6431-6432.
(24) Houston, M. E., Jr.; Kay, C. M.; Hodges, R. S. Fourteenth American
Peptide Symposium, Columbus, OH, June 18-23, 1995. Houston, M. E.,
Jr.; Gannon, C. L.; Kay, C. M.; Hodges, R. S. J. Pept. Sci. 1995, 1, 274-
282.
(25) Yu, C.; Kapurniotu, A.; Taylor, J. W. Fourteenth American Peptide
Symposium, Columbus, OH, June 18-23, 1995.
(26) Ravi, A.; Prasad, B. V. V.; Balaram, P. J. Am. Chem. Soc. 1983,
105, 105-109.
(27) Jackson, D. Y.; King, D. S.; Chmielewski, J.; Singh, S.; Schultz, P.
G. J. Am. Chem. Soc. 1991, 113, 9391-9392.
(28) Callewaert, G. L.; Shipolini, R.; Vernon, C. A. FEBS Lett. 1968, 1,
111-113. Shipolini, R.; Bradbury, A. F.; Callewaert, G. L.; Vernon, C. A.
Chem. Commun. 1967, 679-680. Habermann, E.; Reiz, K. G. Biochem. Z.
1965, 343, 192-203. Hahn, G.; Fernholz, M. E. Berichte 1939, 72, 1281.
(29) Danho, W.; Makofske, R.; Swistok, J.; Hakimi, J.; Kondas, J. A.;
Powers, G.; Biondi, D.; Varnell, T.; Fry, D.; Greeley, D.; Madison, V.
Fourteenth American Peptide Symposium, Columbus, OH, June 18-23,
1995.
Peptides containing a 3-methylene bridge maintained a
consistent helical conformation but showed significant “bending”
(33) Fairman, R.; Anthony-Cahill, S. J.; DeGrado, W. F. J. Am. Chem.
Soc. 1992, 114, 5458-5459.
(30) Brazil, B. T.; Cleland, J. L.; McDowell, R. S.; Skelton, N. J.; Paris,
K.; Horowitz, P. M. Submitted for publication.
(31) Huyghues-Despointes, B. M. P.; Nelson, J. W. Biochemistry 1992,
31, 1476-1483.
(32) Skelton, N. J.; Paris, K. Unpublished observations.
(34) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J. B.; Meyer, E. F.;
Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi, M. J.
Mol. Biol. 1977, 112, 535-542.
(35) Blaney, J. M.; Crippen, G. M.; Dearing, A.; Dixon, J. S. DGEOM;
QCPE No. 590.

Constraining Short Peptides to an R-Helical Conformation
Table 1. Structures of Peptides 1-4
J. Am. Chem. Soc., Vol. 119, No. 3, 1997 457
of the helix axis. Peptides containing a 4-methylene bridge
maintained helicity with little distortion, having backbone
dihedral angles comparable to those of the control peptide; ø1
and ø2 angles of the tethered glutamines did not change during
the simulation. Peptides based on a 5-methylene bridge
transiently escaped out of a helical conformation into nested
turns centered around the i + 5 residue. Multiple side chain
rotamers were also observed in the i + 7 residue. On the basis
of these observations, it seemed that the 4-methylene bridge
would be of optimal length, although the behavior of the
5-methylene-bridged peptide may have been an artifact of the
simple in Vacuo simulation.
Synthesis. Amino acid sequences for trial peptides were
based on the C-terminal helix of apamin36-38 (peptides 1 and
2, “apamin”) and on S-peptide derived from the C-peptide from
RNAse A6,39 (peptides 3 and 4, “C-tide”). The complete
sequences of these peptides are shown in Table 1.
Figure 1. Synthesis of 1b and 1c. a, 20% piperidine/DMA; b, H2-
NCH2CH2CH2NHR (R ) H or BOC), BOP, DIPEA, CH2Cl2; c, Pd-
(PPh3)4, 20% piperidine/DMA; (R ) BOC) TFA/CH2Cl2/anisole/
HSCH2CH2SH 9:9:1:1 v/v; d, BOP, DIPEA, CH2Cl2; e, HF/anisole/
EtSMe 20:2:1 v/v, 0 °C; f, CH3NH2, BOP, CH2Cl2.
Linear protected peptides 1a, 1f, and 2a were synthesized
by standard Merrifield methods using BOC chemistry. Control
peptides 1b and 2b were elaborated from 1f and 2a by
simultaneous deprotection of both glutamate residues followed
by coupling with methylamine (Figure 1). Attempted synthesis
of 1d from 1f by double deprotection and coupling with 1,4-
butanediamine gave an impure product in poor yield. Con-
strained peptides 1c-e and 2c-e were elaborated from 1a and
2a by removal of the 9-fluorenylmethyl ester from Glu3,
coupling with the appropriate alkanediamine, removal of the
allyl ester from Glu10, and cyclization (Figure 1). Yields were
improved by the use of mono-BOC-protected alkanediamine
in the first coupling step (R ) BOC) and by the use of a
polystyrene resin with 2% DVB cross-linker. The completed
peptides were cleaved from the resin with HF and purified by
preparative HPLC.
n
Figure 2. Synthesis of N-Fmoc-S-Acm-D-thiolysine (7): a, BuLi,
THF, -78 °C; Br(CH2)4Br; b, 4-MeOBnSH, KOtBu, THF; c, 0.25 M
HCl, THF/H2O; d, Hg(OAc)2, TFA; H2S; e, acetamidomethanol, TFA;
f, LiOH, THF/H2O; g, Fmoc-OSu, dioxane, NaHCO3.
in Figure 2. The Scho¨llkopf reagent40 was treated with
n-butyllithium followed by 1,4-dibromobutane to give the
known (bromobutyl)pyrazine 5. The bromide was displaced
with the potassium salt of 4-methoxytoluenethiol to give 6. The
pyrazine was hydrolyzed with aqueous HCl, and the thiol was
deprotected with Hg(OAc)2 in TFA followed by H2S. The crude
thiolysine ethyl ester was then reprotected with acetami-
domethanol in TFA. The ester was hydrolyzed with LiOH, and
the free S-protected amino acid was N-protected with Fmoc-
N-hydroxysuccinimide in dioxane to give 7. The same proce-
dures were used for the synthesis of 10.
Structural Characterization. Methods. The peptides were
primarily characterized using 2D 1H NMR; details are provided
in the Supporting Information. Resonance positions were
obtained by standard sequential assignment methods.41 The
degree of helicity of each peptide was judged from TOCSY
Thiolysine-based peptides 3a and 4a were synthesized in the
linear acetamidomethyl-protected form using standard Merrifield
methods and FMOC chemistry, followed by cleavage from the
resin with trifluoroacetic acid/triethylsilane (9:1 v/v) and
purification by preparative HPLC. These were converted into
the disulfide forms 3b and 4b in solution by simultaneous Cys
deprotection and oxidation with acetic acid and molecular iodine
in 2,2,2-trifluoroethanol.
Peptides 1-4 were characterized by mass spectrometry and
by quantitative amino acid analysis. All peptides gave results
consistent with the intended structures. Experimental and
calculated molecular weights and AAA values are reported in
the Supporting Information.
Protected D- (7) and L-thiolysine (10) were prepared as shown
(36) Habermann E.; Reiz, K. G. Biochem. Z. 1965, 343, 192-203.
(37) Callewaert, G. L.; Shipolini, R.; Vernon, C. A. FEBS Lett. 1968, 1,
111-113.
(40) Scho¨llkpf, U.; Groth, U.; Deng, C. Angew. Chem., Int. Ed. Engl.
1981, 20, 798-799.
(41) Wu¨thrich, K. NMR of Proteins and Nucleic Acids; Wiley: New
York, 1986.
(38) Shipolini, R.; Bradbury, A. F.; Callewaert, G. L.; Vernon, C. A.
Chem. Commun. 1967, 14, 679-680.
(39) Brown, J. E.; Klee, W. A. Biochemistry 1971, 10, 470-476.

458 J. Am. Chem. Soc., Vol. 119, No. 3, 1997
Phelan et al.
Figure 3. HN-HR (a) and HN-HN (b) sections of the ROESY spectrum
of 1c. The spectrum was collected at 280 K, pH 5.0, 500 MHz and a
peptide concentration of 1.5 mM with a 4.5 kHz spin-lock mixing pulse
of 200 ms duration. Lines connect the ROEs by which sequential
assignments were made. Rectangular, oval, and diamond-shaped boxes
denote intraresidue, sequential, and (i, i + 3) correlations, respectively.
3
N
R
Figure 4. ROE and JH
data for 1c and 1b. For the dNN and dRN
-H
rows, observation of the sequential ROE is indicated by a bar connecting
two residues, the thickness of the bar indicating the relative intensity
of the ROE. The downward pointing arrows indicate 3JH
less than
N
R
-H
6.0 Hz. Observed medium-range ROEs (HR-HN (i, i + 3) and HR-Hâ
(i, i + 3)) are indicated by the lines in the lower part of the figure;
dotted lines and stars indicate ROEs that could not be unambiguously
observed because of chemical shift degeneracy. The coil motif above
the primary sequence indicates the region deduced to have helical
structure from the NMR data; the dashed coil indicates sections of
peptide where only some of the NMR data indicate helical character.
and ROESY spectra by the presence of intense sequential HN-
HN ROE cross-peaks, the presence of i and i + 3 HR-HN or
HR-Hâ ROE cross-peaks, and 3JH
-
< 6.0 Hz. A represen-
N
R
H
tative ROESY spectrum of diamide-constrained peptide 1c is
shown in Figure 3; a section of the TOCSY spectrum of this
peptide is show in the Supporting Information. Additional
spectra (NOESY and higher-sensitivity ROESY) were acquired
and used as a basis for structure determination using distance
geometry (DG) and restrained molecular dynamics (rMD).
Circular dichroism spectra were acquired on aqueous solutions
of 1-4 between 20 and 120 µM at 280 K and pH 5 to confirm
the results of NMR studies; variable-temperature CD studies
were used to evaluate the helicity of peptides at higher
temperatures.
Table 2. Amide Hydrogen Exchange Rate Constantsa and
Protection Factorsb for Peptides 1b and 1c
residue
log k (1b)
log k (1c)
protection factor
Thr1
-2.44
nd
nd
-2.72
-2.69
-2.73
-2.51
nd
nd
nd
nd
nd
-2.48
nd
∼1
Asn2
Gln3
Asp4
Leu5
Ala6
nd
-2.84
-3.51
-3.77
-3.55
nd
-3.21
-3.29
nd
1.3
6.7
10.9
11.1
> 26
> 20
>25
Helicity of 1c. NMR. Summaries of the HN-HN and HR-
HN ROEs between neighboring residues are depicted in Figure
4 for 1c and for the corresponding control peptide 1b. These
data indicate that peptide 1c is helical between residues Asn2
Ala7
Arg8
Arg9
Gln10
Gln11
Gln12
3
N
R
and Gln10. Beyond Gln10, JH
-
rises above 6.0 Hz, but
nd
H
some medium range ROEs are still present, suggesting partial
or transient helical character. Such fraying at helix termini is
commonly observed in NMR studies of peptides and proteins.
a The rate constants are expressed in units of s-1. nd indicates that
the exchange was sufficiently fast that no peak was observed in the
NMR spectrum acquired 300 s after addition of D2O. In these cases,
assuming that >90% of the hydrogen atoms have exchanged in 300 s
allows a lower limit of 0.013 s-1 (log k ) -1.89) to be calculated for
the rate constant. b Protection factors are calculated as the rate constant
for peptide 1c divided by that of peptide 1b.
1
The H chemical shifts of 1c change by <0.02 ppm over the
concentration range 8.0-0.06 mM; this suggests that the helical
conformation is not stabilized by a self-association event. HR-
HN (i, i + 4) interactions were also observed in higher-sensitivity
ROESY spectra of 1c, indicating that the helical conformation
adopted is not of the 310 type, but rather is of the regular
R-helical variety.42
of 20 structures is depicted in Figure 5. The structures agree
with the input data very well, with no distance restraint
violations above 0.1 Å, no dihedral angle violations above 1.0°,
and a mean restraint violation energy term of 0.10 ( 0.09 kcal
mol-1. The backbone atoms of residues Thr1 to Gln10 are well-
defined (average RMS deviation from the mean structure 0.38
( 0.08 Å), but the two C-terminal glutamine residues are not.
The side chains of Thr1, Gln3, Asp4, Leu6, and Gln10 have
well-defined ø1 values, but only Gln10 has a consistent value
of ø2 in all structures.
Interproton distance restraints were generated from the
ROESY and NOESY data, and were used as a basis for
calculating a structure for 1c. Nearly half (66) of the 141
restraints were between amino acids two to four residues apart
in the primary sequence, as expected for an R-helical conforma-
tion. Dihedral angle restraints, based on observed 3JH
-
and
N
R
H
3
R
â
JH -H , were also used in these calculations, but explicit
hydrogen bond restraints were not utilized. The final ensemble
Hydrogen-bonding interactions observed both in the calcu-
lated structure of 1c and by solvent exchange experiments also
support an R-helical structure for 1c. HN(i)-O(i - 4) hydrogen
(42) Harper, E. T.; Rose, G. D. Biochemistry 1993, 30, 7605-7609.

Constraining Short Peptides to an R-Helical Conformation
J. Am. Chem. Soc., Vol. 119, No. 3, 1997 459
Figure 5. Ensemble of 20 rMD structures calculated using NMR data for peptide 1c. The structures were overlaid using the N, CR, and C atoms
of residue Thr1 to Gln10. Backbone and side chain heavy atoms are connected by solid and dotted lines, respectively. The side chains of Arg8 and
Arg9 have been truncated at Cγ, and all side chain atoms of Gln11 and Gln12 have been omitted for clarity.
bonds are observed to the amide protons of Leu5, Ala6, and
Gln10 in >90% of the structures and to the amide protons of
Ala7, Arg8, and Arg9 in ∼50% of the structures. HN(i)-O(i
- 3) hydrogen bonds are present for the latter protons in 25-
35% of the structures, suggesting that there is a slight distortion
of the helix in this region. Interestingly, hydrogen bonds from
Asp4 HN to Thr1 Oγ2 are present in 80% of the structures,
indicating that an N-cap hydrogen-bonding interaction42 is
present even in this short peptide. Solvent exchange rates for
the amide hydrogen atoms of 1c are up to 25-fold slower than
those of 1b (see Table 2), providing experimental evidence for
the hydrogen bonding observed in the calculated structures.
However, the amide proton of Asp4 is not noticeably protected
from exchange; hence, the N-cap hydrogen bond may be more
transient.
The backbone dihedral angles throughout the tethered region
are close to those expected for an ideal R-helix (mean φ ) -63°
Figure 6. CD spectra of 1c at 280, 310, 330, 350, and 370 K.
( 8°, mean ψ ) -42° ( 8°). Exceptions are ψ of Ala6, which
is approximately 15° lower than expected (closer to that
expected for a 310 helix), and ψ of Gln10, which is higher than
ideal (reflecting fraying beyond the tethered region). The
deviation at Ala6 may be the result of the short tether present
in this peptide. Although the diamide linkage is not well
defined, the side chains of Gln3 and Gln10 do adopt conforma-
tions close to those predicted by the initial modeling experiments
described above (Gln3 ø1 ) -173° ( 17°, ø2 ) 34° ( 47°;
Gln10 ø1 ) -71° ( 7°, ø2 ) 174° ( 22°). Thus, the 1H NMR
data support an R-helical solution structure for 1c from Asp2
to Gln10 with an N-terminal capping box and a very slight
distortion in the central turn of the helix.
helicity for control peptide 1b (Figure 6). The integrity of the
1c helix was retained under thermal denaturation conditions.
Even at 97 °C, the ellipticity of 1c was substantially more
negative than that of 1b at 7 °C, and after recooling most of
the helical character of 1c was restored (data shown in the
Supporting Information).
Comparison of Peptides. A summary of both 1H NMR and
CD data for peptides 1-4 is given in Table 3. Detailed
summaries of 1H NMR and graphs of CD data for these peptides
are given in the Supporting Information. Comparison of
peptides 1c-1e and 2c-2e with corresponding control peptides
1b and 2b shows that incorporation of diamide tethers of three,
four, or five carbon atoms in peptides 1 and 2 clearly reduces
3
N
R
Stability by CD. The helicity of 1c was confirmed by CD
spectrometry. At 7 °C, 1c showed 84% helicity (evaluated by
per-residue molar ellipticity at 222 nm), as compared to 20%
the mean value of JH
-
in the restrained region, increases
H
the number of observable (i, i + 3) ROEs, and increases the
percent helicity observed by CD. The results for peptide 4

460 J. Am. Chem. Soc., Vol. 119, No. 3, 1997
Phelan et al.
Table 3. Evaluation of Peptide Helicitya
variability in all residues except the two directly involved in
the tether and is successful with short (12 and 13 residue)
peptides; comparison of helical peptides 1c-1e with nonhelical
peptide 1b shows that the helicity is achieved by introduction
of the linker rather than being a property of the primary
sequence. The method induces helicity approaching 100% in
aqueous solution at room temperature and substantial helicity
at higher temperatures. The tethered peptides are synthesized
by standard solid-phase (Merrifield) chemistry requiring only
commercially available reagents. This method should be
generally useful for studies of biologically active helical regions
of proteins; for the experimental study of helix formation,
propagation, and stability; and for physical organic experiments
on the interactions of helical peptides with their environments.
mean
3JH
in
fraction of percent
N-HR
pep-
tide
constrained medium-rang helicity
description
region
ROEs obsd by CD
1b apamin, N-methyl-Gln control
1c apamin, 1,3-propanediyl linker
1d apamin, 1,4-butanediyl linker
1e apamin, 1,5-pentanediyl linker
2b C-tide, N-methyl-Gln control
2c C-tide, 1,3-propanediyl linker
2d C-tide, 1,4-butanediyl linker
2e C-tide, 1,5-pentanediyl linker
3a apamin, S-Acm thiolys control
3b apamin, thiolys disulfide
6.0
5.0
5.2
5.0
5.9
4.8
4.8
4.9
6.0
6.0
6.0
6.0
0.14
0.69
0.56
0.69
0.08
0.75
0.43
0.80
0.03
0.08
0.05
0.48
20
84
63
100
32
60
82
63
10
35
19
27
4a C-tide, S-Acm thiolys control
4b C-tide, thiolys disulfide
Experimental Section
a Values below 6 for the mean three-bond HN-HR coupling constant
indicate helicity. Medium-range i - i + 3 ROEs are expressed as the
observed fraction of the total number of such ROEs possible, with very
weak ROEs counted as half. The percent helicity as determined by
CD is derived as by Jackson.42
Details for the syntheses of peptides 1-4 and for D- and L-Boc-
thiolys(S-Acm)OH (7 and 10) and for experimental methods for H
NMR and CD spectroscopy are provided in the Supporting Information.
1
Acknowledgment. Mono-BOC-1,3-propanediamine was pre-
pared by Kirk D. Robarge. Mass spectra were obtained by Dan
Burdick and Kathy O’Connell. Quantitative amino acid analy-
ses were performed by Allan Padua. Purifications of 1c, 1d,
and 1e were performed by Kathryn S. Chan. Kevin Judice,
Mary A. Reppy, Mary Cromwell, and David Y. Jackson
provided helpful discussions.
indicate that the formation of the disulfide bond does constrain
the peptide to be helical (Table 3). However, a number of
3
N
R
medium-range ROEs could not be observed and the JH
-
H
values were >6.0 Hz for the two thiolysine residues and Leu9
in 4b, indicating a distortion from an ideal helical structure in
the region of the D-thiolysine residue. With the apamin
sequence, incorporation of the thiolysine disulfide appears to
impart an increase in helicity as judged by CD (3b Vs 3a);
Supporting Information Available: Details of synthetic and
analytical methods, additional references for the Experimental
Section, tables of 1H NMR data for peptides 1-4 , summaries
of ROE and coupling data for peptides 1-4, a section of the
TOCSY spectrum of peptide 1c, a graph of thermal denaturation
of 1c, and CD spectra of peptides 1-4 (29 pages). See any
current masthead page for ordering and Internet access
instructions.
1
however, the H NMR data for 3b show none of the charac-
teristics of a true helix.
Conclusion
We have devised a simple method for constraining small
peptides to an R helical conformation. This i to i + 7 amide-
based tether is successful as a general method for inducing
R-helicity in small peptides. It allows complete sequence
JA9611654