Crystallographic characterization of helical secondary structures in α/β-peptides with 1:1 residue alternation

Article abstract of DOI:10.1021/ja800355p

Oligomers that contain both α- and β-amino acid residues in a 1:1 alternating pattern have recently been shown by several groups to adopt helical secondary structures in solution. The β-residue substitution pattern has a profound effect on the type of helix formed and the stability of the helical conformation. On the basis of two-dimensional NMR data, we have previously proposed that β-residues with a five-membered ring constraint promote two different types of α/β-peptide helix. The 11-helix contains i,i+3 C=O...H-N hydrogen bonds between backbone amide groups; these hydrogen bonds occur in 11-atom rings. The α/β-peptide 14/15-helix contains i,i+A C=O...H-N hydrogen bonds, which occur in alternating 14- and 15-atom rings. Here we provide crystallographic data for 14 α/β-peptides that form the 11-helix and/or the 14/15-helix. These results were obtained for a series of oligomers containing β-residues derived from (S,S)-trans-2- aminocyclopentanecarboxylic acid (ACPC) and α-residues derived from α-aminoisobutyric acid (Aib) or L-alanine (Ala). The crystallized α/β-peptides range in length from 4 to 10 residues. Nine of the α/β- peptides display the 11-helix in the solid state, three display the 14/15-helix, and two display conformations that contain both i,i+3 and i,i+4 C=O...H-N hydrogen bonds, but not bifurcated hydrogen bonds. Only 3 of the 14 crystal structures presented here have been previously described. These results suggest that longer α/β-peptides prefer the 14/15-helix over the 11-helix, a conclusion that is consistent with previously reported NMR data obtained in solution.

Full text of DOI:10.1021/ja800355p

Published on Web 04/26/2008
Crystallographic Characterization of Helical Secondary
Structures in r/ꢀ-Peptides with 1:1 Residue Alternation
Soo Hyuk Choi, Ilia A. Guzei, Lara C. Spencer, and Samuel H. Gellman*
Department of Chemistry, UniVersity of Wisconsin, Madison, Wisconsin 53706
Received January 17, 2008; E-mail: gellman@chem.wisc.edu
Abstract: Oligomers that contain both R- and ꢀ-amino acid residues in a 1:1 alternating pattern have recently
been shown by several groups to adopt helical secondary structures in solution. The ꢀ-residue substitution
pattern has a profound effect on the type of helix formed and the stability of the helical conformation. On
the basis of two-dimensional NMR data, we have previously proposed that ꢀ-residues with a five-membered
ring constraint promote two different types of R/ꢀ-peptide helix. The “11-helix” contains i,i+3 CdO· · ·H-N
hydrogen bonds between backbone amide groups; these hydrogen bonds occur in 11-atom rings. The
R/ꢀ-peptide “14/15-helix” contains i,i+4 CdO· · ·H-N hydrogen bonds, which occur in alternating 14- and
15-atom rings. Here we provide crystallographic data for 14 R/ꢀ-peptides that form the 11-helix and/or the
14/15-helix. These results were obtained for a series of oligomers containing ꢀ-residues derived from (S,S)-
trans-2-aminocyclopentanecarboxylic acid (ACPC) and R-residues derived from R-aminoisobutyric acid (Aib)
or L-alanine (Ala). The crystallized R/ꢀ-peptides range in length from 4 to 10 residues. Nine of the R/ꢀ-
peptides display the 11-helix in the solid state, three display the 14/15-helix, and two display conformations
that contain both i,i+3 and i,i+4 CdO· · ·H-N hydrogen bonds, but not bifurcated hydrogen bonds. Only
3 of the 14 crystal structures presented here have been previously described. These results suggest that
longer R/ꢀ-peptides prefer the 14/15-helix over the 11-helix, a conclusion that is consistent with previously
reported NMR data obtained in solution.
Introduction
a single class (R-amino acids). A desire to move beyond the
biopolymer prototypes has recently motivated many researchers
The broad array of activities manifested among folded
biopolymers (proteins and RNA) has inspired widespread
interest in unnatural oligomers that display comparable folding
behavior, which are collectively designated “foldamers”.1 A
long-range goal of foldamer research is to endow synthetic
oligomers with sophisticated functions of the types found among
biopolymers or comparably sophisticated functions that are not
part of the biopolymer repertoire. Efficient catalysis, selective
recognition, programmed self-assembly, and other natural
protein activities often require adoption of a specific conforma-
tion by the polypeptide backbone, and it seems likely that
function-based engineering of foldamers will depend upon a
thorough understanding of the conformational propensities of
unnatural backbones. ꢀ-Peptides (oligomers of ꢀ-amino acids)
are among the most widely studied foldamers at present, and
there are now many examples in which ꢀ-peptide function
requires the adoption of a discrete and predictable secondary
structure.2
to explore oligomers with heterogeneous backbones, that is,
backbones containing more than one type of building block.3–9
Several groups have characterized the folding of oligomers
containing both R- and ꢀ-amino acid residues (“R/ꢀ-peptides”).3–6
A number of helical secondary structures have been computa-
tionally predicted5 and/or empirically deduced based on NMR
(2) (a) Cheng, R. P.; Gellman, S. H.; DeGrado, W. F. Chem. ReV. 2001,
101, 3219. (b) Stephens, O. M.; Kim, S.; Welch, B. D.; Hodsdon,
M. E.; Kay, M. S.; Schepartz, A. J. Am. Chem. Soc. 2005, 127, 13126.
(c) Potocky, T. B.; Menon, A. K.; Gellman, S. H. J. Am. Chem. Soc.
2005, 127, 3686. (d) English, E. P.; Chumanov, R. S.; Gellman, S. H.;
Compton, T. J. Biol. Chem. 2006, 281, 2661. (e) Pomerantz, W. C.;
Abbott, N. L.; Gellman, S. H. J. Am. Chem. Soc. 2006, 128, 8730. (f)
Karlsson, A. J.; Pomerantz, W. C.; Weisblum, B.; Gellman, S. H.;
Palecek, S. P. J. Am. Chem. Soc. 2006, 128, 12630.
(3) (a) Hayen, A.; Schmitt, M. A.; Ngassa, F. N.; Thomasson, K. A.;
Gellman, S. H. Angew. Chem., Int. Ed. 2004, 43, 505. (b) Schmitt,
M. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc. 2004, 126,
6848. (c) Schmitt, M. A.; Choi, S. H.; Guzei, I. A.; Gellman, S. H.
J. Am. Chem. Soc. 2005, 127, 13130. (d) Sadowsky, J. D.; Schmitt,
M. A.; Lee, H.-S.; Umezawa, N.; Wang, S.; Tomita, Y.; Gellman,
S. H. J. Am. Chem. Soc. 2005, 127, 11966. (e) Schmitt, M. A.; Choi,
S. H.; Guzei, I. A.; Gellman, S. H. J. Am. Chem. Soc. 2006, 128,
4538. (f) Sadowsky, J. D.; Fairlie, W. D.; Hadley, E. B.; Lee, H. S.;
Umezawa, N.; Nikolovska-Coleska, Z.; Wang, S. M.; Huang, D. C. S.;
Tomita, Y.; Gellman, S. H. J. Am. Chem. Soc. 2007, 129, 139. (g)
Schmitt, M. A.; Weisblum, B.; Gellman, S. H. J. Am. Chem. Soc.
2007, 129, 417. (h) Horne, W. S.; Price, J. L.; Keck, J. L.; Gellman,
S. H. J. Am. Chem. Soc. 2007, 129, 4178. (i) Price, J. L.; Horne, W. S.;
Gellman, S. H. J. Am. Chem. Soc. 2007, 129, 6376. (j) Choi, S. H.;
Guzei, I. A.; Gellman, S. H. J. Am. Chem. Soc. 2007, 129, 13780. (k)
Horne, W. S.; Boersma, M. D.; Windsor, M. A.; Gellman, S. H.
Angew. Chem. Int. Ed., Early View (DOI: 10.1002/anie.200705315).
Our ability to achieve specific functional goals with foldamers
should be enhanced by enlarging the set of backbones for which
folding rules are known, because each new scaffold offers new
ways to orient sets of side chains in space. Proteins have
homogeneous backbones, since all building blocks come from
(1) (a) Gellman, S. H. Acc. Chem. Res. 1998, 31, 173. (b) Hill, D. J.;
Mio, M. J.; Prince, R. B.; Hughes, T. S.; Moore, J. S. Chem. ReV.
2001, 101, 3893–4011. (c) Hecht, S., Huc, I., Eds. Foldamers:
Structure, Properties and Applications; Wiley-VCH: Weinheim,
Germany, 2007. (d) Goodman, C. M.; Choi, S.; Shandler, S.; DeGrado,
W. F. Nat. Chem. Biol. 2007, 3, 252.
9
6544 J. AM. CHEM. SOC. 2008, 130, 6544–6550
10.1021/ja800355p CCC: $40.75
2008 American Chemical Society

Helical Secondary Structures in R/ꢀ-Peptides
A R T I C L E S
analysis.3a,b,4 Our own efforts in this area have focused on
conformationally restricted ꢀ-residues, such as those derived
from trans-2-aminocyclopentanecarboxylic acid (ACPC).10 For
R/ꢀ-peptides with 1:1 residue alternation, we have found that
the five-membered ring constraint favors two different helical
conformations that are named on the basis of the characteristic
hydrogen bonds formed between backbone amide groups.3a The
11-helix contains i,i+3 CdO · · · H-N hydrogen bonds, and
the 14/15-helix contains i,i+4 CdO· · ·H-N hydrogen bonds.
These two R/ꢀ-peptide helices can be regarded as analogues of
the two most common helical secondary structures among
proteins, the 310-helix (i,i+3 CdO· · ·H-N hydrogen bonds)
and the R-helix (i,i+4 CdO· · ·H-N hydrogen bonds). Rela-
tively short R/ꢀ-peptides containing constrained ꢀ-residues (6-8
residues total) appear to form both helices in solution, with
Chart 1. Crystallized R/ꢀ-Peptides
interconversion rapid on the NMR time scale. Longer R/ꢀ-
peptides in this family (15 residues) seem to favor the 14/15-
helix in solution.3b However, these NMR-based conclusions
must be regarded as tentative because all R/ꢀ-peptides examined
to date experience rapid equilibration between folded and
unfolded states on the NMR time scale.
(4) (a) De Pol, S.; Zorn, C.; Klein, C. D.; Zerbe, O.; Reiser, O. Angew.
Chem., Int. Ed. 2004, 43, 511. (b) Sharma, G. V. M.; Nagendar, P.;
Jayaprakash, P.; Krishna, P. R.; Ramakrishna, K. V. S.; Kunwar, A. C.
Angew. Chem., Int. Ed. 2005, 44, 5878. (c) Srinivasulu, G.; Kumar,
S. K.; Sharma, G. V. M.; Kunwar, A. C. J. Org. Chem. 2006, 71,
8395. (d) Seebach, D.; Jaun, B.; Sebesta, R.; Mathad, R. I.; Flogel,
O.; Limbach, M.; Sellner, H.; Cottens, S. HelV. Chim. Acta 2006, 89,
1801. (e) Vilaivan, T.; Srisuwannaket, C. Org. Lett. 2006, 8, 1897.
(f) Jagadeesh, B.; Prabhakar, A.; Sarma, G. D.; Chandrasekhar, S.;
Chandrashekar, G.; Reddy, M. S.; Jagannadh, B. Chem. Commun.
2007, 371.
Crystallographic data for new foldamers provide high-
resolution structural information that serves as a basis for
subsequent function-based design efforts. To date, only three
crystal structures have been reported for R/ꢀ-peptides with 1:1
residue alternation.3c,j Here we report an additional 11 R/ꢀ-
peptide crystal structures, in which all ꢀ-residues are (S,S)-ACPC
(1) and the R-residues are R-aminoisobutyric acid (Aib; 2) or
L-alanine (Ala; 3). Chart 1 shows all 14 R/ꢀ-peptides for which
crystal structures have been obtained (the structures of 8, 8b,
and 9 have been previously described).3c,e The R/ꢀ-peptides
range in length from 4 to 10 residues, and length seems to
influence the type of helix formed in the solid state. Up to the
heptamer length, the R/ꢀ-peptides crystallize in the 11-helical
conformation. Among the four octamers we crystallized, two
are entirely 11-helical; that is, they contain exclusively i,i+3
CdO· · ·H-N hydrogen bonds. The other two display helical
conformations that contain both i,i+3 and i,i+4 CdO· · ·H-N
hydrogen bonds. The two nonamers and the decamer crystallize
in the 14/15-helical conformation; that is, they display only i,i+4
CdO· · ·H-N hydrogen bonds. The resulting set of structures
is large enough to allow a meaningful analysis of the R-residue
and ꢀ-residue torsion angles associated with each type of helix.
(5) (a) Baldauf, C.; Gunther, R.; Hofmann, H. J. Biopolymers 2006, 84,
408. (b) Zhu, X.; Yethiraj, A.; Cui, Q. J. Chem. Theory Comput. 2007,
3, 1538.
(6) For crystal structures of short R/ꢀ-peptides: (a) Karle, I. L.; Handa,
B. K.; Hassall, C. H. Acta Crystallogr., Sect. B: Struct. Sci. 1975,
B31, 555. (b) Pavone, V.; Diblasio, B.; Lombardi, A.; Isernia, C.;
Pedone, C.; Benedetti, E.; Valle, G.; Crisma, M.; Toniolo, C.; Kishore,
R. J. Chem. Soc., Perkin Trans. 2 1992, 1233. (c) Hanessian, S.; Yang,
H. Tetrahedron Lett. 1997, 38, 3155. (d) Rossi, F.; Bucci, E.; Isernia,
C.; Saviano, M.; Iacovino, R.; Romanelli, A.; Di Lello, P.; Grimaldi,
M.; Montesarchio, D.; De Napoli, L.; Piccialli, G.; Benedetti, E.
Biopolymers 2000, 53, 140. (e) Tanaka, M.; Oba, M.; Ichiki, T.;
Suemune, H. Chem. Pharm. Bul. 2001, 49, 1178.
(7) (a) Hagihara, M.; Anthony, N. J.; Stout, T. J.; Clardy, J.; Schreiber,
S. L. J. Am. Chem. Soc. 1992, 114, 6568. (b) Baldauf, C.; Gunther,
R.; Hofmann, H. J. J. Org. Chem. 2006, 71, 1200. (c) Sharma,
G. V. M.; Jadhav, V. B.; Ramakrishna, K. V. S.; Narsimulu, K.;
Subash, V.; Kunwar, A. C. J. Am. Chem. Soc. 2006, 128, 14657. (d)
Baruah, P. K.; Sreedevi, N. K.; Gonnade, R.; Ravindranathan, S.;
Damodaran, K.; Hofmann, H. J.; Sanjayan, G. J. J. Org. Chem. 2007,
72, 636. (e) Vasudev, P. G.; Ananda, K.; Chatterjee, S.; Aravinda, S.;
Shamala, N.; Balaram, P. J. Am. Chem. Soc. 2007, 129, 4039. (f)
Chatterjee, S.; Roy, R. S.; Balaram, P. J. R. Soc. Interface 2007, 4,
587.
(8) (a) Gong, B. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 11583. (b) Yang,
D.; Li, W.; Qu, J.; Luo, S. W.; Wu, Y. D. J. Am. Chem. Soc. 2003,
125, 13018. (c) Chowdhury, S.; Schatte, G.; Kraatz, H. B. Angew.
Chem., Int. Ed. 2006, 45, 6882. (d) Zhao, Y.; Zhong, Z. Q.; Ryu,
E. H. J. Am. Chem. Soc. 2007, 129, 218. (e) Olsen, C. A.; Bonke, G.;
Vedel, L.; Adsersen, A.; Witt, M.; Franzhk, H.; Jaroszewski, J. W.
Org. Lett. 2007, 9, 1549. (f) Angelici, G.; Luppi, G.; Kaptein, B.;
Broxterman, Q. B.; Hofmann, H. J.; Tomasini, C. Eur. J. Org. Chem.
2007, 2713. (g) Zhong, Z.; Zhao, Y. Org. Lett. 2007, 9, 2891. (h)
Delsuc, N.; Godde, F.; Kauffmann, B.; Leger, J. M.; Huc, I. J. Am.
Chem. Soc. 2007, 129, 11348.
Results and Discussion
Synthesis. A previously reported procedure10b,c was used to
prepare the ACPC derivative 2-(1-phenylethylamino)cyclopen-
tanecarboxylic acid ethyl ester. The phenylethyl group was
removed via catalytic hydrogenolysis, and the resulting amino
group was acylated with Boc-Ala-OH or Boc-Aib-OH using a
carbodiimide activating agent. The resulting dipeptide ethyl
esters, Boc-Ala-ACPC-OEt and Boc-Aib-ACPC-OEt, were
either saponified or converted directly to the benzyl ester. All
R/ꢀ-peptides in Chart 1 were prepared by carbodiimide-mediated
coupling of dipeptide segments.
r/ꢀ-Peptide Crystal Structures. 11-Helical conformations are
adopted in the solid state by the R/ꢀ-peptides among our set
that contain from four to seven residues (Figure 1). Tetramer 4
and pentamer 5, with alternating Aib and ACPC residues,
display the maximum number of 11-membered H-bonded rings.
For each of the three hexamers (6, 6a, and 6b) the N-terminal
R-residue does not participate in the helical hydrogen bonding
pattern, although the rest of each hexamer is 11-helical.
Heptamer 7, which contains only Aib R-residues, adopts a fully
(9) For R/ꢀ-peptides with sheet or turn structures, see:(a) Krauthauser,
S.; Christianson, L. A.; Powell, D. R.; Gellman, S. H. J. Am. Chem.
Soc. 1997, 119, 11719. (b) Huck, B. R.; Fisk, J. D.; Gellman, S. H.
Org. Lett. 2000, 2, 2607. (c) Gopi, H. N.; Roy, R. S.; Raghothama,
S. R.; Karle, I. L.; Balaram, P. HelV. Chim. Acta 2002, 85, 3313. (d)
Arnold, U.; Hinderaker, M. P.; Nilsson, B. L.; Huck, B. R.; Gellman,
S. H.; Raines, R. T. J. Am. Chem. Soc. 2002, 124, 8522. (e) Roy,
R. S.; Karle, I. L.; Ragothama, S.; Balaram, P. Proc. Natl. Acad. Sci
U.S.A. 2004, 101, 16478. (f) Roy, R. S.; Gopi, H. N.; Ragothama, S.;
Karle, I. L.; Balaram, P. Chem.-Eur. J. 2006, 12, 3295.
(10) (a) Appella, D. H.; Christianson, L. A.; Klein, D. A.; Richards, M. R.;
Powell, D. R.; Gellman, S. H. J. Am. Chem. Soc. 1999, 121, 7574.
(b) Lee, H.-S.; LePlae, P. L.; Porter, E. A.; Gellman, S. H. J. Org.
Chem. 2001, 66, 3597. (c) LePlae, P. R.; Umezawa, N.; Lee, H.-S.;
Gellman, S. H. J. Org. Chem. 2001, 66, 5629.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6545

A R T I C L E S
Choi et al.
Figure 1. Crystal structures of 11-helical R/ꢀ-peptides: (a) stereoview of heptamer 7. Hydrogen atoms other than amide protons are omitted. Dotted line
indicates intramolecular H-bonds. (b) Overlay of seven structures. The structure of heptamer 7 is drawn in gray. For the other six structures, only backbone
atoms (CR, Cꢀ, CdO, N-H) are shown. C-Terminal ACPCs and nonhelical N-terminal residues (in 6, 6a, 6b) are omitted.
Figure 2. Crystal structures of octamers: (a) fully 11-helical conformation; (b) chimeric conformation (four 11-helical H-bonds and one 14-atom H-bonded
ring). (c) Comparison of 11-atom and 14-atom H-bonded rings in 8a and 8b, respectively (stereoview). Only the first four residues from the N-terminus are
shown.
Figure 3. Crystal structures of 14/15-helical R/ꢀ-peptides: (a) stereoview of decamer 10; (b) overlay of three structures. The structure of decamer 10 is
drawn in gray. For the two nonamers, only backbone atoms (CR, Cꢀ, CdO, N-H) are shown.
11-helical conformation (five intramolecular hydrogen bonds).
Heptamer 7a, on the other hand, forms only four 11-helical
hydrogen bonds in the solid state because the N-terminal Ala
residue is not part of the helix, which parallels the behavior of
the N-terminal R-residues in the three hexamers.
Among the four octamers, 8 and 8a are fully 11-helical in
the solid state (all six of the possible i,i+3 CdO· · ·H-N
hydrogen bonds are formed), while 8b and 8c display confor-
mations that contain both 11-helical and 14/15-helical hydrogen
bonds (Figure 2). Octamers 8b and 8c contain only four 11-
helical hydrogen bonds in the solid state. For 8b, the carbonyl
of the N-terminal Boc group is engaged in a 14-membered ring
hydrogen bond with the amide proton of ACPC-4 (i,i+4
CdO· · ·H-N), which represents the beginning of a 14/15-helix.
A similar interaction seems to occur in 8c, but in this case the
O· · ·H distance, 2.7 Å, is a little longer than expected for a
hydrogen bond. The N-H· · ·O angle for this interaction is
favorable (147°).
The three longest among our R/ꢀ-peptides, which contain 9
or 10 residues (9, 9a, and 10), display 14/15-helical conforma-
tions in the solid state (Figure 3). Nonamers 9 and 9a are 14/
15-helical: five and six i,i+4 CdO· · ·H-N hydrogen bonds are
evident, respectively. Decamer 10, too, contains six i,i+4
CdO· · ·H-N hydrogen bonds; formation of the seventh is
precluded because the N-terminal Ala residue curls away from
the helix in the solid state. This N-terminal R-residue distortion
is similar to those observed for hexamers 6, 6a, and 6b and
heptamer 7.
9
6546 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Helical Secondary Structures in R/ꢀ-Peptides
A R T I C L E S
Table 1. Average Interproton Distances (Å) Corresponding to
Implications from the Crystallographic Data Regarding
Helical Folding in Solution. Our initial studies of R/ꢀ-peptides
containing ACPC and similarly constrained ꢀ-residues focused
on hexamers and octamers.3a These molecules were studied in
solution, and two-dimensional NMR data were used to elucidate
conformational behavior. Numerous medium-range NOEs were
observed for pairs of protons from residues not adjacent in
sequence, which provided strong evidence for helix formation
in methanol solution. However, accommodating all of the i,i+2,
i,i+3, and i,i+4 NOE patterns within a single helix was possible
only by proposing a network of three-center or “bifurcated”
hydrogen bonds, involving simultaneous i,i+3 and i,i+4
CdO· · ·H-N interactions. Alternatively, we could account for
the observed NOEs by proposing that two distinct helical
conformations were populated, one containing only i,i+3
CdO· · ·H-N hydrogen bonds (the 11-helix) and the other
containing only i,i+4 CdO· · ·H-N hydrogen bonds (the 14/
15-helix), with rapid interconversion between these two helical
conformations on the NMR time scale. We favored the latter
hypothesis based on two considerations. First, bifurcated
hydrogen bonds are rare in R-peptide and protein helices.11
Second, interconversion between i,i+3 and i,i+4 CdO· · ·H-N
hydrogen bonding patterns, i.e., between the 310- and R-helices,
has been reported for R-peptides.12 However, the NMR data
did not allow us to disprove either of these competing
hypotheses.
Medium-Range NOE Patterns
11-helix
14/15-helix
NOE type
NOEa crystalb NOEa crystalb
ꢀ-residue CꢀH(i)-ꢀ-residue NH(i+2)
yes 3.2(22) yes 4.2(12)
ꢀ-residue CꢀH(i)-ꢀ-residue CRH (i+2) yes 2.9(22) yes 3.9(11)
ꢀ-residue CRH(i)-ꢀ-residue NH(i+2)
R-residue CRH(i)-R-residue NH(i+2)
ꢀ-residue CꢀH(i)-R-residue NH(i+3)
R-residue CRH(i)-ꢀ-residue NH(i+3)
yes 4.2(22) yes 4.2(11)
yes 3.9(4) no 5.7(6)
yes 3.6(14) yes 2.7(9)
yes 3.8(4)
5.4(4)
yes 4.1(6)
yes 3.6(4)
R-residue CRH(i)-ꢀ-residue CRH (i+3) no
ꢀ-residue CꢀH(i)-ꢀ-residue NH(i+4)
R-residue CRH(i)-R-residue NH(i+4)
no
no
6.2(11) yes 3.4(7)
7.1(2) yes 4.2(4)
a “Yes” indicates that NOEs of this type should be observed in
solution if the indicated helix is formed. “No” indicates that the
indicated helix would not give rise to this type of NOE. b The number
in parentheses indicates number of measured and averaged distances.
is known among R-peptides: the R-helix (i,i+4 hydrogen bonds)
is favored over the 310-helix (i,i+3 hydrogen bonds) by
increasing the number of residues.12a This R-peptide preference
has been suggested to arise from a competition between a drive
to form the maximum number of intramolecular hydrogen bonds
(favored by the smaller hydrogen-bonded ring) and a drive to
minimize backbone conformational energy (apparently favored
by the larger hydrogen-bonded ring).
The proposed length-dependence of helix preference among
R/ꢀ-peptides is strongly supported by the crystallographic data
discussed here. The length effect in our structure set is very
distinct. All 11 R/ꢀ-peptides that contain 8 residues or fewer
are 11-helical in the solid state, while all 3 R/ꢀ-peptides that
contain 9 residues or more are 14/15-helical. The observation
of a single 14-membered-ring hydrogen bond in octamer 8b,
and perhaps another in 8c, seems to hint that the allure of the
14/15-helical conformation becomes just barely perceptible at
this length, at least within this particular set of R/ꢀ-peptides
(Figure 2b). Of course, it is precarious to extrapolate from
conformational behavior observed in crystal structures to
conformational behavior in solution, but the size of our data
set and the clarity of the trend suggest that such extrapolation
is reasonable in this case.
Relationships between H · · ·H Distances and Conforma-
tionally Diagnostic NOEs. Table 1 lists the nine types of
medium-range NOEs between backbone protons that have been
observed for R/ꢀ-peptides containing ACPC and similarly
constrained ꢀ-residues.3a On the basis of simple computational
modeling, we previously proposed that five of these NOE
patterns, three involving i,i+2 sequence relationships and two
involving i,i+3 sequence relationships, could arise from either
the 11-helix or the 14/15-helix, but that the other four NOE
patterns would originate from only one helix or the other.
Specifically, we proposed that i,i+2 R-residue CRH f R-residue
NH NOEs would be characteristic of the 11-helix and that i,i+3
R-residue CRH f ꢀ-residue CRH, i,i+4 ꢀ-residue CꢀH f
ꢀ-residue NH, and i,i+4 R-residue CRH f R-residue NH NOEs
would be characteristic of the 14/15-helix. These hypotheses
constitute powerful tools for interpretation of R/ꢀ-peptide NOE
data in terms of helical folding.
The 14 R/ꢀ-peptide crystal structures reported here provide
strong albeit indirect support for the hypothesis that R/ꢀ-peptides
in solution can interconvert between 11- and 14/15-helical
conformations rather than adopting a hybrid helical conformation
that contains bifurcated hydrogen bonds. If one uses a standard
structural criterion for identifying hydrogen bonds in the solid
state, O· · ·H 2.5 Å,13 then there is not a single intramolecular
bifurcated interaction among the 14 structures. If one relaxes
this constraint slightly, to include O· · ·H distances up to 2.8Å,
then there is one possible bifurcated interaction at the C-terminus
of nonamer 9 and another at the C-terminus of 9a. Overall, the
set of crystal structures suggests that forming two-center O · · ·H
interactions is strongly preferable to forming three-center
interactions, even when the backbone seems to have trouble
deciding between the i,i+3 and i,i+4 CdO· · ·H-N hydrogen-
bonding patterns (cf. structures of 8b and 8c).
Relationship between r/ꢀ-Peptide Length and Helix Pre-
ference. We have previously examined two 15-mer R/ꢀ-peptides
by two-dimensional NMR, and in both cases the set of medium-
range NOEs observed suggested that only the 14/15-helical
conformation was populated to a significant extent.3b,g These
results contrasted with previous conclusions based on the NOEs
displayed by shorter R/ꢀ-peptides, which implied that both the
11- and 14/15-helical conformations are significantly populated.
This contrast led us to propose that the 14/15-helix is increas-
ingly favored relative to the 11-helix as the R/ꢀ-peptide
backbone grows longer. A comparable helix preference trend
(11) (a) Baker, E. N.; Hubbard, R. E. Prog. Biophys. Mol. Biol. 1984, 44,
97. (b) Yang, J.; Gellman, S. H. J. Am. Chem. Soc. 1998, 120, 9090.
(c) Yang, J. H.; Christianson, L. A.; Gellman, S. H. Org. Lett. 1999,
1, 11.
The crystallographic data allow incisive tests of the
previously proposed relationships between backbone NOE
patterns and specific helical conformations of R/ꢀ-peptides.
For each type of medium-range NOE, Table 1 provides the
average of all the relevant H · · · H distances among our 14
structures and the number of measurements contributing to
(12) (a) Bolin, A. K.; Millhauser, G. L. Acc. Chem. Res. 1999, 32, 1027.
(b) Carlotto, S.; Paola, C.; Zerbetto, M.; Franco, L.; Corvaja, C.;
Crisma, M.; Formaggio, F.; Toniolo, C.; Polimero, A.; Barone, V.
J. Am. Chem. Soc. 2007, 129, 11248. (c) Crisma, M.; Saviano, M.;
Moretto, A.; Broxterman, Q. B.; Kaptein, B.; Toniolo, C. J. Am. Chem.
Soc. 2007, 129, 15471.
(13) McDonald, I. K.; Thornton, J. M. J. Mol. Biol. 1994, 238, 777.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6547

A R T I C L E S
Choi et al.
Figure 4. Ramachandran-type plots for R/ꢀ-peptides.
Table 2. Average Backbone Torsion angles (deg)
this average. (It should be noted that an H · · · H pair was
included in this analysis only if both of the H atoms occurred
within a helical conformation; H atoms on nonhelical
N-terminal residues of 6, 6a, 6b, 7a, and 10 were not
included.) In each case, our earlier deductions from modeled
structures have been borne out. Thus, for example, the i,i+2
R-residue CRH f R-residue NH NOE pattern appears to arise
exclusively from the 11-helix, because the four relevant
H · · · H pairs among the 11-helical crystal structures have an
average separation of 3.9 Å, which is within the NOE
detection range of ∼5 Å, while the six relevant H · · · H pairs
among the 14/15-helical crystal structures have an average
separation of 5.7 Å, which is too long for NOE detection.
Overall, these results provide a retrospective validation of
earlier NOE-based conformational analyses of 1:1 R/ꢀ-
peptides and a firm foundation for analogous future studies.
Backbone Torsion Angle Analysis. The set of 14 R/ꢀ-peptide
crystal structures allows us to identify average backbone
torsion angles for both R-residues and ꢀ-residues in the 11-
helical and 14/15-helical conformations. The data are pre-
sented in Ramachandran-type plots14 in Figure 4, and the
average torsion angles are summarized in Table 2. ꢀ-Residues
have an additional torsion, corresponding to rotation about
the CR-Cꢀ backbone bond, relative to R-residues; this
torsion-distinctive ꢀ-residue torsion angle has been designated
θ.2a The ꢁ and ψ torsion angles of ꢀ-residues, corresponding
to rotations about N-Cꢀ and CR-C(dO), respectively, are
defined by analogy to the ꢁ and ψ torsion angles of
R-residues. Because there are three torsion angles of interest
for each ꢀ-residue, the graphical summary includes three two-
11-helixb
14/15-helixb
310-helixa
R-helixa
R-residue
ꢀ-residue
R-residue ꢀ-residue
ꢁ
ψ
θ
-49
-26
-58
-47
-56 (25) -99 (27) -62 (16) -129 (14)
-40 (25) -88 (27) -38 (16) -117 (14)
93 (27)
81 (14)
a Reference 14. b The number in parentheses indicates number of
measured and averaged angles.
dimensional Ramachandran-type plots so that information on
all three torsion angles is available.
The ꢁ and ψ torsion angles of each helical R-residue fall
in the standard R-helical region of the Ramachandran plot,
with no apparent distinction between 11- and 14/15-helical
R-residues. (The few R-residues that do not participate in
helical conformations, because of distortion at R/ꢀ-peptide
termini, are not included in this analysis.) In contrast to the
similarity among R-residue torsion angles, the ꢀ-residue
torsion angles differ in small but distinct ways between
the two helical conformations. This difference is smallest,
∼10° on average, for the θ torsion angle, which is constrained
by the cyclopentane ring. The ꢀ-residue ꢁ and ψ torsion angle
averages differ by ∼30° between the 11- and 14/15-helices.
Overall, these results suggest that it might be possible to
achieve higher stability for either helix, and higher preference
for one over the other, by additional preorganization of the
ꢀ-residues.
Helical Parameter Analysis. Average parameters for both 11-
and 14/15-helices were deduced from the 14 crystal structures.
Every helical parameter was calculated from a set of four
(14) Ramachandran, G. N.; Sasisekharan, V. AdV. Protein Chem. 1968,
23, 283.
9
6548 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008

Helical Secondary Structures in R/ꢀ-Peptides
A R T I C L E S
Table 3. Average Structural Parameters of R/ꢀ-Peptide Helices
drawn from NOE analysis. Overall, this set of high-resolution
structures should facilitate function-based design of R/ꢀ-
peptides.
helix type
res/turn, n
rise/turn, p (Å)
rise/res, d (Å)
radius, r (Å)
11-helix
2.8
3.6
3.2
3.5
5.6
5.0
5.8
5.4
2.0
1.4
1.8
1.5
2.1
2.6
2.0
2.3
14/15-helix
310-helixa
R-helixa
Experimental Section
General Procedure for Dipeptide Segment Preparation.
Dipeptide segments Boc-Ala-ACPC-OH (11) and Boc-Aib-ACPC-
OH (12) were prepared by the previously reported procedure.3j The
known ACPC intermediate (1 equiv), ethyl (1S,2S)-2-[(1′S)-
phenylethyl]aminocyclopentane carboxylate hydrochloride,10b,c
was dissolved in methanol. The mixture was shaken on a Parr
apparatus under H2 (50 psi) with 10% Pd/C (50 wt %) for 48 h.
After the reaction was complete, the mixture was filtered through
Celite, and the filtrate was concentrated to give the HCl salt form
of trans-2-aminocyclopentane carboxylate (1 equiv), which was
added directly to a solution of Boc-Aib-OH or Boc-Ala-OH (1
equiv), EDCI (1.5 equiv), HOBt (1.3 equiv), and DIEA (1.2 equiv)
in DMF. The resulting solution was stirred for 60 h. The mixture
was diluted with excess amount of EtOAc, washed with aqueous
10% citric acid, aqueous saturated NaHCO3, and brine. The organic
layer was dried over MgSO4 and concentrated to give a crude
product, which was purified by silica gel column chromatography
to give the desired dipeptide ethyl ester, Boc-Aib-ACPC-OEt (13)
or Boc-Ala-ACPC-OEt (14).
a Reference 13; mean values in globular proteins.
Boc-Aib-ACPC-OEt (13). 1H NMR (300 MHz, CDCl3): δ 6.70
(br s, 1H), 4.85 (br s, 1H), 4.35 (quintet, J ) 7.2 Hz, 1H), 4.13 (q,
J ) 7.1 Hz, 2H), 2.62 (q, J ) 7.9 Hz, 1H), 2.20-1.50 (m, 6H),
1.47 (s, 3H), 1.46 (s, 3H), 1.44 (s, 9H), 1.24 (t, J ) 7.1 Hz, 1H).
ESI-TOF MS: m/z 343.4 [M + H]+, 365.4 [M + Na]+, 707.7 [2M
+ Na]+.
Boc-Ala-ACPC-OEt (14). 1H NMR (300 MHz, CDCl3): δ 6.34
(br s, 1H), 5.04 (br s, 1H), 4.37 (quintet, J ) 7.5 Hz, 1H), 4.14 (q,
J ) 7.1 Hz, 2H), 4.19-4.03 (m, 1H), 2.60 (q, J ) 8.1 Hz, 1H),
2.20-1.67 (m, 6H), 1.44 (s, 9H), 1.33 (d, J ) 6.8 Hz, 1H), 1.24 (t,
J ) 7.1 Hz, 1H). ESI-TOF MS: m/z 329.4 [M + H]+, 351.4 [M +
Na]+, 679.7 [2M + Na]+.
Figure 5. Cartoon representation of R/ꢀ-peptide helices: (a) 11-helix
(orange, 8a) (b) 14/15-helix (green, 10). Views perpendicular (top) or along
(bottom) the helical axis.
consecutive R-carbons by a previously reported method.15 Only
helical backbone R-carbons were used for these calculations.
For ꢀ-residues, the midpoint between CR and Cꢀ atoms was used
as an imaginary atom for the calculations. Average helical
parameters are listed in Table 3. Trends between the 11- and
14/15-helices are analogous to those between the 310-helix and
R-helix. For example, the number of residues per turn (n) of
the 11- and 14/15-helices is 2.8 and 3.6, respectively, while
the parameters for 310- and R-helices are 3.2 and 3.5.13 The
11-helix has a longer rise per residue (d) and shorter helix radius
(r) than the 14/15-helix (Figure 5), and comparable differences
are observed between the 310- and R-helices.
Dipeptide segment (13 or 14) (1 equiv) was dissolved in MeOH/
H2O (v/v ) 2:1) to generate a 0.1 M solution. LiOH ·H2O (5 equiv)
was added at 0 °C, and the mixture was stirred for 6 h at 0 °C.
After most of the solvent was evaporated by a nitrogen gas stream,
aqueous 1 M HCl was added until pH ∼2. The turbid mixture was
extracted with EtOAc, and the organic fraction was washed with
brine, dried over MgSO4, and concentrated in vacuo to give the
carboxylic acid form of the dipeptide segment (11 or 12), which
was used in peptide coupling reactions without further purification.
General Procedure for Peptide Synthesis. To a solution of
amine (1 equiv) and acid (1 equiv) in DMF (∼0.1 M) were added
EDCI (1.5 equiv) and DMAP (1.1 equiv). The mixture was
stirred at rt for ∼60 h. Workup and purification methods were
similar to those described for dipeptide segment preparation.
Boc-Aib-ACPC-Aib-ACPC-Aib-OBn (5). Boc-Aib-ACPC-OH
(11) was coupled with HCl ·H-Aib-OBn by the general procedure
Conclusions. The large set of R/ꢀ-peptide crystal structures
reported here shows that use of ꢀ-residues with a five-
membered-ring constraint and trans disposition of the amino
and carboxyl groups promotes folding into two distinct helical
conformations,whichcontaineitheri,i+3ori,i+4CdO· · ·H-N
hydrogen bonds. None of the 14 crystal structures display
an alternative ”hybrid” helix containing bifurcated hydrogen
bonds, which suggests that this hypothetical conformation
does not occur in solution. Length-dependent variations in
helical conformation are consistent with the conclusion
previously deduced via NMR that the i,i+4 hydrogen-bonding
pattern is favored for longer R/ꢀ-peptides. Analysis of key
H · · · H distances in the crystal structures validates conclusions
regarding R/ꢀ-peptide folding in solution that were previously
1
described above to give Boc-Aib-ACPC-Aib-OBn (15). H NMR
(300 MHz, CDCl3): δ 8.06 (br s, 1H), 7.38-7.23 (m, 5H), 6.58
(br s, 1H), 5.13 (ABq, JAB ) 12.6 Hz, ∆υ ) 0.06 ppm, 2H), 4.80
(br s, 1H), 4.11 (m, 1H), 2.58 (m, 1H), 2.12 (m, 1H), 1.93-1.58
(m, 6H), 1.55 (s, 3H), 1.54 (s, 3H), 1.46 (s, 3H) 1.43 (br s, 12H).
ESI-TOF MS: m/z 490.5 [M + H]+, 512.5 [M + Na]+, 1001.9
[2M + Na]+.
Tripeptide 15 (1 equiv) was treated with 4.0 M HCl in dioxane
(∼10 equiv). The mixture was stirred for 30 min and then
concentrated under a nitrogen gas stream to give the HCl salt form
of the amine, which was coupled with Boc-Aib-ACPC-OH by a
general coupling method to give the desired product 5. The X-ray
quality crystal was grown from a chloroform/ether/n-heptane
mixture. 1H NMR (300 MHz, CDCl3): δ 7.86 (d, J ) 8.6 Hz, 1H),
7.76 (s, 1H), 7.39-7.22 (m, 6H), 6.51 (d, J ) 8.9 Hz, 1H), 5.58
(s, 1H), 5.16 (ABq, JAB ) 10.0 Hz, ∆υ ) 0.07 ppm, 2H), 4.35
(quintet, J ) 8.1 Hz, 1H), 4.26 (quintet, J ) 8.6 Hz, 1H), 2.67 (q,
(15) (a) Sugeta, H.; Miyazawa, T. Biopolymers 1967, 5, 673. (b) Kahn,
P. C. Comput. Chem. 1989, 13, 185.
9
J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008 6549

A R T I C L E S
Choi et al.
J ) 7.6 Hz, 1H), 2.22-1.63 (m, 13H), 1.60 (s, 3H), 1.55 (s, 3H),
1.51 (s, 3H), 1.48 (s, 3H), 1.46 (s, 9H), 1.39 (s, 3H), 1.37 (s, 3H).
ESI-TOF MS: m/z 686.8 [M + H]+, 708.8 [M + Na]+, 1394.7
[2M + Na]+.
Other R/ꢀ-peptides were prepared by segment coupling analogous
to those described above. Characterization data may be found in
the Supporting Information.
the Samsung Scholarship Foundation. X-ray equipment purchase
was supported in part by grants from the NSF.
Supporting Information Available: Characterization data,
crystallographic information, full table of helical parameters.
This material is available free of charge via the Internet at http://
pubs.acs.org.
Acknowledgment. This work was supported by the NSF grant
CHE-0551920. S.H.C. was supported in part by a fellowship from
JA800355P
9
6550 J. AM. CHEM. SOC. VOL. 130, NO. 20, 2008