Hairpin folding behavior of mixed α/β-peptides in aqueous solution

Article abstract of DOI:10.1021/ja2002346

The invention of new strategies for the design of protein-mimetic oligomers that manifest the folding encoded in natural amino acid sequences is a significant challenge. In contrast to the α-helix, mimicry of protein β-sheets is less understood. We report

Full text of DOI:10.1021/ja2002346

COMMUNICATION
pubs.acs.org/JACS
Hairpin Folding Behavior of Mixed r/β-Peptides in Aqueous Solution
George A. Lengyel, Rebecca C. Frank, and W. Seth Horne*
Department of Chemistry, University of Pittsburgh, Pittsburgh, Pennsylvania 15260, United States
S Supporting Information
b
lagged behind R-helical model systems7 by more than two
ABSTRACT: The invention of new strategies for the design
of protein-mimetic oligomers that manifest the folding
encoded in natural amino acid sequences is a significant
challenge. In contrast to the R-helix, mimicry of protein β-
sheets is less understood. We report here the aqueous
folding behavior of a prototype R-peptide hairpin model
sequence varied at cross-strand positions by incorporation
of 16 different β-amino acid monomers. Our results provide
a folding propensity scale for β-residues in a protein β-sheet
context as well as high-resolution structures of several
mixed-backbone R/β-peptide hairpins in water.
decades. Sheet folding has been demonstrated in pure β-pep-
tides, however, only in organic solvent.8 In mixed backbones, it
has been shown that one class of β-residue (β3) can form sheet
structures in organic solvent, in the solid state, and in two-
dimensional (2D) crystals at an air/water interface.9 These prior
studies showed that β-peptides tend to form sheets with a
different hydrogen-bonding pattern than that of R-peptide β-
sheets due to inversion of hydrogen bonding at the unnatural
residue.8,9 These precedents leave several open questions that we
sought to address: (1) Can a mixed backbone R/β-peptide adopt
a sequence-encoded sheet fold in aqueous solution?10,11 (2)
What is the relationship between the structure of a β-residue and
its propensity to be accommodated into a protein sheet? (3) Can
an appropriately structured β-residue promote a protein-like
hydrogen-bonding pattern?
he development of unnatural-backbone oligomers capable of
The minimal β-sheet model is a two-stranded antiparallel β-
hairpin. The first reported sequence to form a discrete hairpin
fold in water was the C-terminal segment of the Streptococcal
protein GB1.6a We employed a known analogue of this GB1
sequence, 12-residue peptide 1a (Figure 1),12 as a model system
T
protein-like folding is the focus of intensive research efforts,1
due to the prospect of these molecules to deliver protein-like
functions (e.g., biomolecular recognition, catalysis) on protease-
resistant backbones. A fundamental challenge in the design of
unnatural species that fold in defined ways is the same faced in
the de novo design of folded proteins:2 the generation of a
sequence of monomer building blocks that will manifest the
correct fold.
Early efforts to develop folded oligomers with tailored func-
tions relied primarily on structure-based engineering of known
scaffolds to display functional groups in defined arrangements.1,3
Recent work has demonstrated an alternative approach in which
natural protein sequences can serve as blueprints for the design of
unnatural mimics. Specifically, the display of natural side-chain
sequences on heterogeneous backbones comprising mixtures of
R- and β-amino acid residues can generate “R/β-peptides”4 that
adopt folds similar to those of the R-peptide prototypes.5
Termed “sequence-based design,” this strategy has produced
protein-like helix-bundle quaternary structures5a,c,e and protease-
resistant inhibitors of biomedically important protein-protein-
binding interactions.5b,d
Applications of sequence-based design to date have focused
exclusively on R-helical prototype sequences.5 This method has
the potential to provide analogues of more complex protein
tertiary structures; however, an unmet need for the realization of
this potential is the development of complementary strategies for
the mimicry of β-sheets. Here, we explore the ability of R/β-
peptides to adopt a sequence-encoded protein β-sheet fold in
aqueous solution.
Compared to helical folding, sheet structures in unnatural
oligomers have received considerably less attention. An interest-
ing parallel exists in the study of folding in natural proteins where
the development of peptides that form discrete β-sheets in water6
Figure 1. Structures of the peptides synthesized and characterized in
the present study. R-Peptides 1a and 1b served as prototype sequences
for incorporation of matched β-residue (“X”) pairs at sequence positions
3 and 10 to produce R/β-peptides 2a-17a and 2b-17b. “DP” denotes
D-Pro.
Received: January 10, 2011
Published: March 03, 2011
r
2011 American Chemical Society
4246
dx.doi.org/10.1021/ja2002346 J. Am. Chem. Soc. 2011, 133, 4246–4249
|

Journal of the American Chemical Society
COMMUNICATION
Figure 2. Summary of NMR chemical shift data for oligomer pairs 1-
17. (A) Separation (Δδ) of the diastereotopic Gly7 HR signals in 1a-
17a. (B) Sum of the absolute values of the chemical shift deviations
(Σ|Δδ|) of backbone C-H signals with mutation of D-Pro6 (1a-17a)
to L-Pro (1b-17b); turn residues 6-7 were excluded from the analysis.
Spectra were acquired at 4 °C in 100 mM NaOAc-d3 buffer, 9:1 H2O/
D2O, pH 3.8. Dashed lines indicate 50% increments of the bar height for
the R-peptide.
Figure 3. Select backbone NOEs observed for R-peptide hairpin 1a and
R/β-peptide analogues 2a, 8a, and 9a. Spectra were acquired at 4 °C in
100 mM NaOAc-d3 buffer, 9:1 H2O/D2O, pH 3.8. Dashed lines indicate
cases where overlap prevented an unambiguous assignment.
to examine the folding of R/β-peptide hairpins in aqueous
solution. Peptide 1a, which contains a four-residue hydrophobic
cluster from GB1 and a turn-promoting D-Pro-Gly segment,13 is
on the razor’s edge of conformational stability with less than 0.5
kcal/mol separating the folded and unfolded states.12 The
instability of 1a makes it and related hairpin sequences highly
sensitive probes for detecting small energetic contributions to
protein folding.14
preorganization through incorporation of a 5-membered (10-
13) or 6-membered (14-17) ring constraint in all possible
stereochemical configurations. An assortment of known routes
were employed for the asymmetric synthesis of the different
protected β-amino acid monomers.16 Peptides were prepared by
microwave-assisted Fmoc solid-phase peptide synthesis, purified
by reverse-phase HPLC, and the identity of each confirmed by
mass spectrometry.
We compared the folding of R-peptide 1a and R/β-peptides
2a-17a in aqueous solution using NMR spectroscopy. Spectra
of each oligomer were acquired in deuterated acetate buffer (pH
3.8) at a temperatureof 4°C (the use of an acidic medium disfavors
aggregation in this sequence12). A series of solvent-suppressed 1D
and 2D homonuclear experiments (TOCSY, COSY, NOESY)
allowed complete assignment of all backbone resonances.
Separation of the diasterotopic HR atoms in turn residue Gly7
is diagnostic of folded population in 1a and related hairpins.14
These two protons have significantly different chemical shifts in
well-folded structures, but tend to coalesce into a single peak in
random coils. We compared the separation (Δδ) of HR resonances
in Gly7 of prototype R-peptide 1a to the separation observed in
R/β-peptide derivatives 2a-17a (Figure 2A). Among the R/β-
peptides, only four (8a, 9a, 12a, and 16a) showed separation of at
We introduced 16 different β-amino acid residues at cross-
strand sequence positions 3 and 10 midway along the hairpin
formed by R-peptide 1a to generate R/β-peptide derivatives
2a-17a (Figure 1). Identical substitutions were made in un-
folded control sequence 1b to generate 2b-17b. Three mono-
mer structural variables were considered: backbone substitution
pattern (β2 vs β3 vs β2,3), backbone stereochemistry, and
constraint of the CR-Cβ torsion by incorporation into a ring.
Collectively, the compounds examined were intended to system-
atically probe the impact of these β-residue structural properties
on the ability of oligomers to incorporate that residue into the
hairpin fold encoded by parent R-peptide sequence 1a.
Oligomers 2-5 incorporate either enantiomer of a β3-residue
(2, 3) or β2-residue (4, 5) bearing an isopropyl side chain.15
Oligomers 6-9 are substituted with all four possible stereo-
isomers of an acyclic disubstituted β2,3-residue bearing an iso-
propyl group at the β3-position and a methyl group at the
β2-position. Oligomers 10-17 probe the role of backbone
4247
dx.doi.org/10.1021/ja2002346 |J. Am. Chem. Soc. 2011, 133, 4246–4249

Journal of the American Chemical Society
COMMUNICATION
Figure 4. NMR solution structures of R-peptide 1a and R/β-peptides 2a, 8a, and 9a. (A) Overlay of the backbone coordinates for the 20 lowest-energy
structures resulting from simulated annealing and (B) minimized average coordinates; carbon atoms are colored cyan in β-residues. (C) Overlays of 1a/
2a, 8a/9a, 1a/8a, and 1a/9a. Some side chains are omitted for clarity.
least 50% that of R-peptide 1a. Of note, incorporation of a trans-
substituted acyclic β2,3-residue (8a or 9a) led to greater-magni-
tude Δδ than the parent sequence; this observation may indicate
that the trans-β2,3-residue promotes hairpin folding more
strongly than the native R-peptide backbone. The folding
behavior of 8a and 9a is consistent with the high sheet propensity
of trans-β2,3-residues in organic solvent,8a,b,e presumably due to a
strong preference for an anti relationship between the two
continuing peptide chains about the central C-C bond.
trans-β2,3-residue pair might be better matched to the chiral L-R-
peptide backbone, the absolute stereochemical configuration of the
β-residue had a minimal impact on the pattern of NOEs observed or
the extent of folding evident from chemical shift analysis. Of note,
the hydrogen-bonding pattern is inverted by each β-residue sub-
stitution (Figure S1, Supporting Information [SI]) with an accom-
panying change in side-chain display along the strand.8a,9a
We performed simulated annealing with NOE-derived dis-
tance restraints to determine the solution structures of R-peptide
1a and the best folded R/β-peptide analogues 2a, 8a, and 9a
(Figure 4A,B). The ensemble of structures resulting from each
calculation showed good internal agreement (Figures 4A and S2
[SI]). The folded conformation of β3-residue-containing oligo-
mer 2a, although reminiscent of R-peptide 1a, showed significant
deviations from a canonical β-sheet fold. The two oligomers with
trans-β2,3-residues (8a, 9a) exhibited folded conformations very
similar to one another (Figure 4C); despite some differences
around the unnatural residue, both were also similar to the fold
of R-peptide 1a. Among all the oligomers examined, R/β-
peptide 8a showed the fold closest to that of the native backbone
hairpin.
A single Rfβ residue substitution in each strand of an R-
peptide hairpin results in a change in the display of side chains as
a result of inversion of backbone hydrogen-bonding at the
position of the unnatural monomer (Figure S1 [SI]).8a,9a This
difference is subtle in the context of folding in a model β-hairpin,
but would be disastrous in the sequence-based modification of a
β-sheet in a protein tertiary structure. Our initial goal was to try
to restore a native hydrogen-bonding pattern with an appro-
priately structured β-residue; however, none of the 16 monomers
we examined achieved this. Nevertheless, our results suggest to
us a strategy that might turn the natural conformational predis-
position of β-residues in sheets from a problem into an advan-
tage. We have demonstrated here that trans-β2,3 residues
maintain high sheet-folding propensity in aqueous solution,
and the recently developed synthetic method16d we employed
for their preparation allows introduction of diverse side chains at
either backbone carbon. We hypothesize that substitution of two
Mutation of the turn-promoting D-Pro6 in 1a to L-Pro (peptide
1
1b) abolishes folding.12 A significant change in backbone H
chemical shifts upon the mutation of D-Pro6 to L-Pro in otherwise
identical sequences can be indicative of hairpin folding; the
magnitude of these deviations is often used to estimate folded
population.12,17 We acquired NMR data for control peptides
1b-17b under the same conditions described above for 1a-
17a. For each oligomer pair (i.e., 1a/1b, 2a/2b, etc.), we deter-
mined the difference in chemical shifts (Δδ) between corres-
ponding protons in the D- and L-Pro mutants. We summed the
absolute values of Δδ for backbone C-H groups adjacent to
nitrogen (excluding turn residues) for each pair (Figure 2B). Six
of the 16 R/β-peptides examined showed backbone chemical
shift deviations at least 50% of that observed for prototype R-
peptide 1a. Among the R/β-peptides, 8a and 9a showed back-
bone chemical shift deviations closest in magnitude to the natural
backbone 1a, similar to the results of Gly HR separation analysis.
We analyzed the NOESY spectra of peptides 1a-17a to
ascertain basic information about their folded structures in solu-
tion. R-Peptide 1a showed long-range NOE correlations con-
sistent with the expected β-hairpin fold (Figure 3A). Among the
R/β-peptide derivatives, some showed no NOE correlations,
consistent with a nativelike hairpin (5a, 12a, 15a-17a); others
(3a, 4a, 6a, 7a, 10a, 11a, 13a, 14a) showed NOEs in the turn
region but no correlations past the unnatural residue. Only three of
the 16 R/β-peptides examined showed evidence of hairpin folding
along the entire length of the chain (Figure 3B-D): one β3-residue
enantiomer (2a) and each enantiomer of the trans-β2,3-monomer
(8a, 9a). Contrary to the expectation that one enantiomer in the
4248
dx.doi.org/10.1021/ja2002346 |J. Am. Chem. Soc. 2011, 133, 4246–4249

Journal of the American Chemical Society
COMMUNICATION
adjacent R-residues in a protein sheet by a single trans-β2,3-resi-
due with side chains derived from the parent dipeptide sequence
will generate a mixed R/β-peptide analogue with native-like
display of side chains (Figure S3 [SI]). We are currently working
to test this hypothesis and related strategies in a larger β-sheet
prototype.
(5) (a) Horne, W. S.; Price, J. L.; Keck, J. L.; Gellman, S. H. J. Am.
Chem. Soc. 2007, 129, 4178–4180. (b) Horne, W. S.; Boersma, M. D.;
Windsor, M. A.; Gellman, S. H. Angew. Chem., Int. Ed. 2008,
47, 2853–2856. (c) Horne, W. S.; Price, J. L.; Gellman, S. H. Proc. Natl.
Acad. Sci. U.S.A. 2008, 105, 9151–9156. (d) Horne, W. S.; Johnson,
L. M.; Ketas, T. J.; Klasse, P. J.; Lu, M.; Moore, J. P.; Gellman, S. H. Proc.
Natl. Acad. Sci. U.S.A. 2009, 106, 14751–14756. (e) Price, J. L.; Horne,
W. S.; Gellman, S. H. J. Am. Chem. Soc. 2010, 132, 12378–12387.
(6) (a) Blanco, F. J.; Rivas, G.; Serrano, L. Nat. Struct. Biol. 1994,
1, 584–590. For reviews, see:(b) Searle, M. S.; Ciani, B. Curr. Opin.
Struct. Biol. 2004, 14, 458–464. (c) Nowick, J. S. Acc. Chem. Res. 2008,
41, 1319–1330.
(7) Brown, J. E.; Klee, W. A. Biochemistry 1971, 10, 470–476.
(8) (a) Krauthauser, S.; Christianson, L. A.; Powell, D. R.; Gellman,
S. H. J. Am. Chem. Soc. 1997, 119, 11719–11720. (b) Seebach, D.; Abele,
S.; Gademann, K.; Jaun, B. Angew. Chem., Int. Ed. 1999, 38, 1595–1597.
(c) Karle, I.; Gopi, H. N.; Balaram, P. Proc. Natl. Acad. Sci. U.S.A. 2002,
99, 5160–5164. (d) Martinek, T. A.; Toth, G. K.; Vass, E.; Hollosi, M.;
Fulop, F. Angew. Chem., Int. Ed. 2002, 41, 1718–1721. (e) Langenhan,
J. M.; Guzei, I. A.; Gellman, S. H. Angew. Chem., Int. Ed. 2003,
42, 2402–2405. (f) Martinek, T. A.; Mandity, I. M.; Fulop, L.; Toth,
G. K.; Vass, E.; Hollosi, M.; Forro, E.; Fulop, F. J. Am. Chem. Soc. 2006,
128, 13539–13544. (g) Rua, F.; Boussert, S.; Parella, T.; Diez-Perez, I.;
Branchadell, V.; Ortuno, R.M. Org. Lett. 2007, 9, 3643–3645.
(9) (a) Karle, I. L.; Gopi, H. N.; Balaram, P. Proc. Natl. Acad. Sci. U.S.
A. 2001, 98, 3716–3719. (b) Gopi, H. N.; Roy, R. S.; Raghothama, S. R.;
Karle, I. L.; Balaram, P. Helv. Chim. Acta 2002, 85, 3313–3330. (c) Roy,
R. S.; Gopi, H. N.; Raghothama, S.; Gilardi, R. D.; Karle, I. L.; Balaram, P.
Biopolymers 2005, 80, 787–799. (d) Segman, S.; Lee, M. R.; Vaiser, V.;
Gellman, S. H.; Rapaport, H. Angew. Chem., Int. Ed. 2010, 49, 716–719.
(10) For a discussion on the significant differences in β-hairpin
folding energetics in water vs organic solvent, see: Maynard, A. J.;
Sharman, G. J.; Searle, M. S. J. Am. Chem. Soc. 1998, 120, 1996–2007.
(11) Sequence-specific peptoid polymers have recently been shown
capable of forming 2D crystalline sheets in aqueous solution. Nam, K. T.;
Shelby, S. A.; Choi, P. H.; Marciel, A. B.; Chen, R.; Tan, L.; Chu, T. K.;
Mesch, R. A.; Lee, B. C.; Connolly, M. D.; Kisielowski, C.; Zuckermann,
R. N. Nat. Mater. 2010, 9, 454–460.
In summary, we have shown that the ability of a mixed-
backbone R/β-peptide to manifest the sheet fold encoded by a
prototype R-peptide sequence depends critically on the structure
of the β-residues employed. Quantitative comparison of folded
populations is precluded here by the lack of a fully folded control
sequence for each monomer examined; however, qualitative
analysis of NMR chemical shift data, observed long-range NOEs,
and the resulting NMR-derived solution structures leads to some
general conclusions. An Rfβ3 replacement is tolerated, while
Rftrans-β2,3 substitution leads to an oligomer that folds as well
or better than the native R-peptide. Backbone torsional con-
straints in the form of 5- or 6-membered rings do not promote
folding in a protein sheet context. Our results provide a family of
unnatural scaffolds for the mimicry of bioactive β-hairpins.18
Moreover, the insights we have obtained into the sequence-based
mimicry of protein β-sheets, combined with earlier studies on the
R-helix,5 deliver a testable strategy for the sequence-based design
of unnatural-backbone oligomers that fold like a natural protein
tertiary structure.
’ ASSOCIATED CONTENT
S
Supporting Information. Complete ref 18b, Figures
b
S1-S3, Tables S1-S7, coordinates for NMR-derived solution
structures, and experimental methods. This material is available
free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
(12) Espinosa, J. F.; Gellman, S. H. Angew. Chem., Int. Ed. 2000,
39, 2330–2333.
(13) Haque, T. S.; Little, J. C.; Gellman, S. H. J. Am. Chem. Soc. 1996,
118, 6975–6985.
Corresponding Author
horne@pitt.edu
(14) For examples, see: (a) Searle, M. S.; Griffiths-Jones, S. R.;
Skinner-Smith, H. J. Am. Chem. Soc. 1999, 121, 11615–11620. (b) Tatko,
C. D.; Waters, M. L. J. Am. Chem. Soc. 2002, 124, 9372–9373.
(15) An isopropyl side chain was used to maintain comparable shape
and hydrophobicity among the acyclic and cyclic β-residues examined.
(16) (a) β3-Residues used are commercially available. (b) β2-Re-
sidues: Chi, Y. G.; Gellman, S. H. J. Am. Chem. Soc. 2006,
128, 6804–6805. (c) cis-β2,3-Residues: Zhu, C.; Shen, X. Q.; Nelson,
S. G. J. Am. Chem. Soc. 2004, 126, 5352–5353. (d) trans-β2,3-Residues:
Xu, X. A.; Wang, K.; Nelson, S. G. J. Am. Chem. Soc. 2007,
129, 11690–11691. (e) trans-Cyclic residues: LePlae, P. R.; Umezawa,
N.; Lee, H. S.; Gellman, S. H. J. Org. Chem. 2001, 66, 5629–5632. (f) cis-
Cyclic residues: Csomos, P.; Kanerva, L. T.; Bernath, G.; Fulop, F.
Tetrahedron: Asymmetry 1996, 7, 1789–1796.
’ ACKNOWLEDGMENT
We thank Prof. Scott Nelson, Brad Hutnick, and Joanne
Bertonazzi for advice on β2,3 monomer synthesis and Dr.
Damodaran Krishnan and Sage Bowser for assistance configuring
solvent-suppressed NMR experiments. Funding for this work
was provided by the University of Pittsburgh.
’ REFERENCES
(1) For reviews, see: (a) Gellman, S. H. Acc. Chem. Res. 1998,
31, 173–180. (b) Hill, D. J.; Mio, M. J.; Prince, R. B.; Hughes, T. S.;
Moore, J. S. Chem. Rev. 2001, 101, 3893–4011. (c) Seebach, D.; Beck,
A. K.; Bierbaum, D. J. Chem. Biodiversity 2004, 1, 1111–1239. (d)
Bautista, A. D.; Craig, C. J.; Harker, E. A.; Schepartz, A. Curr. Opin. Chem.
Biol. 2007, 11, 685–692. (e) Goodman, C. M.; Choi, S.; Shandler, S.;
DeGrado, W. F. Nat. Chem. Biol. 2007, 3, 252–262.
(17) Wishart, D. S.; Sykes, B. D.; Richards, F. M. J. Mol. Biol. 1991,
222, 311–333.
(18) (a) Robinson, J. A. Acc. Chem. Res. 2008, 41, 1278–1288. (b)
Srinivas, N.; et al. Science 2010, 327, 1010–1013.
(2) Dill, K. A.; Ozkan, S. B.; Weikl, T. R.; Chodera, J. D.; Voelz, V. A.
Curr. Opin. Struct. Biol. 2007, 17, 342–346.
(3) For examples of protein recognition, see: (a) Kritzer, J. A.; Lear,
J. D.; Hodsdon, M. E.; Schepartz, A. J. Am. Chem. Soc. 2004,
126, 9468–9469. (b) Sadowsky, J. D.; Schmitt, M. A.; Lee, H. S.;
Umezawa, N.; Wang, S. M.; Tomita, Y.; Gellman, S. H. J. Am. Chem.
Soc. 2005, 127, 11966–11968.
(4) Horne, W. S.; Gellman, S. H. Acc. Chem. Res. 2008,
41, 1399–1408.
4249
dx.doi.org/10.1021/ja2002346 |J. Am. Chem. Soc. 2011, 133, 4246–4249