Stabilization of folded peptide and protein structures via distance matching with a long, rigid cross-linker

Article abstract of DOI:10.1021/ja075829t

Intramolecular cross-linking is predicted to stabilize the folded state of peptides and proteins most effectively if the cross-linker provides a rigid link that is well-matched in end-to-end distance with attachment sites in the peptide or protein. We des

Full text of DOI:10.1021/ja075829t

Published on Web 10/26/2007
Stabilization of Folded Peptide and Protein Structures via Distance Matching
with a Long, Rigid Cross-Linker
Fuzhong Zhang, Oleg Sadovski, Steven J. Xin, and G. Andrew Woolley*
Department of Chemistry, UniVersity of Toronto, 80 St. George Street, Toronto, ON, M5S 3H6, Canada
Received August 15, 2007; E-mail: awoolley@chem.utoronto.ca
The functions of peptides and proteins often depend critically
on their folded three-dimensional structures. However, the folded
states of most proteins are only e10 kcal/mol lower in energy than
their unfolded states.1 Short peptide sequences are usually unfolded
in aqueous solution, although they adopt specific folded structures
when bound to their biological targets. Stabilizing the active folded
forms of peptide or proteins is thus important for maintaining or
enhancing the functions of these molecules under a wide variety
of conditions. Owing to its prevalence as a secondary structural
element in proteins, the R-helix has received considerable attention
as a target for conformational stabilization. Stabilized helices have
been employed as biological agents, for example, against microbial
Figure 1. (a) Structures of cross-linkers. (b) Model of helical peptide with
Cys residues at i, i+11 positions cross-linked with 1. (c) Distance
distributions of cross-linkers calculated using Langevin dynamics simula-
infections,2 and for targeting tumor suppressor proteins3 and proteins
involved in apoptosis.4
Methods developed to stabilize peptide R-helical structures
include introduction of R,R-dialkyl amino acids,5 cross-linking
amino acid side chains via disulfide bonds,6 metal chelates,7 lactam
bridges,8,9 or via ruthenium-catalyzed ring closing metathesis (RCM)
to form cyclic peptides.10 These methods require peptide synthesis
using non-natural amino acids. Recently, Fujimoto et al. introduced
acetylenic cross-linking agents that could be introduced via pairs
of naturally occurring lysine side chains;11 however, the degree of
helix stabilization was moderate.
In the absence of specific favorable interactions between a cross-
linker and the folded state of a peptide or protein, the mechanism
of stabilization appears to be primarily the decrease in conforma-
tional entropy of the unfolded state caused by an intramolecular
covalent linkage.12,13 We anticipate that maximal stabilization should
occur with (i) optimal matching between the cross-linker length
and the distance distribution between the two attachment points of
the folded peptide or protein, (ii) enhanced rigidity of the cross-
linker, and (iii) a long cross-linker that can bridge sites far apart in
amino acid sequence. Here we introduce a water-soluble, thiol-
reactive cross-linking reagent (EY-CBS) (3,3′-ethyne-1,2-diylbis-
{6-[(chloroacetyl)amino]benzenesulfonic acid} (1) that can be used
to introduce long, rigid bridges into peptides and proteins. We
compare its stabilizing effects on helices as well as on a small
â-sheet protein with the effects of more flexible linkers, including
the commercially available 1,4-di[3′-(2′-pyridyldithio]propionami-
do]butane (DPDPB).
Figure 1 shows the structure of EY-CBS (1), synthesized in four
steps (see Supporting Information), together with an analogue EA-
CBS (2) with a single bond in place of the central triple bond, and
the flexible commercial cross-linker DPDPB (3). EY-CBS is
designed as linear, symmetric, and Cys-reactive to minimize the
number of possible conformations of the cross-link with respect to
the peptide backbone attachment points. Sulfonate groups are
included for water solubility since introduction of a large hydro-
phobic cross-linker to a protein surface may lead to aggregation or
insolubility. Whereas each of these cross-linkers can span a ∼20
Å distance, their end-to-end distance probability distributions as
tions. Cys-Cys sulfur distances corresponding to different side-chain
rotamers in an ideal helix with Cys residues spaced at i, i+11 and a
T14CE33C-FynSH3 mutant are also shown.
calculated using Langevin dynamics14 vary considerably (Figure
1). EY-CBS gives a narrow distance distribution centered at 18.7
Å, and EA-CBS gives a bimodal distribution due to the three
possible rotamers around the central carbon-carbon single bond.
DPDPB, although marketed as a cross-linker with a spacer length
of 19.9 Å,15 gives a broad distribution with numerous short
conformers.
An end-to-end distance of 18.7 Å closely matches the sulfur-
to-sulfur distance in an ideal R-helix with Cys residues spaced at
i, i+11 positions (Figure 1). These positions span three helical turns
and are located on the same face of the helix so that they can be
cross-linked without steric inference between the cross-linker and
protein side chains.
We first applied these cross-linkers to FK11W, a short peptide
sequence that has been used as a model helix previously.16 Cross-
linking reactions were carried out in aqueous solution at mildly
basic pH and ambient temperature. All the cross-linking reactions
proceeded to more than 95% completion when 4 equiv of cross-
linker was present. There was no detectable amount of doubly
reacted peptide (two cross-linkers attached to one peptide), indicat-
ing the cross-linking step, an intramolecular reaction, is much faster
than a second intermolecular reaction.
The effect of cross-linking on peptide structure was examined
by circular dichroism (CD). At room temperature, the CD spectra
of FK11W cross-linked with each of 1-3 showed an increase in
R-helix as compared with the uncross-linked peptide (Figure 2a).
The order in increased helix percentage was 1 > 2 > 3 with the
flexible DPDPB cross-linked peptide only slightly more helical than
uncross-linked FK11W, much less than the rigid EY-CBS-FK11W.
There was no evidence of peptide self-association from 2 to 300
µM as determined by CD.
Next, we examined the response of these cross-linked peptides
to thermal melting. The R-helical contents of the peptides were
evaluated on the basis of the mean residue ellipticity at 222 nm16
9
14154
J. AM. CHEM. SOC. 2007, 129, 14154-14155
10.1021/ja075829t CCC: $37.00 © 2007 American Chemical Society

C O M M U N I C A T I O N S
file as input. This procedure identified two residues T14 and E33
located on two loops of the FynSH3 domain with a spacing that
matched the distance of the cross-linker well (Figure 1, Supporting
Information Figure S1).
Figure 3 shows CD spectra and thermal melting curves for
FK22C and T14CE33C-FynSH3 with and without cross-linkers.
Cross-linking of FK22C increased the helix content to almost 100%
at 2 °C and enhanced helix content by more than 10% from 30 to
90 °C. At 37 °C, the helicity was increased from 49 to 60%. The
CD spectrum of cross-linked T14CE33C-FynSH3 showed little
difference with that of uncross-linked FynSH3 at 25 °C, indicating
that EY-CBS-T14CE33C-FynSH3 is fully folded at this temperature
and that cross-linking does not significantly disturb the native folded
state of SH3. The thermal melting curve indicated that the cross-
linker enhanced the conformational stability of T14CE33C-FynSH3,
causing an increase in the Tm by almost 10 °C, despite the fact that
it links flexible loops in the protein rather than more rigid elements
of secondary structure.
Figure 2. (a) CD spectra of uncross-linked FK11W (black), EY-CBS-
FK11W (red), EA-CBS-FK11W (green), DPDPB-FK11W (blue) at 25 °C.
(b) Thermal melting curves FK11W (black), EY-CBS-FK11W (red), EA-
CBS-FK11W (green), DPDPB-FK11W (blue).
These results demonstrate the effectiveness of a long rigid cross-
linker at conformational stabilization of peptides and proteins.
Flexible linkers that can adopt similar overall lengths, even an
analogue that differs in flexibility only at one bond, are demon-
strably less effective. The strategy is particularly useful for
stabilizing R-helical segments with Cys residues at i and i+11
positions but also appears applicable to larger structures.
Acknowledgment. F.Z. is supported by the CIHR training
program in protein folding. This work was funded by the NSERC
and the CIHR.
Supporting Information Available: Synthesis, 1H NMR, 13C NMR
of compounds 1 and 2. Synthesis of peptides, expression and purifica-
tion of protein and cross-linking reactions. Modeling of cross-linker
distance distributions. This material is available free of charge via the
Internet at http://pubs.acs.org.
Figure 3. (a) CD spectra of FK22C (black) and EY-CBS-FK22C (red) at
25 °C. (b) Thermal melting curves of FK22C (black) and EY-CBS-FK22C
(red). (c) CD spectra of T14CE33C-FynSH3 (black) and EY-CBS-
T14CE33C-FynSH3 (red) at 25 °C (d) Thermal melting curves of
T14CE33C-FynSH3 (black) and EY-CBS-T14CE33C-FynSH3 (red) at
25 °C
References
(1) Rose, G. D.; Fleming, P. J.; Banavar, J. R.; Maritan, A. Proc. Natl. Acad.
Sci. U.S.A. 2006, 103, 16623-16633.
over the temperature range of 2-98 °C (Figure 2b). At 2 °C, the
peptide cross-linked with 1 is almost 100% helical. Even at 62 °C,
where the uncross-linked FK11W has already completely unfolded,
the EY-CBS-FK11W is still 44% helical, more helical than uncross-
linked FK11W at 2 °C. At 37 °C, the EY-CBS cross-link increases
FK11W helicity from 16 to 71%. Introduction of a single bond in
place of the triple bond in an otherwise identical cross-linker
decreases its effectiveness at conformational stabilization (Figure
2b). The very flexible DPDPB-FK11W showed only slight stabi-
lization (up to 14%).
To explore the effectiveness of 1 at conformational stabilization
of larger, more complex targets, we examined its effects on FK22,
a 32-residue R-helical peptide and on the FynSH3 domain, a small
â-sheet protein. We expect that the extent of stabilization in these
cases will in general be smaller and will depend on the extent to
which the cross-linker can influence distinct segments of the protein.
Cys residues were installed near the C-terminal end of FK22 with
an i, i+11 spacing, thereby targeting about half the total length of
the peptide. To identify sites for cross-linking in the FynSH3
domain, we used the program sGAL,17 which identifies surface-
exposed pairs of residues a specified distance apart taking a PDB
(2) Yokum, T. S.; Elzer, P. H.; McLaughlin, M. L. J. Med. Chem. 1996, 39,
3603-3605.
(3) Bernal, F.; Tyler, A. F.; Korsmeyer, S. J.; Walensky, L. D.; Verdine, G.
L. J. Am. Chem. Soc. 2007, 129, 2456-2457.
(4) Walensky, L. D.; Pitter, K.; Morash, J.; Oh, K. J.; Barbuto, S.; Fisher, J.;
Smith, E.; Verdine, G. L.; Korsmeyer, S. J. Mol. Cell. 2006, 24, 199-
210.
(5) Andrews, M. J. I.; Tabor, A. B. Tetrahedron 1999, 55, 11711-11743.
(6) Jackson, D. Y.; King, D. S.; Chmielewski, J.; Singh, S.; Schultz, P. G. J.
Am. Chem. Soc. 1991, 113, 9391-9392.
(7) Ghadiri, M. R.; Fernholz, A. K. J. Am. Chem. Soc. 1990, 112, 9633-
9635.
(8) Bracken, C.; Gulyas, J.; Taylor, J. W.; Baum, J. J. Am. Chem. Soc. 1994,
116, 6431-6432.
(9) Phelan, J. C.; Skelton, N. J.; Braisted, A. C.; McDowell, R. S. J. Am.
Chem. Soc. 1997, 119, 455-460.
(10) Schafmeister, C. E.; Po, J.; Verdine, G. L. J. Am. Chem. Soc. 2000, 122,
5891-5892.
(11) Fujimoto, K.; Oimoto, N.; Katsuno, K.; Inouye, M. Chem. Commun. 2004,
1280-1281.
(12) Nagi, A. D.; Regan, L. Fold. Des. 1997, 2, 67-75.
(13) Zhou, H. X. Acc. Chem. Res. 2004, 37, 123-130.
(14) Green, N. S.; Reisler, E.; Houk, K. N. Protein Sci. 2001, 10, 1293-1304.
(15) Cross-linking reagents: Pierce Chemical Co., 2007.
(16) Flint, D. G.; Kumita, J. R.; Smart, O. S.; Woolley, G. A. Chem. Biol.
2002, 9, 391-397.
(17) Woolley, G. A.; Lee, E. S.; Zhang, F. Bioinformatics 2006, 22,
3101-3102.
JA075829T
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14155