A minimal protein folding model to measure hydrophobic and CH-π effects on interactions between nonpolar surfaces in water

Article abstract of DOI:10.1002/anie.200700932

(Figure Presented) In the balance: A synthetic molecule that exhibits two-state folding behavior in water is described. Quantitative experiments reveal that the microscopic hydrophobic effect is similar in magnitude to the CH-π, interaction. These forces

Full text of DOI:10.1002/anie.200700932

Angewandte
Chemie
DOI: 10.1002/anie.200700932
Molecular Folding
A Minimal Protein Folding Model To Measure Hydrophobic and CH–p
Effects on Interactions between Nonpolar Surfaces in Water
Brijesh Bhayana and Craig S. Wilcox*
Predicting the pathways of protein folding and quantifying
the relative thermodynamic stability of intermediate and final
states along these pathways constitute two of the most
important challenges in modern chemistry. Such predictions
are difficult because the desired relative free-energy differ-
ences among the solvated intermediates depend importantly
on the sum of numerous weak intramolecular forces that
contribute to each folded state, on the partition functions
representative of these states, and on the effects of water and
co-solutes upon these interactions. Paulingꢀs hydrogen-bond-
ing motifs[1] and the canonical hydrophobic effect[2] have been
Scheme 1. Representative chemical structures for the folding models
joined by a new generation of weak intermolecular forces—
each of which has been proposed as a potential contributor to
protein folding and/or drug–receptor binding. Such forces
include a variety of aromatic interactions that may be divided
among neutral CH–p interactions,[3] ion–p interactions,[4] and
OH–p interactions[5] together with CH–O interactions,[6] the
venerable salt bridge,[7] and various halogen bonds.[8]
presented herein. The unfolded (left) and folded (right) conformations
for tert-butyl ester 10b.
The validation of computational methods for predicting
folding behavior requires accurate and precise experimental
data in well-defined contexts.[9,10] A decade ago, we intro-
duced a “molecular torsion balance” for measuring folding
energies to quantify the CH–p interaction and to examine the
effect of electron-withdrawing and electron-donating sub-
stituents on this force.[11] We concluded that the edge-to-face
aromatic interaction was driven principally by London
dispersion forces and that substituents had little effect on
the magnitude of the interaction.[12,13] The average folding
energy found in our model for edge-to-face aromatic inter-
actions in organic solvents was 0.3 kcalmolÀ1 and methyl aryl
p-face interactions led to slightly higher folding energies,
0.5 kcalmolÀ1. The balance we used incorporated a methyl
group counterpoised with an ester. This required that we
correct folding energies because they may have been affected
by dipole moment and solvation differences between ester
groups and methyl groups. In addition, our original balance
was not water soluble and it was therefore not possible to
measure the effects of water on folding energies.
Figure 1. a) Representation of solvated tert-butyl ester 10b in unfolded
(left) and folded (right) conformations. b) Simplified representation
illustrating the changes in hydration surrounding the tert-butyl (tBu)
and arene surfaces caused by folding.
torsion balances, as in our original studies, rotation about the
biphenyl bond is slow enough that individual signals for
folded and unfolded states are observable.[14,15] Here, two
esters are counterpoised and the rotation process exchanges
the position of only two alkyl groups. In the illustration, the
exchange is between a methyl group and a tert-butyl group.
No correction for dipole moment change is required. Fur-
thermore, we have incorporated a water-solubilizing group on
the axis of rotation, a location that minimizes any effects of
this group on the folding equilibria.
The measurement method we describe herein improves
on our earlier methods. The experiments evaluate equilibria
of the type illustrated in Scheme 1 and Figure 1. In these new
[*] B. Bhayana, Prof. Dr. C. S. Wilcox
Department of Chemistry
University of Pittsburgh
Pittsburgh, PA 15260 (USA)
Fax : (+1)412-624-1272
In the folding event, as the larger group moves from the
exo to the endo position, the solvent-accessible nonpolar
surface area of the molecule is reduced and water is expelled.
E-mail: daylite@pitt.edu
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 6833 –6836
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6833

Communications
At the same time, the tert-butyl group
establishes contact with the arene ring
(Figure 1). An X-ray diffraction study of a
similar isopropyl ester established the
methyl–arene distance (C to centroid) to
be 3.1 Š.[12]
The mutual-transfer free energy associ-
ated with this exchange of positions of the
methyl and tert-butyl group is defined
exactly by the equilibrium constant for the
exchange. This method of examining trans-
fer energies reduces ambiguities (for exam-
ple, uncertainties in activity-coefficient
assignments) that are associated with tradi-
tional phase-transfer free-energy studies[16]
and is more direct than attempts to measure
weak hydrophobic association effects
through intermolecular association stud-
ies.[17] The experimental results are also
immediately addressable by free-energy
perturbation calculation methods.[18]
The synthesis of the required water-
soluble torsion balances features the con-
trolled hydrolytic desymmetrization of die-
ster 1 to afford hemiester 2[19] (Scheme 2).
Nitration and esterification of the hemiester
provided the unsymmetrical diesters 3a–3e,
and nitration of 1 provided the symmetrical
ester 3 f. Pinacolatoboronate dibenzodiazo-
cine 7 was readily obtained by using our
unsymmetrical Trögerꢀs base synthesis
method[20] and was completed with a Pd-
catalyzed boronation.[21,22] The two portions
of the torsion balance were united through a
Suzuki[23] reaction to provide nitro diesters
8a–8 f. Reduction of the nitro group and
treatment of the resulting anilines 9a–9 f
with glutaric anhydride provided the final
carboxylic acids 10a–10 f. The acids were
soluble in D2O that contained Cs2CO3 or
K2CO3.[24]
Scheme 2. Synthesis pathway for the water-soluble torsion balances 10a–10 f. DCC=1,3-
dicyclohexylcarbodiimide, DMAP=4-dimethylaminopyridine, DMSO=dimethyl sulfoxide,
dppf=1,1’-bis(diphenylphosphanyl)ferrocene, HMTA=hexamethylenetetramine, TFA=tri-
fluoroacetic acid.
The free-energy changes associated with
folding in CDCl3, DGfold(CDCl3), for the nonpolar esters 8a–
8e and 8a–8e (Table 1) were in the range expected based on
our prior studies.[11,12] They were independent of the sub-
stituent (nitro, amino, or amide) at the position meta to the
esters (on the rotation axis), which is where we planned to
introduce a polar group to enhance water solubility. Folding
energies in water, DGfold(D2O), of 10a–10e were higher
compared with organic solvents but mirrored the trend
0.22 kcalmolÀ1 for the isopropyl ester to nearly 0.35 kcal
molÀ1 for the adamantyl ester torsion balance.
The influence of nonpolar surface exposure on folding in
water may be quantified by the excess solvent free-energy
parameter g, which is defined as the excess free energy per
square angstrom of the nonpolar–water interface.[25,26] The
magnitudes of g reported in prior work range from 7 to
200 calmolÀ1 Š2. To specify a value of g based on our data
requires calculation of the change in exposed surface area in
the unfolded and folded states. This change in the area of the
nonpolar-surface–water interface can be combined with our
measurement of the hydrophobic effect ((DGfold)D2OÀ
(DGfold)CDCl3) to evaluate the microscopic excess free
energy, g.
observed in organic solvents: DGcyclohexyl > DGisopropyl
DGtert-butyl
>
.
In the absence of experimental or theoretical evidence to
the contrary, we must expect the CH–p dispersion interaction
energy to be the same in chloroform and water,[11,12] and one
may logically define the difference between folding in non-
polar solvents and water (DGfold)D2OÀ(DGfold)CDCl3 to be a
measure of the hydrophobic contribution to folding in water.
This difference increases with the size of the alkyl group:
We calculated g by determining the solvent-accessible
surface areas (ASA) for folded and unfolded torsion balances.
Depending on the nature of the molecule, 100 to 50000
6834 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6833 –6836

Angewandte
Chemie
Table 1: Folding data for diesters 8, 9, and 10 at 298 K.[a]
including halogen bonds, cation–p interaction, and salt
bridges.
Entry
Ester
R
DGfold (CDCl3)
DGfold (D2O)
[kcalmolÀ1
]
[kcalmolÀ1
]
Received: March 1, 2007
Published online: August 3, 2007
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
8a
9a
10a
8b
9b
10b
8c
9c
10c
8d
9d
10d
8e
(CH3)2HC-
(CH3)2HC-
(CH3)2HC-
(CH3)3C-
(CH3)3C-
(CH3)3C-
cyclohexyl
cyclohexyl
cyclohexyl
1-adamantyl
1-adamantyl
1-adamantyl
2-adamantyl
2-adamantyl
2-adamantyl
H3C-
0.50
0.50
0.50[c]
0.65
0.65
0.65[c]
0.36
0.36
0.36[c]
0.36
0.36
0.36[c]
0.55
0.55
0.55[c]
0.0
i[b]
i[b]
Keywords: conformation analysis · Gibbs energy ·
0.72
.
i[b]
hydrophobic interactions · protein folding · solvent effects
i[b]
0.92
i[b]
i[b]
[1]a) L. Pauling, R. B. Corey, Nature 1951, 168, 550 – 551; b) Y.
Nozaki, C. Tanford, J. Biol. Chem. 1971, 246, 2211 – 2217;
c) M. D. Joesten, L. J. Schaad, Hydrogen Bonding, Marcel
Dekker, New York, 1974.
[2]a) W. Kauzmann, Adv. Protein. Chem. 1959, 14, 1 – 63; b) C. B.
Anfinsen, Science 1973, 181, 223 – 230; c) R. R. Gardner, L. A.
Christianson, S. H. Gellman, J. Am. Chem. Soc. 1997, 119, 5041 –
5042.
[3]a) S. K. Burley, G. A. Petsko, J. Am. Chem. Soc. 1986, 108, 7995 –
8001; b) K. Kobayashi, Y. Asakawa, Y. Kikuchi, H. Toi, Y.
Aoyama, J. Am. Chem. Soc. 1993, 115, 2648 – 2654; c) C. A.
Hunter, Philos. Trans. R. Soc. London Ser. A 1993, 345, 77 – 85;
d) C. T. Chen, J. S. Siegel, J. Am. Chem. Soc. 1994, 116, 5959;
e) M. Nishio, Y. Umezawa, M. Hirota, Y. Takeuchi, Tetrahedron
1995, 51 8665 – 8701; f) M. L. Waters, Curr. Opin. Chem. Biol.
2002, 6, 736 – 741; g) E. A. Meyer, R. K. Castellano, F. Die-
derich, Angew. Chem. Int. Ed. 2003, 42, 1210 – 1250; h) M.
Nishio, CrystEngComm 2004, 6, 130 – 158; i) S. Lavieri, J. A.
Zoltewicz, J. Org. Chem. 2001, 66, 7227 – 7230; j) W. B. Jennings,
B. M. Farrell, J. F. Malone, Acc. Chem. Res. 2001, 34, 885 – 894;
k) D. E. Anderson, J. H. Hurley, H. Nicholson, W. A. Baase,
B. W. Matthews, Protein Sci. 1993, 2, 1285 – 1290; l) R. O. Gould,
A. M. Gray, P. Taylor, M. D. Walkinshaw, J. Am. Chem. Soc.
1985, 107, 5921 – 5927; m) R. Ehama, M. Tsushima, T. Yuzuri, H.
Suezawa, K. Sakakibara, M. Hirota, Bull. Chem. Soc. Jpn. 1993,
66, 814 – 818.
0.67
i[b]
i[b]
0.68
i[b]
9e
10e
8 f
9 f
10 f
i[b]
0.9
i[b]
H3C-
H3C-
0.0
i[b]
0.0[c]
0.0
[a] Free-energy change upon folding calculated from the observed
equilibrium constant determined by integration and NMR line-shape
analysis. Samples were at 0.1 mm concentration. [b] Not soluble. [c] The
methyl ester of the free acid was used.
starting geometries were generated and geometry optimiza-
tions were carried out on each (MMFF and Eng–Huber force
fields). The ASA for conformations lying within 1 kcalmolÀ1
of the global minimum were Boltzmann averaged to provide
an average ASA for the ensemble of conformations contri-
buting to the folding states.
The values of the calculated g ranged from 5 to 30 calm-
olÀ1 Š2. These values of g lie in the low end of the range that
was expected based on prior work. The breadth of the
calculated values arises owing to uncertainties in the calcu-
lation of the surface areas for small molecules, which is where
we are attempting to see the finest details of the hydrophobic
effect. We look forward to creating molecular torsion
balances that evaluate larger nonpolar surface changes. It
will be especially interesting to evaluate future data in
comparison with predictions made by the Lum-Chandler-
Weeks theory of hydrophobicity. An intriguing aspect of this
important theory is that the value of g is expected to change
with the area of the nonpolar surface—water in contact with
small nonpolar surfaces is predicted to have a lower excess
energy (per square angstrom) than water in contact with more
extensive nonpolar surfaces, a prediction not opposed to our
results.[27]
This study demonstrates that even very small models of
proteins are influenced by the effects of water on folding. We
have described the synthesis and initial evaluation of a water
soluble molecular torsion balance that exhibits two-state
folding. We find that despite the small changes in solvent-
accessible area that accompany folding, the effect of water on
folding is clearly evident. The new torsion balance we present
herein can serve as a versatile tool for precise quantitative
studies of other important effects on folding and drug binding,
[4]a) D. A. Dougherty, D. A. Stauffer, Science 1990, 250, 1558 –
1560; b) M. M. S. L. Mowbray, J. Mol. Biol. 1994, 235, 709 –
717; c) K. S. Kim, J. Y. Lee, S. J. Lee, T.-K. Ha, D. H. Kim, J.
Am. Chem. Soc. 1994, 116, 7399 – 7400; d) S. M. Ngola, D. A.
Dougherty, J. Org. Chem. 1998, 63, 4566 – 4567; e) S. Bartoli, S.
Roelens, J. Am. Chem. Soc. 2002, 124, 8307 – 8315; f) E. Cubero,
M. Orozco, F. J. Luque, J. Phys. Chem. A 1999, 103, 315 – 321;
g) C. Felder, H.-L. Jiang, W.-L. Zhu, K.-X. Chen, I. Silman, S. A.
Botti, J. L. Sussman, J. Phys. Chem. A 2001, 105, 1326 – 1333.
[5]a) D. S. C. Reid, P. F. Lindley, J. M. Thornton, FEBS Lett. 1985,
190, 209 – 213; b) S. K. Burley, G. A. Petsko, FEBS Lett. 1986,
203, 139 – 143; c) K. Oku, H. Watanabe, M. Kubota, S. Fukuda,
M. Kurimoto, Y. Tsujisaka, M. Komori, Y. Inoue, M. Sakurai, J.
Am. Chem. Soc. 2003, 125, 12739 – 12748; d) M. Sulpizi, P.
Carloni, J. Phys. Chem. B 2000, 104, 10087 – 10091; e) N. S.
Scrutton, A. R. C. Raine, Biochem. J. 1996, 319, 1 – 8.
[6]a) G. Xiao, S. Liu, X. Ji, W. W. Johnson, J. Chen, J. F. Parsons,
W. J. Stevens, G. L. Gilliland, R. N. Armstrong, Biochemistry
1996, 35, 4753 – 4765; L. Jiang, L. Lai, J. Biol. Chem. 2002, 277,
37732 – 37740; b) S. Scheiner, T. Kar, Y. Gu, J. Biol. Chem. 2001,
276, 9832 – 9837.
[7]a) W. C. Wimley, K. Gawrisch, T. P. Creamer, S. H. White, Proc.
Natl. Acad. Sci. USA 1996, 93, 2985 – 2990; b) R. Luo, L. David,
H. Hung, J. Devaney, M. K. Gilson, J. Phys. Chem. B 1999, 103,
727 – 736; c) S. Kumar, R. Nussinov, J. Mol. Biol. 1999, 293,
1241 – 1255; d) S. E. Kiehna, M. L. Waters, Protein Sci. 2003, 12,
2657 – 2667; e) Q. Kaas, A. Aumelas, S. Kubo, N. Chino, Y.
Kobayashi, L. Chiche, Biochemistry 2002, 41, 11099 – 11108;
f) M. P. Aliste, J. L. MacCallum, D. P. Tieleman, Biochemistry
Angew. Chem. Int. Ed. 2007, 46, 6833 –6836
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6835

Communications
2003, 42, 8976 – 8987; g) B. Ibarra-Molero, J. A. Zitzewitz, C. R.
Matthews, J. Mol. Biol. 2004, 336, 989 – 996.
[15]H. J. Reich, J. Chem. Ed. Software 1996, 3D, 2.
[16]For discussions, see a) M. Auton, D. W. Bolen, Biochemistry
2004, 43, 1329 – 1342; b) D. R. Robinson, W. P. Jencks, J. Am.
Chem. Soc. 1965, 87, 2462 – 2470.
[17]C. A. Connors, Binding Constants, Wiley, New York, New York,
1987.
[18]a) P. A. Kollman, K. M. Merz, Jr., Acc. Chem. Res. 1990, 23,
246 – 252; b) M. Aschia, F. Mazzaa, A. Di Nolab, J. Mol. Struc.
(Theochem) 2002, 587, 177 – 188.
[19]M. H. Chen, J. G. Davidson, J. T. Freisler, J. Magano, Org. Prep.
Proced. Int. 2000, 32, 381 – 384.
[20]a) T. H. Webb, C. S. Wilcox, J. Org. Chem. 1990, 55, 363 – 366; for
a useful alternative approach to halogenated dibenzodiazocines,
see b) J. Jensen, J. Tejler, K. Wꢁrnmark, J. Org. Chem. 2002, 67,
6008 – 6014; c) K. Wꢁrnmark, J. Jensen, Synthesis 2001, 1873 –
1877.
[21]T. Ishiyama, M. Murata, N. Miyaura, J. Org. Chem. 1995, 60,
7508.
[22]We found that the method of Wꢁrnmark and co-workers is also
effective for preparing 6: J. Jensen, J. Tejler, K. Wꢁrnmark, J.
Org. Chem. 2002, 67, 6008 – 6014.
[8]a) V. Cody, P. Murray-Rust, J. Mol. Struct. 1984, 112, 189 – 199;
b) P. Kollman, A. Dearing, E. Kochanski, J. Phys. Chem. 1982,
86, 1607 – 1610; c) H. Loc Nguyen, P. N. Horton, M. B. Hurst-
house, A. C. Legon, D. W. Bruce, J. Am. Chem. Soc. 2004, 126,
16 – 17; d) J. A. Webb, J. E. Klijn, P. A. Hill, J. L. Bennett, N. S.
Goroff, J. Org. Chem. 2004, 69, 660 – 664; e) P. Auffinger, F. A.
Hays, E. Westhof, P. S. Ho, Proc. Natl. Acad. Sci. USA 2004, 101,
16789 – 16794; f) F. Hof, D. M. Scofield, W. B. Schweizer, F.
Diederich, Angew. Chem. 2004, 116, 5166 – 5169; Angew. Chem.
Int. Ed. 2004, 43, 5056 – 5059.
[9]J. Chen, W. Im, C. L. Brooks III, J. Am. Chem. Soc. 2006, 128,
3728 – 3736.
[10]Gellman and co-workers have reported several interesting
studies of minimal-protein-folding models: a) L. F. Newcomb,
T. S. Haque, S. H. Gellman, J. Am. Chem. Soc. 1995, 117, 6509 –
6519; b) R. R. Gardner, L. A. Christianson, S. H. Gellman, J.
Am. Chem. Soc. 1997, 119, 5041 – 5042; c) R. R. Gardner, S. L.
McKay, S. H. Gellman, Org. Lett. 2000, 2, 2335 – 2338; d) For a
valuable review of such models, see: W. B. Jennings, B. M.
Farrell, J. F. Malone, Acc. Chem. Res. 2001, 34, 885 – 894.
[11]S. Paliwal, S. Geib, C. S. Wilcox, J. Am. Chem. Soc. 1994, 116,
4497 – 4498.
[23]a) N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457; b) A.
Schnyder, A. F. Indolese, M. Studer, H. Blaser, Angew. Chem.
2002, 114, 3820 – 3823; Angew. Chem. Int. Ed. 2002, 41, 3668 –
3671.
[24]Dilution studies of the NMR spectra of the water-soluble torsion
balances revealed that the critical micelle concentrations for
these molecules lies well above 100 mm.
[12]E. Kim, S. Paliwal, C. S. Wilcox, J. Am. Chem. Soc. 1998, 120,
11192 – 11193.
[13]a) W. L. Jorgensen, D. L. Severence, J. Am. Chem. Soc. 1990,
112, 4768 – 4774; b) R. P. LꢀEsperance, D. Van Engen, R. Dayal,
R. A. Pascal, Jr., J. Org. Chem. 1991, 56, 688 – 694; c) R. L. Jaffe,
G. D. Smith, J. Chem. Phys. 1996, 105, 2780 – 2788; d) P. Hobza,
H. L. Selzle, E. W. Schlag, J. Am. Chem. Soc. 1994, 116, 3500 –
3506; e) C. Chipot, R. Jaffe, B. Maigret, D. A. Pearlman, P. A.
Kollman, J. Am. Chem. Soc. 1996, 118, 11217 – 11224; f) K.
Nakamura, K. N. Houk, Org. Lett. 1999, 1, 2049 – 2051.
[14]The barrier to rotation was determined by variable-temperature
NMR spectroscopy on diester 9 f. The WINDNMR lineshape
analysis program[15] was used to model the observed spectra.
Coalescence was observed at 408C in CDCl3.
[25]C. H. Tanford, The Hydrophobic Effect, Wiley, New York, New
York, 1980.
[26]a) R. R. Gardner, L. A. Christianson, S. H. Gellman, J. Am.
Chem. Soc. 1997, 119, 5041 – 5042; b) J. E. Mogensen, H. Ipsen, J.
Holm, D. E. Otzen, Biochemistry 2004, 43, 3357 – 3367.
[27]a) K. Lum, D. Chandler, J. D. Weeks, J. Phys. Chem. B 1999, 103,
4570 – 4580; b) D. M. Huang, D. Chandler, Proc. Natl. Acad. Sci.
USA 2000, 97, 8324 – 8327.
6836 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 6833 –6836