Intramolecular amide stacking and its competition with hydrogen bonding in a small foldamer

Article abstract of DOI:10.1021/ja9054965

(Chemical Equation Presented) Attractive interactions between two carboxamide groups in a "stacked" geometry are explored under isolated molecule conditions. Infrared spectra of single conformations of a small γ-peptide, Ac-γ²-hPhe-NHMe, reveal

Full text of DOI:10.1021/ja9054965

Published on Web 09/16/2009
Intramolecular Amide Stacking and Its Competition with Hydrogen Bonding in
a Small Foldamer
William H. James III,† Christian W. Mu¨ller,† Evan G. Buchanan,† Michael G. D. Nix,†,# Li Guo,‡
Luke Roskop,§ Mark S. Gordon,§ Lyudmila V. Slipchenko,† Samuel H. Gellman,‡ and
Timothy S. Zwier*,†
Department of Chemistry, Purdue UniVersity, West Lafayette, Indiana 47907, Department of Chemistry, UniVersity
of Wisconsin, Madison, Wisconsin 53706, and Department of Chemistry, Iowa State UniVersity, Ames, Iowa 50011
Received July 3, 2009; E-mail: Zwier@purdue.edu
Protein secondary structures, including ꢀ-sheets, R-helices, and
reverse turns, generally feature networks of NH...OdC hydrogen
bonds between backbone amide groups. Non-hydrogen bonded
attractive interactions can be envisaged as well, based on the large
dipole moment associated with the amide group (µamide ) 3.7 D)1
and the prospect of dispersion involving amide π-electrons. These
two factors would seem to favor stacking interactions in which the
amide groups are oriented in an antiparallel fashion. This mode of
amide-amide interaction, however, has received relatively little
attention,2-10 in part because it is more difficult to detect. Formation
of an H-bond is often clearly detectable by methods such as infrared
or nuclear magnetic resonance spectroscopy, and there are well-
established guidelines for inferring the existence of H-bonds in
crystal structures. In contrast, no general spectroscopic signatures
have yet been associated with non-H-bonding interactions between
amide groups. Here a combined experimental and computational
approach is used to identify and characterize an antiparallel
amide-amide stacking interaction that arises in a small foldamer.
We have taken advantage of the amide-amide spacing in the
γ-peptide backbone to obtain experimental evidence for face-to-
face amide stacking that is competitive with intramolecular H-
bonding. In Ac-γ2-hPhe-NHMe (Figure 1, inset), the three-carbon
spacing between the amide groups enables the two planar amide
groups to approach one another in a parallel (“stacked”) geometry
with the amide dipoles antialigned. Single-conformation infrared
and ultraviolet spectroscopy reveal signatures of this amide stacking
interaction, which is particularly evident in the strong coupling
between the antialigned CdO groups.
The experimental methods (SI) used to record single-conforma-
tion ultraviolet and infrared spectra of the isolated molecule cooled
in a supersonic expansion are similar to those recently employed
by several groups to study small polypeptides composed of R-amino
acid,11 ꢀ-amino acid,12,13 or a combination of R- and ꢀ-amino acid
units.14 We studied Ac-γ2-hPhe-NHMe as a prototypical γ-peptide
because it contains only two amide groups and bears an aromatic
ring in the side chain, which is necessary for IR-UV double
resonance studies. Resonant two-photon ionization (R2PI) was used
to record ultraviolet spectra in a mass-selective fashion, thereby
avoiding interference from thermal decomposition products. The
top trace of Figure 1 shows the R2PI spectrum of Ac-γ2-hPhe-
NHMe in the S0-S1 origin region of the phenylalanine moiety.
Two of the conformers have origin transitions that are similar in
intensity, A (37584 cm-1) and B (37484 cm-1), which dominate
the R2PI spectrum. The origin transition of the third conformer, C
(37471 cm-1), is significantly weaker in intensity than the other
two; however, despite the intensity difference, C plays an important
role in the conformational behavior of Ac-γ2-hPhe-NHMe.
Figure 1. Top: (a) R2PI spectrum and (b) scaled, single-point TDDFT
M05-2X/6-31+G(d) vertical excitation energies of Ac-γ2-hPhe-NHMe.
Bottom: RIDIR spectra in the amide I (left) and amide NH stretch (right)
spectral regions for the three conformers of Ac-γ2-hPhe-NHMe, with M05-
2X/6-31+G(d) scaled, harmonic vibrational frequencies, infrared intensities
(as sticks), and relative energies included for comparison (scale factors:
0.96 for amide I and 0.94 for amide NH stretch spectral regions). Table S1
contains the experimental and calculated frequencies and intensities for the
data shown in Figure 1.
Once identified, these transitions can be used as monitor
transitions to record resonant ion-dip infrared (RIDIR) spectra, an
IR-UV double resonance method by which single-conformation
infrared spectra can be obtained. The bottom panels of Figure 1
show the conformation-specific RIDIR spectra of conformers A-C
in the amide I and amide NH stretch spectral regions. Based on
the similarities of the infrared spectra of A and B, we conclude
that the conformations responsible for these spectra are quite similar,
containing a single, strong intramolecular amide-amide hydrogen
bond, as evidenced by the broad NH stretch transitions shifted down
in frequency to ∼3360 cm-1 in both spectra. The weak, sharp NH
stretch transitions at 3481 and 3476 cm-1 in A and B, respectively,
are due to the amide NH groups not involved in the H-bond.
† Purdue University.
# Present address: University of York, York, UK.
‡ University of WisconsinsMadison.
§ Iowa State University.
9
10.1021/ja9054965 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 14243–14245 14243

C O M M U N I C A T I O N S
Conformers A and B also show nearly identical spectra in the amide
I region, as expected based on the similar H-bonding in both
structures.
spacing. Extended conformers were shifted more than 300 cm-1
lower in frequency than the calculated origin of conformer B (Table
S1), which argues against assignment of conformer C to an extended
conformation.
Calculations predict amide stacked structures to have energies
nearly as low as those due to C9 H-bonded conformations, within
2 kJ/mol of one another, a result consistent with observing both
types of structures in the supersonic expansion. Extended and C7
H-bonded conformers were found to be less stable (>10 kJ/mol),
which supports the assignment of the three observed conformers
to C9 or amide stacked conformers.
Stick diagrams showing scaled harmonic vibrational frequencies
and infrared intensities of the NH stretch modes calculated at the
DFT M05-2X/6-31+G(d) level of theory15,16 are included in Figure
1 for comparison with experiment. The computed spectra faithfully
reproduce the experimental spectra of all three conformers, leading
to firm structural assignments. Conformers A and B are both
ascribed to a family of structures containing a nine-atom ring (C9)
that is closed by an NH...OdC H-bond (Figure 2, top panel). The
two structures differ in the position of the phenyl ring relative to
the peptide backbone, gauche (g-) for conformer B and anti (a) for
conformer A.
Figure 3. MRIRPT scans of conformers A-C of Ac-γ2-hPhe-NHMe in
the amide NH stretch region. Dips and gains in each spectrum arise from
loss or gain in population in the monitored conformation downstream caused
by IR excitation upstream. The total signal level of the UV probed conformer
in the absence of the IR pump pulse was set to a known output signal level,
while the difference in ion signal with or without the IR laser was recorded
using active baseline subtraction in a gated integrator.
Figure 2. M05-2X/6-31+G(d) optimized geometries of Ac-γ2-hPhe-NHMe.
Top: The two C9 H-bonded conformers ascribed to A and B (dashed lines
in red denote the intramolecular amide-amide hydrogen bond). Bottom:
The amide stacked structure ascribed to conformer C (left) and an alternate
view of the amide stacked interaction with the aromatic ring replaced by a
hydrogen atom for clarity.
A quantitative measurement of the relative abundances of amide
stacked and H-bonded conformations in Ac-γ2-hPhe-NHMe was
carried out using mass-resolved infrared population transfer spec-
troscopy (MRIRPT), an extension of previous population transfer
spectroscopies.23 MRIRPT spectra were obtained by counter-
propagating the infrared laser along the molecular beam, creating
a column of IR-excited molecules for interrogation. By adjusting
the timing of the IR excitation relative to the UV probe laser (∆t
) 44 µs), molecules initially excited within a few millimeters of
the pulsed valve orifice can isomerize and be recooled during travel
to the ion source region before interrogation with the UV probe
via R2PI. The ion signal due to one conformer is monitored via its
S0-S1 origin transition while the IR laser is tuned through the
spectral region of interest. IR transitions due to the probed
conformer appear as depletions in ion signal, while gains in ion
signal are observed when transitions of other conformers are
encountered that transfer population into the probed conformer zero-
point level. The set of MR-IRPT spectra is used to extract relative
abundances by taking a weighted sum of the spectra (Figure 3,
bottom) that zero-out the net change in population at all wavelengths
(∆Ntot(ν˜) ) 0). These weighting factors are the relative fractional
abundances of the conformers.23 The fractional abundances derived
from MRIRPT give FC ) 0.21 ( 0.01, within a factor of 2 of either
of the amide-amide hydrogen bonded conformers (FA ) 0.38 (
0.01 and FB ) 0.41 ( 0.02). This method avoids the use of S0-S1
origin intensities to infer relative fractional abundances (Figure 1,
top), which assume identical Franck-Condon factors, excited state
lifetimes, and R2PI cross sections for each conformer. The MRIRPT
The spectra of conformer C in the amide NH stretch and amide
I regions provide a striking contrast to those of A and B. Both NH
stretch fundamentals of C occur at frequencies characteristic of non-
H-bonded amide NH groups (3469 and 3480 cm-1). The gains in
the RIDIR spectrum of C (3357, 3372 cm-1) (Figure 1C) are the
result of an IR-created background that contributes to the R2PI
signal at the UV wavelength used to monitor conformer C, when
the strong H-bonded NH stretch fundamentals of conformers A and
B are excited. In the amide I spectral region, the spectrum displays
a single, dominant transition at 1716 cm-1, despite the presence of
two CdO groups in the molecule. Calculations clearly point to a
structure for conformer C in which the carbonyl groups are
antialigned and strongly coupled, producing in-phase and out-of-
phase oscillations that enhance the intensity in one fundamental at
the expense of the other, leading to the assignment of C to an amide
stacked structure (Figure 2, bottom). Fully extended conformers
possessed vibrational frequencies and IR intensities (Figure S1) that
were poorer matches to the experimental data than were those
calculated for the amide stacked conformation, particularly in that
the extended conformations displayed two CdO stretch fundamen-
tals of comparable intensity.
Single-point TDDFT M05-2X/6-31+G(d) vertical excitation
energies of the three conformers were used to compare experimental
and theoretical S0-S1 origin frequencies. The computed frequencies
for C9(a), C9(g-), and S(a), scaled to the experimental S0-S1 origin
of conformer B (Figure 1), are in agreement with the experimental
9
14244 J. AM. CHEM. SOC. VOL. 131, NO. 40, 2009

C O M M U N I C A T I O N S
abundances differ strikingly from those inferred from the intensities
in the R2PI spectrum (IA ) 0.55, IB ) 0.40, IC ) 0.05), pointing
to significant differences in the R2PI efficiencies of the conformations.
formation of the intramolecular H-bond in the C9 structures induces
some torsional strain in the segment linking the two amide groups.
In conclusion, the γ-peptide backbone juxtaposes adjacent amide
groups in a way that facilitates amide stacking, an interaction that
is energetically competitive with nearest-neighbor H-bonding.
Single-conformation IR spectra in the NH stretch and amide I
regions show clear evidence for a planar stacked conformation in
which the amide groups are antialigned. In the case of γ-peptide
foldamers, amide stacking can be used as a design element for
conformational control, a strategy that could lead to unique
secondary structures not yet anticipated.
To evaluate the various contributions to the stabilization present
in the amide stacked structure, the systematic fragmentation method
(SFM)17-19 was used to divide the molecular structure into
fragments whose nonbonded interactions can be evaluated and
characterized using the effective fragment potential (EFP) method.20,21
One of the strengths of the EFP method is that it allows one to
apportion the nonbonded interaction energy into physically mean-
ingful contributions, including Coulombic (electrostatics), induction
(polarization), exchange repulsion, dispersion, and charge transfer.
Acknowledgment. W.H.J., C.W.M., E.G.B., M.G.D.N., T.S.Z.
(NSF-CHE0909619), L.G., and S.H.G. (CHE-0848847) acknowl-
edge support from the NSF. L.V.S. acknowledges support from
Purdue University. L.R. and M.S.G. acknowledge support from the
Air Force Office of Scientific Research.
Comparison of these contributions for the structures assigned to
conformers A-C (Table 1) was carried out with the SFM internal
energies calculated at the M05-2X/6-31+G(d) level of theory. As
anticipated, dispersion plays a larger role in the amide stacked
structure (-18 kJ/mol) than in its C9 H-bonded counterparts (-10
kJ/mol), just as it does in other circumstances in which two π clouds
interact with one another.22 Indeed, although the electrostatic term
favors structures A and B by 5-7 kJ/mol, the dispersion interaction
favors C by a similar amount. The electrostatic contribution from
the SFM/EFP calculations is the single largest attractive contribution
to amide stacking, more than 80% of that in the C9 structures,
indicating that there is also a significant stabilization from the
antialigned polar amide groups.
Supporting Information Available: Complete ref 16. Experimental
and computational methods, conformational analysis details. This
material is available free of charge via the Internet at http://pubs.acs.org.
References
(1) Cantor, C. R. ; Schimmenl, P. R. Biophysical Chemistry: I. The Conforma-
tion of Biological Macromolecules; W. H. Freeman and Co.: San Francisco,
1980.
(2) Choudhary, A.; Gandla, D.; Krow, G. R.; Raines, R. T. J. Am. Chem. Soc.
2009, 131, 7244.
(3) Yonezawa, T.; Morishim., I. B. Chem. Soc. Jpn. 1966, 39, 2346.
(4) Rabinovitz, M.; Pines, A. J. Chem. Soc. B 1968, 1110.
(5) Betson, M. S.; Clayden, J.; Lam, H. K.; Helliwell, M. Angew. Chem., Int.
Ed. 2005, 44, 1241.
Table 1. Results of SFM/EFP Calculations for the Nonbonded
Interactions in the Three Conformers of Ac-γ2-hPhe-NHMe at the
M05-2X/6-31+G(d) Level of Theory
(6) Clayden, J. Chem. Soc. ReV. 2009, 38, 817.
(7) Maccallum, P. H.; Poet, R.; Milnerwhite, E. J. J. Mol. Biol. 1995, 248,
361.
(8) Maccallum, P. H.; Poet, R.; Milnerwhite, E. J. J. Mol. Biol. 1995, 248,
374.
experimental conformer
assigned structure
C
S(a)
B
C9(g-)
A
C9(a)
Nonbonded contributions
electrostatics
exchange repulsion
polarization
dispersion
charge transfer
Total nonbonded energy
Relative nonbonded energy
Total relative energya
Relative through bond energyb
(kJ/mol)
-31.58
36.11
-1.94
-18.06
-4.81
-20.28
12.26
(kJ/mol)
-38.29
27.32
-7.12
-10.82
-3.63
-32.54
0.00
(kJ/mol)
-36.63
25.24
-6.55
-10.61
-3.38
-31.93
0.61
(9) Hinderaker, M. P.; Raines, R. T. Protein Sci. 2003, 12, 1188.
(10) Fischer, F. R.; Wood, P. A.; Allen, F. H.; Diederich, F. P. Proc. Natl.
Acad. Sci. U.S.A. 2008, 105, 17290.
(11) Chin, W.; Piuzzi, F.; Dimicoli, I.; Mons, M. Phys. Chem. Chem. Phys.
2006, 8, 1033.
(12) Baquero, E. E.; James, W. H.; Choi, S. H.; Gellman, S. H.; Zwier, T. S.
J. Am. Chem. Soc. 2008, 130, 4784.
(13) Baquero, E. E.; James, W. H.; Choi, S. H.; Gellman, S. H.; Zwier, T. S.
J. Am. Chem. Soc. 2008, 130, 4795.
(14) James, W. H.; Baquero, E. E.; Shubert, V. A.; Choi, S. H.; Gellman, S. H.;
Zwier, T. S. J. Am. Chem. Soc. 2009, 131, 6574.
1.31
-10.95
0.00
0.00
0.97
-0.36
(15) Zhao, Y.; Truhlar, D. G. J. Chem. Theory Comput. 2007, 3, 289.
(16) Frisch, M. J., Gaussian 03, revision E.01; Gaussian, Inc.: Wallingford, CT,
2004.
a Fully optimized relative energies at the M05-2X/6-31+G(d) level of
theory. b Difference between total relative and relative nonbonded energies
ascribable to through bond contributions to the relative stabilities.
(17) Deev, V.; Collins, M. A. J. Chem. Phys. 2005, 122, 154102.
(18) Gordon, M. S.; Mullin, J. M.; Pruitt, S. R.; Roskop, L. B.; Slipchenko,
L. V.; Boatz, J. A. J. Phys. Chem. B 2009, 113, 9646.
(19) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon,
M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.;
Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993,
14, 1347.
(20) Gordon, M. S.; Freitag, M. A.; Bandyopadhyay, P.; Jensen, J. H.; Kairys,
V.; Stevens, W. J. J. Phys. Chem. A 2001, 105, 293.
(21) Gordon, M. S.; Slipchenko, L. V.; Li, H.; Jensen, J. H. Ann. Rep. Comput.
Chem. 2007, 3, 177.
The net nonbonded contribution to the stabilization of the stacked
structure (-20.3 kJ/mol) is ∼12 kJ/mol smaller than that in C9(g-)
(-32.5 kJ/mol), while the total relative energies differ by only 1.3
kJ/mol. Hence, there is a through-bond contribution that stabilizes
the amide stacked structure relative to C9 structures that is not
captured by the nonbonded EFP energy. This factor arises from
strain in the three-carbon bridge between amide groups; this strain
is minimized in the amide stacked conformation. In contrast,
(22) Slipchenko, L. V.; Gordon, M. S. J. Comput. Chem. 2007, 28, 276.
(23) Dian, B. C.; Longarte, A.; Winter, P. R.; Zwier, T. S. J. Chem. Phys. 2004,
120, 133.
JA9054965
9
J. AM. CHEM. SOC. VOL. 131, NO. 40, 2009 14245