Effects of lysine acetylation in a β-hairpin peptide: Comparison of an amide-π and a cation-π interaction

Article abstract of DOI:10.1021/ja0648460

The acetylation of lysine is a common posttranslational modification of histone proteins, and the interaction of acetylated lysines with aromatic rings is commonly observed in transcriptionally relevant protein-protein interactions. To determine the nature of this interaction and its potential role in protein structure and function, the effect of lysine acetylation on its interaction with tryptophan has been investigated within the context of a β-hairpin peptide. Acetylation of Lys results in the replacement of a cation-π interaction with an amide-π interaction. Despite the loss of positive charge, the interaction energy is not significantly perturbed, although the geometry of interaction is influenced such that the amide NH interacts directly with the Trp ring. Thermodynamic analysis indicates an enthalpic driving force for the stabilization, indicating a polar-π interaction. Acyl lysine analogues formyl lysine and trifluoroacetyl lysine were used to further investigate the sterics and electronics of the interaction.

Full text of DOI:10.1021/ja0648460

Published on Web 09/28/2006
Effects of Lysine Acetylation in a â-Hairpin Peptide:
Comparison of an Amide-π and a Cation-π Interaction
Robert M. Hughes and Marcey L. Waters*
Contribution from the Department of Chemistry, UniVersity of North Carolina,
Chapel Hill, North Carolina 27599-3290
Received July 7, 2006; E-mail: mlwaters@email.unc.edu
Abstract: The acetylation of lysine is a common posttranslational modification of histone proteins, and the
interaction of acetylated lysines with aromatic rings is commonly observed in transcriptionally relevant
protein-protein interactions. To determine the nature of this interaction and its potential role in protein
structure and function, the effect of lysine acetylation on its interaction with tryptophan has been investigated
within the context of a â-hairpin peptide. Acetylation of Lys results in the replacement of a cation-π
interaction with an amide-π interaction. Despite the loss of positive charge, the interaction energy is not
significantly perturbed, although the geometry of interaction is influenced such that the amide NH interacts
directly with the Trp ring. Thermodynamic analysis indicates an enthalpic driving force for the stabilization,
indicating a polar-π interaction. Acyl lysine analogues formyl lysine and trifluoroacetyl lysine were used to
further investigate the sterics and electronics of the interaction.
Introduction
The acetylation of lysine is a common posttranslational
modification of histone proteins. Lysine acetylation weakens
the histone-DNA complex in chromatin by neutralizing the
electrostatic interaction between unmodified lysine in histones
and the negatively charged phosphate backbone of DNA, thereby
activating transcription.1Additionally, lysine acetylation is com-
monly found in cells with DNA damage, where it is also thought
to weaken the DNA-histone complex to make DNA accessible
to repair enzymes,2 and can play a role in protein stability by
preventing lysine ubiquitination, which tags proteins for pro-
teolysis.3 The bromodomain, a protein domain with a specific
affinity for acetylated lysine, is found in most acyltransferases
and in many proteins associated with the regulation of chromatin
structure.4 The bromodomain binding pocket contains highly
conserved aromatic residues that appear to interact with acety-
lated lysine through amide-π interactions (Figure 1).5
Figure 1. (a) Lys-Trp pair in typical cation-π geometry (PDB code:
1ONR). (b) KAc-Tyr pair in typical amide-π geometry (Bromodomain
from GCN5 complexed with acetylated H4 peptide; PDB code: 1E6I).
the faces of aromatic rings.6 Additional investigations have
identified numerous backbone amide-π interactions in protein
and peptide structures,6d,7 as well as a number of amide-π
interactions that occur in the binding of ligands.8 For example,
backbone amide-π interactions have been identified between
Gly/Phe and Gly/Trp pairs in analyses of pairwise residue
The importance of amide-π interactions in structural biology
has been long established through data-mining studies of protein
crystal structures, in which amide-containing side chains (Gln,
Asn) were found to have statistical preference for packing near
(6) (a) Burley, S. K.; Petsko, G. A. FEBS Lett. 1986, 203, 139-143. (b) Burley,
S. K.; Petsko, G. A. AdV. Prot. Chem. 1988, 39, 125-189. (c) Singh, J.;
Thornton, J. M. J. Mol. Biol. 1990, 211, 595-615. (d) Steiner, T.; Koellner,
G. J. Mol. Biol. 2001, 305, 535-557.
(7) (a) van der Spoel, D.; van Buuren, A. R.; Tieleman, P.; Berendsen, H. J.
C. J. Biomol. NMR 1996, 8, 229-238. (b) Honda, S.; Kobayashi, N.;
Munekata, E. J. Mol. Biol. 2000, 295, 269-278. (c) Duan, G.; Smith, V.
H.; Weaver, D. F. Int. J. Quantum Chem. 2000, 80, 44-60. (d) Toth, G.;
Watts, C. R.; Murphy, R. F.; Lovas, S. Prot. Struct. Func. Gen. 2001, 43,
373-381. (e) Toth, G.; Murphy, R. F.; Lovas, S. Prot. Eng. 2001, 14,
543-547. (f) Toth, G.; Murphy, R. F.; Lovas, S. J. Am. Chem. Soc. 2001,
123, 11782-11790. (g) Toth, G.; Kover, K. E.; Murphy, R. F.; Lovas, S.
J. Phys. Chem. B 2004, 108, 9287-9296. (h) Meyer, E. A.; Castellano, R.
K.; Diederich, F. Angew. Chem., Int. Ed. 2003, 42, 1211-1250. (i)
Sengupta, A.; Mahalakshmi, R.; Shamala, N.; Balaram, P. J. Pept. Res.
2005, 65, 113-129. (j) Mahalakshmi, R.; Raghothama, S.; Balaram, P. J.
Am. Chem. Soc. 2006, 128, 1125-1138.
(1) Hong, L.; Schroth, G. P.; Matthews, H.; Yau, P.; Bradbury, E. M. J. Biol.
Chem. 1993, 268, 305-314.
(2) Masumoto, H.; Hawke, D.; Kobayashi, R.; Verreault, A. Nature 2005, 436,
294.
(3) Caron, C.; Boyault, C.; Khochbin, S. BioEssays 2005, 27, 408-415. (b)
Simonsson, M.; Heldin, C.; Ericsson, J.; Gronroos, E. J. Biol. Chem. 2005,
280, 21797-21803.
(4) Zeng, L.; Zhou, M. M. FEBS Lett. 2002, 513, 124-128.
(5) (a) Dhalluin, C.; Carlson, J. E.; Zeng, L.; He, C.; Aggarwal, A.; Zhou, M.
Nature 1999, 399, 491-496. (b) Owen, D. J.; Ornaghi, P.; Yang, J. C.;
Lowe, N.; Evans, P. R.; Ballario, P.; Neuhaus, D.; Filetici, P.; Travers, A.
A. EMBO J. 2000, 19, 6141-6149.
(8) Perutz, M. F. Philos. Trans. Phys. Sci. Eng. 1993, 345, 105-112.
9
13586
J. AM. CHEM. SOC. 2006, 128, 13586-13591
10.1021/ja0648460 CCC: $33.50 © 2006 American Chemical Society

Acylation Effects on a Lys-Trp Interaction
A R T I C L E S
interactions in water within a biologically relevant context and
has previously been used to study cation-π interactions.15 Two
analogues of acetyl lysine (KAc), formyl lysine (KFm) and
trifluoroacetyl lysine (KFAc), were studied to explore the effects
of changes in hydrophobic and electronic character of the acyl
group. In our investigation, we find that the interaction between
acetylated lysine and tryptophan consists of a polar-π interac-
tion that occurs primarily between the polarized amide of KAc
and the electron rich face of the Trp indole ring. Despite the
loss of the positive charge upon N-acetylation, the KAc‚‚‚Trp
interaction is as strong as the unmodified Lys‚‚‚Trp interaction
previously reported in the same model system.15a,b This indicates
that the polar-π interaction is energetically competitive with
the cation-π interaction. This is likely due to a lower desol-
vation penalty for the interaction of acetyl lysine and tryptophan,
while maintaining the electrostatic and van der Waals compo-
nents of the interaction via NH(δ+)-π(δ-) and π-π stacking.
Results and Discussion
Design. The 12-residue â-hairpin sequence used in this study
is based on a sequence that has been previously described
(Figure 2).15 Key design features include a good turn-nucleating
sequence (Val-Asn-Gly-Orn) and a number of hydrophobic
interactions to stabilize the hairpin, including Val3, Val5, Ile10
on one face of the hairpin, and Trp2 and Leu11 on the opposite
face of the hairpin. The diagonal interaction between Trp2 and
position 9 (residue X in Figure 2) is the site of interest for this
study. A Lys residue at position 9 has been previously been
shown to interact favorably with Trp2 via a cation-π interaction
between the ꢀ-CH2 of Lys and the face of the aromatic ring. To
explore the relative favorability of an amide-π interaction, we
investigated the effect of Lys acetylation.
Figure 2. (a) â-Hairpin peptide structure. (b) Amino acids at position X:
lysine (Lys), acetyl lysine (KAc), formyl lysine (KFm), trifluoroacetylated
lysine (KFAc), and norleucine (Nle).
preferences in antiparallel â-sheets,9 and amide-π interactions
have been identified in the binding of the drug bezafibrate by
human deoxyhemoglobin.10 Amide-π interactions have also
been utilized in the selective binding of guests by model systems
in organic solution.11 Additional theoretical studies have been
conducted to demonstrate both the orientation and magnitude
of the amide-π interaction, using variously the formamide-
benzene interaction12 and the ammonia-benzene interaction13
in the gas phase as models for the amide-π interaction in
proteins. Notably, a recent study investigated both amide- and
cation-π interactions using implicit solvent models and pre-
dicted the amide-π interaction of Asn and Gln side chains with
an adenine ring (Ade) to be intermediate in magnitude (-4 kcal/
mol) between the cation-π interactions Arg-Ade (-7 kcal/mol)
and Lys-Ade (-2 kcal/mol).14 In addition, Burley and Pestko’s
survey of 33 protein crystal structures found higher fractions
of polar amide residues interacting with aromatic side chains
(31% Asn, 40% Gln) than for Lys (26%), but lower than that
of Arg (47%).6a These findings suggest that the amide-π
interaction may be more favorable than the cation-π interaction
between lysine and an aromatic ring (Figure 1).
Characterization. The â-hairpin structure and stability was
characterized by a number of standard NMR techniques,
including R-hydrogen (HR) chemical shifts, glycine splitting
(∆δ Gly), and cross-strand NOEs.15,16 The degree of HR
downfield shifting and Gly splitting relative to random coil
values is used as an indicator of the degree of â-sheet structure
at each position along the strand and in the turn, respectively.17
The fraction folded is calculated with the following equation:
%
folded
) (δobs - δU)/(δF - δU), where the unfolded state (U)
is represented by random coil control peptides and the fully
folded state (F) by cyclic peptides (see Supporting Information).
Furthermore, numerous NOEs between cross-strand pairs of side
chains were observed for all peptides, consistent with â-hairpin
formation (see Supporting Information).
To explore the efficacy of amide-π interactions, and in
particular their role in the recognition of KAc, we have
investigated the effects of acetylation of lysine and its interaction
with tryptophan within the context of a â-hairpin peptide (Figure
2). This is a useful model system to study noncovalent
Effects of Lysine Acetylation. Acetylation of lysine results
in a modest enhancement in stability of the â-hairpin peptide.
This is demonstrated by an increase in the HR shifts relative to
random coil, as shown in Figure 3a. Based on the glycine
splitting, fraction folded of WKAc is 87% ((1%) (side chain
HRavg ) 83 ( 8%),18 versus 78% ((1%) for the peptide WK
at 298 K (side chain HRavg ) 76 ( 7%),18 indicating that
(9) (a) Hutchinson, E. G.; Sessions, R. B.; Thornton, J. M.; Woolfson, D. N.
Prot. Sci. 1998, 7, 2287-2300. (b) Merkel, J. S.; Regan, L. Folding &
Design 1998, 3, 449-455.
(10) Perutz, M. F.; Fermi, G.; Abraham, D. J.; Poyart, C.; Bursaux, E. J. Am.
Chem. Soc. 1986, 108, 1064-1078.
(11) (a) Bisson, A. P.; Lynch, V. M.; Monahan, M. C.; Anslyn, E. V. Angew.
Chem. Int. Ed. 1997, 36, 2340-2342. (b) Snowden, T. S.; Bisson, A. P.;
Anslyn, E. V. J. Am. Chem. Soc. 1999, 121, 6324-6325.
(12) (a) Duan, G.; Smith, V. H.; Weaver, D. F. J. Phys. Chem. A 2000, 104,
4521-4532. (b) Duan, G.; Smith, V. H.; Weaver, D. F. Chem. Phys. Lett.
1999, 310, 323-332.
(15) (a) Tatko, C. D.; Waters, M. L. Prot. Sci. 2003, 12, 2443-2452. (b) Tatko,
C. D.; Waters, M. L. J. Am. Chem. Soc. 2004, 126, 2028-2034. (c) Hughes,
R. M.; Waters, M. L. J. Am. Chem. Soc. 2005, 127, 6518-6519.
(16) Maynard, A. J.; Sharman, G. J.; Searle, M. S. J. Am. Chem. Soc. 1998,
120, 1996-2007.
(13) Tsuzuki, S.; Honda, K.; Uchimaru, T.; Mmikami, M.; Tanabe, K. J. Am.
Chem. Soc. 2000, 122, 11450-11458.
(17) Griffith-Jones, S. R.; Maynard, A. J.; Searle, M. S. J. Mol. Biol. 1999,
292, 1051-1069.
(14) Biot, C.; Buisine, E.; Rooman, M. J. Am. Chem. Soc. 2003, 125, 13988-
13994.
(18) Side chain HRavg values were calculated excluding the terminal residues
(Arg, Gln) and the turn residues (Asn, Gly).
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13587

A R T I C L E S
Hughes and Waters
Figure 4. Proposed geometries of interaction for (a) Trp‚‚‚Lys, (b) Trp‚
‚‚KAc, (c) Trp‚‚‚KFm, and (d) Trp‚‚‚KFAc based on the NMR chemical
shifts in Figures 3 and 6.
Figure 5. Thermal denaturation profiles of WK, WKAc, and WNle
peptides as determined by NMR. The fraction folded was determined from
the Gly splitting. Error is (0.5 K in temperature and (1% in fraction folded.
Conditions: 50 mM NaOAc-d3 buffer, pD 4.0 (uncorrected).
NH in KAc versus the ꢀ-CH2 in Lys likely arises from
differences in the desolvation cost for an amide relative to an
ammonium group. Indeed, the calculated solvation energy for
the Lys side chain is -66.2 kcal/mol, versus -8.1 kcal/mol for
KAc.19 Moreover, if the amide in KAc is stacked with the Trp
residue, the NH group may still be able to hydrogen bond with
water, such that there is minimal desolvation penalty, as has
been proposed for the interaction of Arg with Trp.20 In proteins,
the stacked conformation for Asn and Gln side chains with
aromatic residues is favored by a factor of about 2.5.6d Given
that the Trp‚‚‚KAc interaction is highly solvent exposed in this
peptide model system, a stacked conformation is likely.
Figure 3. (a) WK and WKAc RH shifts relative to random coil values.
Glycine shifts reflect the splitting. (b) WK and WKAc side chain upfield
proton shifts relative to random coil values. Conditions: 298K, 50 mM
NaOAc-d3 in D2O, pH 4.0 (uncorrected), referenced to DSS.
acetylation enhances hairpin stability (∆∆G ) 0.38 kcal/mol,
( 0.05 kcal/mol). Because acetylation of Lys may change its
â-sheet propensity, we performed a double mutant cycle to
quantitatively assess the interaction energy between Trp and
KAc. Val and Ser were used as controls for Trp and Lys,
respectively.15 Double mutant cycles give the magnitude of the
Trp‚‚‚X side chain-side chain interaction for WK and WKAc
to be -0.30 and -0.34 kcal/ mol ((0.1 kcal/mol), respectively.
While the differences in the two interaction energies are within
experimental error, this is significant in that acetylation does
not decrease the magnitude of the interaction with Trp, despite
the loss of the positive charge. Rather, the conversion of the
ammonium group to an amide provides an interaction with Trp
of comparable magnitude.
Although the interaction energies for Trp‚‚‚KAc and Trp‚‚‚
Lys are similar, the upfield shifting patterns of the Lys and KAc
side chains indicate a change in the interaction geometry with
Trp for KAc relative to Lys (Figure 3b). While the unmodified
lysine is most upfield shifted at the ꢀ-CH2 protons, KAc is most
upfield shifted at its amide proton. The ꢀ-CH2 of KAc exhibits
some upfield shifting as well, wheareas the acetyl methyl group
is less shifted. Thus, although it is difficult to determine if the
amide group is stacked with the Trp or interacting via an NH‚
‚‚π interaction, it is clear that it interacts in a different geometry
than Lys (Figure 4). The preference for interaction at the amide
The thermodynamics of hairpin folding for WKAc are
consistent with hairpin stabilization due to a polar-π interaction
(Figure 5 and Table 1).21 The driving force for folding of WKAc
is similar to WK, with an enthalpic driving force for the folding
and an associated entropic cost, despite the difference in charge.
In contrast, comparison with a peptide containing a nonpolar
alkyl side chain, WNle, at position X shows that the driving
force folding of WKAc is much more enthalpically favorable
than the nonpolar Nle side chain.15b This is consistent with
significant interaction between the polarized amide NH and
(19) Solvation energies for the Lys and KAc side chains were calculated with
Gaussian 03 (implicit water/IEFPCM model/HF:6-31g**). Gaussian 03,
Frisch, M. J., et al., Gaussian, Inc., Wallingford CT, 2004. The full reference
can be found in the Supporting Information.
(20) (a) Mitchell, J. B. O.; Nandi, C. L.; McDonald, I. K.; Thornton, J. M. J.
Mol. Biol. 1994, 239, 315-331. (b) Gallivan, J. P.; Dougherty, D. A. Proc.
Natl. Acad. Sci. U.S.A. 1999, 96, 9459-9464.
(21) Hughes, R. M.; Waters, M. L. J. Am. Chem. Soc., in press.
9
13588 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Acylation Effects on a Lys-Trp Interaction
A R T I C L E S
Table 1. Thermodynamic Parametersa for Folding at 298 K10
peptide
∆H
°
(kcal/ mol)
∆S
°
(cal/ mol K)
∆Cp
°
(cal/ mol K)
reference
WK
WKAc
WNle
-2.8 (0.03)
-2.91 (0.07)
-0.85 (0.07)
-6.8 (0.1)
-6.1 (0.2)
-0.7 (0.2)
-163 (3)
-250 (40)
-170 (25)
15a,b
this work
15b
a Determined from the temperature dependence of the Gly chemical shift
from 0 to 60 °C. Errors (in parentheses) are determined from the fit. Error
is (0.5 K in temperature and (1% in fraction folded. Conditions: 50 mM
NaOAc-d3 buffer, pD 4.0 (uncorrected).
tryptophan in WKAc as seen in Figure 3b, and argues against
a hydrophobic driving force alone.22
Effects of Acetyl Lysine Analogues. Acetyl lysine analogues,
formyl lysine (KFm) and trifluoroacetyl lysine KFAc), were
used to further examine the effects of acylation. KFm was
investigated in order to determine the contribution of hydro-
phobic or van der Waals interactions, and KFAc was studied
to examine electronic effects on the interaction with Trp
interaction. With these analogues, subtle changes were detected
in the geometry of the interaction, but there was little effect on
stability and driving force for folding. WKAc and its analogues
WKFm and WKFAc are approximately equistable at 298K:
87, 86, and 87% folded (( 1%), respectively, based on the
glycine splitting (83, 82, and 83% ( 8% folded based on the
average side chain HR fraction folded).18 This is supported by
the consistency of the HR shifts relative to random coil (Figure
6a). While these hairpins are of similar stabilities, comparison
of the side chain upfield shifting profiles for KAc, FAc, and
Fm indicates differences in the mode of interaction with Trp
(Figures 4, 6b). The KFm side chain has a similar profile to
the KAc side chain. Both are most upfield shifted at the amide
NH position. However, the KFm has both polarized amide and
aldehyde protons (Figure 7), both of which can interact favorably
with the tryptophan ring. As a result, the upfield shifting is
distributed more evenly over the two sites, while the KAc side
chain has a definite preference for interacting with tryptophan
through the amide position (Figure 4). The interaction of KFm
with Trp is in agreement with predictions by Duan et al. from
gas-phase calculations that a formamide-π interaction is stable
over a range of geometries.12b In contrast, the KFAc side chain
exhibits significant upfield shifting at the δ- and ꢀ-CH2 groups
and very little shifting of the amide nitrogen, suggesting a shift
in the preferred geometry (Figures 4, 6b). Indeed, it has an
interaction profile more similar to lysine (Figure 6c).
The differences in interaction geometries can be elucidated
through a comparison of electrostatic potential maps of the three
side chains (KAc, KFm, and KFAc) as shown in Figure 7.
Clearly the amide proton is the most polarized site on each of
the three side chain analogues, with the carbonyl carbon on the
KFAc analogue also highly polarized due to the inductive effects
of the fluorines. However, the fluorines in the KFAc analogue
may present both a steric and electronic barrier to interaction
of the amide NH with the indole ring. The CF3 group in KFAc
is the largest of the groups on the carbonyl, thus obscuring the
amide proton. Additionally, CF‚‚‚π interactions have been
shown to have weak repulsive character in fluoromethane-
benzene model systems, lending some repulsive electronic
Figure 6. (a) WKAc, WKFm, and WKFAc RH shifts relative to random
coil values. Glycine shifts reflect the splitting. (b) KAc, KFm, and KFAc
side chain upfield proton shifts relative to random coil values. (c) Lys and
KFAc side chain upfield proton shifts relative to random coil. Conditions:
298 K, 50 mM NaOAc-d3 buffer, pH 4.0 (uncorrected), referenced to DSS.
character to the KFAc‚‚‚Trp interaction.23 Due to long-range
inductive effects of the fluorines, the δ and ꢀ positions of the
KFAc side chain are more polarized than the corresponding
positions in the KAc and KFm side chains (Figure 7). This, in
combination with the electronic and steric repulsion of the
trifluoro group, would explain the upfield shifting pattern seen
in Figure 5c, where the most intense shifts are seen at the δ
and ꢀ positions of the KFAc side chain. This suggests that the
interaction energy for the Trp‚‚‚KFAc interaction can only be
(22) The difference in ∆Cp for WK and WKAc may also reflect differences in
desolvation energy, although no clear trend is observed when a comparison
to WNle is also made. Since a large, negative ∆Cp for folding is taken as
evidence of a hydrophobic driving force for folding, the inconsistency
between WKAc and WNle may have to do with differences in chain length.
(23) Kawahara, S.; Tsuzuki, S.; Uchimaru, T. J. Phys. Chem. A. 2004, 108,
6744.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13589

A R T I C L E S
Hughes and Waters
chemical shifts indicate that the amide NH is in close proximity
to the face of the aromatic ring, rather than between the ꢀ-CH2
and Trp that is observed with the unmodified Lys. The change
in geometry is likely due to differences in desolvation costs for
the Lys and KAc side chains. The thermodynamics for folding
indicate that the amide-π interaction contains significant
polar-π character. Thus, both amide-π and cation-π interac-
tions consist of a sum of a number of components, including
charge-quadrupole (for cation-π), dipole-quadrupole, and
quadrupole-quadrupole interactions (for amide-π), as well as
dispersion forces and hydrophobic effects.24 The primary
difference between the two types of interactions is that conver-
sion of the ammonium group to an amide changes the desol-
vation cost of the nitrogen group, such that the NH of the amide
can favorably interact directly with the aromatic ring, which is
too costly for the ammonium group. Thus, the terms “cation-
π” and “amide-π” are useful as general terms that emphasize
the importance of quadrupole interactions rather than simple
hydrophobic effects alone, but by no means do these terms
indicate that quadrupole interactions are the only contributor
to these noncovalent interactions.
Figure 7. Electrostatic potential maps of amino acid side chains: (a) KAc,
(b) KFm, (c) KFAc. Electrostatic potential maps generated with MacSpartan
(HF/6-31g*; Isodensity value ) 0.02; range ) -30 to 30 kcal/mol).
Acetylated lysine analogues KFm and KFAc have further
elucidated our understanding of this interaction through ma-
nipulation of both electronic and sterics of the Lys side chain.
These changes in sterics and electronics result in subtle changes
in the geometery of interaction which are required to maintain
the interaction energy, indicating the fine-tuning that is possible
in such noncovalent interactions. This has significance for
understanding specificity in protein folding as well as in
protein-ligand interactions and may provide useful insight for
medicinal chemists. In summary, these studies provide insight
into the specific recognition of acetylated lysine that plays a
significant biological role in chromatin condensation via the
“histone code” as well as the general role of amide-π
interactions in protein structure and function.
Figure 8. Thermal denaturation profiles of WKAc, WKFm, and WKFAc
peptides as determined by NMR. The fraction folded was determined from
the Gly splitting. Error is (0.5 K in temperature and (1% in fraction folded.
Conditions: 50 mM NaOAc-d3 buffer, pD 4.0 (uncorrected).
Table 2. Thermodynamic Parametersa for Folding at 298 K
peptide
∆H°
(kcal/ mol)
∆S
°
(cal/ mol K)
∆Cp° (cal/ mol K)
WKAc
WKFm
WKFAc
-2.91 (0.07)
-3.06 (0.03)
-2.68 (0.07)
-6.1 (0.2)
-6.7 (0.1)
-5.2 (0.2)
-250 (40)
-195 (30)
-220 (30)
a Determined from the temperature dependence of the Gly chemical shift
from 0 to 60 °C. Errors (in parentheses) are determined from the fit.
Experimental Section
maintained with a change in the geometry of interaction relative
to Trp‚‚‚KAc. This sort of subtle tuning of the interaction
geometry has relevance to specificity in protein folding and
protein-ligand interactions.
The thermal denaturation for each of the three peptides
WKAc, WKFm, and WKFAc were found to be close to or
within error (Figure 8), resulting in very similar values for ∆H°,
∆S°, and ∆Cp° (Table 2). Thus, although changing the sterics
and electronics of the amide group influences the geometry of
the interaction, there is very little effect on its driving force.
This suggests a level of control of the interaction geometry that
is not often observed with noncovalent interactions. Moreover,
it provides insight into the subtle factors that may influence
the folding of a protein into a single, low energy structure.
Synthesis of Fmoc-Trifluoroacetyl Lysine. Fmoc-lysine (0.384 g;
0.97 mmol) was dissolved in THF (10 mL) and cooled in an ice bath
to 0 °C. DIPEA (168 µL, 0.97 mmol) was added dropwise to the cold
solution. Trifluoroacetic anhydride (162 µL, 1.2 mmol) was added in
one portion, and the resulting solution was allowed to warm to room
temperature over 20 min. The reaction was followed by TLC (10:1
CH2Cl2:MeOH). The resulting solution was rotovapped to dryness, taken
up into ethyl acetate, and washed with water and brine. The organic
layer was dried with sodium sulfate, and the solvent was removed in
vacuo giving a clear oil. Further purification with silica gel chroma-
tography (10:1 CH2Cl2:MeOH) gave 0.405 g of a clear oil (0.87 mmol;
90% yield). ESI-MS: calculated ) 464.156; actual ) 464.2 NMR (600
MHz, CDCl3): 7.83 (d, 2H); 7.60 (d, 2H); 7.42 (dd, 2H); 7.33 (dd,
2H); 6.59 (s, 1H); 5.415 (d, 1H); 4.44 (m, 2H); 4.23 (m, 1H), 3.38 (m,
2H); 1.95-1.47 (m, 6H).
Synthesis of Fmoc-Formyl Lysine. Formic acid (534 µL, 14 mmol)
was added dropwise to acetic anhydride (1.086 mL, 11.4 mmol) at 0
°C, followed by gentle heating (60 °C, 2 h).25 The resulting mixture
was cooled to 0 °C in an ice bath. Fmoc-Lys (0.384 g; 1.0 mmol) was
taken up into 5 mL of THF and added dropwise to the mixed anhydride
solution. The resulting solution was allowed to stir overnight at room
temperature. After removal of the solvent in Vacuo, the solution was
taken up into ethyl acetate (25 mL) and washed with water (25 mL)
Conclusion
We have utilized a â-hairpin model system to investigate the
nature of an amide-π interaction between Trp and an acetylated
Lys residue and compared it to the cation-π interaction of Trp
with the unmodified Lys. Our system demonstrates both the
geometric and energetic effects of acetylation upon the Trp‚‚‚Lys
interaction. Surprisingly, while the geometry of the interaction
changes upon acetylation, the magnitude of the interaction does
not decrease, despite the loss of a positive charge. NMR
(24) Ma, J. C.; Dougherty, D. A. Chem. ReV. 1997, 97, 1303-1324.
(25) Krishnamurthy, S. Tetrahedron Lett. 1982, 23, 3315-3318.
9
13590 J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006

Acylation Effects on a Lys-Trp Interaction
A R T I C L E S
and brine (25 mL). The organic layer was dried with sodium sulfate
and the solvent removed in vacuo, resulting in a clear oil (0.390 g;
98% yield; 0.98 mmol). ESI-MS: calculated ) 396.169; actual ) 396.2
NMR (600 MHz, CDCl3): 8.08 (s, 1H), 7.74 (d, 1H), 7.58 (d, 2H),
7.38 (dd, 2H), 7.28 (dd, 2H), 6.07 (s,1H); 5.71 (s, 1H); 4.38 (m, 2H);
4.19 (m, 1H), 3.26 (m, 2H); 1.25-1.9 (m, 6H).
scans using a 1-3 s presaturation. All 2D NMR experiments used pulse
sequences from the Chempack software including TOCSY, DQCOSY,
gCOSY, and NOESY. 2D NMR scans were taken with 16-64 scans
in the first dimension and 64-256 scans in the 2nd dimension. All
spectra were analyzed using standard window functions (sinebell and
Gaussian). Mixing times of 0.5 s were used in the NOESY spectra.
Assignments were made using standard methods as described by
Wu¨thrich.26 Thermal denaturations were run in duplicate, with standard
deviations of less than 0.005 ppm. Samples were allowed to equilibrate
for 7 min at each temperature before the measurement was taken.
Temperature calibrations were made using methanol and ethylene glycol
standards.
Peptide Synthesis. Fmoc-Lys(Ac) was purchased from Bachem.
Fmoc-trifluoroacetyl lysine (Fmoc-KFAc) and Fmoc-formyl lysine
(Fmoc-KFm) were synthesized as described above. The synthesis of
all peptides was performed on an Applied Biosystems Pioneer peptide
synthesizer using Applied Biosystems PEG-PAL amide resin. Peptides
were synthesized on a 0.1 or 0.07 mmol scale. All amino acids with
functionality were protected during synthesis as follows: Arg(Pbf), Asn-
(trt), Lys(Boc), Orn(Boc), Gln(trt), Trp(Boc), and Glu(tBu). Coupling
reagents were HBTU/HOBt. Noncommercially available amino acids
were coupled by hand. The N-terminus was acetylated for all peptides
with a solution of 5% acetic anhydride and 6% 2,6-lutidine in DMF.
Cleavage conditions removed all side chain protection with a cocktail
of 90% TFA/5% Triisopropylsilane/5% H2O. Peptides were purified
by RP-HPLC on a C18 column. Peptides were purified with a gradient
of A and B (A: 95% H2O/5% ACN with 0.1% TFA, B: 95% ACN/
5% H2O with 0.1% TFA). Once purified, peptides were lyophilized to
powder and characterized by ESI-TOF mass spectroscopy and NMR.
NMR Spectroscopy. NMR samples were made in concentrations
of approximately 1 mM and analyzed on a Varian Inova 600 MHz
spectrometer. Samples were dissolved in D2O buffered to pD 4.0
(uncorrected) with 50 mM NaOAc-d3 and referenced to DSS. Amine
and amide resonances were assigned in 60% H2O solutions. 1D NMR
spectra were collected using 32K data points and between 8 and 64
Acknowledgment. M.L.W. gratefully acknowledges an Al-
fred P. Sloan Foundation Fellowship. R.M.H. gratefully ac-
knowledges support from a Burroughs Wellcome foundation
fellowship and an ACS Division of Organic Chemistry Fellow-
ship sponsored by Albany Molecular Research. This work was
funded by an NSF Career Award to M.L.W. and a grant from
the NIH NIGMS (GM071589).
Supporting Information Available: Control peptides, chemi-
cal shift assignments, NOEs, and thermal denaturation data are
available in Supporting Information. This material is available
free of charge via the Internet at http://pubs.acs.org.
JA0648460
(26) Wu¨thrich, K. NMR of Proteins and Nucleic Acids; John Wiley and Sons:
New York, 1986.
9
J. AM. CHEM. SOC. VOL. 128, NO. 41, 2006 13591