Synthesis and β-sheet propensity of constrained N-amino peptides

Article abstract of DOI:10.1016/j.bmc.2017.08.017

The stabilization of β-sheet secondary structure through peptide backbone modification represents an attractive approach to protein mimicry. Here, we present strategies toward stable β-hairpin folds based on peptide strand N-amination. Novel pyrazolidinone and tetrahydropyridazinone dipeptide constraints were introduced via on-resin Mitsunobu cyclization between α-hydrazino acid residues and a serine or homoserine side chain. Acyclic and cyclic N-amino peptide building blocks were then evaluated for their effect on β-hairpin stability in water using a GB1-derived model system. Our results demonstrate the strong β-sheet stabilizing effect of the peptide N-amino substituent, and provide useful insights into the impact of covalent dipeptide constraint on β-sheet folding.

Full text of DOI:10.1016/j.bmc.2017.08.017

Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and b-sheet propensity of constrained N-amino peptides
⇑
Matthew P. Sarnowski, Kyle P. Pedretty, Nicole Giddings, H. Lee Woodcock, Juan R. Del Valle
Department of Chemistry, University of South Florida, Tampa, FL 33620, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The stabilization of b-sheet secondary structure through peptide backbone modification represents an
attractive approach to protein mimicry. Here, we present strategies toward stable b-hairpin folds based
on peptide strand N-amination. Novel pyrazolidinone and tetrahydropyridazinone dipeptide constraints
Received 29 June 2017
Revised 4 August 2017
Accepted 8 August 2017
Available online xxxx
were introduced via on-resin Mitsunobu cyclization between a-hydrazino acid residues and a serine or
homoserine side chain. Acyclic and cyclic N-amino peptide building blocks were then evaluated for their
effect on b-hairpin stability in water using a GB1-derived model system. Our results demonstrate the
strong b-sheet stabilizing effect of the peptide N-amino substituent, and provide useful insights into
the impact of covalent dipeptide constraint on b-sheet folding.
Keywords:
Peptidomimetics
Secondary structure
Amino acids
Ó 2017 Elsevier Ltd. All rights reserved.
Peptide conformation
b-Strands
1. Introduction
Our interest in b-strand/sheet stabilization led us to consider
conformationally restricted peptidomimetic building blocks that
After
a
helices, b-sheets represent the most common secondary
could be readily incorporated into host structures. Ideally, these
subunits would enforce extended dihedral angles, nucleate folding,
and inhibit aggregation. Furthermore, we envisioned that an
approach amenable to synthesis on solid support would enable
the rapid conformational scanning of lead peptides.
We recently developed a class of peptidomimetics featuring a
backbone N-amino substituent (Fig. 1).6 This substituent provides
a nucleophilic handle for subsequent covalent constraint and
may engage in an intraresidue C6 H-bond. In addition to N-amino
peptides (NAPs), our group reported a ‘trans-locked’ (R)-tetrahy-
dropyridazinedione (tpd) variant that results from covalent tether-
ing of the N-amino substituent onto an adjacent aspartic acid side
chain.7 The current work introduces two novel peptidomimetics
based on (R)-tetrahydropyridazinone (tpy) and (R)-pyrazolidinone
(pyz) scaffolds (Fig. 1). These orthotic constraints are readily intro-
duced on solid support via Mitsunobu reactions between an N-
structure in proteins. A number of protein-protein and protein-
DNA interactions implicated in pathogenesis involve b-sheets or
strands at the macromolecular interface.1 The b-sheet-like
oligomerization and fibrilization of amyloid monomers, for exam-
ple, is known to contribute to the progression of Alzheimer’s and
Parkinson’s disease.2 In cancer, Ras pathway hyperactivation relies,
in part, on edge-edge interactions between the b-sheet domains of
Ras and various downstream kinases.3 Modulation of these and
similar interactions with b-strand/sheet mimics has thus attracted
considerable interest. Strategies aimed at stabilizing b-sheet con-
formations within shorter peptides have also played an important
role in elucidating the anatomy and dynamics of biomolecular
associations.4
Residues within parallel b-sheets are defined by / and
w dihe-
dral angles close to À119° and 113°, respectively, whereas those in
antiparallel b-sheets are at or near À139° and 135°.1 In conjunction
aminated residue and an adjacent
chain. Engaging -amino acid side chains in covalent tethering
restricts the dihedral angles to values close to those in b-sheets
D-homoserine or D-serine side
with trans amide (
x
) bonds, these torsions result in a sawtooth
D
amino acid arrangement featuring two H-bonding peptide ‘edges’
and two ‘faces’ harboring side chain functional groups. Enforce-
ment of these backbone conformations has typically been achieved
through judicious incorporation of sheet-forming side chains,
covalent tethering approaches, turn-templated hairpin design, or
the use of conformationally extended backbone isosteres.1,5
w
(+120°). Here, we also evaluate the b-sheet propensity of N-amino,
tpd, tpy, and pyz constraints side-by-side in a biologically relevant
model of b-hairpin folding using NMR.
2. Solution-phase synthesis
Model pyz and tpy constrained dipeptides were first synthe-
⇑
sized in solution starting from
tively (Scheme 1). Following Alloc protection, the carboxylic acids
D-serine and D-homoserine, respec-
Corresponding author.
E-mail address: delvalle@usf.edu (J.R. Del Valle).
http://dx.doi.org/10.1016/j.bmc.2017.08.017
0968-0896/Ó 2017 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Sarnowski M.P., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.08.017

2
M.P. Sarnowski et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
C6
H
H-bond
H
N
N
H
O
H
R
S
S
H
O
GB1
Ac-c[CGEWAYNPATGKFAVTEC]-NH2
N-amino peptide
cyclic (fully folded) control
(NAP)
O
O
O
H
5
OMe
H
NH
N
O
NH2
N
O
R
H
H
H
lactamization
N
H
N
H
H
H
Ac-TGKFAVTE-NH2
O
R
H-GEWAYNPATGKFAVTE-NH2
native (partially folded) control
tetrahydropyridazinedione
random coil (unfolded) control
D-aspartic acid
6
(tpd) constraint
4
HO
diagnostic Gly10 H
Boc
NH
N
O
mutation site
diastereotopic separation
R
HN
O
Mitsunobu
Mitsunobu
H
N
N
H
N
O
O
Thr
O
O
Ala
Lys
N
H
H
N
H
N
H
N
CONH2
Glu
H
O
H
H
N
H
N
H
N
H
Thr
Ala
O
R
H
H
tetrahydropyridazinone
(tpy) constraint
O
O
Val
Trp
O
Phe
Tyr
NH
D-homoserine
O
H
N
O
H
N
O
H
N
Gly
NH2
O
Boc
HO
O
NH
N
N
N
R
HN
O
N
HN
H
H
H
N
N
Asn
O
Ala
O
Glu
O
N
H
O
4
H
O
H
H
O
R
pyrazolidinone
(pyz) constraint
Fig. 2. Model b-hairpins for thermodynamic analysis. The green circle indicates the
diagnostic Gly10H
protons from which diastereotopic 1H NMR separation values
D-serine
a
are derived. The tripeptide region targeted for mutation is highlighted in blue.
Fig. 1. Constrained N-amino peptides.
been utilized to evaluate the impact of various mutations on b-
sheet stability.10 Recently, Horne and co-workers introduced tar-
geted mutations into the GB141-56 model peptide in order to
enhance its folded population and facilitate chemical shift assign-
ments.10d Our studies on the b-sheet propensity of constrained
NAP derivatives make use of this sequence in conjunction with
its cyclic (fully folded) and truncated (random coil) variants, as
control peptides (4–6).
1) Alloc-Cl, aq. NaHCO3,
dioxane
2) BTC, collidine,THF,
TBSO
OTBS
Boc
n HN
N
n
N2Boc-aPhe-OMe
31-37%
(2 steps)
CO2Me
H2N
CO2H
AllocHN
Boc
1a n = 1
O
2a,b
Bn
1b n = 2
N
1) TBAF, THF
n
Previous work with hairpins related to 4 has demonstrated
N
CO2Me
2) DIAD, PPh3, THF
d.r. > 20:1
AllocHN
chemical shift separation of the diastereotopic Gly H
a protons
38-47%
(2 steps)
O
3a,b
Bn
within the turn to be diagnostic of folded population. This diaster-
eotopic separation arises from the distinct chemical shift environ-
ment of each proton within a stable folded hairpin.11 Conversely,
destabilization of secondary structure leads to conformational
averaging and coalescence of these signals. To evaluate our con-
straints, we targeted residues 12–14 as candidate sites for muta-
tional analysis. This region was selected on the basis of its
central position within the northern b-strand, and for ease of syn-
thesis. Moreover, incorporation of a tpd, tpy, or pyz motif results in
loss of the pro-(S) side chain, which may negatively impact folded
population. We thus considered mutations at Ala13 to be the least
disruptive given the low b-sheet propensity of alanine as measured
in a full-length GB1 model system.12
Scheme 1. Synthesis of constrained dipeptides in solution.
were converted to acid chlorides using BTC and condensed with H-
(N2-Boc)-aPhe-OMe. Silyl ether deprotection and intramolecular
Mitsunobu cyclization yielded cyclic peptidomimetics 3a and 3b
in good yield. The diastereopurity of intermediates 2a and 2b were
determined to be >20:1 by comparison of their 1H NMR spectra
with those of diastereomeric (D-aPhe) analogs. The crude Mit-
sunobu reaction mixture of 3a did not reveal any detectable elim-
ination to dehydroalanine or epimerization at the configurationally
Hydrazino acid derivatives 8a and 8b were prepared by oxazir-
idine-mediated electrophilic amination13 of phenylalanine and
valine as shown in Scheme 2. We previously reported a 2-step pro-
labile
D-Ser a-carbon. These results thus established feasibility,
prompting us to adapt our covalent tethering approach to solid-
phase peptide synthesis.
tocol to access these building blocks via
a-amino ester amina-
tion.14 The current method involves direct amination of the
unprotected zwitterions in water, and gives pure products follow-
ing trituration with Et2O/hexanes. Assembly of peptides on solid
support employed standard Fmoc chemistry, with the exception
of coupling to N2-Boc-protected hydrazino amides. In these cases,
acid chloride activation was required. The configurationally stable
Fmoc-Asp(Me)-Cl building block for tpd formation was pre-formed
3. Design and synthesis of b-sheet model peptides
In order to experimentally assess the b-sheet propensity of our
constrained motifs, we employed a 16-residue b-hairpin derived
from the protein G immunoglobulin-binding domain B1 (GB1,
Fig. 2).8 GB141–56, a soluble monomeric antiparallel b-sheet model
peptide featuring a 4-residue turn, is estimated to be ꢀ40% folded
in aqueous solution at 5 °C.9 This peptide and its close analogs have
using SOCl2, as described previously.7 As in Scheme 1, Fmoc-(
D)
Please cite this article in press as: Sarnowski M.P., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.08.017

M.P. Sarnowski et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
3
O
H
N
O
O
Thr
O
O
Lys
NH
N
O
H
N
H
N
H2N
MBHA
resin
pyz
R1
EtO2C
EtO2C
NaHCO3, THF, H2O
NBoc
N
H
CONH2
Glu
R2
N
H
N
H
N
R1
n
Thr
Ala
H
HATU, DIEA, 50oC
O
Phe
Tyr
HN
CO2H
O
Val
Trp
O
NH
H2N
CO2H
NHBoc
8a R = CH2Ph
8b R = i-Pr
7
O
O
O
Gly
NH2
O
H
N
H
N
H
N
R1
O
N
N
N
O
R1
O
H
N
HN
H
H
H
H
N
Asn
O
Ala
O
Glu
O
N
HN
HN
N
H
N
N
H
O
9
O
R2
R1
R3 HN
O
Boc
R2
n
n
Boc
H
O
O
Thr
O
O
Lys
NH
N
TFA,
H2O
N
H
N
H
N
tpy
CONH2
Glu
R3 = Me or (CH2)4NHBoc
R3 = CH2CO2Me
O
O
H
N
H
H
N
H
N
H
Thr
Ala
N
N
O
Val
Trp
O
N
N
H
H
Phe
Tyr
O
NH
R3 = CH2OTBS or (CH2)2OTBS
R3 NH2
O
R2
n
O
O
O
Gly
NH2
1) TBAF/THF
2) DIAD, PPh3
3) TFA/H2O
R3 = Me or (CH2)4NH2
O
H
N
H
N
H
N
-Xaa-aXaa-Xaa-
N
N
N
HN
H
H
Asn
O
Ala
O
Glu
O
TFA,
H2O
R1
O
O
O
O
R1
O
O
10
H
N
H
N
H
H
N
O
N
N
NH
N
H
H
N
NH
N H
H
N
O
H
O
Thr
O
O
Lys
NH
N
R2
O
R2
O
n
H
H
tpd
n
1,2
N
CONH2
Glu
N
-pyz-Xaa-Xaa-
-tpy-Xaa-Xaa-
N
H
N
N
-tpd-Xaa-Xaa-
Thr
Ala
H
H
O
Val
Trp
O
O
Phe
Tyr
NH
Scheme 2. Solid-phase synthesis of constrained NAPs.
O
O
O
Gly
NH2
O
H
N
H
N
H
N
hSer-OH15and Fmoc-( )Ser-OH16 were activated in situ with BTC/-
D
N
N
N
HN
H
H
Asn
O
Ala
O
Glu
O
collidine. Formation of the tpd ring occurred readily during cleav-
age from the resin with 95:5 TFA:H2O. Mitsunobu cyclizations to
produce tpy and pyz constraints were performed on resin, prior
to global deprotection/cleavage. As the N-terminal Gly Fmoc group
was not stable to DIAD/PPh3 (due to base-mediated deprotection),
these peptides were terminated with a Boc-protected Gly residue
prior to Mitsunobu reactions.
O
11
strong
NOE:
medium weak
Fig. 3. Interstrand NOE correlations observed for 9–11.
Table 1 depicts each of the peptides synthesized for the current
study. In addition to the native (unmutated) peptide 4, we pre-
pared its macrocyclic variant 5 (disulfide bridge) as a fully folded
positive control. Truncated linear analog 6 was synthesized to pro-
Examination of the NOESY spectra for each of the constrained NAPs
revealed several key correlations indicative of folded structures.
Fig. 3 depicts interstrand contacts observed for 9–11. Notably,
pyz mutant 9 exhibited few interstrand correlations relative to
vide Gly H
a chemical shift separation values within a relevant ran-
tpy and tpd mutants 10 and 11. The
residues, but not the pyz residue, engages in a close interaction
with the Tyr5H . Correlations between Trp3-Val14, Gly1-Val14,
a proton of the tpd and tpy
dom coil. Covalently constrained derivatives 9–11 display the
orthotic bridge on the outer edge of the hairpin, so as not to inter-
fere with interstrand H-bonding interactions. Compounds 12 and
13 are acyclic NAPs substituted at each of the outer edge amides
within the region of interest. Finally, di-NAP mutant 14 was pre-
pared to assess the impact of multiple backbone aminations,
whereas N-methylated analog 15 allows for a comparison of
NAP-based constraints with N-alkylation.
a
and Tyr5-Phe12 were also consistent in the spectra of 10 and 11.
These data qualitatively suggested enhanced stability of the tpy-
and tpd-contaning b-hairpins relative to 9.
In order to obtain a quantitative assessment of folding, we cal-
culated the Gly10
Ha chemical shift separation for each peptide
from TOCSY spectra (0.010 ppm uncertainty was assumed and
error propagated for subsequent calculations). In our hands, the
fully folded control peptide (5) and the random coil peptide (6)
exhibited dDGly10 values of 0.290 and 0.047 ppm, respectively. This
dynamic range was used to calculate a folded fraction of 73% for
the native (unmutated) peptide 4, which is slightly higher than
that previously reported (Table 2).10d The folded fraction of pyz
4. Structural and thermodynamic analysis
Detailed NMR studies of 9–15 were carried out at 1 mM concen-
tration in D2O (50 mM NaHPO4, pH 6.8). 1H NMR assignments were
made on the basis of TOCSY and NOESY spectra obtained at 4 °C.
Table 1
Synthesized peptides.
Peptide
Sequence
HRMS m/zobs
Yield %
4
5
6
9
10
11
12
13
14
15
H-GEWAYNPATGKFAVTE-NH2
Ac-c[CGEWAYNPATGKFAVTEC]-NH2
Ac-TGKFAVTEC-NH2
H-GEWAYNPATGKF-pyz-VTE-NH2
H-GEWAYNPATGKF-tpy-VTE-NH2
H-GEWAYNPATGKF-tpd-VTE-NH2
H-GEWAYNPATGK-aPhe-AVTE-NH2
H-GEWAYNPATGKFA-aVal-TE-NH2
H-GEWAYNPATGK-aPhe-A-aVal-TE-NH2
H-GEWAYNPATGK-(N-Me)Phe-AVTE-NH2
870.4214 [M+2H]2+
993.9292 [M+2H]2+
893.4715 [M+H]+
25
5
20
4
4
3
9
15
8
876.9194 [M+2H]2+
883.9269 [M+2H]2+
890.9160 [M+2H]2+
877.9276 [M+2H]2+
877.9280 [M+2H]2+
885.4366 [M+2H]2+
877.4299 [M+2H]2+
21
Please cite this article in press as: Sarnowski M.P., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.08.017

4
M.P. Sarnowski et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Table 2
Folding thermodynamics of constrained NAPs.
Peptide
Mutation
D
dGly10 (ppm)
% Folded
D
Gfold (kcal/mol)
DDGfold (kcal/mol)
4
9
10
11
12
13
14
15
native
pyz13
tpy13
tpd13
aPhe12
aVal14
aPhe12/aVal14
(N-Me)Phe12
(N-Me)Phe12trans
(N-Me)Phe12 cis
0.224
0.127
0.171
0.174
0.261
0.252
0.301
73
33
51
52
88
84
>99
24
37
5
5
5
5
6
5
6
5
5
À0.5 0.1
+0.4 0.1
+0.0 0.1
+0.0 0.1
À1.1 0.3
À0.9 0.2
N/A
+0.9
+0.5
+0.5
À0.6
À0.4
N/A
+0.6 0.2
+0.3 0.1
+1.1
+0.8
15trans
15cis
0.096
0.000
mutant 9, harboring a 5-membered ring constraint, was found to
be only 33%, corresponding to destabilization of +0.9 kcal/mol
relative to 4. The tpy and tpd 6-membered ring constraints in 10
and 11 were better accommodated into the b-strand region, as
these peptides were ꢀ50% folded in water. N-Amino substitution
without covalent tethering (12 and 13) resulted in a notable
increase in b-hairpin population. The most dramatic stabilization
of secondary structure was observed with the aPhe12/aVal14
double mutant (14). In this case, amination of two outer edge
amides on the same b-strand yielded a fully folded acyclic b-
hairpin. Conversely, backbone N-methylation in 15 resulted in
+1.1 kcal/mol destabilization relative to
4 and +1.7 kcal/mol
destabilization relative its N-amino counterpart 12.
Fig. 5. Ramachandran plots for ui+1, wi+1 in 16 and 18.
5. Computational analysis
An additional critical factor appears to be the ability for the N-
amino group to act either as a H-bond donor or acceptor. Examin-
In order to gain additional insights into the conformational
preferences of NAP-based constraints, we carried out computa-
tional analysis on the model dipeptides shown in Fig. 4. Two-
dimensional potential energy surfaces were computed as a func-
tion of backbone dihedral angles using CHARMM.17 Subsequent
constrained geometry optimizations were carried out with Q-
ing structures with /i+1 = À105°,
w
i+1 = À60° (i.e., the transition
state wi+1 saddle point region in 16) we found that cyclic NAPs
18 and 19 adopt stable geometries involving an H-bond between
the C-terminal amide NH (donor) and the endocyclic N-amino
nitrogen (acceptor). In contrast, the N-amino group in 16 acts only
as a donor, avoiding stabilization of unproductive conformations
(Fig. 6). NBO analysis on these structures reveals that the NÁ Á ÁHAN
interaction leads to stabilization gains in pyz, and tpy of 1.5, and
4.3 kcal/mol, respectively. While this cannot fully explain the
experimental trends, it clearly illustrates that subtle stereoelec-
tronic effects are playing a role in the conformational preferences
of both cyclic and acyclic NAPs.
Chem 4.4 at the
x
B97X-D/6-31+G⁄⁄ level of theory.18 NBO calcula-
tions were performed for each global minimum and additional points
of interest to evaluate (de)stabilizing intramolecular interactions.19
Ramachandran plots for /i+1, wi+1 in 16 and 18 are shown in
Fig. 5, with additional plots for cyclic NAPs (18 and 19) presented
in the Supplementary Material. As noted above, idealized antipar-
allel b-sheets exhibit /,
w
values in the À139°, 135° range. Exam-
ining the /i+1 i+1 energy surfaces for 16–19 it becomes clear that
,
w
Lastly, we considered that the destabilization observed in hair-
pins 9–11 relative to 4 may be due to unfavorable wi torsions
16 is more likely to adopt sheet-like conformations than 17–19.
Intuition would suggest this results from destabilizing clashes or
ring strain that may occur in the cyclic NAPs, however this is not
the case. Instead, the energy landscape of the wi+1 torsion in 17–
19 is more accessible, allowing cyclic NAP derivatives to more
easily adopt non-ideal conformations.
within the heterocycles. Computational analysis of the
wi dihedral
angles in 17–19 across a range of /i torsions showed a significant
energetic cost associated with antiparallel b-sheet conformations.
The wi angle in the tpd and tpy rings of 17 and 18 are generally
O
i+1
i
NH2
N
O
NH
N
O
AcHN
NHMe
AcHN
NHMe
O
i
O
i+1
Ac-tpd-Ala-NHMe
Ac-Ala-aAla-NHMe
17
16
O
NH
N
O
NH
N
AcHN
NHMe
AcHN
NHMe
O
O
Ac-tpy-Ala-NHMe
Ac-pyz-Ala-NHMe
18
19
Fig. 6. Energy minimized structures of dipeptide models at /i+1 = À105°,
showing stabilizing NÁ Á ÁHAN interactions in 18 and 19.
wi+1 = À60°
Fig. 4. Model dipeptides employed for computational analysis.
Please cite this article in press as: Sarnowski M.P., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.08.017

M.P. Sarnowski et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
5
constrained to values well above 135°. Moreover, the lowest
energy /i conformations in the cyclic NAP derivatives are in the
on-resin Mitsunobu cyclizations. The pyz and tpy subunits are
derived from readily accessible building blocks and their synthesis
on solid support should facilitate their use in various pep-
tidomimetic applications. Our assessment of the b-sheet propen-
sity of 4 distinct NAP derivatives revealed that the 6-membered
tpy and tpd constraints are better accommodated into a b-hairpin
strand than the pyz ring. Though all three cyclic constraints
destabilize the b-hairpin relative to the parent peptide, each pro-
moted folding relative to backbone N-methylation. In the current
model system, mono- and di-N-amination significantly enhanced
b-hairpin stability. These results will inform the design of
+150° to 180° range. A
wi, /i dihedral energy scan of the more flex-
ible acyclic NAP (16) revealed its ability to more readily access Ala
torsions favorable to b-hairpin folding (see Supplementary Mate-
rial for details).
6. Discussion
The current work once again demonstrates the enhanced b-
sheet propensity of hydrazino acid residues using a distinct b-hair-
pin model system. In our previous examination of folded NAPs, we
observed secondary structure stabilization with aTyr, aLys, and
aLeu substitutions within a 12-residue b-hairpin.6 Enhanced stabil-
ity from two other hydrazino acid residues in this study (aPhe,
aVal) suggests that N-amination may be a general strategy for pro-
moting b-sheet folds. Polyamination along a single edge of an
extended peptide appears to be a particularly powerful approach
for b-strand stabilization. Peptide 14 represents the second exam-
ple where strand di-N-amination yields a b-hairpin that is isoener-
getic, in water, with its fully folded macrocyclic control. This
remarkable effect is observed in spite of the presence of two ter-
tiary amide bonds within the peptide strand. We previously attrib-
additional a-hydrazino acid building blocks capable of nucleating
b-sheet-like folds.
Acknowledgments
We gratefully acknowledge the National Science Foundation for
funding (NSF-CHE1709927). We thank the USF Interdisciplinary
NMR Facility and USF Mass Spectrometry Facility for assistance.
Computer time was provided by USF Research Computing, spon-
sored in part by NSF MRI CHE-1531590.
uted this result to
a combination of cis amide rotamer
A. Supplementary data
destabilization, intraresidue C6 H-bonding, and enhanced torsional
strain resulting from the introduction of the N-amino group.6
The data in Table 2 shed some light on the relative importance
of non-covalent interactions for b-hairpin folding within this ser-
ies. Although the minor cis amide rotamer of N-methylated peptide
15 is effectively unstructured, the trans rotamer (65% of the rota-
meric mixture) still exhibits poor folding relative to 4. This is in
agreement with previous data reported for 15 and suggests that
other factors are primarily responsible for the strong b-hairpin
destabilizing effect of N-methylation.10f The trans-locked motifs
(pyz, tpy, tpd) in 9–11 were better accommodated into the b-
hairpin than the N-methyl substituent. However, we did not
observe any cis amide rotamers in the NMR spectra of acyclic
NAPs 12–14, suggesting that covalent tethering does not provide
a significant entropic benefit with respect to amide isomerism.
Acyclic NAPs are known to engage in C6 intraresidue H-bonds
involving the hydrazino acid NH2 donor and the carbonyl O accep-
tor (see Fig. 1).6 Such interactions in pyz- and tpy-constrained pep-
tides would involve an analogous H-bond donor, while the
enhanced acidity of the diacylhydrazide NH in tpd-constrained
peptides results in superior donating capacity. Computational
analysis suggests that both tpy and pyz constraints force the N-
amino lone pair to orient toward the peptide C-terminus, increas-
ing its potential to act instead as an H-bond acceptor. This is in
stark contrast with the acyclic NAP, in which the N-amino lone pair
remains syn-periplanar to the acylhydrazide NAC bond. While the
N-amino group in the tpd motif also acts exclusively as a H-bond
donor (due to N lone pair conjugation), increased ring strain rela-
tive to the acyclic NAP may hinder its ability to adopt b-sheet-like
conformations.
Supplementary data associated (experimental methods, com-
pound characterization, and supplementary figures) with this arti-
cle can be found, in the online version, at http://dx.doi.org/10.
1016/j.bmc.2017.08.017.
References
1. Loughlin WA, Tyndall JDA, Glenn MP, Hill TA, Fairlie DP. Chem Rev. 2010;110:2.
2. (a) Selkoe DJ. Nature. 2003;426(Suppl 1):900–904;
(b) Ross CA, Poirier MA. Nat Med. 2004;10:S10;
(c) Knowles TP, Vendruscolo M, Dobson CM. Nat Rev Mol Cell Biol. 2014;15:384.
3. (a) Nassar N, Horn G, Herrmann C, Block C, Janknecht R, Wittinghofer A. Nat
Struct Biol. 1996;3:723;
(b) Downward J. Nat Rev Cancer. 2003;3:11.
4. (a) Remaut H, Waksman G. Trends Biochem Sci. 2006;31:436;
(b) Cheng P-N, Pham JD, Nowick JS. J Am Chem Soc. 2013;135:5477;
(c) Watkins AM, Arora PS. ACS Chem Biol. 2014;9:1747.
5. (a) Khakshoor O, Nowick JS. Curr Opin Chem Biol. 2008;12:722;
(b) Del Valle JR. In: Lubell W, ed. Peptidomimetics II. Cham: Springer
International Publishing; 2017:25–49.
6. Sarnowski MP, Kang CW, Elbatrawi YM, Wojtas L, Del Valle JR. Angew Chem Int
Ed. 2017;56:2083.
7. Kang CW, Ranatunga S, Sarnowski MP, Del Valle JR. Org Lett. 2014;16:5434.
8. Gronenborn AM, Filpula DR, Essig NZ, et al. Science. 1991;253:657.
9. Blanco FJ, Rivas G, Serrano L. Nat Struc Biol. 1994;1:584.
10. (a) Fesinmeyer RM, Hudson FM, Andersen NH. J Am Chem Soc. 2004;126:7238;
(b) Huyghues-Despointes BMP, Qu X, Tsai J, Scholtz JM. Proteins. 2006;1005:63;
(c) Wei Y, Huyghues-Despointes BMP, Tsai J, Scholtz JM. Proteins. 2007;69:285;
(d) Lengyel GA, Horne WS. J Am Chem Soc. 2012;134:15906;
(e) Lengyel GA, Eddinger GA, Horne WS. Org Lett. 2013;15:944;
(f) Lengyel GA, Reinert ZE, Griffith BD, Horne WS. Org Biomol Chem.
2014;12:5375;
(g) Karnes MA, Schettler SL, Werner HM, Kurz AF, Horne WS, Lengyel GA. Org
Lett. 2016;18:3902.
11. Griffiths-Jones SR, Maynard AJ, Searle MS. J Mol Biol. 1999;292:1051.
12. Minor DL, Kim PS. Nature. 1994;367:660.
13. Armstrong A, Jones LH, Knight JD, Kelsey RD. Org Lett. 2005;7:713.
14. Kang CW, Sarnowski MP, Elbatrawi YM, Del Valle JR. J Org Chem. 2017;82:1833.
15. Dekan Z, Vetter I, Daly NL, Craik DJ, Lewis RJ, Alewood PF. J Am Chem Soc.
2011;133:15866.
It is also worth noting that, in contrast to acylic NAPs 12–14,
each of the covalently tethered motifs in 9–11 lacks the native
Ala13 side chain. One ongoing area of investigation is the confor-
mational impact of quaternary cyclic subunits that restore pro-
(S) Ca substituents. Although these side chains may be important
for b-sheet folding, such designs would still be forced to overcome
16. Fischer PM. Tetrahedron Lett. 1992;33:7605.
ˇ
17. (a) Miller BT, Singh RP, Klauda JB, Hodošcek M, Brooks BR, Woodcock HLJ. Chem
Inf Mod. 1920;2008:48;
(b) Brooks BR, Brooks CL, Mackerell AD, et al. J Comp Chem. 2009;30:1545.
18. (a) Chai J-D, Head-Gordon M. Phys Chem Chem Phys. 2008;10:6615;
(b) Chai J-D, Head-Gordon M. J Chem Phys. 2008;128:084106;
(c) Shao Y, Gan Z, Epifanovsky E, et al. Mol Phys. 2015;113:184;
the energetic penalties associated with cyclic constraint.
7. Conclusions
ˇ
(d) Woodcock HL, Hodošcek M, Gilbert ATB, Gill PMW, Schaefer HF, Brooks BR. J
Comp Chem. 2007;28:1485.
19. Glendening ED, Landis CR, Weinhold F. J Comp Chem. 2013;34:1429.
In summary, we have developed a novel tethering approach
for the constraint of conformationally extended dipeptides via
Please cite this article in press as: Sarnowski M.P., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.08.017