A new tool in peptide engineering: A photoswitchable stilbene-type β-hairpin mimetic

Article abstract of DOI:10.1002/chem.200500648

Peptide secondary structure mimetics are important tools in medicinal chemistry, as they provide analogues of endogenous peptides with new physicochemical and pharmacological properties. The development, synthesis, photochemical investigation, and conformational analysis of a stil-bene-type β-hairpin mimetic capable of light-triggered conformational changes have been achieved. In addition to standard spectroscopic techniques (nuclear Overhauser effects, amide temperature coefficients, circular dichroism spectroscopy), the applicability of self-diffusion measurements (longitudinal eddy current delay pulsed-field gradient spin echo (LED-PGSE) NMR technique) in conformational studies of oligopeptides is demonstrated. The title compound shows photoisomerization of the stilbene chromophore, resulting in a change in solution conformation between an unfolded structure and a folded β-hairpin.

Full text of DOI:10.1002/chem.200500648

FULL PAPER
DOI: 10.1002/chem.200500648
A New Tool in Peptide Engineering: A Photoswitchable Stilbene-type
b-Hairpin Mimetic
MµtØ ErdØlyi,[a, b] Anders KarlØn,[b] and Adolf Gogoll*[a]
Abstract: Peptide secondary structure
mimetics are important tools in medici-
nal chemistry, as they provide ana-
logues of endogenous peptides with
new physicochemical and pharmaco-
logical properties. The development,
synthesis, photochemical investigation,
and conformational analysis of a stil-
bene-type b-hairpin mimetic capable of
light-triggered conformational changes
have been achieved. In addition to
standard spectroscopic techniques (nu-
clear Overhauser effects, amide tem-
perature coefficients, circular dichroism
spectroscopy), the applicability of self-
diffusion measurements (longitudinal
eddy current delay pulsed-field gradi-
ent spin echo (LED-PGSE) NMR
technique) in conformational studies of
oligopeptides is demonstrated. The title
compound shows photoisomerization
of the stilbene chromophore, resulting
in a change in solution conformation
between an unfolded structure and a
folded b-hairpin.
Keywords: beta-hairpin · peptido-
mimetics · photochemistry · self-
diffusion · stilbene
Introduction
we present here a photoswitchable, stilbene-based peptido-
mimetic and its conformational analysis.
Molecular switches are compounds capable of existing in
two or more interconvertible conformational states.[1] Exter-
nal switching between these isomers may be driven by pH
change, chemical reactions, complexation, electron transfer,
heat, or magnetic or electric fields, or it may be triggered by
irradiation at a selected wavelength.[2] The development of
such photoresponsive systems is of outstanding interest be-
cause they have considerable potential for application as
new tools in biomedical engineering.[3] Incorporation of a
molecular switch into bioactive compounds may allow exter-
nal modulation of their biological effect: it may, for exam-
ple, enable photoactivation of a prodrug, in a manner re-
sembling the bioactivation of vitamin D,[4] photoactivation
of a drug in a specific area in the human body, or light-trig-
gered association/dissociation of a ligand–receptor complex.
As a key step towards the development of such compounds,
The b-hairpin motif is involved in numerous vital physio-
logical processes and pathological disorders. Because the
biological activities of some b-hairpins[5] have been shown to
be correlated with the thermodynamic stabilities of their
folded conformations,[6] a b-hairpin mimetic incorporating a
stilbene moiety should be an attractive candidate for a pho-
toswitchable peptidomimetic. Such a mimetic should allow
externally triggered interconversion between a bioactive b-
hairpin and a bioinactive nonhairpin conformation. The in-
corporation of
a diazobenzene dipeptide mimic into
cyclic[3,7] and acyclic[8] peptides was recently reported, but
the hairpin-inducing ability of this turn mimic appeared to
be limited.[7b] In addition, diazobenzenes are in many ways
inferior to stilbene derivatives in that they undergo thermal
cis–trans isomerization and are sensitive to reducing agents,
whilst the separation of their isomers is also often compli-
cated, if not impossible.[7b,8] At present, no attempts at incor-
poration of stilbene derivatives into peptides have been re-
ported. Because of their greater chemical stability, however,
they would be attractive alternatives to azobenzenes.
[a] Dr. M. ErdØlyi, Prof. Dr. A. Gogoll
Department of Chemistry, Organic Chemistry
Uppsala University, Box 599, 751 24 Uppsala (Sweden)
Fax : (+46)18-471-3818
E-mail: adolf.gogoll@kemi.uu.se
Results and Discussion
[b] Dr. M. ErdØlyi, Dr. A. KarlØn
Department of Medicinal Chemistry, Uppsala University
Box 574, 751 23 Uppsala (Sweden)
Here we present the development of a stilbene-type pepti-
domimetic capable of light-triggered conformational
changes between an unfolded structure and a folded b-hair-
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2006, 12, 40030–64W12iꢀley2-VCH Verlag GmbH & Co. KGaA, Weinheim
403

pin form. In this model study, the trans isomers of a cyclic
and an acyclic peptidomimetic—cyclo(Leu-Leu-Val-Ile-stil-
bene-Thr-Thr-Ala-Leu-Gly-DPro) and Leu-Leu-Val-Ile-stil-
bene-Thr-Thr-Ala-Leu-X—were prepared by combination of
solution-phase and solid-phase methodologies (X=diethyl-
aminoacetyl; Scheme 1) and their cis isomers were obtained
by subsequent photoisomerization. The amino acid strands
of the investigated compounds were extracted from the S4
region of the TATA-box binding protein, and are known to
fold into b-sheets in their native environments.[9] Surprising-
ly, no attempts at application of these amino acid strands in
model hairpin studies have yet been reported, although
their sequences are very likely to provide stable b-hairpin
conformations, as aliphatic b-branched amino acids are
known to promote the formation of hydrophobic clusters,[10]
and so are frequent constituents of the strand segments of
native b-hairpins.[11] For comparison, the nonswitchable ana-
logue Leu-Leu-Val-Ile-Gly-DPro-Thr-Thr-Ala-Leu-X was
also prepared by standard solid-phase peptide synthesis
(SPPS). A DPro-Gly dipeptide incorporated into 1 is a well-
known b-turn inducer,[12] so this peptide was designed to
fold into a b-hairpin. The acyclic (2 and 3) and cyclic (4 and
5) peptidomimetics incorporating a photoswitchable stilbene
unit were expected to allow externally triggered folding and
unfolding through irradiation at selected wavelength(s). In
addition, diethylaminoacetyl tails were also attached to the
N termini of all but the cyclic peptides 4 and 5 to increase
their solubility in polar solvents.[13]
Photochemistry: Isomerizations of 2–5 were carried out in
dimethyl sulfoxide by selective use of irradiation at the ab-
sorption maxima (lmax =300 or 280 nm)[14] of the trans- or
cis-stilbene double bonds. In the case of 2, the maximum
percentage of cis isomer 3 at the photostationary state was
reached after 3 h of irradiation (l=300 nm) and was 63%,
1
as determined by H NMR spectroscopy. For 4, the photo-
Scheme 1. The structures of b-hairpin mimetic 1, together with its open-chain (2, 3) and cyclic (4, 5) photoswitchable analogues. Amino acid building
blocks and photochemical transitions are indicated.
404
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 403 – 412

Stilbene-type b-Hairpin Mimetic
FULL PAPER
stationary state (80% cis isomer 5) was reached after 90 min
of irradiation at l=300 nm. Isomerization of the cis isomer
5 at l=280 nm resulted in 16% of the trans compound 4.
Attempts to separate the cis and trans isomers by prepara-
tive reversed-phase HPLC were successful in the case of
compounds 4 and 5, so—in contrast with similar studies on
azobenzene derivatives[15]—the cis peptidomimetic 5 was an-
alyzed separately from its trans isomer, subsequent to their
separation. Because of the lack of separation of the acyclic
compounds, the conformation of 3 was successfully investi-
gated in the cis–trans mixture obtained after isomerization
of 2.
Conformational analysis by NMR spectroscopy: The most
reliable and extensively employed tool used today for explo-
ration of peptide solution conformations is without doubt
the detection of nuclear Overhauser effects (NOEs).[16] The
observation of nonsequential NOEs hence provides the
strongest evidence for the three-dimensional solution struc-
ture of a compound, and so conformational analysis was
first performed with standard NMR techniques in water,
methanol, and dimethyl sulfoxide.[13] As expected, inter-
strand NOEs observed in the solutions of compound 1 in
methanol and dimethyl sulfoxide were characteristic of a b-
hairpin (Scheme 2). However, the absence of such NOEs in
aqueous solution suggests that folded conformations ob-
served in methanol and dimethyl sulfoxide have low thermo-
dynamic stability. No NOEs between nonadjacent residues
were observed in the cases of compounds 2, 3, and 4, thus
clearly indicating unfolded solution structures. It should be
mentioned, though, that the NOESY spectra obtained for 4
showed a few indications of the formation of a DPro-Gly b-
turn. In contrast with its open-chain analogue 3, interstrand
NOEs observed in the case of the cyclic peptidomimetic 5
provided strong evidence of its folding into a b-hairpin con-
formation (Scheme 2).
Scheme 2. NMR evidence for folded b-hairpin conformations in com-
pounds 1 and 5. The building blocks, the observed amide temperature co-
efficients (bold), and interstrand NOEs (arrows) of 1 and 5 in methanol
are indicated.
tion of compound 1 in dimethyl sulfoxide (Table 1) were
consistent with a folded b-hairpin structure in which the
amide protons of Leu-9 and Ile-7 are intramolecularly hy-
drogen bonded and Thr-4 is in equilibrium between hydro-
gen-bonded and nonbonded states.[17] The Dd/DT values ob-
tained for 1 in methanol imply that the folded conformation
has low thermodynamic stability and exists in exchange with
unfolded structures in competitive solvents, an observation
in good agreement with the NOE studies performed with
water solutions (Scheme 2). Moreover, the alternating
amide temperature coefficients of the cis isomer of the
cyclic peptidomimetic 5 are typical of a folded b-hairpin
structure in which the amide protons of Leu-1, Thr-3, Stil-
The absence of any aggregation phenomena was revealed
by the concentration independence of the amide chemical
1
shifts, as well as by the observation of sharp H NMR signals
for all investigated compounds. No chemical shift change or
line-broadening was observed over a period of three weeks.
In addition, pulsed-field gradi-
ent spin echo (PGSE) NMR
Table 1. The amide proton temperature coefficients DdNH/DT [ppbKÀ1] obtained in dimethyl sulfoxide and
translational diffusion studies
performed with dimethyl sulf-
oxide confirmed that all com-
methanol.[a]
1
2
3
4
5
DMSO MeOH DMSO MeOH DMSO MeOH DMSO MeOH DMSO MeOH
pounds were present in
a
L-1
A-25.0
T-3
3.4
8.0
5.3
3.6
4.5
5.7
4.5
7.4
8.7
6.7
5.4
1.7
monomeric, disaggregated form.
5.2 4.7
7.5
4.7
8.8
4.0
6.1
Temperature
(Dd/DT) offer
coefficients
a qualitative
4.0
3.2
2.6
1.6
5.5
2.7
4.3
7.8
4.5
6.8
4.7
10.25.25.1
6.7
3.6
2.2
3.8
4.1
3.7
2.3
3.9
4.27.8
8.23.7
5.9
5.7
4.24.4 .8
5.5
7.3
41.20.7
9.9
2.3
3.9
6.7
4.5
4.3
2.6
3.8
T-4
4.4
3.9
5.6
3.5
3.9
5.6
4.6
X-6[b]
I-7
measure of the involvement of
amide protons in intramolecu-
lar hydrogen bonding and thus
an additional insight into pep-
tide secondary structures.[13a] As
expected, the temperature coef-
2
7.9
7.5
V-8
L-9
L-10
9.7
5.1
4.0
4.3
3.8
3.6
3.7
4.28.1
3.7
5.5
6.3
4.5
7.0
7.3
5.7
4.9
3.9
[a] Temperature coefficients obtained as values (dT,highÀdT,low)/(ThighÀTlow) that are negative numbers, but are
reported as positive values in accordance with accepted literature conventions.[13a] [b] X-6=G-6 or Stilbene-
ficients obtained for the solu- 5/6.
Chem. Eur. J. 2006, 12, 400036–W41il2eꢀy-V2 CH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
405

A. Gogoll et al.
bene-5/6, Val-8, and Leu-10 are intramolecularly hydrogen
bonded (Scheme 2). It should be noted that the DdNH/DT
values for this compound gradually increase in the direction
of the stilbene linker, indicating the inferior turn-inducing
properties of the artificial dipeptide mimic in relation to the
DPro-Gly sequence. In agreement with the NOESY and
ROESY studies, the high DdNH values observed for the solu-
tions of the trans isomers of the acyclic and cyclic mimetics
2 and 4 and the acyclic cis isomer 3 were indicative of un-
folded structures. The low temperature coefficients generally
observed for the NH of Thr-4 might be due to the formation
of a six-membered ring through hydrogen bonding to the
hydroxy group of Thr-3.
through the use of NMR constraints derived from NOE
data obtained for solutions of the compounds in methanol
(Scheme 2) the reliability of the calculation output was con-
siderably improved and the processor time was radically de-
creased in relation to initial unrestricted calculations. The
computations resulted in 402( 1), 204 (4), and 302( 5) struc-
tures within 3 kcalmolÀ1 of the global minima. The ten
lowest-energy conformations for the reference compound 1
and the trans (4) and cis (5) isomers of the photoswitchable
cyclopeptide are depicted in Figure 1 (top, middle, and
bottom, respectively). Our computations indicate that the
folded geometric arrangement is strongly preferred in the
cases of 1 and 5, whereas 4 favors unfolded conformations
and hence, in its lack of stabilizing intramolecular hydrogen
bonds and hydrophobic interactions, is notably more flexible
than its folded analogues.
The magnitude of 3JNH,Ha is known as an easily measurable
parameter furnishing further indications of backbone con-
3
formation.[16] For compounds 2–4 the JNH,Ha coupling con-
stants fall into the ranges typical for random coil (ꢀ6–7 Hz)
or extended (ꢀ9 Hz) conformations, indicating the presence
of a mixture of interconverting conformations in solution.
The largest coupling constants indicative of extended
strands were observed for peptide 1 and 5.[14] It should be
emphasized, though, that coupling constants in themselves
are only indicators of—and do not prove any—peptide sec-
ondary structure.
Amino acid proton chemical shifts are also known indica-
tors of peptide overall conformation.[18] Residues in b-
strands or in extended conformations thus have higher
chemical shifts (0.1–0.6 ppm) than those participating in
random coils, while amino acids in a-helices and b-turns ex-
perience the opposite tendency and have low d values. The
observed chemical shift differences of the NH and Ha pro-
tons in corresponding amino acids of peptides may thus
reveal folding tendencies. The chemical shifts of peptidomi-
metics 2 and 3 are similar and hence indicate that cis–trans
isomerization does not induce folding or unfolding of the
acyclic stilbene-containing peptide.[14] In constrast, the chem-
3
ical shifts of the amide and a-protons, as well as the JaH,NH
coupling constants of most of the amino acids of cis isomer
5 of the cyclic peptidomimetic are higher than those found
for its trans isomer 4,[14] so these changes suggest a photoiso-
merization-triggered folding process in the cyclic peptidomi-
metic. Again, we would like to stress that chemical shifts,
unlike observation of nonresidual NOEs, are in themselves
only indicative but not conclusive regarding secondary struc-
3
tures. It should be noted, though, that our JaH,NH and d data
are in good agreement with the observed NOEs and DdNH
/
DT values.
Figure 1. Representative solution conformations of 1 (top), 4 (middle),
and 5 (bottom). Overlaid backbones of the ten lowest-energy conforma-
tions resulting from NOE-restrained Monte Carlo conformational search
followed by conjugate gradient minimization are shown.
Molecularmodeling : To model the preferred conformations
of our peptides, Monte Carlo conformational searches fol-
lowed by conjugate gradient minimizations were performed.
As the solution structures of the folded conformations of
our peptidomimetics were observable through NOEs and
the established structures were supported by a number of
other NMR parameters, we felt obliged to take advantage
of the way constraint-processing techniques and conforma-
tional search techniques complement each other.[19] Thus,
Circular dichroism (CD): Circular dichroism spectroscopy,
allowing the investigation of overall solution conformations,
is a further commonly applied tool in peptide chemistry.
Compound 1 exhibited a negative band indicative of a b-
hairpin conformation at 220 nm (Figure 2). The observed
low ellipticity and a second minimum at approximately
406
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 403 – 412

Stilbene-type b-Hairpin Mimetic
FULL PAPER
ation upon cis–trans isomerization of 4 and 5. The results of
the CD investigation of 1–5 are hence in good agreement
with the conclusions drawn from NMR studies.
Translational self-diffusion measurements: Translational
self-diffusion measurements using pulsed-field gradient spin
echo (PGSE) NMR methods are known tools in structural
investigation of peptides and proteins, as the translational
diffusion coefficient (Dt) is related to the mass of a com-
pound through its hydrodynamic radius.[21] Diffusion mea-
surements are therefore commonly applied in order to draw
conclusions about monomeric or oligomeric states, but have
in a few cases also been used to follow conformational
changes in polypeptides.[22] In spite of the simplicity and ra-
pidity of this technique, its use for conformational investiga-
tion of small peptides is extremely rare in the literature. We
anticipated that the light-triggered conformational change of
our cyclic peptidomimetic between b-hairpin and random
coil forms might be reflected in different diffusion behavior,
so the self-diffusion coefficients were measured by use of
the LED-PGSE[23] pulse sequence in [D6]DMSO at 258C
and were determined as 1.4”10À6 (trans isomer 4) and 1.8”
10À6 cm2 sÀ1 (cis isomer 5).[24] For comparison, we estimated
the diffusion coefficient of a peptide of comparable molecu-
lar weight by the theoretical method of Gräslund et al.[25]
This model is applicable for prediction of the translational
diffusion coefficient of an aqueous solution of a monomeric,
linear, random-coil-forming peptide of known molecular
mass. The theoretical prediction estimated somewhat slower
diffusion for our peptide than experimentally observed:
Dpt redicted =2.6”10À6 cm2 sÀ1. If allowance is made for the dif-
ferences between the conditions of the prediction and the
experiments—that is, that the measurements were per-
formed on a structured, cyclic peptide in dimethyl sulfox-
ide—the calculated diffusion coefficient is in good qualita-
tive agreement with those observed experimentally. In addi-
tion, the translational self-diffusion coefficients were also es-
timated by use of the program HYDROPRO,[26] developed
for globular proteins by Carrasco et al. The hydrodynamic
modeling predicted a coefficient of 1.1”10À6 cm2 sÀ1 for
both structures depicted in Figure 1 (4 (middle) and 5
(bottom)). The diffusion coefficients estimated by the two
theoretical methods are thus not identical with the experi-
mentally determined values, but are both of comparable
magnitude, hence confirming the absence of any aggregation
for the investigated solutions. Furthermore, in the case of
oligomerization of the b-hairpin-forming cis isomer, a de-
crease in the diffusion coefficient relative to the nonhairpin
trans isomer would be expected,[27] rather than the observed
increase. The experimentally observed small, but still signifi-
cant, increase in the diffusion coefficient upon photoisom-
erization of the trans isomer may thus be an indicator of a
decrease in the compoundꢁs hydrodynamic size.[25] Conse-
quently, the higher diffusion coefficient of the folded cis
isomer may originate from lower solvent accessibility of its
intramolecularly hydrogen-bonded amide protons, a factor
also indicated by low DdNH/DT values. The amide protons of
Figure 2. CD spectra obtained with samples in methanol. The CD spec-
trum of 3 was obtained by subtraction of the absorbance of the trans
isomer 2 from the CD spectrum of the photostationary state.
200 nm indicate that the folded conformation is in equilibri-
um with unfolded structures. The CD of 2 showed a broad
negative band at 214–220 nm, a weak positive absorption
centered at 280 nm, and a negative band at 300–350 nm.
Compound 3 showed two weak negative absorptions at 209
and 217–224 nm. Although the stilbene chromophore does
not interfere with the amide region of the spectrum, because
it has its main absorption region at 300–350 nm,[20] the ob-
tained spectra are difficult to interpret in terms of folding
because of the lack of CD data for similar compounds. The
CDs of 4 and 5 confirm that the DPro-Gly sequence func-
tions as a turn inducer independently of the geometry of the
stilbene double bond (Figure 3). Upon trans to cis isomeri-
Figure 3. The CD spectra of the cyclic photoswitchable mimetics 4 (trans)
and 5 (cis) in methanol.
zation the negative amide band shifts from 214 to 220 nm, a
change that may be indicative of a b-turn to b-hairpin transi-
tion. We would like to stress that, although the obtained
spectra are difficult to interpret, they are in good agreement
with the single available report for CD investigation of stil-
bene derivatives,[20] and the observed change in the chiropti-
cal features is in itself an indicator of a conformation alter-
Chem. Eur. J. 2006, 12, 400036–W41il2eꢀy-V2 CH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
407

A. Gogoll et al.
the unfolded trans isomer are likely to be hydrogen bonded
to solvent molecules, as previously indicated by the high
DdNH/DT values, so the intermolecular NH to DMSO hydro-
gen bonds may increase the effective size of the peptidomi-
metic. In parallel, the good hydrogen-bond-accepting prop-
erties of dimethyl sulfoxide are known to cause a radical de-
crease in the diffusion coefficient of water molecules.[28]
However, the interpretation of the observed alteration in
diffusion properties at the molecular level is rather difficult,
because, unlike in the case of water,[29] there is very little ex-
perience of the behavior of solvent layers in dimethyl sulf-
oxide,[3] whilst the properties of residual water in dimethyl
sulfoxide are also currently not well understood.[31] The dif-
fusion coefficients for all investigated compounds are sum-
marized in Table 2.
isomer (7), most possibly for the following entropic reasons:
1) high flexibility of the (CH2)2 linkers of the incorporated
stilbene dipeptide mimic, and 2) few attractive stabilizing
forces provided by the attached short tetrapeptide strands.[32]
Here we would like to emphasize that this finding does not
necessarily indicate that light-triggered isomerization would
not be applicable for conformational modulation of linear
peptides, a suggestion strongly supported by the observed
low thermodynamic stability of the folded conformation of
the nonswitchable analogue 1, and also by a very recent
report by Hilvert et al.[8,33] In addition, we have shown that
the light-induced conformational change, established by
standard spectroscopic methods, is accompanied by a signifi-
cant alteration in the translational self-diffusion coefficient
of the cyclic oligopeptide. It should also be noted that all ex-
perimental and computational data are in good agreement,
both with each other and with current theories. The fact that
4 and 5 but not 2 and 3 could be separated by reversed-
phase HPLC may further indicate a significant conforma-
tional change upon irradiation of the cyclic peptidomimetic,
resulting in considerable changes in its accessible surface
and physicochemical properties.
Table 2. Experimentally observed and calculated translational diffusion
coefficients of the investigated peptidomimetics.
[25]
Compound
Mr [gmolÀ1
]
D [cm2 sÀ1
]
Predicted D [cm2 sÀ1
]
1
2
3
4
5
1122
1246
1246
1255
1255
0.9”10À6
0.7”10À6
0.7”10À6
1.4”10À6
1.8”10À6
2.8”10À6
2.6”10À6
2.6”10À6
2.6”10À6
2.6”10À6
As b-hairpins are involved in molecular recognition
events in numerous vital physiological processes[34] and
pathological disorders,[35] we believe that their photoswitch-
able mimetics should be of considerable interest in drug de-
velopment,[7] and should also become widely used tools in
nanotechnology.[2] Further work should address the optimi-
zation of the photochemical aspects and the incorporation
of switches into specific biological systems.
As expected, the Dt values for peptides 1–3 were signifi-
cantly lower than those of 4 and 5, as these compounds are
linear and thus have larger hydrodynamic size than their
cyclic analogues. Moreover, in agreement with theory,[21] the
diffusion coefficients of the linear compounds followed the
inverse order of molecular weight. The identical diffusion
coefficients of 2 and 3 agree well with our hypothesis, as
their conformational investigation with standard NMR tools
clearly indicated that both isomers were present in solution
as ensembles of interconverting unfolded structures. In sum-
mary, the advantage of translational self-diffusion measure-
ments in peptide conformational analysis is simplicity and
rapidity. It should be noted, though, that diffusion coeffi-
cients, like CD spectra, provide only limited information
and may only indicate the overall solution conformation of
a compound, but cannot be interpreted in terms of detailed
site-specific information.
Experimental Section
Synthesis: Starting materials were purchased from commercial suppliers
and were used without purification. Solid-phase peptide synthesis of 4
was carried out on 2-chlorotrityl chloride resin on a 500 mg scale by use
of a Fmoc/tBu protection scheme. For compounds 1 and 2 solid-phase
peptide synthesis was carried out on Rink amide MBHA resin on a
500 mg scale (loading rate 0.73 mmolgÀ1) by a Fmoc/tBu protection
scheme. Chain elongation was performed with the Fmoc-protected amino
acids (110 mmol) with PyBOP-mediated (110 mmol) coupling steps (2h)
in a mixture of diisopropylethylamine (220 mmol) and DMF (3.0 cm3).
Removal of the Fmoc groups was achieved by treatment with 20% piper-
idine in DMF for 5+10 min. After introduction of each amino acid, a
Kaiser test[36] was performed and capping was carried out (30 min) by ad-
dition of acetic anhydride (1.5 cm3) in dichloromethane (2.0 cm3) and di-
isopropylethylamine (0.5 cm3). The preparation of the turn mimetic part
of the presented stilbene-type peptidomimetics is outlined in Scheme 3.
The Fmoc-protected turn mimic was then incorporated into the peptide
by standard SPPS techniques as described above. Cleavage of the prod-
ucts (1 and 2) was achieved by addition of 95% trifluoroacetic acid in di-
chloromethane (1 h+2”30 min), followed by filtration and concentration
of the solutions under reduced pressure. In order to obtain the cyclic pep-
tidomimetic, the linear peptide NH2-Thr(tBu)-Thr(tBu)-NH-Stilbene-
CO-Ile-Val-Leu-Leu-DPro-Gly-Leu-Ala-OH was cleaved from the resin
with 0.5% trifluoroacetic acid in dichloromethane (4”5 cm3), followed
by filtration into a flask containing 250 mL pyridine. The combined
phases were washed with H2O and concentrated under reduced pressure.
Preparative HPLC indicated that the purity of the cleaved product was
approximately 95%. Cyclization of the linear peptide was achieved by
HATU-mediated (110 mmol) coupling in a mixture of diisopropylethyl-
Conclusion
We have demonstrated that incorporation of a stilbene-type
dipeptide mimic into a cyclic b-hairpin allows for light-trig-
gered switching between different peptidomimetic confor-
mations. Structural changes established by standard NMR
techniques (NOEs, temperature coefficients) and indicated
by CD measurements demonstrated folded b-hairpin and
unfolded conformations for the trans and cis isomers, re-
spectively, of the presented cyclic stilbene-containing pep-
tide analogue. The comparable acyclic stilbene-type peptido-
mimetic, however, did not fold into a b-hairpin as its cis
408
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 403 – 412

Stilbene-type b-Hairpin Mimetic
FULL PAPER
7.52(d, 3J(H,H)=15.9 Hz, 1H; CH),
7.59 ppm (t, 3J(H,H)=1.8 Hz, 1H;
ArH); 13C NMR (67.5 MHz, CD3OD,
258C): d=119.5, 122.8, 126.5, 130.2,
130.6, 132.9, 136.3, 143.4, 168.6 ppm;
MS (70 eV, EI): m/z (%): 228, 226
[M]+, 147, 102, 91, 75, 51.
Methyl 3-(3-bromophenyl)acrylate (7):
Concentrated aqueous HCl (2–3
drops) was added to 3-(3-bromophe-
nyl)acrylic acid (4.96 g, 21.90 mmol) in
CH3OH (15 cm3), and the solution was
stirred for 12h. The solvent was then
evaporated and the residue was dis-
solved in CH2Cl2. The organic phase
was washed with HCl (0.1m) and
water, and the water phase was re-
extracted three times with dichlorome-
thane. The combined organic phases
were evaporated, yielding white crys-
tals (4.87 g, 20.20 mmol, 93%).
1H NMR (270.2 MHz, CDCl3, 2 58C):
d=3.80 (s, 3H; CH3), 6.42(d,
3J(H,H)=15.9 Hz, 1H; CH), 7.25 (t,
3J(H,H)=7.9 Hz, 1H; ArH), 7.42(dt,
3J(H,H)=1.0, 1.6, 7.9 Hz, 1H; ArH),
7.49 (ddd, 3J(H,H)=1.0, 1.6, 7.9 Hz,
1H; ArH), 7.60 (d, 3J(H,H)=15.9 Hz,
1H; CH), 7.65 ppm (t, 3J(H,H)=
1.6 Hz,
1H;
ArH);
13C NMR
(67.5 MHz, CDCl3, 2 85C): d=51.8,
119.2, 123.0, 126.6, 130.3, 130.7, 133.0,
136.4, 143.1, 166.9 ppm; MS (70 eV,
EI): m/z (%): 242, 240 [M]+, 184, 182,
171, 169, 104, 77, 51.
Scheme 3. Outline of the synthesis of 14. a) CH2(COOH)2, pyridine, piperidine, 1008C, 1.5 h, 95%. b) CH3OH,
conc. HCl, RT, 16 h, 93%. c) Ni(OAc)2, NaBH4, CH3OH, EtOAc, H2 (2bar), 20 min, RT, 68%. d) Bu 3SnCH=
CH2, [Pd(PPh3)2Cl2], Et3N, DMF, 1308C, 25 min, 78%. e) O[CO2C(CH3)3]2, CH2Cl2, K2CO3 in H2O, RT, 24 h,
70%. f) Pd(OAc)2, (CH3C6H4)3P, Et3N, DMF, 1208C, 30 min, 43%. g) 50% CF3COOH in CH2Cl2, 15 min.
h) Fmoc-Cl, dioxane, 10% Na2CO3 (aq), 17 h, 82%. i) CH2Cl2, conc. HCl, 22 h, 91%.
Methyl 3-(3-bromophenyl)propionate
(8): Ni(OAc)2 (2.46 g, 9.90 mmol) dis-
solved in CH3OH (15 cm3) and ethyl
acetate (30 cm3) was added to methyl
3-(3-bromophenyl)acrylate
(1.59 g,
6.60 mmol) and NaBH4 (0.75 g,
19.80 mmol) in a Parr tube. The mix-
amine (220 mmol) and DMF (3.0 cm3) overnight. The mixture was then
concentrated under reduced pressure and the tert-butyl protecting groups
ture was hydrogenated (2bar) in a Parr apparatus for 20 minutes. The
solvent was evaporated, and the residue was dissolved in CH2Cl2 and
washed with water. The organic phase was separated, and the aqueous
phase was reextracted with dichloromethane three times. The combined
organic phases were filtered through Celite and MgSO4 and were then
concentrated under reduced pressure, yielding a yellowish oil, containing
1.5% debrominated byproduct (1.09 g, 4.47 mmol, 68%). 1H NMR
(499.9 MHz, CDCl3, 2 58C): d=2.50 (t, 3J(H,H)=7.6 Hz, 2H; CH2), 2.80
were removed with 50% trifluoroacetic acid in dichloromethane.
Purification of the peptides was performed on
a Gilson 321 HPLC
system connected to a Vydac Protein & Peptide C18 (218TP) column
(10 mm, 22”250 mm) with use of a gradient of acetonitrile in 0.1% aque-
ous trifluoroacetic acid (10–85% MeCN in 75 min) at a flow rate of
5 cm3 minÀ1 and with detection by UV absorbance at 230 nm (LKB 2151
absorbance detector). The fractions were further analyzed by analytical
LC–MS.
3
3
(t, J(H,H)=7.6 Hz, 2H; CH2), 3.55 (s, 3H; CH3), 7.16 (dd, J(H,H)=1.5,
7.5 Hz, 1H; ArH), 7.17 (t, 3J(H,H)=7.5 Hz, 1H; ArH), 7.35 (dd,
3J(H,H)=1.5, 7.5 Hz, 1H; ArH), 7.37 ppm (brt, 3J(H,H)=1.5 Hz, 1H;
ArH); 13C NMR (100.6 MHz, CDCl3, 2 58C): d=30.7, 35.4, 51.8, 127.2,
129.6, 128.7, 130.3, 131.6, 143.2, 172.9 ppm; MS (70 eV, EI): m/z (%):
244, 242 [M]+, 184, 182, 171, 169, 104, 77, 63, 51.
Amino acid analyses were performed at the Department of Bioorganic
Chemistry, Biomedical Centre, Uppsala, Sweden, on 24 h hydrolysates
with an LKB 4151 alpha plus analyzer, with use of ninhydrin detection.
Experimental data: The method of preparation of the peptidomimetic 14
2-(3-Vinylphenyl)ethylamine (9): 3-Bromophenethylamine (500.0 mg,
2.50 mmol), [Pd(PPh3)2Cl2] (52.6 mg, 75.0 mmol), tributylvinyltin (1.1 cm3,
3.75 mmol), and LiCl (264.9 mg, 6.30 mmol) in DMF (1.5 cm3) were stir-
red in a Smith Process Vial at 1308C for 25 min in the microwave cavity.
This procedure was repeated four times, and the combined reaction mix-
ture was then filtered through Celite and extracted with concentrated
aqueous NaHCO3. The organic phase was then extracted with aqueous
HCl (1m). By addition of NaOH pellets, the pH of the aqueous solution
was increased to 14, which was followed by extraction with CH2Cl2. The
organic phase was filtered through CaCO3 and concentrated under re-
is outlined in Scheme 3.
3-(3’-Bromophenyl)acrylic acid (6): A mixture of 3-bromobenzaldehyde
(5.00 g, 27.0 mmol), malonic acid (4.39 g, 42.2 mmol), pyridine (2.93 g,
37.1 mmol), and piperidine (7–8 drops) was heated at reflux on an oil
bath (1008C) for 1.5 h. Ice and concentrated HCl (12cm 3) were then
added and the formed crystals were filtrated and washed with HCl (1m)
and water. The product was recrystallized from 95% ethanol and washed
with cooled methanol, yielding white crystals (4.77 g, 21.0 mmol, 95%).
1H NMR (270.2 MHz, CD3OD, 258C): d=6.34 (d, 3J(H,H)=15.9 Hz,
1H; CH), 7.19 (t, 3J(H,H)=7.8 Hz, 1H; ArH), 7.38 (ddd, 3J(H,H)=1.0,
1.8, 7.8 Hz, 1H; ArH), 7.43 (ddd, 3J(H,H)=1.0, 1.8, 7.8 Hz, 1H; ArH),
duced pressure, yielding
a yellowish oil (1.15 g, 7.81 mmol, 78%).
1H NMR (399.8 MHz, CDCl3, 2 58C): d=1.01 (brs, 2H; NH2), 2.64 (t,
Chem. Eur. J. 2006, 12, 400036–W41il2eꢀy-V2 CH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
409

A. Gogoll et al.
3
3J(H,H)=6.8 Hz, 2H; CH2), 2.86 (t, J(H,H)=6.8 Hz, 2H; CH2), 5.17 (d,
ethyl acetate (2:1) eluent mixture, yielding a white solid (887.7 mg,
1.67 mmol, 82%). 1H NMR (399.8 MHz, CDCl3, 2 85C): d=2.67 (t,
3J(H,H)=7.7 Hz, 2H; CH2), 2.86 (t, 3J(H,H)=6.7 Hz, 2H; CH2), 2.98 (t,
3J(H,H)=7.7 Hz, 2H; CH2), 3.50 (dt, 3J(H,H)=6.7, 6.9 Hz, 2H; CH2),
3.69 (s, 3H; CH3), 4.22 (t, 3J(H,H)=6.8 Hz, 1H; Fmoc-CH), 4.41 (d,
3J(H,H)=6.8 Hz, 2H; Fmoc-CH2), 4.88 (t, 3J(H,H)=6.9 Hz, 1H; NH),
7.09 (m, 2H; CH), 7.11 (d, 3J(H,H)=7.3 Hz, 2H; ArH), 7.25–7.35 (m,
3J(H,H)=10.8 Hz, 1H; CH), 5.68 (d, 3J(H,H)=17.6 Hz, 1H; CH), 6.63
(dd, 3J(H,H)=10.8, 17.6 Hz, 1H; CH), 7.00 (d, 3J(H,H)=6.2Hz, 1H;
ArH), 7.15–7.20 ppm (m, 3H; ArH); 13C NMR (100.5 MHz, CDCl3,
258C): d=40.1, 43.6, 113.8, 124.1, 126.8, 128.4, 128.7, 136.9, 137.7,
140.2ppm; MS (30 eV, ESI): m/z (%): 189, 148 [M+H]+.
[2-(3-Vinylphenyl)ethyl]carbamic acid tert-butyl ester(10) : 2-(3-Vinyl-
phenyl)ethylamine (950.0 mg, 6.50 mmol) and di-tert-butyl dicarbonate
(1.55 g, 7.10 mmol) were dissolved in CH2Cl2 (20 cm3) and mixed with
aqueous potassium carbonate (20 cm3, 2.68 g, 19.4 mmolK2CO3). The het-
erogeneous mixture was rigorously stirred for 24 h, and the organic phase
was then separated, filtered through MgSO4, and concentrated. The resi-
due was purified by column chromatography with a hexane/ethyl acetate
(9:1) eluent mixture, yielding a colorless oil (113.5 mg, 4.5 mmol, 70%).
1H NMR (399.8 MHz, CDCl3, 2 58C): d=1.43 (s, 9H; CH3Boc), 2.78 (t,
3J(H,H)=6.9 Hz, 2H; CH2), 3.36 (t, 3J(H,H)=6.9 Hz, 2H; CH2), 4.61
(brs, 1H; NH), 5.26 (d, 3J(H,H)=11.0 Hz, 1H; CH), 5.74 (d, 3J(H,H)=
17.6 Hz, 1H; CH), 6.69 (dd, 3J(H,H)=11.0, 17.6 Hz, 1H; CH), 7.08 (d,
3J(H,H)=6.0 Hz, 1H; ArH), 7.22–7.29 ppm (m, ArH3H;); 13C NMR
(100.5 MHz, CDCl3, 2 58C): d=28.5 (3C), 36.2, 41.8, 79.3, 114.0, 124.4,
126.8, 128.4, 128.8, 136.8, 137.9, 139.3, 156.0 ppm; MS (30 eV, ESI): m/z
(%): 495 [2M+H]+ 405, 289, 248 [M+H]+, 233.
3
8H; ArH), 7.39 (t, J(H,H)=7.3 Hz, 2H; Fmoc-ArH), 7.58 (d, 3J(H,H)=
7.5 Hz, 2H; Fmoc-ArH), 7.76 ppm (d, 3J(H,H)=7.5 Hz, 2H; Fmoc-
ArH); 13C NMR (100.5 MHz, CDCl3, 2 58C): d=31.0, 35.7, 36.3, 42.3,
47.4, 51.8, 66.7, 120.1, 124.7, 124.9, 125.2, 126.6, 127.1, 127.2, 127.7, 127.8,
128.3, 128.6, 128.9, 129.0, 129.1, 137.6, 137.8, 139.3, 141.0, 141.4, 144.1,
156.4, 173.4 ppm; MS (30 eV, ESI): m/z (%): 1063.1 (22) [2M+H]+, 532.1
(100) [M+H]+.
trans-3-{3-{2-{3-[2-(9H-Fluoren-9-ylmethoxycarbonylamino)ethyl]phe-
nyl}vinyl}phenyl}propionic acid (14): Compound 17 (106.0 mg,
0.20 mmol), dissolved in dichloromethane (4.5 cm3), was added to con-
centrated aqueous HCl solution (40 cm3) and was heated at reflux at
1208C for 22 h. The mixture was then extracted with dichloromethane,
and the organic phase was separated, filtered through MgSO4, and con-
centrated under reduced pressure, yielding
a white solid (639.4 mg,
1.24 mmol, 91%). 1H NMR (399.8 MHz, CD3CN, 258C): d=2.63 (t,
3J(H,H)=8.0 Hz, 2H; CH2), 2.85 (t, 3J(H,H)=7.0 Hz, 2H; CH2), 2.92 (t,
3J(H,H)=8.0 Hz, 2H; CH2), 3.42(brt, 3J(H,H)=7.0 Hz, 2H; CH2), 4.20
(t, 3J(H,H)=7.0 Hz, 1H; Fmoc-CH), 4.31 (d, 3J(H,H)=7.0 Hz, 2H;
Fmoc-CH2), 7.13 (d, 3J(H,H)=7.7 Hz, 1H; ArH), 7.16 (d, 3J(H,H)=
7.7 Hz, 1H; ArH), 7.22 (m, 2H; CH), 7.27 (t, 3J(H,H)=7.7 Hz, 2H;
Methyl trans-3-{3-{2-[3-(2-tert-butoxycarbonylaminoethyl)phenyl]vinyl}-
phenyl}propionate (11): A mixture of 10 (90.0 mg, 0.3 mmol), 8 (80.2mg,
0.33 mmol), Pd(OAc)2 (3.4 mg, 15.0 mmol), tri-o-tolyl-phosphine (9.1 mg,
30.0 mmol), and triethylamine (0.10 cm3, 0.90 mmol) in dimethylform-
amide (1.50 cm3) was stirred in a Smith Process Vial at 1208C for 30 min
in the microwave cavity. The resulting mixture was filtered through
Celite into a separation funnel. Dichloromethane (25 cm3) was added,
3
ArH), 7.29 (t, J(H,H)=7.3 Hz, 2H; ArH), 7.38 (t, 3J(H,H)=7.3 Hz, 2H;
3
ArH), 7.43 (d, J(H,H)=7.7 Hz, 2H; ArH), 7.48 (2”brs, 2H; ArH), 7.66
(d, 3J(H,H)=7.3 Hz, 2H; Fmoc-ArH), 7.83 ppm (d, 3J(H,H)=7.3 Hz,
2H; Fmoc-ArH); 13C NMR (100.5 MHz, CD3CN and 1 drop of CD3OD,
258C): d=30.7, 35.1, 35.9, 42.1, 47.3, 65.9, 120.0, 124.4, 124.5, 125.3,
126.6, 127.0, 127.1, 127.6, 127.7, 127.72, 128.2, 128.6, 128.63, 128.8, 137.6,
137.7, 140.0, 141.3, 141.6, 144.4, 156.4, 173.2ppm; MS (30 eV, ESI): m/z
(%): 1035.0 (14) [2M+H]+, 518.1 (1) [M+H]+.
3
and the organic phase was extracted with HCl (1.0m, 2 5 cm) and concen-
trated aqueous NaHCO3 solution, the aqueous phase being reextracted
twice. The combined organic layers were filtered through MgSO4 and
concentrated under reduced pressure. The residue was purified by
column chromatography with a hexane/ethyl acetate eluent mixture with
a gradient of 9:1 to 1:1, yielding a white precipitate (53.1 mg, 0.13 mmol,
43%). M.p.=80–828C (from EtOAc); 1H NMR (399.8 MHz, CDCl3,
258C): d=1.45 (s, 9H; CH3Boc), 2.66 (t, 3J(H,H)=7.6 Hz, 2H; CH2), 2.69
(brt, 3J(H,H)=6.9 Hz, 2H; CH2), 2.96 (t, 3J(H,H)=7.6 Hz, 2H; CH2),
3.40 (brt, 3J(H,H)=6.9 Hz, 2H; CH2), 3.68 (s, 3H; CH3), 4.60 (brs, 1H;
NH), 7.08 (m, 2H; CH), 7.10 (d, 3J(H,H)=7.3 Hz, 2H; ArH), 7.27 (dt,
3J(H,H)=2.2, 7.6 Hz, 2H; ArH), 7.33–7.40 ppm (m, 4H; ArH); 13C NMR
(100.5 MHz, CDCl3, 2 58C): d=28.5 (3C), 31.0, 35.8, 36.3, 41.8, 51.7, 79.3,
124.6, 124.8, 126.6, 127.0, 127.7, 128.2, 128.7, 128.8, 128.9, 129.0, 137.6,
137.7, 139.5, 141.0, 156.0, 173.4 ppm; IR (CHCl3): n˜max =3446, 3055, 2984,
1725, 1433, 1262 cmÀ1; MS (30 eV, ESI): m/z (%): 819 [2M+H]+, 410
[M+H]+, 354.
(C2H5)2N-(CH2)2CO-Leu-Ala-Thr-Thr-DPro-Gly-Ile-Val-Leu-Leu-NH2
(1): 84.4 mg, 75 mmol, 20.6%; [a]D =À73.58 (methanol, 178C, pH 3.2);
MS (ESI, 30 eV) m/z (%): 1123.4 (11) [M+H]+, 562.6 (100) [M+2H] +
;
D=0.90”10À6 cm2 sÀ1 ([D6]DMSO, 258C); amino acid analysis: Thr 2.01,
Pro 1.03, Gly 1.03, Ala 1.01, Val 0.93, Ile 0.93, Leu 3.05 (71% peptide).
trans-(C2H5)2N-(CH2)2CO-Leu-Ala-Thr-Thr-(CH2)2PhCH=CHPh(CH2)2-
Ile-Val-Leu-Leu-NH2 (2): 29.9 mg, 24 mmol, 6.6%; [a]D =À64.38 (metha-
nol, 198C, pH 3.0); D=0.7”10À6 cm2 sÀ1 ([D6]DMSO, 258C); MS (ESI,
30 eV) m/z (%): 1246.7 (17) [M+H]+, 624.5 (100) [M+2H] +; amino acid
analysis: Thr 2.08, Ala 1.00, Val 0.90, Ile 0.89, Leu 3.00 (71% peptide).
cis-(C2H5)2N-(CH2)2CO-Leu-Ala-Thr-Thr-(CH2)2PhCH=CHPh(CH2)2-
Ile-Val-Leu-Leu-NH2 (3): Quantum yield=2.9%; MS (ESI, 30 eV) m/z
(%): 1246.7 (0.5) [M+H]+, 624.5 (100) [M+2H] +; D=0.7”10À6 cm2 sÀ1
([D6]DMSO, 258C).
Methyl
trans-3-{3-{2-[3-(2-aminoethyl)phenyl]vinyl}phenyl}propionate
(12): Compound 11 (202.5 mg, 0.49 mmol) was mixed with trifluoroacetic
acid in dichloromethane (50%) and stirred for 15 min. The solution was
then concentrated under reduced pressure, yielding
a yellowish oil
trans-Cyclo(-DPro-Gly-Leu-Ala-Thr-Thr-(CH2)2PhCH=CHPh(CH2)2-Ile-
(202.5 mg, 6.55 mmol, 93%). 1H NMR (399.8 MHz, CDCl3, 2 58C): d=
2.72 (t, 3J(H,H)=7.7 Hz, 2H; CH2), 2.96 (t, 3J(H,H)=7.7 Hz, 2H; CH2),
3.02(t, 3J(H,H)=7.4 Hz, 2H; CH2), 3.41 (brt, 3J(H,H)=7.4 Hz, 2H;
Val-Leu-Leu-) (4): [a]1D9 =À42.9 (methanol); MS (30 eV, ESI): m/z (%):
1256.4 (17) [M+H]+, 629.1 (100) [M+2H]2+
;
D=1.4”10À6 cm2 sÀ1
([D6]DMSO, 258C); amino acid analysis: Thr 2.01, Pro 1.06, Gly 1.03,
Ala 0.96, Val 0.99, Ile 0.94, Leu 3.01 (54% peptide).
cis-Cyclo(-DPro-Gly-Leu-Ala-Thr-Thr-(CH2)2PhCH=CHPh(CH2)2-Ile-
Val-Leu-Leu-) (5): Quantum yield=1.2% (300 nm); [a]1D9 =À66.8 (meth-
anol); MS (30 eV, ESI): m/z (%): 1256.4 (17) [M+H]+, 629.1 (100)
[M+2H] +; D=1.8”10À6 cm2 sÀ1 ([D6]DMSO, 258C).
CH2), 3.72(s, 3H; CH ), 6.91 (brs, 2H; NH2), 7.06 (m, 2H; CH), 7.08 (d,
3
3J(H,H)=7.3 Hz, 2H; ArH), 7.25–7.32 (m, 4H; ArH), 7.34 (t, 3J(H,H)=
7.7 Hz, 1H; ArH), 7.43 (d, 3J(H,H)=7.7 Hz, 1H; ArH), 12.10 ppm (brs,
1H; COOH); 13C NMR (100.5 MHz, CDCl3, 2 58C): d=30.9, 33.2, 35.9,
42.0, 52.7, 124.9, 126.1, 126.5, 126.8, 127.7, 127.8, 128.0, 129.1, 129.4,
129.8, 135.1, 137.4, 138.5, 140.3, 176.5 ppm; MS (30 eV, ESI): m/z (%):
620.1 (1.4) [2M+H]+, 310.1 (100) [M+H]+.
Photoisomerization: Photochemical reactions were performed in dimeth-
yl sulfoxide under N2 gas flow with use of an Oriel 1000 W Xe ARC light
source and a 300 nm or a 280 nm Oriel UV filter. The emitted light inten-
sity was determined at the wavelengths of isomerizations (280 or 300 nm)
with use of a UV enhanced Silica photodiode (5.8 mm2) attached to a cur-
rent meter.
Methyl trans-3-{3-{2-{3-[2-(9H-fluoren-9-ylmethoxycarbonylamino)ethyl]-
phenyl}vinyl}phenyl}propionate
(13):
Compound
12
(630.4 mg,
2.04 mmol) and 9-fluorenylmethyl chloroformate (582.3 mg, 2.25 mmol)
were dissolved in a mixture of dioxane (30 cm3) and aqueous Na2CO3 so-
lution (10%), and the mixture was stirred for 17 h at room temperature.
The mixture was then extracted with CH2Cl2, and the organic phase was
separated, filtered through MgSO4, and concentrated under reduced pres-
sure. The residue was purified by column chromatography with a hexane/
NMR measurements: NMR spectra were recorded on Varian INOVA
(1H at 499.9 MHz), Jeol EX-400 (1H at 399.8, and 13C at 100.5 MHz), or
Jeol EX-270 (1H at 270.2, and 13C at 67.5 MHz) spectrometers. Signal as-
410
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 403 – 412

Stilbene-type b-Hairpin Mimetic
FULL PAPER
signment was carried out by use of COSY,[37] NOESY,[38] ROESY,[39]
wettntocsy,[40] and wetroesy[40] experiments performed at 258C. NOE ef-
fects were measured with mixing times of between 0.3 and 1.0 s. The pH
values of the NMR samples were in the 3.2–3.5 range for CH3OH/
CD3OD and the 4.6–4.8 range for [D6]DMSO solutions. These values
were measured with an NMR electrode and are uncorrected. Amide
proton temperature coefficients DdNH/DT (ppbKÀ1) were measured for
3 mmoldmÀ3 samples in DMSO (298–388 K), and CH3OH/CD3OD (1:1,
188–328 K) solutions. For the PGSE experiments performed in
[D6]DMSO at 258C, z-gradients were employed and 16 scans were ac-
quired. 1 s relaxation delay, 9 ms gradient pulse duration, 20 ms diffusion
delay, 5 ms storage delay was used and the gradient pulse strength was ar-
rayed between 0 and 20 gausscmÀ1 (20 steps). The diffusion coefficients
were calculated from the known coefficient of H2O in [D6]DMSO (9”
10À6 cm2 sÀ1).[28]
bradt, S. Deindl, E.-K. Sinner, D. Oesterhelt, C. Renner, L. Morod-
er, Chem. Biol. 2003, 10, 487; d) M. Sisido, M. Harada, K. Kawashi-
ma, Y. Okahata, Biopolymers 1998, 47, 159; e) O. Pieroni, A. Fissi,
N. Angelini, F. Lenci, Acc. Chem. Res. 2001, 34, 9; f) F. D. Lewis, X.
Liu, J. Am. Chem. Soc. 1999, 121, 11928; g) F. D. Lewis, Y. Wu, X.
Liu, J. Am. Chem. Soc. 2002, 124, 12165; h) M. Schutt, S. S. Krupka,
A. G. Milbradt, S. Deindl, E. K. Sinner, D. Oesterhelt, C. Renner, L.
Moroder, Chem. Biol. 2003, 10, 487; i) T. Muraoka, K. Kinbara, Y.
Kobayashi, T. Aida, J. Am. Chem. Soc. 2003, 125, 5612; j) D. G.
Flint, J. R. Kumita, O. S. Smart, G. A. Woolley, Chem. Biol. 2002, 9,
391; k) V. Borisenko, D. C. Burns, Z. Zhang, G. A. Woolley, J. Am.
Chem. Soc. 2000, 122, 6364; l) D. C. Burns, D. G. Flint, J. R. Kumita,
H. J. Feldman, L. Serrano, Z. Zhang, O. S. Smart, G. A. Woolley,
Biochemistry 2004, 43, 15329.
[4] R. S. Becker, Photochem. Photobiol. 1988, 48, 369.
CD spectroscopy: Optical rotation was measured with
a Perkin–
[5] Unlike in the case of b- and g-turns, for b-hairpins it is not the
amino acids of the turn region but those positioned in the extended
peptide strands that are most likely to be responsible for biological
activity. See, for example: S. M. Butterfield, M. L. Waters, J. Am.
Chem. Soc. 2003, 125, 9580.
[6] a) B. A. Armitage, Annu. Rep. Prog. Chem. Sect. B 2000, 96, 187;
b) J. R. Lai, B. R. Huck, B. Weisblum, S. H. Gellman, Biochemistry
2002, 41, 12835.
[7] a) L. Ulysse, J. Cubillos, J. Chmielewski, J. Am. Chem. Soc. 1995,
117, 8466; b) C. Renner, J. Cramer, R. Behrendt, L. Moroder, Bio-
polymers 2000, 54, 501.
[8] A. Aemissegger, V. Kräutler, W. F. van Gunsteren, D. Hilvert, J.
Am. Chem. Soc. 2005, 127, 2929.
Elmer 241 polarimeter and samples in methanol. Circular dichroism spec-
tra were obtained on a JASCO J-810 spectropolarimeter from 190 to
400 nm in a 0.2mm pathlength cell. Five scans were accumulated at am-
bient temperature with a scanning speed of 100 nmminÀ1, with sample
solutions (0.56 mmoldmÀ3) in methanol (pH 4.6–4.8). UV spectra were
measured with a Varian Cary 3 spectrometer.
Mass spectrometry: ESI-mass spectra (ESI=electrospray ionization)
were obtained with a Finnigan ThermoQuest AQA mass spectrometer
(ESI 30 eV, probe temperature 1008C) fitted with a Gilson 322-H2 Gra-
dient Pump system and a SB-C18 column. A water/acetonitrile/formic
acid (0.05%) mobile phase was used with a gradient of 20 to 100% ace-
tonitrile during 3–5 min.
[9] D. B. Nikolov, H. Chen, E. D. Halay, A. Hoffman, R. G. Roeder,
S. K. Burley, Proc. Natl. Acad. Sci. USA 1996, 93, 4862.
[10] a) A. J. Maynard, G. J. Sharman, M. S. Searle, J. Am. Chem. Soc.
1998, 120, 1996; b) K. Y. Tsang, H. Diaz, N. Graciani, J. W. Kelly, J.
Am. Chem. Soc. 1994, 116, 3988.
[11] a) D. L. Minor, Jr., P. S. Kim, Nature 1994, 367, 660; b) K. Gunsekar-
an, C. Ramakrishnan, P. Balaram, Protein Eng. 1997, 10, 1131.
[12] J. D. Fisk, D. R. Powell, S. H. Gellman, J. Am. Chem. Soc. 2000, 122,
5443.
Computational chemistry: Theoretical conformational analysis without
experimental constraints was performed first for observation of folding
tendencies. In addition, calculations for description of the solution struc-
tures were carried out with constraints derived from ROESY/NOESY
cross peaks. The NMR structures obtained for solutions in methanol
were hence visualized by use of the program Macromodel 7.0.[41] The
OPLS-AA all-atom force field and the General Born/Solvent Accessible
(GB/SA) surface area method developed by Still[42] were used. The
number of torsion angles allowed to vary during each Monte Carlo step
ranged from 1 to nÀ1, where n equals the total number of rotatable
bonds. Amide bonds were fixed in trans configurations. Structural con-
straints derived from NOESY cross peaks (illustrated in Scheme 2) were
introduced by use of the DISC command (1.0–5.0 Š) as implemented in
Macromodel 7.0. Conformational searches were conducted by use of the
Systematic Unbound Multiple Minimum (SUMM) search method[43] im-
plemented in the Batchmin program. 10000 Monte Carlo step runs were
performed and those conformations within 25 kJmolÀ1 of the global mini-
mum were kept. PR Conjugate Gradient minimization with 5000 itera-
tions was used in the conformational search. The threonine hydroxy
groups were protected with methyl groups during the computation in
order to compensate for Macromodelꢁs well-known overestimation of
electrostatic forces. The structures of lowest energy are depicted in
Figure 1.
[13] As the low water solubility of compound 4 and 5 did not allow their
spectroscopic conformational analysis in aqueous solutions, similarly
to other model studies (see the following references [13a–n]), the
NMR conformational investigations of 4 and 5 were performed in
methanol and dimethyl sulfoxide. Although such investigations were
performed for peptidomimetics 1–3, the details of those aqueous
studies are sparingly reported in this article and will be published
elsewhere. Here, it must be emphasized that these compounds are
model systems only signifying the usability of stilbenes for control-
led modulation of peptide conformations. a) M. ErdØlyi, V. Langer,
A. KarlØn, A. Gogoll, New J. Chem. 2002, 26, 834; b) C. Das, S. Ra-
gothama, P. Balaram, J. Am. Chem. Soc. 1998, 120, 5812; c) J. M.
Soth, J. S. Nowick, J. Org. Chem. 1999, 64, 276; d) T. S. Haque, J. C.
Little, S. H. Gellman, J. Am. Chem. Soc. 1996, 118, 6975; e) S. C.
Shankaramma, S. K. Singh, A. Sathyamurthy, P. Balaram, J. Am.
Chem. Soc. 1999, 121, 5360; f) Y. J. Chung, B. R. Huck, L. A. Christi-
anson, H. E. Stanger, S. Kräuthauser, D. R. Powell, S. H. Gellman, J.
Am. Chem. Soc. 2000, 122, 3995; g) D. S. Kemp, Z. Q. Li, Tetrahe-
dron Lett. 1995, 36, 4175; h) D. S. Kemp, Z. Q. Li, Tetrahedron Lett.
1995, 36, 4179; i) J. S. Nowick, F. Insaf, J. Am. Chem. Soc. 1997, 119,
10903; j) J. S. Nowick, Acc. Chem. Res. 1999, 32, 287; k) T. S. Haque,
J. C. Little, S. H. Gellman, J. Am. Chem. Soc. 1994, 116, 4105;
l) G. P. Dado, S. H. Gellman, J. Am. Chem. Soc. 1993, 115, 42 2 8;
m) E. Gallo, S. H. Gellman, J. Am. Chem. Soc. 1993, 115, 9774;
n) E. Gallo, S. H. Gellman, J. Am. Chem. Soc. 1994, 116, 11560.
[14] Further experimental details for the photochemistry, NMR assigna-
tion, and conformational investigation of the presented compounds
are available as Supporting Information.
Acknowledgements
We thank the Swedish Research Council for financial support. We are
also grateful to Johanna Nurbo, Ida Niklason, and Šsa Persson for their
contributions during their undergraduate work in our group.
[1] B. L. Feringa, R. A. van Delden, M. K. J. ter Wiel, Pure Appl. Chem.
2003, 75, 563.
[2] B. L. Feringa, R. A. van Delden, N. Koumura, E. M. Geertsema,
Chem. Rev. 2000, 100, 1789.
[15] a) C. Renner, R. Behrendt, N. Heim, L. Moroder, Biopolymers
2002, 63, 382; b) C. Renner, J. Cramer, R. Behrendt, L. Moroder,
Biopolymers 2000, 1, 501; c) R. Behrendt, C. Renner, M. Schenk, F.
[3] a) C. Dugave, L. Demange, Chem. Rev. 2003, 103, 2 475; b) L. G.
Ulysse, J. A. Chmielewski, Pept.: Chem. Struct. Biol., Proc. Am.
Pept. Symp., 14th 1996, 203; c) M. Schütt, S. S. Krupka, A. G. Mil-
Chem. Eur. J. 2006, 12, 400036–W41il2eꢀy-V2 CH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
411

A. Gogoll et al.
Wang, J. Wachtveitl, D. Oesterhelt, L. Moroder, Angew. Chem.
1999, 111, 2941; Angew. Chem. Int. Ed. 1999, 38, 2771.
drews, B. M. Pettitt, Biopolymers 1998, 45, 469; e) M. Gerstein,
R. M. Lynden-Bell, J. Phys. Chem. 1993, 97, 2898.
[16] H. J. Dyson, P. E. Wright, Annu. Rev. Biophys. Biophys. Chem. 1991,
20, 519.
[30] a) E. S. Baker, J. Jonas, J. Phys. Chem. 1985, 89, 1730; b) P. P. Wie-
wior, H. Shirota, E. W. Castner, Jr., J. Chem. Phys. 2002, 116, 4643;
c) J. T. Cabral, A. Luzar, J. Teixeira, M. C. Bellissent-Funnel, J.
Chem. Phys. 2000, 113, 8736.
[31] a) I. Demetropoulos, A. Tsibiris, V. Tsikaris, M. Sakarellos-Daitsio-
tis, C. Sakarellos, J. Biomol. Struct. Dyn. 1995, 12, 755; b) V. Tsika-
ris, M. Sakarellos-Daitsiotis, N. Theophanidis, C. Sakarellos, M. T.
Cung, M. Marraud, J. Chem. Soc. Perkin Trans. 2 1991, 1353.
[32] H. E. Stanger, F. A. Syud, J. F. Espinosa, Proc. Natl. Acad. Sci. USA
2001, 98, 12015.
[33] It should be noted that extensive aggregation of the trans isomer of
the diazobenzene-based system deprived the authors of its spectro-
scopic conformational analysis and so their conclusions are mainly
based on computational simulations.
[34] a) A. G. Blanco, M. Sola, F. X. Gornis-Rüth, M. Coll, Structure 2002,
10, 701; b) W. S. Somers, S. E. V. Phillips, Nature 1992, 359, 387;
c) J. D. Puglishi, L. Chen, S. Blanchard, A. D. Frankel, Science 1995,
270, 1200; d) G. Variani, Acc. Chem. Res. 1997, 30, 189; e) N. C. J.
Strynadka, S. E. Jensen, P. M. Alzari, M. N. G. James, Nat. Struct.
Biol. 1996, 3, 290.
[35] a) C. P. Jaroniec, C. E. MacPhee, V. S. Bajaj, M. T. McMahon, C. M.
Dobson, R. G. Griffin, Proc. Natl. Acad. Sci. USA 2004, 101, 711;
b) J. Collinge, Annu. Rev. Neurosci. 2001, 24, 519; c) J. M. McDon-
nell, D. Fushman, S. M. Cahill, B. J. Sutton, D. Cowburn, J. Am.
Chem. Soc. 1997, 119, 5321; d) J. P. Derrick, G. J. Davies, Z. Dauter,
K. S. Wilson, D. B. Wigley, J. Mol. Biol. 1992, 227, 1253; e) P. D.
Kwong, R. Wyatt, J. Robinson, R. W. Sweet, J. Sodroski, W. A. Hen-
drickson, Nature 1998, 393, 648.
[36] V. K. Sarin, S. B. Kent, J. P. Tam, P. B. Merrifield, Anal. Biochem.
1981, 117, 147.
[37] a) A. Wokaun, R. R. Ernst, Chem. Phys. Lett. 1977, 52, 407; b) A. J.
Shaka, R. Freeman, J. Magn. Reson. 1983, 51, 169.
[17] Observation of amide temperature coefficients below 3 ppbKÀ1 was
interpreted as an indication of the involvement of a particular
amide proton in intramolecular hydrogen bonding. For details see
ref. [13a].
[18] a) D. S. Wishart, B. D. Sykes, F. M. Richards, J. Mol. Biol. 1991, 222,
311; b) M. Ramirez-Alvarado, F. J. Blanco, L. Serrano, Nat. Struct.
Biol. 1996, 3, 604.
[19] a) K. Wüthrich, G. Wider, G. Wagner, W. Braun, J. Mol. Biol. 1982,
155, 311; b) T. F. Havel, K. Wüthrich, J. Mol. Biol. 1985, 182, 281;
c) T. F. Havel, K. Wüthrich, Bull. Math. Biol. 1984, 46, 673; d) M. P.
Williamson, T. F. Havel, K. Wüthrich, J. Mol. Biol. 1985, 182, 295;
e) A. D. Gossert, S. Bonjour, D. A. Lysek, F. Fiorito, K. Wüthrich,
Proc. Natl. Acad. Sci. USA 2005, 102, 646; f) D. A. Lysek, C. Schorn,
L. G. Nivon, V. Esteve-Moya, B. Christen, L. Calzolai, L. C. von
Schroetter, F. Fiorito, T. Herrmann, P. Güntert, K. Wüthrich, Proc.
Natl. Acad. Sci. USA 2005, 102, 640; g) S. Hiller, A. Kohl, F. Fiorito,
T. Herrmann, G. Wider, J. Tschopp, M. Grütter, K. Wüthrich, Struc-
ture 2003, 11, 1198; h) C. Fernµndez, C. Hilty, G. Wider, P. Güntert,
K. Wüthrich, J. Mol. Biol. 2004, 336, 1211; i) R. Horst, F. Damberg-
er, P. Luginbühl, P. Güntert, G. Peng, L. Nikonova, W. S. Leal, K.
Wüthrich, Proc. Natl. Acad. Sci. USA 2001, 98, 14374.
[20] A. Fissi, J. L. Houben, N. Rosato, S. Lopes, O. Pieroni, Makromol.
Chem. Rapid Commun. 1982, 3, 2 9.
[21] a) W. S. Price, Annu. Rep. Prog. Chem. Sect. C 2000, 96, 3; b) W. S.
Price, Concepts Magn. Reson. 1997, 9, 299.
[22] a) R. E. Moll, J. Am. Chem. Soc. 1968, 90, 4739; b) J. A. Jones, D. K.
Wilkins, L. J. Smith, C. M. Dobson, J. Biomol. NMR 1997, 10, 199.
[23] a) S. J. Gibbs, C. S. Johnson, J. Magn. Reson. 1991, 93, 395; b) G. A.
Morris, H. Barjat in Methods for Structure Elucidation by High Res-
olution NMR (Eds.: Gy. Batta, K. E. KçvØr, Cs. Szµntay), Elsevier,
1997, pp. 209.
[38] a) A. Kumar, R. R. Ernst, K. Wüthrich, Biochem. Biophys. Res.
Commun. 1980, 95, 1; b) G. Bodenhausen, H. Kogler, R. R. Ernst, J.
Magn. Reson. 1984, 58, 370.
[39] A. Bax, D. G. Davis, J. Magn. Reson. 1985, 58, 370.
[40] S. H. Smallcombe, S. L. Patt, P. A. Keifer, J. Magn. Reson. 1995, 117,
295.
[24] We would like to emphasize here that these measurements were per-
formed in order to evaluate the applicability of diffusion measure-
ments for use in peptide conformational studies by measuring the
diffusion coefficients of compounds of already established confor-
3
mation (NOE, DdNH/DT, JNH,Ha, d, CD).
[25] J. Danielsson, J. Jarvet, P. Damberg, A. Gräslund, Magn. Reson.
Chem. 2002, 40, S89.
[26] J. G. de la Torre, M. L. Huertas, B. Carrasco, Biophys. J. 2000, 78,
719.
[27] S. Yao, G. J. Howlett, R. S. Norton, J. Biomol. NMR 2000, 16, 109.
[28] K. J. Packer, D. J. Tomlinson, Trans. Faraday Soc. 1971, 67, 1302.
[29] a) C. Rocchi, A. R. Bizzarri, S. Cannistraro, Phys. Rev. E 1998, 57,
3315; b) Y. Komeiji, M. Uebayasi, J. Someya, I. Yamato, Proteins
Struct. Funct. Genet. 1993, 16, 268; c) V. Lounnas, B. M. Pettitt,
G. N. Phillips, Biophys. J. 1994, 66, 601; d) V. A. Makarov, B. K. An-
[41] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J.
Comput. Chem. 1990, 11, 440.
[42] W. C. Still, A. Tempczyk, R. C. Hawley, T. Hendrickson, J. Am.
Chem. Soc. 1990, 112, 6127.
[43] W. C. Goodman, W. C. Still, J. Comput. Chem. 1991, 12, 1110.
Received: March 20, 2005
Published online: September 27, 2005
412
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 403 – 412