Direct structural comparison of a rigid cyclic peptidic scaffold using crystallography and NMR in strained PH polymer gels

Article abstract of DOI:10.1002/ejoc.201000440

A small series of biaryl ether containing cyclic peptidic scaffolds was synthesized and cyclized by an SNAr reaction. The structure of one rigid scaffold was solved, by X-ray crystallography and also determined in solution by NMR spectroscopy. Molecular a

Full text of DOI:10.1002/ejoc.201000440

FULL PAPER
DOI: 10.1002/ejoc.201000440
Direct Structural Comparison of a Rigid Cyclic Peptidic Scaffold Using
Crystallography and NMR in Strained PH Polymer Gels
Christopher J. Arnusch,[a][‡] Johannes H. Ippel,[a][‡] Huub Kooijman,[b] Anthony L. Spek,[b]
Rob M. J. Liskamp,[a] Johan Kemmink,*[a] and Roland J. Pieters*[a]
Keywords: NMR spectroscopy / Structure elucidation / Peptides / Amino acids
A small series of biaryl ether containing cyclic peptidic scaf-
folds was synthesized and cyclized by an SNAr reaction. The
structure of one rigid scaffold was solved by X-ray crystal-
lography and also determined in solution by NMR spec-
troscopy. Molecular alignment of the peptidic scaffold in
strained PH polymer gels in [D6]DMSO was applied to ex-
tract residual dipolar couplings (RDCs). The RDC values
were used to obtain a structure that was compared to the
crystal structure. Good correlation was obtained, indicating
that the RDC method represents a very precise structure de-
termination method for small organic molecules in solution.
macrolide archazolide, RDCs were used in combination
with NOEs and J couplings to determine the conformation
and also the configuration of its 7 stereocenters.[7] In an-
other example, RDCs allowed the distinction to be made
between two NMR-derived structures of hormaomycin, a
20-membered ring depsipeptide.[8] Here we describe the de-
tailed molecular structure of a peptidic 14-membered biaryl
ether construct (2) determined by using RDCs. The deter-
mination was performed in DMSO for which strained poly-
mer gels were recently reported.[9] The used gel was a copol-
ymeric crosslinked polyacrylamide gel (PH gel). To gain
specific information about the stereochemistry and to have
more precise details on the relative orientation of the two
Introduction
Residual dipolar couplings are a valuable tool in NMR
spectroscopy for structure determination of biological mac-
romolecules.[1] The use of media that induce partial align-
ment of the molecules reduces the dipolar couplings down
to manageable magnitudes of a few Hertz. These so-called
residual dipolar couplings (RDCs) are obtained from spec-
tra relatively easily. More recently, the method has been ex-
tended for the use of small organic molecules.[2] This too
was due to the use of novel media that provide a suitable
degree of alignment for the small molecules.[3] The method
is most successfully used for rigid compounds, but applica-
tions for flexible compounds are also being developed. Use-
ful applications for rigid molecules were found in determin-
ing the relative stereochemistry in compounds such as
strychnine,[4] menthol,[5] and others.[2b] The method is par-
ticularly useful in cases where conventional NMR tech-
niques fail. Besides this, RDCs can in principle also be used
for the determination of the 3D structure of small organic
molecules and at least they can improve the precision of
solution NMR structures. The RDCs determined for cy-
closporin A allowed structural refinement of an NOE-de-
termined structure.[6] Furthermore, for the 24-membered
3
rings, “classic” (ROE/NOE and J couplings) NMR con-
straints are of limited use. Values of distances based on
ROE/NOE intensities cannot be measured precisely enough
in small molecules to get accurate local angular information
on, for example, the two aryl rings. The RDC-derived struc-
ture was compared to the structural information obtained
by crystal structure analysis and proved to be in excellent
agreement, demonstrating the value and utility of the
method. Peptidic compound 2 is part of a study involving
conformationally restricted biaryl ether containing peptides
inspired by natural compounds such as the antibiotic van-
comycin[10] and its structural relative complestatin, which
has multiple activities including anti-HIV activity.[11]
A
[a] Department of Medicinal Chemistry and Chemical Biology,
Utrecht Institute for Pharmaceutical Sciences, Utrecht University,
P. O. Box 80082, 3508 TB Utrecht, The Netherlands
Fax: +31-30-2536655
large spectrum of biological activities is displayed by com-
pounds of the class of cyclic peptides containing biaryl
ether linkages.[12] The strained 14-membered cyclophane
ring system of cycloisodityrosine[13] is found in several natu-
ral products including the antifungal antibiotic piperazino-
mycin,[14] the series of antitumor antibiotics including bou-
vardin and RA I–RA XVIII,[12,15] and the metallopeptidase
inhibitors K-13[16,17] and OF4949 I-IV[17,18] (see Supporting
E-mail: R.J.Pieters@uu.nl
J. Kemmink@uu.nl
[b] Crystal Structural Chemistry, Utrecht University
Padualaan 8, 3584 CH Utrecht, The Netherlands
[‡] Both authors contributed equally to this work.
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201000440.
Eur. J. Org. Chem. 2010, 4501–4507
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4501

J. Kemmink, R. J. Pieters et al.
FULL PAPER
Information). Furthermore, a related ring structure is pres- by cryocrystallography at 150 K. The asymmetric unit of
ent in the antitumor macrocyclic lactones combretasta- the crystal contained two molecules and one methanol sol-
tin DI and D2.[19] An SNAr cyclization was used to synthe- vent molecule (Figure 1). Ring strain in the small cyclic mo-
size the biaryl ether linkage.[20]
lecule leads to the deformation of the six-membered rings
C13–C18 towards a boat conformation. The average abso-
lute endocyclic torsion angle is 4.8 and 5.4° for structures A
and B, respectively. The substituent atoms O19 and C12 are
located approximately 0.4 Å from the least-squares planes
through the six-membered rings. The orientation of the bi-
aryl rings is similar to that found in the crystal structure of
piperazinomycin.[14]
Results and Discussion
Synthesis
The solution-phase synthesis of the cyclic peptides began
with tyrosine, which was protected as a methyl ester on the
C-terminus (Scheme 1). Three derivatives were made by
coupling this compound to Boc-protected alanine, phenyl-
alanine, and valine with BOP and iPr2NEt in CH2Cl2. Re-
moval of the protecting groups was achieved by TFA/
CH2Cl2 (1:1) treatment. The N-terminus of the dipeptide
was first linked to 1,5-difluoro-2,4-dinitrobenzene by using
NEt3 as a base to give compounds 4–6. These were isolated
in overall yields of 65-76% over four steps after column
chromatography. 1,5-Difluoro-2,4-dinitrobenzene is highly
reactive in nucleophilic aromatic substitutions because of
the nitro functionalities in the ortho and para positions to
the carbon atoms bearing a fluorine. Upon substitution
with the first nucleophile, the second point of substitution
was observed to be slightly less electrophilic, as compounds
4–6 were stable yellow solids and could be stored at room
temperature for months without degradation. Cyclization
was achieved by using K2CO3 (4 equiv.) in DMF, which
gave highly constrained compounds 1–3 in isolated yields
of 46–64%. These compounds proved to be very stable in
air at room temperature.
Scheme 1. Synthesis of compounds 1–3. Reagents and conditions:
(a) Boc-Xaa-OH, BOP, iPr2NEt, CH2Cl2; (b) TFA/CH2Cl2, 1:1;
(c) 1,5-difluoro-2,4-dinitrobenzene, NEt3, 30 min; (d) K2CO3,
DMF, 3 d, r.t.
Figure 1. (a) Content of the asymmetric unit of the crystal structure
of 2. Displacement ellipsoids are drawn at the 50% probability
level. (b) The two crystallographically independent main molecules
have slightly different conformations as shown in a fit-plot.
Crystal Structure of 2
Crystals for compound 2 were retrieved from a concen-
trated solution of methanol, and the structure was solved (c) Numbering scheme of 2.
4502
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4501–4507

Direct Structural Comparison of a Rigid Cyclic Peptidic Scaffold
Solution Structure of 2 Determined by NMR Spectroscopy
rium NMR spectrum and amounts to 2.5 Hz and less than
1.6 Hz (within the natural line width) for the 12 and 8% gel
sample, respectively. To measure the RDCs, t2-coupled 13C–
1H HSQC spectra and t2-coupled 15N–1H HSQC spectra
were recorded in the presence (anisotropic solution) and ab-
sence (isotropic solution) of the orienting gel. The scalar
1JCH and 1JNH and total couplings 1JCH + 1DCH (and 1JNH
Compound 2 was extensively studied by NMR spec-
troscopy. [D6]DMSO was chosen as the solvent for reasons
of solubility, but also to retain a stable solution when dis-
solved in strained polymer gels. Complete proton and car-
bon resonance assignments in the isotropic phase were ob-
tained by a combination of standard ROESY, TOCSY, 13C–
1H HSQC, and 13C–1H HMBC 2D spectra. In addition,
nitrogen amide resonances could be assigned from a natural
abundance 15N–1H HSQC 2D spectrum (see Supporting In-
formation). Analysis of the ROE patterns in the ROESY
spectrum of 2 showed that the structure in solution is con-
sistent with that in the crystal. Short distances are observed
between H–N10 and H17–H18, but not between H–N10
and H14–H15 (see Figure 1c for the numbering scheme).
These distances are consistent with a rigid orientation of
the ring C13–C18, in which H17 and H18 are positioned
over the amide backbone at position N10, whereas on the
back of the molecule H11 is placed close to H14, but not
H18. Moreover, two ROEs with equal intensity are found
between either H4 and H15 or H4 and H17, which indicates
that the two aryl rings are more or less placed perpendicular
to each other. This perpendicular orientation also explains
the unusual shifted resonance of the aromatic H4 proton
signal at 4.109 ppm. The H4 proton is considerably shifted
upfield, because of its position in the strongly shielded ring
current field of the opposite aryl ring C13–C18. Stereospec-
ific assignment of the methylene H12 and H12Ј protons fol-
lows from the large vicinal coupling constant of 12.3 Hz
observed between H11 and H12Ј (and between H11–H12,
J = 5.3 Hz) and the large ROE between H11 and H12 (with
ROE H11–H12Ј absent). Such a combination is only pos-
sible with a trans configuration between protons H11 and
H12Ј.
1
+ DNH) were measured individually after extraction of 1D
F2 traces from the HSQC spectra. The difference between
1
the two coupling constants then results in the RDCs DCH
1
(or DNH). An overlay of the t2-coupled 13C–1H HSQC
spectrum of 2 in the isotropic solution and the anisotropic
solution (8% PH gel) is shown in Figure S4 (Supporting
Information). The carbon-bound proton signals of the PH
gel matrix show up as residual broad doublet signals in the
spectrum in the 8% PH gel solution. Interference of the
broad polymer signals (NMe and CH2 from DMAA, and
NH and CH2 from AMPS; for the structures, see Figure S3,
Supporting Information) is also visualized by the corre-
sponding 1D proton spectrum inside the 2D HSQC spec-
trum. The large polymer signals appear rather high in the
1D spectrum but are strongly reduced in intensity due to
efficient spin relaxation when measured by 2D correlation
spectra. Figure 2 shows slices through the H18 carbon reso-
nance of the t2-coupled 13C–1H HSQC spectra of the iso-
tropic and two anisotropic solutions, with their respective
RDC values obtained after curve-fitting the line shape of
the 13C-coupled proton resonance signals. As shown in Fig-
ure 2, increased alignment was associated with higher an-
isotropy; hence, larger line widths were present along the
proton dimension. Table 1 summarizes the final determined
NMR in Strained PH Gels
Even though the NMR structure is consistent with the
crystal structure, it cannot provide the same level of struc-
tural detail. The RDCs have the potential to greatly im-
prove the level of detail. Partial alignment of 2 in strained
polymer gels was used to extract the RDCs. The alignment
was achieved by use of the copolymeric cross-linked poly-
acrylamide (PH) gel as the preferred alignment medium
that is compatible with DMSO.[9] The amount of alignment
can be tuned by reswelling the dried polymerized gel in
smaller diameter tubes and/or by increasing the percentage
of monomer concentration in the copolymerization step. In
this study, 8 and 12% PH gels were made inside 5 mm
NMR tubes and equilibrated for 7 and 14 d, respectively,
with a solution of 2 in [D6]DMSO. The diffusion process
into the clear transparent gel was easily monitored because
of the dark yellow colored solution of 2. After sufficient
time, duplicate measurements of 2 in the 8% PH gel showed
that a stable solution was maintained in the gel over time.
The degree of alignment in the media can be measured from
Figure 2. Cross-section taken at the ω1 frequency of carbon C18 of
the isotropic t2-coupled 1H-13C HSQC spectrum aligned with cross-
sections of the spectra recorded in the 8 and 12% PH gels. The
1
observed JHC
+
1DHC coupling constants of each spectrum are
1
indicated together with the dipolar DHC coupling constants in the
8 and 12% PH gel, respectively. Peaks indicated with an asterisk
the quadrupolar splitting of the solvent line in the deute- originate from the H14–C14 correlation.
Eur. J. Org. Chem. 2010, 4501–4507
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4503

J. Kemmink, R. J. Pieters et al.
FULL PAPER
RDC values of all protons in 2. In both alignment media, just above 1.02 Hz. At this point, it is good to know that the
1
relative weak alignment was observed with values of DCH PALES fit relies on correct covalent atomic bond lengths to
1
1
(or DNH) that are always substantially smaller than JCH calculate the alignment tensor and scale the relative size of
1
(or JNH).
1
Table 1. Isotropic one-bond 15N–1H coupling constants JNH and
1
13C–1H coupling constants JCH (Hz) of 2 in [D6]DMSO, together
with corresponding RDC (defined as 1Dij – 1JCH) (Hz) determined
in duplo from two different aligned PH media, all recorded at
27 °C.
Proton pair
1JNH/1JCH
RDC (8%PH gel) RDC (12%PH gel)
[Hz]
[Hz]
[Hz]
H–N7
H–N10
H1–C1
H4–C4
94.01
92.01
–1.95
3.61
–1.17
–1.32
1.31
9.66
1.38
9.53
2.42
–10.21
(+29)[a,b]
–10.34
–11.30
13.23
38.57
13.92
43.86
11.84
169.72
163.48
161.11
164.47
164.89
163.25
140.91
141.49
132.93
131.42
130.88
147.74
126.04
126.17
H14–C14
H15–C15
H17–C17
H18–C18
H8–C8
H11–C11
H12–C12
H12Ј–C12
H21–C21
Me20–C20
Me22–C22
Me23–C23
10.01
–1.65
(+14)[a]
1.92
43.58
–4.27
(N.D.)[a,c]
1.31
0.75[d]
–0.16[d]
0.06[d]
5.00[d]
–1.08[d]
–0.09[d]
[a] Partial overlap with signals of the polymer PH matrix. The absolute
size and sign of the RDC is in agreement with predicted values, but was
not used in the RDC fitting procedure. [b] No accurate value of RDC due
to low signal-to-noise ratio. Estimated error limit is 29Ϯ10 Hz. [c] Not
determined, too broad signal overlapping with resonances of the polymer
PH matrix. [d] Time-averaged RDC values of methyl protons, not used in
the RDC fitting procedure.
PALES Alignment Fit Procedure
The single-value decomposition module (SVD)[21] of the
program PALES[22] was used to fit 12 (in case of the 8%
gel) or 11 (in case of the 12% gel) experimental dipolar
couplings to the reference crystal structures A and B (Fig-
ure 1). Due to partial overlap of H12Ј with residual signals
of the polymer matrix, the RDC of H12Ј was not included
in the fit. The predicted RDC values do show the correct
sign and amplitude as estimated from the experiment.
1
Moreover, the imprecise value of DNH (H-N10), obtained
from the spectrum in the 12% PH gel solution, was ex-
cluded from the fit.
Typical fits of the experimental RDCs, derived from both
the 8 and 12% PH gel NMR spectroscopic data versus pre-
dicted RDCs calculated from different types of reference
structures are displayed in Figure 3. Correlation factors and
statistics for these fits are given in the Supporting Infor-
mation. In the first panel (Figure 3a), the 8% PH gel corre-
lation plot is shown with respect to the original atomic co-
ordinates of the reference crystal structures: structure B.
The correlation factor R is 0.970 with a root mean squares
Figure 3. Correlation plots between the calculated (vertical axis)
and the observed RDCs in case of the 8% PH gel (horizontal axis).
(a) The calculated RDCs derived from X-ray structure B (R =
0.970) and (b) from the SA energy-minimized computer model (R
= 0.998); (c) same as (b) but with an intentional erroneous switch
between the assignments of H12 and H12Ј to illustrate the sensitiv-
deviation (RMSD) of the experimental RDC data points ity of the method (R = 0.739).
4504
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4501–4507

Direct Structural Comparison of a Rigid Cyclic Peptidic Scaffold
the dipolar couplings. Normally, the electron density at the rigid compound 2 is remarkably precise, as it is possible
center of the hydrogen atoms in crystallography is too low to distinguish between very similar structures of 2. Small
to accurately determine hydrogen bond lengths. Not sur- imperfections in the match between calculated and ob-
prisingly, a significant improvement in the PALES fit (R served RDCs readily show up as a lower correlation, that
= 0.992 and RMSD = 0.575 Hz) is obtained after energy- is, a lower R value. Therefore, RDC-based structure deter-
minimization (Amber99 force field) of the hydrogen atom mination is a good alternative to X-ray crystallography for
positions (with all heavy atoms fixed). Comparison to struc- conformationally homogeneous cyclic peptides.
ture A yielded almost the same fits. In the next step, the
newly hydrogen-built reference crystal structures of 2 were
subjected to a low temperature (300 K) simulated annealing
(SA) step in a pre-equilibrated box of DMSO molecules,
again by applying a sophisticated force field calculation.
Improvements in the fits are observed, but mainly for the
structure starting from structure B of the crystal structure.
In this case, the correlation factor increased to 0.998,
whereas at the same time the RMSD decreased to 0.24 Hz,
leading to a perfect fit that falls well within the experimen-
tal error limits (Figure 3b).
Experimental Section
General Remarks: Chemicals were obtained from commercial
sources and used without further purification unless stated other-
wise. DMF, CH2Cl2, and CH3OH were purchased from Biosolve,
the Netherlands. CH2Cl2 and DMF were stored over 4 Å molecular
sieves, and CH3OH was stored over 3 Å molecular sieves. iPr2NEt
was obtained from Acros Organics, and HPLC-grade TFA was ob-
tained from Merck. The bases iPr2NEt and Et3N were distilled
from ninhydrin and KOH. The coupling reagent benzotriazol-1-
yloxy-tris(dimethylamino)phosphonium
hexafluorophosphate
A similarly good correlation (R = 0.997 and RMSD =
1.67 Hz) was obtained between the RDC values that are
independently determined from the 12% PH gel data and
the RDC values calculated from the computer energy-mini-
mized structure B. The energy-minimized crystal struc-
ture A led to a slightly less optimal fit. To illustrate how
critical the PALES SVD procedure works to get a unique
solution for 2 based on experimental RDC values alone is
demonstrated in Figure 3c. This plot shows that the good
correlation between experimental and predicted RDCs
drops dramatically when RDCs are not correctly assigned,
such as in the case of the reversal of H12 and H12Ј assign-
ments and the 8% PH gel data.
(BOP)[23] was obtained from Richelieu Biotechnologies Inc. (Mon-
treal, Canada). Column chromatography was performed on ICN
Silica 32-63, 60 Å. Thin-layer chromatography (TLC) was per-
formed on Merck precoated Silica 60 plates. Visualization was ac-
complished with UV light and staining with ninhydrin. Amino ac-
ids were purchased from Multisyntech. Electrospray ionization
(ESI) mass spectrometry was carried out using electrospray MS
(nano ESI-TOF-MS) run with a Micromass LC–TOF mass spec-
trometer by spraying a 90% MeOH solution of 1 from a gold-
coated glass capillary in a Z-spray nanospray ionization source or
by using a Shimadzu LCMS QP-8000 single quadrupole bench top
mass spectrometer (m/z range Ͻ 2000), coupled with a QP-8000
data system.
1
General NMR Spectroscopy: H NMR spectra were recorded with
a Varian G-300 spectrometer (300 MHz) or a Varian INOVA spec-
trometer (500 MHz). 13C NMR spectra were recorded with a Var-
ian G-300 spectrometer at 75.4 MHz. All spectra were referenced
to the solvent signal of CD3OD(H) [δ(1H) = 3.3 ppm, δ(13C) =
Conclusions
The conclusion is that one particular energy-minimized
conformer derived from one of the two crystal structures
(structure B) agrees best with the experimentally deter-
mined RDC data in solution. Figure 1b shows an overlay
of the two crystal substructures, which reveals a small dif-
ference in the relative orientation of the biaryl rings and the
two NO2 groups (RMSD heavy atom 0.26 Å). The com-
parison between the structure that best fits the RDC values,
that is, the energy minimized computer model derived from
structure B, and structure B itself is shown in Figure 4.
Here, the structural differences are even less (RMSD heavy
atom 0.22 Å). The RDC-based structure determination for
49.0 ppm] or [D6]DMSO [δ(1H)
= =
2.49–2.50 ppm, δ(13C)
39.5 ppm]. Complete assignment of proton, carbon, and amide ni-
trogen atoms for 1 in [D6]DMSO were made by using a combina-
tion of 2D TOCSY (mixing times 20 and 60 ms), 2D ROESY (mix-
ing times 150, 300, and 500 ms), 2D 1H–13C HSQC, 2D 1H–13C
1
HMBC (60 ms filter delay), and 2D H–15N HSQC spectra. Scalar
coupling constants were measured from a high-resolution 1D
NMR spectrum. All spectra were processed and analyzed with
MestRe-C.
Strained Gel NMR Spectroscopy: After recording the reference
NMR spectra of 2 in the isotropic phase, two different strained PH
gel samples were prepared. The PH gel was made by copolymeriza-
tion of 2-(acryloamido)-2-methylpropanesulfonic acid (AMPS)
monomer with N,N-dimethylacrylamide (DMAA) as comonomer
in the presence of N,NЈ-methylenebisacrylamide (BIS) as
crosslinker and ammonium persulfate (APS) as radical initiator[24,9]
(see also Supporting Information for structural formulas). The pro-
cedure for making the PH gel and letting the dried gel reswell in
[D6]DMSO inside a 5 mm NMR tube is described in the litera-
ture.[9] In this case, the isolated PH gel samples were first measured
to check that no residual NMR peaks from unreacted monomers
are present in the PH polymer gel after the washing steps. Two
different strengths of strained gel samples were used: 8% (w/v) and
12% (w/v) made by addition of a concentrated stock solution of 2
Figure 4. Overlap of the X-ray structure of 2 (structure B, black)
with the optimal RDC-derived structure (gray).
Eur. J. Org. Chem. 2010, 4501–4507
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4505

J. Kemmink, R. J. Pieters et al.
FULL PAPER
in [D6]DMSO on top of the preswollen gel. Diffusion of the yellow-
ish colored compound into the gel was achieved in 7 and 14 d for
the 8 and 12% gel sample, respectively, a process that could be
optically followed in time to ensure a homogeneous solution for
that part of the sample that is positioned at the height of the NMR
receiver coil.
equipped with a stir bar. CH2Cl2 (15 mL) and iPr2NEt (0.538 mL,
3.1 mmol) were added, and the solution was stirred at room tem-
perature for 3 h. CH2Cl2 was removed in vacuo, and EtOAc was
added (50 mL). The organic solution was washed with 1  KHSO4,
5% NaHCO3, and brine (each 2ϫ50 mL) and then dried with
Na2SO4. The resulting products were tested with TLC [Rf = 0.21
(Ala, Val), Rf = 0.24 (Phe); CH2Cl2/MeOH, 19:1] and were used
without further purification. Each product was dissolved in CH2Cl2
(10 mL) and to it was added TFA (10 mL) at 0 °C. The solutions
were stirred for 4 h at room temperature. After evaporation and
multiple coevaporation with EtOH and CH2Cl2 in vacuo, the crude
products were dissolved in EtOAc (20 mL). To each was added
Et3N (0.43 mL, 3.1 mmol) and a solution of 1,5-difluoro-2,4-dini-
trobenzene (210 mg, 1.03 mmol) in EtOAc (20 mL). The pH was
tested and observed as basic. After 45 min the solvents were re-
moved in vacuo and 4–6 were purified by column chromatography
(EtOAc/hexane, 1.5:1). Data for 4: 321 mg (69%). Rf = 0.67
(EtOAc). 1H NMR (CD3OD): δ = 1.49 (d, J = 7 Hz, 3 H), 2.89
(dd, J = 9, 14 Hz, 1 H), 3.11 (dd, J = 6, 14 Hz, 1 H), 3.69 (s, 3 H)
4.27 (q, J = 7 Hz, 1 H), 4.66 (q, J = 5 Hz, 1 H), 6.62 (m, 2 H, 1
H), 6.97 (d, J = 8 Hz, 2 H), 9.04 (d, J = 8 Hz, 1 H) ppm. MS: m/z
= 431 [M + H]+. Data for 5: 375 mg (76%). Rf = 0.72 (EtOAc). 1H
NMR (CD3OD): δ = 0.98 (d, J = 6 Hz, 6 H), 2.19 (m, 1 H), 2.84
(dd, J = 10, 14 Hz, 1 H), 3.14 (dd, J = 5, 14 Hz, 1 H), 3.69 (s, 3
H), 3.97 (d, J = 5 Hz, 1 H), 4.75 (m, 1 H), 6.56 (d, J = 9 Hz, 2 H),
6.65 (d, J = 14 Hz, 1 H), 6.94 (d, J = 9 Hz, 2 H), 9.05 (d, J = 8 Hz,
1 H) ppm. Data for 6: 350 mg (65%). Rf = 0.77 (EtOAc). 1H NMR
(CD3OD): δ = 2.86 (dd, J = 9, 14 Hz, 1 H), 3.03-3.11 (m, 2 H),
3.23 (dd, J = 5, 14 Hz, 1 H), 3.69 (s, 3 H), 4.46 (dd, J = 5, 8 Hz, 1
H), 4.70 (dd, J = 5, 9 Hz, 1 H), 6.47 (d, J = 14 Hz, 1 H), 6.59 (d,
J = 8 Hz, 2 H), 6.94 (d, J = 8 Hz, 2 H), 7.20-7.27 (m, 5 H), 8.97
(d, J = 8 Hz, 1 H) ppm.
Proton t2-13C-coupled HSQC spectra were acquired at natural
abundance as 2048ϫ512 data point matrices (2048ϫ86 matrix for
t2-15N-coupled HSQC spectra), processed by sine bell square multi-
plication and zero-filled in both dimensions to a final digital resolu-
tion of 0.73 Hz/pt in the proton acquisition dimension, and to a
resolution of 12.2 and 3.9 Hz/pt in the heteronuclear 13C and 15N
dimension, respectively. Extraction of scalar (1JC/N–H) and/or the
1
sum of scalar and residual dipolar couplings (1JC/N–H + DC/N–H
)
were done by taking 1D slices of proton cross peaks at the carbon
or nitrogen frequency of interest and compare them. Precise values
of the RDC were determined by curve-fitting the relative position
of proton resonance lines from the multiplet patterns in both the
isotropic reference and the PH gel spectra. Error limits are esti-
mated (based on an independent duplo measurement of the 8%
PH gel sample and based on different ways of processing the NMR
spectroscopic data) to Ϯ0.2 Hz for RDCs, belonging to intense aro-
matic and methyl protons, and to Ϯ0.3 Hz for RDCs of other sig-
nals, unless stated otherwise.
Pales Calculations: The single-value decomposition module
(SVD)[21] of the program PALES[22] was used to analyze the experi-
mental determined RDC values. Tensor alignment was done by
simultaneous fitting of the Saupe alignment matrix and the degree
of alignment (Da, Dr), a situation well applicable due to the high
number (11 to 12) of available experimental RDC restraints with
respect to the degrees of fit variables. The four Pales parameters:
correlation factor r, RMSD, Bax slope, and Chi square, were evalu-
ated together in the judgment of best fit.
Synthesis of 1: To a round-bottomed flask equipped with a stir bar
was added 4 (39 mg, 0.09 mmol) in DMF (50 mL). K2CO3 (43 mg,
0.31 mmol) was added, and the yellow solution became dark red.
This solution was stirred protected from light for 5 d. AcOH was
added (≈1 mL), and the color changed from a dark red to a light
orange. The solvents were removed in vacuo. Product 1 was puri-
fied by column chromatography (EtOAc/toluene, 1:1) to yield
17.4 mg (47%) of a yellow solid. Rf = 0.47 (EtOAc). 1H NMR
(CDCl3): δ = 1.57 (d, J = 8 Hz, 3 H), 2.70 (t, J = 13 Hz, 1 H), 3.47
(dd, J = 6 Hz, 1 H), 3.70 (dd, J = 6, 13.5 Hz, 1 H)), 3.84 (s, 3 H)
4.14 (s, 1 H), 4.98 (m, 1 H), 5.63 (d, J = 10.5 Hz, 1 H), 6.99 (dd,
J = 2, 9 Hz, 1 H), 7.24 (m, 2 H), 7.49 (dd, J = 2, 9 Hz, 1 H), 7.93
(d, J = 3.5 Hz, 1 H), 9.04 (s, 1 H) ppm. 13C NMR ([D6]DMSO): δ
= 18.9, 36.4, 52.2, 52.6, 54.9, 102.0, 122.9, 123.3, 125.1, 126.6,
127.5, 132.0, 132.9, 137.3, 147.0, 154.2, 160.9, 169.4, 171.4 ppm.
MS: m/z = 431 [M + H]+.
Amber99 Energy-Minimization Calculations: Crystal structures of 2
in mmCIF format were converted to PDB format using Yasara
Structure (www.yasara.org), with both the residue- and atom nam-
ing convention largely restructured to -amino acid topology. An
accurate description of the molecular geometry for 2 has been ob-
tained by applying the Amber99 force field in Yasara energy calcu-
lations, making use of self-parametrized force field parameters and
quantum-mechanically derived AM1 derived atomic charges. En-
ergy-minimization (simulated annealing from 300 K to zero Kelvin)
of the two crystal molecules in the unit cell were carried in two
ways: (i) optimization of protons with all heavy atoms fixed in
vacuo; (ii) optimization of all atoms in explicit DMSO solvent. In
the latter case, the fixed solute was placed into a pre-equilibrated
periodic box of DMSO solvent molecules (density 1.10 g/mL) and
equilibrated for another 500 ps of molecular dynamics under per-
iodic boundary conditions (time step 1 fs, nonbonded cutoff
7.86 Å, Ewald summation of long-range electrostatics, constant
pressure at a density of 1.10 g/mL) before slowly free up the solute
and running a full simulated annealing energy minimization on the
total system. Distance and angle measurements, as well as RMSD
fits were also carried out in Yasara Structure.
Synthesis of 2: To a round-bottomed flask equipped with a stir
bar was added 5 (259 mg, 0.54 mmol) in DMF (275 mL). K2CO3
(300 mg, 2.17 mmol) was added, and the yellow solution became
dark red. This solution was stirred protected from light for 5 d.
AcOH was added (≈10 mL), and the color changed from a dark
red to a light orange. EtOAc (500 mL) and H2O (250 mL) were
added, and the two phases were separated. The EtOAc layer was
washed with H2O (3ϫ400 mL) and brine, and dried with Na2SO4.
The solvent was removed in vacuo, and the residue was purified by
column chromatography (EtOAc/toluene, 1:1) to yield 158 mg
X-ray Crystallography: CCDC-769665 (for 2) contains the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
1
(64%) of a yellow solid. Rf = 0.53 (EtOAc). H NMR (CD3OD):
Synthesis of 4–6: H-Tyr(OH)-OMe (200 mg, 1.03 mmol), Boc-Xaa-
δ = 0.99 (d, J = 7 Hz, 3 H), 1.09 (d, J = 7 Hz, 3 H), 2.27 (m, 1 H),
OH (Boc-Ala-OH 195 mg, 1.03 mmol; Boc-Val-OH 224 mg, 2.82 (t, J = 13 Hz, 1 H), 3.44 (d, J = 4 Hz, 1 H), 3.66 (dd, J = 8,
1.03 mmol; Boc-Phe-OH 273 mg, 1.03 mmol), and BOP (455 mg,
1.03 mmol) were measured into a 100-mL round-bottomed flask
13 Hz, 1 H), 3.79 (s, 3 H), 4.19 (s, 1 H), 4.88 (m, 1 H), 6.93 (dd, J
= 2, 9 Hz, 1 H), 7.30 (dd, J = 2, 9 Hz, 1 H), 7.37 (dd, J = 2, 9 Hz,
4506
www.eurjoc.org
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2010, 4501–4507

Direct Structural Comparison of a Rigid Cyclic Peptidic Scaffold
1 H), 7.50 (dd, J = 2, 9 Hz, 1 H), 8.92 (s, 1 H) ppm. 13C NMR
([D6]DMSO): δ = 16.8, 19.4, 30.8, 36.2, 52.2, 52.6, 61.7, 101.8,
123.0, 123.2, 125.2, 126.7, 127.6, 132.0, 132.8, 137.3, 147.6, 154.3,
161.0, 167.9, 171.3 ppm. MS: m/z = 459 [M + H]+.
[10] G. H. Singh, H. Jayasuriya, D. L. Hazudda, P. Felock, C. F.
Homnick, M. Sardana, A. Patane, Tetrahedron Lett. 1998, 39,
8769–8770.
[11] J. Garfunkle, F. S. Kimball, J. D. Trzupek, S. Takizawa, H. Shi-
mamura, M. Tomishima, D. L. Boger, J. Am. Chem. Soc. 2009,
131, 16036–16038.
Synthesis of 3: To a round-bottomed flask equipped with a stir bar
was added 6 (54 mg, 0.10 mmol) in DMF (50 mL). K2CO3 (43 mg,
0.31 mmol) was added, and the yellow solution became dark red.
This solution was stirred protected from light for 5 d. AcOH was
added (≈1 mL), and the color changed from a dark red to a light
orange. The solvents were removed in vacuo. Compound 3 was
purified by column chromatography (EtOAc/toluene, 1:1) to yield
24 mg (46%) of a yellow solid. Rf = 0.60 (EtOAc). 1H NMR
(CD3OD): δ = 2.86 (t, J = 13 Hz, 1 H), 3.03 (dd, J = 9, 13 Hz, 1
H), 3.24 (dd, J = 4, 14 Hz, 1 H), 3.69 (dd, J = 4, 14 Hz, 1 H), 3.78
(dd, J = 4, 9 Hz, 1 H), 3.84 (s, 3 H), 4.19 (s, 1 H), 4.91 (m, 1 H),
6.93 (dd, J = 2, 9 Hz, 1 H), 7.24 (m, 1 H), 7.31 (m, 5 H), 7.41 (dd,
J = 2, 9 Hz, 1 H), 7.52 (dd, J = 2, 9 Hz, 1 H), 8.87 (s, 1 H) ppm.
13C NMR ([D6]DMSO): δ = 36.5, 37.6, 52.3, 52.7, 58.1, 102.0,
123.1, 123.3, 124.9, 126.7, 127.1, 127.6, 128.6, 129.1, 132.1, 132.7,
135.9, 137.3, 146.9, 154.2, 160.9, 167.9, 171.4 ppm. MS: m/z = 507
[M + H]+.
[12] L. Feliu, M. Planas, Int. J. Pept. Res. Ther. 2005, 53–97.
[13] a) A. Bigot, J. Zhu, Tetrahedron Lett. 1998, 38, 551–554; b)
P. J. Krenitsky, D. L. Boger, Tetrahedron Lett. 2002, 43, 407–
410; c) O. Poupaardin, F. Ferreira, J. P. Genet, C. Greck, Tetra-
hedron Lett. 2001, 42, 1523–1526; d) A. Bigot, M. E. T. H.
Dau, J. Zhu, J. Org. Chem. 1999, 64, 6283–6296.
[14] a) S. Tamai, M. Kaneda, S. Nakamura, J. Antibiot. 1982, 35,
1130–1136; b) S. Nishiyama, K. Nakamura, Y. Suzuki, S. Yam-
amura, Tetrahedron Lett. 1986, 37, 4481–4484; c) S. Ghosh,
A. S. Kumar, G. N. Mehta, R. Sounderarajan, S. Sen, Arkivoc
2009, 7, 72–78.
[15] a) S. D. Jolad, J. J. Hoffmann, S. J. Torrance, R. M. Wiedhopf,
J. R. Cole, S. K. Arora, R. B. Bates, R. L. Gargiulo, G. R.
Kriek, J. Am. Chem. Soc. 1977, 99, 8040; b) J.-E. Lee, Y. Hitot-
suyanagi, I.-H. Kim, T. Hasuda, K. Takeya, Bioorg. Med.
Chem. Lett. 2008, 18, 808–811.
[16] H. Kase, M. Kaneko, K. Yamado, J. Antibiot. 1987, 40, 450–
454.
[17] J. W. Janetka, D. H. Rich, J. Am. Chem. Soc. 1997, 119, 6488–
Supporting Information (see footnote on the first page of this arti-
cle): X-ray structures of current compound and related natural
products, chemical shift and coupling constants of 2, components
of the PH gel, and correlation factors between calculated and ob-
served RDCs.
6495.
[18] a) S. Sano, K. Ikai, H. Kuroda, T. Nakamura, A. Obayashi, Y.
Ezure, H. Enomoto, J. Antibiot. 1986, 39, 1674–1684; b) S.
Sano, K. Ikai, K. Katayama, K. Takesako, T. Nakamura, A.
Obayashi, Y. Ezure, H. Enomoto, J. Antibiot. 1986, 39, 1685–
1696; c) S. Sano, M. Ueno, K. Katayama, T. Nakamura, A.
Obayashi, J. Antibiot. 1986, 39, 1697–1703; d) S. Sano, K. Ikai,
Y. Yoshikawa, T. Nakamura, A. Obayashi, J. Antibiot. 1987,
40, 512–518; e) S. Sano, H. Kuroda, M. Ueno, Y. Yoshikawa,
T. Nakamura, A. Obayashi, J. Antibiot. 1987, 40, 519–525.
[19] G. R. Pettit, P. D. Quistorf, J. A. Fry, D. L. Herald, E. Hamel,
J.-C. Chapuis, J. Nat. Prod. 2009, 72, 876–883.
[20] a) C. J. Arnusch, R. J. Pieters, Tetrahedron Lett. 2004, 45, 4153–
4156; b) C. J. Arnusch, R. J. Pieters, Eur. J. Org. Chem. 2003,
3131–3138.
[21] J. A. Losonczi, M. Andrec, M. W. F. Fischer, J. H. Prestegard,
J. Magn. Reson. 1999, 138, 334–342.
[1] a) N. Tjandra, A. Bax, Science 1997, 278, 1111–1114; b) J. H.
Prestgard, A. I. Kishore, Curr. Opin. Chem. Biol. 2001, 5, 584–
590.
[2] a) R. M. Gschwind, Angew. Chem. Int. Ed. 2005, 44, 4666–
4668; b) C. M. Thiele, Eur. J. Org. Chem. 2008, 5673–5685.
[3] C. M. Thiele, Concepts Magn. Reson. A 2004, 30A, 65–80.
[4] C. M. Thiele, J. Org. Chem. 2004, 69, 7403–7413.
[5] L. Verdier, P. Sakhaii, M. Zweckstetter, C. Griesinger, J. Magn.
Reson. 2003, 163, 353–359.
[6] J. Klages, C. Neubauer, M. Coles, H. Kessler, B. Luy, ChemBi-
oChem 2005, 6, 1672–1678.
[7] C. Farès, J. Hassfeld, D. Menche, T. Carlomagno, Angew.
Chem. Int. Ed. 2008, 47, 3722–3726.
[22] M. Zweckstetter, A. Bax, J. Am. Chem. Soc. 2000, 122, 3791–
3792.
[23] B. Castro, J. R. Dormoy, G. Evin, C. Selve, Tetrahedron Lett.
1975, 16, 1219–1222.
[24] X. Liu, Z. Tong, F. Gao, Polym. Int. 1998, 31, 215–220.
Received: April 1, 2010
[8] U. Reinscheid, J. Farjon, M. Radzom, P. Haberz, A. Zeeck, M.
Blackledge, C. Griesinger, ChemBioChem 2006, 7, 287–296.
[9] P. Haberz, J. Farjon, C. Griesinger, Angew. Chem. Int. Ed. 2005,
44, 427–429.
Published Online: July 7, 2010
Eur. J. Org. Chem. 2010, 4501–4507
© 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
4507