Asymmetric synthesis and conformational analysis by NMR spectroscopy and MD of Aba- and α-MeAba-containing dermorphin analogues

Article abstract of DOI:10.1002/cmdc.201100314

Dermorphin analogues, containing a (S)- and (R)-4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one scaffold (Aba) and the α-methylated analogues as conformationally constrained phenylalanines, were prepared. Asymmetric phase-transfer catalysis was unable to provide the (S)-α-Me-o-cyanophenylalanine precursor for (S)-α-MeAba in acceptable enantiomeric purity. However, by using a Schoellkopf chiral auxiliary, this intermediate was obtained in 88% ee. [(S)-Aba3-Gly4]dermorphin retained μ-opioid affinity but displayed an increased δ-affinity. The corresponding Repimer was considerably less potent. In contrast, the [(R)-α-MeAba3-Gly4]dermorphin isomer was more potent than its Sepimer. Tar-MD simulations of both non-methylated [Aba3-Gly4]dermorphin analogues showed a degree of folding at the C-terminal residues toward the Nterminus of the peptide, without however, adopting a stabilized β-turn conformation. The α-methylated analogues, on the other hand, exhibited a typeI/I' β-turn conformation over the α-MeAba3 and Gly4 residues, which was stabilized by a hydrogen bond involving Tyr5-H_N and D-Ala2-CO.

Full text of DOI:10.1002/cmdc.201100314

DOI: 10.1002/cmdc.201100314
Asymmetric Synthesis and Conformational Analysis by
NMR Spectroscopy and MD of Aba- and a-MeAba-
Containing Dermorphin Analogues
Bart Vandormael,[a] Rien De Wachter,[a] Josꢀ C. Martins,[b] Pieter M. S. Hendrickx,[b]
Attila Keresztes,[c] Steven Ballet,[a] Jayapal R. Mallareddy,[c] Fanni Tꢁth,[c] Gꢀza Tꢁth,[c] and
Dirk Tourwꢀ*[a]
Dermorphin analogues, containing a (S)- and (R)-4-amino-
1,2,4,5-tetrahydro-2-benzazepin-3-one scaffold (Aba) and the
a-methylated analogues as conformationally constrained phe-
nylalanines, were prepared. Asymmetric phase-transfer cataly-
sis was unable to provide the (S)-a-Me-o-cyanophenylalanine
precursor for (S)-a-MeAba in acceptable enantiomeric purity.
However, by using a Schçllkopf chiral auxiliary, this intermedi-
ate was obtained in 88% ee. [(S)-Aba3-Gly4]dermorphin re-
tained m-opioid affinity but displayed an increased d-affinity.
The corresponding R epimer was considerably less potent. In
contrast, the [(R)-a-MeAba3-Gly4]dermorphin isomer was
more potent than its S epimer. Tar-MD simulations of both
non-methylated [Aba3-Gly4]dermorphin analogues showed a
degree of folding at the C-terminal residues toward the N ter-
minus of the peptide, without however, adopting a stabilized
b-turn conformation. The a-methylated analogues, on the
other hand, exhibited a type I/I’ b-turn conformation over the
a-MeAba3 and Gly4 residues, which was stabilized by a hydro-
gen bond involving Tyr5-HN and d-Ala2-CO.
Introduction
Three types of opioid receptors, m, d and k, have been identi-
fied. These receptors are involved in pain perception, regula-
tion of mood, reward, motivation and response to stress.[1]
They can be activated by endogenous opioid peptides or exo-
genously administered opiates, which act as analgesic drugs
for moderate to severe pain. It is well known that the chronic
use of common opioid analgesics results in the development
of analgesic tolerance.[2] Therefore, the search for new strong
analgesics with suppressed adverse side effects and abuse po-
tential is still ongoing. Conformational, topographical and ste-
reoelectronic structural features of the opioid peptides are im-
portant for interaction with the m, d, and k opioid receptors.[3]
The heptapeptide dermorphin, H-Tyr-d-Ala-Phe-Gly-Tyr-Pro-
Ser-NH2 (1) was isolated from the skin of the South American
frog Phyllomedusa Sauvagei.[4] It is one of the most potent and
selective m-opioid receptor agonists among naturally occurring
opioids.[5] The minimal sequence required for opiate-like activi-
ty in vivo was identified as the N-terminal tetrapeptide.[6] It
was shown by NMR analysis and X-ray studies that the optimal
message domain for both m and d opioid receptors is the tri-
peptide Tyr1-d-Ala2-Phe3 in a b-turn conformation. This turn
conformation is induced by the d-chirality of the second resi-
due.[7,8] A general feature of opioid peptides containing the
Tyr 1-d-Xaa2-Phe3 message domain, was the unusually high-
field position (<1 ppm) of the side chain protons of the
b-turn conformation around Xaa2-Phe3 in which the side
chain of the second residue is “sandwiched” between the aro-
matic rings of Tyr1 and Phe3.[9,12–16] The m/d selectivity was
mainly attributed to the conformation and polarity of the
C terminus.[12] In addition, conformational restrictions in the
flexibility of the peptide side chains of Tyr1 and Phe3 in der-
morphin, and short dermorphin analogues, have been shown
to cause shifts in selectivity and affinity.[17–19]
The 4-amino-1,2,4,5-tetrahydro-2-benzazepin-3-one scaffold
(Aba; Figure 1) is able to fix the side chain orientation of the
Phe3 residue into the trans (c1 =1808) or gauche(+) (c1 =608)
staggered conformation.[17,18] The incorporation of the confor-
mationally constrained dipeptide, Aba-Gly, (2) in positions 3
and 4 of dermorphin, resulted in a major increase in d affinity
with substantial loss of selectivity.[18,19] In contrast, the Tic struc-
ture 3 (Figure 1) is able to fix the gauche(+) and gauche(ꢀ)
conformations, but excludes the trans conformation of the Phe
[a] Dr. B. Vandormael, Dr. R. De Wachter, Prof. S. Ballet, Prof. D. Tourwꢀ
Department of Organic Chemistry, Vrije Universiteit Brussel
Pleinlaan 2, 1050 Brussels (Belgium)
E-mail: datourwe@vub.ac.be
[b] Prof. J. C. Martins, Dr. P. M. S. Hendrickx
Organic Chemistry Department, Universiteit Gent
Krijgslaan 281S4, 9000 Ghent (Belgium)
[c] Dr. A. Keresztes, J. R. Mallareddy, Dr. F. Tꢁth, Prof. G. Tꢁth
Biological Research Center
1
second residue in the H NMR spectra. This is observed for der-
morphin,[9] its analogues[7] and in the d-selective heptapeptide
deltorphin.[10,11] Thus, a common preferred solution conforma-
tion of the N-terminal tripeptide was proposed to be a type I/I’
Temesvꢂri Kçrfflt 62, 6701 Szeged (Hungary)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201100314.
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2035

D. Tourwꢀ et al.
MED
Results and Discussion
Synthesis
The synthesis of Boc-(S)- and Boc-(R)-Aba-Gly-OH was per-
formed starting from phthaloyl-protected (S)-Phe-Gly and (R)-
Phe-Gly, respectively, by using a N-acyliminium ion cyclization,
as described earlier.[17,19] Both enantiomers were incorporated
into the dermorphin sequence to give 5 and 6 by standard
solid phase peptide synthesis by using Boc chemistry and 4-
methyl-benzhydrylamine (MBHA) resin as solid support. Race-
mic Boc-a-MeAba-Gly was prepared starting from (R,S)-a-Me-
o-cyanophenylalanine, as described earlier.[21] After incorpora-
tion into the dermorphin sequence, the resulting peptide epi-
mers, 7 and 8, were separated by semipreparative RP-HPLC. In
order to assign the absolute configuration of these [a-
MeAba3]-containing epimers, asymmetric synthesis of a-
MeAba-Gly was performed. Since (R,S)-a-Me-o-cyanophenylala-
nine was conveniently obtained by phase-transfer catalyzed
(PTC) alkylation of N-benzylidene alanine ethyl ester with o-cy-
anobenzyl bromide,[21] and since we obtained o-cyanophenyla-
lanine with high enantiomeric purity using a cinchonidinium-
based chiral phase-transfer catalyst,[20] this method was our
first choice for the asymmetric synthesis of a-MeAba
(Scheme 1).
Figure 1. Structures of the conformationally constrained Phe analogues and
definition of the dihedral angles.
side chain. Since a profound decrease in m- and d affinity and
activity can be observed for the [Tic3]dermorphin analogue,[17]
it has been speculated that the bioactive conformer at the
d opioid receptor binding site is likely to be the trans rotamer
of Phe.[17]
It was shown by both molecular modeling and NMR analysis
that Aba3 adopts the trans conformation (c1 =1808) in [Aba3-
Gly4]dermorphin-NH2 (5).[17] We have previously reported that
Numerous methods have already been reported for the
asymmetric synthesis of mono-alkylated a-amino acids from
prochiral protected glycine derivatives by PTC alkylation with a
chiral quaternary ammonium salt as catalyst.[24–30] In a pioneer-
ing work, O’Donnell et al. described the asymmetric synthesis
of a-methyl-a-amino acids through enantioselective phase-
transfer alkylation of the p-chlorobenzaldimine of (S)-alanine
tert-butyl ester 9a with enantiomeric excesses up to only
50%.[31]
By using the more efficient N-antracenylmethyl dihydrocin-
chonidium bromide catalyst 19 Lygo et al. improved the enan-
tioselectivity for the preparation of a-MePhe to 87%.[32] More
the Aba-Gly structure adopts a chair-like ring conformation,
but is not able to induce a turn structure in the Ac-Aba-Gly-
NHMe model.[20] In contrast, the a-Me-substituted Aba ana-
logue adopts a boat-like ring conformation, and was shown to
be a strong b-turn inducer in this model.[21] This is an analogy
to the finding that a-MePro is a stronger turn inducer than
Pro.[22,23]
Herein we compare the effects of S and R stereoisomers of
Aba and a-MeAba incorporated into the dermorphin sequence
on m- and d-opioid receptor affinity, functionality and on the
conformation of the dermorphin analogues. Whereas asym-
metric synthesis of Aba-Gly has been described,[18–20] only a rac-
emic synthesis of a-MeAba-Gly was reported.[21] We now also
describe the asymmetric variant of the latter.
2036
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
Scheme 1. Asymmetric synthesis of Boc-(S)-a-MeAba-Gly-OH (16). Conditions: a) see Table 1; b) TFA/H2O (95:5), room temperature, overnight; c) tert-BuLi
(1.6m), dry THF, Ar, ꢀ788C, 1 h, o-CN-BnBr, 10 h (72%); d) i) TFA AcN/water (3:1), room temperature, overnight (quantitative), ii) HCl (6n), 658C, 50 h (65%);
e) NaOH (1n), MocCl, 1 h (70%); f) i) Pd/C, H2, EtOH/water, AcOH (10%), 0.34 mpa, 48 h, room temperature (85%), ii) EDC, pyridine, AcN/H2O (3:1), room tem-
perature, overnight (87%); g) NaH, DMF, BrCH2COOtBu, 1 h, room temperature (47%); h) i) HBr (33%) in AcOH, 558C, 1.5 h, and ii) Boc2O, Et3N, dioxane/water
(9:1), room temperature, overnight (67%).
recently, Chinchilla and colleagues optimized the reaction con-
ditions for the benzylation of 9a using catalyst 20. When the
2-naphthyl aldimine 9b was used instead of imine 9a, the ee
increased from 82 to 87%.[33] A C2-symmetric non-Cinchona
phase-transfer catalyst, derived from (S)-binaphthol, was used
by the Maruoka group for the benzylation of 9a, with an enan-
tioselectivity of 98%.[34,35]
CsOH as a base at ꢀ788C. These conditions were already
proven to yield (S)-o-cyanophenylalanine in our laboratory
with an enantioselectivity of 96%.[20] However, a considerably
lower ee was obtained for the alkylation of the alanine deriva-
tive (Table 1, entry 1, 40% ee). We, therefore, changed to the
third generation Cinchona catalyst of Lygo et al.,[28] 20, and
used the 2-naphthyl aldimine alanine tert-butyl ester 9b, as de-
scribed by Chincilla et al.[33] However, this procedure did not
improve the enantioselectivity (Table 1, entry 2, 36% ee). In ac-
cordance with the results reported by Jew and co-workers, the
aldimine 9a, RbOH and O(9)-allyl-N-1-cyanobenzylhydrocincho-
nidinium bromide (24)[39] catalyst were used in the alkylation
reaction, but only a moderate ee of 60% was obtained
(Table 1, entry 3).
With catalyst 22, which is also commercially available, the
Maruoka group obtained the same enantioselectivity (98%) for
this reaction.[36] Jew et al. reported on the synthesis of a-alkyl
alanines by phase-transfer alkylation of the 2-naphthyl aldi-
mine tert-butyl ester 9b with RbOH and O(9)-allyl-N-2’,3’,4’-tri-
fluorobenzylhydrocinchonidinium bromide (23) as a catalyst,
with up to 96% ee.[37] Catalyst 23 and the N(1)-2’-cyanobenzyl
derivative 24 were also successfully applied to the synthesis of
a-amino acids with enantioselec-
tivities varying from 94 to
>99% ee.[38,39]
Table 1. Enantioselective catalytic phase-transfer alkylation of 9a or 9b with o-cyanobenzyl bromide under
specified conditions.
We have investigated the alky-
lation of aldimines 9a and 9b
with o-cyanobenzyl bromide (10;
Scheme 1) using various chiral
catalysts and conditions de-
scribed for the reaction with
benzyl bromide. The results are
shown in Table 1.
Entry
Aldimine
Catalyst
Base
T [8C]
t [h]
Solvent
ee[a] [%]
1
2
3
4
5
6
7
8
9a
9b
9a
9a
9a
9a
9b
9b
9a
9a
9b
21
20
24
23
24
24
24
23
24
23
22
CsOH
RbOH
RbOH
RbOH
RbOH
CsOH
RbOH
RbOH
CsOH
CsOH
CsOH
ꢀ78
ꢀ20
ꢀ35
ꢀ35
ꢀ78
ꢀ35
ꢀ35
ꢀ35
ꢀ35
ꢀ35
ꢀ208C
27
24
26
26
20
4
21
21
4
CH2Cl2
40
36
60
54
13
68
34
58
60
70
40
toluene/CHCl3 (7:3)
toluene/CH2Cl2 (5:2)
toluene
toluene/CH2Cl2 (5:2)
toluene
toluene/CH2Cl2 (5:2)
toluene/CH2Cl2 (5:2)
CH2Cl2
First, the third generation cat-
alyst of Corey et al.,[26,29] 21, was
used for the enantioselective
catalytic phase-transfer alkylation
of the p-chlorobenzaldimine of
alanine tert-butyl ester 9a with
9
10
11
4
1
CH2Cl2
toluene
[a] Enantiomeric purity was determined by HPLC analysis after hydrolysis of 11 a,b to 12 and (S)-NIFE derivatiza-
tion.[40]
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2037

D. Tourwꢀ et al.
MED
Switching to the O(9)-allyl-N-2’,3’,4’-trifluorobenzylhydrocin-
chonidinium bromide catalyst (23),[37,38] under the same reac-
tion conditions, did not improve the enantioselectivity (entry 4,
54% ee). Lowering the temperature to ꢀ788C (entry 5, 13%
ee), switching to CsOH as a base (entry 6, 68% ee) by using the
2-naphthyl aldimine 9b (entry 7, 34% ee) or/and another cata-
lyst (entry 8, 58% ee) did not improve the enantioselectivity
dramatically in comparison with entry 3. The use of only CH2Cl2
as a solvent in the enantioselective alkylation also did not lead
to further improvements (entry 9, 60% ee; entry 10, 70% ee).
Finally, the commercially available catalyst used by the Maruo-
ka group, 22,[36] was applied in the phase-transfer catalysis re-
action of 9b with o-cyanobenzyl bromide. Unfortunately, this
did also not improve the enantioselectivity (entry 11, 40% ee).
These experiments demonstrate that the o-cyano substituent
drastically lowers the obtained enantiomeric excess, compared
to those obtained for alkylations by using unsubstituted
benzyl bromide. Therefore, another asymmetric synthesis
method for a-Me-o-cyanophenylalanine (12) had to be chosen.
Fortunately, various other methodologies for the asymmetric
synthesis of quaternary a-amino acids have been reported.
These include Schçllkopf’s bis-lactim ether strategy,[41,42] Wil-
liams’ diphenyloxazinones method,[43] and Seebach’s oxazolidi-
none-based methodology,[44] or one of its variants.[45–50] By
using an optimized Schçllkopf protocol, as described by Vassi-
liou et al.,[51] (S)-a-Me-o-CN-Phe·HCl (12) was synthesized with
good enantioselectivity (Scheme 1).
using a 33% solution of HBr in AcOH to provide the amino
acid, which was immediately converted into Boc-(S)-MeAba-
Gly-OH (16) by use of Boc2O. This enantiomerically enriched
building block was introduced into the dermorphin peptide se-
quence as described for the racemate. After cleavage from the
resin, the first eluting epimer 7 was identified to contain (S)-a-
MeAba, and hence the second eluting peptide was the corre-
sponding R epimer, 8.
Biological evaluation
The affinity and potency of compounds 1 and 5–8 for the m-
and d-opioid receptors were determined by competition bind-
ing experiments of the ligands with the receptor selective radi-
oligands [3H]DAMGO (m) and [3H][Ile5,6]deltorphin-2 (d) in rat
brain membrane homogenates (Table 2). Comparative analysis
of the binding results revealed that (S or R)-a-MeAba3- and (S
or R)-Aba3- modifications on the dermorphin structure, in gen-
eral, yielded analogues that retained m-receptor preferences
and decreased receptor selectivities. The native ligand (1)
showed affinity values consistent with the literature data,[52]
and exhibited a far higher selectivity than those of analogues
5 to 8. Among the analogues, two compounds (5 and 8)
proved to be as potent as the parent dermorphin (1) on the
basis of binding results; this suggests that they are more likely
to adopt conformations suitable for high-affinity ligand bind-
ing than compounds 6 and 7. Compared to the native dermor-
phin 1, [(S)-Aba3-Gly4]dermorphin (5) maintained subnanomo-
lar affinity for the m-opioid receptor, but showed a tenfold in-
crease in d affinity. This is in agreement with our previous re-
sults.[18] The R epimer 6 was substantially less potent, although
it still possessed low nanomolar affinity for the m-receptor. Sur-
prisingly, the [(R)-a-MeAba3-Gly4]dermorphin epimer (8) was
much more potent than the S epimer, 7. Compound 8 is about
four-times less potent with the m- and d-receptors relative to
the most potent ligand [(S)-Aba3-Gly4]dermorphin (5). Taking
these findings together, it should be noted that the introduc-
tion of the methyl substituent at the a-position of Aba is, in
the case of compound 8, well tolerated by both opioid recep-
tors.
Deprotonation of the substituted bis-lactim ether chiral aux-
iliary 17 with tBuLi, followed by an alkylation of the resulting
anion with o-cyanobenzyl bromide (10) provided compound
18 with a yield of 72% after purification by flash column chro-
matography. Hydrolysis of 18 by using TFA at room tempera-
ture, and subsequent hydrolysis of the methyl ester in HCl
(6n) at 658C, yielded (S)-a-Me-o-CN-Phe·HCl (12) with an ee of
88%, as determined by HPLC analysis after (S)-NIFE derivatiza-
tion[40] (S)-a-Me-o-CN-Phe·HCl was methyloxycarbonyl (Moc)-
protected to 13 by means of MocCl, as previously described
for the racemic compound.[21] Reduction of the nitrile in 13
with H2/Pd on C provided the aminomethyl derivative, which
underwent intramolecular cyclization to 14 after activation of
the carboxylic acid by using N-(3-methylaminopropyl)-N-ethyl-
carbodiimide (EDC). N-Alkylation of Moc-(S)-a-MeAba (14) with
tBu-bromoacetate and NaH as a base provided the ester 15.
Moc deprotection and ester hydrolysis was performed by
Functional properties of the new ligands were assessed by
the widely used ligand-stimulated [35S]GTPgS binding assay.
With regard to functionality, each ligand stimulated [35S]GTPgS
binding over the basal activity. The resulting values clearly indi-
Table 2. Receptor binding affinities and selectivities.
Peptide
[3H]DAMGO
Kim [nm]
[3H][Ile5,6]deltorphin-2
Selectivity
Kid/Kim
EC50 [nm]
Emax [%]
Kid [nm]
Dermorphin (1)
0.2ꢁ0.05
0.3ꢁ0.04
10.3ꢁ0.9
252ꢁ40
1.1ꢁ0.1
381ꢁ31
33. 3ꢁ2.2
612ꢁ18
1905
111
59
–
121
46ꢁ2
42ꢁ7
155ꢁ2
175ꢁ6
149ꢁ2
123ꢁ4
153ꢁ3
H-Tyr-d-Ala-[(S)-Aba-Gly]-Tyr-Pro-Ser-NH2 (5)
H-Tyr-d-Ala-[(R)-Aba-Gly]-Tyr-Pro-Ser-NH2 (6)
H-Tyr-d-Ala-[(S)-a-MeAba-Gly]-Tyr-Pro-Ser-NH2 (7)
H-Tyr-d-Ala-[(R)-a-MeAba-Gly]-Tyr-Pro-Ser-NH2 (8)
855ꢁ151
2436ꢁ397
184ꢁ35
>10000
134ꢁ10
Receptor binding affinities and selectivities of native dermorphin (1) and its constrained heptapeptide analogues 5–8, and their efficacies determined by
the ligand-stimulated GTPgS functional assays. Data are means ꢁSEM of 3–5 independent experiments. No selectivity was determined when the Ki value
proved to be higher than 10000.
2038
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
cate that the ligands bind the opioid receptors and activate
G proteins (Table 2). These data were compared to those of the
Hydrogen–hydrogen distances were extracted from 200 ms
off-resonance ROESY spectra. The 47 interproton distances,
consisting of 22 intraresidue and 25 inter-residue (i, i+1) were
derived for 5. For 6, 62 proton–proton distance restraints were
obtained including 31 inter-residue (i, i+1) contacts. Impor-
tantly, no long-range (i, i+n; n>1) contacts were observed for
both 5 and 6; this indicates that these molecules most likely
adopt an extended structure, as expected. For 7, 52 distance
restraints were extracted corresponding to 25 intraresidue and
27 inter-residue NOE contacts including eight long-range con-
tacts. These involve Ala2 Hb, for which contacts to Tyr5 HN, Hb,
Hd and Ser7 HN, Ha, Hb were found. Additionally, contacts from
Ala2 HN to Tyr5 HN and Hb were observed. Finally, 46 NOE con-
tacts were assigned for 8, consisting of 21 intraresidue, 21
inter-residue (i, i+1) and four long-range contacts. Two of
these correlate with the Tyr5 Hb protons with Ala2 HN and
Ser7 HN. Furthermore, contacts from Ala2 Hb to Tyr5 Hd and
from Aba3 Hi (Figure S2, Supporting Information) to Pro6 Hd
were present. From these observations, both 7 and 8 are ex-
pected to adopt a turn-like feature. The complete set of re-
straints used for all molecules is presented in Table S8 in the
Supporting Information.
prototypic m-receptor ligand DAMGO (EC50 =122 nm, Emax
=
178%). As can be seen, each ligand showed an agonist charac-
ter providing Emax values of 123 up to 175. Interestingly, the
tendency of ligand potencies and efficacies correlated well
with those of the Ki values, that is, ligands that had low or
high m-receptor affinities, displayed similarly low or high EC50
and proportional Emax values, suggesting profiles ranging from
partial to full agonists. Considering these facts, it can be con-
cluded that (S or R)-a-MeAba3- or (S or R)-Aba3- modifications
of the heptapeptide dermorphin do not influence functionality
and the ligands retain their G-protein binding and activating
properties.
Conformational analysis of 5 to 8
Our previous studies on dipeptide amides containing Aba-Gly
and a-MeAba-Gly demonstrated that the former adopted
mainly extended structures and the latter showed a well-de-
fined b turn.[21] To investigate the persistence of these features
in the novel dermorphin compounds 5 to 8, conformational
studies by using NMR-based molecular modeling techniques
were performed.
The results from this qualitative analysis of the data were va-
lidated by using NMR-based molecular modeling. Initially, a si-
mulated annealing (SA) restrained MD protocol was used for 5
and 8. However, analysis of the conformations sampled by SA
for 5 revealed that on average 21 distance restraints (45%)
were violated with maximum violations between 0.23 and
0.45 ꢃ. The simulation reveals an extended N-terminal confor-
mation with its C terminus folding back over the Aba3 residue.
The fact that the simulated annealing protocol does not gener-
ate any conformations that satisfy the majority of the NOE re-
straints at the same time, suggests some degree of conforma-
tional flexibility of 5. Indeed, it is well known that flexible
linear peptides that are able to adopt many low-energy confor-
mations in solution, lead to averaging of the various NMR con-
straints. By imposing these averaged constraints, an averaged,
virtual, and therefore, nonexistent conformation is generated,
which does not represent the actual conformational space
available to the peptide.
From the initial analysis of the 1D proton spectrum two dis-
tinct sets of signals were found to occur for each of the com-
pounds. The population of the major set in 5 to 8 was deter-
mined at 82–86%. The presence in each compound of clear
exchange correlation cross-peaks in 2D ROESY that connect
both sets of resonances, indicates the presence of two confor-
mations in slow exchange (kex ~1 sꢀ1) on the NMR timescale.
Their presence results from cis–trans isomerization of the Tyr5-
Pro6 amide bond as proven by NOE contacts correlating Tyr5
Ha with either the Hd (major), or the Ha (minor; Figure S1, see
the Supporting Information).
Assignments and conformational studies of 5–8 were re-
stricted to the major conformation due to the low signal inten-
sities of the minor conformation and were obtained from the
analysis of 2D 1H,1H TOCSY and 1H,13C HSQC/HMBC 2D NMR
spectra (Tables S3–S6 in the Supporting Information). Addition-
al conformational dependent NMR spectroscopy parameters,
In order to address this, time averaged MD simulations (tar-
MD)[53] were performed instead on compounds 5–8. Such sim-
ulations generate a conformational ensemble that represents
the conformational space, that is, on average over time, in
3
that is, JHaHN scalar couplings, HN temperature coefficients and
HN hydrogen deuterium exchange rates (Table 3) were deter-
mined as described in the Experimental Section.
Table 3. Temperature coefficients and deuterium exchange rates of 5–8.
Dd/DT [ppbKꢀ1
]
t1/2 (d-exchange)
3JHaHN [Hz]
Dd/DT [ppbKꢀ1
]
t1/2 (d-exchange)
3JHaHN [Hz]
5
6
5
6
5
6
7
8
7
8
7
8
[a]
[a]
[a]
Tyr 1
ꢀ1.47
ꢀ4.47
ꢀ6.10
ꢀ6.53
ꢀ4.23
ꢀ1.60
ꢀ3.94
ꢀ5.86
ꢀ6.48
ꢀ4.29
1
1
1
100
100
29
5.64
8.04
7.65
8.22
7.85
5.16
7.79
7.02
8.32
8.86
Tyr1
ꢀ1.40
ꢀ4.07
ꢀ5.50
ꢀ5.03
ꢀ4.40
–
1
–
5.75
7.76
–
8.43
7.97
–
d-Ala2
Aba3
Tyr 5
92
100
58
d-Ala2
Aba3
Tyr5
ꢀ3.00
ꢀ6.01
ꢀ4.73
ꢀ5.03
190
100
87
1
7.72
[b]
[b]
100
4000
400
–
8.22
7.96
Ser7
17
Ser7
33
Temperature coefficients in the 298 to 328 K interval, deuterium exchange half-lifetimes (normalized to Aba3) and scalar coupling constants of amide pro-
tons (and ammonium protons for Tyr1). [a] Ammonium protons not observed; [b] Ca-methylated residue.
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2039

D. Tourwꢀ et al.
MED
agreement with the available NMR data. Therefore, NMR con-
straints need not be instantaneously and simultaneously satis-
fied and individually sampled conformations are allowed to
violate part of the NMR data at any specific time point in the
trajectory. Yet, this protocol forces each restraint to be satisfied
on average over a certain simulation time interval (here
chosen to be 8 ns).
A view of the conformational space sampled during the tar-
MD trajectories for 5–8 is shown in Figure 2, and the distribu-
tion of main chain F,Y dihedral angles for the five internal res-
idues are presented by using Ramachandran plots (Figure 3).
These charts allow a better appreciation of the extent of the
conformational space sampled by each individual amino acid
backbone during the trajectory. Comparison with the F,Y di-
hedral angles obtained from the simulated annealing restrain-
ed MD protocol (Figure 3) clearly shows the detrimental
impact of imposing averaged constraints, thereby generating
virtual conformations with unfavorable energies. For 5, for in-
stance, the simulated annealing protocol results in a single
Pro6 population in an unfavorable area of the Ramachandran
map, whereas application of tar-MD populates the two ener-
getically favored conformations. For other residues, the local
conformations, generated by the SA approach, fit quite well to
the conformational space in the tar-MD ensemble (e.g., Aba3).
Figure 2. Overlay of 70 conformations of 5–8 uniformly sampled from the
last 78 ns of 80 ns time-averaged restrained MD simulations. Superpositions
are based on all heavy atoms of Aba3 and Gly4 residues. No side chains are
shown except for Aba3.
Figure 3. Ramachandran plots of dermorphin analogues 5 to 8 resulting from tar-MD (gray) and SA (red) simulations.
2040
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
The most evident and unambiguous difference in conforma-
tion between the compounds that can be observed when
comparing these plots is evident for the Aba3 residue. Com-
paring 5, 6 with 7, 8 clearly shows the former pair to adopt
mainly (and 6 exclusively) an extended conformation in the en-
semble, whereas the latter pair adopt mostly (and 7 exclusive-
ly) a helical conformation, in agreement with the qualitative
analysis above. The C-terminal Pro6 is shown to cover the
entire energetically allowed F,Y space, indicating flexibility at
the C terminus. For 7, however, the Y angle remains confined
to the helical area. Within each group, the impact of the
change in stereochemistry of the Aba3 residue results mostly
in the inversion of the populated areas, as can be expected.
Similar conformational spaces are covered by d-Ala2, with
both extended and helical character present, except for 8, in
which only extended conformations appear (Figure 3). Other
trends that can be observed are less general. While Tyr5 popu-
lates clearly different areas when comparing 5 and 6, which
suggests a link to the Aba3 stereochemistry, no such differ-
ence is apparent in 7 and 8. Finally, the behavior of the Gly4
residue is least amenable to interpretation. For 6–8, mostly a
single helical region is populated for this residue, but no evi-
dent link with the stereochemistry of the preceding Aba3 is
apparent.
pling constants of d-Ala2 and Ser7 were reproduced within
0.4 and 0.2 Hz, respectively. For the Tyr5 residue a scalar cou-
pling of 7.3 Hz was obtained during the tar-MD simulation (ex-
perimental: 8.22 Hz). This simulation clearly shows well-defined
major and minor conformations for the d-Ala2, Aba3, Gly5 res-
idues and a single conformation for Gly5, while Pro6 popu-
lates two conformations similar to those observed for 5 and 6.
Overall, a clear turn structure is visible over Aba2-Gly3, which
is closed by a hydrogen bond involving Tyr5 HN and d-Ala2
CO. In the major conformation, a second hydrogen bond is ad-
ditionally present between d-Ala2 HN and Tyr5 CO. For the
minor conformation the C-terminal residues adopt a wider
range of conformations. The distance and angles defining the
geometry of the Tyr5 HN to d-Ala2 CO hydrogen bond is pre-
sented in Figure 4, which shows that values leading to an ener-
While in 5, the d-Ala2-Aba3 dipeptide is rather well-defined
and mostly occupies an extended conformation, the larger
F,Y space accessed by the Gly4-Tyr5 dipeptide in 5 suggests
a more flexible C-terminal segment, whereby the C terminus is
folded back towards the N terminus by two distinct conforma-
tions on opposing sides of the molecule, as clearly evident
from Figure 2. Despite this back-folding, no turn-like structure
is adopted. Furthermore, no hydrogen bonds were observed in
this set of molecules, which is in good agreement with the
measured temperature coefficients and hydrogen–deuterium
exchange rates.
Figure 4. NꢀH$O=C distances (ꢃ) and angles (8) of the d-Ala2 HN···Tyr5 CO
hydrogen bond observed during tar-MD simulations of 7 (top panel) and 8
(lower panel).
Considering that on average 19 distance restraints (40%)
were still violated with a maximum violation of 0.35 ꢃ in the
tar-MD trajectory of 5, no significant drop in violations is ob-
served when compared to the previous simulated annealing
results. In all, 5 can be considered a flexible molecule without
any regular structural elements.
getically favorable geometry (i.e., 2.4 ꢃ and 1508 for the hydro-
gen bond distance and angle, respectively) is on average
adopted in the ensemble.
For the second hydrogen bond, values of 2.6 ꢃ and 1208
were obtained, indicating a less optimal geometry. Based on
the mean f and y angles of Aba3 (478, 558) and Gly4 (1188,
ꢀ108), this turn structure can be classified as a type I’ b-turn.
The stability of the secondary structure element is supported
by the very slow exchange rate of Tyr5 HN, which also features
On average, 13 distance restraints (21%) were violated for 6
with an average violation of 0.11 ꢃ, giving a satisfactory agree-
ment. Here, the backbone of Gly4 and Tyr5 is restricted to one
single conformation. Compared to 5, this results in a peptide
conformation in which the C terminus is still folded toward the
N terminus of the peptide but only along one side of the
Aba3 residue, resulting in a more defined overall conforma-
tion. Again, no stabilizing hydrogen bonds were observed in
the set of conformations modeled and no classical turn-like
structure is adopted. This agrees with the experimental tem-
perature coefficients and hydrogen–deuterium exchange rates
measured. In all, the core tripeptide of 6 can be considered to
adopt a single but rather flexible conformation without any
regular structural elements.
a
lower temperature coefficient of ꢀ4.73 ppbKꢀ1 (cfr.
ꢀ6.48 ppbKꢀ1 for 6). An even smaller temperature depend-
ence of ꢀ3.00 ppbKꢀ1 is observed for the d-Ala2 residue. How-
ever, the d-Ala2 HN can be efficiently exchanged. This is in line
with the fact that this hydrogen bond is not persistent since it
is not featured in the minor conformation.
Finally, the most populated conformation of the Aba3 and
Gly4 residues in 7 presents an inverse geometry with respect
to 8 (Figure 3). This is further reflected in the seven-membered
ring geometry (Table 4).
A turn-like structure is again apparent, closed by the same
Tyr 5 HN to d-Ala2 CO hydrogen bond as in 8. Given the inver-
sion of all torsion angles of the Aba3-Gly4 dipeptide, com-
For 8, 19 distance restraints were violated on average, but
3
none more than 0.37 ꢃ while the measured JHNHa scalar cou-
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2041

D. Tourwꢀ et al.
MED
dermorphin analogues 5 and 6. In agreement with previous re-
sults, the [(S)-Aba3-Gly4]dermorphin (5) retained m-opioid af-
finity but displayed an increased d affinity. The corresponding
R epimer 6 was much less potent. In contrast the [(R)-a-
MeAba3-Gly4]dermorphin isomer 8 was more potent than its
S epimer, 7. This is in agreement with previous results for the
related [spiro-Aba3-Gly4]endomorphin isomers, for which the
R epimer is the most potent, whereas in [Aba3-Gly4]endomor-
phin the S enantiomer is the most potent.[54]
Table 4. Aba3 conformation of 5–8.
%
Aba3 f Aba3 y Aba3 c1 Aba3 c2 Aba3 c5 Aba3 c4
5
6
7
8
94.1 ꢀ165
175
ꢀ165
ꢀ50
60
180
180
0
0
ꢀ60
50
50
60
ꢀ65
ꢀ60
55
100
100
165
ꢀ50
60
60ꢁ20 15ꢁ20
92.8
ꢀ65ꢁ20 ꢀ5ꢁ20 ꢀ45
Population (%) and geometry (dihedral angles, deg, see Figure 1) of the
major conformation of the seven-membered ring in four dermorphin ana-
logues 5–8 from the 80 ns tar-MD simulations.
For the non-methylated dermorphin analogues 5 and 6, tar-
MD simulations showed a degree of folding of the C terminus
residues toward the N terminus of the peptide, without adopt-
ing a stabilizing b-turn conformation, however, and with con-
siderable flexibility, especially for 5. Both 7 and 8, on the other
hand, exhibited a type I/I’ b-turn conformation over the Aba3
and Gly4 residues stabilized by a hydrogen bond involving
Tyr 5 HN and d-Ala2 CO. For 8 this is nicely correlated to the
hydrogen–deuterium exchange rate. For 7 the modeling re-
sults indicate that this hydrogen bond is less persistent and is
therefore not reflected in its hydrogen–deuterium exchange
rate. A second hydrogen bond involving d-Ala2 HN and
Tyr5 CO has been observed in 8, although also in this case its
persistence is not clearly reflected in its hydrogen–deuterium
exchange rate. This conformational study clearly shows the in-
fluence of the methylation of Aba3 Ca on the conformational
behavior of these dermorphin analogues in [D6]DMSO solution.
However, the Aba3 Ca stereochemistry has been shown to
have less effect on the overall conformational preference of
the peptide.
pared to 7, this turn can be classified as a type I b turn. The hy-
drogen bond has a similar and favorable geometry as the cor-
responding one in 8 (Figure 3), but is less persistent during the
simulation, since it is clearly absent during a short time in the
tar-MD trajectory. The existence of a second hydrogen bridge
involving d-Ala2 HN was only observed during about 20% of
the total simulation. The existence of the hydrogen bond in-
volving Tyr5 HN and d-Ala2 CO is corroborated by the many
long-range NOE contacts observed in this compound. On the
other hand, the measured temperature dependence of the
amide protons and their hydrogen–deuterium exchange rates
are not clearly indicative for this feature as a considerably fast
deuterium exchange rate was recorded for Tyr5 HN in 7, sup-
porting a lower persistence than in 8. Even though the latter
experimental parameters average over a much longer time
scale than the one sampled during the simulation, the com-
bined result does suggest a more stable b-turn closing hydro-
gen bond in 8 than in 7.
In none of the analogues 5–8, evidence for a stacking of the
d-Ala2 methyl between the Tyr1 and Aba3 aromatic rings, as
found for the parent dermorphin (1), was observed. The chem-
ical shift of the d-Ala2 methyl signals was found at 1.11 to
1.13 ppm and not below 1 ppm as for 1. Remarkably, the ring
conformation of the seven-membered azepinone ring of Aba3
is different in the non-a-methylated analogues 5 and 6, in
which the c1 is trans, compared to the a-methylated analogues
7 and 8, in which c1 is gauche (Table 3). In contrast, previous
conformational studies on Ac-a-MeAba-Gly-NH-Me have
shown a trans ring conformation in solution and in the crys-
tal.[21] Theoretical conformational studies showed that the
gauche conformation was the lowest energy one, the trans
having 2.34 kJmolꢀ1 higher energy. This indicates that the pep-
tide sequence is able to switch preference between the two
conformations.
The opioid receptor affinities of analogues 5 and 8 were
shown to be very similar. Nevertheless, both analogues show a
different conformational preference. This might indicate that
these flexible ligands are able to adapt their conformation on
binding to the receptor in order to obtain a common binding
topography of the pharmacophoric groups.
Experimental Section
General: All Boc-protected amino acids-AA-OH were obtained
from Fluka (Bornem, Belgium) as well as diisopropylcarbodiimide
(DIC), 1-hydroxybenzotriazole (HOBt), trifluoroacetic acid (TFA), and
dichloromethane. The MBHA resin (0.95 mmolgꢀ1) was purchased
from Neosystem (Strasbourg, France). All other reagents were from
Sigma–Aldrich (Bornem, Belgium). 2-1H(Benzotriazol-1-yl)-1,1,3,3-
tetramethyluronium tetrafluoroborate (TBTU) was purchased from
Senn Chemicals (Dielsdorf, Switzerland). H-Tyr-d-Ala-NMePhe-Gly-ol
(DAMGO), tris(hydroxymethyl)aminomethane (Tris, free base), bovin
serum albumin (BSA), guanosine-5’-[g-thio]triphoshate (GTPgS), in-
organic salts, organic solvents and HCl (37%) were purchased from
Sigma–Aldrich and Molar, Kft. (Budapest, Hungary). The tritiated
opioid ligands were prepared in the Isotope Laboratory of the Bio-
logical Research Center (HAS-BRC, Hungary). The radiolabeled
GTPgS ([35S]GTPgS; specific activity ꢂ1000 Cimmolꢀ1) was ordered
from the Izotop Intezet, Kft. (Hungary).
Conclusions
Asymmetric phase-transfer catalysis was not able to provide
(S)-a-Me-o-cyanophenylalanine (12) in acceptable enantiomeric
purity. However, by using a Schçllkopf chiral auxiliary, this im-
portant intermediate was obtained in 88% ee, and it was fur-
ther transformed into (S)-a-MeAba-Gly (16). This allowed the
preparation of the Phe3 constrained analogues of dermorphin
7 and 8, and to compare the opioid affinities, selectivities and
functional properties to those of the non-a-methylated Aba-
RP-HPLC (Waters Breeze analytical HPLC system) was performed by
using a RP C-18 column (Supelco Discovery BIO Wide Pore C-18,
l=25 cm, d=0.45 cm, PS=5 mm) with a mobile phase consisting
of water/acetonitrile containing TFA (0.1%). Products were eluted
2042
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
by using the gradient: t=0 min, 3% CH3CN, t=20 min, 97%
treated according to the European Communities Council Directives
(86/609/ECC) and the Hungarian Act for the Protection of Animals
in Research (XXVIII.tv. 32.§). The protein content of the homoge-
nate was determined by the method of Bradford[57] by using BSA
as standard. The protein concentration ranged between 0.2–
0.4 mg per test tube. The following conditions were applied to
assess inhibitory constants: [3H]DAMGO (258C, 1 h, GF/C filter,
glass tube), [3H][Ile5,6]deltorphin-2 (358C, 45 min, GF/B filter, plas-
tic tube, 0.25 mg BSA per tube). The incubation mixtures were
filled to a final volume of 1 mL with Tris-HCl buffer (50 mm, pH 7.4)
and samples were incubated in a shaking water-bath at the appro-
priate temperature. Competition binding experiments were per-
formed by incubating the membranes with [3H]DAMGO (0.5 nm),
[3H][Ile5,6]deltorphin-2 (2 nm) and increasing concentrations
(10ꢀ11–10ꢀ5 m) of the unlabeled dermorphin analogues. Nonspecific
binding was determined with naloxone (10 mm) and subtracted
from the total binding to yield specific binding. Incubation was ini-
tiated by the addition of the membrane suspension and stopped
by rapid filtration over Whatman GF/C or GF/B glass fiber filters, by
using a Brandel cell harvester. Filters were washed with ice-cold
Tris-HCl buffer (pH 7.4; 3ꢅ5 mL), and then the filter-bound radioac-
tivity was measured in an Optiphase Supermix scintillation cocktail
(PerkinElmer) by using a TRI-CARB 2100TR liquid scintillation coun-
ter. Each experiment was performed in duplicate and analyzed by
the one/two-site binding competition fitting option of the Graph-
Pad Prism software (Version 4).
CH3CN; flow-rate: 1.0 mLminꢀ1
, l=215 nm. Preparative HPLC
(Gilson semipreparative HPLC system) was performed by using a
RP C-18 column (Supelco Discovery BIO Wide Pore C-18, l=25 cm,
d=2.12 cm, PS=10 mm) with the above-mentioned gradient at a
flow-rate of 20.0 mLminꢀ1. TLC analysis was performed on plastic
sheets precoated with silica gel 60F254 (Merck). Melting points were
measured with a Bꢄchi B 540 melting point apparatus, with a tem-
1
perature increment of 18Cminꢀ1. H and 13C NMR spectra were re-
corded at 250.13 and 62.90 MHz, respectively, with a Bruker
Avance DRX250 spectrometer, by using TMS or the residual solvent
signal as internal reference. Mass spectra were recorded on a VG
QuattroII spectrometer by using electrospray ionization in positive
ion mode (Micromass, Manchester, UK). For LCMS recording, a
Waters Breeze analytical HPLC with manual injector, UV detector
(2487, 215 nm) and Waters 600E pump was coupled to the VG
QuattroII spectrometer. The runs were performed on a reversed
phase C-18 column (Supelco Discovery BIO Wide Pore C-18, l=
25 cm, d=4.6 mm, PS=5 mm) with a mobile phase consisting of
water/acetonitrile containing TFA (0.1%). Products were eluted by
using the gradient: t=0 min, 3% CH3CN, t=20 min, 97% CH3CN;
flow-rate: 1.0 mLminꢀ1, l=215 nm.
Peptide synthesis
Peptide synthesis was performed manually in fritted polypropylene
syringes (10 mL), which were shaken during coupling, washing, de-
protection and cleavage steps. The N-terminal dermorphin ana-
logues 1 and 5–8 were manually prepared by solid-phase synthesis
by using Boc chemistry and the MBHA resin (0.95 mmolgꢀ1) as
solid support. The resin was swollen in CH2Cl2 for 1 h followed by
treatment with 20% DIPEA/CH2Cl2 to neutralize. Side chain protect-
ing groups were used in the case of Ser(Bn) and Tyr(2,6-Cl2Bn).
Peptides 5 and 6 were prepared by using a threefold excess of the
amino acids and a twofold excess of the (S)-Aba-Gly and (R)-Aba-
Gly building blocks together with TBTU (4 and 3 equiv, respective-
ly) as a coupling reagent and DIPEA (8 and 6 equiv, respectively) as
base in a mixture of anhydrous DMF/anhydrous CH2Cl2 (1:1). A
threefold excess of the amino acids and coupling reagent (DIC)
were used together with a threefold excess of HOBt anhydrous in
a DMF/anhydrous CH2Cl2 (1:1) solvent mixture to prepare peptides
7 and 8. The completeness of the couplings was checked with the
NF-31 test.[55] If incomplete, the coupling was repeated. Boc depro-
tections were performed in 50% TFA/2% anisol/CH2Cl2 (1ꢅ10 and
1ꢅ20 min). When the SPPS was completed, the resin was washed
with CH2Cl2 three times and then neutralized with 20% DIPEA/
CH2Cl2 solution (3ꢅ1 min). Between every coupling, deprotection
and cleavage step the resin was washed three times with DMF,
iPrOH and CH2Cl2. Final cleavage of the peptide from the resin was
done by treatment with HFliq (Asti, France) for 1 h at 08C. Anisol
(2 mL) and HFliq (10 mL) were used for 1 g peptide–resin. After
evaporation of the HF under vacuum the crude peptide was pre-
cipitated with cold dry ether and filtered. The precipitated peptide
and resin beads were washed with glacial acetic acid. The filtrate
was lyophilized and these crude mixtures were purified by prepara-
tive HPLC with a final purity >99%. Peptide purity and identity
was determined by LCMS.
Ligand-stimulated [35S]GTPgS functional assay: Ligand-stimulated
[35S]GTPgS binding was performed as described elsewhere.[58] Brief-
ly, rat brain membranes (15 mgprotein per tube) were incubated
with [35S]GTPgS (0.05 nm) and opioids (10ꢀ10–10ꢀ5 m) in the pres-
ence of GDP (30 mm), NaCl (100 mm), MgCl2 (3 mm) and EGTA
(1 mm) in Tris-HCl buffer (50 mm, pH 7.4) for 60 min at 308C. Basal
binding activity was measured in the absence of opioids and was
corrected with the nonspecific binding to yield specific binding.
Nonspecific binding was determined with unlabeled GTPgS
(10 mm). The reaction was initiated by the addition of the protein
and terminated by the addition of ice-cold buffer (5 mL, 50 mm
Tris-HCl, pH 7.4); then the samples were filtered through a What-
man GF/B glass fiber filter by using a Brandel cell harvester. Vials
were washed three times with ice-cold buffer (5 mL), and then im-
mersed in Optiphase Supermix scintillation cocktail. The radioactivi-
ty was measured by using a TRI-CARB 2100TR liquid scintillation
counter. Ligand stimulations were expressed as a percentage of
the specific [35S]GTPgS binding over the basal activity. Each mea-
surement was performed in triplicate and analyzed by the sigmoid
dose-response curve fitting option of the GraphPad Prism software
(Version 4).
NMR investigation of 5–8
Samples are prepared by using standard high performance 5 mm
NMR spectroscopy tubes with solutions of 5–8 (~5 mm) in
[D6]DMSO. All experiments used for assignments were performed
either on a Bruker DRX spectrometer (5, 8) equipped with a 5 mm
1
TXI-Z probe and operating at a H and 13C frequency of 500.13 and
125.77 MHz, respectively, or on a Bruker AvanceII spectrometer (6,
1
7) equipped with a 5 mm TXI-Z probe and operating at a H and
13C frequency of 700.13 and 176.06 MHz, respectively. The temper-
ature used was 298 K throughout. All 1D and 2D experiments were
performed by using pulse sequences available from the standard
Bruker library.[59] Gradient enhanced sequences were used for the
COSY and heteronuclear 2D experiments. TOCSY mixing times of
200 ms range were used. The HMBC spectra were optimized for
Biology
Radioligand binding assay: For competitive binding experiments,
rat brain membrane homogenate (Wistar, male, 250–300 g body
weight) was prepared as described elsewhere.[56] Animals were
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2043

D. Tourwꢀ et al.
MED
n
8 Hz long-range JCH couplings. Typically, 2D spectra consisted of
512 t1 increments of 8–32 scans each, sampled with 4 k points.
Prior to Fourier transformation, the data were zero-filled along t1
and suitably apodized by using a squared cosine bell in both di-
mensions, except for the COSY and HMBC, for which a squared
field. AM1-BCC charges[68] generated by the antechamber module
were used as partial atomic charges for 5–8. Averaging of restrain-
ed proton–proton distances was performed by using an Rꢀ6 rela-
tionship and an exponentially decaying function [Eq. (1)]. The
memory decay constant, t, was set to 8 ns. These restraints were
used in a single 80 ns MD simulation at 300 K by using a scaling
factor of 57 (corresponding to DMSO) for electrostatic interactions
as basic implicit solvent model.
1
sine bell was used. The H chemical shifts were obtained by using
1D proton NMR spectroscopy if resolved or deduced from 2D NMR
cross-peaks otherwise.
Temperature coefficients were measured from additional 1D
t
R
0
Rðt0Þꢀ6eðt ꢀtÞ=tdt0
3
proton NMR spectra at 308, 318 and 328 K. The JHH scalar coupling
1
0
RTARðtÞꢀ6
¼
ð1Þ
constants were extracted from the 1D H spectrum at 298 K. The
t
R
1H–2H exchange rates of amide protons were obtained by integrat-
ing amide resonances from 1D 1H spectra recorded after spiking
the sample with a small amount of D2O. All processing and analysis
of these spectra was performed by using Bruker TopSpin 2.1 soft-
ware suite.
0
eðt ꢀtÞ=tdt0
0
NOE distance restraints were introduced by using 2 kcalmolꢀ1 ꢃꢀ2
flat-bottomed harmonic potentials. During all simulations, dihedral
angle restraints were introduced by keeping each peptide bond in
the trans configuration. No scalar couplings were used as restraints
within the molecular modeling scheme.
1
NMR distance restraints for 5–8 were extracted from H–1H off-res-
onance ROESY spectra[60] recorded at a magnetic field strength of
700.13 MHz for protons. The 4 k by 1 k data points were recorded
in the direct and indirect dimension, respectively. Then 64 scans
were recorded for each increment by using a total interscan relaxa-
tion delay of 1.9 s. Off-resonance irradiation was performed during
200 ms by using a irradiation strength of 10 kHz and an offset of
ꢁ5773.5 Hz. Spectra were Fourier transformed after apodization
with squared cosine window functions in both dimensions result-
ing in a spectrum of 4 k by 4 k data points.
Synthesis
2-(((2R,5S)-2-Isopropyl-3,6-dimethoxy-5-methyl-2,5-dihydropyra-
zin-5-yl)methyl)benzonitrile (18):
15.75 mL, 25.2 mmol) was added dropwise to
A
solution of tBuLi (1.6m;
solution of
a
(2R,5SR)-2-isopropyl-3,6-dimethoxy-5-methyl-2,5-dihydropyrazine
(17; 5.0 g, 25.2 mmol, 5 mL, 1=1 gmLꢀ1) in dry THF (50 mL) under
argon atmosphere at ꢀ788C. The resulting solution was stirred at
ꢀ788C for 1 h and treated carefully with a solution of o-cyanoben-
zyl bromide (6.5 g, 33.3 mmol, 1.32 equiv) in dry THF (70 mL). The
reaction mixture was stirred at ꢀ788C for 10 h. The mixture was al-
lowed to warm slowly to room temperature and was subsequently
quenched with a saturated NH4Cl solution (80 mL). The aqueous
layer was twice extracted with Et2O (80 mL). The combined organic
phases were washed with brine (150 mL) and dried over MgSO4.
After removal of the solvent in vacuo, the crude was purified by
flash column chromatography (CH2Cl2/petroleum ether 1:1!100%
CH2Cl2) to give 18 (5.7 g, 72%): Rf =0.40 (CH2Cl2/CHCl3 9:1); mp:
51.2–53.98C; [a]D = +15 (c=1 in CH2Cl2); 1H NMR (DMSO,
250 MHz) dH =0.51 (d, J=6.8 Hz, 3H), 0.89 (d, J=6.9 Hz, 3H), 1.43
(s, 3H), 2.04 (m, 1H), 3.03 (d, J=13.1 Hz, 1H), 3.17+3.24 (AB
system, JAB =13.1 Hz, 2H), 3.63 (s, 6H), 7.17 (d, J=7.4 Hz, 1H), 7.37
(t, J=7.6 Hz, 1H), 7.56 (t, J=7.7 Hz, 1H), 7.70 ppm (d, J=7.7 Hz,
1H); 13C NMR (DMSO, 63 MHz) dC =16.2, 19.1, 28.4, 29.9, 44.5, 52.0,
52.1, 59.1, 59.6, 113.4, 118.2, 127.4, 130.6, 132.4, 132.5, 140.3, 161.5,
163.1 ppm; MS (ES+): m/z 314 [M+H]+, 282, 197; HPLC tR =
19.9 min.
Cross-peaks were assigned and integrated manually within the
CC NMR software suite[61] and converted to interproton distances
by using an Rꢀ6 conversion. Contacts between methylene moieties
were used as a calibration for this conversion. Upper boundaries
were set at 110% of the experimental NOE distance while lower
boundaries were fixed at 1.8 ꢃ. Non-stereospecific NOE contacts
were treated by introducing pseudoatoms and increasing the
upper bound of the restraint by a fixed value.[62]
Molecular modeling of 5–8
Molecular modeling studies by using the classical simulated an-
nealing[63,64] approach were performed within the Discover3
module of InsightII (Accelrys) by using the built-in consistent force
field (CFF91).[53] Partial atomic charges were generated by using
the standard module within InsightII. A distance dependant scaling
for electrostatic interactions was used as a basic implicit solvent
model in all calculations. During an initial unrestrained MD simula-
tion of 5.5 ns at 1000 K, 100 conformations of 5 to 8 were uniform-
ly sampled over the last 5 ns of the trajectory. While subjecting
each of these 100 conformations to the experimental NOE distance
constraints short MD (1.5 ps) simulations were subsequently ran at
decreasing temperatures down to 300 K. At 300 K, an additional
4 ps MD simulation was performed. NOE distance restraints were
introduced by using 40 kcalmolꢀ1 ꢃꢀ2 harmonic potentials when
the experimental distance boundaries were exceeded. Additionally,
each conformation obtained was subjected to a full restrained
energy minimization by using the same potential for the NOE re-
straints. During all simulations, dihedral angle restraints were intro-
duced by keeping each peptide bond in the trans configuration.
No scalar couplings were used as restraints within the molecular
modeling scheme.
a-Methyl-(S)-o-cyanophenylalanine hydrochloride (12): Trifluoro-
acetic acid (10.3 mL) was added to a solution of 2-(((2R,5S)-2-iso-
propyl-3,6-dimethoxy-5-methyl-2,5-dihydropyrazin-5-yl)methyl)ben-
zonitrile (18; 5.6 g, 17.9 mmol) in acetonitrile/water (82 mL, 3:1)
and the resulting solution was stirred, overnight (~10 h). After
evaporation, the residue was diluted with ethyl acetate (100 mL)
and water (80 mL). The aqueous layer was neutralized with a 5%
solution of NaHCO3 and extracted with CH2Cl2 (4ꢅ80 mL). The or-
ganic phase was dried over MgSO4 and evaporated to give a-
methyl-(S)-o-cyanophenylalanine methyl ester hydrochloride (4.1 g,
100%); Rf =0.66 (EBAW); mp: 59.8–62.48C; [a]D =ꢀ3.8 (c=1 in
1
CH2Cl2); H NMR (DMSO, 250 MHz) dH =0.83 (brs, 2H), 1.19 (s, 3H),
Molecular modeling studies by using the time-averaged restraints
molecular dynamics (Tar-MD) approach[65] were performed within
the Sander module of AMBER9[66] with the built-in GAFF[67] force
3.04 (d, J=3.7 Hz, 1H), 3.09 (d, J=5.4 Hz, 1H), 3.63 (s, 3H), 7.39–
7.46 (m, 2H), 7.61 (t, J=7.6 Hz 1H), 7.75 ppm (d, J=7.7 Hz, 1H);
13C NMR (DMSO, 63 MHz) dC =24.7, 44.1, 51.9, 58.3, 113. 2, 118.40
2044
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
(ꢀCN), 127.4, 131.5, 132.4, 132.6, 140.5, 176.4 ppm; MS (ES+): m/z
437, 219 [M+H]+, 159; HPLC tR =9.5 min.
combined aqueous phases were evaporated again, redissolved in
H2O (300 mL), and extracted with CH2Cl2 (3ꢅ300 mL). The organic
layer was washed with H2O (150 mL), dried (MgSO4), filtered and
evaporated to yield a white solid (1.3 g, 87%): Rf =0.13 (EtOAc);
Next, a solution of a-methyl-(S)-o-cyanophenylalanine methyl ester
hydrochloride (3.9 g, 17.9 mmol, 1.0 equiv) in HCl (6n, 70 mL) was
stirred for 55 h at 658C and subsequently cooled to room tempera-
ture, which induced partial crystallization. The white crystals were
filtered and rinsed with a minimum amount of cold HCl (6n). The
remaining product was obtained after evaporation under high
pressure (2.8 g, 0.8 g crystals+2.0 g crude; 65%): Rf =0.51 (EBAW);
mp: 192.8–194.18C; [a]D =ꢀ5.0 (c=1 in H2O); 1H NMR (DMSO,
250 MHz) dH =1.46 (s, 3H), 3.34 (d, J=14.2 Hz, 1H), 3.40 (d, J=
14.4 Hz, 1H), 7.48 (d, J=7.4 Hz, 1H), 7.54 (t, J=7.7 Hz, 1H), 7.68 (t,
3J=7.4 Hz, 1H), 7.82 (d, J=7.6 Hz, 1H), 8.85 ppm (brs, 3H);
13C NMR (DMSO, 63 MHz) dC =20.7, 40.0, 59.1, 113.3, 117.8, 128.4,
131.8, 133.1, 137.4, 171.4 ppm; MS (ES+): m/z 205 [M+H]+, 159;
HPLC tR =8.3 min; ee (%) 88.
1
mp: 187.2–188.68C; [a]D = +31.7 (c=1 in CH2Cl2); H NMR (DMSO,
250 MHz) dH =0.99 (s, 3H), 2.52 (d, J=13.2 Hz, 1H), 3.53 (s, 3H),
3.79 (dd, J=14.5, 7.7 Hz, 1H), 3.99 (d, J=13.8 Hz, 1H), 4.58 (d, J=
14.3 Hz, 1H), 7.12–7.26 (m, 4H), 7.64 (s, 1H), 7.80 ppm (d, J=
6.1 Hz, 1H); 13C NMR (DMSO, 63 MHz) dC =25.5, 40.9, 44.0, 51.1,
58.2, 126.7, 127.1, 127.7, 130.2, 137.5, 138.6, 155.8, 174.3 ppm; MS
(ES+): m/z 535, 519, 497, 287, 271, 249 [M+H]+, 174; HPLC tR =
11.4 min.
Moc-(S)-a-MeAba-Gly-OtBu (15): NaH (333 mg, 14.4 mmol,
3.0 equiv) was added to a cooled solution of 4-(S)-methyloxycarbo-
nylamino-4-methyl-1,2,4,5-tetrahydro-2-benzazepin-3-one
(14;
1.2 g, 4.8 mmol, 1.0 equiv) in N,N-dimethylformamide (90 mL), fol-
lowed by the addition after 20 min of tert-butyl-bromoacetate
(3.1 mL, 19.2 mmol, 4.0 equiv). After 20 min the ice-bath was re-
moved and the mixture was stirred for 1 h. The mixture was dilut-
ed with EtOAc (300 mL) and the organic phase was subsequently
extracted with a saturated NaHCO3 solution (3ꢅ230 mL) and H2O
(2ꢅ300 mL), dried (MgSO4), filtered and evaporated. The product
was purified by flash column chromatography (EtOAc/cyclohexane
2:5), to yield a colorless oil (800 mg, 46%): Rf =0.47 (EtOAc); [a]D =
N-Methyloxycarbonyl-a-methyl-(S)-o-cyanophenylalanine
(13):
Methylchloroformate (8.7 mL, 113.0 mmol, 10.0 equiv) was added
to a solution of a-methyl-(S)-o-cyanophenylalanine hydrochloride
(12; 2.7 g, 11.3 mmol, 1.0 equiv) in a NaOH solution (1n, 80 mL).
After being stirred for 1 h, the reaction mixture was cooled by use
of an ice-bath and acidified to pH 2–3 with HCl (6n). The aqueous
phase was extracted with cold CH2Cl2 (4ꢅ140 mL). The combined
organic layers were washed with water (pH 2, 1ꢅ140 mL). The or-
ganic phase was dried (MgSO4), filtered and evaporated. A yellow
oil was obtained and used in the next step without further purifi-
cation (2.1 g, 70%): Rf =0.83 (EBAW); [a]D =ꢀ17.6 (c=1 in CH2Cl2);
1H NMR (DMSO, 250 MHz) dH =1.14 (s, 3H), 3.24 (d, J=13.7 Hz,
1H), 3.49 (d, J=14.3 Hz, 1H), 3.56 (s, 3H), 7.27 (d, J=7.7 Hz, 1H),
7.44 (t, J=7.6 Hz, 1H), 7.64 (t, J=7.6 Hz, 1H), 7.78 ppm (d, J=
7.7 Hz, 1H); 13C NMR (DMSO, 63 MHz) dC =22.3, 38.5, 51.4, 58.7,
113.6, 118.4, 127.6, 132.0, 132.7, 132.9, 140.5, 155.6, 174.9 ppm; MS
(ES+): m/z 809, 547, 285 [M+Na]+, 217, 188; HPLC tR =11.8 min.
1
+23.6 (c=1 in CH2Cl2); H NMR (DMSO, 250 MHz) dH =0.94 (s, 3H),
1.17 (s, 9H), 2.46 (d, J=14.1 Hz, 1H), 3.43 (s, 3H, CH3 Moc), 3.77 (d,
1H, Ha Gly, 2J=16.5 Hz), 3.93 (d, 1H, J=13.9 Hz), 4.03 (d, J=
14.9 Hz, 1H), 4.16 (d, J=16.7 Hz, 1H), 4.94 (d, J=14.8 Hz, 1H),
7.10–7.21 (m, 4H), 7.64 ppm (s, 1H); 13C NMR (DMSO, 63 MHz)
dC =26.2, 27.5, 41.5, 51.1, 51.9, 53.3, 58.9, 80.2, 126.6, 127.8, 129.8,
137.3, 137.7, 155.8, 168.3, 173.3 ppm; MS (ES+): m/z 401, 385, 363
[M+H]+, 307, 289, 261; HPLC tR =15.5 min.
Boc-(S)-a-MeAba-Gly-OH (16): Moc-(S)-a-MeAba-Gly-OtBu (15;
718 mg, 1.98 mmol, 1.0 equiv) was dissolved in HBr (33%) in AcOH
(40 mL) and stirred at 558C for 1.5 h. The reaction mixture was
evaporated and repeatedly redissolved in AcOH and evaporated,
to yield brown oil of HBr·H-(S)-a-MeAba-Gly-OH (quantitative): Rf =
4-(S)-Methyloxycarbonylamino-4-methyl-1,2,4,5-tetrahydro-2-
benzazepin-3-one (14): Pd/C (10%, 40 wt.%, 0.796 g) and an aque-
ous AcOH solution (10%, 5.9 mL, 11.4 mmol, 1.5 equiv) were added
to a suspension of N-Moc-a-methyl-(S)-o-cyanophenylalanine (13;
2.0 g, 7.6 mmol, 1.0 equiv) in EtOH/H2O (3:1, 65 mL). The suspen-
sion was hydrogenated in a Parr apparatus (0.34 mpa, room tem-
perature, 2 days). The mixture was filtered over dicalite and rinsed
with water. After evaporation of the solvent, the residue was crys-
tallized from a minimum amount of hot EtOH and gave N-methyl-
oxycarbonylamino-a-methyl-(S)-o-aminomethylphenylalanine
(1.7 g, 85%): Rf =0.69 (EBAW); mp: 219.2–220.18C; [a]D = +2.4 (c=
1
0.6 (MeOH/CH2Cl2); [a]D =ꢀ7.1 (c=1 in dioxane); H NMR (DMSO,
250 MHz) dH =1.89 (s, 3H), 3.13 (d, J=14.6 Hz, 1H), 3.55 (d, J=
14.6 Hz, 1H), 4.15 (d, J=17.2 Hz, 1H), 4.29 (d, J=17.2 Hz, 1H), 4.56
(d, J=15.1 Hz, 1H), 4.69 (d, J=15.1 Hz, 1H), 7.25–7.41 (m, 4H),
8.43 ppm (s, 3H); 13C NMR (DMSO, 63 MHz) dC =25.0, 38.4, 51.8,
51.9, 59.3, 127.5, 128.5, 129.7, 135.3, 136.9, 170.1, 170.8 ppm; MS
(ES+): m/z 249 [M+H]+, 232, 204, 214; HPLC tR =9.1 min.
1
1 in H2O); H NMR (DMSO, 250 MHz) dH =1.52 (s, 3H), 3.01 (d, J=
Et3N (238 mL, 1.98 mmol, 1.0 equiv) and di-tert-butyldicarbonate
(3.1 mL, 15.8 mmol, 8.0 equiv) were added to a cooled solution of
HBr·NH2-(S)-a-MeAba-Gly-OH (1.98 mmol) in dioxane/H2O (17 mL,
9:1). After 30 min the pH was checked and adjusted to 9, if neces-
sary. The reaction mixture was stirred, overnight, at room tempera-
ture. The solution was evaporated and the residue was redissolved
in H2O (20 mL), followed by adjustment of the pH to 3–4 with
KHSO4aq (10%) solution in the presence of EtOAc (10 mL). The
phases were separated and the aqueous layer was extracted with
EtOAc (3ꢅ10 mL). The combined organic phases were dried
(MgSO4), filtered and evaporated and yielded compound 16
(460 mg, 67%): Rf =0.55 (EtOAc/MeOH 49:49+2% AcOH); mp:
98.6–99.58C; [a]D = +9.7 (c=1 in CH2Cl2); 1H NMR (DMSO,
250 MHz) dH =1.26 (s, 3H), 1.45 (s, 9H), 2.72 (d, J=14.3 Hz, 1H),
3.92–4.17 (m, 3H), 4.79 (d, J=16.9 Hz, 1H), 5.24 (d, J=14.5 Hz, 1H),
5.52 (brs, 1H), 7.24–7.35 ppm (m, 4H); 13C NMR (DMSO, 63 MHz)
dC =26.3, 28.3, 42.3, 53.3, 53.9, 59.9, 127.5, 127.9, 129.0, 130.1,
13.8 Hz, 1H), 3.30 (d, J=13.7 Hz, 1H), 3.57 (s, 3H), 3.80 (d, J=
13.0 Hz, 1H), 4.02 (d, J=13.1 Hz, 1H), 6.65 (brs, 1H), 7.07 (d, J=
6.9 Hz, 1H), 7.16–7.27 (m, 3H), 8.45 ppm (brs, 2H); 13C NMR
(DMSO, 63 MHz) dC =24.8, 36. 7, 40.1, 51.0, 60.5, 126.1, 128.1,
130.6, 131.3, 132.8, 138.4, 154.4, 175.1 ppm; MS (ES+): m/z 289,
267 [M+H]+, 250, 204, 190; HPLC tR =8.1 min.
Next, a solution of N-Moc-a-methyl-(S)-o-aminomethylphenylala-
nine (1.64 g, 6.2 mmol, 1.0 equiv) and pyridine (1 mL, 12.4 mmol,
2.0 equiv) in CH3CN/H2O (3:1, 500 mL) was cooled in an ice-bath
for 10 min, followed by the addition of N-(3-methylaminopropyl)-
N’-ethylcarbodiimide (1.55 g, 8.1 mmol, 1.3 equiv, EDC). After
30 min the ice-bath was removed and the reaction mixture was
stirred, overnight. After evaporation of the solvent, a white solid
was obtained. This residue was redissolved in water (300 mL) and
extracted with CH2Cl2 (4ꢅ300 mL). The organic layer was washed
with H2O (2ꢅ240 mL), dried (MgSO4), filtered and evaporated. The
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2045

D. Tourwꢀ et al.
MED
[19] S. Ballet, A. Frycia, J. Piron, N. N. Chung, P. W. Schiller, P. Kosson, A. W.
Likpowski, D. Tourwꢀ, J. Pept. Res. 2005, 66, 222–230.
[20] K. Van Rompaey, S. Ballet, C. Tomboly, R. De Wachter, K. Vanomme-
slaeghe, M. Biesemans, R. Willem, D. Tourwꢀ, Eur. J. Org. Chem. 2006,
2899–2911.
136.4, 136.8, 156.4, 170.7, 175.1 ppm; MS (ES+): m/z 387, 371, 349
[M+H]+, 293, 249, 235, 218; HPLC tR =15.0 min.
Peptides 5–8
[21] R. De Wachter, L. Brans, S. Ballet, I. Van den Eynde, D. Feytens, A. Ker-
esztes, G. Toth, Z. Urbanczyk-Lipkowska, D. Tourwꢀ, Tetrahedron 2009,
65, 2266–2278.
[22] C. Bisang, C. Weber, J. Inglis, C. A. Schiffer, W. F. van Gunsteren, I. Jelesar-
ov, H. R. Bosshard, J. A. Robinson, J. Am. Chem. Soc. 1995, 117, 7904–
7915.
H-Tyr-d-Ala-(S)-Aba-Gly-Tyr-Pro-Ser-NH2·TFA (5): White powder,
25.0 mg, 45%; Rf =0.58 (EBAW); HRMS-ES+: m/z [M+H]+ calcd for
C41H51N8O10: 816.3801, found: 816.3770; HPLC tR =10.5 min.
H-Tyr-d-Ala-(R)-Aba-Gly-Tyr-Pro-Ser-NH2·TFA (6): White powder,
24.0 mg, 43%; Rf =0.63 (EBAW); HRMS-ES+: m/z [M+H]+ calcd for
C41H51N8O10: 816.3801, found: 816.3811; HPLC; tR =10.0 min.
[23] A. Flores-Ortega, A. I. Jimenez, C. Cativiela, R. Nussinov, C. Aleman, J. Ca-
sanovas, J. Org. Chem. 2008, 73, 3418–3427.
H-Tyr-d-Ala-(S)-MeAba-Gly]-Tyr-Pro-Ser-NH2·TFA
(7):
White
[24] M. J. O’Donnell, W. D. Bennett, S. D. Wu, J. Am. Chem. Soc. 1989, 111,
2353–2355.
[25] M. J. O’Donnell, S. D. Wu, J. C. Huffman, Tetrahedron 1994, 50, 4507–
4518.
powder, 20.4 mg, 3%: Rf =0.43 (EBAW); HRMS-ES+: m/z [M+H]+
calcd for C42H53N8O10: 830.3958, found: 830.3930; HPLC tR =
10.42 min.
[26] E. J. Corey, F. Xu, M. C. Noe, J. Am. Chem. Soc. 1997, 119, 12414–12415.
[27] B. Lygo, P. G. Wainwright, Tetrahedron Lett. 1997, 38, 8595–8598.
[28] T. Ooi, M. Kameda, K. Maruoka, J. Am. Chem. Soc. 1999, 121, 6519–
6520.
[29] E. J. Corey, M. C. Noe, F. Xu, Tetrahedron Lett. 1998, 39, 5347–5350.
[30] B. Lygo, J. Crosby, J. A. Peterson, Tetrahedron Lett. 1999, 40, 1385–1388.
[31] M. J. O’Donnell, S. D. Wu, Tetrahedron: Asymmetry 1992, 3, 591–594.
[32] B. Lygo, J. Crosby, J. A. Peterson, Tetrahedron Lett. 1999, 40, 8671–8674.
[33] R. Chinchilla, C. Najera, F. J. Ortega, Eur. J. Org. Chem. 2007, 6034–6038.
[34] T. Ooi, M. Takeuchi, M. Kameda, K. Maruoka, J. Am. Chem. Soc. 2000,
122, 5228–5229.
[35] K. Maruoka, J. Fluorine Chem. 2001, 112, 95–99.
[36] S. S. Jew, B. S. Jeong, J. H. Lee, M. S. Yoo, Y. J. Lee, B. S. Park, M. G. Kim,
H. G. Park, J. Org. Chem. 2003, 68, 4514–4516.
[37] S. S. Jew, M. S. Yoo, B. S. Jeong, I. Y. Park, H. G. Park, Org. Lett. 2002, 4,
4245–4248.
H-Tyr-d-Ala-(R)-MeAba-Gly]-Tyr-Pro-Ser-NH2·TFA
(8):
White
powder, 10.0 mg, 2%: Rf =0.40. (EBAW); HRMS-ES+: m/z [M+H]+
calcd for C42H53N8O10: 830.3958, found: 830.3955; HPLC tR =
10.73 min.
Acknowledgements
This work was supported by the Research Foundation, Flanders
(FWO-Vlaanderen, Belgium, grant G.0008.08), the I.W.T. (Institute
for the Promotion of Innovation by Science and Technology in
Flanders, SBO-project IWT 040083) and the European grant “Nor-
molife” (LSHC-CT-2006-037733).
[38] M. S. Yoo, B. S. Jeong, J. H. Lee, H. G. Park, S. S. Jew, Org. Lett. 2005, 7,
1129–1131.
[39] M. Kitamura, S. Shirakawa, K. Maruoka, Angew. Chem. 2005, 117, 1573–
1575; Angew. Chem. Int. Ed. 2005, 44, 1549–1551.
Keywords: beta-turn stabilization
ligand design · ligand effects · opioid peptides
· molecular dynamics ·
[40] E. Olajos, A. Pꢀter, R. Casimir, D. Tourwꢀ, Chromatographia 2001, 54,
77–82.
[1] B. L. Kieffer, J. Pain 2000, 1, 45–50.
[41] U. Schçllkopf, W. Hartwig, U. Groth, K. O. Westphalen, Liebigs Ann.
Chem. 1981, 696–708.
[2] S. M. Johnson, W. W. Fleming, Pharmacol. Rev. 1989, 41, 435–488.
[3] V. J. Hruby, R. S. Agnes, Biopolymers 1999, 51, 391–410.
[42] U. Schçllkopf, Tetrahedron 1983, 39, 2085–2091.
[43] R. M. Williams, M. N. Im, J. Am. Chem. Soc. 1991, 113, 9276–9286.
[44] D. Seebach, A. Fadel, Helv. Chim. Acta 1985, 68, 1243–1250.
[45] S. Karady, J. S. Amato, L. M. Weinstock, Tetrahedron Lett. 1984, 25,
4337–4340.
[46] A. B. Smith III, R. C. Holcomb, M. C. Guzman, T. P. Keenan, P. A. Sprengel-
er, R. Hirschmann, Tetrahedron Lett. 1993, 34, 63–66.
[47] F. Alonso, S. G. Davies, A. S. Elend, J. L. Haggitt, J. Chem. Soc. Perkin
Trans. 1 1998, 257–264.
[4] P. C. Montecucchi, R. De Castiglioni, S. Piani, L. Gozzini, V. Erspamer, Int.
J. Pept. Protein Res. 1981, 17, 275–283.
[5] M. Amiche, S. Sagan, A. Mor, A. Delfour, P. Nicolas, Int. J. Pept. Prot. Res.
1988, 32, 506–511.
[6] M. Broccardo, V. Erspamer, G. Falconieri-Erspamer, G. Improta, G. Linari,
P. Melchiorri, P. C. Montecucchi, Br. J. Pharmacol. 1981, 73, 625–631.
[7] M. A. Castiglione-Morelli, F. Lelj, A. Pastore, S. Salvadori, T. Tancredi, R.
Tomatis, E. Trivellone, P. A. Temussi, J. Med. Chem. 1987, 30, 2067–2073.
[8] M. M. Słabicki, M. J. Potrzebowski, G. Bujacz, S. Olejniczak, J. Olczak, J.
Phys. Chem. B 2004, 108, 4535–4545.
[48] K. Nebel, M. Mutter, Tetrahedron 1988, 44, 4793–4796.
[49] S. R. Kapadia, D. M. Spero, M. Eriksson, J. Org. Chem. 2001, 66, 1903–
1905.
[9] A. Pastore, P. A. Temussi, T. Tancredi, S. Salvadori, R. Tomatis, Biopolymers
1984, 23, 2349–2360.
[50] D. M. Coe, R. Perciaccante, P. A. Procopiou, Org. Biomol. Chem. 2003, 1,
1106 – 1111.
[10] P. A. Temussi, D. Picone, T. Tancredi, R. Tomatis, S. Salvadori, M. Marasto-
ni, G. Balboni, FEBS Lett. 1989, 247, 283–288.
[51] S. Vassiliou, P. A. Magriotis, Tetrahedron: Asymmetry 2006, 17, 1754–
1757.
[52] a) S. A. Krumins, Neuropeptides 1987, 9, 93–102; b) M. Amiche, S. Sagan,
A. Mor, A. Delfour, P. Nicolas, Int. J. Pept. Protein Res. 1988, 32, 506–511;
c) M. Amiche, S. Sagan, A. Mor, D. Pelaprat, W. Rostene, A. Delfour, P.
Nicolas, Eur. J. Biochem. 1990, 189, 625–635.
[53] J. R. Maple, U. Dinur, A. T. Hagler, Proc. Natl. Acad. Sci. USA 1988, 85,
5350–5354.
[54] C. Tçmbçly, S. Ballet, D. Feytens, K. E. Kçver, A. Borics, S. Lovas, M. Al-
Khrasani, Z. Fꢄrst, G. Toth, S. Benyhe, D. Tourwꢀ, J. Med. Chem. 2008, 51,
173–177.
[55] A. Madder, N. Farcy, N. G. C. Hosten, H. De Muynck, P. J. De Clercq, J.
Barry, A. P. Davis, Eur. J. Org. Chem. 1999, 2787–2791.
[56] A. Keresztes, M. Szꢄcs, A. Borics, E. K. Kçvꢀr, E. Forrꢁ, F. Fꢄlçp, C. Tçmb-
çly, A. Pꢀter, A. Pꢆhi, G. Fꢆbiꢆn, M. Murꢆnyi, G. Tꢁth, J. Med. Chem.
2008, 51, 4270–4279.
[11] G. Balboni, M. Marastoni, D. Picone, S. Salvadori, T. Tancredi, P. A. Temus-
si, R. Tomatis, Biochem. Biophys. Res. Commun. 1990, 169, 617–622.
[12] T. Tancredi, P. A. Temussi, D. Picone, P. Amodeo, R. Tomatis, S. Salvadori,
M. Marastoni, V. Santagada, G. Balboni, Biopolymers 1991, 31, 751–760.
[13] M. A. Castiglione-Morelli, T. Tancredi, E. Trivellone, G. Balboni, M. Maras-
toni, S. Salvadori, R. Tomatis, P. A. Temussi, Biopolymers 1988, 27, 1353–
1361.
[14] B. C. Wilkes, P. W. Schiller, Biopolymers 1990, 29, 89–95.
[15] M. D. Shenderovich, G. V. Nikiforovich, A. A. Golbraikh, Int. J. Pept. Prot.
Res. 1991, 37, 241–251.
[16] P. Amodeo, A. Motta, T. Tancredi, S. Salvadori, R. Tomatis, D. Picone, G.
Saviano, P. A. Temussi, Pept. Res. 1992, 5, 48–55.
[17] D. Tourwe, K. Verschueren, A. Frycia, P. Davis, F. Porreca, V. J. Hruby, G.
Toth, H. Jaspers, P. Verheyden, G. VanBinst, Biopolymers 1996, 38, 1–12.
[18] D. Tourwꢀ, K. Verschueren, G. VanBinst, P. Davis, F. Porreca, V. J. Hruby,
Bioorg. Med. Chem. Lett. 1992, 2, 1305–1308.
2046
www.chemmedchem.org
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 2035 – 2047

Aba- and a-MeAba-Containing Dermorphin Analogues
[57] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
[58] G. Fꢆbiꢆn, B. Bozꢁ, M. Szikszay, G. Horvꢆth, C. J. Coscia, M. Szucs, J. Phar-
macol. Exp. Ther. 2002, 302, 774–780.
[59] S. Berger, S. Braun, 200 and more NMR experiments, Wiley-VCH, Wein-
heim, 2004, and references therein.
[60] N. Birlirakis, P. Berthault, H. Desvaux, M. Goldman, Bruker Report 1996,
142, 34–38.
[61] W. F. Vranken, W. Boucher, T. J. Stevens, R. H. Fogh, A. Pajon, P. Llinas,
E. L. Ulrich, J. L. Markley, J. Ionides, E. D. Laue, Proteins Struct. Funct.
Bioinf. 2005, 59, 687–696.
16; d) P. M. S. Hendrickx, F. Corzana, S. Depraetere, D. A. Tourwꢀ, K. Au-
gustyns, J. C. Martins, J. Comput. Chem. 2010, 31, 561–572.
˝
[66] D. A. Case, T. A. Darden, I. T. E. Cheatham, C. L. Simmerling, J. Wang, R. E.
Duke, R. Luo, K. M. Merz, D. A. Pearlman, M. Crowley, R. C. Walker, W.
Zhang, B. Wang, S. Hayik, A. Roitberg, G. Seabra, K. F. Wong, F. Paesani,
X. Wu, S. Brozell, V. Tsui, H. Gohlke, L. Yang, C. Tan, J. Mongan, V.
Hornak, G. Cui, P. Beroza, D. H. Mathews, C. Schafmeister, W. S. Ross,
P. A. Kollman, AMBER 9, University of California, San Francisco, 2006.
[67] J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman, D. A. Case, J. Comput.
Chem. 2004, 25, 1157–1174.
[62] D. Neuhaus, M. P. Williamson, The Nuclear Overhauser Effect in Structural
and Conformational Analysis, 2 ed., Wiley-VCH, Weinheim, 2000.
[63] S. Kirkpatrick, C. D. Gelatt, M. P. Vecchi, Science 1983, 220, 671–680.
[68] a) A. Jakalian, B. L. Bush, D. B. Jack, C. I. Bayly, J. Comput. Chem. 2000,
21, 132–146; b) A. Jakalian, D. B. Jack, C. I. Bayly, J. Comput. Chem.
2002, 23, 1623–1641.
ˇ
´
[64] V. Cerny, J. Optimiz. Theory App. 1985, 45, 41–51.
[65] a) A. E. Torda, R. M. Scheek, W. F. van Gunsteren, Chem. Phys. Lett. 1989,
157, 289–294; b) A. P. Nanzer, W. F. van Gunsteren, A. E. Torda, J. Biomol.
NMR 1995, 6, 313–320; c) D. A. Pearlman, J. Biomol. NMR 1994, 4, 1–
Received: June 28, 2011
Published online on August 29, 2011
ChemMedChem 2011, 6, 2035 – 2047
ꢂ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
2047