Role of the sugar moiety on the opioid receptor binding and conformation of a series of enkephalin neoglycopeptides

Article abstract of DOI:10.1016/j.bmc.2017.02.052

Glycosylation by simple sugars is a drug discovery alternative that has been explored with varying success for enhancing the potency and bioavailability of opioid peptides. Long ago we described two O-glycosides having either β-Glucose and β-Galactose of (D-Met², Pro⁵)-enkephalinamide showing one of the highest antinociceptive activities known. Here, we report the resynthesis of these two analogs and the preparation of three novel neoglycopeptide derivatives (α-Mannose, β-Lactose and β-Cellobiose). Binding studies to cloned zebrafish opioid receptors showed very small differences of affinity between the parent compound and the five glycopeptides thus suggesting that the nature of the carbohydrate moiety plays a minor role in determining the binding mode. Indeed, NMR conformational studies, combined with molecular mechanics calculations, indicated that all glycopeptides present the same major conformation either in solution or membrane-like environment. The evidences provided here highlight the relevance for in vivo activity of the conjugating bond between the peptide and sugar moieties in opioid glycopeptides.

Full text of DOI:10.1016/j.bmc.2017.02.052

Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Role of the sugar moiety on the opioid receptor binding and
conformation of a series of enkephalin neoglycopeptides
Mònica Rosa a, Verónica Gonzalez-Nunez b, Katherine Barreto-Valer b, Filipa Marcelo c,
Julia Sánchez-Sánchez d, Luis P. Calle e,1, Juan C. Arévalo d, Raquel E. Rodríguez b, Jesús Jiménez-Barbero e,f,g
,
Gemma Arsequell a, Gregorio Valencia a,
⇑
a Instituto de Química Avanzada de Cataluña (IQAC-CSIC), E-08034 Barcelona, Spain
b Department of Biochemistry and Molecular Biology, Faculty of Medicine, Instituto de Neurociencias de Castilla y León (INCyL), University of Salamanca, Institute of Biomedical
Research of Salamanca (IBSAL), E-37007 Salamanca, Spain
c UCIBIO, REQUIMTE Faculdade Ciências e Tecnologia, Universidade Nova de Lisboa, 2829-516 Caparica, Portugal
d Department of Cell Biology and Pathology, Faculty of Medicine, Instituto de Neurociencias de Castilla y León (INCyL), University of Salamanca, Institute of Biomedical Research
of Salamanca (IBSAL), E-37007 Salamanca, Spain
e CIC bioGUNE, Bizkaia Technological Park, E-48160 Derio, Spain
f Ikerbasque, Basque Foundation for Science, Bilbao E-48013, Spain
g Department of Organic Chemistry II, Faculty of Science and Technology, University of the Basque Country, EHU/UPV, E-48940 Leioa, Bizkaia, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Glycosylation by simple sugars is a drug discovery alternative that has been explored with varying suc-
Received 28 November 2016
Revised 21 February 2017
Accepted 24 February 2017
Available online xxxx
cess for enhancing the potency and bioavailability of opioid peptides. Long ago we described two O-gly-
cosides having either b-Glucose and b-Galactose of (
highest antinociceptive activities known. Here, we report the resynthesis of these two analogs and the
preparation of three novel neoglycopeptide derivatives -Mannose, b-Lactose and b-Cellobiose).
D
-Met2, Pro5)-enkephalinamide showing one of the
(
a
Binding studies to cloned zebrafish opioid receptors showed very small differences of affinity between
the parent compound and the five glycopeptides thus suggesting that the nature of the carbohydrate moi-
ety plays a minor role in determining the binding mode. Indeed, NMR conformational studies, combined
with molecular mechanics calculations, indicated that all glycopeptides present the same major confor-
mation either in solution or membrane-like environment. The evidences provided here highlight the rel-
evance for in vivo activity of the conjugating bond between the peptide and sugar moieties in opioid
glycopeptides.
Keywords:
Neuropeptide
Opioid receptors
Neoglycopeptides
Enkephalin-related
Glycosylation
Pharmacology
Ó 2017 Published by Elsevier Ltd.
1. Introduction
of both selective and non-selective
these peptides have played an important role in defining the phar-
macological effects of combined /d opioid receptor activation,
natural and synthetic opioid peptides have historically been poor
drug candidates, mainly because of their limited ability to cross
the blood-brain barrier (BBB) that impairs their access to the corre-
sponding central nervous system (CNS) sites of action after sys-
temic administration.3
l
/d active peptides.2 Although
Endogenous opioid neurotransmitters such as Met- and
Leu-enkephalin represent one class of opioid receptor ligands that
l
produce relatively non-selective agonist effects at both
l and d
receptors1 but have inspired the design and development of many
Abbreviations: BBB, Blood brain barrier; CNS, Central nervous system; [3H]-DPN,
[3H]-Diprenorphine; DMEM, Dulbecco’s modified Eagle’s medium; ENK, Enkepha-
lin; HRMS, High resolution mass spectrometry; HPLC, High performance liquid
chromatography; icv, intracerebroventricular administration; ip, intraperitoneal
administration; it, intrathecal administration; iv, Intravenous administration; sc,
Subcutaneous administration; Nx, Naloxone; NMR, Nuclear magnetic resonance;
TSP, 2,2,3,3-Tetradeutero-3-trimethylsilylpropionic acid.
In spite of these hurdles, a number of structure-activity studies
on opioid peptides suggest that glycosylation may facilitate biodis-
tribution of these peptides across the BBB and enhance their
potency after iv, ip or sc administration by producing centrally
mediated behavioral effects.4–6 A recent example is MMP2200,
which is a glycosylated derivative of a Leu-enkephalin analog, that
is approximately 10-fold more potent than the parent unglycosy-
lated compound and twice as potent as morphine in producing
antinociception in mice after systemic iv administration.7–9
⇑
Corresponding author.
E-mail address: gregorio.valencia@iqac.csic.es (G. Valencia).
New address: Eli Lilly, Avenida de la Industria, 30, E28108 Alcobendas, Madrid,
1
Spain.
http://dx.doi.org/10.1016/j.bmc.2017.02.052
0968-0896/Ó 2017 Published by Elsevier Ltd.
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052

2
M. Rosa et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
However, when systemically administered in rhesus monkeys
MMP2200 acted as peripheral /d opioid receptor agonist with
glycoopioid results that indicate that disaccharides seem to yield
more dual d and
active compounds,6 we have also prepared the
l
l
limited distribution to the central nervous system, suggesting the
existence of species differences in the pharmacokinetics and BBB
penetration of glycopeptides.10
corresponding O-linked glycopeptide analogs of lactose (b-Lac) 5
and cellobiose (b-Cel) 6. The pharmacological properties of these
new and the two old glycosylated derivatives (3 and 4) were
These and other results on the opioid glycopeptide field are in
good agreement with the more general realization that post-trans-
lational glycosylation of proteins and synthetic glycosylation of
other peptides extend their half-life in serum,11,12 enhance their
potency,13,14 and even influence their biodistribution,15 including
BBB penetration.3,12–14,16–18
assessed by radioligand binding assays on isolated l and d opioid
receptors from zebrafish owing that this organism presents high
receptor opioid homology respect to the human counterparts. A
pilot study of the antinociceptive properties of the Man glycopep-
tide (2) was conducted on the tail-flick test, after ip administration
in mice, to assess the potential permeability of this compound
across the BBB. Finally, in an attempt to correlate their binding
properties to their conformation in solution, NMR conformational
studies combined with molecular mechanics calculations were also
conducted.
Following this principle, a more than 20 years-old search for
systemically active glycoopioids is now still under way and pro-
ducing interesting compounds such as the above mentioned
MMP2200. During this time, the initially difficult attempts to con-
jugate enkephalin analogs to sugar moieties have led to more
amenable synthetic procedures to screen different glycosides of
various complexities, most of them naturally occurring mono-,
di- and trisaccharides. Also, different peptide to carbohydrate con-
jugation methods have been explored that range from native O-
and N-glycosyl bond formation to chemical ligation at suitable
functional groups on a rather limited number of enkephalin
sequences. The analgesic potency of many of these compounds in
mice is comparable to, or sometimes greater than, the potency of
morphine after iv administration.4 Regrettably, up to now, clear
correlations between identity of the sugars, type and site of linkage
to the peptide and enkephalin sequence with biological activity
seem still far away.
2. Results and discussion
2.1. Chemistry
The five neoglycopeptides 2–6 depicted in Fig. 1 and their par-
ent compound 1 were prepared by stepwise manual solid-phase
peptide synthesis following standard Fmoc protocols from previ-
ously synthesized suitable glycosyl amino acid building blocks
(Fig. 1).24–27 The final products were purified (>98% by HPLC) and
characterized by UPLC-TOF/MS. The analytical data from already
reported products 3 and 4 which were previously prepared by
solution phase methods was consistent with the new solid-phase
synthesized materials.
We have contributed to these efforts by providing early exam-
ples of neoglycoenkephalins. In one case, we prepared a couple
of neoglycopeptides by amide-linking a glucosyl-19 and a galacto-
syl-amine residue20 to the C-terminal of the
l
-selective sequence
H-Tyr-DMet-Gly-Phe-Pro-NH2, that produces significant analgesic
activity after systemic administration by acting upon both and
2.2. Biological activity
As pointed out, CNS bioavailability is on the main focus on cur-
rent glycoopioid studies, however, owing the mounting evidence of
l
d receptors.21 The Glc analog was about two orders of magnitude
less potent by peripheral (ip) than by central (icv or it) administra-
tion in rats, but still about 2000 times more potent than morphine.
By central administration (icv) the Gal analog was one order of
magnitude more potent than the Glc analog and about 5000 times
more potent than morphine. These data were obtained after exper-
iments using the tail immersion analgesic test and were in fair
agreement with the ones obtained on the paw pressure test.19,20
In a second case, we synthesized a couple of positional analogs of
the previous glycopeptides by O-glycosyl bond formation of Glc
and Gal with the hydroxyl function of Hyp that substitutes Pro.5
The most striking results were that the Gal analog was about
57000 times more potent than morphine, 1700 times more potent
than the Glc analog and 10000 times more potent than the parent
unglycosylated peptide, as assessed by the tail immersion anal-
gesic test after icv administration in rats.22 More recently, we have
a functional interaction between
l and d opioid receptors and a
possible regulatory role for d agonists, interests in opiod research
found that a a-mannoside of morphine is 100 fold more potent and
twice long lasting as compared to morphine when ip injected to
rats and assessed by the tail-flick and paw pressure analgesic
tests.23
At the time of these early experiments with the glycopeptides,
the molecular and structural characterization of opioid receptors
have not been achieved yet, and cloning as well as the modern
expression techniques were not available either. Thus, to shed light
on the affinity and selectivity of these class of glycopeptides to
bind to individual opioid receptor types and to study the role of
the nature of the sugar moiety on their binding features, we have
resynthesized the Glc and Gal O-linked Hyp5 derivatives (3 and
4) above discussed and expanded these family with new com-
pounds. Thus, owing to the improved pharmacological profile
observed with our mannoside of morphine we have chosen to pro-
duce the enkephalin mannoside derivative 2. Also, following other
Fig. 1. Structures of the peptide (1) and glycopeptides (2–6) prepared in this work.
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052

M. Rosa et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
3
Table 1
Binding affinities, expressed as Ki-values (mean SEM), found for a series of glycopeptides (2–6) and their parent compound (1) when tested on the
l
and d opioid receptors from
zebrafish. Data taken from our previous work: ⁄(Ref. 28), #(Ref. 29) and y(Ref. 30). n.d. not determined.
No. Peptide and glycopeptides and references
Sequence
(DMet2,Pro5)-enkephalinamide parent compound H-Tyr-DMet-Gly-Phe-Pro-NH2
-Man H-Tyr-DMet-Gly-Phe-Hyp(Man)-NH2 239.4 37.0
m receptor KI (nM)
d1a receptor KI (nM) d1b receptor KI (nM)
1
2
3
4
5
6
269.0 40.9
565.8 38.0
608.6 58.4
389.5 118.8
468.2 105.0
634.6 109.7
843.1 116.5
223#
256.4 39.4
536.2 5.3
158.9 18.9
518.5 39.9
559.4 1.1
528.3 16.2
1427y
a
b-
b-
b-Lac
b-Cel
-
D
D
-Glc
-Gal
H-Tyr-DMet-Gly-Phe-Hyp(Glc)-NH2
H-Tyr-DMet-Gly-Phe-Hyp(Gal)-NH2
H-Tyr-DMet-Gly-Phe-Hyp(Lac)-NH2
H-Tyr-DMet-Gly-Phe-Hyp(Cel)-NH2
312.6 67.6
383.3 57.1
202.5 3.8
223.1 24.8
187⁄
D
Morphine
Met-ENK
Leu-ENK
H-Tyr-Gly-Gly-Phe-Met-OH
H-Tyr-Gly-Gly-Phe-Leu-OH
684⁄
n.d.
45y
1317⁄
73
175y
have also shifted to the d opioid receptor. Thus, more contempo-
rary efforts have also focused to produce compounds possessing
dual actions at d and
trum of analgesic efficacy with less side effects as compared with
-selective agonists.31 Accordingly, we have examined opioid
l receptors that may show a broader spec-
l
receptor selectivity in our series of compounds.
The binding affinity of the different glycosylated derivatives for
the
l and d receptors has been assessed by competition binding
assays using [3H]-DPN. The obtained Ki-values are summarized
in Table 1; and the binding affinities of morphine, Met-ENK, and
Leu-ENK have also been included for comparison (See also Supple-
mentary information Figs. S3 and S4). All glycopeptides tested
showed Ki-values on the nanomolar range, and were able to dis-
place up to 100% of the specific binding, thus confirming their opi-
oid nature. Statistical analysis has been performed to determine
whether the differences in the binding affinities for these gly-
coenkephalins are statistically significant (See Supplementary
information Table S5). Disaccharides, especially the Lac derivative
Fig. 2. Antinociceptive actions of morphine and glycopeptide a-D-Man 2 following
ip administration in the tail-flick test. Pre-drug (PD) measurement was taken before
the saline or drug administration, followed up by the four subsequent post-
injection measurements every 30 min (30, 60, 90 and 120 min). Each point
(5), display higher affinities for the
l opioid receptor than the ung-
lycosylated parent compound (1). In contrast, glycosylation with
monosaccharides does not greatly improve the binding of these
represents the mean SEM of 4–6 animals. At 30 min post-injection, GP
2 (5 mg/kg) (⁄) and GP -Man 2 (8 mg/kg) (##) latency was significantly higher
than saline. After 60 min, morphine (§), GP -Man 2 (5 mg/kg) (⁄) and GP
-Man
2 (8 mg/kg) (#) also showed a statistical difference versus saline. At 90 min time
point, only GP
-Man 2 (5 mg/kg) dose (⁄) showed a significant analgesic effect
a-D-Man
analogs to the
l opioid receptor; just the Man derivative (2) yields
a-D
a-
D
a-D
a slight lower Ki-value. In the case of the d receptors, glycosylation
does not improve the binding affinity of the parent compound,
except for the Glc analog (3). Therefore, it seems that disaccharides
shift the binding affinity of the unglycosylated peptide towards the
a-D
compared to saline treatment. To assess the statistical differences (p-value ꢂ 0.05)
between saline and the corresponding treatment, a non-paired, one tailed Student’s
t-test was performed (⁄, #, § < 0.05; ## < 0.01).
l
opioid receptor, while the presence of a single Glc shifts the bind-
ing profile towards the d receptor.
The observed affinities of these glycopeptides are not higher
than those found for longer enkephalin-containing peptides, such
l opioid receptor. This led us
to think that the improvement in binding affinity is not due to
(5 mg/kg) were performed to minimize the use of test animals.
The compounds were ip administered in mice and the antinocicep-
tive effects monitored by the tail-flick test for two hours. As seen in
Fig. 2 and corresponding Table 2, morphine showed a 50% of max-
imum activity while the mannoside (2) a modest 62% at the higher
assayed dose.
as MEGY28 or b-endorphin, for the
the disaccharide per se, but to the fact that bulkier ligands yield
better affinities for the
l
opioid receptor.1 In any event, this simi-
larity among the binding data is puzzling owing that when some of
these products where directly administered into the CNS large dif-
ferences in their antinociceptive properties where observed. Thus,
in our early experiments the Gal analog was more potent than
morphine (57000 times), the Glc analog (1700 times) and the par-
ent unglycosylated peptide (10000 times).22 In an attempt to
rationalize these discrepancies and to better understand the mech-
anisms of action of these products we have carried out a series of
new in vivo and new conformational studies in solution.
To explore the possibility that the glycosyl moieties may be
facilitating the BBB passage of these glycopeptides a pilot study
of their activity after peripheral administration was conducted.
The Man derivative (2) was chosen because, in our hands, an anal-
ogous derivative of morphine was found to be more potent, long
lasting and with less side effects than morphine after ip peripheral
administration.23 A set of experiments with just two doses of the
compound (5 and 8 mg/kg) and a reference dose of morphine
Table 2
Percentage of analgesia in response to different compounds administration calculated
following the formula below included.
Compound
ANALGESIA (%)
Time after injection (min)
30
60
90
120
Saline
Morphine (5 mg/kg)
ꢀ7
30
40
45
3
10
45
50
35
ꢀ8
30
40
45
51
60
62
GP
GP
a
a
-
-
D
-Man 2 (5 mg/kg)
-Man 2 (8 mg/kg)
D
ðPost ꢀ drug score ꢀ ðcontrol scoreÞÞ
% Analgesia ¼
ꢁ 100
ðcut ꢀ off valueÞ ꢀ ðcontrol scoreÞ
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052

4
M. Rosa et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
Macromodel32 as integrated in the Maestro package.33 The global
minimum structure found for 3 is shown in Fig. 3.
This narrow potency difference was not expected from a compound
that has a similar binding affinity than others such as Glc (3) and
Gal (4) which have potencies of various orders of magnitude above
morphine when the BBB is bypassed by central administration.
These findings seem suggestive of a poor BBB permeability of these
O-linked glycopeptides and are in clear opposition to our own pre-
vious results observed with N-linked positional analogs that show
substantial BBB permeability.19,20 These results are reasonable
proof of the critical role of the nature of the linking bond between
the sugar and the peptide moieties.
The rest of the analogs displayed similar features. In addition,
and independently of its chemical nature, in all cases, the sugar
moiety was always observed to move freely and not contacting
the peptide backbone nor the amino acid side chains of these gly-
copeptides. Thus, the presentations of the sugar or the peptide
moieties of these glycopeptides to possible targets are likely to
be unrelated.
3. Conclusions
2.3. Conformational studies in solution
The small differences of binding affinity along this series of gly-
copeptides suggest that the nature of the carbohydrate moiety
plays a minor role in determining the binding mode, as the affini-
ties found for the glycopeptides are of the same order of magnitude
than the parent compound. This has been corroborated by the fact
that no major conformation changes on the peptide backbone
structure were observed by NMR along the series, thus suggesting
that the saccharide part of the molecule does not modulate this
enkephalin receptor recognition. Also, these minor differences in
binding affinities cannot explain the substantial potency changes
found for the in vivo antinociceptive activities that we reported
for the Glc (3) and Gal (4) analogs as well as for the ones of the
Man (2) analog here disclosed. These results suggest that these
O-linked glycopeptides do not easily diffuse through the BBB in
contrast to the N-linked series previously reported by us, thus
highlighting the relevance for activity of the bonding nature
between the peptide and sugar moieties in opioid glycopeptides.
The apparent contradiction seen in some of these glycopeptides
that in spite of having similar binding affinities as morphine dis-
play an unusual antinociceptive activity when centrally adminis-
tered, prompted us to examine the possibility that certain
structural features of some of these glycopeptides may be in the
cause of this phenomena. Thus, NMR-based conformational studies
in solution or membrane model media have been conducted. A
particular aim was to evaluate the effects of the sugar moieties
on the peptide backbone conformation and side chain orientations
in solution.
Careful analysis of NMR data, including 1H chemical shifts
(Table S9 and Figs. S5-S9), coupling constants (Table S10) and
NOEs, did not show any significant differences between enkephali-
namide 1 and its hydroxyproline derivatives 2–6. In fact, no signif-
icant differences were apparent in the NOE fingerprints in the
absence (glycopeptide 3, Fig. S8) and in the presence of SDS mem-
brane-like environment (glycopeptide 3, Fig. S9). This evidence
strongly suggests that the conformation of these peptides in the
membrane-like environment corresponds to the major conforma-
tion existing in water solution, which was deduced by standard
NOE analysis. In particular, the global 3D model was generated
using the key NOE data for all the glycopeptides. In particular,
the following NOEs were employed to build the 3D model shown
4. Materials and methods
4.1. Chemicals
Naloxone (Nx) was purchased from Sigma-Aldrich and [3H]-
diprenorphine ([3H]-DPN) 50 Ci/ mmol from Perkin-Elmer. All
other reagents used were from analytical grade.
in Fig. 3: NH
Hb2
-Met2; NH Gly3 Hb3
Hb Hyp5 and H1 –H Hyp5. Thus, no long-range NOEs were
D
-Met2 - HdTyr1; NH Gly3 – NH Phe4; NH Gly3
–
D
D
-Met2; NH Phe4 – H -Met2; H1 –
a
D
c
4.2. Cell culture
observed, strongly suggesting that these peptides do not adopt a
well defined secondary or tertiary structure in solution or in the
presence of SDS micelles. The measured coupling constants also
displayed medium size values, also in agreement with the above
mentioned lack of a defined structure. From these experimental
data, model 3D structures for the peptides were generated by
Stably transfected HEK293 cells expressing the isolated
l recep-
tor (dre-oprm1, UniProt entry name Q98UH1_DANRE28), the d1a
receptor (dre-oprd1a, UniProt entry name O57585_DANRE29) or
the
d1b
receptor
(dre-oprd1b,
UniProt
entry
name
B3DH72_DANRE30) from zebrafish were used for this study. Cell
lines were maintained in Dulbecco’s modified Eagle’s medium
(DMEM) supplemented with 10% (v/v) fetal calf serum, 2 mM glu-
employing
a
standard conformational search protocol with
tamine, 100 U/mL penicillin, 0.1 mg/mL streptomycin and 250 lg/
mL geneticin (G-418) (all from Gibco-BRL, Life Technologies), at
37 °C in humidified atmosphere containing 5% (v/v) CO2 in a Forma
incubator. Cells were grown to 80% confluence, harvested with
2 mM EDTA in PBS and collected by centrifugation at 500 g. The
cell pellet was frozen at ꢀ80 °C until use.
Hyp 5
Tyr 1
Glc
4.3. Membrane preparation
Gly 3
Cell pellets were resuspended and homogenized with a Potter-
Elvehjem tissue grinder in assay buffer (Tris HCl 50 mM pH 7.4
Phe 4
with protease inhibitors: 0.1 mg/mL bacitracin, 3.3 lM captopril
DMet 2
and protease inhibitor cocktail, all from Sigma-Aldrich), and homo-
genates were centrifuged at 500g for 10 min at 4 °C. The nuclear
pellet was homogenized again, centrifuged and discarded. The
two supernatants were combined, homogenized again with the tis-
sue grinder and the membrane pellet was collected upon centrifu-
gation at 18000g for 1 h at 4 °C. The crude membrane fraction was
Fig. 3. Representative model structure for glycopeptide 3, the major conformation
obtained from NMR-derived data after molecular mechanics calculations.
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052

M. Rosa et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
5
resuspended in ice-cold assay buffer with protease inhibitors and
protein concentration was determined by Bradford (BioRad).
(20 and 60 ms), assisted with 2D-ROESY experiments (200 ms).
The concentration of the natural peptide enkephalin and its glyco-
sylated analogs was set to 2 mM. The pH was adjusted in order to
detect the amide exchangeable protons. Additional NMR experi-
ments were performed in the presence of SDS micelles to mimic
a membrane-like environment. In this case, the NMR samples were
prepared with ca. 1 mM of peptide in 105 mM of SDS in a H2O/D2O
(90:10) solution. These experiments were recorded on the same
spectrometer and at 288 K. The NMR assignments were accom-
plished combining 2D-TOCSY (mixing times of 20, 60 ms) with
2D-NOESY experiments (mixing time of 300 ms). The resonance
of 2,2,3,3-tetradeutero-3-trimethylsilylpropionic acid (TSP) was
used as a chemical shift reference in the 1H NMR experiments
(dTSP = 0 ppm).
4.4. Competition binding assays and data analysis
Radioligand binding was performed as previously described28
7–30 lg protein were incubated with different concentrations of
unlabelled ligand ranging from 0.3 nM to 10 l
M, and using [3H]-
DPN as radioligand. The working concentration for [3H]-DPN was
similar to the affinity constant obtained for the different receptors:
1 nM for the
l
opioid receptor28 and 3.4 nM for both d receptors.34
Reactions were incubated for 1 h in the case of the
l
and d1a recep-
tors and for 4 h for d1b receptor at 25 °C in 250
lL assay buffer.
10 lM Nx was used to determine nonspecific binding. After incu-
bation, the reaction was stopped by adding 4 mL of ice-cold
50 mM Tris HCl buffer pH 7.4, the mixture was rapidly filtrated
using a Brandel Cell Harvester and washed two times onto GF/B
glass-fiber filters that were pre-soaked with 0.2% (v/v)
polyethylenimine for at least 1 h. The filters were placed in scintil-
lation vials and incubated overnight at room temperature in EcoS-
cint A scintillation liquid (National Diagnostics). Radioactivity was
counted using a Beckman Coulter 6500 scintillation counter.
Experiments were performed in duplicate and repeated three
times.
Specific binding was defined as the difference between total
binding and non-specific binding, as measured in presence of
10 lM Nx. Radioligand binding data were analyzed by computer-
assisted non linear regression analysis using GraphPad PRISM soft-
ware (GraphPAd Software Inc., San Diego, U.S.A.), and inhibition
constants1 were obtained for each ligand using Cheng and Prusoff’s
equation, which corrects for the concentration of radioligand used
in each experiment as well as for the affinity of the radioligand for
its binding site (KD).35
4.7. Molecular mechanics calculations
Molecular mechanics were conducted using Macromodel 9.6,
(MacroModel, version 9.6; Schrödinger, LLC: New York, 2008) as
implemented in version 8.5.110 of the Maestro suite32,33 using
OPLS-200536 as the force field. The starting coordinates for confor-
mational search calculations (OPLS-2005 as the force field) were
those obtained after energy minimization. The continuum GB/SA
solvent model37 was employed, and the general PRCG (Polak-
Ribière conjugate gradient) method for energy minimization was
used. A conformational search protocol was then performed, using
the Monte Carlo torsional sampling (MCMM) method, with the
same force field and minimization conditions.
Acknowledgments
This work was supported by a Grant 080530/31/32 from the
Fundació Marató de TV3, Barcelona, Spain. F. M. thanks to FCT-Por-
tugal for the research grant SFRH/BPD/65462/2009. J. C. A. is
funded by the Spanish MINECO Grant (BFU2014-51846-R). Julia
Sánchez-Sánchez acknowledges a contract thanks to the Spanish
MINECO Grant (P.I.: Juan Carlos Arévalo, BFU2014-51846-R). Due
recognition is paid to Prof. Manuel Martín-Lomas for encouraging
his friend and our former mentor Prof. Juan José García Domínguez
to start this work in line of Prof. Raymond A. Dwek coining the Gly-
cobiology term.
4.5. Antinociceptive tail-flick assay
The tail-flick test was performed in male C57Bl6/J mice weigh-
ing 22–28 g. They were housed in a room with controlled temper-
ature on a 12 h light cycle and access to water and food ad libitum.
Prior to the antinociceptive test, habituation was carried out for a
week, always by the same operators, during the following periods
of time: 11:00h–12:00h and 15:00h–17:30 h. Experiments were
always performed during the afternoon timeframe. The Ugo Basile
tail-flick testing apparatus was set up to emit a 129 mW/cm2 infra-
red light intensity with a cut-off time of 12 s (approximately twice
the control response latency).
Mice were divided into four different groups depending on the
drug to be administrated. Thirty and fifteen minutes before the
intraperitoneal injection, two pre-drug measurements which cor-
respond to the basal response were taken. Following treatment, a
single reading for each mouse was performed after 30, 60, 90
and 120 min. The observers were blind to the treatment. To assess
the statistical differences (p-value ꢂ 0.05) between every group, a
non-paired, one tailed Student’s t-test was performed.
A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2017.02.052.
References
1. Mansour A, Hoversten MT, Taylor LP, Watson SJ, Akil H. Brain Res.
1995;700:89–98.
2. Hruby VJ, Mossberg HI. Endogenous peptides for delta opioid receptors and
analogues. In: Chang K-J, Porreca F, Woods JH, eds. The delta receptor. New
York: Marcel Dekker Inc; 2004:159–174.
All animals were housed and bred in an Animal Facility of the
University of Salamanca. Proper measures were taken to reduce
the pain or discomfort of experimental animals. All animal care
and procedures were done in accordance with protocols approved
by the Bioethics Committee of the University of Salamanca and fol-
lowing the European Community guidelines.
3. Egleton RD, Davis TP. NeuroRx. 2005;2:44–53.
4. Bilsky EJ, Egleton RD, Mitchell SA, et al. J Med Chem. 2000;43:2586–2590.
5. Egleton RD, Mitchell SA, Huber JD, Palian MM, Polt R, Davis TP. J Pharmacol Exp
Ther. 2001;299:967–972.
6. Lowery JJ, Yeomans L, Keyari CM, et al. Chem Biol Drug Des. 2007;69:41–47.
7. Elmagbari NO, Egleton RD, Palian MM, et al.
2004;311:290–297.
8. Lowery JJ, Raymond TJ, Giuvelis D, Bidlack JM, Polt R, Bilsky EJ. J Pharmacol Exp
Ther. 2011;336:767–778.
9. Mabrouk OS, Falk T, Sherman SJ, Kennedy RT, Polt R. Neurosci Lett.
2012;531:99–103.
J Pharmacol Exp Ther.
4.6. NMR conformational studies
10. Do Carmo GP, Polt R, Bilsky EJ, Rice KC, Negus S. T. J Pharmacol Exp Ther.
2008;326:939–948.
11. Kontermann RT. Curr Opin Biotechnol. 2011;22:868–876.
12. Sola RJ, Griebenow K. BioDrugs. 2010;24:9–21.
Experiments were recorded in H2O/D2O 90:10 on a Bruker
Avance 500 MHz at 278 K. NMR assignments were accomplished
using standard 2D-TOCSY experiments at different mixing times
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052

6
M. Rosa et al. / Bioorganic & Medicinal Chemistry xxx (2017) xxx–xxx
13. Aldrich JV, McLaughlin JP. Drug Discov Today Technol. 2012;9:e23–e31.
14. Fichna J, Mazur M, Grzywacz D, et al. Bioorg Med Chem Lett.
2013;23:6673–6676.
15. Sinclair AM, Elliott S. J Pharm Sci. 2005;94:1626–1635.
16. Polt R, Dhanasekaran M, Keyari CM. Med Res Rev. 2005;25:557–585.
17. Jones EM, Polt R. Front Chem. 2015;3:40.
18. Lefever M, Li Y, Anglin B, et al. J Med Chem. 2015;58:5728–5741.
19. Torres JL, Reig F, Valencia G, Rodriguez RE, Garcia-Anton JM. Int J Pept Prot Res.
1988;31:474–480.
20. Rodriguez RE, Rodriguez FD, Sacristan MP, et al. Psychopharmacology.
1990;101:222–225.
26. Lee DJ, Kowalczyk R, Muir VJ, Rendle PM, Brimble MA. Carbohydr Res.
2007;342:2628–2634.
27. Taylor CM, Karunaratne CV, Xie N. Glycobiology. 2012;22:757–767.
28. Gonzalez-Nunez V, Jimenez Gonzalez A, Barreto-Valer K, Rodriguez RE. Mol
Med. 2013;19:7–17.
29. Rodriguez RE, Barrallo A, Garcia-Malvar F, McFadyen IJ, Gonzalez-Sarmiento R,
Traynor JR. Neurosci Lett. 2000;288:207–210.
30. Pinal-Seoane N, Martin IR, Gonzalez-Nunez V, de Velasco EM, Alvarez FA,
Sarmiento RG, Rodriguez RE. J Mol Endocrinol. 2006;37:391–403.
31. Schiller PW. Life Sci. 2010;86:598–603.
32. MacroModel, version 9.6. New York: Schrödinger, LLC; 2008.
21. Bajusz S, Ronai AZ, Szekely JI, Graf L, Dunai-Kovacs Z, Berzetei I. FEBS Lett.
1977;76:91–92.
33. Maestro, a powerful all-purpose molecular modeling environment, version
8.52008. New York: Schrödinger, LLC; 2008.
22. Rodriguez RE, Rodriguez FD, Sacristan MP, Torres JL, Valencia G, Garcia Anton
JM. Neurosci Lett. 1989;101:89–94.
34. Gonzalez-Nunez V, Toth G, Rodriguez RE. Peptides. 2007;28:2340–2347.
35. Cheng Y, Prusoff WH. Biochem Pharmacol. 1973;22:3099–3108.
23. Arsequell G, Salvatella M, Valencia G, et al. J Med Chem. 2009;52:2656–2666.
24. Arsequell G, Sàrries N, Valencia G. Tetrahedron Lett. 1995;36:7323–7326.
36. Jorgensen WL, Maxwell DS, Tirado-Rives J.
1996;118:11225–11236.
J
Am Chem Soc.
25. Owens NW, Stetefeld J, Lattová E, Schweizer F.
2010;132:5036–5042.
J
Am Chem Soc.
37. Still WC, Tempczyk A, Hawley RC, Hendrickson T.
1990;112:6127–6129.
J Am Chem Soc.
Please cite this article in press as: Rosa M., et al. Bioorg. Med. Chem. (2017), http://dx.doi.org/10.1016/j.bmc.2017.02.052