Synthesis of 7β-hydroxy-8-ketone opioid derivatives with antagonist activity at mu- and delta-opioid receptors

Article abstract of DOI:10.1016/j.ejmech.2018.02.074

Despite extensive years of research, the direct oxidation of the 7,8-double bond of opioids has so far received little attention and knowledge about the effects of this modification on activity at the different opioid receptors is scarce. We herein report

Full text of DOI:10.1016/j.ejmech.2018.02.074

Accepted Manuscript
Synthesis of 7β-hydroxy-8-ketone opioid derivatives with antagonist activity at mu-
and delta-opioid receptors
Tiina J. Ahonen, Maiju Rinne, Peter Grutschreiber, Kert Mätlik, Mikko Airavaara,
Dieter Schaarschmidt, Heinrich Lang, David Reiss, Henri Xhaard, Claire Gavériaux-
Ruff, Jari Yli-Kauhaluoma, Vânia M. Moreira
PII:
S0223-5234(18)30211-3
10.1016/j.ejmech.2018.02.074
EJMECH 10251
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 November 2017
Revised Date: 31 January 2018
Accepted Date: 22 February 2018
Please cite this article as: T.J. Ahonen, M. Rinne, P. Grutschreiber, K. Mätlik, M. Airavaara, D.
Schaarschmidt, H. Lang, D. Reiss, H. Xhaard, C. Gavériaux-Ruff, J. Yli-Kauhaluoma, Vâ.M. Moreira,
Synthesis of 7β-hydroxy-8-ketone opioid derivatives with antagonist activity at mu- and delta-opioid
receptors, European Journal of Medicinal Chemistry (2018), doi: 10.1016/j.ejmech.2018.02.074.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Synthesis of 7β-hydroxy-8-ketone opioid derivatives with antagonist activity at
mu- and delta-opioid receptors
Tiina J. Ahonen,1 Maiju Rinne,1 Peter Grutschreiber,1 Kert Mätlik,2 Mikko Airavaara,2 Dieter
Schaarschmidt,3,4 Heinrich Lang,3 David Reiss,5 Henri Xhaard,1 Claire Gavériaux-Ruff,5 Jari Yli-
Kauhaluoma,1 Vânia M. Moreira1,6,*
1Drug Research Program, Division of Pharmaceutical Chemistry and Technology, Faculty of
Pharmacy, University of Helsinki, Finland. Email: vania.moreira@helsinki.fi
2Institute of Biotechnology, University of Helsinki, Finland
3Technische Universität Chemnitz, Faculty of Natural Sciences, Institute of Chemistry, Inorganic
Chemistry, 09107 Chemnitz, Germany
4University of Hamburg, Institute of Inorganic and Applied Chemistry, Germany
5Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France; Université de
Strasbourg, Illkirch, France; Centre National de la Recherche Scientifique, UMR7104, Illkirch,
France; Institut National de la Santé et de la Recherche Médicale, U964, Illkirch, France.
6Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, UK
Abstract
Despite extensive years of research, the direct oxidation of the 7,8-double bond of opioids has so far
received little attention and knowledge about the effects of this modification on activity at the
different opioid receptors is scarce. We herein report that potassium permanganate supported on
iron(II) sulfate heptahydrate can be used as a convenient oxidant in the one-step, heterogeneous
conversion of ∆7,8-opioids to the corresponding 7β-hydroxy-8-ketones. Details of the reaction
mechanism are given and the effects of the substituent at position 6 of several opioids on the
reaction outcome is discussed. The opioid hydroxy ketones prepared are antagonists at the mu- and
delta-opioid receptors. Docking simulations and detailed structure-activity analysis revealed that the
presence of the 7β-hydroxy-8-ketone functionality in the prepared compounds can be used to gain
activity towards the delta opioid receptor. The 7β-hydroxy-8-ketones prepared with this method can
also be regarded as versatile intermediates for the synthesis of other opioids of interest.
Keywords: morphine, oxidation, supported permanganate, hydroxy ketone, antagonist, opioid
receptor
1

ACCEPTED MANUSCRIPT
1. Introduction
Pain is the most common symptom for which patients seek medical attention [1–3]. The impact of
chronic pain for the society is dramatic due to the high cost of medical treatment and loss in
productivity. Opioids are a group of well-known pain relieving medicines, of which morphine (1) is
the classical example (Figure 1), used to treat moderate to severe pain conditions such as cancer-
related pain, post-surgery pain, chronic low-back pain and pain associated to certain inflammatory
diseases like osteoarthritis [4–6]. However, the adverse side effects and the development of
tolerance and hyperalgesia, especially in long-term clinical use, hamper the widespread use of
opioids for relieving pain. Opioids act by binding to and activating the opioid receptors (ORs)
which belong to the superfamily of G protein-coupled receptors [7–9]. The effects of opioids such
as morphine, either beneficial (analgesia) or adverse (sedation, respiratory depression, constipation,
addiction), are mediated by the mu opioid receptor (MOR), by means of different downstream
signalling and regulatory pathways. This receptor has two closely related receptor types (~70%
sequence identity) named delta (DOR) and kappa (KOR) receptors.
Figure 1. Structures of common opioids 1-8.
Since the isolation of morphine (1) in 1806 to the complete elucidation of its chemical structure,
concluded in 1927 [10], and the reports on detailed crystal structures of the ORs co-crystallized
with their respective ligands in 2012 [7–9], a plethora of semisynthetic derivatives of morphine (1)
and related opioids has been made aiming at improving the analgesic potential while minimizing the
side effects [11–14]. Consequently, solid structure-activity relationships have been built and it is
generally accepted that morphine-like activity and addiction liability is associated with the presence
of the same stereochemistry as that of morphine (1), a relatively small alkyl group on the tertiary
nitrogen at position 17 and a free aromatic hydroxy group at position 3 [11–14]. Among the many
chemical modifications of opioid compounds, oxidation at position 14 to give 14-hydroxy
derivatives has successfully resulted in an increase in analgesic potency as in oxymorphone (2) and
oxycodone (3), and variation of the alkyl substituents at the nitrogen atom has led to the general
conclusion that phenylethyl groups improve the analgesic activity whereas allyl or
cyclopropylmethyl groups result in compounds such as naloxone (4) and naltrexone (5) that block
the effect of opioids, through antagonistic behavior at the ORs (Figure 1).
2

ACCEPTED MANUSCRIPT
Despite the large number of semisynthetic derivatives of opioid compounds available up to now,
which have mainly resulted from derivatization at positions 3, 6 and 17, chemical modification of
the 7,8-double bond of opioids is seldom reported in the literature. This double bond has been
mostly removed by reduction, in the preparation of some 7,8-dihydro derivatives of interest [15–
20]. Attempts to oxidize it using, for instance, peroxy acids yield side products derived from attack
on the aromatic ring together with the double bond [21,22]. The epoxidation of morphine proved
unsuccessful and a 7β,8β-epoxide could only be obtained from 3-methoxymethylmorphinone, using
hydrogen peroxide in aqueous methanol containing sodium hydroxide, revealing the relevance of
the C6 substituent on the outcome of this oxidation [23,24]. Oxidations with osmium tetroxide
succeeded to give the respective diols [25–28]. However, they were also highly dependent on the
substituent present at C6, with more electron withdrawing substituents giving lower product yields.
Overall, oxidations of opioids are challenging and almost exclusive to the production of 14-hydroxy
derivatives from thebaine (6) and oripavine (7), in route to the preparation of drugs such as
naloxone (4) and naltrexone (5) (Figure 1) [19,29]. Nonetheless, some oxidized products of opioids,
namely 7,8-epoxides, have been suggested to bear less liability to dependence, with activity
comparable to the parent compounds both in vitro and in vivo [23,24]. We therefore envisioned that
the establishment of a new protocol for the oxidation of the 7,8-double bond of opioids would be a
useful means to produce versatile synthetic intermediates for opioid-related syntheses and to tackle
the biological effects that resulted from the oxidation of this particular double bond.
Although the oxidation of codeine (8) (Figure 1) with aqueous permanganate has been reported
[30], to the best of our knowledge, there are no reports available on the use of solid-supported
permanganate-based systems for the oxidation of opioids. These systems are advantageous as they
involve heterogeneous reaction conditions, which are typically milder, allowing good to excellent
product yields and recovery. In the particular case of potassium permanganate, they avoid handling
of this reagent in the presence of water which results in unpractical and tedious work up procedures
[31]. In addition, permanganate-mediated oxidations are usually recognized as eco-friendly because
the co-product formed, manganese dioxide, can be recycled. Solid supports such as molecular
sieves, graphite reagent, polystyrene, alumina, silica gel and copper sulfate pentahydrate have been
previously used for permanganate-based oxidation of alcohols to carbonyl compounds, sulfides to
sulfones, diols to lactones, and oximes to ketones [32–44]. The oxidation of olefins to diketones, α-
hydroxy ketones and epoxides has been described with permanganate-supported on copper sulfate
pentahydrate and iron(III) perchlorate hydrate [45,46]. Herein we report the use of potassium
permanganate supported on iron(II) sulfate heptahydrate for the oxidation of opioids as a
convenient, one-step procedure for the preparation of previously unreported opioid 7β-hydroxy-8-
ketones. Their structure and stereochemistry have been determined with the support of 1D and 2D
nuclear magnetic resonance (NMR) and X-ray diffraction studies. Moreover, the affinity and
activity of the new compounds towards ORs were determined by radioligand binding and receptor
activation assays, and herein we provide novel insights into structure-activity relationships resulting
from chemical modification at positions 7 and 8.
2. Results and discussion
2.1. Chemistry
3

ACCEPTED MANUSCRIPT
Previous research by the authors has shown that potassium permanganate (KMnO4) supported on
Fe(ClO4)3·nH2O successfully converts ∆5-steroids to their corresponding steroidal α-hydroxy
ketones, in high yields [46]. Therefore, we set out to investigate the use of this same system on the
model opioid compound 11, which was prepared according to the synthesis depicted on Scheme 1.
Scheme 1. Synthesis of compound 12. Reagents and conditions: (a) NaHCO3, acetic anhydride, r.t.,
15 min, 99%. (b) NaHCO3, methyl chloroformate, anhydrous CHCl3, 62 °C, 3 h, 79%. (c) Pyridine,
acetic anhydride, 90 °C, 2.5 h, 79%. (d) KMnO4, FeSO4·7H2O, t-BuOH, H2O, CH2Cl2, r.t., 45 min,
61%.
Compound 11 was then oxidized using a mixture of Fe(ClO4)3·nH2O and KMnO4 with
dichloromethane as the solvent, in the presence of water and tert-butanol (t-BuOH), according to
the reaction conditions depicted on Table S1, Entry 1 (Supporting info). However, the reaction
mixture became very dense and quickly dried, and the reaction could not be effected even by
addition of fresh amounts of solvent to the reaction medium. We next used permanganate-supported
on CuSO4·5H2O, according to the reaction conditions in Table S1, Entry 2. This method has been
previously reported for the oxidation of olefins to either α-hydroxy ketones, diketones or epoxides
[45]. This time, the reaction proceeded in a relatively short time and we were able to isolate a single
reaction product, corresponding to the 7β-hydroxy-8-ketone (12) (Scheme 1), however at a low
yield of 52%. Therefore, we set out to screen the reaction with different metal salts including
sulfates, nitrates and acetates in an attempt to improve the yield of desired product.
4

ACCEPTED MANUSCRIPT
All metal sulfates tested were able to complete the reaction to give compound 12 as the single
product with the exception of Ag2SO4, where mixtures of products, with the hydroxy ketone 12 as
the major product, were obtained (Table S1, Entries 3-6). High resolution mass spectrometry
(HRMS) analysis of the crude reaction mixture revealed the presence of the 7,8-diol derivative and
of unreacted starting material. Mixtures of products were also obtained with metal nitrates (Table
S1, Entries 7-8) and with Cu(OAc)2·H2O (Table S1, Entry 9). In the presence of these salts, the
isolation of product 12 was more difficult as t-BuOH was retained in significant amounts in the
crude mixture. The best yield was obtained using FeSO4·7H2O, with mild extension of the reaction
time when compared to CuSO4·5H2O (Table S1, Entry 3). This is the first report on the use of
FeSO4·7H2O supported on KMnO4 for the preparation of α-hydroxy ketones from olefins.
We next investigated which was the minimum amount of KMnO4 that could be used in the presence
of FeSO4·7H2O to give compound 12, as this would most likely lower the amounts of MnO2 formed
at the end of the reaction and could help to improve the isolation procedure and final yield (Table
S1, Entries 10-11). With amounts of oxidant lower than 0.5 g the reaction was slower and could not
be completed. Nonetheless, under the conditions described in Table S1, Entry 12, the reaction
proceeded well with easier isolation of the final product, small compromise in time, and no
compromise in yield. Two final runs were made in the absence of water or t-BuOH, under the best
conditions given in Table S1 (Entry 12). Without t-BuOH, the reaction proceeded much slower and
after 3 hours the expected compound 12 could be isolated, in 45% yield (Table S1, Entry 13).
However, without water (Table S1, Entry 14), we observed decomposition of the starting material
and no product was recovered, even after extension of the reaction time to 3 hours. These
observations show that the presence of water is crucial for the success of the reaction and are
consistent with a role of t-BuOH as a catalyst. 2D NMR experiments (Supporting info, Table S2)
and single crystal X-ray diffraction measurements (Figure 2) confirmed the structure and
stereochemistry of compound 12, and provided important information for elucidating the reaction
mechanism.
5

ACCEPTED MANUSCRIPT
Figure 2. Proposed reaction mechanism for the formation of the 7β-hydroxy-8-ketone 12.
On the newly formed hydroxy ketone 12, the stereochemistry of the chiral carbon 7 is (S) with the
hydroxy group assuming an anti configuration in relation to the C6 acetoxy group, which suggests
that the approximation of the permanganate ion to the double bond occurred on the less hindered
face of the opioid 11 (Figure 2). The oxidation of olefins with permanganate has been extensively
studied and it is generally accepted that the reaction proceeds by formation of a cyclic
manganate(V) intermediate, which is the limiting step (Figure 2) [47].
Scheme 2. Preparation of opioid derivatives 15-19. Reagents and conditions: (a) Compound 13:
TBDPSCl, imidazole, CH2Cl2, r.t., 1 h, quant. Compound 14: TBDMSCl, imidazole, dry DMF, 90
°C, 4 h, 63%. (b) NaHCO3, methyl chloroformate, anhydrous CHCl3, 62 °C, 82% (15), 87% (16),
97% (18). (c) Aq. NaOH sol., MeOH:CHCl3 4:1, r.t., 30 min, 87%. (d) NaH, MeI, dry THF, r.t., 2
h, 49%.
6

ACCEPTED MANUSCRIPT
In reaction media where the basicity is low, disproportionation of this intermediate occurs leading
to the formation of ketols. Systems comprising permanganate supported on metal salts such as the
one in our work have been reported to proceed by Ω-phase catalysis where the water/t-BuOH
mixture forms a third phase in the reaction medium, by surrounding the inorganic solids, where the
reaction actually takes place [45]. We reason that this effect may account for an overall low basicity
in our reaction medium leading the reaction to progress towards the formation of the hydroxy
ketone 12.
To study the effect of the substituent present at C6 on the reaction outcome, we synthesized the
opioid derivatives 15, 16 and 19 (Scheme 2). Compounds 15, 16 and 19 were then oxidized using
the reaction conditions of Table S1, Entry 12 (Scheme 3). The reaction was successful in all opioids
and compounds 20, 21 and 22 were prepared with this method in 36, 34 and 15% yields,
respectively, after chromatographic purification. The electron withdrawing effect of the oxygen
atom is likely to account for the observed difference in the reactivity. This effect is more
pronounced in the methoxy derivative 19 than on the siloxy derivatives 15 and 16. In compound 11,
bearing an acetoxy group, this effect is the least pronounced due to electron delocalization between
the two oxygen atoms. Of note, when the reaction was made using the 6-oxo derivative of
compound 10, we observed degradation of the starting material and no products could be recovered.
As a result of deacetylation of compound 12 with sodium hydroxide, compound 23 was obtained, in
90% yield.
Scheme 3. Synthesis of opioid hydroxy ketones. (a) KMnO4, FeSO4·7H2O, t-BuOH, H2O, CH2Cl2,
r.t., 36% (20), 34% (21), 15% (22). (b) Aq. NaOH sol., MeOH:CHCl3 4:1, r.t., 2.5 h, 90%.
2.2. Pharmacology
We tested the activity of the new opioid hydroxy ketones 12, 22 and 23 and the respective 7,8-
double bond bearing compounds 11, 19 and 17 on the human (h) MOR, DOR and KOR. We started
by determining the binding of the compounds to hMOR, hDOR and hKOR using membrane
preparations obtained from stably transfected cells, by competition binding assay with 1 nM
7

ACCEPTED MANUSCRIPT
[3H]diprenorphine as the radioligand. The KD values for [3H]diprenorphine binding were 0.48 nM to
hMOR, 1.1 nM to hDOR and 0.67 nM to hKOR, as previously reported [48,49]. Receptor
expression levels, determined as Bmax were 3.1, 4.7, and 6.9 pmol/mg of protein for hMOR, hDOR,
and hKOR receptors, respectively.
Table 1. Inhibition of [3H]diprenorphine binding by opioid 7β-hydroxy-8-ketones (12, 22, 23) and
compounds 11, 19 and 17. Values are mean ± SEM from three separate experiments (n=3)
performed in triplicate. Statistical significance (*, p < 0.05) was determined compared to the 0.05%
DMSO control using one-way ANOVA with Dunnett’s multiple comparison test.
Compound
MOR % inhibition
DOR % inhibition
KOR % inhibition
1 µM
1 ± 2
7 ± 1
3 ± 1
5 ± 2
16 ±3
8 ± 3
10 µM
3 ± 3
9 ± 2
11 ± 2
5 ± 3
58 ± 6 *
48 ± 6 *
1 µM
1 ± 2
22 ± 4 *
2 ± 1
5 ± 2
21 ± 1 *
24 ± 2 *
10 µM
9 ± 2
51 ± 2 *
10 ± 1
19 ± 7 *
59 ± 1 *
63 ± 1 *
1 µM
0 ± 3
4 ± 3
6 ± 3
5 ± 1
7 ± 3
6 ± 1
10 µM
1 ± 3
8 ± 1
2 ± 3
0 ± 1
11
12
19
22
17a
23b
27 ± 3 *
21 ± 2 *
a Ki value of 17 on MOR 0.814–1.72 µM and on DOR 1.69–2.63 µM
b Ki value of 23 on MOR 1.08–2.58 µM and on DOR 1.13–1.83 µM
Compounds 17 [File compound17.mol here] and 23 [File compound23.mol here] showed the
highest competition to OR binding among the six compounds tested on all three receptor subtypes.
They were able to block 48-58% of [3H]diprenorphine binding at the hMOR and 59-63% at hDOR,
at 10 µM (Table 1). We further evaluated these two compounds on a concentration range of 0.1 to
50 µM (hMOR) or 0.01 to 50 µM (hDOR) (Figures 3A and B). The Ki values ranged between
0.814–1.72 µM (compound 17) and 1.08–2.58 µM (compound 23) for hMOR, and 1.69–2.63 µM
(compound 17) and 1.13–1.83 µM (compound 23) for hDOR (Table 1). At hKOR the compounds
17 and 23 blocked only 21-27% of radioligand binding at 10 µM, displaying lower binding affinity.
As the binding affinity at hKOR was anticipated to be in the high µM range, the Ki was not
determined in this case. Compound 12 was also quite effective at hDOR with 51% inhibition of
[3H]diprenorphine binding at 10 µM (Table 1), and an estimated Ki value of about 10 µΜ. To a
much lesser extent, compound 22 also showed activity at the hDOR, with 19% inhibition of
[3H]diprenorphine binding at 10 µM. Although it is difficult to judge selectivity based on simple
screening, preliminary observations can be provided based on Table 1. Compounds 17 and 23
appeared equipotent at the hMOR and hDOR, while having a lower affinity at the hKOR.
Compounds 12 and 22 appear more selective for hDOR than for hMOR or hKOR.
We next investigated the agonist or antagonist activities of compounds 17 and 23 using the
[35S]GTPγS binding assay. This assay scores the agonist-evoked GDP/GTP exchange that occurs at
the G-protein within the receptor-G protein complex, and is classically used to quantify receptor
activation. Neither of the compounds acted as agonists in the [35S]GTPγS binding assays on hMOR
8

ACCEPTED MANUSCRIPT
and hDOR (Figures 3C and 3E). Conversely they acted as antagonists of DAMGO-mediated
activation of hMOR at the concentrations tested, as well as antagonists of SNC80-mediated
activation of hDOR at the highest concentration of 50 µM (Figures 3D and 3F).
Figure 3. Binding and functional [35S]GTPγS binding assay of 17 and 23 at hMOR and hDOR. A:
Competition binding curves of 17 and 23 on hMOR. B: Competition binding curves of 17 and 23 on
hDOR. The competition was determined to the general opioid antagonist [3H]diprenorphine. The
values are mean ± SEM from three separate experiments (n=3) performed in triplicate. The results
were analyzed using non-linear regression analysis. C: No MOR activation was detected either with
compound 17 or 23 at 20 or 50 µM concentration, as compared to the MOR agonists DAMGO and
morphine. D: Compounds 17 and 23 inhibited MOR activation by DAMGO (1 µM) at 20 and 50
µM concentrations. E: No DOR activation was detected either with compound 17 or 23 at 20 or 50
9

ACCEPTED MANUSCRIPT
µM concentration, as compared to the DOR agonist SNC80 or to morphine as the initial starting
material of the synthesis. F: Compounds 17 and 23 inhibited DOR activation by SNC80 (1 µM) at
the 50 µM concentration.
2.3 Computational chemistry
2.3.1 The orthosteric binding sites of DOR, MOR, and KOR
The ORs belong to the rhodopsin-like GPCR family characterized by seven α-helical
transmembrane segments. Crystallographic high-resolution structures have been solved for the three
opioid receptor subtypes. The mouse DOR (mDOR) and hKOR have been co-crystallized with
selective antagonists, respectively naltrindole (PDB code 4EJ4) and JDTic (PDB code 4DJH) [8,9].
The mouse MOR has been crystallized with β-FNA (PDB code 4DKL), a covalently bound
irreversible antagonist [7], as well as with BU72 (PDB code 5C1M), a high efficacy agonist [50].
The three-dimensional structures allow a detailed atomistic view of the orthosteric binding site,
formed by a crevice surrounded by the transmembrane segments 3-7, where the ligands are
similarly bound (Figures 4A and 4B). The N-terminus is not part of the binding site in the
antagonist-bound structures, but in the structure of the active MOR forms a lid on top of the binding
cavity and interacts with the co-crystallized ligand [50].
Figure 4. Close-up view of the orthosteric binding cavities for (a, b) co-crystallized ligands and (c,
d) docked compounds 17 and 23 in the superimposed MOR, DOR and KOR X-ray structures. Focus
is on the binding site conserved amino acids (a, c, message site) and varying amino acids (b, d,
10

ACCEPTED MANUSCRIPT
address site). MOR is co-crystallized with β-FNA, DOR with naltrindole and KOR with JDTic.
Selected amino acids are labeled using the Ballesteros-Weinstein convention.
The co-crystallized ligands occupy the same region of the binding pocket and accordingly, the
amino acids at the binding site are highly conserved among the three OR subtypes (Table 2): out of
19 amino acids at close distance to the co-crystallized ligands, 13 are identical and three
conservatively changed. The conserved amino acids are located closer to the intracellular surface of
the binding cavity. They form the so-called message site (Figures 4A and 4C) [51,52], which plays
a role in the efficacy of ligands. Among those, aspartic acid3.32 anchors the positive charge of
ligands through a salt bridge, and methionine3.36 and tryptophan6.48 have hydrophobic contacts.
Although amino acids at the message site are conserved, they do not necessarily form the same
interactions with ligands: for example, tyrosine3.33‘s phenolic group forms hydrogen-bonds to
naltrindole and β-FNA directly at DOR and MOR, but via a bridging water molecule with JDTic at
KOR. At DOR and MOR tyrosine7.43 is at hydrogen-bonding distance of aspartic acid3.32, holding in
place the salt bridge to co-crystallized ligands. In contrast, at the KOR the contact between
tyrosine7.43 and aspartic acid3.32 is not direct but bridged by the ligand JDTic. Polar groups of
ligands are in addition to side-chains participating in hydrogen bond networks that involve water
molecules [7–9].
Table 2. Amino acid composition of the binding pocket of OR subtypes. Address site differences
are in bold. Ballesteros-Weinstein numbering is indicated on the left.
Ballesteros-
Weinstein MOR DOR KOR
2.53
2.56
2.60
2.63
ECL1
3.28
3.29
3.32
3.33
3.36
ECL2
5.39
6.48
6.51
6.52
6.55
6.58
7.35
A117 A98
V108
T120 T101 T111
Q124 Q105 Q115
N127 K108 V118
W133 W114 W124
V143 V124 V134
L125 L135
I144
D147 D128 D138
Y148 Y129 Y139
M151 M132 M142
C217 C198 C210
K233 K214 K227
W293 W274 W287
I296 I277 I290
H297 H278 H291
V300 V281
I294
K303 W284 E297
W318 L300 Y312
11

ACCEPTED MANUSCRIPT
7.39
7.43
I322 I304 I316
Y326 Y308 Y320
In contrast to the message site, the amino acids located at the binding cavity’s extracellular surface
are more variable across the three OR subtypes, and consequently likely to play a significant role in
selectivity. These amino acids have been named at the address site [51,52]. The most variable
positions, differing by size, charge, and polarity, are 2.63 (Lys/Val/Asn), 6.58 (Trp/Glu/Lys) and
7.35 (Leu/Tyr/Trp) (Table 2). Differences at the intracellular surface represent more conservative
changes, for example DOR and MOR have alanine2.53, but KOR has valine2.53. While the co-
crystallized ligands occupy fully the message site, whether and how they would occupy the address
site differs. At DOR and MOR, co-crystallized ligands with a morphine-like skeleton, naltrindole
and β-FNA, are binding towards the transmembrane segments 5-7, in a region surrounded by amino
acids 5.39, 6.58, 7.35. Naltrindole bound to DOR is placing indole in a sub-pocket formed by
lysine5.39, trypthophan6.58 and leucine7.35. In contrast, JDTic binds to KOR by extending towards
transmembrane segment 2 and its carbonyl group directly interacts with tyrosine7.35. A similar
interaction cannot take place at DOR and MOR that have respectively leucine7.35 and
trypthophan7.35, possibly explaining partly its low affinity for these receptors. Another determinant
of the preference of JDTic for the KOR is the small amino acid valine2.63 instead of large and polar
amino acids in other subtypes, lysine2.63 at DOR and asparagine2.63 at MOR.
Docking simulations of compounds 17 [File compound17.mol here] and 23 [File compound23.mol
here] suggest that the morphine-like skeleton roughly occupies the message site (Figure 4C and D,
Supporting info Figure S2). At the MOR [Files complex_compound17_MOR.pse and
complex_compound23_MOR.pse] and DOR [Files complex_compound17_DOR.pse and
complex_compound23_DOR.pse] subtypes (low µM activity in MOR and DOR), this binding mode
is similar to naltrindole/DOR and β-FNA/MOR (Figure 4A and B, Supporting info, Figure S2). In
comparison, at KOR the docking simulations suggest a small (~1.5-2.5 Å) shift around the ligand’s
positions 6 to 8, probably driven by steric clashes with tyrosine7.35 (Supporting information, Figure
S2). The binding modes are overall in agreement with docking studies of the selective antagonists
norbinaltorphimine and 5’-guanidinonaltrindole [8]. Docking simulations of the other four
compounds, 11, 19, and 22 (mostly inactive) as well as 12 (mid-range µM activity at DOR) to the
three subtypes led also to similar binding modes. An exception was found for compounds 11, 12
and 22 at KOR, where the compounds docked closer to TM2 than classical opioids, which may be
due to the larger size of the acetoxy groups in positions 3 and 6 for compounds 11 and 12.
12

ACCEPTED MANUSCRIPT
Figure 5. Binding modes at the DOR of (a) co-crystallized naltrindole, (b) docked compound 17
and (c) docked compound 23. Positions 7 and 8 are shown in green in docked ligands. The 7β-
hydroxy-8-ketone is pointing towards the solvent and facing the hydrophobic wall of the binding
pocket. Amino acids labeled with Ballesteros-Weinstein numbering, lysine5.39, isoleucine6.51
,
valine6.55 and isoleucine7.39
.
2.3.2 Structure-activity relationships
We assessed the binding affinities (competition binding assay) of six compounds for the MOR,
DOR and KOR in comparison to morphine 1, with focus on four positions according to Table 3.
Positions 7 and 8 had either a double bond or were modified into a 7β-hydroxy-8-ketone moiety
whereas positions 3 and 6 had hydroxy, methoxy, or acetoxy substituents. Position 17 had an N-
methyl carbamate in all cases.
Table 3. Overview of the SARs of the studied compounds. Boxes, estimates of Ki ranges from %
binding inhibition. White: high µM, light grey: less than 100 µM, dark grey: less than 10 µM,
black: less than 3 µM.
Comparison of the binding results obtained for the pairs of compounds 17 and 23, 19 and 22, and
11 and 12 demonstrates that introduction of a 7β-hydroxy-8-ketone is associated with a gain in
activity at the DOR, but not at the other OR types. Inhibitory activities at 10 µM increase from 9 to
51% for compounds 11 and 12, from 10 to 19% for compounds 19 and 22, and from 59 to 63 % for
compounds 17 and 23. For these two later pairs the change is of lesser magnitude than for the 11
and 12 pair, but the confidence intervals do not overlap. When these compounds (17, 19, 22, 23) are
docked to the crystal structure of the DOR (as well as in MOR and with the exception at 22 for
KOR), positions 7 and 8 are located within an hydrophobic environment surrounded by amino acids
isoleucine6.51, valine6.55 and isoleucine7.39, while the extracellular region of the binding cavity funnel
13

ACCEPTED MANUSCRIPT
is exposed to the solvent (Figure 5). With the exception of tyrosine7.35‘s phenolic group in KOR,
which is at hydrogen bonding distance from the hydroxy group at position 7, the compound’s 7 and
8 newly introduced positions do not seem to form polar interactions with the receptors. Solvent
effects may thus be at the source of the inhibitory activity observed with introduction of 7β-
hydroxy-8-ketone at DOR, however, these are very difficult to predict. In general, previous work in
the literature does not provide a good framework to rationalize the effect of modifying positions 7
and 8. Reduction of the 7,8 double bond in morphine, leading to dihydromorphine, increases
activity approximately by two fold at the DOR and MOR, but decreases it at the KOR by 8-fold
[53–55]. When comparing dihydrocodeine to codeine (8), the activity is in contrast decreasing at
DOR and KOR (respectively 6- and 2-fold) and increasing 8.5-fold at MOR [56].
Considering the changes at positions 3 and 6, the hydroxy group is clearly preferred compared to
methoxy or acetoxy. This can be seen by comparing the inhibitory activities of compounds 11, 12,
19, and 22 with the activities of compounds 17 and 23, i.e, at MOR, 3-11% vs 48-58%; at DOR 9-
51% vs 59-63%; and to a lesser extent at KOR 0-10 % vs 21-27%. An interesting observation is that
the conserved side chain nitrogen of lysine5.39 is located near (3-4 Å) position 6 in most docked
binding modes at MOR and DOR, where it would potentially form hydrogen bonds that support
favoring of the hydroxy groups. Activities of compounds 17 and 23 at all three subtypes are
unaffected by introduction of 7β-hydroxy-8-ketone, which may be due to this direct hydrogen
bonding of position 6 with lysine5.39. For compounds 11, 12, 19 and 22, methoxy and acetoxy
groups can be tolerated (or their activities even rescued) at the DOR upon introduction of a 7β-
hydroxy-8-ketone, but this mechanism is not clarified by the docking study. Altogether the data
presented is in line with the literature, where replacing both hydroxy groups with methoxy groups at
positions 3 and 6 (e.g. morphine 1 and thebaine 6) leads to decreased binding (90-fold at MOR, 20-
fold at DOR) [57]. If replacing the hydroxy with methoxy occurs solely at position 3 (morphine 1
and codeine 8), the decrease in activity is even more noticeable at MOR and KOR (700-fold and
300-fold, respectively), however lower at DOR (3-fold) [58]. In contrast, modifying position 6
(morphine 1 and heterocodeine) leads to increased binding by 3.6-fold at the MOR, 5-fold at the
DOR, but is unchanged at KOR [17].
Concerning position 17, compound 1 is easily compared to compound 17, which are otherwise
identical At MOR, in 17 the loss of the positive charge and the associated loss of salt bridge with
.
aspartate3.32 led to a 250-fold increase in Ki and significant loss of radioligand binding inhibition,
from 85% at 10 µM for 1 (Ki ~0.004 µM, [54]) to 58 % for 17 (Ki ~1 µM) (Supporting info, Table
S3). At KOR Ki has not been determined, but the fold difference is probably even higher than at
MOR, since morphine (1) displays 80% inhibition of radioligand binding at 10 µM (Ki 0.028 µM)
while compound 17 displays only 27% inhibition. At DOR there is surprisingly only limited loss of
activity associated with the loss of the positive charge, i.e., only 6-fold: morphine has 63%
inhibition (Ki ~0.34 µM) at DOR, while 17 has 59% inhibition (Ki ~2 µM). In addition to charge, in
most binding modes, the carbamate fits well into a hydrophobic pocket lined by amino acids
tryptophan6.48, glycine7.42, tyrosine7.43 and aspartic acid3.32. Altogether N-alkylation should thus be
associated with a large gain in activity. Nonetheless, we were unable to produce N-alkylated
compounds within the time frame of the project that supported this work, and especially regulations
about opioid substances hindered further work in this direction. Of note, charge is not a requirement
14

ACCEPTED MANUSCRIPT
for high activity at ORs. Nanomolar ligands without protonable nitrogen atoms are active at the
KOR such as salvinorin A, which partially occupies the orthosteric binding pocket as demonstrated
by covalent salvinorin A analogs binding at the KOR [8].
Altogether, concerning determinants of selectivity, the presence of tyrosine7.35 in the vicinity of
positions 7 and 8 at KOR may be linked to the lower activity at that receptor in comparison to MOR
and DOR (respectively tryptophan /leucine7.35). Docking simulations suggest that selectivity could
be driven through a small shift in ligand binding mode at KOR relative to the other subtypes (see
above) that would prevent interactions between the hydroxyl group in position 6 and lysine5.39
Glutamate6.58 in KOR (lysine/tryptophan6.58) is ~10 Å further away but may also play a role.
.
3. Conclusion
7β-Hydroxy-8-ketone opioids can be prepared in one-step and easily isolated, under heterogeneous
conditions, using potassium permanganate supported on metal salts. Our study reveals that the most
suitable metal salt for this transformation is iron(II) sulfate heptahydrate and that a number of
opioids, bearing different functional groups, can be converted into the expected hydroxy ketones.
Better yields are obtained with opioids bearing electron-donating groups at position 6.
Mechanistically, the reaction is thought to proceed via formation of a cyclic manganate(V)
intermediate that undergoes disproportionation to give the hydroxy ketone. The compounds act as
antagonists at MOR and DOR. Overall, our work provides valuable tools for those working in the
field of opioids chemistry, as the 7β-hydroxy-8-ketones can be regarded as a good starting point for
further syntheses. Herein we have shown that chemical modification at positions 7 and 8 of opioids
can lead to increased selectivity at the DOR.
4. Experimental
Chemistry. The reagents were obtained from Sigma Aldrich Co., VWR International Oy, or Merck
Reagenzien. For thin layer chromatography (TLC) Silica gel 60 F254 was used. Flash column
chromatography (FCC) was made with a Biotage High-Performance Flash Chromatography Sp4-
system (Uppsala, Sweden) using a 0.1-mm path length flow cell UV detector/recorder module
(fixed wavelength: 254 nm), and 10 g, 25 g or 50 g SNAP cartridges (10–50 mL/min flow rate).
The melting points were measured with Stuart SMP40 automated melting point apparatus and are
uncorrected. IR spectra were obtained using a Vertex 70 (Bruker Optics Inc., MA, USA) FTIR
instrument. The FTIR measurements were made with a horizontal attenuated total reflectance
(ATR) accessory (MIRacle, Pike Technology, Inc, WI, USA). The transmittance spectra were
recorded at a 4 cm-1 resolution between 4000 and 600 cm-1 using the OPUS 5.5 software (Bruker
Optics Inc., MA, USA). NMR spectra were obtained using a Varian Mercury Plus 300 spectrometer
or Bruker Ascend 400 spectrometer, in CDCl3 or CD3OD, with tetramethylsilane (TMS) as the
internal standard. The chemical shifts were reported in parts per million (ppm) and on the
δ scale
from tetramethylsilane (TMS) as an internal standard. The coupling constants J are quoted in Hertz
13
(Hz). If rotamers are observed in the C spectrum, the minor rotamer peaks are labelled with *.
15

ACCEPTED MANUSCRIPT
Exact mass analysis were performed by Waters UPLC-ESI/QTOF-MS using a Synapt G2 HDMS
(Waters, MA, USA) instrument. Purity analyses were executed with Waters Acquity® UPLC system
(Waters, Milford MA, USA) attached to Acquity PDA detector and Waters Synapt G2 HDMS mass
spectrometer (Waters, Milford MA, USA) via an ESI ion source. The narcotic substances were
synthesized under the permission by the Finnish Medicines Agency Fimea (3/2016, 9.3.2016,
Helsinki, Finland).
General procedure for oxidations (Table S1)
The respective starting material (0.5 mmol) was dissolved in dichloromethane. The metal salt and
KMnO4 were ground to fine powder and added to the stirred solution, followed by water and t-
BuOH. The mixture was stirred at room temperature and monitored with TLC. When the reaction
was complete, the reaction mixture was diluted with diethyl ether (10 mL) and stirred for 3 minutes.
The mixture was filtered through a Celite pad and the pad was washed several times with diethyl
ether (150 mL in total). The filtrate was washed with water (30 mL), dried with anhydrous Na2SO4,
filtered and concentrated to dryness.
N-Methyl 3,6-di-acetyl-4,5-epoxy-7-hydroxy-8-oxo-(5α,6α,7β)-morphinan carbamate (12).
Compound 12 was prepared according to the general procedure described above under the reaction
conditions described on Table S1, Entry 12. Colorless solid (137 mg, 61%). The crystals for X-ray
analysis were obtained from methanol, directly from the crude. Mp 190 °C (decomp.); FTIR-ATR
3423 (OH), 1770, 1722, 1693, 1446, 1259, 1190, 1109, 766 cm-1. 1H NMR (CDCl3, 400 MHz):
δ
6.89 (d, J = 8.2 Hz, 1H, H-2), 6.65 (d, J = 8.2 Hz, 1H, H-1), 5.38 (m, 1 H, H-9), 5.15 (d, J = 4.7 Hz,
1H, H-5), 5.08 (m, 1H, H-6), 4.07 (m, 1H, H-16a), 3.75 (brs, 3H, NCOOCH3), 3.57 (d, J = 12.2 Hz,
1H, H-7), 2.89 (m, 4H, H-10a, H-14, H-16b, OH), 2.71 (d, J = 18.9 Hz, 1H, H-10b), 2.31 (s, 3H, 3-
13
OCOCH3), 2.18 (s, 3H, 6-OCOCH3), 1.98 (m, 2 H, 2 × H-15a and H-15b). C NMR (CDCl3, 101
MHz):
δ ppm 205.9 (C8), 205.5*, 170.6 (6-OCOCH3), 168.4 (3- OCOCH3), 155.7 (NCO), 155.5*,
149.2 (C4), 132.4 (C3), 131.9 (C11), 129.3 (C12), 124.4 (C2), 121.0 (C1), 87.5 (C5), 71.4 (C6),
71.1 (C7), 53.1 (NCOOCH3), 50.4 (C14), 50.1*, 47.9 (C9), 41.9 (C13), 37.3*, 37.1 (C16), 35.4*,
35.2 (C15), 29.6 (C10), 29.5*, 21.0 (OCOCH3), 20.7 (OCOCH3).. HRMS calcd for C22H24NO9.
[M+1]+ 446.1451 found 446.1455. Crystal data: C22H23NO9, M = 445.41 g·mol–1, crystal
dimensions 0.25×0.20×0.05 mm, T = 110 K, orthorhombic, P212121, a = 7.2918(4), b = 12.5640(7),
c = 22.0421(16) Å, V = 2019.4(2) Å3, Z = 4, F(000) = 936, ρcalcd = 1.465 g·cm–3, µ = 0.971 mm–1, θ
range = 4.01–71.30°, reflections collected: 7119, independent: 3771 (Rint = 0.0403), R1 = 0.0448,
wR2 = 0.1035 [I>2σ(I)], S = 1.036, Flack parameter = 0.09(18).
N-methyl
3,6-bis[[(1,1-dimethylethyl)diphenylsilyl]oxy]-4,5-epoxy-7-hydroxy-8-oxo-
(5α,6α,7β)-morphinan carbamate (20). Compound 20 was prepared according to the general
procedure described above under the reaction conditions described on Table S1, Entry 12. The
reaction time was 1.5 h. Purification with automated column chromatography by eluting with an n-
heptane/ethyl acetate gradient (10 → 25% EtOAc) gave compound 20 as a white solid (150 mg,
36%). Mp 116 °C (decomp.); FTIR-ATR 1694 (C=O), 1498, 1445, 1262, 1174, 1113, 998, 700 cm-
1. 1H NMR (CDCl3, 400 MHz):
δ ppm 7.96 (d, J = 6.8 Hz, 2H), 7.82 (m, 6H), 7.42 (m, 12H), 6.39
(d, J = 8.2 Hz, 1H, aromatic-H), 6.18 (d, J = 8.2 Hz, 1H, aromatic-H), 5.19 (m, 1H), 4.49 (d, J = 4.5
Hz, 1H), 3.90 (m, 1H), 3.88 (dd, J1 = 11.9 Hz and J2 = 4.6 Hz, 1H), 3.67 (brs, 3H, NCOOCH3), 3.56
16

ACCEPTED MANUSCRIPT
(d, J = 11.9 Hz, 1H), 2.76 (m, 3H), 2.49 (m, 2H), 1.65 (m, 1H), 1.55 (td, J1 = 12.7 Hz and J2 = 5.2
13
Hz, 1H) 1.16 (s, 9H, t-Bu), 1.13 (s, 9H, t-Bu). C NMR (CDCl3, 101 MHz):
δ ppm 207.2 (CO),
206.9*, 155.6*, 155.4 (NCO) 148.4, 138.1, 136.4, 136.4, 135.8, 135.8, 134.9, 134.9, 133.9, 133.5,
132.6, 132.4, 130.2, 130.1, 128.8, 128.0, 127.9, 127.8, 126.5*, 126.4, 122.3, 119.9, 89.4, 73.6, 72.6,
53.0, 50.1, 49.8* 47.9, 41.2, 37.3*, 37.1, 36.0*, 35.7, 29.3, 29.1*, 27.0, 26.8, 19.7, 19.6. HRMS
calcd for C50H55NO7NaSi2. [M+23]+ 860.3415 found 860.3417.
N-methyl
3,6-bis[[(1,1-dimethyl)-t-butylsilyl]oxy]-4,5-epoxy-7-hydroxy-8-oxo-(5α,6α,7β)-
morphinan carbamate (21). Compound 21 was prepared according to the general procedure
described above under the reaction conditions described on Table S1, Entry 12. The reaction time
50 min. Purification with automated column chromatography by eluting with an n-heptane/ethyl
acetate gradient (5 → 20% EtOAc) gave compound 21 as a white solid (100 mg, 34%). Mp 167-169
1
°C; FTIR-ATR 3431 (OH), 1690 (C=O), 1495, 1444, 1254, 1127, 998, 837, 778 cm-1. H NMR
(CDCl3, 400 MHz):
δ ppm 6.69 (d, J = 8.1 Hz, 1H, aromatic-H), 6.48 (d, J = 8.1 Hz, 1H, aromatic-
H), 5.32 (m, 1H), 4.76 (d, J = 4.7 Hz, 1H), 4.02 (m, 2H), 3.74 (brs, 3H, NCOOCH3), 3.43 (d, J =
11.7 Hz, 1H), 2.90 (m, 2H), 2.77 (d, J = 3.3 Hz, 1H), 2.71 (brs, 1H), 2.67 (m, 2H), 1.88 (m, 2H),
0.99 (s, 9H, t-Bu), 0.95 (s, 9H (t-Bu), 0.22 (s, 3H, SiCH3), 0.19 (s, 3H, SiCH3), 0.16 (s, 3 H,
SiCH3), 0.13 (s, 3 H, SiCH3). 13C NMR (CDCl3, 101 MHz):
δ ppm 206.7 (CO), 206.5* 155.7*,
155.6 (NCO), 148.4, 137.7, 128.9, 126.8*, 126.6, 123.6, 120.3, 89.9, 73.6, 72.3, 53.1, 50.4, 50.1*,
48.1, 41.4, 37.5*, 37.3 36.3*, 36.0, 29.4, 29.3* 26.0, 25.9, 18.5, 18.6, -4.2 (SiCH3), -4.4 (SiCH3), -
4.4 (SiCH3), -4.7 (SiCH3). HRMS calcd for C30H48NO7Si2. [M+1]+ 590.2969 found 590.2970.
N-Methyl 4,5-epoxy-7-hydroxy-3,6-dimethoxy-8-oxo-(5α,6α,7β)-morphinan carbamate (22).
Compound 22 was prepared according to the general procedure described above under the reaction
conditions described on Table S1, Entry 12. The reaction time 1 h. Purification with automated
column chromatography by eluting with n-heptane/ethyl acetate (1:5) gave compound 22 as a
colorless solid (29 mg, 15%). Mp 125 °C (decomp.); FTIR-ATR 3396 (OH), 1690 (C=O), 1503,
1
1444, 1274, 1098, 1030, 753, 731 cm-1. H NMR (CDCl3, 400 MHz):
δ ppm 6.75 (d, J = 8.3 Hz,
1H, aromatic-H), 6.57 (d, J = 8.3 Hz, 1H, aromatic-H), 5.35 (m, 1H), 5.07 (d, J = 4.1 Hz, 1H), 4.03
(m, 1H), 3.86 (s, 3H, OCH3), 3.74 (brs, 3H, NCOOCH3), 3.62 (s, 3H, OCH3), 3.52 (m, 2H), 2.91
13
(m, 3H), 2.77 (d, J = 3.5 Hz, 1H), 2.65 (d, J = 18.6 Hz, 1H), 1.92 (m, 2H). C NMR (CDCl3, 101
MHz):
δ ppm 205.8 (CO), 205.6*, 155.6*, 155.4 (NCO), 146.1, 142.6, 128.5, 125.9*, 125.7, 120.6,
115.4, 86.1, 78.9, 72.9, 58.1 (OCH3), 56.6 (OCH3), 53.0 (NCOOCH3), 50.1, 49.8* 48.0, 41.3,
37.3*, 37.2, 35.9*, 35.7, 29.1, 28.9*. HRMS calcd for C20H24NO7. [M+1]+ 390.1553 found
390.1557.
N-Methyl 4,5-epoxy-3,6,7-trihydroxy-8-oxo-(5α,6α,7β)-morphinan carbamate (23). Compound
12 (297 mg, 0.667 mmol) was dissolved in a mixture of methanol: chloroform (4:1, 10 mL). A 0.1
M aqueous solution of NaOH (4.5 mL) was added and the resulting white suspension was stirred at
room temperature for 2.5 h. The organic solvents were evaporated and the pH of the residue was
adjusted to 6 with a 1 M aqueous solution of HCl. The aqueous phase was extracted with a mixture
of chloroform: isopropanol 4:1 (3 × 20 mL) and the combined organic phases were washed with
brine (18 mL), dried with anhydrous Na2SO4, filtered and concentrated in vacuo to give a light
brown solid. The crude product was purified with automated column chromatography by eluting
17

ACCEPTED MANUSCRIPT
with dichloromethane/MeOH (19:1) to give 23 as a colorless solid (216 mg, 90%). Mp 145 °C
(decomp.); FTIR-ATR 1724, 1693, 1649, 1452, 1396, 1236, 1086, 949, 758 cm-1.1H NMR
(Methanol-d4, 400 MHz):
δ ppm 6.70 (d, J = 8.2 Hz, 1H, aromatic-H), 6.53 (d, J = 8.2 Hz, 1H,
aromatic-H), 5.24 (m, 1H), 4.35 (d, J = 4.9 Hz, 1H), 4.01 (m, 1H), 3.99 (dd, J1 = 11.9 Hz and J2 =
4.9 Hz, 1H), 3.73 (d, J = 14.2 Hz, 3H, NCOOCH3), 3.16 (d, J = 11.9 Hz, 1H), 2.89 (m, 3H), 2.62 (d,
J = 18.6 Hz, 1H), 2.04 (td, J1 = 12.9 Hz and J2 = 5.3 Hz, 1H), 1.84 (m, 1H). 13C NMR (Methanol-
d4, 101 MHz):
δ ppm 208.3 (CO), 157.5 (NCO), 157.2* 146.4, 140.5, 130.5, 126.4, 121.7, 119.9,
91.2, 74.5, 71.3, 53.6*, 53.5 (NCOOCH3), 51.4*, 51.2, 49.6, 42.5, 38.6, 38.4*, 36.4, 36.2*, 30.0*,
29.8. HRMS calcd for C18H20NO7. [M+1]+ 362.1240 found 362.1240.
7,8-Didehydro-4,5-epoxy-6-hydroxy-17-methyl-(5α,6α)-morphinan-3-carboxylate
(9).
Morphine (1, 0.500 g; 1.75 mmol) was added to a stirred mixture of NaHCO3 (8.00 g; 95.3 mmol)
and water (50 mL), at room temperature. Acetic anhydride (4 × 0.83 mL; 8.76 mmol) was added
slowly in 4 portions in 10-minute intervals and after the last addition the stirring was continued for
15 min. The reaction mixture was extracted with dichloromethane (3 × 60 mL), washed with brine
(40 mL) and dried with anhydrous Na2SO4, filtered and concentrated in vacuo to give 9 as light
brown foam (643 mg, 99%), which was used without further purification. FTIR-ATR: 3502 (OH),
1760 (COOCH3), 1614, 1211, 1193, 1033, 941, 784 cm-1. 1H NMR (CDCl3, 300 MHz):
δ ppm 6.73
(d, J = 8.2 Hz, 1H, aromatic-H), 6.59 (d, J = 8.2 Hz, 1H, aromatic-H), 5.74 (d, J = 10.1 Hz, 1H),
5.28 (d, J = 10.1 Hz, 1H), 4.91 (d, J = 6.9 Hz, 1H), 4.15 (brs, 1H), 3.35 (m, 1H), 3.24 (brs, 1H),
3.05 (d, J = 18.9 Hz, 1H), 2.69 (s, 1H), 2.59 (dd, J1 = 12.2 and J2 = 4.5 Hz, 1H), 2.43 (s, 3H,
NCH3), 2.35 (dd, J1 = 14.4 and J2 = 4.9 Hz, 2H), 2.28 (s, 3H, OCOCH3), 2.07 (td, J1 = 12.4 and J2 =
4.9 Hz, 1H), 1.90 (m, 1H). 13C NMR (CDCl3, 75 MHz):
δ ppm 168.5 (OCOCH3), 148.7, 134.2,
132.8, 132.3, 131.8, 127.7, 121.0, 119.9, 92.3, 65.9, 58.9, 46.4, 43.0, 42.6, 40.4, 35.2, 20.8, 20.7.
HRMS calcd. for C19H22NO4. [M+1]+ 328.1549 found 328.1549.
N-Methyl 3-acetyl-7,8-didehydro-4,5-epoxy-6-hydroxy-(5α,6α)-morphinan carbamate (10).
Compound 9 (638 mg; 1.95 mmol) was dissolved in dry chloroform (60 mL) under argon. NaHCO3
(2.46 g; 29.2 mmol) and methyl chloroformate (2.56 mL; 33.1 mmol) were added and the mixture
was refluxed at 62 °C for 3 hours. The reaction mixture was cooled to room temperature, diluted
with dichloromethane (60 mL) and washed with water (40 mL). The aqueous phase was extracted
with dichloromethane (3 × 35 mL). The organic phases were combined, washed with brine (40 mL),
dried with anhydrous Na2SO4, filtered and concentrated in vacuo. The crude product was purified
by automated column chromatography by eluting with a gradient of n-hexane and ethyl acetate (0
→ 100% EtOAc) to give compound 10 as a white solid (574 mg, 79%). Mp 67 °C (decomp.);
FTIR-ATR: 3515 (R-OH), 1751 (COOCH3), 1689 (NC=O), 1442, 1213, 1128, 1064, 939, 781 cm-1.
1H NMR (CDCl3, 300 MHz):
δ ppm 6.77 (d, J = 8.2 Hz, 1H, aromatic-H), 6.61 (d, J = 8.2 Hz, 1H,
aromatic-H), 5.81 (d, J = 10.1 Hz, 1H), 5.29 (brs, 1H), 4.91 (m, 1H), 4.91 (dd, J1 = 7.0 and J2 = 1.1
Hz, 1H), 4.08 (m, 2H), 3.74 (br s, 3H, NCOOCH3), 3.10 (m, 2H), 2.87 (dd, J1 = 19.5 and J2 =5.8
Hz, 1H), 2.74 (d, J = 18.8 Hz,, 1H), 2.56 (m, 1H), 2.29 (s, 3H, OCOCH3), 1.94 (m, 2H). 13C NMR
(CDCl3, 75 MHz):
δ ppm 168.7 (OCOCH3), 156.0 (NCO), 155.7* (NCO), 149.0, 135.5*, 135.3,
132.3, 131.7, 131.6*, 131.3, 126.5, 126.4*, 121.8, 120.4, 92.3, 65.7, 52.9, 50.7*, 50.4, 43.2, 39.6*,
39.5, 37.6, 35.1, 34.8, 30.0*, 29.8, 20.9 (OCOCH3). HRMS calcd. for C20H22NO6. [M+1]+ 372.1447
found 372.1447.
18

ACCEPTED MANUSCRIPT
N-Methyl 3,6-diacetyl-7,8-didehydro-4,5-epoxy(5α,6α)-morphinan carbamate (11). Compound
10 (0.100 g, 0.269 mmol) was dissolved in pyridine (0.62 mL) and acetic anhydride (0.11 mL, 1.13
mmol) was slowly added. The reaction mixture was stirred at 90 °C for 2.5 h, after which the
solvent was co-evaporated with toluene. The residue was dissolved in dichloromethane (30 mL) and
washed with aqueous saturated NaHCO3 solution (10 mL). The aqueous phase was extracted with
dichloromethane (3 × 10 mL), the organic phases were combined and washed with brine (8 mL),
dried with anhydrous Na2SO4, filtered and concentrated in vacuo to give brown oil. The crude
product was purified with automated column chromatography by eluting with an n-hexane/ethyl
acetate gradient (3:2 to 1:1) to give 11 as a white solid (88.0 mg, 79%). Mp 181-182 °C; FTIR-ATR
1684, 1587, 1313, 1217, 1175, 1067, 831, 754 cm-1. 1H NMR (CDCl3, 300 MHz):
δ ppm 6.80 (d, J
= 8.2 Hz, 1 H, aromatic-H), 6.59 (d, J = 8.2 Hz, 1 H, aromatic-H), 5.69 (d, J = 10.1 Hz, 1 H), 5.45
(d, J = 8.9 Hz, 1 H), 5.11 (m, 2 H), 4.89 (m, 1 H), 4.06 (m, 1H), 3.74 (br s, 3 H, NCOOCH3), 2.91
(m, 2 H), 2.75 (d, J = 18.8 Hz, 1 H), 2.61 (m, 1H), 2.27 (s, 3 H, OCOCH3), 2.13 (s, 3 H, OCOCH3),
1.92 (m, 2 H). 13C NMR (CDCl3, 75 MHz):
δ ppm 170.5 (OCOCH3), 168.4 (OCOCH3), 156.0
(NCO), 149.6, 132.2, 131.1, 131.0*, 130.4, 129.8*, 129.5, 128.2, 128.1*, 122.6, 119.9, 88.5, 67.7,
52.9, 50.7*, 50.4, 43.2, 39.7*, 39.6, 37.8, 34.9, 34.7*, 29.9*, 29.7, 20.7 (OCOCH3), 20.7
(OCOCH3). HRMS calcd for C22H24NO7. [M+1]+ 414.1553 found 414.1554.
7,8-Didehydro-3,6-bis[[(1,1-dimethylethyl)diphenylsilyl]oxy]-4,5-epoxy-17-methyl-(5α,6α)-
1
morphinan (13). Compound 13 was prepared according to the literature [59]. H NMR (CDCl3,
300 MHz): δ ppm 7.85 (t, J = 6 Hz, 4H), 7.75 (dd, J1 = 24 and J2 = 7.2 Hz, 4H), 7.40 (m, 12H), 6.26
(d, J = 8 Hz, 1H, aromatic-H), 6.12 (d, J = 8 Hz, 1H, aromatic-H), 5.79 (d, J = 10.7 Hz, 1H), 5.15
(d, J = 9.5 Hz, 1H), 4.52 (d, J = 5.6 Hz, 1H), 4.14 (brs, 1H), 3.22 (brs, 1H), 2.89 (d, J = 18.5 Hz,
1H), 2.45 (m, 1H), 2.35 (s, 3H, NCH3), 2.16 (d, J = 19.1 Hz, 1H), 1.72 (brs, 2H), 1.12 (d, J = 3.3
Hz, 18H, 2 × t-Bu). HRMS calculated for C49H56NO3Si2. [M+1]+ 732.3799 found 762.3803.
7,8-Didehydro-3,6-bis[[(1,1-dimethyl)-t-butylsilyl]oxy]-4,5-epoxy-17-methyl-(5α,6α)-
morphinan (14). Morphine 1 (0.600 g; 2.10 mmol) and imidazole (726 mg; 10.7 mmol) were
dissolved in dry dimethylformamide (3.6 mL) under argon. t-Butyldimethylsilyl chloride (1.20 g;
7.96 mmol) was added and the reaction mixture was heated at 90 °C for 4 hours. The reaction
mixture was diluted with dichloromethane (40 mL), washed with water (2 × 20 mL), dried with
anhydrous Na2SO4 and concentrated in vacuo. The crude product was purified using automated
column chromatography by eluting with a chloroform → methanol gradient (0 → 4%) to give
compound 14 as light brown solid (677 mg, 63%). Mp 117-119 °C ; FTIR-ATR 1494, 1444, 1253,
1
1124, 1033, 979, 835, 773 cm-1. H NMR (CDCl3, 300 MHz):
δ ppm 6.56 (d, J = 8.0 Hz, 1H,
aromatic-H), 6.41 (d, J = 8.0 Hz, 1H, aromatic-H), 5.59 (m, 1H), 5.23 (m, 1H), 4.66 (dd, J1 = 5.7
Hz, J2 = 1.4 Hz 1H), 4.21 (m, 1H), 3.34 (dd, J = 6.4, 3.2 Hz, 1H), 3.02 (d, J = 18.6 Hz, 1H), 2.59
(m, 2H), 2.44 (m, 1H), 2.44 (s, 3H, NCH3), 2.30 (dd, J = 18.6, 6.4 Hz, 1H), 2.02 (td, J = 12.3, 5.3
Hz, 1H), 1.89 – 1.80 (m, 1H), 0.97 (s, 9H, t-Bu), 0.93 (s, 9H, t-Bu), 0.21 (s, 3H, SiCH3), 0.14 (s,
13
3H, SiCH3), 0.13 (s, 3H, SiCH3), 0.10 (s, 3H, SiCH3). C NMR (CDCl3, 75 MHz):
δ ppm 150.2,
137.2, 133.9, 131.3, 128.5, 127.6, 121.4, 118.6, 92.9, 69.6, 59.0, 46.6, 44.2, 43.3, 41.6, 36.2, 26.1,
26.0, 21.0, 18.5, 18.4, -4.2, -4.5, -4.6, -4.7. HRMS calcd. for C29H47NO3Si2. [M+1]+ 514.3173
found 514.3173.
19

ACCEPTED MANUSCRIPT
N-Methyl
7,8-didehydro-3,6-bis[[(1,1-dimethylethyl)diphenylsilyl]oxy]-4,5-epoxy-(5α,6α)-
morphinan carbamate (15). Compound 13 (0.760 g; 0.997 mmol) was dissolved in dry
chloroform (32 mL) under argon. NaHCO3 (1257 mg; 15.0 mmol) and methyl chloroformate (1.31
mL; 17.0 mmol) were added and the mixture was refluxed at 62 °C for 2 h 20 min. The reaction
mixture was diluted with dichloromethane (20 mL) and washed with water (35 mL). The aqueous
phase was extracted with dichloromethane (2 × 20 mL). The organic phases were combined,
washed with brine (30 mL), dried with anhydrous Na2SO4, filtered and concentrated in vacuo. The
crude product was purified by automated column chromatography by eluting with n-heptane/ethyl
acetate (5:1) to give compound 15 as a colorless solid (659 mg, 82%). Mp 75 °C (decomp.); FTIR-
1
ATR 1697 (C=O), 1498, 1448, 1324, 1269, 1172, 1108, 1087, 981, 825, 698 cm-1. H NMR
(CDCl3, 400 MHz):
δ ppm 7.84 (m, 4H), 7.78 (m, 2H), 7.70 (m, 2H), 7.40 (m, 12H), 6.30 (d, J =
8.1 Hz, 1H, aromatic-H), 6.13 (d, J = 8.2 Hz, 1H, aromatic-H), 5.83 (d, J = 9.8 Hz, 1H), 5.15 (m,
1H), 4.75 (m, 1 H), 4.52 (d, J = 5.5 Hz, 1H), 4.15 (m, 1H), 3.94 (m, 1H), 3.67 (br s, 3H,
NCOOCH3), 2.96 (m, 1H), 2.75 (m, 1H), 2.59 (d, J = 18.6 Hz, 1H), 2.24 (brs, 1H), 1.63 (m, 2H),
13
1.13 (s, 9H, t-Bu), 1.10 (s, 9H, t-Bu). C NMR (CDCl3, 101 MHz):
δ ppm 155.8 (NCO), 155.5*,
150.2, 137.9, 136.1, 135.9, 135.9, 135.8, 134.4, 134.1, 133.9, 133.6, 133.3, 130.6, 130.0, 129.9,
129.9, 129.8, 127.9, 127.8, 127.8, 127.8, 126.3, 126.1, 120.7, 118.7, 92.4, 69.8, 52.8, 50.4*, 50.1,
44.6, 40.4*, 40.2, 37.4, 35.7, 35.4*, 30.0*, 29.8, 27.0, 26.9, 19.7, 19.4. HRMS calcd. for
C50H55NO5NaSi2. [M+23]+ 828.3516 found 828.3515.
N-Methyl 7,8-didehydro-3,6-bis[[(1,1-dimethyl)-t-butylsilyl]oxy]-4,5-epoxy-(5α,6α)-morphinan
carbamate (16). Compound 14 (538 mg; 1.05 mmol) was dissolved in dry chloroform (25 mL)
under argon. NaHCO3 (1.32 g; 15.7 mmol) and methyl chloroformate (1.37 mL; 17.8 mmol) were
added and the reaction mixture was refluxed at 62 °C for 4 hours. The reaction mixture was diluted
with dichloromethane (60 mL) and washed with water (40 mL). The aqueous phase was extracted
with dichloromethane (2 × 50 mL), the organic phases were washed with brine (40 mL), dried with
anhydrous Na2SO4 and concentrated in vacuo. The crude product was purified by automated
column chromatography by eluting with n-hexane/ethyl acetate (5:1) to give compound 16 as a pale
yellow solid (508 mg, 87%). Mp 160-162 °C; FTIR-ATR 1697 (C=O), 1496, 1444, 1249, 1128,
1
1091, 979, 838, 775 cm-1. H NMR (CDCl3, 300 MHz):
δ ppm 6.59 (d, J = 8.1 Hz, 1H, aromatic-
H), 6.41 (d, J = 8.1 Hz, 1H, aromatic-H), 5.65 (d, J = 9.7 Hz, 1H), 5.23 (m, 1H), 4.87 (m, 1H), 4.66
(dd, J = 5.7, 1.4 Hz, 1H), 4.07 (m, 2H), 3.73 (brs, 3H, NCOOCH3), 3.06 (m, 1H), 2.86 (m, 1H),
2.71 (d, J = 18.6 Hz, 1H), 2.47 (brs, 1H), 1.89 (m, 2H), 0.97 (s, 9H, t-Bu), 0.93 (s, 9H, t-Bu), 0.21
(s, 3H, SiCH3), 0.14 (s, 3H, SiCH3), 0.12 (s, 3H, SiCH3), 0.10 (s, 3H, SiCH3) ppm. 13C NMR
(CDCl3, 75 MHz):
δ ppm 156.0 (NCO), 155.6*, 150.3, 137.5, 135.0*, 134.7, 130.3, 127.4, 127.3*,
126.4, 126.2*, 121.9, 119.0, 92.8, 69.3, 52.8, 50.6*, 50.4, 44.7, 40.6*, 40.4, 37.5, 36.0, 35.7*,
30.1*, 30.0, 26.1, 26.0, -4.2, -4.5, -4.6, -4.7. HRMS calcd. for C30H47NO5Si2Na. [M+1]+ 580.2890
found 580.2889.
N-Methyl 7,8-didehydro-4,5-epoxy-3,6-dihydroxy-(5α,6α)-morphinan carbamate (17).
Compound 11 (127 mg, 0.307 mmol) was dissolved in a mixture of methanol: chloroform (7:2, 4.5
mL). A 0.1 M aqueous solution of NaOH (2.2 mL) was added to the solution and the resulting white
suspension was stirred at room temperature for 30 min. The organic solvents were evaporated,
water (5 mL) was added to the residue, and the pH of the aqueous phase was adjusted to 5 with a 1
20

ACCEPTED MANUSCRIPT
M aqueous solution of HCl. The aqueous phase was extracted with dichloromethane (3 × 30 mL)
and the combined organic phases were washed with brine (15 mL), dried with anhydrous Na2SO4,
filtered and concentrated in vacuo to give yellowish oil. The crude product was purified with
automated column chromatography by eluting with n-heptane/ethyl acetate (1:2) to give 17 as a
colorless solid (87.9 mg, 87%). Mp 88 °C (decomp.); FTIR-ATR 3283, 1666, 1450, 1232, 1126,
785, 725 cm-1. 1H NMR (CDCl3, 400 MHz):
δ ppm 6.67 (d, J = 8.1 Hz, 1 H, aromatic-H), 6.49 (d, J
= 8.1 Hz, 1 H, aromatic-H), 5.71 (m, 1 H), 5.26 (d, J = 9.8 Hz, 1 H), 5.87 (m, 2 H), 4.21 (m, 1 H),
4.06 (m, 1H), 3.73 (br s, 3 H, NCOOCH3), 3.04 (m, 1 H), 2.84 (m, 1 H), 2.71 (d, J = 18.6 Hz, 1 H),
13
2.51 (m, 1H), 1.92 (m, 2 H). C NMR (CDCl3, 101 MHz):
δ ppm 156.1 (NCO), 155.8*, 145.5,
138.3, 133.9*, 133.6, 129.9, 127.6, 127.3*, 125.5, 125.3*, 120.5, 117.5, 91.5, 66.5, 53.0, 50.7*,
50.4, 43.7, 40.0*, 39.8, 37.7, 35.5, 35.2*, 29.8*, 29.6. HRMS calcd for C18H20NO5. [M+1]+
330.1341 found 330.1342.
N-Methyl 7,8-didehydro-4,5-epoxy-6-hydroxy-3-methoxy-(5α,6α)-morphinan carbamate (18).
Codeine (8, 0.500 g; 1.67 mmol) was dissolved in dry chloroform (42 mL) under argon. NaHCO3
(2.11 g; 25.1 mmol) and methyl chloroformate (2.19 mL; 28.4 mmol) were added and the mixture
was refluxed at 62 °C for 2 h 15 min. The reaction mixture was cooled to room temperature, diluted
with dichloromethane (60 mL) and washed with water (30 mL). The aqueous phase was extracted
with dichloromethane (3 × 20 mL). The organic phases were combined, washed with brine (35 mL),
dried with anhydrous Na2SO4, filtered and concentrated in vacuo. The crude product was purified
by automated column chromatography by eluting with n-hexane and ethyl acetate (1:1) to give
compound 18 as a white solid (555 mg, 97%). Mp 120 °C (decomp.); FTIR-ATR 3439 (OH), 1688
(C=O), 1503, 1443, 1130, 1052, 942, 784 cm-1. 1H NMR (CDCl3, 300 MHz):
δ ppm 6.69 (d, J = 8.1
Hz, 1H, aromatic-H), 6.57 (d, J = 8.1 Hz, 1H, aromatic-H), 5.77 (d, J = 9.9 Hz, 1H), 5.29 (m, 1H),
4.88 (m, 1H), 4.88 (dd, J1 = 6.6 Hz and J2 = 1.2 Hz, 1H), 4.08 (m, 2H), 3.85 (s, 3H, OCH3), 3.74
(brs, 3H, NCOOCH3), 3.02 (m, 1H), 2.86 (m, 2H), 2.73 (d, J = 18.6 Hz, 1H), 2.53 (m, 1H), 1.94 (m,
2H). 13C NMR (CDCl3, 75 MHz):
δ ppm 156.0 (NCO), 155.7*, 146.6, 142.7, 134.7*, 134.5, 130.1,
127.1, 126.9*, 126.0, 125.9*, 120.2, 113.7, 91.2, 66.21, 56.55, 52.9, 50.7*, 50.5, 43.5, 39.9*, 39.8,
37.7, 35.7, 35.4*, 29.7*, 29.6. HRMS calcd for C19H22NO5. [M+1]+ 344.1498 found 344.1498.
N-Methyl 7,8-didehydro-4,5-epoxy-3,6-dimethoxy-(5α,6α)-morphinan carbamate (19). A 60%
dispersion of NaH in mineral oil (175 mg, 4.37 mmol) was suspended in dry THF (20 mL) under
argon. A solution of compound 18 (0.500 g, 1.46 mmol) in dry tetrahydrofuran (20 mL) was added
to the stirred suspension over 90 min, after which the stirring was continued for additional 70 min.
Iodomethane (413 mg, 2.19 mmol) was added and the mixture was stirred for 2 h. The reaction was
quenched by slow addition of a 25 wt-% solution of NaOMe in methanol (2.0 mL, 8.7 mmol). The
mixture was diluted with water (30 mL) and the organic solvents were evaporated. The remaining
aqueous phase was extracted with chloroform (3 × 50 mL). The combined organic phases were
washed with water (30 mL), dried with anhydrous Na2SO4, filtered and concentrated in vacuo. The
crude product was purified by automated column chromatography by eluting with n-hexane and
ethyl acetate (2:1) to give compound 19 as a white solid (256 mg, 49%). Mp 158 °C (decomp.);
1
FTIR-ATR 1688 (C=O), 1502, 1444, 1174, 1057, 941, 789 cm-1. H NMR (CDCl3, 300 MHz):
δ
ppm 6.66 (d, J = 8.2 Hz, 1H, aromatic-H), 6.51 (d, J = 8.2 Hz, 1H, aromatic-H), 5.78 (d, J = 9.9 Hz,
1H), 5.32 (m, 1H), 4.98 (dd, J1 = 6.0 Hz and J2 = 1.3 Hz, 1H), 4.88 (m, 1H), 4.06 (m, 1H), 3.83 (s,
21

ACCEPTED MANUSCRIPT
3H, OCH3), 3.75 (m, 1H) 3.73 (brs, 3H, NCOOCH3), 3.52 (s, 3H, OCH3), 3.04 (m, 1H), 2.88 (m,
13
1H), 2.72 (d, J = 18.6 Hz, 1H), 2.51 (brs, 1H), 1.92 (m, 2H). C NMR (CDCl3, 75 MHz):
δ ppm
156.0 (NCO), 155.7*, 147.8, 142.6, 131.9*, 131.6, 129.9, 127.6, 127.5*, 125.8, 125.7*, 119.4,
114.1, 89.5, 75.9, 57.3, 56.7, 52.9, 50.7*, 50.5, 44.0, 40.3*, 40.2, 37.7, 35.9, 35.6*, 29.8*, 29.7.
HRMS calcd for C20H24NO5. [M+1]+ 358.1654 found 358.1656.
Computational modeling
We studied possible binding modes of compounds 17, 23, 19, 22, 11, 12 computationally by
docking them to inactive forms of the MOR, (PDB code 4DKL), DOR (PDB code 4EJ4), and KOR
(PDB code 4DJH). Docking was conducted with Schrödinger Maestro 10.3 [60], Glide XP. The
docking protocol was validated by trying to recreate the binding modes of co-crystallized ligands.
Grid generated was centered on the centroid of the co-crystallized ligands that indicates well the
center of the orthosteric binding pocket. For DOR and MOR, the ligand docking method was based
on a rigid protein. For KOR the rigid protein did not lead to reasonable solutions (KOR ligands
located very high in the binding pocket) but reasonable solutions were delivered by the induced-fit
protocol; induced fit dockings at the KOR are therefore presented in this manuscript. Prior to
docking ligand preparation was performed with standard parameters with LigPrep (pH 7.0 ± 2). The
poses were selected based on the best docking score, but alternative poses were also considered.
Water molecules were not considered during the docking runs. PyMOL was used for visualizing the
results [61].
Biology
Materials - DAMGO, naltrindole, naloxone and GDP were obtained from Sigma. GTP was from
Thermo Scientific. [3H]Diprenorphine (specific activity, 37 Ci/mmol) and [35S]GTPgS (specific
activity, 1250 Ci/mmol) were obtained from PerkinElmer Life Sciences, SNC80 and U50488 from
Tocris (Bio-Techne, Lille, France), and morphine from Francopia (Gentilly, France). Stably
transfected cells expressing human (h) MOR (HEK), DOR (HEK) and KOR (CHO) were obtained
and cultured as described [48,49]. KD, Ki and Bmax values were determined using the Prism software
(GraphPad, San Diego, CA).
Cell viability - The toxicity of the compounds 12, 17, 22 and 23 was evaluated on HEK-293 cells by
cell viability assay. The cells 9000 cells/well) were incubated in the presence of the compounds
(0.1-100 µM) for 24 hours. The growth medium was changed to Opti-MEM prior to the addition of
the compounds. CellTiterGlo reagent was used for cell lysis and determination of the ATP levels of
the lysate from luminescence. The results were analyzed using Prism software (GraphPad, San
Diego, CA). No effect on cell viability was seen at concentrations lower than 100 µM (Supporting
info, Figure S3), showing that these compounds do not have any inherent properties that would
interfere strongly with the basic functions of living cells.
Cell Membrane Preparations – Membranes from transfected cells were obtained as previously
described [48,49]. Briefly, cells were washed with phosphate-buffered saline (PBS, pH 7.4),
scrapped off the plates with PBS, pelleted by centrifugation at 1500 rpm for 10 min at room
temperature and frozen at -80 °C. All the following steps were performed at 4 °C. The pellet was
22

ACCEPTED MANUSCRIPT
thawed and suspended in 30 mL of cold 50 mM Tris-HCl, pH 7.4, 1 mM EDTA and 0.1 mM
phenylmethylsulfonyl fluoride (added extemporaneously). The cell lysate was Dounce-
homogenized (30 ×) and centrifuged at 2500 rpm for 10 min. The pellet was resuspended in the
buffer (15 mL), Dounce-homogenized (30 ×), and centrifuged again at 2500 rpm 10 min. Both
supernatants were combined and ultracentrifuged at 40,000 rpm for 25 min. The pellet was re-
suspended in 5 mL of 50 mM Tris-HCl, pH 7.4, 1 mM EDTA and 0.32 M sucrose and Dounce-
homogenized (10 ×). The membrane preparations were aliquoted and stored at -80 °C. Protein
concentration was determined using the Bradford assay.
Receptor Binding Assay - For saturation experiments, a range of [3H]diprenorphine concentrations
(from 7.8 × 10-11 to 5 × 10-9 M) was used. Nonspecific binding was determined in the presence of
10 µM naloxone. For competition experiments, membrane proteins were diluted in 50 mM Tris-HCl
pH 7.4 and 1 mM EDTA, and incubated with [3H]diprenorphine (1 nM), and several concentrations
of competitor ligand (1 × 10-7 to 5 × 10-5 M, for hDOR naltrindole 1 × 10-11 to 1 × 10-9 M) in a total
volume of 0.2 mL at 23 °C for 1 h. Incubation mixtures were washed (8 ×) using a cell harvester
(PerkinElmer FilterMate Harvester) with cold 50 mM Tris-HCl, pH 7.5 on GF/B filters (Whatman)
presoaked with 50 mM Tris HCl pH 7.4 and 0.1% polyethyleneimine. Bound radioactivity was
determined by scintillation counting using TopCount (Perkin Elmer). Ki and KD values were
determined using the Prism software (GraphPad, San Diego, CA).
[35S]GTPgS Binding Assay – From each hMOR- and hDOR-transfected cell type, 5 µg of
membrane preparations were incubated at 23 °C for 1 h in 50 mM Tris-HCl, pH 7.3, 5 mM MgCl2,
100 mM NaCl, 1 mM EDTA, 30 µM GDP, 0.2 nM [35S]GTPgS, and ligands (1 or 10 µM for the
reference opioid ligand, and 20 or 50 µM for the competitor ligands) in a final volume of 0.2 mL.
Nonspecific binding was determined in the presence of 10 µM GTP. Incubation mixtures were
washed (6 ×) using a cell harvester (PerkinElmer FilterMate Harvester) with cold 50 mM Tris-HCl,
pH 7, 5 mM MgCl2, 50 mM NaCl on H2O-presoaked GF/B filters. Bound radioactivity was
determined by scintillation counting. DAMGO and SNC80 were used as MOR and DOR activators,
respectively, because DAMGO and SNC80 are full high-affinity MOR and DOR agonists and are
classically used for examining antagonism by new compounds.
Acknowledgments
The research leading to these results has received funding from the European Union Seventh
Framework Programme (FP7/2007 - 2013) under grant agreement no 602919. We thank Frederic
Simonin, Valérie Utard, Armand Drieu La Rochelle, Thibaud Tranchant (UMR7242 – Université de
Strasbourg) and Dominique Filliol (CNRS UMR7364 – Université de Strasbourg) for advice on the
biology part.
References
[1] S.M. Parsadaniantz, C. Rivat, W. Rostene, A.R.-L. Goazigo, Opioid and chemokine receptor
crosstalk: a promising target for pain therapy?, Nat. Rev. Neurosci. 16 (2015) 69–78.
23

ACCEPTED MANUSCRIPT
http://dx.doi.org/10.1038/nrn3858.
[2] A.S. Yekkirala, D.P. Roberson, B.P. Bean, C.J. Woolf, Breaking barriers to novel analgesic
drug development, Nat. Rev. Drug Discov. 16 (2017) 545–564.
http://dx.doi.org/10.1038/nrd.2017.87.
[3] H. Breivik, E. Eisenberg, T. O’Brien, The individual and societal burden of chronic pain in
Europe: the case for strategic prioritisation and action to improve knowledge and availability
of appropriate care, BMC Public Health 13 (2013) 1229. doi:10.1186/1471-2458-13-1229.
[4] A. Moore, S. Derry, C. Eccleston, E. Kalso, Expect analgesic failure; pursue analgesic
success, BMJ 346 (2013) f2690. doi:10.1136/bmj.f2690.
[5] E. Kalso, J.E. Edwards, A.R. Moore, H.J. McQuay, Opioids in chronic non‐cancer pain:
systematic review of efficacy and safety, Pain 112 (2004) 372-380.
10.1016/j.pain.2004.09.019.
[6] G.W. Hanks, F. de Conno, N. Cherny, M. Hanna, E. Kalso, H.J. McQuay, S. Mercadante, J.
Meynadier, P. Poulain, C. Ripamonti, L. Radbruch, J.R. i Casas, J. Sawe, R.G. Twycross, V.
Ventafridda, Morphine and alternative opioids in cancer pain: the EAPC recommendations,
Br. J. Cancer 84 (2001) 587–593. http://dx.doi.org/10.1054/bjoc.2001.1680.
[7] A. Manglik, A.C. Kruse, T.S. Kobilka, F.S. Thian, J.M. Mathiesen, R.K. Sunahara, L. Pardo,
W.I. Weis, B.K. Kobilka, S. Granier, Crystal structure of the µ-opioid receptor bound to a
morphinan antagonist., Nature 485 (2012) 321–326. doi:10.1038/nature10954.
[8] H. Wu, D. Wacker, M. Mileni, V. Katritch, G.W. Han, E. Vardy, W. Liu, A.A. Thompson,
X.-P. Huang, F.I. Carroll, S.W. Mascarella, R.B. Westkaemper, P.D. Mosier, B.L. Roth, V.
Cherezov, R.C. Stevens, Structure of the human κ-opioid receptor in complex with JDTic,
Nature 485 (2012) 327–332. http://dx.doi.org/10.1038/nature10939.
[9] S. Granier, A. Manglik, A.C. Kruse, T.S. Kobilka, F.S. Thian, W.I. Weis, B.K. Kobilka,
Structure of the δ-opioid receptor bound to naltrindole, Nature 485 (2012) 400–404.
http://dx.doi.org/10.1038/nature11111.
[10] J.W. Reed, T. Hudlicky, The Quest for a Practical Synthesis of Morphine Alkaloids and
Their Derivatives by Chemoenzymatic Methods, Acc. Chem. Res. 48 (2015) 674–687.
doi:10.1021/ar500427k.
[11] S. Fürst, S. Hosztafi, The chemical and pharmacological importance of morphine analogues,
Acta Physiol. Hung. 95 (2008) 3–44. doi:10.1556/APhysiol.95.2008.1.1.
[12] G.L. Patrick, An Introduction to Medicinal Chemistry, fifth ed., Oxford University Press,
New York, 2013.
[13] H. Schmidhammer, M. Spetea, Synthesis of 14-Alkoxymorphinan Derivatives and Their
Pharmacological Actions, in: H. Nagase (Ed.), Chemistry of Opioids, Springer Berlin
Heidelberg, Berlin, Heidelberg, 2011, pp. 63–91. doi:10.1007/128_2010_77.
[14] J.W. Lewis, S.M. Husbands, 14-Amino-4,5-Epoxymorphinan Derivatives and Their
Pharmacological Actions, in: H. Nagase (Ed.), Chemisty of Opioids, Springer Berlin
Heidelberg, Berlin, Heidelberg, 2011, pp. 93–119. doi:10.1007/128_2010_89.
[15] I. Hutchinson, S. Archer, K.P. Hill, J.M. Bidlack, Synthesis and opioid binding properties of
24

ACCEPTED MANUSCRIPT
2-chloroacrylamido derivatives of 7,8-dihydromorphinans, Bioorg. Med. Chem. Lett. 6
(1996) 1563–1566. doi:https://doi.org/10.1016/S0960-894X(96)00274-0.
[16] P. Magnus, B. Ghavimi, J.W. Coe, Access to 7α-analogs of codeine with mixed µ/δ agonist
activity via 6,7-α-epoxide opening, Bioorg. Med. Chem. Lett. 23 (2013) 4870–4874.
doi:10.1016/j.bmcl.2013.06.084.
[17] C.W. Cunningham, S.L. Mercer, H.E. Hassan, J.R. Traynor, N.D. Eddington, A. Coop,
Opioids and efflux transporters. Part 2: P-glycoprotein substrate activity of 3- and 6-
substituted morphine analogs, J. Med. Chem. 51 (2008) 2316–2320. doi:10.1021/jm701457j.
[18] S. Béni, G. Tóth, B. Noszál, S. Hosztafi, Preparation of benzoate esters of morphine and its
derivatives, Monatsh. Chem. 143 (2012) 1431–1440. doi:10.1007/s00706-012-0803-8.
[19] G.B. Kok, P.J. Scammells, Improved synthesis of 14-hydroxy opioid pharmaceuticals and
intermediates, RSC Adv. 2 (2012) 11318-11325. doi:10.1039/c2ra21693k.
[20] R. Csuk, G. Vasileva, A. Barthel, Towards an Efficient Preparation of Hydromorphone,
Synthesis 44 (2012) 2840–2842. doi:10.1055/s-0031-1291151.
[21] M.A.A. Endoma-Arias, D.P. Cox, T. Hudlicky, General method of synthesis for naloxone,
naltrexone, nalbuphone, and nalbuphine by the reaction of grignard reagents with an
oxazolidine derived from oxymorphone, Adv. Synth. Catal. 355 (2013) 1869–1873.
doi:10.1002/adsc.201300284.
[22] H. Rapoport, E.C. Galloway, The Action of Peracids on the Desoxycodeines, J. Am. Chem.
Soc. 77 (1955) 5753–5754. doi:10.1021/ja01626a096.
[23] K. Uba, N. Miyata, K. Watanabe, M. Hirobe, Stereospecific Synthesis of Codeine 7, 8-Oxide
and Codeinone 7, 8-Oxide, Chem. Pharm. Bull. 27 (1979) 2257–2258.
doi:10.1248/cpb.27.2257.
[24] N. Miyata, K. Uba, K. Watanabe, M. Hirobe, Synthesis and Analgesic Activity of Morphine-
7, 8-oxide and Heroin-7, 8-oxide, Chem. Pharm. Bull. 28 (1980) 3722–3724.
doi:10.1248/cpb.28.3722.
[25] H. Rapoport, M.S. Chadha, C.H. Lovell, Preparation and Cleavage of Some Codeine
Glycols, J. Am. Chem. Soc. 79 (1957) 4694–4697. doi:10.1021/ja01574a031.
[26] H. Rapoport, C.H. Lovell, H.R. Reist, M.E. Warren, The Synthesis of Thebaine and
Northebaine from Codeinone Dimethyl Ketal, J. Am. Chem. Soc. 89 (1967) 1942–1947.
doi:10.1021/ja00984a032.
[27] L.J. Sargent, J.H. Ager, Transformation of Codeine to an Analog of the Potent Analgesic
Phenazocine, J. Med. Chem. 6 (1963) 569–572. doi:10.1021/jm00341a024.
[28] L.J. Sargent, L.H. Schwartzman, L.F. Small, Hydroxylated Codeine Derivatives, J. Org.
Chem. 23 (1958) 1247–1251. doi:10.1021/jo01103a003.
[29] A. MacHara, D.P. Cox, T. Hudlicky, Direct synthesis of naltrexone by palladium-catalyzed
N-demethylation/ acylation of oxymorphone: The benefit of C-H activation and the
intramolecular acyl transfer from C-14 hydroxy, Adv. Synth. Catal. 354 (2012) 2713–2718.
doi:10.1002/adsc.201200677.
25

ACCEPTED MANUSCRIPT
[30] R.S. Cahn, R. Robinson, CXXVI.-The morphine group. Part IV. A new oxidation product of
codeine, J. Chem. Soc. 129 (1926) 908–912. doi:10.1039/JR9262900908.
[31] A.J. Fatiadi, The Classical Permanganate Ion: Still a Novel Oxidant in Organic Chemistry,
Synthesis 1987 (1987) 85–127. doi:10.1055/s-1987-27859.
[32] S.L. Regen, C. Koteel, Activation through impregnation. Permanganate-coated solid
supports, J. Am. Chem. Soc. 99 (1977) 3837–3838. doi:10.1021/ja00453a053.
[33] F.M. Menger, C. Lee, Oxidations with solid potassium permanganate, J. Org. Chem. 44
(1979) 3446–3448. doi:10.1021/jo01333a051.
[34] N.A. Noureldin, W.B. McConnell, D.G. Lee, Heterogeneous permanganate oxidations. 4.
The oxidation of sulfides and selenides, Can. J. Chem. 62 (1984) 2113–2116.
doi:10.1139/v84-360.
[35] J. Tercio B Ferreira, W.O. Cruz, P.C. Vieira, M. Yonashiro, Carbon-carbon double bond
cleavage using solid-supported potassium permanaganate on silica gel, J. Org. Chem. 52
(1987) 3698–3699. doi:10.1021/jo00392a041.
[36] J.H. Clark, D.G. Cork, Synthesis of 1,4-diketones by fluoride-catalysed Michael addition and
supported permanganate oxidation, J. Chem. Soc., Chem. Commun. (1982) 635–636.
doi:10.1039/C39820000635.
[37] S. Wolfe, C.F. Ingold, Oxidation of organic compounds by zinc permanganate, J. Am. Chem.
Soc. 105 (1983) 7755–7757. doi:10.1021/ja00364a054.
[38] C.W. Jefford, Y. Wang, Selective, heterogeneous oxidation of alcohols and diols with
potassium permanganate, J. Chem. Soc., Chem. Commun. (1988) 634–635.
doi:10.1039/C39880000634.
[39] T.X.T. Luu, T.-T.T. Nguyen, T.N. Le, J. Spanget-Larsen, F. Duus, Fast and efficient green
synthesis of thiosulfonate S-esters by microwave-supported permanganate oxidation of
symmetrical disulfides, J. Sulfur Chem. 36 (2015) 340–350.
doi:10.1080/17415993.2015.1025404.
[40] L.-Y. Zhu, C. Huang, C. Shi, An Improved Method for Oxidation of Oximes with Potassium
Permanganate Adsorbed on Graphite Reagent under Viscous Conditions, Oxid. Commun. 35
(2012) 389–393.
[41] M.M. Lakouraj, S.M. Hossaini, V. Hasantabar, M. Soleimani, 2, 4-Ionene Supported
Permanganate: A Novel, Efficient and Selective Solid Phase Oxidizing Agent for Oxidation
of Organic Compounds, J. Adv. Oxid. Technol. 14 (2011) 251. doi:10.1515/jaots-2011-0210.
[42] T.X.T. Luu, P. Christensen, F. Duus, T.N. Le, Microwave- and Ultrasound-Accelerated
Green Oxidation of Alcohols by Potassium Permanganate absorbed on Copper(II) Sulfate
Pentahydrate, Synth. Commun. 38 (2008) 2011–2024. doi:10.1080/00397910801997819.
[43] A. Chacko, B. Mathew, Polystyrene-supported polyoxyethylene bound potassium
permanganate as a heterogeneous oxidizing agent, J. Appl. Polym. Sci. 90 (2003) 3708–
3717. doi:10.1002/app.13069.
[44] S. Lai, D.G. Lee, Heterogeneous Permanganate Oxidation of Styrene and Cinnamic Acid
Derivatives: A Simple and Effective Method for the Preparation of Benzaldehydes, Synthesis
26

ACCEPTED MANUSCRIPT
2001 (2001) 1645–1648. doi:10.1055/s-2001-16760.
[45] S. Baskaran, J. Das, S. Chandrasekaran, Heterogeneous permanganate oxidations: an
improved procedure for the direct conversion of olefins to α-diketones/α-hydroxy ketones, J.
Org. Chem. 54 (1989) 5182–5184. doi:10.1021/jo00282a042.
[46] J.A.R. Salvador, V.M. Moreira, J.R. Hanson, R.A. Carvalho, One-pot, high yield synthesis of
α-ketols from ∆5-steroids, Steroids 71 (2006) 266–272. doi:10.1016/j.steroids.2005.11.002.
[47] D.G. Lee, T. Chen, Oxidation of hydrocarbons. 18. Mechanism of the reaction between
permanganate and carbon-carbon double bonds, J. Am. Chem. Soc. 111 (1989) 7534–7538.
doi:10.1021/ja00201a039.
[48] J.A.J. Becker, A. Wallace, A. Garzon, P. Ingallinella, E. Bianchi, R. Cortese, F. Simonin,
B.L. Kieffer, A. Pessi, Ligands for κ-opioid and ORL1 receptors identified from a
conformationally constrained peptide combinatorial library, J. Biol. Chem. 274 (1999)
27513–27522. doi:10.1074/jbc.274.39.27513.
[49] K. Befort, D. Filliol, F.M. Décaillot, C. Gavériaux-Ruff, M.R. Hoehe, B.L. Kieffer, A Single
Nucleotide Polymorphic Mutation in the Human µ-Opioid Receptor Severely Impairs
Receptor Signaling, J. Biol. Chem. 276 (2001) 3130–3137. doi:10.1074/jbc.M006352200.
[50] W. Huang, A. Manglik, A.J. Venkatakrishnan, T. Laeremans, E.N. Feinberg, A.L. Sanborn,
H.E. Kato, K.E. Livingston, T.S. Thorsen, R.C. Kling, S. Granier, P. Gmeiner, S.M.
Husbands, J.R. Traynor, W.I. Weis, J. Steyaert, R.O. Dror, B.K. Kobilka, Structural insights
into µ-opioid receptor activation, Nature 524 (2015) 315–321.
http://dx.doi.org/10.1038/nature14886.
[51] C. Chavkin, A. Goldstein, Specific receptor for the opioid peptide dynorphin: structure--
activity relationships, Proc. Natl. Acad. Sci. 78 (1981) 6543–6547.
http://www.pnas.org/content/78/10/6543.abstract.
[52] A.W. Lipkowski, S.W. Tam, P.S. Portoghese, Peptides as receptor selectivity modulators of
opiate pharmacophores, J. Med. Chem. 29 (1986) 1222–1225. doi:10.1021/jm00157a018.
[53] A.-K. Gilbert, S. Hosztafi, L. Mahurter, G.W. Pasternak, Pharmacological characterization of
dihydromorphine, 6-acetyldihydromorphine and dihydroheroin analgesia and their
differentiation from morphine, Eur. J. Pharmacol. 492 (2004) 123–130.
doi:https://doi.org/10.1016/j.ejphar.2004.03.050.
[54] V.O. Nikolaev, C. Boettcher, C. Dees, M. Bünemann, M.J. Lohse, M.H. Zenk, Live Cell
Monitoring of µ-Opioid Receptor-mediated G-protein Activation Reveals Strong Biological
Activity of Close Morphine Biosynthetic Precursors, J. Biol. Chem. 282 (2007) 27126–
27132. doi:10.1074/jbc.M703272200.
[55] P.F. Zaratin, G. Petrone, M. Sbacchi, M. Garnier, C. Fossati, P. Petrillo, S. Ronzoni, G.A.M.
Giardina, M.A. Scheideler, Modification of Nociception and Morphine Tolerance by the
Selective Opiate Receptor-Like Orphan Receptor Antagonist (–)-cis-1-Methyl-7-[[4-(2,6-
dichlorophenyl)piperidin-1-yl]methyl]-6,7,8,9-tetrahydro-5H-benzocyclohepten-5-ol (SB-
612111), J. Pharmacol. Exp. Ther. 308 (2004) 454-461. doi:10.1124/jpet.103.055848.
[56] H. Schmidt, S. V Vormfelde, K. Klinder, U. Gundert-Remy, C.H. Gleiter, G. Skopp, R.
Aderjan, U. Fuhr, Affinities of Dihydrocodeine and its Metabolites to Opioid Receptors,
Pharmacol. Toxicol. 91 (2002) 57–63. doi:10.1034/j.1600-0773.2002.910203.x.
27

ACCEPTED MANUSCRIPT
[57] M.D. Aceto, L.S. Harris, M.E. Abood, K.C. Rice, Stereoselective µ- and δ-opioid receptor-
related antinociception and binding with (+)-thebaine, Eur. J. Pharmacol. 365 (1999) 143–
147. doi:https://doi.org/10.1016/S0014-2999(98)00862-0.
[58] C. Mignat, U. Wille, A. Ziegler, Affinity profiles of morphine, codeine, dihydrocodeine and
their glucuronides at opioid receptor subtypes, Life Sci. 56 (1995) 793–799.
doi:https://doi.org/10.1016/0024-3205(95)00010-4.
[59] M.P. Wentland, W. Duan, D.J. Cohen, J.M. Bidlack, Selective protection and
functionalization of morphine: Synthesis and opioid receptor binding properties of 3-amino-
3-desoxymorphine derivatives, J. Med. Chem. 43 (2000) 3558–3565.
doi:10.1021/jm000119i.
[60] Schrödinger Release 2015-3: Maestro, version 10.3, Schrödinger, LLC, New York, NY,
(2015).
[61] The PyMOL Molecular Graphics System, Version 1.7.0.0 Schrödinger, LLC.
28