Inclusion Complexes of Gold(I)-Dithiocarbamates with β-Cyclodextrin: A Journey from Drug Repurposing towards Drug Discovery

Article abstract of DOI:10.1002/chem.202101366

The gold(I)-dithiocarbamate (dtc) complex [Au(N,N-diethyl)dtc]₂ was identified as the active cytotoxic agent in the combination treatment of sodium aurothiomalate and disulfiram on a panel of cancer cell lines. In addition to demonstrating pronounced differential cytotoxicity to these cell lines, the gold complex showed no cross-resistance in therapy-surviving cancer cells. In the course of a medicinal chemistry campaign on this class of poorly soluble gold(I)-dtc complexes, >35 derivatives were synthesized and X-ray crystallography was used to examine structural aspects of the dtc moiety. A group of hydroxy-substituted complexes has an improved solubility profile, and it was found that these complexes form 2 : 1 host–guest inclusion complexes with β-cyclodextrin (CD), exhibiting a rarely observed “tail-to-tail” arrangement of the CD cones. Formulation of a hydroxy-substituted gold(I)-dtc complex with excess sulfobutylether-β-CD prevents the induction of mitochondrial reactive oxygen species, which is a major burden in the development of metallodrugs.

Full text of DOI:10.1002/chem.202101366

Accepted Article
Accepted Article
Title: Inclusion Complexes of Gold(I)-Dithiocarbamates with ß-
Cyclodextrin: A Journey from Drug Repurposing towards Drug
Discovery
Authors: Michael Morgen, Piotr Fabrowski, Eberhard Amtmann,
Nikolas Gunkel, and Aubry Kern Miller
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Chem. Eur. J. 10.1002/chem.202101366
Link to VoR: https://doi.org/10.1002/chem.202101366
0
1/2020

Chemistry - A European Journal
10.1002/chem.202101366
FULL
PAPER
Inclusion Complexes of Gold(I)-Dithiocarbamates with
-Cyclodextrin: A Journey from Drug Repurposing
towards Drug Discovery

Michael Morgen,[a] Piotr Fabrowski,[a] Eberhard Amtmann,[a] Nikolas Gunkel[a,b] and Aubry K.
Miller*[a,b]
[
a]
Dr. M. Morgen, Dr. P. Fabrowski, Dr. E. Amtmann, Dr. N. Gunkel, Dr. A. K. Miller
Cancer Drug Development Group (A390)
German Cancer Research Center (DKFZ)
Im Neuenheimer Feld 280
69120 Heidelberg (Germany)
E-mail: aubry.miller@dkfz.de
Dr. N. Gunkel, Dr. A. K. Miller
German Cancer Consortium (DKTK)
[
b]
69120 Heidelberg (Germany)
Supporting information for this article is given via a link at the end of the document.
Abstract: We identified the gold(I)-dithiocarbamate (dtc) complex
Au(N,N-diethyl)dtc] as the active cytotoxic agent in the combination
copper dithiocarbamate (dtc) complex Cu((N,N-diethyl)dtc)
2
,
which forms in vivo with endogenous or supplemented Cu2
+
[
2
treatment of sodium aurothiomalate and disulfiram on a panel of
cancer cell lines. In addition to demonstrating pronounced differential
cytotoxicity to these cell lines, the gold complex showed no cross-
resistance in therapy-surviving cancer cells. In the course of a
medicinal chemistry campaign on this class of poorly soluble gold(I)-
dtc complexes, we synthesized >35 derivatives and used x-ray
crystallography to examine structural aspects of the dtc moiety. A
group of hydroxy-substituted complexes has an improved solubility
(Figure 1). The development of metal-containing drugs remains a
vibrant area of research, and we found the idea of forming
organometallic complexes in vivo a unique opportunity for drug
repurposing. Among the late transition metals, gold has a long
history in medicine and is found in approved drugs like
[
4]
aurothiomalate (ATM) and auranofin. As gold complexes are
widely regarded to be promising anti-cancer agents,[ we
wondered what the effects from combined administration of DSF
and ATM would be.[
5]
6]
profile, and we found that these complexes form 2:1 host–guest
inclusion complexes with -cyclodextrin (CD), exhibiting a rarely
observed “tail-to-tail” arrangement of the CD cones. Formulation of a
hydroxy-substituted gold(I)-dtc complex with excess sulfobutylether-
S
S
S
S
S
S
N
N
S
N
Cu
N
-CD prevents the induction of mitochondrial reactive oxygen species,
S
which is a major burden in the development of metallodrugs.
disulfiram(DSF)
Cu((N,N-diethyl)dtc)2
OAc
Introduction
O
S
OAc
AcO
AcO
S
ONa
Au
NaO
aurothiomalate (ATM)
Au
O
PEt
3
O
Drug repurposing, the development of approved or clinically-
investigated drugs for new indications, is a powerful approach to
bring new therapy options to patients.[ By harnessing the world’s
vast chemical equity for which toxicological and pharmacokinetic
data in human subjects already exists, drug repurposing can
progress compounds from pre-clinical proof of concept studies to
Phase 2 clinical trials more quickly and economically than
traditional drug development projects. One substance that has
been investigated extensively for repurposing in oncology is
n
auranofin
1]
Figure 1: Metal-containing and metal-coordinating drugs.
Results and Discussion
Combination of DSF and ATM produces a cytotoxic gold-
complex. In initial screening experiments, we treated cells with
ATM and DSF at different ratios and always observed the
formation of a precipitate, presumably the result of rapid reaction
between the two compounds. As cyclodextrins are commonly
used to solubilize lipophilic substances and to facilitate cellular-
uptake of small molecules, we repeated the experiment with DSF
disulfiram (DSF), a drug which is clinically used to treat alcohol
dependence (Figure 1).[
2]
In
a
seminal retrospective
epidemiological study, Skrott et al. found a significant benefit for
cancer patients who continued to take DSF after their diagnosis.[3]
Interestingly, the researchers demonstrated that disulfiram’s anti-
cancer activity derives not from disulfiram itself, but from the
1
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
pre-formulated in 2-hydroxypropyl--cyclodextrin.[7] In this
experimental set-up no precipitate formed and a wide range of
combination ratios was cytotoxic against a panel of cell lines. For
example, a 1:1 combination of ATM/DSF proved to be at least
with DSF (1) is redox neutral, presumably starting with a ligand
exchange to give 3 and 4. Compound 4 can stabilize itself through
dimerization, forming 5. The poor solubility of 5 shifts the
equilibrium of the initial ligand exchange and serves to drive the
reaction to completion. Calu-6 cells responded to treatment with
5 formulated in 2-hydroxypropyl--cyclodextrin (EC50 = 1.6 M)
similarly to when treated with DSF and ATM (EC50 = 2.9 M, see
Figure 2A, Figure S2), suggesting that 5 is the active cytotoxic
substance.
10-fold more efficient against each cell line than ATM or DSF
alone (Figure 2A). Some cells were more sensitive than others to
the combination treatment, with the most sensitive cell lines (HL-
6
0 and HCT 116) showing between one and two orders of
magnitude lower EC50 values than the more resistant cell lines
HeLa, SK-OV-3, Calu-6, COLO-357 and MeWo). The
combination of DSF and ATM also proved more effective against
therapy-surviving” oxaliplatin-resistant HT-29 cells than four
(
S
S
ONa
Au
“
S
N
+
N
S
O
other standard chemotherapy agents, suggesting a potential to
overcome acquired drug resistance (Figure 2B). This superior
efficacy on drug-resistant cells was not accompanied by high
general toxicity as non-transformed human fibroblasts showed
NaO
O
S
DSF (1)
ATM (2)
3
0-fold lower sensitivity to ATM/DSF than both naïve and resistant
S
HT-29 cells (Figure 2C).
S
S
ONa
N
+
S
N
Au
O
NaO
3
O
S
4
a. Dimerization
b. Precipitation
S
S
Au
Au
S
S
N
N
[
Au(N,N-diethyl)dtc] (5)
2
Scheme 1: Proposed pathway of the reaction between DSF and ATM to give
5.
The cytotoxic properties of gold(I)-dtc complexes like 5 and also
of gold(III)-dtc complexes have previously been studied; however,
to the best of our knowledge, no reports of selective anti-tumoral
properties of such complexes have been discussed.[ Also, the
poor solubility of 5 in all common solvents, including DMSO,
makes studying the cellular activity of such a compound virtually
impossible in the absence of formulation. Encouraged by our
initial findings, we decided to expand our initial repurposing
concept and launched a medicinal chemistry campaign to prepare
new analogs of 5 for two reasons: 1. Solubilization of a DSF/ATM
combination or 5 requires >30 molar equivalents of cyclodextrin,
a very high amount for pre-clinical or clinical studies. We
hypothesized that analogs of 5 containing different dtc moieties
would require reduced amounts of cyclodextrin for solubilization.
10]
Figure 2: Cytotoxic effects of disulfiram (DSF) combined with aurothiomalate
(
ATM). A. EC50 values measured in cell lines covering seven different cancer
entities for DSF, ATM, and the combination DSF/ATM (1:1). For all experiments,
-hydroxypropyl--cyclodextrin was used as formulation in 30-fold excess. B.
2
Factor of resistance measured in “therapy-resistant” and naïve HT-29 cells for
four clinically approved drugs and the combination DSF/ATM. The HT-29 cells
were rendered “therapy-resistant” by treating them with EC80 concentrations of
oxaliplatin, and then re-cultivating the surviving population. The factor of
resistance is defined as the ratio of the EC50 values measured in “therapy-
resistant” and naïve cells, respectively. C. EC50 values for the DSF/ATM
combination in naïve and “therapy-resistant” HT-29 cells, and human fibroblast
cells. For B and C, DSF/ATM was used in a 1:2 ratio and 2-hydroxypropyl--
cyclodextrin was used as formulation in 30-fold excess.
2
. We wanted to explore the role of the dtc moiety in the biological
activity of such complexes. To this end, we synthesized and
tested a series of dithiocarbamate complexes for solubility with
the FDA-approved formulant sulfobutylether--cyclodextrin (SBE-

-CD)[
7a,11]
and for efficacy and selectivity in a panel of cancer
cells lines. Concurrent with these biological studies, we became
interested in the structural aspects of these gold complexes and
their interaction with cyclodextrins in general.
Encouraged by these results, we decided to investigate the
chemical outcome of the combination of DSF and ATM further.
Combining the two in a 1:1 ratio rapidly produced an orange
precipitate, which we confirmed to be the gold(I)-dithiocarbamate
complex 5, isolated in a yield of 84% (Scheme 1).[8] In contrast to
the reaction of Cu(II) salts with DSF, where a multi-step redox
Structural aspects of gold(I)-dithiocarbamate complexes.
Numerous reports deal with the synthesis and characterization of
dimeric gold(I)-dtc complexes such as 5. These molecules can
have R1 and R2 groups that are identical[8b,12] or different from
[10a,13]
each other,
either as distinct substituents or connected in a
cyclic system (Figure 3A, central circle).[
12c,14]
Most of the known
process produces Cu((N,N-diethyl)dtc)
2
,[9] the reaction of ATM (2)
examples are highly symmetric complexes with identical R and
1
2
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
Figure 3: Synthesis and experimental structures of dimeric gold(I)-dtc complexes. A. Synthesized gold(I)-dtc complexes with dithiocarbamates containing acyclic
orange), aromatic (blue), cyclic (green) and hydroxy-substituted (red) residues. Isolated yields are shown in parenthesis; also shown is the primarily-used synthetic
(
route (grey middle circle). B. Experimental structure of 23 highlighting the co-linear alignment of the gold atoms due to aurophilic bonding. C. Experimental structure
of 17; view of the asymmetric unit along the crystallographic a-axis. For B and C: Ellipsoids are shown at a 50% probability level. Hydrogens are omitted for clarity.
Color code: carbon, grey; nitrogen, blue; sulfur, yellow; gold, orange.[
17]
2
R groups. These complexes are typically synthesized by mixing
an alkali dithiocarbamate salt and a source of Au(I), such as
piperazine derivative 23 stack their gold atoms to form linear
“wires” through the crystal lattice with a distance of 3.24 Å to
adjacent dimeric gold complexes (Figure 3B). Such “aurophilic
bonding” is frequently found in the solid state of closed-shell Au(I)
[8b]
S)AuCl,[13,14] or in situ reduced Au(III)
(
Ph
salts.
3
P)AuCl
or (Me
We found using ATM as a source of Au(I) to be
2
[
10a,12a]
10
superior. This approach takes advantage of the poor water
solubility of most dimeric gold(I)-dtc complexes and the excellent
water solubility of dithiocarbamate salts and ATM. As such, a
lithium or potassium dithiocarbamate salt, prepared from an
(5d ) complexes (and gold(I)-dtc complexes in general, see
Table S1).[ In contrast, the gold atoms in benzyl derivative 17
make no intermolecular contacts (Figure 3C). This situation is
rarely observed for gold(I)-dtc complexes – in a thorough CCDC
search, we found ten similar gold(I) complexes with different dtc
moieties, of which only two showed no aurophilic bonding. In
these two specific cases the steric bulk of the dtc moiety or co-
crystallized compounds prevent this aggregation (see Table S1).
In the case of 17 both situations do not apply. Presumably,
favorable hydrophobic interactions of the benzyl groups
energetically outweigh stabilization via intermolecular aurophilic
contacts.
16]
2
amine and CS , either in a separate step or in situ, is added to an
aqueous solution of ATM. The resulting gold(I)-dtc complexes
precipitate from the solution and can be isolated in analytically
pure form by filtration and washing with water. This convenient
method provided
a variety of functionalized complexes in
moderate to excellent yields and high purity. We categorized the
complexes into four different groups based on their dtc
substituents: acyclic (orange), aromatic (blue), cyclic (green) and
hydroxy-substituted (red) (Figure 3A).[15]
Many of the synthesized complexes are poorly soluble in organic
solvents at ambient temperature. Some derivatives could be
solubilized at high temperatures and, upon cooling, crystallized.
In this manner, x-ray quality crystals of 23 (Figure 3B) and 17
Second, the eight-membered rings defined by the two gold atoms
and the two S-C-S fragments of the dithiocarbamates have quite
different conformations in the two structures, indicating flexibility
in the ring. This can best be described by the twist angle , which
we define to be the angle, when viewed down the Au-Au axis,
formed by lines made by connecting S1 with S4 and S2 with S3
(Figure 4A). While piperazine derivative 23 has a twisted eight-
membered ring with  = 33.6°, benzyl derivative 17 has an almost
flat eight-membered ring with   2.3°.[ Another distinguishing
(
Figure 3C) were obtained, complementing the set of available
experimental structures (Table S1). Three interesting structural
features of gold(I)-dtc complexes are exemplified when
comparing these two structures. First, multiple molecules of
18]
3
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
-1
absorption peaks at around 1500 cm . These can be attributed to
the “thioureide-band”, which is between a carbon-nitrogen single
-
1
bond at ꢀꢁ = 1250 – 1350 cm
and
a
double bond at
-
1 [10a]
ꢀꢁ = 1640 – 1690 cm .
A consequence of this amide-like
double bond character are “cis/trans-conformers” that, due to
restricted bond rotation constitute energetic minima, will exist for
1
2
compounds where R and R are different (Figure 4C). In the case
1
2
of 23, R and R are identical, but rotation around the thioureide
C-N bond still produces two distinct conformations, because the
cyclic piperazine moiety adopts a chair-conformation with the N-
methyl group in a favorable equatorial position. One conformation
orients both piperazines “above” the eight-membered ring (not
observed), whereas the other (observed) displays the piperazines
on opposite sides in an anti-conformation, as shown in Figure 4A.
Dimeric gold(I)-dtc complexes form inclusion complexes
with -Cyclodextrin. We noticed that the hydroxy-substituted
complexes are, in general, more soluble than the other classes,
with many (29–39) being readily soluble in some organic solvents
such as DMSO (see Figure 3A). More importantly, less SBE--
CD was required to prepare standard aqueous solutions of these
complexes. To better quantify the amount of SBE--CD that was
needed for each gold(I)-dtc complex, we performed a modified
kinetic solubility experiment on a selection of compounds (Figure
S3). While there is no apparent solubility trend of the complexes
with respect to the size or constitution of the dithiocarbamate side
Figure 4: A. Depiction of the twist angle , which can be calculated by taking
the mean of both dihedral angles S(1)-Au(1)-Au(2)-S(2) and S(3)-Au(2)-Au(1)-
S(4), for 17 and 23. B. Depiction of the S-Au-S angles for 17 and 23. C.
Depiction of the possible cis/trans-conformers in gold(I)-dtc complexes bearing
different R1 and R2 groups.
[17]
feature of the eight-membered ring is the angle of the S-Au-S
bonds: in compound 23, this is nearly linear with a mean angle of
Figure 5: -cyclodextrin as
a host for host–guest complexes. A. Sketch
1
79.7°, but it significantly deviates from linearity at 171.9° for
visualizing different common modes of interaction of -CD with a lipophilic
compound with a 2:1 host (blue)/guest (yellow) ratio – formation of an inclusion
complex with a “head-to-head”-arrangement of the CD cones and encapsulation
of the guest by -CD in the interstitial space. B. Proposed less-ordered
aggregates formed with insoluble compounds like 5 (yellow) in the presence of
complex 17, presumably due to an intramolecular aurophilic
interaction between the two gold atoms (Figure 4B). Lastly, the
R R N—CS bond shows considerable double bond character.
2
1
2
This is apparent from the R-N-C-S torsion angle being close to 0°,
and also from the IR spectrum of 5 (Figure S1) that shows
high excess of -CD (blue) as
C. Schematic depiction of -cyclodextrin.
a
special case of encapsulation.
4
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
chains, we were intrigued that some complexes only require
cyclodextrin ratios as low as 2 to 3 (for 34 and 32, respectively)
for solubilization. These relatively low ratios suggested to us that
such compounds might be capable of forming distinct 2:1 host–
guest inclusion or encapsulation complexes with cyclodextrins
corresponding cyclodextrin, whereas the gold atoms of the eight-
membered ring system are located within the cleft of the
cyclodextrin cones.
The primary hydroxy groups of the narrow “tail” side of each CD
cone are all in a (–)-gauche orientation, as in uncomplexed -
cyclodextrin (i.e. pointing towards the outside of the cavity;
Figure 5C) indicating no direct interaction with the complex on the
inside. Both -CD cones of the inclusion complexes are arranged
almost parallel to each other – the angle between the mean plane
through the seven glycosidic oxygens of both cyclodextrins is 1.1°
(
(
Figure 5A),[19] as opposed to less-ordered aggregates
Figure 5B), which might be more operative under high
cyclodextrin concentrations as required for 5.
To explore this hypothesis, we attempted to grow crystals of
selected gold complexes out of cyclodextrin solutions. Because
SBE--CD is not a homogenous material, we used solutions of

gold complexes, and allowed crystals to form via slow evaporation
of water. From these experiments, we were able to isolate crystals
from seven different gold complexes (30, 32, 33, 35, 37, 39, and
for [33/(-CD)
2
], 5.7° for [39/(-CD)
2
2
] and 3.7° for [43/(-CD) ] –
-cyclodextrin (Figure 5C)[ that were saturated with specific
20]
and are not precisely stacked on top of each other but are shifted
sideways by approximately 2.2 to 2.4 Å (see Figure S5 for a
detailed description of the determination of the structural
parameters).
4
3). NMR analysis of crystals re-dissolved in DMSO-d
6
revealed
In complex 33 (Figure 6A) the gold atoms are disordered (three-
fold disorder of the eight-membered ring system; not shown). As
a consequence, one of the carbamate moieties is fixed in its
position within the cavity of the host, whereas the other one
adopts different positions due to different twist angles of the eight-
membered ring system (twist angle  = 33.1° for the major
component; site occupancy factor (s.o.f.) 71%). In contrast, the
eight-membered ring system of 39 (Figure 6B) shows no disorder.
It has a twist angle  = 29.1°, and the gold-complex exists as a
nearly 1:1 mixture of cis/trans-conformers (s.o.f. 52% for the
major component). Complex 43 (Figure 6C) also shows no
disorder and is solely present in an anti-conformation similar to 23
(Figure 3B). It has a twist angle  = 9.9°, resulting in the most
planar eight-membered ring system of the three gold(I)-dtc
inclusion complexes.
[21]
all had 2:1 host/guest ratios (Figure S4). For the crystals grown
from solutions of 33, 39 and 43, we were able to obtain well-
resolved X-ray structures, thereby confirming the formation of
inclusion complexes.[22]
The experimental structures were elucidated in resolutions of
0
complexes share common features (Figure 6A–C). The gold(I)-
dtc complexes are included into two -CDsthat arrange in a “tail-
to-tail” manner (2:1 host/guest ratio) (Figure 6D). This
counterintuitive arrangement of the CD cones is only rarely
observed in inclusion complexes.[ The organic residues of each
gold(I)-dtc complex reach out of the wide “head”-side of the
.91 Å (33), 0.91 Å (39) and 0.84 Å (43), and all three inclusion
23]
Interestingly, the stacking of the inclusion complexes in the crystal
lattice is different for all obtained experimental structures. In
[
33/(-CD)
channel-like structures, stacking up to make “infinite” pipes
Figure 7A). In [39/(-CD) ] the CD cones are staggered in pairs
with only every second unit stacking-up to one adjacent unit
Figure 7B). In [43/(-CD) ] the inclusion complexes are arranged
2
] the CD cones of the inclusion complexes form
(
2
(
2
in a brick-like structure (Figure 7C). In each case the interstitial
space of the CD-cones as well as between the inclusion
complexes is filled with water molecules stabilizing this assembly
by an extensive network of hydrogen bonds (not shown). No water
molecules were found within the cavity of the -cyclodextrins,
indicating that they are completely filled by the organic ligands of
the gold complexes.
SBE--CD reduces mitochondrial ROS formation from
dimeric gold(I)-dtc complexes. Many cancer cells are sensitive
to elevated reactive oxygen species (ROS) levels, and metal
complexes are known to increase ROS by disrupting proteins that
regulate cellular redox homeostasis. Normal cells are not immune
to ROS-induced damage, and bursts of ROS in mitochondria can
cause cardiac toxicity. This is particularly problematic for the safe
use of metallodrugs in animals and patients.[ We therefore
measured the effects of our gold(I)-dtc complexes on cellular ROS
levels using a genetically encoded ROS-sensor (roGFP2-Orp1),
which can be selectively localized to either the cytoplasm or the
24]
Figure 6: Experimental structures of the inclusion complexes of 33, 39 and 43
with -CD. A. Depiction of [33/(-CD)
occupancy factor (s.o.f.) 71%). B. Depiction of [39/(-CD)
component (s.o.f. 52%). C. Depiction of [43/(-CD) ]. D. Sketch visualizing the
2
]; shown is the major component (site
2
]; shown is the major
2
[25]
mitochondria. Surprisingly, many of our compounds produced
obtained inclusion complexes of gold(I)-dtc complexes (yellow) with -
cyclodextrin (blue) in a 2:1 host/guest ratio and a “tail-to-tail” arrangement of the
CD cones. For A–C: Hydrogens and water molecules are omitted for clarity and
the -cyclodextrin cones are shown as capped-stick model whereas 33, 39 and
a rapid and persistent increase in cytoplasmic ROS while having
little effect on mitochondrial ROS (data not shown). In order to
discern the source of this effect, we measured ROS-induction by
4
3 are depicted as ellipsoids (on a 50% probability level). Color code: carbon,
33 in the moderately sensitive non-small cell lung cancer H838
[
17]
grey; oxygen, red; nitrogen, blue; sulfur, yellow; gold, orange.
5
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
2 2
Figure 7: Lateral and axial views of the inclusion complexes. A. Depiction of [33/(-CD) ]; shown is the major component (s.o.f. 71%). B. Depiction of [39/(-CD) ];
shown is the major component (s.o.f. 52%). C. Depiction of [43/(-CD) ]. For A–C: Hydrogens and water molecules are omitted for clarity and the -cyclodextrin
2
cones are shown as capped-stick model whereas 33, 39 and 43 are shown in ellipsoids (on a 50% probability level). Color code: carbon, grey; oxygen, red; nitrogen,
blue; sulfur, yellow; gold, orange.[
17]
cell line in the presence of different amounts of SBE--CD.
Formulations of 33/SBE--CD in ratios of 1:32, 1:16, 1:8, and 1:2
produced high cytoplasmic ROS at similar levels (Figure 8A). In
contrast, only a 1:2 33/SBE--CD formulation generated high
ROS in mitochondria, while all other ratios showed no difference
to a control sample. This suggests that SBE--CD is not a passive
solubilizing agent in these experiments, but that it provides a
“mito-protective” effect. Interestingly, this effect appears not to be
universal: in an identical experimental setup, auranofin produced
high levels of ROS in both the cytoplasm and mitochondria up to
a 32-fold excess of SBE--CD (Figure 8B).
To the best of our knowledge,
a mito-protective role of
cyclodextrins has not been described before. Cyclodextrins in
general don’t cross cell membranes, and SBE--CD, which has
on average seven sulfonates per CD molecule, should also have
Figure 8: Measurement of ROS levels in the cytoplasm and mitochondria using
the roGFP2-Orp1 sensor. A. Kinetic data after treatment of H838 cells with 33
(12.5 M) formulated with different amounts of SBE--CD. B. Kinetic data after
treatment of H838 cells with auranofin (AUR) (12.5 M) either from a DMSO
stock solution or formulated with different amounts of SBE--CD. For A and B:
no passive cell permeability.[
25,26]
The possibility that formulation
in a cell-impermeable host can selectively impact a sub-cellular
drug phenotype is enticing, but the exact mechanism awaits
elucidation. The effect may result from gold(I)-dtc/SBE--CD
inclusion complexes, which then form higher-order oligomers
“No drug” is a control experiment with only vehicle to illustrate the background
oxidation state of the sensor. OxD is the degree of probe oxidation with the data
normalized to 1.0 (fully oxidized) defined by the signal from diamide (2 mM) and
0.0 (fully reduced) defined by DTT (10 mM). Data is presented as the mean of
consisting of drug-loaded and free SBE--CD molecules. Such
oligomeric structures have been previously described,[
27]
and if
3
technical replicates.
active transport across the membrane (e.g. endocytosis) was
possible, an explanation for the effect could be envisaged. The
fact that auranofin produces mitochondrial ROS in the presence
of excess SBE--CD suggests that formation of an inclusion
complex, which is unlikely for auranofin due to its size, is required
for mito-protection. It cannot be excluded, however, that auranofin
induces mitochondrial ROS via a different mechanism of action
than gold(I)-dtc complexes. Studies to understand this effect in
greater detail are currently underway in our laboratory, and will be
reported along with pre-clinical studies of gold(I)-dtc complexes
as anti-cancer agents.
Conclusion
We found that the two drugs disulfiram and aurothiomalate, when
combined in the presence of -cyclodextrins, are selectively
cytotoxic against a collection of cancer cell lines, and show no
cross-resistance in therapy-surviving HT-29 cells. The two drugs
react to make the dimeric gold(I)-dtc complex 5, which we showed
to be the active component, and subsequently prepared >35
analogs for further study. Crystallographic analysis of selected
6
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
(i.e. [AuIII + 2 dtc ] ) specific for these gold(I)-dtc complexes was regularly
found.
– +
dimeric gold(I)-dtc complexes alone, or as inclusion complexes
with -cyclodextrin, revealed a diversity of conformations for the
central eight-membered scaffold, as well as interesting
stereochemical considerations with respect to cyclic and non-
symmetric dtc ligands. Finally, we discovered that when
formulated with an excess of SBE--CD, the gold(I)-dtc complex
Crystallization of inclusion complexes: A mixture of the gold(I)-dtc
complexes (0.05 mmol) and -cyclodextrin (0.05 mmol) in water (5.0 mL)
HPLC grade), resulting in a 10 mM mixture of the components in a 1:1
(
host/guest ratio, was vigorously stirred at 50 °C for 3 h. Then the
mixtures/suspensions were taken up in a 10 mL syringe and filtered
through a syringe filter (Millex ®-HV, 0.45 m, Merck-Millipore) into a
20 mL glass vial. Those vials were left for slow evaporation in a ventilated
fume hood for several days. When initial crystal formation was detected,
the vials were capped and transferred to a 4 °C refrigerator for several
days (and left there until the actual X-ray measurements). For the NMR
experiments of these inclusion complexes, the crystalline material was
filtered through a round filter paper and washed with a minimal amount of
water.
33 does not induce mitochondrial ROS. The same formulation of
the related gold(I) complex auranofin provides no protection
against mitochondrial ROS, suggesting that SBE--CD has
untapped potential beyond just a passive solubilizing agent.
Experimental Section
Chemistry
Biology
General: Reactants and reagents were obtained from Sigma-Aldrich
(
(
Germany), Alfa Aesar (Germany), Fisher Scientific (Germany), Carl Roth
Germany) and Enamine (Ukraine), and were used without further
Quantification of drug sensitivity (Figure 2A): Cell lines (including HT-
2
9) were obtained from ATCC or as gifts from other research groups at the
purification. Deionized water was used from in-house supply. Microwave
reactions were done using a Biotage Initiator+ Microwave Synthesizer.
NMR spectra were recorded on a Bruker 400 MHz instrument (DKFZ) and
DKFZ and checked for mycoplasma contamination and authenticity before
use. Cell lines were seeded in 96-well plates in RPMI-1640 medium,
supplemented with 10% BCS (bovine calf serum) and 1% Pen/Strep at a
density of 2 x 104 cells/well. Disulfiram (DSF) and sodium aurothiomalate
on
a Bruker Avance III 600 MHz instrument (Inorganic Chemistry
Department, University of Heidelberg). High resolution mass spectrometry
was recorded on a Bruker ApexQe FT-ICR instrument, (Department of
Organic Chemistry, University of Heidelberg). Elemental analyses were
performed on a vario Micro Cube by the “Microanalysis Laboratory”
(ATM) were added at a fixed ratio of 1:1, formulated in a 30-fold molar
access of 2HP--CD (2-hydroxypropyl--cyclodextrin). ATM and DSF
monotherapy was also done in the presence of 2HP--CD, to control for
possible combination effects by 2HP--CD. All concentrations were
applied in triplicate. After incubation at 37 °C for 72 h, CellTiter-Blue
(Department of Organic Chemistry, University of Heidelberg). X-ray
crystallography was either performed on a Bruker APEX II Quazar (with
Mo-Microsource) or a Stoe Stadivari (with Cu-Microsource and Pilatus-
detector) instrument (Department of Organic Chemistry, University of
Heidelberg).
(
Promega) was added (10 µL/well). After incubation at 37 °C for 4 h,
fluorescence (excitation filter: 550 nm, emission filter 590 nm) was
determined in an Optima FLUOstar microplate reader (BMG Labtech)
Mean values were calculated for each treatment and EC50 values were
calculated from non-fitted dose response curves
.
.
Synthesis of 5 via ATM and DSF: A solution of ATM (195.0 mg,
0
0
.5 mmol, 1.0 equiv.) in water (10.0 mL) and a solution of DSF (148.3 mg,
.5 mmol, 1.0 equiv.) in EtOH (10.0 mL) were mixed and vigorously stirred
Quantification of factor of resistance (Figure 2B and 2C): 10 million
HT-29 human colon cancer cells were seeded in 175 mL Falcon flasks.
After 24 h, 10 µM oxaliplatin was added. After 48 h incubation at 37 °C,
the medium was replaced with fresh medium. After 48 h, cells were
trypsinized and seeded in 96-well plates at a density of 2 x 104 cells
per well. Human skin fibroblasts (obtained from the Darai lab, Institute
of Virology, University of Heidelberg) and untreated HT-29 cells were
seeded in 96-well plates at the same density. After 24 h incubation at
at room temperature for 4 h. The resulting precipitate was filtered using a
glass frit (pore 4), thoroughly washed with water (200 mL) and afterwards
dried under high vacuum overnight. The pure product was obtained as an
orange-colored powder in a yield of 84% (145.6 mg, 0.21 mmol). MALDI-
MS (m/z) [AuIII + 2 dtc–]+ calcd. for
20AuN
2 2 4
93.026. Anal. calcd for C10H20Au N S : C, 17.39; H, 2.92; N, 4.06; Found:
C
10
H
2 4
S : 493.018; Found:
4
C, 17.69; H, 3.20; N, 3.83. The precipitate was recrystallized from hot DMF
to afford orange-colored needles. MALDI-MS (m/z) [AuIII + 2 dtc–]+ calcd.
3
7 °C, cells were treated in triplicates with 1:2 serial dilutions of
vincristine sulfate (starting concentration 1 µg/mL; 1.1 µM), oxaliplatin
for
C
10
H
20AuN
: C, 17.39; H, 2.92; N, 4.06; Found: C, 17.54; H, 3.14; N,
.74. The IR-spectra of both the initially obtained and recrystallized
2
S
4
:
493.018; Found: 493.018. Anal. calcd for
(
(
starting concentration 50 µg/mL; 126 µM), 5-fluorouracil (5-FU)
starting concentration 500 µM), doxorubicin (starting concentration
10 2 2 4
C H20Au N S
3
1
3
0 µg/mL; 17.2 µM) and DSF/ATM (molar ratio of 1:2), formulated in
0-fold molar excess of 2HP--CD (starting concentration 2 mM ATM).
material are identical (Figure S1).
After incubation for 72 h at 37 °C, 10 µL of CellTiter-Blue was added
per well. After incubation at 37 °C for 4 h, fluorescence (excitation filter:
General procedure for the synthesis of gold(I)-dtc complexes: To a
solution of the appropriate amine (0.92 mmol, 1.2 equiv.) and potassium
hydroxide (0.92 mmol, 1.2 equiv.; for amine hydrochlorides 2.4 equiv. of
KOH was used) in water (10.0 mL) was added carbon disulfide
5
50 nm, emission filter 590 nm) was determined in an Optima
FLUOstar microplate reader (BMG Labtech). Mean values were
calculated for each treatment and IC50 values were calculated from
non-fitted dose response curves. Calculation of "Factor of resistance":
For each treatment EC50 values in oxaliplatin treatment surviving cells
were divided by the EC50 values found with the same treatment in
(
0.92 mmol, 1.2 equiv.). The solution was stirred at room temperature for
in order to assure complete formation of the corresponding
dithiocarbamate (dtc). Then a solution of sodium aurothiomalate (ATM)
0.77 mmol, 1.0 equiv.) in water (15.0 mL) was added to the solution. A
2
h
(
unselected HT-29 cells.
A factor of resistance > 1 indicates
yellow to orange-colored precipitate formed immediately. The suspension
was stirred at room temperature overnight, filtered through a glass frit
"
resistance". A factor of resistance = 1 indicates "no resistance".
(
pore 4), and thoroughly washed with distilled water (2 x 20 mL). This
Measurement of ROS levels in cytoplasm and mitochondria
Figure 8): The effects of 33 and auranofin on roGFP2-Orp1 oxidation
material was dried under high vacuum overnight (and, if necessary,
recrystallized from an appropriate solvent) to afford the pure product. In
most cases, due to the poor solubility of the gold(I)-dtc complexes,
analytics are based on elemental analysis and high-resolution mass
spectrometry (ESI). Concerning HRMS, a diagnostic fragmentation ion
(
were quantified in H838 cells stably expressing roGFP2-Orp1 targeted
either to cytoplasm or mitochondria as described in Ezerina et al.[28]
Specifically, a 5 mM stock solution of 33 was prepared by directly
dissolving solid 33 into 10 mM, 40 mM, 80 mM, and 160 mM solutions of
SBE--CD (Biomol) in PBS (Gibco). A 100 mM auranofin (Sigma-Aldrich)
7
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
solution was prepared in DMSO (Sigma-Aldrich). This was further diluted
to 0.5 mM stock solutions of auranofin in 0 mM, 4 mM, 8 mM, and 16 mM
SBE--CD in PBS. Cells were seeded into a black clear-bottomed 96-well
imaging plate (Corning) at a density of 10,000 cells/well in 100 µL
Fluorobrite medium (Gibco) 16 to 18 h before the measurement. Non-
transduced H838 cells were seeded on the same plate to establish
background fluorescence. The fluorescence intensity from these wells was
subtracted from all the measurement results. To establish the signal
baseline, the entire plate was measured every 3 min for 8 cycles in a
CLARIOstar fluorescence plate reader (BMG Labtech), with excitation
wavelengths of 400 nm and 485 nm, while measuring emission at 520 nm.
Next, 100 µL of 2 times concentrated drug was added, and measurement
was continued for up to 100 cycles. 2 mM diamide (3-
The Supporting Information of this article contains: Experimental
details for the synthesis of each compound and corresponding
NMR spectra, details of the kinetic solubility and X-ray
crystallographic measurements, depiction of the experimental
structure of [37/(-CD)
the inclusion complexes,
2
], the IR spectrum of 5, NMR spectra of
detailed description of the
a
determination of structural parameters of the inclusion complexes
as well as a list of all gold(I)-dtc complexes deposited at
Cambridge Crystallographic Data Centre (CCDC). Deposition
numbers CCDC 2068758 (for 17), CCDC 2068757 (for 23), CCDC
2
068754 (for [33/(-CD)
CCDC 2068755 (for [39/(-CD)
CD) ]) contain the supplementary crystallographic data for this
paper. These data are provided free of charge by the joint
Cambridge Crystallographic Data Center and
2
]), CCDC 2068759 (for [37/(-CD)
2
]),
2
]) and CCDC 2068756 (for [43/(-
(
(
dimethylcarbamoylimino)-1,1-dimethylurea)
dithiothreitol) were used as positive and negative controls, respectively,
and
10 mM
DTT
2
for probe oxidation. Fluorobrite medium was used as a non-treated control.
The readout of the roGFP2 measurement is normalized to the positive
(1.0) and the negative (0) controls using the following equation:
Fachinformationszentrum Karlsruhe Access Structures service.
ꢌꢍꢎꢎꢏꢐꢑꢒꢓꢔ∗ꢌꢍꢕꢖꢗꢔꢘ ꢙ ꢌꢍꢎꢎꢗꢔꢘ ∗ꢌꢍꢕꢖꢏꢐꢑꢒꢓꢔ
ꢂꢃꢄꢅꢆꢇꢈꢉꢊ ꢋ ꢌꢍꢎꢎ
ꢏꢐꢑꢒꢓꢔ∗ꢚꢌꢍꢕꢖꢗꢔꢘꢙꢌꢍꢕꢖꢛꢜꢝ ꢞ ꢌꢍꢕꢖꢏꢐꢑꢒꢓꢔ∗ꢚꢌꢍꢎꢎꢛꢜꢙꢌꢍꢎꢎꢗꢔꢘ
ꢝ
Abbreviations
where I400 and I485 are the signal intensities measured at 520 nm after
excitation at 400 nm and 485 nm, respectively. The subscripts “ox” and
“
red” indicate the signal intensity of fully oxidized and fully reduced
2-HP--CD, 2-hydroxypropyl--cyclodextrin; ATM, sodium
aurothiomalate; CD, cyclodextrin; dtc, dithiocarbamate; DSF,
disulfiram; s.o.f., site occupancy factor; DTT, dithiothreitol.
controls, respectively. All treatments were performed in technical triplicate.
Acknowledgements
Keywords: gold dithiocarbamates • beta-cyclodextrin • inclusion
complexes • host–guest interactions • drug repurposing
We gratefully acknowledge Margit Brückner for setting up the X-
ray structure measurements as well as Frank Rominger
[1]
[2]
a) F. S. Collins, Nature Rev. Drug Discov. 2011, 10, 397; b) Z.
Zhang, L. Zhou, N. Xie, E. C. Nice, T. Zhang, Y. Cui, C. Huang, Sig.
Transduct. Target. Ther. 2020, 5, 113.
E. Ekinci, S. Rohondia, R. Khan, Q. P. Dou, Recent Pat. Anticancer
Drug Discov. 2019, 14, 113–132.
Z. Skrott, M. Mistrik, K. K. Andersen, S. Friis, D. Majera, J. Gursky,
T. Ozdian, J. Bartkova, Z. Turi, P. Moudry, M. Kraus, M. Michalova,
J. Vaclavkova, P. Dzubak, I. Vrobel, P. Pouckova, J. Sedlacek, A.
Miklovicova, A. Kutt, J. Li, J. Mattova, C. Driessen, Q. P. Dou, J.
Olsen, M. Hajduch, B. Cvek, R. J. Deshaies, J. Bartek, Nature 2017,
552, 194–199.
During the course of our studies the marketing of sodium
aurothiomalate (ATM) was ended by several pharma companies
(https://go.drugbank.com/drugs/DB09276; retrieved on April 9th
(Department of Organic Chemistry, University of Heidelberg) for
elucidating X-ray structural data and very fruitful discussions. We
thank the Redox Regulation Division (Tobias Dick, DKFZ) for the
roGFP2-Orp1 cell lines and for helpful discussions in the
interpretation of data. We thank Peter Sehr (EMBL, Heidelberg)
for help with kinetic solubility measurements and Gabriele
Schwebel (DKFZ) as well as Karel Klika (DKFZ) for NMR
spectroscopy support. We also thank Peter Comba (Department
of Inorganic Chemistry, University of Heidelberg) for use of the
FT-IR instrument. Financial support from the Helmholtz Validation
Fund (HVF L18608) and the Else-Kröner-Fresenius Stiftung
[
3]
[
[
[
4]
5]
6]
2021).
a) E. R. T. Tiekink, Inflammopharmacol. 2008, 16, 138–142; b) E.
R. T. Tiekink, Crit. Rev. Oncol. Hematol. 2002, 42, 225–248; c) C. I.
Yeo, K. K. Ooi, E. R. T. Tiekink, Molecules 2018, 23, 1410; d) J. H.
Kim, E. Reeder, S. Parkin, S. G. Awuah, Sci. Rep. 2019, 9, 12335.
For conceptionally related work using DSF in combination with
auranofin see: a) R. E. Kast, J. A. Boockvar, A. Brüning, F. Cappello,
W.-W. Chang, B. Cvek, Q. P. Dou, A. Duenas-Gonzalez, T. Efferth,
D. Focosi, S. H. Ghaffari, G. Karpel-Massler, K. Ketola, A.
Khoshnevisan, D. Keizman, N. Magné, C. Marosi, K. McDonald, M.
Muñoz, A. Paranjpe, M. H. Pourgholami, I. Sardi, A. Sella, K. S.
Srivenugopal, M. Tuccori, W. Wang, C. R. Wirtz, M.-E. Halatsch,
Oncotarget 2013, 4, 502–530; b) H. Huang, Y. Liao, N. Liu, X. Hua,
J. Cai, C. Yang, H. Long, C. Zhao, X. Chen, X. Lan, D. Zang, J. Wu,
X. Li, X. Shi, X. Wang, J. Liu, Oncotarget 2016, 7, 2796–2808.
a) M. E. Davis, M. E. Brewster, Nature Rev. Drug Discov. 2004, 3,
(2019_T16) is gratefully acknowledged.
Author Contributions
M. M. synthesized compounds and analyzed inclusion
complexes. P. F. and E. A. performed biological experiments. N.
G. supervised biological experiments. A. K. M. supervised
chemistry experiments. M. M. and A. K. M. wrote the manuscript.
All authors have given approval to the final version of the
manuscript.
[7]
1023–1035; b) B. dos Santos Lima, S. Shanmugam, J. de Souza
Siqueira Quintans, L. J. Quintans-Júnior, A. A. de Souza Araújo,
Phytochem. Rev. 2019, 18, 1337–1359; c) S. R. Gandhi, J. D. S. S.
Quintans, G. R. Gandhi, A. A. D. S. Araújo, L. J. Quintans Júnior,
Expert Opin. Drug Deliv. 2020, 17, 1069–1080.
[
8]
a) D. D. Heinrich, J.-C. Wang, J. P. Fackler. Jr., Acta Cryst. 1990,
C46, 1444–1447; b) D. Paliwoda, P. Wawrzyniak, A. Katrusiak, J.
Phys. Chem. Lett. 2014, 5, 2182–2188; c) The precipitate was
analyzed by MALDI-MS and elemental analysis. Furthermore, the
precipitate was recrystallized from hot DMF affording orange
needles that were identified as complex 5 by X-ray crystallography
Conflict of Interest
The authors declare no competing financial interests.
(see reference 8b). IR-spectra of both the precipitate and
recrystallized material are identical (see Figure S1) confirming that
the product was not transformed during the recrystallization
procedure.
D. J. Lewis, P. Deshmukh, A. A. Tedstone, F. Tuna, P. O'Brien,
Chem. Commun. 2014, 50, 13334–13337.
Associated Content
[
9]
8
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
[
10]
a) L. Ronconi, L. Giovagnini, C. Marzano, F. Bettìo, R. Graziani, G.
Pilloni, D. Fregona, Inorg. Chem. 2005, 44, 1867–1881; b) X. Zhang,
M. Frezza, V. Milacic, L. Ronconi, Y. Fan, C. Bi, D. Fregona, Q. P.
Dou, J. Cell. Biochem. 2010, 109, 162–172; c) M. Altaf, M. Monim-
ul-Mehboob, A. A. Isab, V. Dhuna, G. Bhatia, K. Dhuna, S. Altuwaijri,
New J. Chem. 2015, 39, 377–385; d) A. A. A. Sulaiman, M. Altaf, A.
A. Isab, A. Alawad, S. Altuwaijri, S. Ahmad, Z. Anorg. Allg. Chem.
2016, 642, 1454–1459; e) C. K. Adokoh, RSC Adv. 2020, 10, 2975–
2988.
[
[
11]
12]
a) S. Jacob, A. B. Nair, Drug Dev. Res. 2018, 79, 201–217; b) M.
Fukuda, D. A. Miller, N. A. Peppas, J. W. McGinity, Int. J. Pharm.
2008, 350, 188–196.
a) H. J. A. Blaauw, R. J. F. Nivard, G. J. M. Van Der Kerk, J.
Organomet. Chem. 1964, 2, 236–244; b) J. Dobrowolski, Z.
Badkowska, I. Kwiatkowska, Roczniki Chemii, Ann. Soc. Chim.
Polonorum 1976, 50, 53–60; c) J. B. Miller, J. L. Burmeister, Synth.
React. Inorg. Met.-Org. Chem. 1985, 15, 223–233; d) P. Bishop, P.
Marsh, A. K. Brisdon, B. J. Brisdon, M. F. Mahon, J. Chem. Soc.,
Dalton Trans. 1998, 675–682; e) M. A. Mansour, W. B. Connick, R.
J. Lachicotte, H. J. Gysling, R. Eisenberg, J. Am. Chem. Soc. 1998,
1
20, 1329–1330.
S. Han, O.-S. Jung, Y.-A. Lee, Transition Met. Chem. 2011, 36,
91–697.
a) F. Baril-Robert, M. A. Radtke, C. Reber, J. Phys. Chem. C 2012,
16, 2192–2197; b) S. Naeem, S. A. Serapian, A. Toscani, A. J. P.
White, G. Horgarth, J. D. E. T. Wilton-Ely, Inorg. Chem. 2014, 53,
404–2416.
[
[
13]
14]
6
1
2
[
15]
a) E. Amtmann, N. Gunkel, A. K. Miller, M. Morgen,“Compositions
Comprising a Metal Source, Dithiocarbamate and Cyclodextrin”,
WO 2018/069525, 2018; b) E. Amtmann, N. Gunkel, A. K. Miller, M.
Morgen, “Improved Dithiocarbamate Based Compounds, Therapies
and Diagnostics”, WO 2020/161337, 2020.
[
[
16]
17]
H. Schmidbaur, A. Schier, Chem. Soc. Rev. 2008, 37, 1931–1951.
The experimental structures are depicted using Mercury 4.2.0
(
http://www.ccdc.cam.ac.uk/mercury/).
[
[
18]
19]
 is shown to adopt values between 0° and ~34° (see Table S1).
a) L. X. Song, L. Bai, X. M. Xu, J. He, S. Z. Pan, Coord. Chem. Rev.
2009, 253, 1276–1284; b) F. Hapiot, S. Tilloy, E. Monflier, Chem.
Rev. 2006, 106, 767–781.
[
20]
a) W. Saenger, T. Steiner, Acta Cryst. Sec. 1998, A54, 798–805; b)
J. Szejtli, Encyclopedia Nanosci. Nanotechnol. 2004, 2, 283–304; c)
G. Crini, Chem. Rev. 2014, 114, 10940–10975.
[
[
[
21]
22]
23]
H.-J. Schneider, F. Hacket, V. Rüdiger, H. Ikeda, Chem. Rev. 1998,
9
8, 1755–1785.
A fourth, less well-resolved X-ray structure of the inclusion complex
37/(-CD) ] was also obtained (see Figure S6).
[
2
A thorough CCDC search on experimental structures with (plain) -
cyclodextrin performed on November 24th 2020 revealed 336
entries. We identified 107 entries of inclusion complexes with a 1:1
host/guest ratio (of which 9 inclusion complexes were host/metal
complexes), 16 entries with a 2:1 host/guest ratio, 38 entries with a
2:2 host/guest ratio, 1 entry with a 3:2 host/guest ratio, and 33
entries which we considered encapsulation interactions of -
cyclodextrin with the guest molecule. Among the sixteen 2:1
host/guest complexes, only
2 entries showed a "tail-to-tail"
arrangement of the CD cones (CCDC reference codes OCIGAK and
SATFEB). In this evaluation we only included inclusion complexes
with small organic or metal-organic guest molecules - solvates or
salt adducts of -cyclodextrin were excluded.
S. J. Berners-Price, A. Filipovska, Metallomics 2011, 3, 863–873.
M. Gutscher, M. C. Sobotta, G. H. Wabnitz, S. Ballikaya, A. J.
Meyer, Y. Samstag, T. P. Dick, J. Biol. Chem. 2009, 284, 31532–
[
[
24]
25]
3
1540.
P. Jansook, N. Ogawa, T. Loftsson, Int. J. Pharm. 2018, 535, 272–
84.
[
[
26]
27]
2
a) T. Loftsson, A. Magnúsdóttir, M. Másson, J. F. Sigurjónsdóttir, J.
Pharm. Sci. 2002, 91, 2307–2316; b) T. Loftsson, M. Másson, H. H.
Sigurdsson, Int. J. Pharm. 2002, 232, 35–43.
[
28]
D. Ezeriņa, Y. Takano, K. Hanaoka, Y. Urano, T. P. Dick, Cell Chem.
Biol. 2018, 25, 447–459.
9
This article is protected by copyright. All rights reserved.

Chemistry - A European Journal
10.1002/chem.202101366
FULL PAPER
Entry for the Table of Contents
Inclusion complex of a complex: The two drugs disulfiram and aurothiomalate react to produce a gold(I)-dithiocarbamate complex,
which selectively kills cancer cells. Hydroxy-substituted derivatives of this complex form inclusion complexes with -cyclodextrin in a
rare “tail-to-tail” arrangement. Formulation of one gold complex with a -cyclodextrin derivative prevents formation of reactive oxygen
species in mitochondria.
Institute and/or researcher Twitter usernames: @DrugDiscov_DKFZ @DKFZ
This article is protected b y1 0c opyright. All rights reserved.