Comparative biological evaluation and G-quadruplex interaction studies of two new families of organometallic gold(I) complexes featuring N-heterocyclic carbene and alkynyl ligands

Article abstract of DOI:10.1016/j.jinorgbio.2019.110844

Experimental organometallic gold(I) compounds hold promise for anticancer therapy. This study reports the synthesis of two novel families of gold(I) complexes, including N1-substituted bis-N-heterocyclic carbene (NHC) complexes of general formula [Au(N1-T

Full text of DOI:10.1016/j.jinorgbio.2019.110844

Journal Pre-proof
Comparative biological evaluation and G-quadruplex interaction
studies of two new families of organometallic gold(I) complexes
featuring N-heterocyclic carbene and alkynyl ligands
Samuel M. Meier-Menches, Brech Aikman, Daniel Döllerer, Wim
T. Klooster, Simon J. Coles, Nicolò Santi, Louis Luk, Angela
Casini, Riccardo Bonsignore
PII:
S0162-0134(19)30300-9
DOI:
https://doi.org/10.1016/j.jinorgbio.2019.110844
JIB 110844
Reference:
To appear in:
Journal of Inorganic Biochemistry
Received date:
Revised date:
Accepted date:
20 May 2019
7 September 2019
8 September 2019
Please cite this article as: S.M. Meier-Menches, B. Aikman, D. Döllerer, et al.,
Comparative biological evaluation and G-quadruplex interaction studies of two new
families of organometallic gold(I) complexes featuring N-heterocyclic carbene and
alkynyl ligands, Journal of Inorganic Biochemistry (2018), https://doi.org/10.1016/
j.jinorgbio.2019.110844
This is a PDF file of an article that has undergone enhancements after acceptance, such
as the addition of a cover page and metadata, and formatting for readability, but it is
not yet the definitive version of record. This version will undergo additional copyediting,
typesetting and review before it is published in its final form, but we are providing this
version to give early visibility of the article. 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.
© 2018 Published by Elsevier.

Journal Pre-proof
Comparative biological evaluation and G-quadruplex interaction
studies of two new families of organometallic gold(I) complexes
featuring N-heterocyclic carbene and alkynyl ligands
Samuel M. Meier-Menches,a,+ Brech Aikman,a+ Daniel Döllerer,a Wim T. Klooster,b Simon J.
Coles,b Nicolò Santi,a Louis Luk,a Angela Casini,a,c* and Riccardo Bonsignorea,*
a School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF103AT, UK.
b School of Chemistry, University of Southampton, Southampton SO17 1BJ, United Kingdom.
c
Department of Chemistry, Technical University of Munich, Lichtenbergstr. 4, 85748 Garching,
Germany.
+shared first authors
* Corresponding authors: E-mail: casinia@cardiff.ac.uk; bonsignorer@cardiff.ac.uk;
Ph/Fax: +44 (0)29 2087 6364
Keywords: gold(I) organometallics; N-heterocyclic carbenes; alkynyl ligands; G-quadruplexes;
cancer.
Abstract
1

Journal Pre-proof
Experimental organometallic gold(I) compounds hold promise for anticancer therapy. This study
reports the synthesis of two novel families of gold(I) complexes, including N1-substituted bis-N-
heterocyclic carbene (NHCs) complexes of general formula [Au(N1-TBM)2]BF4 (N1-TBM = N1-
substituted 9-methyltheobromin-8-ylidene) and mixed gold(I) NHC-alkynyl complexes, [Au(N1-
TBM)alkynyl]. The compounds were fully characterized for their structure and stability in aqueous
environment and in the presence of N-acetyl cysteine by nuclear magnetic resonance (NMR)
spectroscopy. The structures of bis(1-ethyl-3,7,9-trimethylxanthin-8-ylidene)gold(I), (4-
ethynylpyridine)(1,9-dimethyltheobromine-8-ylidene)gold(I)
and
of
(2,8-Diethyl-10-(4-
ethynylphenyl)-5,5-difluoro-1,3,7,9-tetramethyl-5H-4λ4,5λ4-dipyrrolo[1,2-c:2',1'-
f][1,3,2]diazaborinine)(1,3,7,9-tetramethylxanthin-8-ylidene)gold(I) were also confirmed by X-ray
diffraction analysis. The compounds were studied for their properties as DNA G-quadruplexes
(G4s) stabilisers by fluorescence resonance energy transfer (FRET) DNA melting. Only the cationic
[Au(N1-TBM)2]BF4 family showed moderate G4 stabilization properties with respect to the
previously reported benchmark compound [Au(9-methylcaffein-8-ylidene)2]+ (AuTMX2).
However, the compounds also showed marked selectivity for binding to G4 structures with respect
to duplex DNA in competition experiments. For selected complexes, the interactions with G4s were
also confirmed by circular dichroism (CD) studies. Furthermore, the gold(I) complexes were
assessed for their antiproliferative effects in human cancer cells in vitro, displaying moderate
activity. Of note, among the mixed gold(I) NHC-alkynyl compounds, one features a fluorescent
boron-dipyrromethene (BODIPY) moiety which allowed determining its uptake into the cytoplasm
of cancer cells by fluorescence microscopy.
2

Journal Pre-proof
1. INTRODUCTION
The medicinal use of gold is known since ancient Egyptian times and in the 17th century colloidal
gold was applied for treatment of fever and sickness [1,2]. At the beginning of the 1900s, Koch and
co-workers described the bacteriostatic activity of K[Au(CN)2], and 30 years later, a gold(I)
phosphane-thiolate complex, Auranofin (Fig. 1), was recognised as an anti-rheumatic
metallopharmaceutical capable of reducing inflammation and the progression of this chronic disease
[3]. Subsequently, Auranofin has been evaluated in vitro for its anticancer properties [4,5]. It is
nowadays actively being repurposed in pilot trials and clinical studies for anticancer therapy against
chronic lymphocytic leukaemia and ovarian cancer among others [6,7].
Potential gold-based pharmaceuticals are prepared in the oxidation states +1/+3. In both
redox states, gold complexes can be easily reduced to colloidal gold, which leads to metallodrug
inactivation or to unselective cellular damage. In order to achieve the control of the metal
complex’s reactivity and speciation in biological environment, organometallic compounds,
featuring a direct metal-carbon bond which further stabilises the metal centre, have been widely
explored in medicinal chemistry [8,9]. In this context, various families of gold(I) and gold(III)
organometallic compounds have been recently designed as a source of stable and biologically active
molecules [9–11]. For example, gold(I) N-heterocyclic carbenes (NHCs) are of particular interest
since the strong electron-donating properties of the carbene ligand stabilises the gold(I) centre
against reduction, even under physiological conditions [10,12]. Moreover, the NHC core structure
can be chemically modified conferring different physicochemical and functional properties to the
resulting metallopharmaceutical. Thus, since their first report, several gold(I) NHC complexes have
been shown to possess promising cytotoxic properties against cancer cells, or even antibacterial
effects [13–24]. The proposed mechanism of action of this class of oganometallic complexes has
been linked to their affinity for binding to thiol and selenol groups in protein targets. For example,
the seleno-enzyme thioredoxin reductase (TrxR), involved in maintaining the redox homeostasis in
3

Journal Pre-proof
cells, has been shown to be efficiently inhibited by gold(I)/gold(III) complexes [25–27], including
gold(I) NHCs [21,28–30].
In the last years, metal NHCs complexes derived from xanthines, including of platinum(II)
and silver(I), have attracted much attention for therapeutic applications. In fact, the easy
functionalization of xanthines, combined with their biological activity, makes their metal
derivatives excellent candidates for the development of novel metallodrugs[31]. Recently, we
reported on a cationic gold(I) bis-NHC complex [Au(9-methylcaffein-8-ylidene)2]+ (AuTMX2, Fig.
1) that is selectively cytotoxic against the ovarian cancer cell line A2780. Moreover, the compound
was reported to selectively stabilize human telomeric (hTelo) G-quadruplexes (G4s), a type of DNA
secondary structure present in the telomeric regions of the chromosomes [32]. G4s are non-
canonical DNA secondary structures formed in guanine rich sequences and have been identified in
human telomeres and promoter regions of many genes. They are supposed to regulate important
biological processes such as telomere homeostasis, gene transcription and DNA replication [33–35].
Stabilizing these DNA secondary structures by small molecules may interfere with their
physiological function and induce potential anticancer effects, e.g. by inhibiting telomere extensions
or oncogene expression [34–36].
Initial structure activity relationship studies revealed that any modification at the N9-
position of the xanthine in the AuTMX2 scaffold leads to a loss of G4 binding potency and
selectivity. From X-ray diffraction analysis, it was found that any bulky modification at the N9-
position of the xanthine scaffold would force the two carbenes ligands coordinating the gold(I)
centre out of planarity, probably reducing the strength of the interaction with the flat aromatic
hTelo-G4 surface. This was also corroborated by experimental evidences in solution [37].
Furthermore, via a combination of computational and experimental assays, the binding energy of
AuTMX2 with hTelo was elucidated [37].
In the present study, we reasoned that modifications at the N1-position of the xanthine
ligand would retain the planar geometry of the resulting gold(I) bis-NHCs, constituting a promising
4

Journal Pre-proof
path for obtaining further structure-activity relationships for this class of G4-stabilising compounds.
Therefore, eight new complexes were synthesized and characterized by different techniques. Next
to four cationic N1-modified AuTMX2 analogues [Au(N1-TBM)2]BF4 (N1-TBM = N-1 substituted
9-methyltheobromin-8-ylidene), we explored the effect of replacing one of the NHC ligands with an
alkynyl moiety on the G4s stabilizing potency in the new series of general formula [Au(N1-
TBM)alkynyl]. It should be noted that the latter family of compounds is neutral. In a previous work,
a mixed NHC-alkynyl gold(I) complex showed potent cytotoxic activity in human cancer cells
while being scarcely toxic in healthy rat liver tissues ex vivo [20]. Thus, four gold(I)
organometallics featuring mixed NHC-alkynyl ligands were also synthesized and characterized by
different methods. The stability of all the new compounds was studied in aqueous environment and
in the presence of N-acethyl-cysteine (NAC), used as model nucleophile, by NMR spectroscopy.
Afterward, the two families were comparatively evaluated with respect to their G4-stabilization in
vitro. It should be noted that, since the G4-stabilization of AuTMX2 was previously assessed only
on telomeric G4s, the present study evaluates the new gold(I) compounds towards both telomeric
(hTelo) and promoter G4s (C-KIT1 and hTERT) by fluorescence resonance energy transfter (FRET)
DNA melting, and in selected cases also by circular dichroism (CD). Finally, the cytotoxic
properties of the compounds have been studied in a small panel of human cancer cells in vitro, and
one of the derivatives, endowed with fluorescence properties, also charcaterized for its uptake in
melanoma cells by fluorescence microscopy.
Figure 1.
2. MATERIALS AND METHODS
General.
Solvents and reagents (reagent grade) were all commercially available and used without further
1
purification. H, 11B and 13C NMR spectra were recorded in CD3CN (deuterated acetonitrile),
CDCl3 (detuterated chloroform), DMSO-d6 (deuterated dimethyl sulfoxide) or acetone-d6 solution,
5

Journal Pre-proof
with tetramethylsilane (TMS) as an internal reference, on a Bruker Avance NMR spectrometer
(300–500 MHz, Fig. S1-S33). Electrospray ionization (ESI) mass spectra were recorded on a
Synapt G2-Si time-of-flight mass spectrometer (Waters Corporation, Milford, USA). Elemental
analysis were performed by the Microanalytical Laboratory at the Technical University of Munich.
Au(tht)Cl (tht = tetrahydrothiophene) and AuTMX2, AuTMXI ( [Au(9-methylcaffein-8-
ylidene)Iodide], Fig. 1), and the BODIPY-alkynyl ligand were synthesised according procedures
already published in literature [37,38].
Synthesis and characterisation.
1-Ethyl-3,7-dimethylxanthine (1a): Theobromine (200 mg, 1.1 mmol) and potassium carbonate
(228 mg, 2.4 mmol) were suspended in dimethylformamide (DMF, 4 mL) in a round bottom flask.
30 min later, iodoethane (176 μL, 1.6 mmol) was added dropwise and the reaction stirred overnight
at room temperature. The resulting suspension was filtered to achieve a clear solution which was
completely evaporated. The off-white solid was washed 3 times with isopropanol to obtain the
product as a white solid (150 mg, 65%).
1H NMR (400 MHz, CDCl3) δ 7.43 (s, 1H, Ar), 4.02 (q, J = 7.1 Hz, 2H, CH2), 3.92 (d, J = 0.6 Hz,
3H, CH3), 3.51 (s, 3H, CH3), 1.19 (t, J = 7.1, Hz, 3H, CH3).
13C NMR (101 MHz, DMSO-d6) δ 154.71 (Ar), 151.13 (Ar), 148.68 (Ar), 143.36 (Ar), 107.17 (Ar),
35.99 (CH2), 33.64 (CH3), 29.81 (CH3), 13.60 (CH3).
1-(2-Cyclohexylethyl)-3,7-dimethylxanthine (1b): Theobromine (200 mg, 1.1 mmol) and
potassium carbonate (500 mg, 3.6 mmol) were suspended in DMF (8 mL) for 30 min prior to the
addition of (2-bromoethyl)cyclohexane (208 μL, 1.3 mmol). The reaction mixture was heated to
reflux overnight. The day after the suspension was filtered off and the resulting clear yellow
solution was dried to achieve a sticky solid. A white solid was isolated after recrystallization in
isopropanol (240 mg, 75%).
6

Journal Pre-proof
1H NMR (400 MHz, DMSO-d6) δ 7.99 (s, 1H, Ar), 3.86 (s, 3H, CH3), 3.39 (s, 3H, CH3), 1.66 (dt,
J = 28.2, 15.6 Hz, 6H, CH2), 1.43 – 1.36 (m, 2H, CH2), 1.34 – 1.05 (m, 5H, CH and CH2), 0.97 –
0.83 (m, 2H, CH2).
13C NMR (101 MHz, DMSO-d6) δ 154.77 (Ar), 151.19 (Ar), 148.62 (Ar), 143.33 (Ar), 107.11 (Ar),
35.66 (CH2), 35.41 (CH2), 33.61 (CH3), 33.39 (CH2), 33.13 (CH2), 29.81 (CH3), 26.53 (CH2), 26.13
(CH2).
Unfortunately, the synthesis of this compound was succesfull only once and could not be
reproduced to perform complete characterization by mass spectrometry and elemental analysis.
1-Benzyl-3,7-dimethylxanthine (1c): Theobromine (200 mg, 1.1 mmol) and potassium carbonate
(228 mg, 2.4 mmol) were suspended in DMF (4 mL) in a round bottom flask. 30 min later, benzyl
bromide (142 μL, 1.6 mmol), was added dropwise and the reaction stirred overnight at room
temperature. The resulting suspension was completely evaporated and water (20 mL) was added.
The product was afterwards extracted with dichloromethane (DCM, 3 ×15 mL), dried over
magnesium sulfate and fully evaporated. The resulting product was obtained after recrystallization
in methanol (268 mg, 89%).
1H NMR (400 MHz, DMSO-d6) δ 8.05 (s, 1H, Ar), 7.29 (t, J = 4.2 Hz, 4H, Ar), 7.27 – 7.21 (m, 1H,
Ar), 5.06 (s, 2H, CH2), 3.89 (s, 3H, CH3), 3.43 (s, 3H, CH3).
13C NMR (101 MHz, DMSO-d6) δ 154.89 (Ar), 151.45 (Ar), 148.91 (Ar), 143.65 (Ar), 138.09 (Ar),
128.74 (Ar), 127.97 (Ar), 127.50 (Ar), 107.14 (Ar), 44.01 (CH2), 33.68 (CH3), 29.97 (CH3).
3,7-Dimethyl-1-(4-trifluoromethyl)benzylxanthine (1d): A suspension of theobromine (360 mg,
2.0 mmol) in DMF (15 mL) was heated to 50 °C for 10 min before adding 3 equiv. of potassium
carbonate (830 mg, 6 mmol). After 30 min, an equimolar amount of 4-(trifluoromethyl)benzyl
bromide (309 μL, 2.0 mmol) was added and the reaction mixture stirred overnight at 50 °C. Thus,
7

Journal Pre-proof
the suspension was filtered off and the resulting clear solution was dried to achieve a white solid.
Further recrystallization in isopropanol allowed to isolate the final product as an off-white solid
(423 mg, 63%).
1H NMR (300 MHz, CDCl3) δ 7.65 – 7.49 (m, 4H, Ar), 5.25 (s, 2H, CH2), 4.01 (s, 3H, CH3), 3.60
(s, 3H, CH3).
13C NMR (101 MHz, CDCl3) δ 154.12 (Ar), 150.54 (Ar), 148.03 (Ar), 140.75 (Ar), 140.19 (Ar),
128.74 (q, J = 32.32 Hz, Ar, Cα-CF3), 128.02 (Ar), 126.03 (q, J = 272.7 Hz, CF3), 124.37 (q, J =
3.8 Hz, Ar, Cβ-CF3), 106.59 (Ar), 43.02 (CH2), 32.62 (CH3), 28.82 (CH3).
1-Ethyl-3,7,9-trimethylxanthinium tetrafluoroborate (2a): Compound 1a (214 mg, 1.0 mmol)
was dissolved in acetonitrile (MeCN, 8 mL) prior to the addition of trimethyl-oxonium
tetrafluoroborate (243 mg, 1.6 mmol). After 5 min, sodium carbonate (200 mg, 1.9 mmol) was
added and the mixture stirred for 1 h at room temperature under an inert atmosphere. The resulting
suspension was filtered to achieve a clear solution which was further dried. The solid was washed
with Et2O and DCM to afford a sticky solid. The final white solid was obtained after
recrystallization in isopropanol (226 mg, 74%).
1H NMR (400 MHz, CD3CN) δ 8.47 (s, 1H, Ar), 4.08 (s, 6H, CH3), 4.02 (q, J = 7.1 Hz, 2H, CH2),
3.73 (s, 3H, CH3), 1.21 (t, J = 7.1 Hz, 3H, CH3).
13C NMR (126 MHz, CD3CN) δ 154.21 (Ar), 151.02 (Ar), 140.55 (Ar), 139.58 (Ar), 109.48 (Ar),
38.16 (CH3), 37.86 (CH3), 36.59 (CH2), 31.99 (CH3), 12.92 (CH3).
Elemental analysis for C10H19BF4N4O4 (2a  2H2O): C 35.02%, H 5.68%, N 16.37% (calc. C
34.70%, H 5.53%, N 16.19%)
1-(2-Cyclohexylethyl)-3,7,9-trimethylxanthinium tetrafluoroborate (2b): Compound 1b
(500 mg, 1.7 mmol) was dissolved in MeCN (7 mL) prior to the addition of trimethyloxonium
tetrafluoroborate (267 mg, 1.8 mmol). After 5 min, sodium carbonate (288 mg, 2.7 mmol) was
8

Journal Pre-proof
added and stirred for 4 h at 40 °C under an inert atmosphere. The resulting suspension was filtered
to achieve a clear solution which was further dried. A final recrystallization in isopropanol allowed
to obtain the final product as a white solid (304 mg, 89%).
1H NMR (400 MHz, CD3CN) δ 8.43 (s, 1H, Ar), 4.05 (t, J = 2.6 Hz, 6H, CH3), 3.99 – 3.93 (m, 2H,
CH2), 3.69 (s, 3H, CH3), 1.83 – 1.61 (m, 5H, CH and CH2), 1.50 – 1.42 (m, 2H, CH2), 1.38 – 1.13
(m, 4H, CH2), 1.03 - 0.90 (m, 2H, CH2).
13C NMR (101 MHz, CD3CN) δ 154.37 (Ar), 151.16 (Ar), 140.57 (Ar), 139.53 (Ar), 109.50 (Ar),
41.25 (CH2), 37.94 (CH3), 36.68 (CH2), 36.41 (CH3), 35.58 (CH2), 33.82 (CH2), 32.09 (CH3), 27.20
(CH2), 26.93 (CH2).
Unfortunately, due to the impossibility to reproduce the synthesis of compound 1b (see above), this
compound was not obtained in sufficient quantity to perform complete characterization by mass
spectrometry and elemental analysis.
1-Benzyl-3,7,9-trimethylxanthinium tetrafluoroborate (2c): Compound 1c (273 mg, 1.0 mmol)
was dissolved in MeCN (8 mL) prior to the addition of trimethyloxonium tetrafluoroborate
(243 mg, 1.6 mmol). After 5 min, sodium carbonate (288 mg, 2.7 mmol) was added and stirred for
4 h at 40 °C under an inert atmosphere. The resulting suspension was filtered to achieve a clear
solution which was further dried. The solid was washed with Et2O and DCM. Recrystallization
from isopropanol yielded a white solid (357 mg, 93%).
1H NMR (400 MHz, CD3CN) δ 8.48 (s, 1H, Ar), 7.47 – 7.24 (m, 5H, Ar), 5.16 (s, 2H, CH2), 4.14 –
4.02 (m, 6H, CH3), 3.73 (s, 3H, CH3).
13C NMR (101 MHz, CD3CN) δ 154.09 (Ar), 150.97 (Ar), 140.40 (Ar), 139.35 (Ar), 137.10 (Ar),
129.04 (Ar), 128.57 (Ar), 128.23 (Ar), 109.14 (Ar), 45.56 (CH2), 37.59 (CH3), 36.36 (CH3), 31.87
(CH3).
Elemental analysis for C15H25BF4N4O6 (2c  4H2O): C 40.73%, H 5.41%, N 12.55% (calc. C
40.56%, H 5.67%, N 12.61%)
9

Journal Pre-proof
3,7,9-Trimethyl-1-(4-trifluoromethyl)benzylxanthinium tetrafluoroborate (2d): Compound 1d
(202 mg, 0.6 mmol) was dissolved in MeCN (6 mL) prior to the addition of trimethyloxonium
tetrafluoroborate (262 mg, 1.8 mmol). After 5 min, sodium carbonate (288 mg, 2.7 mmol) was
added and stirred for 5 h at 50 °C under an inert atmosphere. The resulting suspension was filtered
to achieve a clear solution which was further dried. The final white solid was obtained after
recrystallization in isopropanol (154.7 mg, 62%).
1H NMR (400 MHz, CD3CN) δ 8.47 (s, 1H, Ar), 7.63 (dd, J = 45.5, 8.1 Hz, 4H, Ar), 5.23 (s, 2H,
CH2), 4.09 (s, 3H, CH3), 4.08 (s, 3H, CH3), 3.74 (s, 3H, CH3).
13C NMR (101 MHz, CD3CN) δ 154.06 (Ar), 150.97 (Ar), 141.62 (Ar), 140.49 (Ar), 139.50 (Ar),
129.69 (q, J = 32.32 Hz, Ar, Cα-CF3), 125.88 (q, J = 3.0 Hz, Ar, Cβ-CF3), 125.25 (q, J = 272.7 Hz,
CF3), 125.88 (Ar), 125.84 (Ar), 123.55 (Ar), 109.15 (Ar), 45.20 (CH2), 37.63 (CH3), 36.37 (CH3),
31.93 (CH3).
Elemental analysis for C16H20BF7N4O4 (2d  2H2O): C 40.14%, H 4.17%, N 11.43% (calc. C
40.36%, H 4.23%, N 11.77%)
Bis(1-ethyl-3,7,9-trimethylxanthin-8-ylidene)gold(I) tetrafluoroborate (3a): Ligand 2a (100 mg,
0.32 mmol) and Ag2O (59.79 mg, 0.26 mmol) were suspended in dry acetonitrile (10 mL) with
molecular sieves (4 Å) under an inert atmosphere. After 6 h at room temperature, Au(tht)Cl
(51.7 mg, 0.16 mmol) was added and the suspension stirred overnight. The day after potassium
iodide (27 mg, 0.16 mmol) was added and the mixture was left under vigorous stirring for further
30 min. The final dark grey suspension was filtered over Celite® and washed with acetonitrile and
DCM to give a clear colourless solution. After concentration to ca. 1 mL, an excess of Et2O was
added to precipitate the final white solid (46 mg, 39%).
10

Journal Pre-proof
1H NMR (400 MHz, CD3CN) δ 4.22 (s, 3H, CH3), 4.19 (s, 3H, CH3), 4.03 (d, J = 6.9 Hz, 2H, CH2),
3.77 (s, 3H, CH3), 1.21 (t, J = 7.0 Hz, 3H, CH3).
13C NMR (126 MHz, CD3CN) δ 188.37 (Ar), 154.35 (Ar), 151.45 (Ar), 141.62 (Ar), 110.28 (Ar),
39.70 (CH3), 38.04 (CH3), 37.91 (CH2), 32.30 (CH3), 13.12 (CH3).
ESI-MS (CH3CN, pos. mode) for C20H27AuN8O4: exp. 641.1909 (calc. 641.1875).
Elemental analysis for C20H32AuBF4N8O6 (3a  2H2O): exp. C 31.87%, H 4.34%, N 14.33% (calc.
C 31.43%, H 4.22 %, N 14.66%)
Bis(1-(2-cyclohexylethyl)-3,7,9-trimethylxanthin-8-ylidene)gold(I) tetrafluoroborate (3b):
Ligand 2b (133 mg, 0.32 mmol) and Ag2O (67 mg, 0.29 mmol) were suspended in dry MeCN
(10 mL) with molecular sieves (4 Å) under an inert atmosphere. After 6 h at room temperature,
Au(tht)Cl (81 mg, 0.25 mmol) was added and the suspension stirred overnight. Thus, potassium
iodide (42 mg, 0.25 mmol) was added and the mixture was left under vigorous stirring for further
30 min. The final dark grey suspension was filtered over Celite® and washed with MeCN and
DCM to give a clear colourless solution. After concentration to ca. 1 mL, an excess of Et2O was
added to precipitate the final white compound (57 mg, 38%).
1H NMR (400 MHz, CD3CN) δ 4.19 (s, 3H, CH3), 4.15 (s, 3H, CH3), 4.13 (s, 3H, CH3), 4.04 (s, 3H,
CH3), 3.95 (ddd, J = 9.8, 7.8, 5.5 Hz, 4H, CH2), 3.73 (s, 3H, CH3), 3.69 (s, 3H, CH3), 1.82 – 1.60
(m, 10H, CH and CH2), 1.50 – 1.40 (m, 4H, CH2), 1.37 – 1.10 (m, 9H, CH2), 1.03 – 0.89 (m, 4H,
CH2).
13C NMR (126 MHz, CD3CN) δ 188.34 (Ar), 176.95 (Ar), 154.46 (Ar), 154.42 (Ar), 151.53 (Ar),
141.60 (Ar), 141.16 (Ar), 110.25 (Ar), 109.71 (Ar), 109.24 (Ar), 40.98 (Ar), 40.90 (CH3), 39.82
(CH3), 39.70 (CH2), 38.16 (CH2), 38.05 (CH2), 36.47 (CH2), 35.75 (CH2), 33.84 (CH3), 32.33 (CH
and CH3), 27.20 (CH3), 26.92 (CH3).
11

Journal Pre-proof
Unfortunately, due to the impossibility to reproduce the synthesis of ligand 2b (see above), this
compound was not obtained in sufficient quantity to perform complete characterization by mass
spectrometry and elemental analysis.
Bis(1-benzyl-3,7,9-trimethylxanthin-8-ylidene)gold(I) tetrafluoroborate (3c): Ligand 2c
(159 mg, 0.42 mmol) and Ag2O (80 mg, 0.35 mmol) were suspended in dry MeCN (10 mL) with
molecular sieves (4 Å) under an inert atmosphere. After 6 h at room temperature, Au(tht)Cl (89 mg,
0.28 mmol) was added and the suspension stirred overnight. Thus, potassium iodide (48 mg,
0.28 mmol) was added and the mixture was left under vigorous stirring for further 30 min. The final
dark grey suspension was filtered over Celite® and washed with acetonitrile and DCM to give a
clear colourless solution. After concentration to ca. 1 mL, an excess of Et2O was added to
precipitate out the final white compound (34 mg, 20%).
1H NMR (400 MHz, CD3CN) δ 7.34 – 7.15 (m, 5H, Ar), 5.06 (s, 2H, CH2), 4.12 (s, 3H, CH3), 4.08
(s, 3H, CH2), 3.67 (s, 3H, CH3).
13C NMR (101 MHz, CD3CN) δ 188.61 (Ar), 154.57 (Ar), 151.76 (Ar), 141.85 (Ar), 137.98 (Ar),
129.36 (Ar), 128.89 (Ar), 128.45 (Ar), 110.27 (Ar), 45.72 (CH2), 39.76 (CH3), 38.11 (CH3), 32.50
(CH3).
Elemental analysis for C30H36AuBF4N8O6 (3c  2H2O): exp. C 40.34%, H 3.75%, N 12.90% (calc.
C 40.56%, H 4.08%, N 12.61%)
Bis(3,7,9-trimethyl-1-(4-trifluoromethyl)benzylxanthin-8-ylidene)gold(I)
tetrafluoroborate
(3d): Ligand 2d (130 mg, 0.30 mmol) and Ag2O (54.8 mg, 0.24 mmol) were suspended in dry
MeCN (10 mL) with molecular sieves (4 Å) under inert atmosphere. After 6 h at room temperature,
Au(tht)Cl (48 mg, 0.15 mmol) was added and the suspension stirred overnight. Thus, potassium
iodide (25 mg, 0.15 mmol) was added and the mixture was left under vigorous stirring for further
30 min. The final dark grey suspension was filtered over Celite® and washed with acetonitrile and
12

Journal Pre-proof
DCM to give a clear colourless solution. After concentration to ca. 1 mL, an excess of Et2O was
added to precipitate out the final white compound (100 mg, 66%).
1H NMR (500 MHz, CD3CN) δ 7.60 (dd, J = 52.1, 8.1 Hz, 2H, Ar), 5.20 (s, 2H, CH2), 4.21 (s, 3H,
CH3), 4.16 (s, 3H, CH3), 3.75 (s, 3H, CH3).
13C NMR (126 MHz, CD3CN) δ 188.75 (Ar), 154.52 (Ar), 151.76 (Ar), 142.53 (Ar), 141.96 (Ar),
129.94 (q, J = 32.76 Hz, Ar, Cα-CF3), 129.51 (Ar), 126.23 (q, J = 3.8 Hz, Ar, Cβ-CF3), 125.32 (q,
J = 270.9 Hz, CF3) 110.29 (Ar), 45.40 (CH2), 39.80 (CH3), 38.14 (CH3), 32.58 (CH3).
ESI-MS (CH3CN, pos. mode) for C32H30AuF6N8O4: exp. 901.1976 (calc. 901.1960).
Elemental analysis for C32H40AuBF10N8O9 (3d  H2O): C 38.37%, H 3.13%, N 11.35% (calc. C
38.19%, H 3.21%, N 11.13%)
(Ethynylphenyl)(1,3,7,9-tetramethylxanthin-8-ylidene)gold(I) (4a): Phenylacetylene (16 μL,
0.14 mmol), was added to NaOH (40 mg, 1 mmol) in methanol (10 mL). After stirring for 15 min at
room temperature, AuTMXI (75 mg, 0.14 mmol) was added and the mixture stirred under reflux
overnight. After filtration, the final product was collected as an off-white solid (50 mg, 69%).
1H NMR (400 MHz, DMSO-d6) δ 7.25 (dd, J = 7.8, 4.1 Hz, 4H, Ar), 7.23 – 7.16 (m, 1H, Ar), 4.20
(s, 3H, CH3), 4.05 (s, 3H, CH3), 3.73 (s, 3H, CH3), 3.23 (s, 3H, CH3).
13C NMR (101 MHz, DMSO-d6) δ 190.70 (Ar), 153.68 (Ar), 150.92 (Ar), 140.82 (Ar), 132.36 (Ar),
131.60 (Ar), 128.67 (Ar), 126.60 (Ar), 126.49 (Ar), 108.80 (C-sp), 104.47 (C-sp), 38.92 (CH3),
37.33 (CH3), 31.92 (CH3), 28.67 (CH3).
ESI-MS (CH3CN, pos. mode) for for C17H18AuN4O2: exp. 507.1096 (calc. 507.1095).
Elemental analysis for C17H19AuN4O3 (4a  ½ H2O): exp. C 39.47%, H 3.45%, N 10.72% (calc. C
39.62%, H 3.52%, N 10.87%)
13

Journal Pre-proof
(4-Ethynylpyridine)(1,3,7,9-tetramethylxanthin-8-ylidene)gold(I)
(4b):
4-Ethynylpyridine
(16 mg, 0.14 mmol) was added to NaOH (40 mg, 1 mmol) in methanol (10 mL). After stirring for
15 min at room temperature, AuTMXI (75 mg, 0.14 mmol) was added and the mixture stirred
under reflux overnight. After filtration, the final product was collected as an off-white solid (33 mg,
46%).
1H NMR (400 MHz, DMSO-d6) δ 8.41 (dd, J = 4.5, 1.6 Hz, 2H, Ar), 7.19 (dd, J = 4.5, 1.6 Hz, 2H,
Ar), 4.19 (s, 3H, CH3), 4.03 (s, 3H, CH3), 3.71 (s, 3H, CH3), 3.21 (s, 3H, CH3).
13C NMR (101 MHz, DMSO-d6) δ 190.00 (Ar), 153.67 (Ar), 150.89 (Ar), 149.86 (Ar), 140.81 (Ar),
139.34 (Ar), 133.97 (Ar), 126.06 (Ar), 108.89 (C-sp), 102.18 (C-sp), 38.95 (CH3), 37.35 (CH3),
31.91 (CH3), 28.65 (CH3).
Elemental analysis for C16H18AuN5O3 (4b  H2O): exp. C 36.17%, H 3.39%, N 13.07% (calc. C
36.58%, H 3.45%, N 13.33%)
(3-Ethynylpyridine)(1,3,7,9-tetramethylxanthin-8-ylidene)gold(I)
(4c):
3-Ethynylpyridine
(16 mg, 0.14 mmol) was added to NaOH (40 mg, 1 mmol) in methanol (10 mL). After stirring for
15 min at room temperature, AuTMXI (75 mg, 0.14 mmol) was added and the mixture stirred
under reflux overnight. After filtration, the final product was collected as an off-white solid (43 mg,
60%).
1H NMR (400 MHz, DMSO-d6) δ 8.46 – 8.41 (m, 1H, Ar), 8.36 (dd, J = 4.8, 1.5 Hz, 1H, Ar), 7.65
– 7.60 (m, 1H, Ar), 7.27 (dd, J = 7.9, 4.8 Hz, 1H, Ar), 4.20 (s, 3H, CH3), 4.04 (s, 3H, CH3), 3.71 (s,
3H, CH3), 3.20 (s, 3H, CH3).
13C NMR (101 MHz, DMSO-d6) δ 189.79 (Ar), 153.20 (Ar), 151.67 (Ar), 150.42 (Ar), 146.47 (Ar),
140.35 (Ar), 137.98 (Ar), 136.01 (Ar), 123.22 (Ar), 122.74 (Ar), 108.37 (C-sp), 100.55 (C-sp),
38.49 (CH3), 36.89 (CH3), 31.45 (CH3), 28.20 (CH3).
Elemental analysis for C16H18AuN5O3 (4c  H2O): exp. C 36.91%, H 3.21%, N 13.12% (calc. C
36.58%, H 3.45%, N 13.33%)
14

Journal Pre-proof
(2,8-Diethyl-10-(4-ethynylphenyl)-5,5-difluoro-1,3,7,9-tetramethyl-5H-4λ4,5λ4-dipyrrolo[1,2-
c:2',1'-f][1,3,2]diazaborinine)(1,3,7,9-tetramethylxanthin-8-ylidene)gold(I) (4d): 2,8-Diethyl-
10-(4-ethynylphenyl)-5,5-difluoro-1,3,7,9-tetramethyl-5H-4λ4,5λ4-dipyrrolo[1,2-c:2',1'-
f][1,3,2]diazaborinine (55 mg, 0.14 mmol) was added to NaOH (40 mg, 1 mmol) in MeOH
(10 mL). After stirring for 15 min at room temperature, AuTMXI (75 mg, 0.14 mmol) was added
and the mixture stirred under reflux overnight. After filtration, the final product was collected as a
dark red solid. (75.0 mg, 68%).
1H NMR (400 MHz, acetone-d6) δ 7.53 (d, J = 8.4 Hz, 2H, Ar), 7.26 (d, J = 8.2 Hz, 2H, Ar), 4.35
(s, 3H, CH3), 4.13 (s, 3H, CH3), 3.91 (s, 3H, CH3), 3.31 (s, 3H, CH3), 2.49 (s, 6H, CH3), 2.36 (q,
J = 7.6 Hz, 4H, CH2), 1.40 (s, 6H, CH3), 0.99 (t, J = 7.5 Hz, 6H, CH2CH3).
13C NMR (126 MHz, acetone-d6) δ 192.51 (Ar), 154.43 (Ar), 154.29 (Ar), 151.70 (Ar), 141.58 (Ar),
139.14 (Ar), 134.59 (Ar), 133.77 (Ar), 133.57 (Ar), 133.22 (Ar), 133.15 (Ar), 131.47 (Ar), 129.00
(Ar), 128.73 (Ar), 109.60 (C-sp), 103.89 (C-sp), 39.22 (CH3), 37.55 (CH3), 32.13 (CH3), 28.58
(CH3), 17.49 (CH3), 14.95 (CH3), 12.61 (CH2), 12.08 (CH3).
11B NMR (128 MHz, acetone-d6) δ 0.77.
ESI-MS (CH3CN, pos. mode) for C34H39AuBF2N6O2: exp. 809.2877 (calc. 809.2867).
Elemental analysis for C34H44AuBF2N6O5 (4d  4H2O): exp. C 46.81%, H 4.85%, N 9.58% (calc. C
46.38%, H 5.27%, N 9.54%)
X-ray diffraction analysis. Complex 3a, 4b and 4d were allowed to form crystals by slow
evaporation of acetonitrile at room temperature. The crystals were analysed at the UK National
Crystallography Service in Southampton. A suitable colourless irregular-shaped crystal of 3a
(0.260×0.110×0.060 mm3) and a colourless plate-shaped crystal of 4b (0.150×0.080×0.020 mm3)
were selected and mounted on a MITIGEN holder in perfluoroether oil on a Rigaku 007HF
15

Journal Pre-proof
diffractometer equipped with Varimax confocal mirrors and an AFC11 goniometer and HyPix
6000HE detector. The crystal was kept at a steady T = 100.01(10) K during data collection. In the
case of complex 4d, several orange plate-shaped crystals with approximate dimensions
0.040×0.025×0.015 mm3 were mounted on a glass fibre in perfluoroether oil. Data were collected at
the Diamond beamline I19 diffractometer equipped with a Helix cryostat low-temperature device
operating at T = 30(2) K. The data was collected using 360 degree phi scans at 5%, and crystal 1
also using a 45 degree phi scan at 2%. Since the first crystal was affected by radiation damage, data
sets for several crystals were collected, and had to be scaled accordingly. Data were measured using
monochromatic Si(111) radiation ( = 0.6889 Å). The maximum resolution that was achieved was
 = 23.905° (0.85 Å).The structures of 3a and 4 b were solved with the ShelXD [39] structure
solution program using the Dual Space solution method and by using Olex2 [40] as the graphical
interface. Each model was refined with version 2014/7 of ShelXL [41] using Least Squares
minimisation. The structure of 4d was solved and the space group P21/c (# 14) determined by the
ShelXT (Sheldrick, 2015) structure solution program [41] using Intrinsic Phasing and by using
Olex2 (Dolomanov et al., 2009) as the graphical interface [40]. The model was refined by Least
Squares using version 2014/7 of ShelXL (Sheldrick, 2015) [42]. All non-hydrogen atoms were
refined anisotropically. Hydrogen atom positions were calculated geometrically and refined using
the riding model. There was a large volume (1094 Å3) for solvents. No obvious solvents were
identified, and the structure was SQUEEZEd. CCDC codes are 1916760, 1913564 and 1940532.
1
1H NMR stability studies. H NMR were recorded in a mixture DMSO-d6 : D2O (80 : 20), with
TMS as an internal reference, on a Bruker Avance 400 NMR spectrometer (Bruker Daltonics,
Bremen, Germany). The gold complexes 3a-d and 4a-c (6 mmol) have been dissolved in the solvent
1
mixture chosen with or without an equimolar amount of NAC. H NMR spectra were recorded at
time 0 and after 24 h with each of the sample freshly prepared before the first run.
16

Journal Pre-proof
Stability studies by UV-visible spectroscopy. UV-visible absorption spectra to investigate the
stability of compound 4d in solution were recorded on a Cary 60 UV-Vis spectrometer (Agilent
Technologies, Santa Clara, USA). A stock solution of 4d (3 mM) was prepared in DMSO. An
aliquot was diluted to 40.6 μM either in 1× PBS (pH 7.4) or deionized water, and the UV-vis
spectra acquired at room temperature over 24 h at different intervals (every 15 min during the first
hour and every hour for the remaining 23 h).
Quantum yield determination. Emission spectra were recorded on a Cary Eclipse fluorescence
spectrophotometer (Agilent Technologies, Santa Clara, USA). Compound 4d was dissolved in
degassed DMSO to a concentration corresponding to UV-vis absorbance 0.8 A.U. (λ(max) = 523–535
nm; 25 °C). The solution was transferred to a fluorescence spectrophotometer and an emission
spectrum was recorded (excitation wavelength 595 nm). The quantum yield of complex 4d was
calculated by comparison to a reference standard (Rhodamine 6G in ethanol, φ = 94%).
FRET DNA melting assay. Fluorescence resonance energy transfer (FRET) experiments were run
on an Applied Biosystems® QuantumStudio 5 Real-Time PCR thermocycler (Thermo Fisher
Scientific, Waltham, USA) equipped with a FAM filter (λex = 492 nm; λem = 516 nm). The
thermocycler was set to perform a stepwise increase of 0.3 °C every 30 s, from 25 °C to 95 °C, and
measurements were acquired after each step.
All the oligonucleotides were purchased from Eurogentec (Belgium) in HPLC purity grade. The
FRET probes used were FAM (6-carboxyfluorescein) and TAMRA (6-carboxy-
tetramethylrhodamine). The lyophilized fluorolabelled hTelo (21-mer), d[GGG(TTAGGG)3], C-
KIT1,
d[GGGAGGGCGCTGGGAGGAGGG]
and
hTERT,
d[GGGGGCTGGGCCGGGGACCCGGGAGGGGTCGGGACGGGGCGGGG] oligonucleotides were
firstly diluted in deionized water to obtain 100 μM stock solutions. Stock solutions were diluted to a
concentration of 400 nM in potassium cacodylate buffer (54 mM, pH 7.4), and then annealed to
17

Journal Pre-proof
form G-quadruplex (G4) structures by heating to 95 °C for 5 min, followed by cooling to room
temperature overnight.
Experiments were carried out in a 96-well plate with a total volume of 30 μL. The final
concentration of the G4-oligonucleotide was set to 200 nM in potassium cacodylate buffer (54 mM,
pH 7.4). Stock solutions of the gold compounds 3a-d and 4a-c in DMSO (1 mM) were freshly
prepared prior to the experiments. The stock solutions were further diluted to a final concentration
of 2 μM (with a total percentage of DMSO of approx. 0.1%) in potassium cacodylate buffer (54
mM, pH 7.4) to achieve G4 : gold compound stoichiometry of 1 : 5.
To perform the competition assays with duplex DNA, a 100 μM stock solution in MilliQ water of
annealed
double-stranded
DNA
(dsDNA),
d[GGGTTACTACGAACTGG]/d[CCAGTTCGTAGTAACCC], was purchased from Eurogentec
(Belgium) in HPLC purity grade. This solution was further diluted to 40 μM in 54 mM potassium
cacodylate buffer (pH 7.4). Stock solutions of the gold compounds 3a-d and 4a-c were diluted to 4
μM (with a total percentage of DMSO of approx. 0.1%) in potassium cacodylate buffer (54 mM, pH
7.4). The final concentration of the G4-oligonucleotide was 200 nM, with a stoichiometry of
1 : 5 : 50 of G4 : gold compound : dsDNA. Experiments were carried out in a 96-well plate with a
total volume of 30 μL. To compare different sets of data, FAM emission was normalised (0 to 1).
Tm is defined as the temperature at which the normalised emission is 0.5 and ΔTm is defined as the
difference of Tm between treated samples and untreated controls. Independent experiments were run
in triplicates.
Circular dichroism. Circular Dichroism (CD) spectra were recorded on a ChiraScan
(AppliedPhotophysics, Surrey, UK) using the following parameters: range 210–400 nm, bandwidth
1 nm, step 1 nm, accumulation 3, temperature 25 °C. Stock solutions of unlabelled 21-mer hTelo,
d[GGG(TTAGGG)3] and C-KIT1, d[GGGAGGGCGCTGGGAGGAGGG] (each 100 μM), were
prepared by dissolving the lyophilised sample in deionized water. Samples were further diluted in a
18

Journal Pre-proof
Tris-HCl/KCl buffer (10/50 mM, pH 7.4) to obtain a final concentration of 4 μM. Stock solutions of
AuTMX2, 3a and 3c were prepared as described in the previous section and further diluted in buffer
to achieve a G4 : gold compound stoichiometry of 1 : 5.
Cell lines and culture maintenance. The human cancer cell lines corresponding to breast
carcinoma (MCF-7), ovarian adenocarcinoma (SKOV-3) and skin malignant melanoma (A375)
were obtained from the ATCC. The SKOV-3 and A375 cells were cultured in Dulbecco`s Modified
Eagle Medium (DMEM, 4,5 g/L glucose, Corning), supplemented with 10% fetal bovine serum
(One-Shot FBS, EU-approved South American Origin, Thermo Fisher Scientific) and 1%
penicillin/streptomycin (Gibco). MCF-7 cells were grown in Roswell Park Memorial Institute
medium (RPMI, L-glutamine, Corning), supplemented with 10% fetal bovine serum (One Shot
FBS, EU-approved South American Origin, Thermo Fisher Scientific) and 1%
penicillin/streptomycin (Gibco). All cell lines were cultured at 37 °C, in a humidified atmosphere of
5% CO2 and passage diluted upon reaching confluence.
Antiproliferative assays. To evaluate the inhibition of cell growth by the gold complexes, cells
were seeded in 96-well tissue culture-treated plates (Corning) at 8’000 cells/well in 200 µL full
medium. Working solutions of the gold compounds 3a–c, 4a–d were prepared in the required
concentration by diluting freshly prepared stock solutions (10-2 M in DMSO) in complete medium.
The stock solutions were protected from light in order to avoid potential light-induced degradation.
Cells were allowed to adhere for 24 h. Then, the medium was refreshed and cells were incubated for
72 h in 200 µL complete medium containing different concentrations of the gold compounds. The
antiproliferative effects of complexes were evaluated using the 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide (MTT) assay. Following 72 h exposure, the medium was removed and
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Fluorochem) in 10× PBS
(Corning) was added to the cells, at a final concentration of 0.3 mg/mL. After 3–4 h incubation at
19

Journal Pre-proof
37 °C and 5% CO2, the supernatant was discarded and the formazan crystals were dissolved with
DMSO. The optical density was quantified in quadruplicates for each experiment at 550 nm using a
multi-well plate reader (VICTOR X, Perking Elmer).
Complex 4d was evaluated using the CellTiter-Blue® assay due to the interference with the MTT
absorbance. After 72 h incubation with the compound, CellTiter-Blue® reagent (20 µL/well) was
added to the assay plate, shaken for 10 s and incubated for 4 h at 37 °C and 5% CO2. Fluorescence
intensity (531Ex/595Em nm) from each well was measured in quadruplicates for each experiment
using a multi-well plate reader (VICTOR X5, Perking Elmer).
The EC50 value for each compound in both assays was calculated as the concentration showing 50%
decrease in cell growth, when compared to controls, using a nonlinear fitting. In the case of
compound 4d fluorescence intensity was corrected for interference of the BODIPY-alkyne. Data is
presented as mean ± standard deviation of at least three independent experiments.
Fluorescence microscopy. Round glass coverslips (Ø 13mm, VWR) sterilized by UV-light were
inserted in 24-well tissue culture-treated plates (Corning). A375 cells were seeded at a
concentration of 120’000 cells/mL and incubated at 37 °C under humidified atmosphere with 5%
CO2 for 24 h. The medium was discarded and fresh medium containing 15 µM of complex 4d was
added. Following 2 h incubation at 37 °C, the glass coverslips were removed from the wells,
washed four times with 1× PBS (Corning) and fixed with 4% formaldehyde (Alfa-Aesar) for 20 min
at room temperature. The coverslips were then washed three times with 1× PBS and incubated for 1
min at room temperature with 4’,6-diamidino-2-phenylindole dihydrochloride (DAPI, 40 µL of 1
µg/mL, Sigma-Aldrich/MERCK). After washing the coverslips three times with 1× PBS they were
mounted on glass microscope slides (VWR) using Mowiol® 4-88 (Sigma-Aldrich). Fluorescence
images were obtained using a Zeiss Axio Vert.A1 epifluorescence microscope and analysed using
ImageJ software [43].
20

Journal Pre-proof
3. RESULTS AND DISCUSSION
3.1 Synthesis and characterisation
Initially, we synthesized a family of cationic gold(I) bis-NHC complexes related to the lead
compound AuTMX2. Specifically, the N1 position of theobromine was substituted with ethyl (a),
(2-cyclohexyl)ethyl (b), benzyl (c) and 4-(trifluoromethyl)benzyl (d) moieties (Scheme 1).
Scheme 1.
The synthesis of the functionalised ligands (1a–d) was confirmed by 1H NMR by the disappearance
of the broad peak of the N1-proton and by the appearance of the signals related to the newly
introduced groups (Fig. S1-4). Next, the N1-substituted ligands were methylated in the N9-position
using Merweein’s salt. This is a crucial step since only the positively charged tetraalkylxanthinium
precursor facilitates carbene formation and subsequent complexation to Au(I) [37]. The xanthinium
1
13
ligands (2a-d) were isolated and the methylation was confirmed by H, C NMR and elemental
analysis. As evident from the 1H NMR spectra (Fig. S9-12), the methylation in N9-position resulted
in the appearance of a singlet with nearly identical chemical shift compared to the N7-methyl
signal, together integrating for 6 protons (4.08 ppm), whereas the C8 proton experienced a
downfield shift to 8.48 ppm. Furthermore, a new methyl-specific carbon signal appeared at 37.5
ppm (Fig. S13-S16).
The final four gold(I) complexes (3a-d) were obtained via a transmetalation reaction
(Scheme 1) [37]. The NHCs were formed in situ in the presence of Ag2O under inert atmosphere
and were stabilized by Ag-complexation. Addition of Au(tht)Cl induced transmetallation and silver
chloride precipitated from solution. The gold(I) bis-NHC complexes were obtained in moderate
1
yields. Deprotonation of the C8 position was confirmed by H NMR. Upon complexation, the
methyl peaks at 4.08 ppm split more clearly into two distinct signals for 3a, 3c and 3d. Moreover,
13
in the C NMR spectra these three complexes also experienced a ca. 2 ppm downfield shift of the
21

Journal Pre-proof
1
N9-methyl signal (see Fig. S17,19, 20). The H NMR spectrum of 3b showed the presence of two
simultaneously existing coordination states, as indicated by the appearance of two sets of methyl
signals with constant ratios, but only one set of signals corresponding to the (2-cyclohexy)ethyl
moiety, probably corresponding to rotamers (Figures S18 and S22).
The second series of compounds included the mixed NHC-alkynyl gold(I) complexes (4a-
d), which were readily obtained from reacting the neutral mono-carbene complex AuTMXI with
the respective deprotonated alkyne in the presence of a base (Scheme 2) [20]. Interestingly,
complex 4d features a boron-dipyrromethene (BODIPY)-based alkynyl ligand with luminescence
properties which can allow the compound’s visualization in cells by fluorescence microscopy.
Scheme 2.
1
13
The successful synthesis of the complexes 4a–d was confirmed by H and C NMR (Fig. S25-
S33), elemental analysis and in some cases by ESI-MS (Fig. S36, S37). The appearance of a signal
13
at 108 ppm in the C NMR spectra is characteristic for the coordinated alkynyl. Single crystals of
3a, 4b and 4d were obtained by slow evaporation of acetonitrile. The structures are shown in Figure
2 while the relevant crystal data are reported in the supplementary material (see Tables S2-S26).
Figure 2.
The three compounds show a linear geometry with the angle C–Au–C being 178.54°, 179.10° and
176.79° for 3a, 4b and 4d, respectively. A monoclinic crystal system was resolved for 3a, where the
complex forms intramolecular hydrogen bonds with a water molecule (see Figure S54), whereas 4b
is characterised by a triclinic crystal system with three molecules are staggered upon each other. In
the latter, the distance between each molecule is 3.610 Å, which is compatible with the occurrence
of π-π interactions between the complexes. Moreover, aurophilic interactions [44] among the three
4b complexes were also found, as evidenced by the gold–gold distance of 3.286 Å. Noteworthy, the
22

Journal Pre-proof
triclinic crystal system resolved for 4d shows the existence of two distinct arrangements of the
molecule within the asymmetric unit (see Figure 2C, Figure S55 and tables S18-S26 for more
information). In each complex bond distances between the gold(I) centre and either the C8 of TBM
and/or the alkynyl carbon are 2.0 Å, similarly to those reported for other TMX-based gold(I)
organometallics [37]. Noteworthy, the crystal of 3a confirms that, differently from what previously
reported for the N9 substitution, the N1-modified complexes retains their planarity.
Next, the quantum yield of the gold compound 4d and the BODIPY-alkyne ligand precursor
were determined in aqueous solution (see Table 1 and experimental section for details).
Complexation of the fluorophore to the gold(I) NHC complex via the alkynyl bridge did not quench
the fluorescence of the resulting metal complex, whose quantum yield was found to be 74%.
Moreover, gold(I) complexation did not change the absorption and emission wavelengths of the
BODIPY fluorophore.
Table 1.
3.2 Stability studies
The stability of the gold(I) bis-NHCs and mixed NHC-alkynyl complexes was studied over 24 h in
1
a mixture DMSO-d6 : D2O (80 : 20) by H NMR (Figures S38-S44). The compounds were largely
stable, except for 4a, for which a new set of peaks appeared in the region from 2.5 to 4.0 ppm after
24 h, most likely due to the displacement of the alkynyl moiety by DMSO (see Figure S42).
In order to characterize the reactivity of the gold(I) complexes with model nucleophiles, the
compounds 3a–d and 4b–c were exposed to an equimolar amount of N-acetylcysteine (NAC) and
1
analysed by H NMR spectroscopy. The stability of NAC alone was also followed over time by
1H NMR. NAC underwent auto-oxidation to cystine to an extent of ca. 30% after 24 h as observed
by the formation of a new set of signals at 4.45, 3.09 and 1.87 ppm (Fig. S45). All samples showed
a similar cystine : cysteine ratio in the presence or absence of the gold(I) complexes.
23

Journal Pre-proof
The results obtained here show that the compounds are largely stable (Figures S46-51).
1
Compounds 3b, 3c and 4b displayed some changes in their H NMR spectra over 24 h (Fig S47,
S48 and S50). In detail, for 3b over 24 hours the signals related to the methyl protons are shifted to
3.99 and 3.66 ppm while the third one disappears probably hidden below the solvent signal (Fig.
S47). Moreover, a new small signal appears at 2.97 which can be attirbuted to the Au-bound
cysteine. The peaks at 5.05 and 3.71 ppm for 3c are split over time while two more signals for the
methyl groups appear at 3.99 and 3.96 respectively. Analysis of the integrals suggests the existence
of a 1:2 ratio between the original compound and a reaction product that may tentatively be
attributed to a cysteine adduct (Fig. S48). The latter is also confirmed by the appearance of new
peaks at 3.05 and 2.97 ppm whose signals are related to Au-bound cysteine. Similarly, as detailed in
Fig. S50, 4b experiences a splitting of the methyl signals at 4.10, 3.97 and 3.20 ppm, while a new
signal appears for the Au-bound cysteine (2.95 ppm) likely due to the protons close to the sulphur
atom. Two additional peaks are found in the aromatic region, respectively at 8.54 and 7.45 ppm,
which clearly show the displacement of the aromatic-alkynyl moiety by NAC, in a 4b : cysteine-
adduct = 70 : 30. Overall, compounds 3a, 3d and 4c are inert with respect to ligand exchange
reactions with NAC over 24 h, while 3b, 3c and 4b are prone to react with NAC, leading to the
probable formation of sulfur-adducts.
Due to the scarce solubility in DMSO, the stability of the BODIPY-based complex 4d could
not be evaluated by NMR, and thus, UV-vis stability studies were conducted. The experiments were
carried out over 24 h in 1× Phosphate Buffered Saline (PBS, pH = 7.4) and deionized water, with
data collection each 15 min for the first hour and then each hour for the remaining 23 hours. The
UV-vis spectra of 4d are characterised by two intense bands at 533 and 293 nm and a shoulder in
between at 380 nm (black solid line in Fig. S52). Over 24 h these bands experienced hypochromic
effects and markedly decreased in PBS buffer (Fig. S52). A similar hypochromism is evident in
deionised water, where 4d is more stable than in PBS. In both solvents, the intensity of the
absorption is reduced by ca. 30-40% after 6 h. These hypochromic effects may be attributable to the
24

Journal Pre-proof
hydrolysis of the B-F moiety in the BODIPY scaffold leading to precipitation. In fact, a similar and
even more prominent trend can be observed for the free BODIPY ligand (Fig. S53).
3.4 G4-stabilisation studies
The DNA G-quadruplex (G4) stabilising properties of the newly synthesised Au(I) NHC
organometallics were assessed by means of fluorescence resonance energy transfer (FRET) DNA
melting assays [45]. Earlier studies with cationic gold(I) bis-NHC complexes achieved after N9
substitution on the lead compound AuTMX2 focused on telomeric G4s [37]. Here, we extended
these studies to two oncogenic promoter G4 structures, namely, C-KIT1 and hTERT. The former is
an oncogenic receptor tyrosine kinase involved in proliferation [46], and the latter is the catalytic
subunit of the human telomerase, which is involved in cancer immortality [47].
The gold compounds 3a-d and 4a-c were reacted with each G4 in a 5:1 stoichiometry as detailed in
the experimental section. Table S1 shows the ΔTm for each compound in comparison to the
previously reported complexes, the cationic bis-carbene derivative AuTMX2 and the mono-carbene
AuTMXI.
Interestingly, AuTMX2 is most potently stabilising hTelo (ΔTm = 13.1 ± 1.8 °C) in
comparison to the gene promoter G4s of C-KIT1 and hTERT (Table S1). In contrast, AuTMXI can
similarly stabilize C-KIT1 (ΔTm = 4.0 ± 0.1 °C) and hTelo (ΔTm = 3.2 ± 0.6 °C), while it is
practically inactive towards hTERT (ΔTm = 1.54 ± 0.15 °C). Concerning the newly synthesized
compounds, extending the N1-methyl in 3a–d also led to a marked reduction of their G4 stabilizing
potency, similarly to N9 modifications [37]. The best stabilizer of this series is complex 3a,
showing a similar trend compared to AuTMX2, while 3c was able to slightly stabilize the hTelo
DNA sequence (Figures 3A and 4B). It may be noted that 3a differs from AuTMX2 only by ethyl to
methyl substituents at the N1-position.
The mixed Au(I) NHC-alkynyl complexes 4a–c did not show any stabilising effects on the
G4 structures. Despite the fact that the flat aromatic surface of these mixed NHC-alkynyl
25

Journal Pre-proof
complexes could have allowed favourable π-π interactions with the G4s, it can be hypothesized that
the steric demand of 4a-c does not allow the compounds’ alignment with the G4-quartet. The
intrinsic fluorescence of 4d did not allow its screening by FRET.
In addition to their stabilizing potency, the selectivity of the compounds’ for G4s and duplex
DNA binding was investigated for AuTMX2, 3a and 3c with hTelo and C-KIT1 in the presence of a
50-fold excess of dsDNA. The three organometallics showed a clear selective interaction towards
both hTelo and C-KIT1 over dsDNA, as their stabilising properties were unperturbed when
challenged with the excess of dsDNA (Fig. 3C, D). In the previously reported study, N9-substituted
Au(I) bis-NHC complexes showed increased affinity towards dsDNA [37]. Consequently, although
the N1-modified Au(I) bis-NHC complexes are less potent in stabilizing the G4 structures
compared to N9-modified analogues, as assessed by ΔTm, the compounds retain their high
selectivity for G4 binding.
Figure 3.
Furthermore, circular dichroism (CD) spectra of hTelo and C-KIT1 in the presence of 5 equiv. of
the G4 stabilising compounds AuTMX2, 3a and 3c (only hTelo) were recorded to assess if any
changes of the G4-folding would occur upon interaction of the polynucleotide with the gold(I) NHC
complexes. In the UV-vis range, the G4-DNA features two π-π* transitions (at ca. 280 and 248 nm).
However, within a G-quadruplex, the tetrads are rotated with respect to each other giving rise to a
chiral exciton coupling between transition dipole moments located in near-neighbour guanines: the
final G4-DNA is therefore, CD active [48]. In detail, the CD spectrum of hTelo in our experimental
conditions is characterised by an intense positive band at 290 nm with a shoulder at 270 nm and by
a negative absorption at 240 nm (black solid line in Fig. 4A). This spectrum is typical of a hybrid
(3+1) G4-conformation [49]. Conversely, in the same conditions, C-KIT1 folds in a parallel
conformation with a positive band at about 263 nm and a negative centred at 240 nm (black solid
26

Journal Pre-proof
line in Fig. 4B).
Figure 4.
Interestingly, when a 4 μM solution of hTelo in Tris-HCl/KCl (10 mM/50 mM, pH 7.4) is
challenged with 5 equiv. of AuTMX2, 3a or 3c the positive band at 290 nm increases, as a result of
the interaction of the three compounds with the G4-DNA as previously reported for other G4-
binders (Fig. 4A) [50–52]. The addition of AuTMX2 also causes a reduction of the shoulder at 270
nm which might suggest a slight structural change of the G4-DNA as previously observed with a
nickel(II) complex [52]. The other two complexes, 3a and 3c do not show a similar decrease of the
band located at 270 nm, while 3c even induces an opposite modification of the signal. On the
contrary, with respect to C-KIT1, the CD analysis reported in Fig. 4B clearly shows that the
interaction of the two complexes with the G4-DNA, proven by the FRET DNA melting assays,
occurs without affecting the secondary structure of the polynucleotide, as reported in literature for
some nickel(II) and zinc(II) Schiff-base complexes [53].
3.5 Antiproliferative assays
The cytotoxicity of the compounds 3a–c and 4a–d was evaluated in human melanoma (A375),
ovarian (SKOV-3) and breast (MCF-7) cancer cell lines in comparison to AuTMX2 (Table 2). It
may be noted that AuTMX2 is completely inactive in these cancer cells, and AuTMXI also showed
scarce cytotoxicity in all the tested cell lines.
Table 2.
Similarly to AuTMX2, the cationic Au(I) bis-NHC complex 3a, featuring an ethyl substituent at the
N1-position, is inactive in all tested cell lines. Interestingly, when a bulkier modification is
27

Journal Pre-proof
introduced, as in 3b and 3c, a considerable increase in the cytotoxicity is observed towards all
cancer cells, irrespective of being saturated (3b) or aromatic (3c). The tested Au(I) mixed NHC-
alkynyl complexes, although scarcely active in all cell lines, showed a distinct activity profile with
4a–c moderately cytotoxic in the MCF-7 breast cancer cell line. Especially the 3-pyridine derivative
4c was slightly more active (IC50 ca. 16 M) than the other derivatives of this family. Finally, 4d
has no cytotoxic effect against any of the three cancer cell lines.
3.6. Fluorescence microscopy
Compound 4d bears a fluorescent BODIPY-moiety, but showed no cytotoxicity against the
investigated cancer cell lines. In order to investigate whether the reduced cytotoxicity might
correlate with a reduced cellular accumulation, we studied its uptake by fluorescence microscopy.
Compound 4d (15 µM) was incubated with A375 cells for 2 h at 37 °C. The compound was found
to be efficiently taken up by the cancer cells, and to accumulate in the cytoplasm (Figure 5), while
cell nuclei did not show any signal corresponding to the BODIPY fluorophore. Thus, the low
cytotoxicity of 4d might rather be related to a reduced stability as indicated by UV-vis experiments
than by a reduced cellular accumulation. Since the absorption and emission wavelengths of 4d and
the BODIPY-alkyne are very similar, it is challenging to judge whether the fluorescence is caused
by intact 4d or free BODIPY ligand.
Figure 5.
4. CONCLUSIONS
The present study comparatively evaluated two novel families of gold(I) bis-N-heterocyclic carbene
and mixed NHC-alkynyl organometallics with respect to their stability in aqueous environment, G-
quadruplex (G4) stabilising effect and cytotoxicity. To the best of our knowledge this is one of the
few reports in the literature on the interactions of metal NHCs featuring xhantine type of ligands
28

Journal Pre-proof
with G4 DNA structures[31]. The representatives of both families form linear and highly planar
gold(I) organometallic complexes, which were reasonably stable in aqueous solution. Upon
interaction with NAC, complexes 3b, 3c and 4b showed some ligand exchange reaction, involving
substitution of the alkynyl ligand in the latter case. The cationic gold(I) N1-modified bis-NHCs
exhibited mild G4 stabilisation properties, while the mixed NHC-alkynyl family was devoid of any
G4 stabilising effects. Although less potent than the lead compound AuTMX2, the cationic gold(I)
bis-NHC derivatives retained their selectivity for the G4 structures even in the presence of excess
double-stranded DNA. While the complexes 3b-c are moderately cytotoxic in all cancer cell lines,
the gold(I) mixed NHC-alkynyl complexes, though being less active, were selectively cytotoxic
against the MCF-7 breast cancer cell line. Finally, a non-cytotoxic fluorescent gold(I) derivative
(4d) was shown to efficiently distribute into the cytoplasm of intact cells after 2 h incubation with
melanoma cells, but did not enter the nuclei. Overall, these investigations provide further insight
into the distinct biological activity of gold(I) organometallics and underline their promise for
tuneable biological applications by appropriate chemical modifications.
Abbreviations
AuTMX2
AuTMXI
BODIPY
C-KIT1
CD
[Au(9-methylcaffein-8-ylidene)2]+
[Au(9-methylcaffein-8-ylidene)I]
boron-dipyrromethene
C-KIT1 oncogenic receptor tyrosine kinase
circular dichroism
DCM
dichloromethane
DMF
dimethylformamide
DMSO
dsDNA
ESI-MS
dimethylsulfoxide
double-stranded DNA
electrospray ionization-mass spectrometry
29