Coordination chemistry of N-heterocyclic carbene gold complexes and their catalytic and anticancer activities

Over the past two decades, N-heterocyclic carbenes (NHCs) have gained significant attention as ligands, offering a versatile alternative to traditional phosphorus-based (e.g., phosphines, phosphites) and nitrogen-based (e.g., amines, imines) ligands. NHCs have been extensively studied in coordination chemistry, particularly for their strong σ-donor properties, making them exceptional ligands for coinage metals, especially gold(I). NHC-gold(I) complexes, with gold in a d10 electronic configuration, have found widespread applications in catalysis and as anticancer agents. In contrast, square planar NHC-gold(III) complexes (d8 electronic configuration) remain relatively underexplored. Although the practical halogen oxidative addition to gold(I) represents the predominant approach to obtaining gold(III) compounds, several observations remain unexplained, such as the occasional absence of expected products, the occurrence of reversible reduction, or the manifestation of geometrical isomerism in mixed halide Au(III) complexes. Furthermore, recent theoretical and experimental studies have emphasized the distinction between formal and physical oxidation states, better described as electronic distribution, undermining the concept of oxidation state +3 for square planar d8 complexes. Indeed, oxidation state formalisms often fall short in describing the electronic partitioning in metal-ligand bonds. This new reinterpretation shakes up several aspects of gold(III) chemistry that have been considered widely established. For example, gold-catalyzed reactions traditionally attributed to oxidation state variations, such as the cyclization of propargyl amides, may instead be influenced by other factors, like the presence of adventitious coordinating species. Further studies are therefore necessary, considering the new interpretation between formal and physical oxidation states. Thus, this PhD thesis investigates the synthesis, characterization, catalytic activity, and anticancer potential of a series of NHC-gold(I/III)X (X = Cl, Br or I), with particular emphasis on understanding the influence of ligand structures, oxidation states and halides as counterions on the cytotoxic and catalytic activities. This thesis can be divided into four main topics: 1) Synthesis and characterization of NHC-gold(I/III) complexes. 2) Gold-catalyzed cycloisomerization of propargyl amides. 3) Mechanistic insights into oxidative addition reaction of halogen to NHC-gold(I) complexes. 4) Anti-cancer applications of NHC-gold(I/III) complexes. 1) Synthesis and characterization of NHC-gold complexes In this first part of the thesis, two linear NHC-gold(I) chloride complexes displaying or not a plane of symmetry, i.e., symmetric 1,3-dimethyl-imidazolyl-2yl gold chloride (1-Cl) and asymmetric 1-benzyl- 3-methyl-imidazolyl-2-yl gold chloride (2-Cl), have been initially synthesized through the "weak base route."[1,2] The systems were kept as simple as possible to relegate the studies on the gold metal centre by feasible computational calculations. The corresponding NHC-Au(III)Cl3 were then obtained through oxidation with iodobenzene dichloride; both 1,3-dimethyl-imidazolyl-2yl gold trichloride (1- Cl3) and 1-benzyl-3-methyl-imidazolyl-2-yl gold trichloride (2-Cl3) were obtained with high yield (up to 90%).[3,4] The homolog series was then completed by changing halide (bromine or iodide) directly bonded to the gold centre and by oxidizing the corresponding NHC-gold(I)X (X =Br or I). Hence, the metathesis reaction of 1-Cl with sodium bromide led to the formation of a bis-carbene [(1,3-dimethyl- NHC)2Au][AuBr2] complex (1bis-AuBr2) with a 74% yield.[ 5 ] The substitution of chloride with bromide facilitated ligand scrambling, yielding the bis-carbene species with AuBr2 – as counterion. Conversely, the reaction of 2-Cl with an excess sodium bromide produced a mono-carbene 2-Br with a 91% yield.[6] Oxidation of 1bis-AuBr2 and 2-Br with liquid bromine led to the formation of a mono- carbene NHC-Au(III) tribromide, 1-Br3 and 2-Br3, with a 93% and 84% yield, respectively.[7] The NHC-gold(I) iodide analogs were achieved through a metathesis reaction of 1-Cl or 2-Cl with a large excess of sodium iodide, leading in both cases to a bis-carbene species with AuI2 - as counterion, namely 1bis-I (43% yield) and 2bis-I (68%).[8] Finally, the oxidation of 1bis-I with solid iodine resulted in a mixture of mono- and bis-carbene species; indeed, a mixture of 1,3-dimethyl-NHC-AuI3 (1mono- I3) and [(1,3-dimethyl-NHC)2AuI2][AuI2] (1bis-AuI2) was produced when 1bis-I was oxidized with solid iodine. The oxidation of 2bis-I produced a mixture of 1-benzyl-3-methyl-NHC-AuI3 (2mono-I3) and [(1-benzyl-3-methyl-NHC)2AuI2][I3/AuI4] (2bis-I3/AuI4) species. All the complexes were isolated as crystals and completely characterized by infrared, 1H, 13C NMR spectroscopies and X-ray crystal diffraction. Particularly, X-ray crystallography revealed notable features in the crystal structures: for example, the Au–Ccarbene bond is longer in bis-carbene species than in mono-carbene species. Additionally, despite the increase in the oxidation state of the gold center, the Au–Xtrans bond length is shorter in gold(I) complexes than in gold(III) complexes. Finally, the square planar geometry of the Au-X3 complexes is strikingly regular, leading to a slight compression of the Ccarbene-Au-Xcis bond angles from 90°, along with a corresponding expansion of the Xcis-Au-Xtrans bond angles. This argument is included in Chapter II., Sections 2.2 and 2.3. 2) Gold-catalyzed cycloisomerization of propargyl amides In the second topic, the cycloisomerization of propargylic amides (oxazoles synthesis) was studied as a benchmark reaction to assess the catalytic activities of the homologous series of NHC-Au(I) and NHC-Au(III) complexes. This oxazole synthesis was found to be gold oxidation dependent, allowing to selectively produce an aromatic oxazole with gold(III) catalysts or a methylenedihydrooxazole via gold(I) catalysts. Hence, the selectivity of this gold-catalyzed reaction was studied in terms of the gold oxidation state and the counterions. The previous NHC-gold(I) and NHC-gold(III) complexes were tested in the cyclization reaction at room temperature, with and without the addition of silver salt, AgPF6. Notably, only the 1-benzyl-3-methyl-gold(III) trichloride (2-Cl3) selectively yielded the aromatic oxazole without the use of silver salt. Monitoring the reaction catalyzed by 2-Cl3 via 1H NMR highlighted the initial formation of the methylenedihydrooxazole in the first two days that finally converted to the aromatic oxazole over a week. With the addition of silver salt, the oxidation state of the gold catalyst (gold(I) or gold(III)) did not affect the chemoselectivity of the cyclization. Indeed, both methylenedihydrooxazole and the aromatic oxazole were obtained regardless of the gold oxidation state. This observation removes any potential involvement of the oxidation state, while the distinct halides directly bonded to the gold centre significantly impacted the selectivity. Finally, the oxazole synthesis of an internal alkyne, specifically 4-methoxy-N-(3-phenylprop-2-ynyl) benzamide, was also explored both in the presence and absence of silver salts under the previously established conditions. However, no conversion was observed under any tested conditions. This result suggested that the active gold metal fragment likely substitutes the terminal alkyne proton to initiate the catalytic cycle rather than merely coordinating to the triple bond. 3) Mechanistic insights into oxidative addition reaction of halogen to NHC-gold(I) complexes. For late transition metals, the d orbital energies can sometimes be lower than those of the ligand orbitals, contrary to what is stated in the Ligand Field Theory. This phenomenon is notably observed in certain square planar “d8” complexes. In such cases, the resulting σ-bonding orbital is primarily localized on the metal center, while the antibonding orbital exhibits a stronger ligand contribution, with electron density shifted toward the metal center. This reinterpretation of metal-ligand bonding in square planar complexes is known as the Inverted Ligand Field (ILF) model. This model challenges traditional views of oxidation states, suggesting that the reactivity of gold(III) complexes is strongly governed by ligand-centered electronic effects. From this perspective, the complete lack of influence of the gold oxidation state on the selectivity of the previous cycloisomerization prompted further investigation into the oxidative addition mechanism to gold(I) complexes and the actual oxidation state of gold, especially in the context of the ILF. Hence, in the third part of this work, the iodine and bromine addition to the previous linear NHC-gold(I) chloride (1-Cl and 2-Cl) was investigated at room temperature and at 75°, both experimentally and computationally. For the iodine addition, the molecular structures highlighted the formation of trans-NHCAuI2Cl isomers for 1-Cl and 2-Cl, both at room temperature and at 75°C. Conversely, the bromine addition was strongly case-sensitive; at room temperature, the bromine addition to 1-Cl led to a mixture of cis and trans-NHC-AuBr2Cl isomers in a 17:83 ratio while, in the same condition, a mixture composed of trans-isomer and a completely brominated product was obtained for 2-Cl. Finally, performing the bromine addition at 75°C, a NHC-gold(III) tribromide product was achieved for both 1-Cl and 2-Cl. All the mixed halide gold(III) complexes were isolated as crystals and characterized by elemental analysis, IR, UV-visible, 1H and 13C NMR spectroscopy, and X-ray crystal diffraction. Computational studies were then performed by the CNR of Florence. DFT analysis reveals a stepwise addition of the halogen atom to the linear gold(I) complex, occurring without a net change in the electronic population of the metal. This observation aligns with the Inverted Ligand Field (ILF) model. Furthermore, the DFT study indicates that this halogen addition mechanism proceeds through the formation of a tri-coordinated intermediate. The stability of these intermediates is governed by the spatial arrangement of the three ligands, carbene, the initial halide, and the newly added halide. The intermediate stability is, in turn, influenced by the electronegativity of the coordinated halide leading to preferential geometrical isomers as final outputs. 4) Anti-cancer applications of NHC-gold(I/III) complexes Lung cancer remains a leading cause of cancer-related mortality worldwide, necessitating the development of novel therapies to address severe side effects and the emergence of drug resistance. Among the different types of lung cancer, non-small lung cancer cells (NSCLC) are the most diffused. In this context, NHC-gold complexes have merged as significant promises in lung cancer therapies. The cytotoxic activity of gold complexes is believed to be their capacity to target proteins. For example, auranofin, the first gold-based drug FDA approved, is known to specifically manipulate the redox environment by inhibiting redox enzymes[9] and is a potent inhibitor of the thioredoxin reductase system (TrxR).[10] Indeed, the TrxR system has been confirmed as a critical target in the proposed mechanisms of action of gold complexes. Interestingly, the overexpression of thioredoxin reductase systems as a defense response against oxidative stress is associated with several types of cancers, including lung cancer. Therefore, NHC-gold compounds are highly considered for anticancer therapies. Hence, in the last part of the thesis, the cytotoxic activity of the previous NHC-gold complexes was explored thanks to our collaboration with the University of Cincinnati. Two gold(I)- triphenylphosphane complexes were also added to the cytotoxic study: complex 3, already tested against breast cancer, and a new compound obtained by the direct reaction between 1-Cl and triphenylphosphine (complex 4). A panel of NSCLC cell lines was utilized as cancer models in this study. Gold compounds featuring triphenylphosphine ligands exhibited greater cytotoxicity compared to homoleptic [(NHC)2-Au(I)]X complexes or heteroleptic NHC-Au(I)X and NHC-Au(III)X3 complexes. Additionally, the gold(I) halide complexes (1bis-AuBr2 and 2-Br) were found to be more active than the corresponding NHC-gold(III) compound. At the same time, the presence of the benzyl substituent enhanced the overall cytotoxic activity with respect to the methyl one. Among the tested compounds, mixed-ligand gold(I) complexes with linear NHC-AuPPh3 (compound 3) and trigonal NHC-Au(Cl)PPh3 (compound 4) arrangements around the central metal demonstrated the highest cytotoxicity against lung cancer cells. Complexes 3 and 4 were also evaluated in normal non-tumor human lung fibroblasts (IMR90 cell line) to investigate their selectivity against healthy cells. Interestingly, both compounds exhibited higher selectivity against the NSCLC cell lines over the lung fibroblasts, suggesting fewer side effects in future anticancer therapies. Finally, analysis of thioredoxin reductase (TrxR) activity in treated cells revealed that these compounds effectively inhibit TrxR, the gold compound's most widely recognized molecular target. Notably, compound 4 achieved over 80% reduction in TrxR activity in lung cancer cells.