Cationic gold ( I ) heteroleptic complexes bearing a pyrazole-derived N-heterocyclic carbene : syntheses , characterizations , and cytotoxic activities †

A series of cationic gold(I) heteroleptic complexes bearing the pyrazole-derived N-heterocyclic carbene (NHC) FPyr (1,2,3,4,6,7,8,9-octahydropyridazino[1,2-a]indazolin-11-ylidene), and either a 1,3-disubstituted benzimidazole-derived NHC of the type RR’-bimy (3: R = R’ = CHPh2; 4: R = CHPh2, R’ = Pr; 5: R = R’ = CH2Ph; 6: R = R’ = Bu; 7: R = R’ = n-Pr; 8: R = R’ = Et; 9: R = R’ = 2-propenyl) or a non-NHC co-ligand L (10: L = PPh3; 11: L = P(OPh)3; 12: L = DMAP) (DMAP = 4-dimethylaminopyridine) have been synthesized from [AuCl(FPyr)] (1). Complexes 3–12 have been characterized using multinuclei NMR spectroscopies, ESI mass spectrometry, and elemental analysis. X-ray diffraction analyses have been performed on complexes 5, 6, and 9–11. To the best of our knowledge, 11 represents the first gold–NHC complex to bear the P(OPh)3 ligand. The cytotoxic activities of complexes 3–12 have been studied in vitro with the NCI-H1666 non-small cell lung cancer cell line.


Introduction
Interest in the cytotoxic activities of gold complexes has grown tremendously over the years. 1,2In particular, gold complexes bearing N-heterocyclic carbenes (NHCs) have been gaining much attention, 3 given the ease with which NHC precursors can be tuned, both electronically and sterically, by varying the N-substituents. 4This in turn allows for synthetic control of the hydro-and lipophilic properties of the resultant complexes; properties that have been shown to be important factors in the development of gold-based anti-cancer agents. 2,3However, only gold mono-and homo-bis(carbene) complexes bearing imidazolin-2-ylidenes have been extensively studied, 2,3 whereas complexes bearing other types of NHCs, as well as heteroleptic complexes, have been scarcely considered. 5herefore, as our maiden contribution to this field of research, we had recently published the cytotoxic activities of a range of gold(I) and gold(III) mono-, homo-bis-and hetero-bis-(carbene) complexes bearing benzimidazole-and/or pyrazolederived NHC ligands on the NCI-H1666 non-small cell lung cancer cell line. 6The preliminary study revealed that the cationic bis(carbene) complexes are far more active as cytotoxic agents when compared with the neutral monocarbene complexes.The gold(I) hetero-bis(carbene) complex 2 (Fig. 1), in particular, shows superior performance with an IC 50 value of 0.241 μM.Based on these findings, it was proposed that a heteroleptic system in which a strongly donating ligand is situated trans to a relatively weaker donating ligand was necessary for the labilization of the latter.This would create a vacant coordination site that may be essential for the cytotoxic activity of the complex.For example, donor atoms on target proteins, such as thioredoxin reductase, 2,5,7 could bind to the complex fragment when a free coordination site is made available.
Spurred by the findings of our initial study, we have extended the investigation to consider a range of other cationic gold(I) heteroleptic complexes bearing the strongly donating pyrazole-derived FPyr (1,2,3,4,6,7,8,9-octahydropyridazino[1,2-a]indazolin-11-ylidene) ligand, and either a 1,3-disubstituted benzimidazole-derived NHC, or a non-NHC co-ligand.Benzimidazolin-2-ylidenes bearing a variety of N-substituents have been considered, in an effort to study the effect on the cytotoxicity of the resultant complexes.For the purpose of comparison with our previous work, the cytotoxic activities of the complexes were studied in vitro with the NCI-H1666 non-small cell lung cancer cell line.We herein report on the syntheses, characterizations, and cytotoxic activities of the aforementioned complexes.

Syntheses and characterizations
The gold(I) chlorido-monocarbene complex [AuCl(FPyr)] (1) was synthesized according to the previously reported procedure. 6Reaction of 1 with the appropriate 1,3-disubstituted benzimidazolium salts in the presence of K 2 CO 3 afforded the gold(I) hetero-bis(carbene) complexes 3-9 in very good yields ranging from 72 to 82% (Scheme 1).The heteroleptic complexes 10-12, which bear non-NHC co-ligands, were synthesized via straightforward ligand exchange reactions.Complex 1 was reacted with the appropriate pro-ligand in the presence of a large excess of KPF 6 in acetone, and the products were obtained in very good yields of 77 to 94% (Scheme 1).All ten complexes were obtained as white powders, and are soluble in most polar organic solvents.It should be noted that 11 represents the first gold-NHC complex to bear a phosphite co-ligand.
Complexes 3-12 were characterized using MS (ESI) and multinuclei NMR spectroscopies.The MS (ESI) spectra of the complexes were particularly useful in confirming the formation of the compounds, given that the base peak observed in all cases corresponds to the [M − PF 6 ] + fragment. 1H NMR spectroscopy corroborated the successful formation of the complexes, whereby signals from both ligands are observed in the respective spectra.For complexes 3-9, the 13 C NMR spectra correspondingly featured two downfield signals as a result of the presence of two carbene carbon atoms in the molecules.These two signals were assigned to their respective carbene carbon atoms by using the previously reported complex 2 as a reference. 6The chemical shift(s) of the carbene carbon atom(s) in complexes 3-12 are presented in Table 1.
Considering the chemical shift of the FPyr carbene carbon, it is overt that as compared to the gold(I) chlorido monocarbene complex 1 (cf.167.8 ppm), 6 the FPyr carbene carbon in complexes 3-11, resonate at a much more downfield region, ranging from 178.8-182.4ppm.This is due to the replacement of the chlorido co-ligand in 1 with a stronger donating ligand situated trans to the FPyr ligand in the mentioned complexes, leading to the observed downfield shift of the signal. 6,8On the other hand, complex 12, with its DMAP co-ligand, shows a more upfield FPyr carbene carbon resonance as compared to 1, which is the result of DMAP being a weaker donating ligand as compared to the chlorido ligand.
The changes in the chemical shift of the FPyr carbene carbon also reveal other interesting, albeit unexpected, findings.For example, when focusing on the relevant data for hetero-bis(carbene) complexes 3-9, it is noteworthy that no real correlation can be drawn between the chemical shift of the FPyr carbene carbon and the N-substituents of the RR′bimy ligands, where the latter would inadvertently influence the overall donating ability of said RR′-bimy ligands.This is Scheme 1 Syntheses of cationic gold(I) hetero-bis(carbene) complexes of the type [Au(FPyr)(RR'-bimy)]PF 6 (3-9), and heteroleptic complexes of the type [Au(FPyr)-(L)]PF 6 (10-12).unlike the 13 C NMR spectroscopic methodology for determining ligand donor strengths that has been reported by our group, wherein even small changes to the N-substituents of various NHCs are shown to affect the donating ability of the ligands, and can be detected by the i Pr 2 -bimy probe in complexes of the type [PdBr 2 ( i Pr 2 -bimy)(NHC)].4a Therefore, the data presented in this work seems to suggest that the Au(I) system is less sensitive than the Pd(II) system when it comes to detecting small changes in donor strength.Furthermore, while the Pd(II) system is able to confirm the stronger donating ability of NHCs as compared to phosphines, the Au(I) system studied in this work suggests the reverse, with the FPyr carbene carbon in [Au(FPyr)(PPh 3 )]PF 6 (10) resonating at 182.4 ppm, which is significantly more downfield as compared to the hetero-bis(carbene) complexes 3-9.This may be due to some unique interactions that the phosphorus donor in PPh 3 has with the electron-rich gold(I) metal center, and warrants further investigation.
Single crystals suitable for X-ray diffraction analysis were obtained for complexes 5, 6, and 9-11 via vapour diffusion of diethyl ether into solutions of the compounds in either CHCl 3 or CH 2 Cl 2 .The determined molecular structures were as expected, with all complexes adopting a linear geometry about the gold(I) metal center (Fig. 2).Complex 5 crystallised as the chloroform solvate, while all other complexes crystallised unsolvated.It is interesting to note that unlike 5, where both the benzyl N-substituents are arranged in an anti arrangement across the benzimidazole plane, the N-substituents in 6 and 9 are arranged in a syn arrangement, which is a result of achieving optimum packing in the unit cell.Two of the carbon atoms (C29 and C30) in the alicyclic ring of the FPyr ligand in 5 were also disordered into two positions, with an occupancy ratio of 50 : 50.The molecular structure of complex 11 is particularly interesting, in that two of the phenyl rings of the P(OPh) 3 ligand are pointed towards the gold metal center.The remaining one is pointed away as a result of the close proximity of a PF 6 − anion.The potential steric shielding of the metal center accorded by the phenyl rings is unique to this ligand, and could prove beneficial for the overall stability of the complex.Furthermore, it is worth noting that none of the complexes showed evidence for intermolecular aurophilic interactions in the solid state. 9

Cytotoxicity study
The cytotoxic activities of complexes 3-12 were investigated through a cell proliferation assay conducted on the NCI-H1666 non-small cell lung cancer cell line, and IC 50 values were obtained from best-fit dose response curves (see ESI †).The performance of each complex in comparison to the previously reported complex 2, as well as cisplatin, is tabulated in Table 2.
Considering the hetero-bis(carbene) complexes 3-9, it is noted that they do not perform as well as the previously reported complex 2, although 3 does come close with an IC 50 value of 0.346 μM.Also, there seems to be no consistent relationship between the nature and length of the

Dalton Transactions Paper
N-substituents on the benzimidazole-derived NHC and the activity of the complexes.However, it is worth noting that complexes 2 and 3, which both bear symmetrically substituted benzimidazole-derived NHCs, with both N-substituents being 2°c arbon atoms, show superior performance as compared to complexes that bear benzimidazole-derived NHCs with either asymmetric N-substituents (i.e.complex 4), or 1°carbon atom N-substituents (i.e.complexes 5-9).This does suggest some form of structure-activity relationship, although further studies would be necessary to better understand the relative performance.Among the hetero-bis(carbene) complexes, 9 is an exception, and this may be attributed to the presence of reactive allyl moieties that make the complex more susceptible to addition and redox reactions, possibly degrading and/or deactivating the complex.
While the presence of a labile ligand seems necessary for the superior activity of these gold(I) complexes, it is apparently also important that the ligand trans to the FPyr ligand is not too weakly bound to the metal center.Evidence for this is found when considering complexes 10-12.As we progress from the stronger donating PPh 3 to the much weaker donating DMAP, there is an increase in the IC 50 values.Since the mode of action of other cationic gold-NHC complexes has been reported to involve permeation into the mitochondria, 1,2,7 it is possible that in the case of complex 12, for example, the DMAP ligand detaches from the complex cation prematurely and prior to entering the mitochondria.This may potentially deactivate the compound even before it reaches its target site.Therefore, a fine balancing of these various properties seems necessary for the development of an optimum complex.

Conclusion
We have reported on the syntheses and characterizations of a series of cationic gold(I) heteroleptic complexes bearing the pyrazole-derived FPyr ligand, and either a benzimidazolederived NHC of the type RR′-bimy (3-9), or a non-NHC coligand L (10-12).Complexes 3-9 were synthesized using [AuCl-(FPyr)] (1) and the appropriate disubstituted benzimidazolium salt in the presence of K 2 CO 3 .10-12, on the other hand, were synthesized using 1 and the necessary pro-ligand in the presence of an excess of KPF 6 .The cytotoxic activities of all ten complexes were studied with the NCI-H1666 non-small cell lung cancer cell line, and their performances were compared with the previously reported activity of the gold(I) hetero-bis (carbene) complex [Au(FPyr)( i Pr 2 -bimy)]PF 6 (2), and cisplatin.While the complexes reported in this work are not as cytotoxic as 2, some do show comparable performance, and most complexes perform better than cisplatin.Our lab is currently in the midst of further investigations of some of the complexes, focussing primarily on their physiochemical and pharmacokinetic properties, as well as their in vitro toxicities to healthy cell lines.We look forward to reporting our findings from these studies in the near future.

General considerations
All operations were performed without taking precautions to exclude air and moisture, and all solvents and chemicals were used as received. 1H, 13 C, 19 F and 31 P NMR spectra were recorded on a Bruker ACF 300 spectrometer or a Bruker AMX 500 spectrometer.The chemical shifts (δ) were internally referenced to the residual solvent signals relative to tetramethylsilane ( 1 H and 13 C), or externally to CF 3 CO 2 H ( 19 F) and 85% H 3 PO 4 ( 31 P).ESI mass spectra were measured using a Finnigan LCQ spectrometer.Elemental analyses were performed on an Elementar Vario Micro Cube elemental analyser at the Department of Chemistry, National University of Singapore.Complex 1 was synthesized as previously reported. 6The azolium salts 1,3dibenzhydrylbenzimidazolium bromide, 10 1,3-dibenzylbenzimidazolium bromide, 10 1,3-diisobutylbenzimidazolium bromide, 10 1,3-dipropylbenzimidazolium bromide, 11 1,3-diethylbenzimidazolium bromide, 12 and 1,3-di(2-propenyl)benzimidazolium bromide 13 were synthesized according to literature procedures, albeit with slight modifications.1-Benzhydryl-3-isopropylbenzimidazolium bromide was synthesized from 1-benzhydrylbenzimidazole, and the procedures for the syntheses of both are reported in this section.All azolium salts were converted to their hexafluorophosphate analogues via salt metathesis reaction with KPF 6 in acetone, stirred overnight at ambient temperature.
1-Benzhydrylbenzimidazole. Benzimidazole (1.18 g, 10 mmol) was suspended in CH 3 CN (20 mL).NaOH (1.60 mL, 10 mmol, 6.25 M) was added to the suspension, and the reaction mixture was stirred for 30 min before a solution of benzhydryl bromide (2.47 g, 10 mmol) dissolved in CH 3 CN (10 mL) was added.The reaction mixture was heated under reflux overnight, following which it was filtered through Celite.The filtrate was dried over anhydrous Na 2 SO 4 , filtered, and all volatiles were removed under vacuum affording a beige oil.The crude product was purified by flash column chromatography (silica gel, ether) (1.45 g, 5.1 mmol, 51%). 1   1-Benzhydryl-3-isopropylbenzimidazolium bromide.1-Benzhydrylbenzimidazole (1.42 g, 5.0 mmol) was suspended in CH 3 CN (2 mL).Isopropyl bromide (0.5 mL, 5.3 mmol) was added to the suspension, and the reaction mixture was refluxed for 24 h.After cooling to ambient temperature, all volatiles were removed under vacuum.The crude product was washed several times with ethyl acetate and dried under vacuum to yield a white solid (1.58 g, 3.9 mmol, 78%). 1  General procedure for the synthesis of gold(I) hetero-bis-(carbene) complexes 3-9 Complex 1 (1 equiv.)and the appropriate azolium hexafluorophosphate salt (1 equiv.)were dissolved in acetone.K 2 CO 3 (1.3 equiv.) was added to the solution, and the resulting mixture was stirred for 24 h at ambient temperature.The solvent of the reaction mixture was then removed under vacuum.The residue was suspended in CH 2 Cl 2 and filtered over Celite.The solvent of the filtrate was removed under vacuum, and the resulting residue was washed thrice with ethyl acetate or diethyl ether.The crude product was purified by crystallization via vapour diffusion of diethyl ether into a solution of the compound in CHCl 3 or CH 2 Cl 2 .All complexes were obtained as colourless crystals.

X-ray diffraction studies
X-ray data for 5•CHCl 3 , 6, and 9-11 were collected with a Bruker AXS SMART APEX diffractometer, using Mo Kα radiation at 100(2) K with the SMART suite of programs. 14Data were processed and corrected for Lorentz and polarization effects with SAINT, 15 and for absorption effect with SADABS. 16tructural solution and refinement were carried out with the SHELXTL suite of programs. 17The structures were solved by direct methods to locate the heavy atoms, followed by difference maps for the light, non-hydrogen atoms.All hydrogen atoms were placed in calculated positions.All non-hydrogen atoms were generally given anisotropic displacement parameters in the final model.A summary of the most important crystallographic data is provided in Table 3.

Cytotoxicity studies
In vitro toxicities of 3-12 were determined by CellTiter 96 AQueous Non-Radioactive Cell Proliferation Assay (MTS).NCI-H1666 cells were maintained in complete RPMI medium with 10% FBS and 1% PenStrep.Standard DMSO solutions of 3-12 were diluted with complete RPMI medium, with the final concentration of DMSO in each diluted solution being less than 0.5%.The cells were seeded in 96-well plates at a density of 2000 cells per well in 100 μL of complete RPMI medium without antibiotics and cultured overnight at 37 °C in a humidified atmosphere containing 5% CO 2 .A 100 μL portion of media containing different concentrations of 3-12 was added in the wells to a final concentration from 0.0005 to 50 μM.Compound-free solvent controls were also included.