Internal and External Influences on Stability and Ligand Exchange Reactions in Bromido[3-ethyl-4-aryl-5-(2-methoxypyridin-5-yl)-1-propyl-1,3-dihydro-2H-imidazol-2-ylidene]gold(I) Complexes

The ligand scrambling reaction of gold(I) complexes is a phenomenon occurring primarily in L–AuI–X (L = phosphine, N-heterocyclic carbene (NHC), and thiol; X = halide and thiol) complexes and has been observed among others for e.g., the bromido[3-ethyl-4-(4-methoxyphenyl)-5-(2-methoxypyridin-5-yl)-1-propyl-1,3-dihydro-2H-imidazol-2-ylidene]gold(I) complex (7a), which underwent ligand rearrangement in aqueous solutions. In this study, we investigated the influence of substituents on the 4-aryl ring of the related (NHC)AuIBr complexes (1a–9a) in terms of the conversion to the [(NHC)2AuI]+ (1b–9b) and [(NHC)2AuIIIBr2]+ (1c–9c) species. Furthermore, the influence of external factors such as solvent, temperature, concentration, and presence of halides (Cl–, Br–, and I–) or hydroxyl ions was studied to gain a deeper understanding of the ligand rearrangement reaction. The substituent on the 4-aryl ring has a marginal impact on the scrambling reaction. Out of the investigated organic solvents (dimethylformamide (DMF), dimethyl sulfoxide (DMSO), ethanol (EtOH), methanol (MeOH), and acetonitrile (ACN)), only ACN separates single complex molecules. In all other solvents, relatively stable ((NHC)AuIBr)2 dimers are present. The addition of water to ACN solutions forces the formation of such dimeric units, starting the transformation to [(NHC)2AuI]+ and [(NHC)2AuIIIBr2]+. The rate-determining step is the release of Br– from a T-shape intermediate because an excess of KBr terminates this reaction. Furthermore, it is obvious that only single molecules react with halides. The aurophilic interactions between two (NHC)AuIBr molecules are too strong in the presence of water and largely impeded reaction with halides. As a single molecule, the reaction with Cl– (e.g., in a 0.9% NaCl solution) is notable, while I– even leads to a fast and quantitative conversion to (NHC)AuII and finally to [(NHC)2AuI]+.


■ INTRODUCTION
Ligand exchange reactions of metal complexes are a phenomenon with increasing interest in the scientific community. 1−8 It is well known that rhodium, ruthenium, platinum, and gold complexes suffer ligand replacement reactions. 2,6,8 For instance, cisplatin has to hydrolyze to reactive aqua species 9−12 prior to binding to its targets. 2,13−19 Additionally, ligand exchange is used to coordinate metallodrugs, e.g., gold complexes, to carrier ligands for high accumulation within cells. 18,20−23 The stability of metal complexes depends on the central ion and the used ligands. Complexes of Mg 2+ , Ca 2+ , K + , and Na + are less stable than those of transition metal ions and exchange the ligands rapidly. In contrast, Ru 2+ , Os 2+ , Ir 3+ , and Pt 2+/4+ complexes require hours or even days. 24,25 Gold(I) complexes are also susceptible to undergo rearrangement reactions. 4,6−8 Ligands can be categorized as carrier ligands or leaving groups, depending on their binding strength to the metal. Suitable carriers represent N-heterocyclic carbenes (NHCs). Resulting (NHC)gold(I) complexes are promising candidates for the application in medicinal and inorganic chemistry because of their anticancer activity 26−29 as well as luminescence 30−32 and catalytic 33 properties.
NHCs form strong σ-donor bonds to a number of metals, 34 and the electronic factors stabilize the NHC-metal bond in a push−pull mechanism as a result of the σand π-frameworks. 35,36 The electron-donor effects are higher than those of phosphines 37 and the (NHC)gold(I) complexes are regarded to be stable in solution under standard conditions (room temperature (rt), protection from light, and dry atmosphere). Nevertheless, ligand exchange reactions are observed for a variety of complexes. 38,39 Of high interest is the stability in the presence of water because it allows the estimation of their behavior under physiological conditions.
The data are important for the interpretation of the in vitro results since 1a−9a exhibited antiproliferative activity in ovarian cancer and leukemia cell lines at low micromolar concentrations, 41 whereas 1b−9b caused these effects already at nanomolar concentrations. 42 Therefore, it is of interest to know more about the conditions for the transformation of 1a− 9a to the higher active 1b−9b. Furthermore, the reaction with chloride is pivotal because of the high concentration in the cell culture medium, which leads to the formation of (NHC)Au I Cl complexes. Thus, we investigated the stability of (NHC)Au I Br complexes in the presence of 0.9% NaCl on the examples of 7a and 8a.
In a second approach, the ligand exchange reactions in 8a were studied using Cl − , Br − , I − , and OH − as model nucleophiles. Such information is of relevance to estimate the reactivity of (NHC)Au I Br complexes against bionucleophiles.
■ RESULTS Sample Preparation. A solution of 1a−9a (1 mM) in an appropriate mixture of ACN and water was monitored by HPLC for 72 h. The complexes were dissolved in ACN (or other organic solvents) and then diluted with water. The samples were filtered through a 0.20 μm membrane filter and analyzed after various incubation times (0, 24, 48, and 72 h) in HPLC experiments using the Shimadzu Prominence HPLC system equipped with a SIL-20A HT autosampler, a CTO-10AS VP column oven, a DGU-20A degasser, an SPD-M20A detector, LC-20AD pumps, and a KNAUER 250 × 4 nm 2 Eurospher 100-C18 column. The mobile phase consisted of ACN and water with 0.1% trifluoroacetic acid (TFA). Separation of the complexes from the reaction product was possible with gradient elution from 70 to 90% ACN and a flow rate of 1 mL/min at an oven temperature of 35°C. All solvents were degassed before use. The injection volume was 20 μL, and UV−vis detection was performed at 254 nm. Each measurement was performed in triplicates. Three-dimensional (3D) graphs of HPLC chromatograms were prepared using OriginPro 2016 (Northampton, MA). The peaks were assigned by the analysis of their UV−vis spectra or comparison with synthesized reference compounds ((NHC)Au I X; X = I: 8d, X = Cl: 8e, X = OH: 8f). 41 Internal Influences. Substituent Effects. To study the influence of the substituents on the 4-aryl ring, the bromido (NHC)gold(I) complexes 1a−9a (1 mM), each dissolved in ACN/water (50:50, v/v), were analyzed by HPLC for their degradation profile during 72 h of incubation at rt.
The amount of water determined the ligand scrambling of (NHC)Au I Br and the following oxidation of the resulting [(NHC) 2 Au I ] + to [(NHC) 2 Au III Br 2 ] + (Scheme 1). The highest degradation caused a 50% portion of water. 40 Therefore, 1a−9a were investigated under these conditions. The results are depicted in Figure 1 and data are listed in Table  1 (see also Figures S3−S11 and Table S1 in the Supporting Information). Most complexes degraded in the range of 10−20% during 24 h. Only 3a and 6a showed higher stability (5.15 and 7.25% degradation, respectively). Through 72 h of incubation, the peak areas of the initial peaks (1a−9a) further decreased in the HPLC chromatograms. 9a underwent the highest transformation, with 24.30% to 9b and 6.97% to 9c. It should be mentioned that the solubility of 9a in ACN/water (50:50, v/v) was too low to realize a 1 mM solution. Therefore, the results must be handled with care.
External Influences. For extended investigations on the dependence of external parameters on the ligand scrambling, complexes 7a and 8a were selected because they showed the best physicochemical properties, especially solubility in various solvents, and high resolution of the degradation products in the HPLC chromatograms.
In ACN, the transformation to 7b (2.75% at t = 0 h) was suppressed for 72 h (2.85%), while it increased in other solvents after 72 h by 2−3% (Table S2 in the Supporting Information). It is noteworthy that the solubility of 7a was insufficient in EtOH (<1 mM), indicated by the reduced area of the main peak (Figures 2 and S18 (Supporting Information)).
The addition of water to the respective solutions of 7a decreased the solubility. The complex was insufficiently soluble in DMF/water (50:50, v/v) ( Figure S15 in the Supporting Information), while in the case of the other mixtures, the target concentration of 1 mM could be realized (t = 0 h), but the peak area in the HPLC chromatograms (Figures S17, S19, and S21 in the Supporting Information) drastically decreased within 24 h. Only in ACN/water ( Figure S13 in the Supporting Information), sufficient solubility was guaranteed.
Concentration and Temperature Dependence. The dependence of the scrambling reaction on the concentration was studied in ACN/water (50:50, v/v). Table 2 Table  S3 (Supporting Information).
The transformation of 8a to the [(NHC) 2 Au I ] + species 8b and the oxidized [(NHC) 2 Au III Br 2 ] + form 8c increased with concentration. At 0.5 mM, the proportion of 8a amounted to 86.27%, at 1 mM to 80.58%, and at 2 mM to 76.17%. Unfortunately, the concentrations could not be reduced to those relevant for in vitro studies. The sensitivity of the HPLC limited the analysis at concentrations < 50 μM.
Effect of KCl, KBr, KI, and KOH Addition. Besides ligand scrambling reactions, the exchange of the bromido ligand in gold(I) is possible. Especially, the reactions with other halides are of interest. Therefore, 7a and 8a solutions (1 mM) in ACN/water (50:50, v/v) were incubated with an excess of KCl, KBr, or KI (20 equiv each) for 72 h at rt. Figure 4 illustrates the chromatograms (t = 72 h) exemplarily for 8a. Analytical data for 8a are listed in Table  3 (for 7a, see Table S4 and Figures S25−S28 in the Supporting Information).
KBr stabilized 8a and the amount of 8b remained unchanged during the incubation for 72 h at 22°C (t = 0 h: 2.31%; t = 72 h: 2.46%; Table S4 in the Supporting Information). Without KBr, degradation of 8a strongly increased (8b: 11.75%; 8c: 7.66%; Table 1). An opposite effect was observed upon addition of KI. The peak of 8a completely disappeared and 8b was built as the main peak (76.49%). (NHC)Au I I complex 8d was visible as a small shoulder. In contrast, only 15.44% of 8a reacted immediately after dissolution with chloride (KCl) to (NHC)Au I Cl (8e). This proportion remained constant for 72 h. Increased amounts of 8b and 8c resulted from degradation of remaining 8a.
The hydroxide ion (KOH) was used as another nucleophile. It reacted with 8a only in small amounts (2.0%) to form (NHC)Au I OH (8f) upon dissolution. The ligand scrambling reaction was nearly independent of the presence of KOH. After 72 h, 15.86% of 8b (11.75% without KOH) and 8.95% of 8c (7.66% without KOH) were detected.
Finally, it is of interest to know the extent of Br − /Cl − exchange under physiological NaCl conditions. For this purpose, the most active complex 7a 41 was dissolved in ACN and a 1.8% NaCl solution was added to obtain a 50:50 (v/v) mixture with 0.9% NaCl at a complex concentration of 1 mM. In agreement with the above-described results, 34.27% of 7a reacted initially (t = 0 h) to (NHC)Au I Cl (7e) (Table S7 and Figure S38 in the Supporting Information). Besides, the chromatogram exhibited peaks of 7a (58.06%) and 7b (7.16%). While the amount of 7e remained unchanged (33.46%) for 72 h, about 5% of 7a was transformed to 7b (Table S7 in the Supporting Information). Complex 7b was investigated in the same way and proved to be stable in the NaCl solution ( Figure S39 in the Supporting Information) and no oxidation to 7c took place.
Effects of the substituents on the C−Au I Br bond can be excluded because N−C−N resonances in the 13 C NMR spectra (CDCl 3 ) are nearly identical in the range of 173.3− 174.6 ppm, indicating a comparable influence on the strength of the Au I −Br bond.
The solubility of 7a in dry ACN, DMF, DMSO, and MeOH allowed the preparation of 1 mM stock solutions. Only in EtOH, this concentration could not be realized. In aqueous mixtures of these solvents, the solubility was strongly reduced. Indeed, it was possible to achieve (with exception of DMF) initially 1 mM concentrations in 50:50 (v/v) mixtures, but 7a crystallized during the incubation for 72 h, indicated by a   Inorganic Chemistry pubs.acs.org/IC Article strong reduction of the peak area in the chromatograms (Figures S15, S17, S19, and S21 in the Supporting Information). The X-ray analysis of the precipitate confirmed the presence of 7a dimers with strong aurophilic bonds. 41 Hence, it can be deduced that the higher polarity of the solvents/water mixtures forces the formation of these "hydrophobic" gold(I)−gold(I) interactions in solution, which is accompanied by lower solubility than under water-free conditions. Furthermore, upon the building of the dimers, ligand scrambling took place and the peak area of the [(NHC) 2 Au I ] + species increased to 8−10% related to the initial area of the (NHC)Au I Br peak. It was further documented that dry ACN stabilizes (NHC)Au I Br complexes as monomers in solution, preventing degradation. In DMF, DMSO, EtOH, and MeOH, the portion of the [(NHC) 2 Au I ] + complex increased within 72 h by 2−3%. Therefore, it can be assumed that dimer formation through aurophilic interactions is possible and these organic solvents are inappropriate for the preparation of stock solutions and long-time storage. Unfortunately, most scientists use DMSO and DMF as solvents for in vitro experiments of gold(I) complexes. An increased transformation to the more active [(NHC) 2 Au I ] + species upon storage is possible and has to be excluded prior to the start of the experiments because it sophisticates the collected data.
We already presented a plausible mechanism for the ligand scrambling, 40 starting from dimeric (NHC)Au I Br adducts, stabilized by strong gold(I)−gold(I) interactions, 43−46 originating from the lanthanide contraction and the relativistic effects in gold. 35,47,48 A simplified schematic drawing is depicted in Scheme 2B.
To gain a better knowledge about the conditions for this reaction, we investigated its dependence on concentration and temperature in the ACN/water mixture (50:50, v/v) on the example of 8a. The proportion of the [(NHC) 2 Au I ] + (8b) and [(NHC) 2 Au III Br 2 ] + (8c) species increased with the concentration. After 72 h of incubation at rt, 86.27% of 8a remained unchanged at 0.5 mM, 80.58% at 1 mM, and only 76.17% at 2 mM. Thus, it can be assumed that higher concentrations forced the aurophilic interactions.
In the next step, it was of interest to study the reaction of 8a with nucleophiles in ACN/water (50:50, v/v). For this purpose, the complex dissolved in ACN and water containing These reactions are principally also possible in the presence of KBr. However, it was observed that the excess of Br − delays the ligand scrambling. We postulate the interference in successive reaction steps after dimer formation. Rate determining seems to be the dissociation of Br − from the Tshape intermediate (IV), which is suppressed by the excess of Br − .
Iodide is a very interesting anion to study the reactivity of metal complexes. It represents on the one hand a strong nucleophile and on the other hand an excellent leaving group. 49 The nucleophilic potency is comparable to that of guanosine and can therefore also be used to estimate the reactivity against bionucleophiles in a model reaction.
After For the interpretation of the biological results, it is of importance to know more about the transformation of drugs in a 0.9% NaCl solution. The reaction of (NHC)Au I Br with biomolecules follows the pathway depicted in Scheme 2A. Possible Br − /Cl − exchange at the gold(I) center prior to the binding to the target influences the biological outcome. In the 0.9% NaCl solution, bromido (NHC)gold(I) complex 7a is subjected to about 35% transformation to the related chlorido (NHC)gold(I) complex 7e immediately after dissolution. This proportion remained unchanged during the incubation for 72 h, while the amount of [(NHC) 2 Au I ] + complex 7b increased from 7.16 to 12.16%. Therefore, 7a, 7b, and 7e participated in the biological activity.
These data clearly point to a considerable role of the ligand (see also Tacke et al. 55,56 ) that acts as a leaving group in the biological activity. Therefore, regarding the interpretation of the biological effects of (NHC)Au I Br complexes, the transformation under cell culture conditions must be considered. The complexes react within minutes with NaCl to less active chlorido (NHC)gold(I) complexes but also to bis(NHC)gold(I) species with higher activity. This degradation allows only an insufficient evaluation of the contribution of the bromido derivatives on the biological effects. Thus, it is necessary to examine the reactivity of (NHC)Au I X (X = halide, NHC) derivatives in more detail, immediately after dissolution in the cell culture medium and during the first 12 h of incubation, which is relevant to cellular accumulation. Such investigations will be part of a forthcoming paper.

■ CONCLUSIONS
In this structure−activity relationship study, we investigated internal and external parameters essential for the ligand exchange reactions in bromido[3-ethyl-4-aryl-5-(2-methoxypyridin-5-yl)-1-propyl-1,3-dihydro-2H-imidazol-2-ylidene]gold(I) complexes. An increase in concentration and a temperature of 37°C favor the formation of (NHC)Au I Br dimers, followed by ligand scrambling. The ligand exchange reaction as investigated with Cl − and I − occurs only with (NHC)Au I Br monomers (not with the [(NHC) 2 Au I ] + species), resulting in (NHC)Au I X (X = Cl, I) complexes. The presence of bromide renders the dissociation of Br − from T-shape intermediate IV (Scheme 2B) and prevents the transformation to the [(NHC) 2 Au I ] + complex. The Br − /X − exchange is of high relevance because it takes place at the moment when stock solutions (organic solvents) come in contact with aqueous media, e.g., used in in vitro assays. For the interpretation of the biological results, it must be considered that not only (NHC)Au I Br but also the formed (NHC)Au I Cl and [(NHC) 2 Au I ] + complexes participate in the observed effects. Therefore, we studied the solution behavior and the cytotoxicity of (NHC)Au I X (X = Cl, Br, I) complexes in more detail. The results are of interest for the interpretation of in vitro data and will be part of a forthcoming paper.
Materials and methods; syntheses; 1 H and 13 C spectra of complex 3a; data (at t = 0−72 h) on the stability of 1a− 9a in ACN/water; stability data of 7a/8a in ACN/water dependent on solvent, concentration, salt addition, and temperature; stability in 0.9% NaCl solutions; corresponding HPLC chromatograms; UV−vis spectra of the NHC ligand, 7a, and 7b (PDF)