Antimicrobial Activity of Silver Camphorimine Complexes against Candida Strains

Hydroxide [Ag(OH)L] (L = IVL, VL, VIL, VIIL), oxide [{AgL}2}(μ-O)] (L = IL, IIL, IIIL, VL, VIL) or chloride [AgIIL]Cl, [Ag(VIL)2]Cl complexes were obtained from reactions of mono- or bicamphorimine derivatives with Ag(OAc) or AgCl. The new complexes were characterized by spectroscopic (NMR, FTIR) and elemental analysis. X-ray photoelectron spectroscopy (XPS), ESI mass spectra and conductivity measurements were undertaken to corroborate formulations. The antimicrobial activity of complexes and some ligands were evaluated towards Candida albicans and Candida glabrata, and strains of the bacterial species Escherichia coli, Burkholderia contaminans, Pseudomonas aeruginosa and Staphylococcus aureus based on the Minimum Inhibitory Concentrations (MIC). Complexes displayed very high activity against the Candida species studied with the lowest MIC values (3.9 µg/mL) being observed for complexes 9 and 10A against C. albicans. A significant feature of these redesigned complexes is their ability to sensitize C. albicans, a trait that was not found for the previously investigated [Ag(NO3)L] complexes. The MIC values of the complexes towards bacteria were in the range of those of [Ag(NO3)L] and well above those of the precursors Ag(OAc) or AgCl. The activity of the complexes towards normal fibroblasts V79 was evaluated by the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay. Results showed that the complexes have a significant cytotoxicity.


Introduction
The resistance of microorganisms to conventional antimicrobials is presently a serious threat to public health worldwide, representing a huge financial burden for public health systems. The group of bacterial pathogens known as ESKAPE is of particular concern, which includes Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp. [1]. In addition, fungal infections, and notably candidiasis caused by members of the Candida genus, are also of increasing concern worldwide. These infections range from superficial infections to life-threatening disseminated mycoses [2,3]. Although Candida albicans remains the major

Results
A new set of camphorimine silver complexes were synthesized using silver acetate (AgOAc) or silver chloride (AgCl) as metal precursors to tune the properties and the antimicrobial activity of the complexes. The objective was to keep the structure of the complexes while replacing the nitrate ion by a less acidic anionic co-ligand (OAc − or Cl − ), aiming to overcome the lack of antifungal activity displayed by the nitrate complexes [Ag(NO 3 )L] towards C. albicans. The absence of antifungal activity was accompanied by formation of silver nanoparticles (AgNPs) [18]. The acidic character of the nitrate ion was hypothesized to favor the reduction processes mediated by a protein existing in C. albicans, but not at other Candida species. Less acidic co-ligands (OAc − , Cl − ) are expected to be less efficient in the activation of AgNPs formation.
The low solubility of silver acetate (AgOAc) in solvents other than water requires that reactions with camphor compounds are carried out in H 2 O/EtOH since the camphor derivatives used as ligands are not soluble in H 2 O. Solutions of silver acetate in water have acidic character (pH ca. 4), consistent with release of acetic acid (pKa = 4.76) [23] and formation of silver hydroxide (AgOH) or silver oxide (Ag 2 O) solutions (Scheme 1). Since different forms of silver species co-exist in solution, complexes with different metal cores can be obtained, depending on the characteristics of the camphorimine derivatives ( Figure 1) and the experimental conditions. The absence of OH stretches in the IR spectra of complexes 1-3 (Table 1) support their formulation as oxide rather than hydroxide complexes. The hydroxide complexes (4,5,7) fit in a 1:1 ligand to metal ratio, consistent with a coordination polymer character. At complex 4 ([Ag( IV L´)(OH)]) the ligand (Y=NH2) is protonated ( IV L´ = IV L·HOOCCH3), as confirmed by elemental analysis and FTIR (Table 1) through bands at 3455 and 3340 cm -1 (attributed to the OH − and NH4 + groups) and at 1593, Since different forms of silver species co-exist in solution, complexes with different metal cores can be obtained, depending on the characteristics of the camphorimine derivatives ( Figure 1) and the experimental conditions. Since different forms of silver species co-exist in solution, complexes with different metal cores can be obtained, depending on the characteristics of the camphorimine derivatives ( Figure 1) and the experimental conditions.  Tables 1 and 2 for details) The absence of OH stretches in the IR spectra of complexes 1-3 (Table 1) support their formulation as oxide rather than hydroxide complexes. The hydroxide complexes (4,5,7) fit in a 1:1 ligand to metal ratio, consistent with a coordination polymer character. At complex 4 ([Ag( IV L´)(OH)]) the ligand (Y=NH2) is protonated ( IV L´ = IV L·HOOCCH3), as confirmed by elemental analysis and FTIR  Since different forms of silver species co-exist in solution, complexes with different metal cores can be obtained, depending on the characteristics of the camphorimine derivatives ( Figure 1) and the experimental conditions.  Tables 1 and 2 for details) The absence of OH stretches in the IR spectra of complexes 1-3 (Table 1) support their formulation as oxide rather than hydroxide complexes. The hydroxide complexes (4,5,7) fit in a 1:1 ligand to metal ratio, consistent with a coordination polymer character. At complex 4 ([Ag( IV L´)(OH)]) the ligand (Y=NH2) is protonated ( IV L´ = IV L·HOOCCH3), as confirmed by elemental analysis and FTIR (Table 1) through bands at 3455 and 3340 cm -1 (attributed to the OH − and NH4 + groups) and at 1593, Scheme 2. Types of complexes obtained from ligands I L-VI L (see Tables 1 and 2 for details). The absence of OH stretches in the IR spectra of complexes 1-3 (Table 1) support their formulation as oxide rather than hydroxide complexes. The hydroxide complexes (4,5,7) fit in a 1:1 ligand to metal ratio, consistent with a coordination polymer character. At complex 4 ([Ag( IV L´)(OH)]) the ligand (Y=NH 2 ) is protonated ( IV L´= IV L·HOOCCH 3 ), as confirmed by elemental analysis and FTIR (Table 1) through bands at 3455 and 3340 cm −1 (attributed to the OH − and NH 4 + groups) and at 1593, 1567 cm −1 (attributed to the acetate (COO − ) group). The metal to ligand ratio at 6 (3:2) differs from that of all the other complexes, conceivably due to steric demands of the bicamphorimine ligand. All complexes (1-7) were characterized by spectroscopic (FTIR, NMR) and analytical techniques. The relevant spectroscopic characteristics are highlighted in Table 1 (see experimental section for further details). Complex 1 is not sufficiently soluble to obtain NMR data, thus formulation was supported by X-ray photoelectron spectroscopy (XPS).

Analysis of Complex 1 by XPS
Complex 1 was characterized by XPS. Besides the survey spectrum (not shown), the detailed regions C 1s, N 1s, O 1s, and Ag 3d were analyzed and are shown in Figure 2. 1567 cm -1 (attributed to the acetate (COO − ) group). The metal to ligand ratio at 6 (3:2) differs from that of all the other complexes, conceivably due to steric demands of the bicamphorimine ligand. All complexes (1-7) were characterized by spectroscopic (FTIR, NMR) and analytical techniques. The relevant spectroscopic characteristics are highlighted in Table 1 (see experimental section for further details). Complex 1 is not sufficiently soluble to obtain NMR data, thus formulation was supported by X-ray photoelectron spectroscopy (XPS).

Analysis of Complex 1 by XPS
Complex 1 was characterized by XPS. Besides the survey spectrum (not shown), the detailed regions C 1s, N 1s, O 1s, and Ag 3d were analyzed and are shown in Figure 2.  Ag 3d region displays a doublet with the main component, Ag 3d5/2, centered at 368.4 ± 0.2 eV and the minor component, Ag 3d3/2, at a BE 6 eV higher. N 1s is fittable with a single peak centered at 399.9 ± 0.2 eV. C 1s was fitted with a main peak (used to correct all the binding energies for charge accumulation effects) assigned to all the carbons just bound to other carbon atoms and/or hydrogen atoms set at 285 eV. Other peaks at 285.6 ± 0.2, 286.6 ± 0.

Silver Chloride Derived Complexes
The solubility of silver chloride in common solvents is even lower than that of silver acetate, however it is reasonably soluble in ammonia. Thus, the reactions of AgCl with the camphorimine derivatives (L) were performed in NH3·H2O/EtOH. In such basic medium, silver oxide exists in solution, thus accounting for the formation of the oxide complexes (3, 9, 10A).  Ag 3d region displays a doublet with the main component, Ag 3d5/2, centered at 368.4 ± 0.2 eV and the minor component, Ag 3d3/2, at a BE 6 eV higher. N 1s is fittable with a single peak centered at 399.9 ± 0.2 eV. C 1s was fitted with a main peak (used to correct all the binding energies for charge accumulation effects) assigned to all the carbons just bound to other carbon atoms and/or hydrogen atoms set at 285 eV. Other peaks at 285.6 ± 0.2, 286.6 ± 0.

Silver Chloride Derived Complexes
The solubility of silver chloride in common solvents is even lower than that of silver acetate, however it is reasonably soluble in ammonia. Thus, the reactions of AgCl with the camphorimine derivatives (L) were performed in NH 3 ·H 2 O/EtOH. In such basic medium, silver oxide exists in solution, thus accounting for the formation of the oxide complexes (3, 9,   The cationic character of 8 was further confirmed through conductivity measurement (138 Ω −1 .cm 2 .mole −1 ) in acetonitrile. The value is within the range (120-160 Ω −1 .cm 2 .mole −1 ) expected for a 1:1 electrolyte [24]. Such as for the above complexes 1-7, the characterization of complexes 8-10A was achieved by elemental analysis, FTIR and NMR. Some relevant spectroscopic details are displayed in Table 2.

Antimicrobial Activity Assessment
The antimicrobial properties of the above complexes were assessed for C. albicans and C. glabrata, as well as for the bacterial pathogens E. coli ATCC25922, S. aureus Newman, B. contaminans IST408 and P. aeruginosa 477, based on the evaluation of the values of Minimum Inhibitory Concentration (MIC). The selected bacterial strains represent pathogens of medical relevance, difficult to treat and eradicate worldwide, mainly due to their resistance to multiple antibiotics. B. contaminans IST408 was isolated from a Portuguese Cystic Fibrosis patient [24]. P. aeruginosa and S. aureus are members of the ESKAPE group, responsible for many hospital-and community-acquired infections [1]. E. coli ATCC25922 is a commonly used reference in antimicrobial activity assays. The antimicrobial activities of complexes 1, 4, and 5 were not assessed due to their low solubility (1) or stability (4,5).

Antimicrobial Activity Assessment
The antimicrobial properties of the above complexes were assessed for C. albicans and C. glabrata, as well as for the bacterial pathogens E. coli ATCC25922, S. aureus Newman, B. contaminans IST408 and P. aeruginosa 477, based on the evaluation of the values of Minimum Inhibitory Concentration (MIC). The selected bacterial strains represent pathogens of medical relevance, difficult to treat and eradicate worldwide, mainly due to their resistance to multiple antibiotics. B. contaminans IST408 was isolated from a Portuguese Cystic Fibrosis patient [24]. P. aeruginosa and S. aureus are members of the ESKAPE group, responsible for many hospital-and community-acquired infections [1]. E. coli ATCC25922 is a commonly used reference in antimicrobial activity assays. The antimicrobial activities of complexes 1, 4, and 5 were not assessed due to their low solubility (1) or stability (4,5).
All the complexes essayed display anti-Candida activity that range from 15.6 µg/mL (1, 2, 6, 9, 10A) to 125-250 µg/mL (8, 10). The antibacterial activity of the complexes (Table 3) ranged from 19 µg/mL (P. aeruginosa, 2) to ≥112 µg/mL (6, 8, 10). The MIC values measured for the ligands ( II L, III L, V L and VI L) display very high MIC values (≥500 µg/mL) consistent with their lack of antimicrobial activity. The MIC values (Table 3) show that the complexes are active against bacterial strains E. coli ATCC25922, B. contaminans IST408, P. aeruginosa 477, and S. aureus Newman, although at relatively high MIC values. In contrast, the MIC 50 values obtained for the two Candida species are very low, thereby showing that the complexes have higher antifungal than antibacterial activity. Although previous work showed that silver camphorimine complexes [Ag(NO 3 )L] have high anti Candida spp. activity (MIC 50 2.0-15.6 µg/mL for C. parapsilosis) the complexes 2, 6, 9, 10A, display values (7.8 µg/mL, 9) that are even lower than those formerly reported against C. glabrata (MIC 50 ≥15.6 µg/mL) [18]. More important, all complexes (except 8) are efficient against C. albicans, a feature not observed for the nitrate complexes [Ag(NO 3 )L] [18]. So, by replacing nitrate by hydroxide or oxide co-ligands, we were able to synthesize complexes that inhibit growth of C. albicans even more efficiently (MIC 50 3.9 µg/mL; 9, 10A) than C. glabrata. The lack of activity of [Ag( II L)]Cl (8) towards the two Candida species and the bacterial strains under study is attributed to its cationic character. Complex 10, which is also cationic, displays a relatively low activity (Table 3). These results reinforce those previously obtained for the cationic [Ag(OC 10 H 13 NOH) 2 ]NO 3 [19]. The ionic character of the complexes decreases their lipophilicity and conceivably makes it more difficult for their penetration into the intracellular environment of Candida spp. or bacterial cells, thereby reducing their activity.
To obtain insights into the toxicity of the complexes towards mammalian cells, the IC 50 measurements of representative complexes were evaluated towards V79 normal fibroblasts which are cells commonly used to assess the toxicological effects of drugs [25]. Data shows that IC 50 values of the complexes are low and comparable or even lower than those of the MIC values obtained for fungi (Table 3).
These results were not completely unexpected since fungi and mammalian cells are eukaryotes and therefore some of the mechanisms by which the complexes exert toxicity against the Candida spp. may be conserved in the mammalian cells [26]. This difficulty in achieving specificity is a recognized challenge in the design of new compounds selectively targeting fungal cells. Future work will focus on the design of silver camphorimine complexes with both reduced cytotoxicity and enhanced antimicrobial activities.

General Procedures
The camphorimines were obtained from camphorquinone by reaction with the appropriate amine or hydrazine in ethanol using reported procedures [18,[27][28][29]. In the case of air sensitive complexes, Schlenk and vacuum techniques were used. Ethanol was purchased from Fisher Scientific, ammonia from Sigma-Aldrich and acetonitrile from Carlo Erba. The amines and hydrazines were purchased from Sigma-Aldrich and silver acetate and silver chloride from Merck.
The ESI mass spectrum was obtained on a LCQFleet ion trap mass spectrometer equipped with an electrospray source (Thermo Scientific TM , Waltham, MA USA), operating in the positive ion mode. The XPS data was obtained using a Kratos XSAM800 equipment.

Synthesis
Complexes 1-7 were obtained from reaction of the suitable ligand with silver acetate (1:1). Air was partially excluded by bubbling of nitrogen (3 minutes). The reaction mixtures were protected from light to preclude reduction of Ag + to Ag 0 . The typical procedure is described for 1. Complexes 8-10 were obtained from reaction of the suitable ligand with silver chloride (1:1). A typical procedure is described for 8.
[ Compound 3 can alternatively be obtained from reaction of AgCl (50 mg; 0.35 mmol in 5 mL of ammonia 33%) with III L (89 mg; 0.35mmol in 5 mL EtOH) upon stirring overnight, filtration to eliminate residues of silver followed by solvent evaporation until precipitation which is then dried under vacuum to obtain 3. The yield (60%) is lower than that in reaction of III L with Ag(OAc).

Antibacterial Activity Determinations
The antibacterial activity of compounds was assessed by determining their Minimal Inhibitory Concentration (MIC) towards the Gram-positive Staphylococcus aureus Newman and the Gram-negative Escherichia coli ATCC 25922, Pseudomonas aeruginosa 477 and Burkholderia contaminans IST408. These bacterial strains are clinical isolates and were chosen as representatives of important bacterial pathogens [1,25,[30][31][32]. MICs were determined using Mueller Hinton Broth (MHB; Becton, Dickinson and Company) as growth medium, based on microdilution assays, using previously described methods [20,21]. In brief, a colony from a bacterial culture freshly grown in MHB solid medium was transferred into MHB liquid medium and incubated for 4-5 h with agitation (250 rpm) at 37 • C. Bacterial cultures were then diluted using fresh MHB to obtain ca. 10 6 colony forming units (CFUs) per mL. These cultures were used to inoculate approximately 5 × 10 5 CFUs per mL in 96-well polystyrene microtiter plates containing 100 µL of MHB supplemented with different concentrations of each compound, obtained by 1:2 serial dilutions ranging 0.5 to 512 µg/mL. Stock solutions of compounds were prepared with DMSO. After inoculation, microtiter plates were incubated at 37 • C for 20 h. Bacterial growth was assessed by measuring the cultures optical density at 640 nm, in a SPECTROstarNano (BMG LABTECH) microplate reader. Experiments were carried out at least four times in duplicates. Wells containing 100 µL of 1× concentrated MHB and 100 µL of 10 6 CFUs per mL were used as positive controls, while wells containing 200 µL of sterile MHB 1 × concentrated were used as negative controls.

Assessment of Complexes Anti-Candida Activity
The ability of the complexes 1-10A or of the ligands to inhibit growth of C. albicans or C. glabrata was assessed using the highly standardized microdilution method recommended by EUCAST (European Committee on Antimicrobial Susceptibility Testing). The MIC 50 values were considered to be the concentration of drug that reduced yeast growth by more than 50% of the growth registered in drug-free medium [33]. The strains used in this work were C. albicans SC5314 and C. glabrata CBS138, largely used as reference strains. Briefly, cells of the different species were cultivated (at 30 • C and with 250 rpm orbital agitation) for 17 h in Yeast Potato Dextrose (YPD) growth medium and then diluted in fresh Roswell Park Memorial Institute (RPMI) growth medium (Sigma) to obtain a cell suspension having an OD 530nm of 0.05. From these cell suspensions, 100 µL aliquots were mixed in the 96-multiwell polystyrene plates with 100 µL of fresh RPMI medium (control) or with 100 µL of this same medium supplemented with 0.98-500 µg/mL of the different compounds. As a control we also examined the inhibitory effect of Ag(OAc) or of AgCl. After inoculation, the 96-multiwell plates were incubated without agitation at 37 • C for 24 h. After that time, cells were re-suspended and the OD 530nm of the cultures was measured in a SPECTROstarNano (BMG LABTECH) microplate reader. The MIC 50 value was taken as being the highest concentration tested at which the growth of the strains was 50% of the value registered in the control lane.

Toxicity Assessment
The toxicity of the compounds was evaluated towards normal fibroblasts V79, obtained from the American Type Culture Collection (ATCC). The cell lines were grown in Dulbeco´s Modified Eagle Medium (DMEM) + Glutamax ® medium supplemented with 10% Fetal Bovine serum (FBS) and maintained in a humidified atmosphere at 37 • C using an incubator (Heraeus, Germany) with 5% CO 2 . Cell viability was measured by the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyl tetrazolium bromide) assay, based on the conversion of the tetrazolium bromide into formazan crystals by living cells which determines mitochondria activity [34]. For the assay, cells were seeded in 96-well plates at a density of 10 4 cells per well in 200 µL medium and allowed to attach overnight. Complexes were first diluted in DMSO to solubilize and then in medium to prepare the serial dilutions in the range 10 −7 -10 −4 M. The maximum concentration of DMSO in the medium (1%) had no toxicity effect. After careful removal of the medium, 200 µL of each dilution were added to the cells, and incubated for another 48 h at 37 • C. At the end of the treatment, the medium was aspirated and 200 µL of MTT solution (1.5 mM in PBS) was applied to each well. After 3 h at 37 • C, the medium was discarded and 200 µL of DMSO was added to solubilize the formazan crystals. The cellular viability was assessed by measuring the absorbance at 570 nm using a plate spectrophotometer (Power Wave Xs, Bio-Tek). The IC 50 values were calculated using the GraphPad Prism software (version 5.0). Results are mean ± SD of at least two independent experiments done with six replicates each and represent the percentage of cellular viability related to the controls (no treatment).

X-ray Photoelectron Spectroscopy
For X-ray photoelectron spectroscopy characterization (XPS), a XSAM800 XPS dual anode spectrometer from KRATOS was used. The unmonochromatic Mg Kα radiation (main line at hν = 1256.6 eV) was used. Operating conditions, data acquisition and data treatment are described elsewhere [35]. For charge correction purposes, carbon bound to carbon and hydrogen in C 1s peak was set to a binding energy (BE) of 285 eV. For quantification purposes, the following sensitivity factors were used: 0.318 for C 1s, 0.736 for O 1s, 0.505 for N 1s, and 6.345 for Ag3d.

Conclusions
Silver hydroxide [Ag(OH)L] (L = IV L, V L, VI L, VII L), oxide [{AgL} 2 }(µ-O)] (L = I L, II L, III L, V L, VI L) and homoleptic [Ag II L]Cl, [Ag( VI L) 2 ]Cl camphorimine complexes were synthesized using Ag(OAc) or AgCl as metal sources. The basic characteristics of the reaction medium prompted the formation of hydroxide or oxide rather than acetate or chloride complexes. The selection of the camphorimine ligands (L) that encompass mono-and bi-camphors, allowed the design of complexes with considerable distinct electronic and steric properties and thus different biological activities.
In summary, the most relevant achievement of this study is that the new oxo and hydroxo silver camphorimine complexes overreach the resistance of C. albicans. Additionally, the complexes reach MIC 50 values for C. glabrata even lower than those previously reported for the camphorimine nitrate complexes [Ag(NO 3 )L]. In fact, the antifungal activity of the oxo and hydroxo silver camphorimine complexes is even higher towards C. albicans than C. glabrata.