Functionalized Derivatives of Benzo-Crown Ethers. Part 4. Antifungal Macrocyclic Supramolecular Complexes of Transition Metal Ions Acting as Lanosterol-14-α-Demethylase Ihibitors

Poly- and mononuclear metal complexes of 2,3,11,12-bis[4-(10-aminodecylcarbonyl)]benzo-18- crown-6 (L) and Cu(II); Ni(II); Co(II) and Cr(III) have been synthesized and characterized by standard physico-chemical procedures. In the newly prepared complexes the crown moiety oxygen atoms of the macrocyclic host did not generally interact with metal ions, whereas the two amino groups of the ligand always did. Several of the newly synthesized compounds act as effective antifungal agents against Aspergillus and Candida spp., some of them showing activities comparable to ketoconazole, with minimum inhibitory concentrations in the range of 0.3−0.5 μg/mL. The mechanism of antifungal action of these coordination compounds is probably connected to an inhibition of lanosterol-14-α-demethylase, a metallo-enzyme playing a key role in sterol biosynthesis in fungi, bacteria and eukariotes.


Introduction
Since the discovery of the crown-ether ligands in the late 1960s, these types of molecules have been recognized as specific complexing agents for alkaline and alkaline-earth metal ions Il[:l as well as for organic ammonium derivatives such as amines, amino acids, and related compounds (obviously in protonated state). [31 Two type of complementarities between the metal ion and the crown ether can be distinguished. Firstly, when a metal ion directly fits into the cavity, interacting equally with all donor atoms present, spherical molecular recognition has been evidenced, leading thus to stable supramolecular complexes. [41 When the cavity size and the radius of the metal ions are not compatible with each other, out of cavity coordination or fractional coordination by other available sites has been evidenced in the isolated supramolecular complexes. [51161 The latter case is observed for many transition metal ions: because of their small size (non-compatible with the relatively large crown cavity such as that of 18-crown-6) or their preferences for lower coordination numbers, they generally do not fit into such a macrocyclic cavity. For this reason only a small number of supramolecular complexes of transition metal ions with crown-ethers have been reported up to now. [5116] Generally in such complexes, the hydrated metal ions form hydrogen bonded networks with heteroatoms present in the crown ether moiety, giving polymeric oneor two-dimensional [51161 derivatives.
In such cases, crown-ether molecules are hydrogen bonded through the metal bound water molecules. [6] We describe here another approach for obtaining some novel complexes of transition metal ions with a functionalized macrocyclic crown-ether. According to this approach, the coordination sphere of the metal ions is completed by secondary coordination sites, chemically grafted on the macrocyclic cavity, which then participate in the molecular recognition processes together with the macrocycle. In this way the metal ion macrocyclic receptor interactions become more intense, leading to new specific supramolecular arrays.
We have previously reported some ditopic macrocyclic derivatives containing different moieties in their second coordination sphere (such as amino, ammonium, pyrilium, pyridinium or L-amino acid groups). [ [12] Their complexation properties evidenced new multiple molecular recognition processes of the metal ions or of certain amino acids, due to the combination of different types of non-covalent interactions. [9][10] A selective membrane transport of Ag+/Cu ++ by using functionalized dibenzo-18-crown-6 derivatives chemically grafted in a solid heteropolysiloxane matrix was also evidenced as being due to a synergetic supramolecular effect. [ Elemental analysis data for compounds 2-9 were within +0.5% of theoretical values calculated for the proposed formulae, evidencing different stoichiometries of combinations between the macrocyclic receptor 1 and the transition metal salts. In its IR spectrum, the diamine exhibits two weak absorption bands at 3369 and 3438 cm 1 attributable to the free (unassociated amino), antisymmetrical and symmetrical N-H vibrations, respectively, and at 1423-1456 cm 1 characteristic for the VC-N vibration. The shift with 100-150 cm 1 towards lower wave numbers of the -NH2 vibrations in the IR spectra of all the complexes reported here is indicative for metal coordination through the nitrogen atom(s). The VC-N vibration also undergoes a downfield shift with 5-10 cm in the new complexes as compared to the same vibration in the free ligand. Changes of the ether vibrations are also observed (table 1) on going from bis-amine 1 to its metal complexes, both for the symmetrical (Vc-o-c sm) as well as the antisymmetrical (Vc-o-c Vc-o-c A,.) aromatic ether stretchings, probably due to the participation of ether-oxygen atoms in the coordination process, or due to hydrogen bonding with water molecules. [3] OH2 NH2 0 _ _ _ --_ _ -[ >o =_2 H2N OH2 H20 4 5 Stereochemical information for the new complexes 2-9 was obtained from the absorption electronic spectra of these compounds. Thus, in the electronic spectra of the Cu(II) derivatives CuCI4L 2 and Cu3CI4L(H2Oh 3 a large band located in the range 10-18 x 103 cm 1 has been evidenced. By deconvoluting it using the Gauss method, [14] the following transitions bands were observed: at 11.35 x 103, 14.37 x 103, 15.7 x 10 cm for complex 2 and 11.89 x 10, 14.59 x 10, 14.63 x 10" cm for complex 3, respectively, 22 22 22 according to the dz ---d .y, dxy-+dx .y and dz, dyz--)dx .y transitions. (Table 2). A second charge transfer band located at 18-25 x 103 cm has been evidenced, which overlaps with some characteristic bands of the free ligand. [14] These spectra are characteristic for copper(II) ions in a (pseudo)-octahedral configuration and acting as Lanosterol-14-fl-Demethylase Inhibitors are consistent with a distorted Oh geometry, the chromophores being of the CuNO3CI2 type for complex 2 and the CuNO3CIz and CuCl402 types for complex 3.
Biochemistry and biological activity of the new complexes Opportunistic fungal infections are an increasingly important cause of morbidity and mortality, with Aspergillus and Candida species being the most common such pathogens. [7' 181 Members of the genus Aspergillus are associated with an impressive spectrum of diseases in humans, ranging from benign colonization of the lung to severe pathologies such as invasive aspergillosis or allergic bronchopulmonary aspergillosis, t91 Although A. fumigatus has been identified as the most common etiological agent in the human diseases, being considered a pathogen and allergen at the same time, [191' I201 recent data showed the Table,,2" Electr,onic spectroscopic data for, complexes 2-9.

Complex
Absorbtion band (kK) Assigned transition 4gt---)4Y 1,u(P) apparently benign A. niger and flavus to be involved in life-threatening conditions such as fungal endocarditis I21 as well as endogenous endophthalmitis, leading in many cases to an irreversible loss of visual outcome. I221 Moreover, these and other fungi developed resistance to many of the clinically used drugs, such as ketoconazole 10 or itraconazole 11, so that novel pharmacological agents of this type are permanently needed.  The mechanism of action of many fungistatic drugs, such as the widely clinically used azoles mentioned above, 10 and 11, [23][24][25][26] consists in inhibition of a metallo-enzyme, sterol 14-ot-demethylase (CYP51A1), which is a microsomal cytochrome P-450 dependent protein belonging to a gene superfamily involved in niger, [4''71 possessing thus a similar mechanism of action with the azole antifungals. It appeared thus of great interest to test whether the coordination compounds of the crown ether derivative 1, of the type 2-8, possess antifungal activity. The rationale of looking for the antifungal activity of compounds of the type reported in the present paper is the following one: substrates of CYP51A1 and related enzymes are generally bulky molecules (sterols possessing several bulky side chains) which must be oxidized at a specific methyl group. Thus, their interaction with the active site Fe(III) ion of the prosthetic group of cytochrome P450 is governed by a multitude of both hydrophobic as well as polar interactions. I71 The same mechanism is valid for the interaction between CYP51A1 and its inhibitors (such as the azole antifungals). Moreover, many antifungal compounds (see for instance the structure of itraconazole) possess themselves a bulky and relatively rigid structure, together with one or more highly polar moieties which are important for their interaction with the prosthetic group as well as with other active site residues in its vicinity. Thus, our hypothesis was that compounds of the type 2-8 prepared by us, might possess such structural elements, which would confer them the desired biological activity, i.e., antifungal action. Mention should be made that the crown-ether possesses modest antifungal properties, which were discovered casually by us during a normal screening program for detecting new lead molecules against the three fungi used in the experiments, i.e., A.
niger, A. flavus and C. albicans (Table 3).  From data of Table 3 it can be seen that the ligand 1 possesses modest but net antifungal activity against the three fungi species used in our experiments, whereas the metal complexes of 1 act as much more effective antifungals than the parent ligand, with potencies sometimes comparable to ketoconazole against the Aspergilli, but being much less efficient against Candida as compared to the azole antifungals. The metal ion present in the coordination compound is an important factor for the biological activity, with Cu(II) derivatives more active than Co(ll) complexes, which in turn were more active than the Ni(II) and Cr(III) derivatives. It is also interesting to note that the di-or mononuclear complexes were more active than the corresponding trinuclear derivatives (compare 2 and 3, or 4 and 5, 6 and 7, 8 and 9). The species most susceptible to inhibition by these antifungals was A. niger, followed by A. flavus, whereas C. albicans was the most resistant. In fact, against the first fungus, the Cu(lI) derivative 2 for instance is more effective than ketoconazole, a widely used drug, being only slightly less effective than itraconazole, one of the most potent antifungals known up to now. [17,23,24] In order to test the hypothesis that the compounds reported here act as ergosterol biosynthesis inhibitors, similarly to the azole antifungals, the amounts of ergosterol present in A. niger cultures after treatment with different concentrations of new inhibitors (itraconazole 11, a potent CYP51A1 inhibitor [7'23'-41 has also been included in the study as standard) have been determined by means of a HPLC method (Table 4). t391 These acting as Lanosterol-14-fl-Demethylase Inhibitors data show that at low concentrations of inhibitor, around 80-96 % of ergosterol (as compared to the amount of sterol formed in cultures in which inhibitors have not been added, and which was considered 100%) is still synthesized. By increasing the concentrations of inhibitors used in the experiments, the amount of synthesized ergosterol decreased dose-dependently. A similar effect has been observed for the well-known CYP51A1 inhibitor itraconazole 11 as well as for the new antifungal compound 2 synthesized in the present study (the most active inhibitor in the series). These data allow us to propose a similar mechanism of action for the two classes of antifungal compounds, i.e., the inhibition of lanosterol-14--demethylase, although it is not improbable that our compounds might interfere with (an)other enzyme(s) involved in the ergosterol biosynthetic pathway. Thus, we presume that our original hypothesis that bulky coordination compounds of the type described in the present paper might interfere with the prostethic group of CYP51A1, is correct. Presently it is difficult to make a hypothesis regarding the precise mode of interaction between the inhibitor and the CYP51 A1 active site, but we consider the carbonyl oxygen or the amino group of derivatives 1-9 as the most probable candidates for binding to the Fe(III) ion of this metallo-enzyme. In this way one could also explain why the metal complexes are much more active as compared to the parent ligand 1: thus, 1 possesses a much more flexible molecule as compared to the metal complexes 2-9, in which the presence of the metal ion(s) somehow produces much more rigid structures. Such rigid structures probably fit better and bind tighter to the active site of CYP51A1 as compared to 1, and in consequence the metal complexes act as better antifungals.

Assay of sterols present in thefungi cultures
A reverse-phase HPLC method adapted from the literature, [39] has been used to determine the amount of sterols (ergosterol 12 and lanosterol 13) present in the fungi cultures. HPLC was performed with a Beckman instrument, using a Rheodyne pump and column (reverse phase -Bondapack-C18). The fungi have been cultivated as mentioned above for 5 days, with or without inhibitors added in the nutrient broth. Culture media were suspended in a small volume of MOPS buffer (pH 7.4) and the cells centrifuged at 20,000g for 30 min. Cells were weighed (wet paste) and broken by sonication. Sterols present in the homogenate were then extracted in chloroform, the solvent has been evaporated to a small volume and the extracts applied on a l.t-Bondapak-C18 column, with acetonitrile as eluting solvent. Authentic ergosterol and lanosterol (from Sigma) were used as standards. The flow rate was of 3 mL/min. The retention times were: 8.87 min for ergosterol; and 7.62 min for lanosterol, respectively. Blank assays were done for cultures which were not treated with inhibitors in order to assess the normal levels of sterols present. The amount of ergosterolpresent in the same amount of wet cells from the culture grown in the absence of inhibitor was taken as 100%. I4gJ