www.elsevier.nl/locate/ica Inorganica Chimica Acta 321 (2001) 42 – 48
Transition metal complexes of two new imino-dihydroxamic acids M. Ame´lia Santos a,*, Raquel Grazina a, Margarida Pinto a, Etelka Farkas b a
b
Centro de Quı´mica Estrutural, Instituto Superior Te´cnico, A6. Ro6isco Pais, 1049 -001 Lisbon, Portugal Department of Inorganic and Analytical Chemistry, Uni6ersity of Debrecen, P.O.B. 21, Debrecen, Hungary Received 14 February 2001; accepted 25 May 2001
Abstract Two new iminodihydroxamic acids [N-benzyl-iminobis(propionohydroxamic acid) and N-benzyl-iminobis(butyrohydroxamic acid)] were prepared and studied as specific binders for the transition M2 + ions due to their potential interest as inhibitors of metalloproteinases. Their architecture is based on aliphatic backbones, as spacers connecting two hydroxamate chelating units, with an N-benzyl group inserted in that skeleton to simulate the protein lipophilic subset. Herein, we first report the synthetic procedure that basically involves the formation of the corresponding intermediates with two nitrile groups, which were then converted to the CONHOH moieties. Then, the acid–base and the chelating properties of these ligands towards Cu2 + , Ni2 + and Zn2 + ions, studied by potentiometric and spectrophotometric techniques, are described. Both the ligands form quite stable complexes with these metal ions, presenting a preferential M2 + coordination to the hydroxamate over the amine groups, according to the order Zn2 + ]Ni2 + \Cu2 + . © 2001 Elsevier Science B.V. All rights reserved. Keywords: Copper– hydroxamate; Nickel–hydroxamate; Zinc– hydroxamate; Amino-dihydroxamate; Hydroxamic acid
1. Introduction Hydroxamic acids and amino-hydroxamic acids are known to play an important role in living systems. They have been found as constituents of therapeutics, mostly related with the microbial transport of iron and the iron-overload chelating therapy [1]. More recently, they have also been extensively studied as metal binding ligands in the field of zinc and nickel metalloenzyme inhibition, namely as inhibitors of matrix metalloproteinase (MMP) [2,3] and urease [4,5], respectively. Several of them have even advanced into human clinical trials for the treatment of diseases such as cancer and arthritis. Therefore, studies in solution on Cu2 + , Ni2 + and Zn2 + complex formation with aminohydroxamic [6] and dihydroxamic acids [7] have attracted considerable attention. We have been involved in the development of diaminohydroxamic acids (secondary hydroxamic acids) with different cyclic backbones, mostly addressed for
the siderophore analogy [8,9]. However, in the present work we have developed a set of two linear amino-dihydroxamic acids (with primary hydroxamic acids), aimed at obtaining potential mimetics for enzyme inhibition. The structure of these synthetic compounds is shown in Scheme 1 (H2L1 is a b-aminodihydroxamic acid and H2L2 is a g-amino-dihydroxamic acid). Their architecture is based on the assumption that the backbone skeleton should contain spacers connecting the two hydroxamate-chelating units. The optimisation of the connecting-chain size has to be addressed towards the effectiveness of the M2 + complexation. On the other hand, an amine group was inserted in that spacer backbone to account for good water solubility
* Corresponding author. Tel.: + 351-21-841-9000; fax: + 351-21846-4455. E-mail address:
[email protected] (M. Ame´lia Santos). 0020-1693/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 2 0 - 1 6 9 3 ( 0 1 ) 0 0 5 0 5 - 9
Scheme 1.
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
43
Fig. 1. Potentiometric titration curves (experimental and calculated) for the ligand H2L1 and its metal M2 + complexes (M2 + =Cu2 + , Ni2 + , Zn2 + ). CL/CM =1, CL =3× 10 − 3 M, I =0.1 M (KNO3), T= 25.0 90.1 °C; a= mols of base added per mol of ligand present.
and also its eventual function as a bridging donor group. The presence of such a N-donor group adjacent to the main hydroxamate groups could enable the formation of polynuclear complexes with mixed hydroxamate–amine nitrogen coordination [6]. Furthermore, the introduction of the N-benzyl substituent could simulate the amine interaction with proteins (lipophilic subset) [10]. The present paper reports the synthesis and full characterisation of these two ligands [N-benzyl-iminobis(propionohydroxamic acid), H2L1, and N-benzyliminobis(butyrohydroxamic acid), H2L2], as well as their complexation behaviours towards the first series of transition metal ions (Cu2 + , Ni2 + and Zn2 + ), in aqueous solution, using pH-metric and UV– Vis spectrophotometric techniques.
2. Results and discussion
2.1. Protonation studies The ligands, in their fully protonated form [H3L]+, have three dissociable protons. The log Ki values for equilibrium (1) were obtained from potentiometric measurements (Figs. 1 and 2) with the aid of the SUPERQUAD program [11]. Their stepwise protonation constants are shown in Table 1, together with literature values of some amino/hydroxamate compounds for the purpose of comparison. Hi − 1Li − 3 + H+ ? Hi Li − 2
where i =1, 2 or 3
(1)
The three logarithmic stepwise protonation constants of the ligands (two for the hydroxamate groups and one for the amine group) are in the ranges 6.1– 9.7 and 8.0–9.9 for the H3L1 and H3L2 ligands, respectively (hereafter the charges are omitted for clarity). Because of the similarity between the basicity of hydroxamate and amino groups, the corresponding three protonation processes may overlap each other and the calculated values are just macroconstants, so that their attribution to individual sites is not possible. However, at least for the ligand H3L1, according to the values expected from the chemical evidence [12], the two first-protonation processes should mainly correspond to the protonation of the hydroxamate groups. The ammonium proton of the ligand H3L1 is more acidic than that of the ligand H3L2, probably due to the inductive effect of the hydroxamate group and also to some hydrogen-bond interaction between the hydroxamate protons and the nitrogen atom, thus inhibiting the amine protonation. In fact, such interactions are expected to be higher for the first ligand than for the second one because in H2L1 the interacting groups are closer.
2.2. Metal complexation Potentiometric titrations of Cu(II), Ni(II) and Zn(II) were performed at both 1:1 and 1:2 metal/ligand molar ratios. Representative potentiometric titration curves for the various metal ion–ligand systems with 1:1 stoichiometry are shown in Figs. 1 and 2 for MH3L1 and MH3L2, respectively, together with the titration curves for the free ligands. A comparative analysis of Figs. 1 and 2 indicates considerably different behaviours which
44
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
are associated with differences between the ligands, as well as between the metal ions. The Cu(II) complexation starts at approximately pH 3 for both the ligands, although the Ni(II) and Zn(II) complexation with L2 starts at significantly lower pH (approximately pH 4)
than with L1 (approximately pH 6). Titrations were stopped when precipitation occurred. For the copper– ligand titration, the formation of insoluble material was observed at an earlier stage for L1 (pH\ 3.5) than for L2 (pH\ 4.5). The behaviour of the Cu(II)–H3L1 sys-
Fig. 2. Potentiometric titration curves (experimental and calculated) for the ligand H2L2 and its metal M2 + complexes (M2 + =Cu2 + , Ni2 + , Zn2 + ). CL/CM =1, CL =1.5× 10 − 3 M, I =0.1 M (KNO3), T= 25.0 90.1 °C; a = mols of base added per mol of ligand present.
Table 1 Stepwise protonation constants of the ligands H2L1 and H2L2 (or other relevant analogues), global formation constants and electronic spectral data of the corresponding metal(II) complexes
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
Scheme 2.
tem can probably be attributed to the formation of polynuclear complexes, whereas the behaviour of the Cu(II)–H3L2 system could be rationalised in terms of hydrolytic processes. The titration data obtained before precipitation was fitted using the SUPERQUAD program. The calculated overall formation constants for various species as well as spectral data are presented in Table 1. This table also includes literature data for some model systems [M(II)– b-alaninehydroxamic acid, M(II)– acetohydroxamic acid and M(II)– nonano-dihydroxamic acid] to aid the interpretation of the results. Complexes with stoichiometry MHL were found in all cases but water-soluble ML species were only detected in the Ni(II)– and Zn(II) – H3L1 systems. Comparison between the logarithmic overall formation constants of the corresponding monoprotonated complexes (MHL) of the two ligands is difficult because
45
they differ in proton dependence. Furthermore, the fact that these ligands have different size for the methylenic chain connecting the amine to the hydroxamate groups gives rise to the hypothesis of having different coordination modes around the metal ion. In order to make easier the interpretation of the experimental results, namely in terms of the complex structures and the metal–ligand coordination in the monoprotonated species (MHL), we present a diagram (Scheme 2) containing the four main theoretical hypotheses that can be admitted. The first type (a) has two nitrogen atoms coordinated to the metal ion, one from the amine and the other from the hydroxamate group, whereas the remaining hydroxamate site is protonated. In this case, the formation of a six- and seven-membered chelate ring would be expected for the complexes with the ligands HL1 and HL2, respectively. In a second theoretical hypothesis (b), one hydroxamate site is protonated while the other is involved in a {O,O} coordination around the metal ion, through a five-membered ring chelate. Hypothesis (c) has the amine nitrogen protonated, while the coordination environment of the metal ion involves both the hydroxamate groups (2×{O,O}). Finally, in the fourth hypothesis (d), a mixed coordination {O,O,N} is admitted, with one amine nitrogen and one normal {O,O} hydroxamate bound to the metal ion, while the remaining hydroxamate group is protonated. In an attempt to make a comparison between our experimental results and their interpretation in terms of hypothetical structures, a rough evaluation of the derived stability constants (log K) for the following process can be made: M(II)+ HL ? MHL
(2)
The log K values for Eq. (2) can be calculated from the log i values by subtracting the corresponding pKi: log K= log iMHL − pKi The calculated log K values are shown in Table 2. Although the microconstants are not known, evaluation of the ligand behaviours leads to the following
Table 2 Derived stability constants (log K), corresponding to the equilibrium presented by Eq. (2), roughly calculated on the assumption of each type of hypothetical protonated site in the MHL species Ligand
H2L1 H2L2
Protonated group
one of the hydroxamates amine-N one of the hydroxamates amine-N
log iMHL−pKi Cu(II)
Ni(II)
Zn(II)
19.75−9.66= 10.09 19.75−6.05= 13.70 21.44−9.88=11.56 21.44−7.97= 13.47
15.50−9.66= 5.84 15.50−6.05= 9.45 17.20−9.88= 7.32 17.20−7.97= 9.23
14.68−9.66=5.02 14.68−6.05=8.63 17.53−9.88=7.65 17.53−7.97=9.56
46
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
conclusion: if one of the hydroxamates is protonated in the MHL species, complexes with structure (a), (b) or (d) can be formed, whereas if the dissociable proton is on the amine-N, structure (c) is the most probable. Concerning hypothesis (a) with {N,N} coordination: although it could be theoretically admitted, namely for the Ni(II) and Cu(II) metal ions due to their usual preference for nitrogen donors, it seems excluded by the experimental results. In fact, the spectrophotometric data of the copper complexes in solution (before precipitation) shows a d– d band at about 710 and 728 nm for the assumed CuHL1 and CuHL2 species, which, by comparison with the calculated value, approximately 690 nm (calculated by the ‘rule of the average surrounding’ [13], with values u = 574 nm from the Cu(1,3-diaminopropane)2 for the {N,N} coordination [14], and u approximately 800 nm from aqueous Cu(II) complex) suggests that somehow oxygen coordination should be involved in the copper complexes. On the other hand, the binding of Cu(II) to hydroxamate oxygens is also indicated in the UV region of the absorption spectra (345– 350 nm) since a charge-transfer band at approximately 350 nm was unambiguously assigned to hydroxamate {O,O} coordination in copper(II)–aminohydroxamate complexes [15]. Besides, coordination (a) does not fit the relationship between the log K values roughly calculated by Eq. (2). In fact, the corresponding stability constant should be larger for the complex with L1 than with L2, because they will form six- and seven-membered ring chelates, respectively. For the Ni(II) complexes, spectrophotometric data (Table 1) exclusively support the octahedral geometry in all the Ni(II) complexes, as indicated by their d– d transition bands which are much closer to those reported for the octahedral complex of Ni(II) with acetohydroxamic acid (umax =671, 387 nm) [16], than those for the square planar Ni(II)– aminohydroxamate complexes (ex.: Ni(II)– Alaha, umax =505, 415 nm) [17]. However, the formation of a planar geometry {N,N} bis-chelate complex with L1 could also be expected, as it was found in Ni(II)– (b-alaninehydroxamate)2 species containing two six-membered N,N-chelates [18]. Hypothesis (b), involving just a monohydroxamate coordination, does not fit the results because the corresponding derived constants, for the CuHL, NiHL, and ZnHL2 complexes (10– 11.5, 6– 7 and approximately 7.7, respectively), are much higher than the expected values for a normal monohydroxamate coordination (approximately 8, 5 and 5, respectively), according to the literature log iML values for the corresponding acetohydroxamate complexes [19]. Only the monohydroxamate coordination seems to fit the result of ZnHL1 (log K approximately 5) [20]. However, the fact that ZnH2L1 was not found, in measurable concentration at
approximately pH 6, contradicts structure (b) even in this case. Complexes with bonding mode (c) involve both the hydroxamates in the coordinated form and the amineN in the protonated form. The corresponding log K values in Table 2 are comparable with the stability constants of nonano-dihydroxamic acid 1:1 complexes in all cases. It means that admission of structure (c) is reasonable. However, if complexes with this structure are exclusively formed, the log K values should be higher for the complexes with L2 than with L1. This can be rationalised by the difference between the lengths of chains connecting the hydroxamate moieties in the ligands. In fact, studies on dihydroxamate complexes with different 3d metal ions showed that the maximum stability of complexes occurs when the number of connecting atoms is seven [21]; in our case, that number is five and seven for L1 and L2, respectively. Thus, L1, having a shorter connecting chain, might not allow a complete wrapping of the metal ion with both the hydroxamate groups of the ligand, and a lower stability of complexes formed with L1 be expected. The trend of results obtained for the zinc complexes seems to fit with this assumption but not those of the copper(II) and nickel(II) complexes. Besides, the short connecting chain (not long enough for the formation of monomeric species) in L1 might be the reason for the formation of precipitate in the copper(II)–H3L1 system below pH 4. Therefore, according to the discussion above, neither the structure (a) nor structure (b) are supposed to be relevant in the studied systems. The coordination around the Ni(II) and Cu(II) metal ions for the MHL species seems to involve the hydroxamate groups in a mixed coordination type, (c) and (d). The existence of an apical coordination in structure (d) might account for the higher umax values of Cu(II) complexes than the expected ones [19]. In conclusion, the present study on the complexation behaviour of the imino-b- and imino-g-dihydroxamic acids with the bicharged metal ions Cu2 + , Ni2 + and Zn2 + showed that the hydroxamate groups can compete with the amino groups for the coordination. The effectiveness of that competition seems to increase with the number (n= 2, 3) of the methylenic groups connecting the amino group to the hydroxamate chelating moieties, and also with the type of metal ions (Zn2 + ] Ni2 + \ Cu2 + ). The competition between the different donor sets ({N,N} or {O,O}) seemed to be more determined by steric reasons (number of atoms of the chelating ring) than by the affinity in terms of softness between the donor atoms and the metal ions. The presence of certain mixed coordination modes in the Cu2 + complexes may favour the formation of polynuclear insoluble species.
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
3. Experimental
3.1. Chemicals Analytical grade reagents were used as supplied. Whenever necessary reagents and solvents were dried according to standard methods [22].
3.2. Synthesis of the ligand N-benzyl-3,3 %-iminobis(proionohydroxamic acid), H2L 1 3.2.1. N-Benzyl-3,3 %-iminobis(propiononitrile) To a mixture of 3,3%-iminobis(propiononitrile) (5.0 g, 41 mmol) and potassium carbonate (5.7 g, 41 mmol) in dry dimethylformamide (dmf) (150 ml) at 90 °C, a solution of benzylchloride (8.2 g, 115 mmol) in dry dmf (30 ml) was added dropwise and the reaction mixture was left in reflux under nitrogen for 4 h. After cooling the mixture, the solid residue was filtered off and the solvent of the filtrate was evaporated under vacuum to dryness to give a pale oil. This was purified by column chromatography silica gel 60 [eluent: dichloromethane– ethyl acetate (3:0.1)] to give a pure product as an oil with 62% yield. 1H NMR (CDCl3, l, ppm): 2.46 (t, 4H, CH2N), 2.91 (t, 4H, CH2CN), 3.72 (s, 2H, CH2Ph), 7.32 (m, 5H, Ph). FAB MS; m/z: 214 (M + 1). 3.2.2. N-Benzyl-3,3 %-iminobis(methylpropionate) The N-benzyl-3,3%-iminobis(propiononitrile) (5.3 g, 25 mmol) in dry methanol (80 ml) was saturated with HCL gas for 1 h. The mixture was refluxed for 1 h and then it was left stirring at room temperature under nitrogen atmosphere for 18 h. After filtration to take off the inorganic material, the solvent of the filtrate was evaporated under reduced pressure and the oil residue was purified by column chromatography silica gel 60 [eluent: dichloromethane– methanol (3:2)], to give a pure product as an oil with 20% yield. 1H NMR (CDCl3, l, ppm): 2.47 (t, 4H, CH2N), 2.80 (t, 4H, CH2CN), 3.58 (s, 2H, CH2Ph), 3.65 (s, 6H, COOCH3), 7.32 (m, 5H, Ph). FAB MS; m/z: 280 (M +1). 3.2.3. N-Benzyl-3,3 %-iminobis(propionohydroxamic acid), H2L 1 To a solution of hydroxylamine hydrochloride (2.4 g, 35 mmol) in dry methanol (20 ml) cooled in an ice bath at 0 °C, potassium hydroxide (2.1 g, 35 mmol) was added under nitrogen and the mixture was left stirring for 30 min. The inorganic material of the reaction mixture was then filtered off under nitrogen. Activated A4 molecular sieves were added to the filtrate which was cooled in an ice bath at 0 °C under nitrogen followed by simultaneous dropwise addition of N-benzyl-3,3%-iminobis(methylpropionate) (1.0 g, 3.6 mmol) dissolved in dry methanol (20 ml) and potassium hydroxide (0.6 g, 11 mmol) dissolved in dry methanol (10
47
ml). The mixture was left stirring for 48 h, the inorganic material was filtered off and the filtrate was evaporated in vacuum. The solid obtained was purified by column chromatography silica gel 60 using a gradient of the eluent dichloromethane– methanol to give a pure product as pale yellow crystals with 27% yield. 1H NMR (D2O, l, ppm): 2.27 (t, 4H, CH2N), 2.74 (t, 4H, CH2CN), 3.64 (s, 2H, CH2Ph), 7.39 (m, 5H, Ph). FAB MS; m/z: 282 (M+ 1).
3.3. Synthesis of the ligand N-benzyl-iminobis(butyrohydroxamic acid), H2L 2 3.3.1. N-Benzyl-iminobis(butyronitrile) A mixture of benzylamine (5.4 g, 50 mmol) and potassium carbonate (20.73 g, 150 mmol) in dry dmf and acetonitrile (50:50) previously dried was heated up to 70 °C. Then a solution of bromobutyronitrile (15 g, 100 mmol) in dry acetonitrile (50 ml) was added. The final mixture was refluxed at the mentioned temperature with nitrogen flow for 24 h. After cooling the mixture was filtered and the filtrate was then evaporated in vacuum to give a product as an oil. This was then purified by column chromatography silica gel 60 with petroleum ether–ethyl acetate (3:1.8) as eluent to give the pure product as an oil with 48% yield. 1H RMN (CDCl3, l, ppm): 1.79 (t, 4H, CH2CH2CH2), 2.34 (t, 4H, CH2CN), 2.53 (t, 4H, CH2N), 3.53 (s, 2H, CH2Ph), 7.28 (m, 5H, Ph). FAB MS; m/z: 242 (M+1). 3.3.2. N-Benzyl-iminobis(methylbutyrate) The synthesis is similar to that described above for the previous analogue [N-benzyl-3,3%-iminobis(methylpropionate)]. The product was obtained as an oil with 29% yield. 1H NMR (CDCl3, l, ppm): 1.77 (t, 4H, CH2CH2CH2), 2.33 (t, 4H, CH2CN), 2.42 (t, 4H, CH2N), 3.53 (s, 2H, CH2Ph), 3.64 (s, 6H, COOCH3), 7.28 (m, 5H, Ph). FAB MS; m/z: 308 (M+1). 3.3.3. N-Benzyl-iminobis(butyrohydroxamic acid), H2L 2 This ligand was synthesised by a procedure similar to that used previously for the H2L1. The pure product was obtained as pale yellow crystals with 25% yield. 1H NMR (D2O, l, ppm): 1.79 (t, 4H, CH2CH2CH2), 2.01 (t, 4H, CH2CN), 2.46 (t, 4H, CH2N), 3.66 (s, 2H, CH2Ph), 7.40 (m, 5H, Ph). FAB MS; m/z: 310 (M+1). 3.3.4. Potentiometric measurements The pH potentiometric titrations were conducted at 25.09 0.1 °C with ionic strength (I) 0.1 M (KNO3), using a Crison Digital 517 instrument with an Ingold U1330 glass electrode and an Orion 90-00.11 Ag/AgCl reference electrode. The electrode calibration was carried out daily from the titration of a strong acid (HNO3, 0.1 M) with a strong base (KOH, 0.1 M) at the same ionic strength to assure that we got adequate
48
M. Ame´ lia Santos et al. / Inorganica Chimica Acta 321 (2001) 42–48
responses in the studied pH range and to control the exact concentration of the ligand (Gran’s method) [23]. The potentiometric titrations for the binary systems were performed at 1:1 and 1:2 metal ion– ligand ratios, with ligand concentrations 3×10 − 3 (L1) and 1.5× 10 − 3 M (L2). The ligand was weighted directly to the potentiometric cell and the M(II) metal ions were pipetted from stock solutions: Cu(NO3)2 and Zn(NO3)2 (0.01 M); and Ni(NO3)2 (0.009 M). Calculations from potentiometric data were performed with the SUPERQUAD program [11]. Selection of the equilibrium models was based on a critical evaluation of the leastsquares fitting results, namely analysis of the weighted residuals and the statistical parameters ( 2 and |) [24].
3.3.5. Spectrophotometric measurements All spectra were measured on a Lambda 9 Perkin– Elmer spectrophotometer at 25 °C and at a constant ionic strength (I =0.1 M, KNO3). Solutions of the metal complexes were generated in situ by addition, to the ligand, of a standard metal ion solution (1:1 ligandto-metal ratio): Cu(NO3)2 0.01 M and Ni(NO3)2 0.009 M. The pH measurements were carried out using a 420A Orion pH-meter, equipped with an Orion 91-03 glass calomel combination electrode. 3.3.6. Other measurements 1 H NMR spectra were recorded on a Varian Unity 300 spectrometer at 25 °C. Chemical shifts are reported in ppm (l) from internal references (tetramethylsilane (TMS) in CDCl3 solutions and sodium 3-(trimethylsilyl)-[2,2,3,3-2H4]propionate) (DSS) in D2O solutions. The following abbreviations are used: s=singlet; t= triplet; m=multiplet. Mass spectra were performed in a VG TRIO-2000 GC/MS instrument.
Acknowledgements The authors thank the Portuguese Fundac¸ a˜ o para a Cieˆ ncia e Tecnologia (FCT) (project Praxis/PCEX/ QUI/85/96), the Hungarian Scientific Research Fund (OTKA T034674) and the COST D8/0010 program for financial support.
.
References [1] J.B. Neilands, J. Biol. Chem. 270 (1995) 2672. [2] M. Arnold, D.A. Brown, O. Deeg, W. Errington, W. Haase, K. Herlihy, T.J. Kemp, H. Nimir, R. Werner, Inorg. Chem. 37 (1998) 2920. [3] M.A. Pearson, L.O. Michel, R.P. Hausinger, P. Andrew Karplus, Biochemistry 36 (1997) 8164. [4] M.H. Parker, E.A. Lunney, D.F. Ortwine, A.G. Pavlovsky, C. Humblet, C.G. Brouillette, Biochemistry 38 (1999) 13592. [5] S. Toba, K.V. Damodaran, K.M. Merz Jr., J. Med. Chem. 42 (1999) 1225. [6] B. Kurzak, H. Kozlowsky, E. Farkas, Coord. Chem. Rev. 114 (1992) 169. [7] E. Farkas, E.A. Enyedy, H. Cso´ ka, Polyhedron 18 (1999) 2391. [8] M.A. Santos, M. Gaspar, M.T. Amorim, Inorg. Chim. Acta 278 (1998) 51 (see also 284 (1999) 20). [9] M. Gaspar, M.A. Santos, K. Krauter, G. Winkelmann, BioMetals 12 (1999) 209. [10] A. Patel, H.J. Smith, J. Sturzebecher, Design of enzyme inhibitors as drugs, in: H.J. Smith (Ed.), Drug Design and Action, 3rd ed., Harwood Academic, Amsterdam, 1998, pp. 283 – 329. [11] P. Gans, A. Sabatini, A. Vacca, J. Chem. Soc., Dalton Trans. (1985) 1195. [12] M.A. Santos, M.A. Esteves, M.C. Vaz, J.J.R. Frau´ sto da Silva, B. Noszal, E. Farkas, J. Chem. Soc., Perkin Trans. 2 (1997) 1977. [13] (a) E.J. Billo, Inorg. Nucl. Chem. Lett. 10 (1974) 613; (b) A.A. Kurganov, V.A. Davankov, Inorg. Nucl. Chem. Lett. 12 (1976) 746. [14] A. Gergely, I. So´ va´ go´ , Inorg. Chim. Acta 20 (1976) 19. [15] E. Farkas, J. Szoke, T. Kiss, H. Kozlowski, W. Bal, J. Chem. Soc., Dalton Trans. (1989) 2247. [16] E. Farkas, D. Brown, R. Cittaro, W.H. Glass, J. Chem. Soc., Dalton Trans. (1999) 2803. [17] B. Kurzak, W. Bal, H. Kozlowski, J. Inorg. Biochem. 38 (1990) 9. [18] B. Kurzak, E. Farkas, J. Coord. Chem. 28 (1993) 203. [19] E. Farkas, E.A. Enyedy, G. Micera, E. Garribba, Polyhedron 19 (2000) 1727. [20] E. Farkas, E. Kozma, M. Petho, K.M. Herlihy, G. Micera, Polyhedron 17 (1998) 3331. [21] A. Evers, R.D. Hancock, A.E. Martell, R.J. Motekaitis, Inorg. Chem. 28 (1989) 2189. [22] W.L.F. Armarego, D.D. Perring, Purification of Laboratory Chemicals, 4th ed., Butterworth-Heinemann Press, Oxford, 1999. [23] G. Gran, Analyst 77 (1952) 661. [24] A. Sabatini, A. Vaca, P. Gans, Talanta 21 (1974) 53.
