Cyanonitrosylmetallates as potential NO-donors

Journal of Inorganic Biochemistry 69 (1998) 121±127

Cyanonitrosylmetallates as potential NO-donors Janusz Oszajca a, Gra_zyna Stochel a, Ewa Wasielewska a, Zo®a Stasicka a,*, Ryszard J. Gryglewski b, Andrzej Jakubowski b, Katarzyna Cieslik b a

b

Department of Inorganic Chemistry, Jagiellonian University, Ingardena 3, 30-060 Krak ow, Poland Department of Pharmacology, Medical College of the Jagiellonian University, Grzeg orzecka 16, 31-531 Krak ow, Poland Received 25 August 1997; received in revised form 20 October 1997; accepted 23 October 1997

Abstract The [M(CN)x NOy ]nÿ complexes (where M ˆ Cr(I), Mn(I), Mn(II), Fe(I), Fe(II), Fe(III)) were studied as potential NO-donors using both pharmacological and theoretical semi-empirical methods. Only iron complexes appeared to be pharmacologically active. The quantum chemical calculations indicated that these complexes have the highest predisposition to undergo a nucleophilic attack followed by the NO‡ release. The results allowed us to interpret the metabolism of the [M(CN)x NOy ]nÿ complexes in terms of the NO‡ -donation. Ó 1998 Elsevier Science Inc. All rights reserved.

1. Introduction The discovery of the essential role of endogenous nitric oxide (NO) in controlling vascular tone [1] triggered many studies on physiological and pathophysiological functions of NO [1±8]. The biological source of endogenous NO is the conversion of L -arginine to L -citrulline and NO, the reaction which is catalyzed by a family of homodimeric dioxygenases named NO-synthases (NOS, EC 1.14.13.39) [4,9±12]. One of the feasible sources of exogenous NO are so called ``nitrovasodilators'', i.e. drugs which release NO, and therefore exert anti-anginal and hypotensive actions. Molecules of NO-donors accommodate hidden NO moiety in various casts. The best known are groups of organic nitrates e.g. glyceryl trinitrate (GTN), sydnonimines e.g. molsidomine, and pentacyanonitrosylmetallates e.g. sodium nitroprusside (SNP). Although GTN and SNP have long been applied therapeutically, the mode of release of NO from their molecules is still poorly understood [5,6,13±24]. Especially important is this in the case of SNP, frequently used ``reference nitrovasodilator'' [14,25]. In a plain bu€er of pH 7, relatively small amounts of NO are released from SNP [14]. One of us [25] suggested that in the pharmacological action of SNP intracellular nucleophilic components may facilitate transfer of NO *

Corresponding author. Tel.: +48 12 6335392; fax: +48 12 6335392; e-mail: [email protected]. 0162-0134/98/$19.00 Ó 1998 Elsevier Science Inc. All rights reserved. PII: S 0 1 6 2 - 0 1 3 4 ( 9 7 ) 1 0 0 3 2 - 0

from SNP to the haeme centre of cytosolic guanylate cyclase. Moreover, unlike GTN and other organic NO-donors, the reaction of nitroprusside is augmented by methylene blue and pyocyanin which otherwise are inhibitors of guanylate cyclase [25]. To elucidate the NO-donation mechanism, the in vitro pharmacological activity of nitroprusside was compared with those of other pentacyanonitrosylmetallates of the ®rst transition metal series. The [M(CN)5 NO]nÿ family was chosen because, despite the apparent similarity in stereochemistry and bonding, the complexes demonstrate a variety of redox and nucleophilic or electrophilic properties, which can mimic various biochemical behaviours of NO [13,26±30]. The di€erences in their reactivities were interpreted basing on the electronic structure determined by a semi-empirical INDO-type method (SINDO) [31,32].

2. Materials and methods 2.1. Chemicals Most of the complexes used in the work were synthesized and puri®ed using literature methods: K2 [Mn(CN)5 NO]á2H2 O [33], K3 [Mn(CN)5 NO]á2H2 O [33], K3 [Cr(CN)5 NO]á2H2 O [34], K3 [Fe(CN)5 NO2 ]á 2H2 O [35]. Purity of the samples was checked by chemical analysis and their IR and UV/Visible spectra. The [Fe(CN)4 NO]2ÿ complex was obtained in situ in reaction

122

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127

between [Fe(CN)5 NO]2ÿ (0.002 M) and ascorbic acid (0.02 M) or Na2 S2 O4 (0.02 M). The synthesis was performed under Ar ¯ow and the product was stored under inert atmosphere (Ar). The reaction progress was followed and purity of the solution was veri®ed by UV/Visible, and for Cr(I), Mn(II) and Fe(I) complexes, also by EPR spectroscopy. The reaction product was kept under inert atmosphere and was diluted to the chosen concentration just before experiments (pH was kept at 7.4). Other reagents, of analytical grade, were purchased from Sigma, (Poole, Dorset, UK), Merck (Darmstadt, Germany) and POCh (Gliwice, Poland). 2.2. Organ bath experiments New Zealand rabbit of either sex (2±2.5 kg) were anaesthetized with an overdose of pentobarbital and their thoracic aortas were excised, cut into rings 2 mm wide, suspended between stainless-steel hooks and mounted in a 5-ml organ bath ®lled with warmed (37°C) and oxygenated (95% O2 , 5% CO2 ) Krebs bu€er (pH ˆ 7.6) containing 5.6 lM indomethacin. The Krebs bu€er had the following composition: 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH2 PO4 , 1.17 mM MgSO4 , 2.25 mM CaCl2 , 3.25 mM NaHCO3 and 5.6 mM glucose. Denuded rings were prepared by gently rotating the rings on a steel wire before mounting. Isometric force was measured with Biogestab K 30 type 351 transducers (Hugo Sachs Electronic, Heidelberg Germany). A tension of 4 g was applied and the rings were equilibrated (60 min) adjusting the preload to 4 g every 15 min. After a stable base line was obtained (ca. 90 min), the rings were contracted with KCl (60 lM) followed by a 30 min equilibration in the fresh Krebs bu€er. The rings were then contracted with phenylephrine (0.2 lM) typically producing 70±80% of maximal contraction, and acetylcholine (ACh) (0.3 lM) was added to the baths to test the functional integrity of endothelium. ACh did not change the tone of the endothelium-denuded rings. The [M(CN)x NOy ]nÿ complex solution in saline was added to the bath in portions to get cumulative concentration±response curve (from 30 nM to 10 lM). The [M(CN)x NOy ]nÿ responses were expressed as the percentage change in the phenylephrine-induced tone. IC50 value was calculated by means of linear regression analysis plotting log concentration against percent of inhibition or percent of relaxation e€ect. IC50 denotes the inhibitor concentration corresponding to 50% inhibition. Solutions containing only potassium cyanide (3 ´ 10ÿ8 ) 1 ´ 10ÿ5 M) appeared to be inactive. 2.3. Preparation of the human platelet rich plasma Venous blood of healthy men aged 25±45 years was used from Blood Donation Centre of the National Clinical Hospital in Krakow (Poland). Blood was collected in plastic test tubes containing 3.15% sodium citrate (1 : 9 v/v). Platelet rich plasma (PRP) was prepared by centrifugation (800 g for 20 min at 4°C).

2.4. Platelet aggregation Platelet aggregation was performed in 0.5 ml samples of PRP in Payton Dual Channel aggregometer calibrated with PRP (for which optical density, OD ˆ 100%), and water (OD ˆ 0%). Collagen (0.5±5 lg/ml) was used as a proaggregatory agent. PRP was preincubated in aggregometer for 3 min at 37°C with stirring at 1100 rpm. Then collagen was added with a Hamilton syringe and the aggregation curves were recorded up to 15 min. The threshold concentration of collagen was assessed separately for each blood sample. Near maximal concentrations of collagen were de®ned as those which caused 80±90% change in OD. A solution of the [M(CN)x NOy ]nÿ complex (of initial concentration varied from 50 lM to 50 mM) was added to a test tube before adding PRP (®nal complex concentration ccompl was within the range 10 nM±3 mM). To minimize the error caused by the changes in susceptibility of platelets to the proaggregatory compounds, the blank aggregation curves (without any [M(CN)x NOy ]nÿ complex) were made before and after the experiment. Relative inhibition of platelet aggregation (IPA) was then calculated [36]: X ÿY IPA…%† ˆ 100%; …1† X where X and Y denote the decrease in OD of collagentreated PRP in the absence and presence of a complex, respectively. From the correlation between IPA and the complex concentration, the IC50 parameter was determined as the concentration needed to reach IPA ˆ 50%. 2.5. Irradiation Full light of a HBO 200 (NARVA, Germany) mercury lamp was used to irradiate 5 ´ 10ÿ2 M complex solutions in the Krebs bu€er at pH 7.4, placed in a 2 mm rectangular quartz cell (lamp housing K0001 Cobrabid, Poland, power supply K0-400-01, Kabid, Poland). The solutions irradiated 15 min just before the biochemical experiments, were immediately diluted with the Krebs bu€er to obtain the same total complex concentrations as in the experiments with unirradiated complexes. Photoconversion was calculated from changes in the UV/ Visible spectra and for NP was assessed as about 25%. Physiological tests were made according to identical procedure for irradiated and nonirradiated solutions. 2.6. Method of calculation The electronic structures were obtained by SINDO method. The details of the method are described elsewhere (see for example Refs. [31,32] and references therein). We give only a quick overview here. The SINDO method is a scaled version of the semiempirical INDO method in which the integrals depend on the actual charge and con®guration of the constituent atoms. All integrals referring to the reference con®g-

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127

uration (ground state con®guration of the neutral atom) are calculated with the use of approximate Hartree± Fock orbitals. The charge and con®guration e€ect, on the other hand, are always estimated with the use of Slater-type orbitals with Burn's rules. Two-centre repulsion integrals are approximated using Ohno's formula. As long as the electronic structure is concerned there are no parameters of molecular origin in the treatment. For the core±core repulsion terms, which are calculated by the MINDO/2 procedure, the molecular parameters are estimated from experimental bond energies and bond lengths. The electronic structure is analyzed using Mulliken population analysis [31,32]. 3. Results For the pharmacological studies besides nitroprusside, the [M(CN)5 NO]nÿ complexes of Cr(I), Mn(I), Mn(II), Fe(I), and Fe(III) were chosen because of their similar structures and divergent reactivities [29,30,37± 39]. Amongst the complexes, those of Fe(I) and Fe(III) are involved in the pH-dependent equilibria: the Fe(I) compound undergoes a dissociation [40±42], 3ÿ

2ÿ

ÿ

‰Fe…CN†5 NOŠ ¢ ‰Fe…CN†4 NOŠ ‡ CN …2† and the iron(III) complex is converted into the nitroform [27,35,43] ÿ

‰Fe…CN†5 NOŠ ‡ 2OHÿ ¢ ‰Fe…CN†5 NO2 Š

3ÿ

‡ H2 O:

…3† At the physiological temperature and pH the [Fe(CN)4 NO]2ÿ form contributes to equilibrium (2) in 70±80% [21,41], whereas equilibrium (3) is almost completely shifted to the right [35,43]. This was the reason for including the [Fe(CN)4 NO]2ÿ and [Fe(CN)5 NO2 ]3ÿ ions into the studied family of complexes, reformulated now as [M(CN)x NOy ]nÿ . Relaxation of rabbit aortic strips (RbA) by the [M(CN)x NOy ]nÿ ions was performed in the dark and after irradiation of the compounds. In the dark only the

123

iron complexes showed well expressed vasorelaxant activity (IC50 0.1 lM), whereas the complexes of chromium and manganese were found to have very low pharmacological activity (IC50  10 lM, Table 1 and Fig. 1(a)). The same was true for the anti-aggregatory potencies of the complexes. Only the iron complexes inhibited blood platelet aggregation, although at concentrations 10 times higher than those required for the vasorelaxant e€ect (Table 1). The chromium and manganese complexes were practically inactive as anti-aggregatory agents. Irradiation diminished vasorelaxant and anti-aggregatory potencies of the Fe(II) and Fe(III) complexes as a result of their photodecomposition and generation of other complexes and/or NO-forms (cf. Eqs. (7a), (7b), (8a), (8b). The e€ect of irradiation on the vasorelaxant properties of the chromium and manganese complexes was at the limit of detection (Table 1 and Fig. 1(b)). In the dark both the investigated pharmacological properties of [M(CN)x NOy ]nÿ decreased in the order: 2ÿ

‰Fe…CN†5 NOŠ

 ‰Fe…CN†5 NO2 Š

3ÿ

 ‰Fe…CN†4 NOŠ2ÿ  ‰Mn…CN†5 NOŠ2ÿ  ‰Mn…CN†5 NOŠ3ÿ  ‰Cr…CN†5 NOŠ3ÿ ;

whereas the activity series of the irradiated solutions was only little modi®ed (cf. Fig. 1). The results showed that the iron complexes play a special role in the process of NO-donation. To interpret this ®nding we ®rst analyzed the M±NO bond characteristics within the [M(CN)x NOy ]nÿ family (Table 2). The analysis included the distances between the atoms (RNO and RM±NO ), the stretching vibrations (mNO ) and two-centre contributions to the total energies (ENO and EM±N…NO† ). The calculated parameters located the iron complexes among those of the strongest bonds between the metal and nitrogen donor atom of the NO group. Despite the strong M±NO bond, nitroprusside is well known for its tendency to release the NO-ligand as a result of a nucleophilic attack [27±29]. The reaction pathway implies the importance of energy and characteristics

Table 1 Relaxation of RbA (IC50 in lM) and anti-platelet activity (IC50 in lM) of unirradiated and irradiated [M(CN)x NOy ]nÿ solutions [M(CN)x NOy ]nÿ

[Cr(CN)5 NO]3ÿ [Mn(CN)5 NO]2ÿ [Mn(CN)5 NO]3ÿ [Fe(CN)5 NO2 ]3ÿ [Fe(CN)5 NO]2ÿ [Fe(CN)4 NO]2ÿ NOÿ 2 a b

…4†

Unirradiated

Irradiated

Relaxation

Inhibition of aggregation

Relaxation

Inhibition of aggregation

IC50; lM

IC50 , lM

IC50; lM

IC50 , lM

10 10 10 0.16 ‹ 0.1 a 0.10 ‹ 0.07 a [0.22 b >10

>3000 700 ‹ 300 a >3000 1.4 ‹ 0.9 a 1.4 ‹ 0.7 a

>10 >10 >10 5.3 ‹ 3.0 1.3 ‹ 1.0

>3000 1000 ‹ 500 1200 ‹ 500 1.8 ‹ 1.0 a 3.3 ‹ 2.5 a

a a

>3000

Mean ‹ standard deviation. Due to partial oxidation of the Fe(I) complex under experimental conditions, its IC50 value could only be roughly estimated.

a a

124

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127 Table 3 Characteristics of the LUMO (8e or 9e) orbitals in the [M(CN)x NOy ]nÿ complexes calculated by SINDO method (partly adapted from Refs. [29,50]) Complex

Energy (eV)

qN (NO) % NO Dpp (M±NO)

[CrI (CN)5 NO]3ÿ [MnII (CN)5 NO]2ÿ [MnI (CN)5 NO]3ÿ [FeIII (CN)5 NO]ÿ [FeII (CN)5 NO]2ÿ [FeI (CN)5 NO]3ÿb [FeI (CN)4 NO]2ÿ b [FeIII (CN)5 NO2 ]3ÿ

+5.4 +0.1 +5.0 )4.8 )0.2 +3.2 )0.4 +2.5

0.18 0.33 0.19 0.54 0.50 0.28 0.44

a b c

c c

c

17 36 26 40 44 48 30

c

)0.31 )0.58 )0.36 )0.67 )0.63 )0.61 )0.54

a

c

Changes in M±NO p bond order resulting from a nucleophilic attack. In this case a SOMO 9e orbital was taken into consideration. Results of this study.

is reasonably high when the LUMO energy is low and qN (NO) is adequately positive. Both these factors point to pentacyanonitrosylferrates(II) and (III) as the most prone to undergo the attack (Table 3). From the two Fe(I) complexes, i.e. [Fe(CN)5 NO]3ÿ and [Fe(CN)4 NO]2ÿ , only the latter is expected to undergo a nucleophilic attack. This theoretical prediction is consistent with behaviour of the complexes in alkaline medium [13,24]. The results are in agreement with the previously found correlation between the LUMO characteristics and electrophilic properties of the [M(CN)5 NO]nÿ family [29,30,51] Cr…0† < Cr…I† < Mn…I† < Fe…I† Fig. 1. Dependence of relaxation of superfused RbA on cumulative concentration (log Ccompl ) of the [M(CN)x NOy ]nÿ ions: (h) ± [Cr(CN)5 NO]3ÿ , (D) ± [Mn(CN)5 NO]3ÿ , (s) ± [Mn(CN)5 NO]2ÿ , (¨) ± [Fe(CN)5 NO]2ÿ , (+) ± [Fe(CN)5 NO2 ]3ÿ in unirradiated (a) and irradiated (b) solutions.

of the lowest unoccupied orbitals. The predisposition of the [M(CN)x NOy ]nÿ to undergo a nucleophilic attack depends on the LUMO energy and net charge on the N(NO) atom, which is the attack site, i.e. the tendency

6 Cr…II† < Mn…II† < Fe…II† < Fe…III†: …5† The irregularity shows only the Mn(II) complex which should behave similarly to the Fe(II) and Fe(III) compounds. However, the pentacyanonitrosylmanganate(II) is too strong oxidant (E ˆ 0.66 V [52,53]) to survive in the cellular reducing milieu. To release the NO±ligand, the nucleophilic attack had to be followed by the e€ective M±NO bond breaking. The process is dependent on contribution of the NO orbitals to the LUMO (%NO) and on the e€ect of the nucleophilic attack on the M±NO bond order (Dpp

Table 2 Characteristics of the M±N±O bonds in [M(CN)x NO]nÿ

[CrI (CN)5 NO]3ÿ [MnII (CN)5 NO]2ÿ [MnI (CN)5 NO]3ÿ [FeIII (CN)5 NO]ÿ [FeII (CN)5 NO]2ÿ [FeI (CN)5 NO]3ÿ [FeI (CN)4 NO]2ÿ a b

RNO (A)

mNO (cmÿ1 )

ENO (eV)

1.21 [44]

1682 [45,46] 1880 [45,46] 1730 [45,46]

)20.55 [37] )23.45 [37] )20.52 [37] )26.81 b )25.26 b )20.87 b )23.70 b

1.21 [47] 1.13 [48] 1.16 [49]

1937 [45,46,48] 1580 [49] 1755 [49]

a

EAB ± two-atomic contributions to the total energy calculated by SINDO method [31,32]. This work.

a

RM±NO (A)

EM±N…NO†

1.71 [44]

)52.68 [37] )54.08 [37] )53.62 [37] )55.62 b )55.16 b )54.15 b )57.80 b

1.66 [47] 1.63 [48] 1.56 [49]

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127

125

(M±NO)): the probability of the NO-release increases with an increase in the NO-contribution and a decrease in the M±NO p-bond order [29,30,51]. Both these parameters change approximately in the same manner following the sequence (5) thus demonstrating that the NO-release is the main chemical consequence of the nucleophilic attack in the studied systems. This conclusion is supported by positive changes in the M±CN p-bond order, which indicates that the M±CN bonds, both in equatorial and axial positions, become even stronger as a result of the attack [29]. Consequently, the sequence (5) may also be treated as the NO-donation series [30,51]. 4. Discussion The pharmacological behaviour of the pentacyanonitrosyl complexes (Table 1) is consistent with the NOdonation series (5). Amongst the Cr(I), Mn(I), Mn(II) and Fe(II) complexes only the last one demonstrates the genuine pharmacological activity. The characteristics of iron(I) cyanonitrosyl complexes strongly depend on the coordination number i.e. this is the [Fe(CN)4 NO]2ÿ , and not [Fe(CN)5 NO]3ÿ complex, that demonstrates the NO-donor properties, which, however, are somewhat weaker than those of SNP (Table 1). The role of [FeIII (CN)5 NO2 ]3ÿ needs a more detailed discussion: the results revealed its physiological e€ectiveness (Table 1), whereas its high LUMO energy (Table 3) provides an argument against the nucleophilic attack. The NO-donation can originate either from [Fe(CN)5 NO]ÿ complex, related to [Fe(CN)5 NO2 ]3ÿ by the equilibrium (3), or, more probably, from [Fe(CN)5 NO]2ÿ generated by reduction and profanation processes (Eq. (6)). 3ÿ e

ÿ

‰Fe…CN †5 NO2 Š !‰Fe…CN †5 NO2 Š

4ÿ H2 O;K

2ÿ

¢ ‰Fe…CN †5 NOŠ ‡ 2OHÿ :

…6†

The cellular reducers (ascorbic acid and thiols) were found [21] to reduce [FeII (CN)5 NO2 ]3ÿ to [FeII (CN)5 NO2 ]4ÿ , which in physiological pH is transformed immediately to nitroprusside (pK ˆ 6 [13]). Thus, the NO-donation of the [Fe(CN)5 NO2 ]3ÿ complex, preceded by its reduction, should follow the same metabolism as that of the [Fe(CN)5 NO]2ÿ complex injected directly. In the metabolism of SNP initiated by a nucleophilic attack (Fig. 2) the fate of the intermediate complex, [Fe(CN)5 (NONucl)]3ÿ had to be considered. This complex is nearly always unstable [24] and produces either Fe(I) compounds and an oxidized nucleophile or/and Fe(II) complexes and NONucl species [5,13,17,23,28]. The latter may be a source of NO-species in both forms NO and NO‡ . The above conclusion is supported by a decrease in the physiological activity of the iron(II) and (III) complexes after irradiation. The photochemical reactions are reported to generate mostly the NO-species other

Fig. 2. The suggested pathways of SNP metabolism.

than NO‡ [35,39,51,54]: in innersphere photoredox processes the complexes generate NO or NO2 molecules (Eqs. (7a) and (8a)), ‰Fe…CN†5 NOŠ2ÿ ! ‰Fe…CN†5 H2 OŠ2ÿ ‡ NO;

…7a†

‰Fe…CN†5 NOŠ2ÿ ! ‰Fe…CN†5 NOŠ3ÿ ‡ Solv‡ ;

…7b†

3ÿ

‰Fe…CN†5 NO2 Š

! ‰Fe…CN†5 H2 OŠ

3ÿ

3ÿ

‡ NO2 ;

…8a†

2ÿ

‰Fe…CN†5 NO2 Š ! ‰Fe…CN†5 H2 OŠ ‡ NOÿ …8b† 2; in other photochemical reactions (Eqs. (7b) and (8b)) nitroprusside undergoes outersphere photoreduction to [Fe(CN)5 NO]3ÿ (Eq. (7b)), whereas the Fe(III) complex generates NOÿ 2 anions in the photosubstitution mode (Eq. (8b)). The much lower vasorelaxant activity of NOÿ 2 ion as compared to that of SNP was reported in the literature [24,56] and con®rmed by our studies (Ta‡ ble 1). Both NOÿ 2 and SNP can be treated as NO -donors but in physiological conditions there is an essential di€erence in their NO‡ -donating capacities. [Fe(CN)5 NO]2ÿ is a very good NO‡ -donor at pH 7.4, ‡ whereas NOÿ 2 is a good NO donor only in an acidic solutions according to the equilibrium: ‡ ÿ …9† H‡ ‡ NOÿ 2 ¡ HNO2 ¡ NO ‡ OH : Another pathway of the [Fe(CN)5 NO]2ÿ metabolism presented recently [6,13,15,26], includes its reduction in cells or tissues as the ®rst step, followed by the release of one of the CN-ligands and formation of [Fe(CN)4 NO]2ÿ , which was claimed to release spontaneously an NO molecule [13], eÿ

ÿCNÿ

‰Fe…CN†5 NOŠ2ÿ !‰Fe…CN†5 NOŠ3ÿ ! ‰Fe…CN†4 NOŠ2ÿ 4ÿ

!! NO ‡ ‰Fe…CN†6 Š : …10† Deoxyhemoglobin [13,19], ascorbic acid [13,20] and thiols [13,17] were considered as cellular reducers.

126

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127

The ®rst two steps of reaction (10) are well documented [13,24,42], whereas the spontaneous loss of an NO molecule seems to be very unlikely taking into account the electronic structure of the [Fe(CN)4 NO]2ÿ ion and the character of its ground state (2 A1 ) [57]. Also other arguments both of structural and kinetic nature are against this conception. X-ray analysis of (Ph4 P)2 [Fe(CN)4 NO] demonstrated the square pyramidal geometry of the anion [49]. This structure and ESR parameters [13,40,47,58±61] support the formulation FeI ±NO‡ for the anion rather than FeII ±NO. Also theoretical calculations by SINDO method point to NO‡ character of the nitrosyl ligand in [Fe(CN)4 NO]2ÿ complex (Table 3). The charge of the NO group in [Fe(CN)4 NO]2ÿ complex is only slightly smaller then that in the nitroprusside. The potential energy curve for [Fe(CN)4 NO]2ÿ is extremely ¯at, suggesting a small barrier to bending [57], and the characteristics of the LUMO orbitals are similar to those of nitroprusside (Table 3), which is why the tendency to undergo a nucleophilic attack is expected. The complex is really unstable in alkaline media [42], which is consistent with the conception that a nucleophilic attack is the ®rst step of pharmacological activity of [Fe(CN)4 NO]2ÿ . Both the pathways of the SNP metabolism are summarized in Fig. 2. If the metabolism of SNP starts from reduction, the key step is nucleophilic attack on the [Fe(CN)4 NO]2ÿ complex. It seems, however, that more important than the pathway induced by reduction, is the pathway initiated by the direct nucleophilic attack. This conclusion is supported both by an extremely high rate of the reaction between SNP and some cellular thiolates [13,23,24] and by the NO-donation capacity of [Fe(CN)4 NO]2ÿ somewhat weaker than that of SNP (Table 1). Moreover, the mechanistic interpretation allows us to elucidate the e€ect of methylene blue or pyocyanin on the SNP pharmacological behaviour reported earlier [25]. These two compounds are suciently strong oxidants [62,63] to prevent the reduction of SNP, which is why they do increase its pharmacological e€ects. There is much current interest in the design of metal complexes as drugs and diagnostic agents and in understanding the molecular mechanism of action of metallopharmaceuticals being already in clinical use. The attempts in establishing the metabolism of SNP and molecular nature of its pharmacological activity, exemplify well the problem. Although di€erent approaches have been applied to solve the question [5,6,13±24], the relation between electronic structure of cyanonitrosylmetallates (including nitroprusside) and their biological activity has not been reported as yet. The present data show that only those iron complexes, which are prone to undergo a nucleophilic attack, can act as potent NOdonors in blood vessels and somewhat weaker blood platelet-suppressants. The nucleophilic attack, followed by NO‡ donation, plays a crucial role in the biotransformation of the iron complexes. In contrast to hydrophobic NO , which is a dominating biological form of both endogenous NO and nitrate-derived NO, hydrophilic NO‡ -species do not easily pass biomembranes. Conse-

quently, the NO‡ -compounds released intracellularly [25] stay there ``imprisoned'' and react preferentially with intracellular thiols [3]. The dominant route of pentacyanonitrosylferrates transformation to NO‡ proposed in this study, constitutes the major di€erence between pharmacology of SNP and organic nitrates [25].

Acknowledgements Financial support from KBN, grants Nos 0976/P-3/ 93/04 and 147/T09/95/09 and from COST/PECO grant No D3/0005/93 is highly acknowledged. References [1] S. Moncada, R.M.J. Palmer, E.A. Higgs, Pharmacol. Rev. 43 (1991) 109. [2] Y. Henry, C. Ducrocq, J.C. Drapier, D. Servent, C. Pellat, A. Guissani, Eur. Biophys. J. 20 (1991) 1. [3] J.S. Stamler, D.J. Singel, J. Loscalzo, Science 258 (1992) 1898. [4] P.L. Feldman, O.W. Grith, D.J. Stuehr, Chem. Eng. News 71 (1993) 26. [5] A.R. Butler, D. Lyn, H. Williams, Chem. Soc. Rev. 22 (1993) 233. [6] K.D. Kr oncke, K. Fehsel, V. Kolb-Bachofen, Biol. Chem. Hoppe-Seyler 376 (1995) 327. [7] R.J. Gryglewski, R.M. Botting, J.R. Vane, Hypertension 12 (1988) 530. [8] R.J. Gryglewski, Thromb. Haemostas 19 (1993) 158. [9] H.G. Korth, R. Sustmann, C. Thater, A.R. Butler, K.V. Ingold, J. Biol. Chem. 269 (1994) 17776. [10] P.L. Feldman, D.J. Stuehr, O.W. Grith, J.M. Fukuto, in: B.A. Weissman, N. Allon, S. Sepira (Eds.), Biochemical, Pharmalogical and Clinical Aspects of Nitric Oxide, Plenum Press, New York, 1995, pp. 13±20. [11] B. Mayer, in: B.A. Weissman, N. Allon, S. Sepira (Eds.), Biochemical, Pharmalogical and Clinical Aspects of Nitric Oxide, Plenum Press, New York, 1995, pp. 37±48. [12] E. Wasielewska, M. Witko, G. Stochel, Z. Stasicka, Eur. J. Chem. 3 (1997) 609. [13] A.R. Butler, C. Glidewell, Chem. Soc. Rev. 16 (1987) 361. [14] M. Feelisch, E. Noack, Eur. J. Pharmacol. 142 (1987) 465. [15] J.N. Bates, M.T. Baker, R. Guerra, D.G. Harrison, Biochem. Pharmacol. 42 (1991) 157. [16] F.W. Flitney, G. Kennovin, J. Physiol. 392 (1987) 43. [17] A.R. Butler, A.M. Calsy-Harrison, C. Glidewell, P.E. S orensen, Polyhedron 7 (1988) 1197. [18] J. Mc Aninly, D.L.H. Williams, S.C. Askew, A.R. Butler, C. Russell, J. Chem. Soc. Chem. Commun. 23 (1993) 1758. [19] A.R. Butler, C. Glidewell, I.L. Johnson, A.S. Mc Intosh, Inorg. Chim. Acta. 138 (1987) 159. [20] P.A. Croven, F.R. De Rubertis, J. Biol. Chem. 235 (1978) 8433. [21] G. Stochel, J. Oszajca, A. Herdegen, E. Wasielewska, Z. Stasicka, 12th Summer School of Coordination Compounds, Book of Abstracts, 1993, p. 146. [22] G. Stochel, G. Stopa, Z. Stasicka, Bull. Acad. Sci. Polon. 42 (1994) 476. [23] M.D. Johnson, R.G. Wilkins, Inorg. Chem. 23 (1984) 231. [24] K. Szacilowski, G. Stochel, H. Kisch, Z. Stasicka, New J. Chem. 21 (1997) 893. [25] R.J. Gryglewski, A. Zembowicz, D. Salvemini, G.W. Taylor, and J.R. Vane, Br. J. Pharmacol. 106 (1992) 838.

J. Oszajca et al. / J. Inorg. Biochem. 69 (1998) 121±127 [26] J. Regli nski, A.R. Butler, C. Glidewell, Appl. Organometallic Chem. 8 (1994) 25. [27] H. Swinehart, Coord. Chem. Rev. 2 (1967) 385. [28] J.A. Mc Cleverty, Chem. Rev. 79 (1979) 53. [29] E. Wasielewska, Z. Stasicka in: J.J. Zi olkowski (Ed.), Coordination Chemistry and Catalysis, World Scienti®c, Singapore, 1987, pp. 440±457. [30] Z. Stasicka, G. Stochel, E. Wasielewska, in: E. Pruchnik, M. Zuber (Eds.), Progress in Inorganic and Organometallic Chemistry, Wroclaw, Poland, 1994, pp. 274±288. [31] A. Goø e biewski, M. Witko, Acta Phys. Polon. A 57 (1980) 585. error [32] A. Goø e biewski, M. Witko, ibid A 51 (1977) 629. error [33] F.A. Cotton, R.R. Monchamp, R.J.M. Henry, R.C. Young, J. Inorg. Nucl. Chem. 10 (1959) 28. [34] W.P. Grith, J. Lewis, G.N. Wilkinson, J. Chem. Soc. (1959) 872. [35] E. Hejmo, E. Porcel-Ortega, T. Senkowski, Z. Stasicka, Bull. Pol. Acad. Sci. Chem. 36 (1988) 351. [36] D. Salvemini, A. Radziszewski, R. Korbut, J. Vane, Br. J. Pharmacol. 101 (1990) 991. [37] E. Wasielewska, A. Goø e biewski, Pol. J. Chem. 55 (1981) 1099. error [38] G. Stopa, Z. Stasicka, Polyhedron 3 (1984) 247. [39] G. Stochel, R. van Eldik, Z. Stasicka, Inorg. Chem. 25 (1986) 3663. [40] R.P. Cheney, M.G. Simic, M.Z. Ho€man, I.A. Taub, K.D. Asmus, Inorg. Chem. 16 (1977) 2187. [41] J.D.W. Van Voorst, P. Hemmerich, J. Chem. Phys. 45 (1966) 3914. [42] G. Stochel, Z. Stasicka, Polyhedron 4 (1985) 1887. [43] G. Stochel, R. van Eldik, E. Hejmo, Z. Stasicka, Inorg. Chem. 27 (1988) 2767.

127

[44] J.H. Enemark, M.S. Quinby, L.L. Recol, M.I. Stenck, K.K. Walthers, Inorg. Chem. 9 (1970) 2397. [45] B. Folkesson, Acta Chem. Scand. A 28 (1974) 491. [46] J.H. Enemark, R.D. Feltham, Coord. Chem. Rev. 13 (1974) 339. [47] A. Tullberg, N.G. Vannerberg, Acta Chem. Scand. 21 (1967) 1462. [48] P.T. Manoharan, W.C. Hamilton, Inorg. Chem. 2 (1963) 1043. [49] J. Schmidt, H. K uhr, W.L. Dorn, J. Koph, Inorg. Nucl. Chem. Lett. 10 (1974) 55. [50] E. Wasielewska, Inorg. Chim. Acta 113 (1986) 115. [51] Z. Stasicka, E. Wasielewska, Coord. Chem. Rev. 159 (1997) 271. [52] T. Senkowski, Roczniki Chemii 38 (1964) 1751. [53] J. Mocak, D.D. Bustin, Electroanal. Chem. Interfacial Electrochem. 57 (1974) 369. [54] S.K. Wolfe, J.H. Swinehart, Inorg. Chem. 14 (1975) 1049. [55] A. Goø e biewski, E. Wasielewska, J. Mol. Struct. 67 (1980) 183. error [56] K. Matsunaga, R.F. Furchgott, J. Pharmacol. Exp. Ther. 259 (1991) 140. [57] T.W. Hawkins, M.B. Hall, Inorg. Chem. 19 (1980) 1735. [58] D. Mulvey, W.A. Walters, J. Chem. Soc. Dalton. Trans. 10 (1975) 951. [59] H.P. Misra, J. Biol. Chem. 259 (1984) 12678. [60] R.P. Cheney, S.D. Pell, M.Z. Ho€man, J. Inorg. Nucl. Chem. 41 (1979) 489. [61] J. Fiedler, J. Masek, Inorg. Chim. Acta 81 (1984) 117. [62] W. Clark, B. Cohen, H.D. Gibbs, Chem. Zentr. Band 11, 25 (1929) 3153. [63] L. Michaelis, E.S. Hill, P.M. Schubert, Biochem. Zeitschr. 225 (1932) 66.