Iron-sulphur cluster nitrosyls, a novel class of nitric oxide

Transcription

1 Br. J. Pharmaol. (1992), 17, Iron-sulphur luster nitrosyls, a novel lass of nitri oxide generator: mehanism of vasodilator ation on rat isolated tail artery 'F.W. Flitney, I.L. Megson, td.e. Flitney & *A.R. Butler (D Mamillan Press Ltd, 1992 Caner Biology Researh Group, Shool of Biologial & Medial Sienes and the *Department of Chemistry, University of St Andrews, St Andrews, Fife, Sotland, KY16 9TS and tuniversity Laboratory of Physiology, University of Oxford, Parks Rd., Oxford OXl 3PT 1 Two iron-sulphur luster nitrosyls have been investigated as potential nitri oxide (O-) donor drugs (A: tetranitrosyltetra-jl3-sulphidotetrahedro-tetrairon; and B: heptanitrosyltri-ji3-thioxotetraferrate(1-)). Both ompounds are shown to dilate preontrated, internally-perfused rat tail arteries. 2 Bolus injetions (1 Ll) of ompound A or B generate two kinds of vasodilator response. Doses below a ritial threshold onentration (DT) evoke transient (or T-type) responses, whih resemble those seen with onventional nitrovasodilators. Doses > DT produe sustained (or S-type) responses, omprising an initial, rapid drop of pressure, followed by inomplete reovery, resulting in a plateau of redued tone whih an persist for several hours. 3 T- and S-type responses are attenuated by ferrohaemoglobin (Hb) and by methylene blue (MB), but not by inhibitors of endothelial O synthase. Addition of either Hb or MB to the internal perfusate an restore agonist-indued tone when administered during the plateau phase of an S-type response. Moreover, subsequent removal of Hb auses the artery to re-dilate fully. 4 We onlude that T- and S-type responses are both mediated by O. It is postulated that S-type responses represent the sum of two vasodilator omponents: a reversible omponent, superimposed upon a non-reoverable omponent. The former is attributed to free O, preformed in solution at the time of injetion; and the latter to O generated by gradual deomposition of a 'store' of iron-sulphur-nitrosyl omplexes within the tissue. This hypothesis is supported by histohemial studies whih show that both lusters aumulate in endothelial ells. Keywords: itri oxide donors; iron-sulphur-nitrosyls; vasodilator responses; nitrovasodilators; endothelium Introdution Endothelial ells release a labile fator whih relaxes vasular smooth musle (Furhgott & Zawadski, 198). Endotheliumderived relaxing fator (or EDRF) has reently been identified as either nitri oxide (O-; Palmer et al., 1987; Ignarro et al., 1988); a labile nitrosothiol, possibly S-nitrosoysteine (Myers et al., 199); or a nitrosyl-iron omplex with thiol ligands (Vanin, 1991). O is formed from the terminal guanidino nitrogen atom of L-arginine by a itrulline-forming enzyme, referred to as O synthase (Palmer et al., 1988a,b). Endothelial O synthase (OS) is ADPH- (Palmer & Monada, 1989) and almodulin-dependent (Bredt & Snyder, 199). It an be inhibited by several L- (but not D-) analogues of arginine (Rees et al., 199). Haemodynami studies using stereospeifi OS inhibitors have firmly established the importane of O in ontrolling peripheral resistane in vivo (Valiane et al., 1989; Aisaka et al., 1989; Rees et al., 1989; Gardiner et al., 199; Chu et al., 1991). Endothelium-dependent relaxations of vasular smooth musle are attenuated in some diseased states, notably in hypertension (Winquist et al., 1984; Lusher & Vanhoutte, 1986; Otsuka et al., 1988; Tesfamariam & Halpern, 1988; Sunano et al., 1989) and in atheroslerosis (Harrison et al., 1987; Chappell et al., 1987; Henry et al., 1987; Forstermann et al., 1988, Guerra et al., 1989), though it remains unlear whether these onditions are neessarily assoiated with impaired synthesis and/or release of O (Van der Voorde & Leusen, 1986; Hoeffner & Vanhoutte, 1989; Chester et al., 199; Jaobs et al., 199; Fozard & Part, 1991). Author for orrespondene. Clinial 'nitrovasodilators', inluding amyl nitrite, glyeryl trinitrate, nitroprusside (P) and molsidomine, at independently of the endothelium, augmenting its funtion temporarily by releasing O- in vivo (Waldman & Murad, 1984; Feelish & oak, 1987). Their hypotensive ations are generally of short duration and ontinuous drug infusions are neessary to effet sustained responses (Kreye, 198). Here we show that two tetrairon-sulphur luster nitrosyls (ompound A: tetranitrosyl-tetra-fi3-sulphidotetrahedro-tetrairon; heptanitrosyl-tri-li3-thioxotetraferrate(i-), and ompound B: also known as Roussin's Blak Salt; Roussin, 1858) are able to generate unusually protrated vasodilator responses when tested on rat isolated, internally-perfused tail artery preparations. The strutures of both lusters, based on X-ray diffration studies (Chu & Dahl, 1977; Chu el al., 1982), are shown in Figure la,b. The vasodilator ations of A and B are bloked by haemoglobin and also by methylene blue, but not by agents whih suppress OS. Histohemial studies reveal that both lusters are able to penetrate ell membranes with extraordin~ary ease and that they aumulate inside the endothelium, forming a moleular 'store' of O. Our results suggest that their gradual deomposition from sites within the endothelium generates free O and that its slow release serves to sustain their vasodilator ations. Preliminary aounts of some aspets of this work have appeared elsewhere (Flitney et al., 199; 1992) Methods Preparation Experiments were performed on segments of tail artery taken from normotensive adult Wistar rats (25-46 g). Animals

2 VASODILATOR PROPERTIES OF IRO-SULPHUR ITROSYLS 843 were killed by ervial disloation. A length of artery ( m) was disseted free, annulated (Portex annula) and transferred to a Perspex bath. Apparatus The apparatus is shown in Figure 1. The annula (C) formed part of a onstant flow perfusion iruit, driven by a peristalti pump (P1; Gilson Minipuls). The vessel (V) was perfused internally (flow rate: 2 ml min-') with solution pre-warmed by passage through a heat exhanger. Drugs were introdued into the lumen of the artery by bolus injetion through a side tube (I). The outer surfae of the vessel was superfused ontinuously with solution driven by a seond peristalti pump (P2). The temperature of the hamber was held at 3-35C by adjustment of the flow rate in the external iruit (a. 8 ml min-'). Light from an Argon ion laser (L; Spetra Physis Ltd., type 168-9) ould be made to irradiate the artery diretly, so as to indue photorelaxation of vasular smooth musle (Furhgott et al., 1961). The beam (A = nm; output intensity I mw) was refleted onto the preparation by means of a front-silvered mirror (M; beam diam. at preparation: approx 2 m). A differential pressure transduer (T: Sensym type SCX 1 5OC; Farnell Eletroni Components, Leeds) deteted hanges in bak pressure due to hanges in arterial tone. Responses were displayed on a hart reorder. T P1P P.-= a t O OFe--A S O FeK- ZO S- Fe S SO---Fe O Fe O\,S/-t - O -K IFe6O' Fe' -~~Fe M Figure 1 (a, b) Moleular strutures of lusters A and B, as determined by X-ray diffration studies. Compound A ontains four and B seven ligated O groups per moleule. () Apparatus used for perfusing isolated segments of rat tail artery and for irradiating preparations with laser light. See text for full desription and explanation of lettering. im I' L Experimental protool Arteries were perfused internally with oxygenated Krebs solution (omposition, mm: acl 118, KCl 4.7, ahco3 25, ah2po4 1.15, CaC12 2.5, MgCl2 1.1, gluose 5.6, gassed with 95% 2/5% CO2 to maintain ph 7.4), initially at a low flow rate whih was inreased gradually over the next 1-2 min to reah a final value of 2 ml min-'. The preparation was allowed to stabilize for 2-3 min, after whih it was preontrated with Krebs plus phenylephrine HCI ( = Krebs- + PE: mean ( ± s.e.) [PE] = 6.5 ±.5 JIM. Mean ( ± s.e.) agonist-indued perfusion pressure= 11 ± 3.5mmHg). Experiments were made to establish the mode of ation of iron-sulphur luster nitrosyls. Responses to bolus injetions (1 l) of A and B were ompared with those evoked by (i) arbahol (CCh), an endothelium-dependent vasodilator; and (ii) either nitroprusside (P), S-nitroso--aetylpeniillamine (SAP) or exposure to laser light (L) all of whih relax vasular smooth musle independently of the endothelium. The effets of adding the following to the internal perfusate were examined: (i) ferro-haemoglobin (Hb) a nitri oxide 'savenger'; (ii) methylene blue (MB); and (iii) either -monomethyl-l-arginine (L-MMA) or -nitro-l-arginine methyl ester (L-AME), both stereospeifi inhibitors of OS. Comparisons were made with responses evoked by P or SAP. Drugs: ommerial soures and syntheti proedures L-phenylephrine HCI (mol. wt. 23.7; Sigma Chemials) was used at onentrations ranging from 2-12x1O-6M. Stok solutions (1O-2M) of P (mol. wt. 298; BDH Ltd., 'Analar' grade) were made and the dose required adjusted by serial dilution immediately prior to use. SAP (mol. wt. 22) was synthesized by reating -aetylpeniillamine with sodium nitrite (Field et al., 1978). Compound B was prepared by reating iron(ii) sulphate heptahydrate, sodium nitrite and sodium sulphide in hot aqeous solution under nitrogen gas (Brauer, 196). Compound A was prepared by addition of elemental sulphur to a solution of B dissolved in re-distilled toluene and refluxed overnight (16 h) under dry nitrogen gas (Gall et al., 1974). Purities of iron sulphur nitrosyl ompounds were heked by (a) infra red spetrosopy (A had a single IR absorption maximum at wave number 179 m-'; B had three maxima, at 1795 m-', 1747 m-' and at 177 m-') and (b) "-nulear magneti resonane (MR) spetrosopy. Solutions of B (mol. wt. 553) were made up in Krebs solution, immediately before use. Compound A (mol. wt. 472) is not soluble in water and so working solutions were made by serial dilution of 5 mg A dissolved in 1 ml dimethyl sulphoxide (a. 1-2M). An important note onerning experimental proedures All experiments were onduted in a darkened laboratory with a red safelight (6W) as the sole means of illumination. There were two reasons for this. First, the ambient lighting under normal laboratory onditions is suffiient to indue photorelaxation of preontrated vasular smooth musle, ausing a progressive loss of vessel tone. Seond, P, A and B are photosensitive and deompose when exposed to light, releasing free O- (Wolfe & Swinehart, 1975; Flitney & Kennovin, 1988; Flitney et al., unpublished results). Drug solutions were therefore made up in the dark and stored in glass vials overed with aluminium foil prior to use. Results Bolus injetions of either P or SAP invariably produe fully-reversible (transient or T-type) vasodilator responses: an

3 844 F.W. FLITEY et al. initial, rapid drop of pressure, followed by omplete reovery, often with some positive 'overshoot'. Compounds A and B produed T-type responses only at doses below a ritial threshold onentration (DT). Doses> DT generate long-lasting (sustained or S-type) responses, omprising an initial, rapid derease of pressure, followed either by partial reovery only, or no reovery at all. The hallmark of the S-type response is a remarkably stable plateau of redued vessel tone whih an persist for several hours. Mediation of T-type responses by O- Figure 2 ompares vasodilator responses to arbahol (CCh), nitroprusside (P) and laser light (L; = nm; 1 mw intensity) with T-type responses produed by A and B. Control responses are shown in Figure 2a. The responses to A and B were bloked by either methylene blue (b; 75 gm) or by a CCh P B A L b ferro-haemoglobin (; 15 tlm), a O- savenger. However, neither L-MMA (d; 1 pm) nor L-AME (not illustrated), both of whih suppress endothelial OS (Rees et al., 199), inhibited responses to A or B. Indeed, it was usual to find that treatment with either L-MMA or L-AME inreased the responsiveness of arteries to A and B. S-type responses produed by sequential injetions of inreasing doses of A or B Pressure reordings made by injetion of inreasing doses of P (Figure 3) or SAP (not illustrated) onfirm that both ompounds produe T-type responses only. Similar experiments with A or B show that T-type responses give way to S-type responses when the injeted dose exeeds DT: both the rate and extent of the reovery following suessive injetions beome progressively redued, resulting in a stepwise and persistent loss of vessel tone (Figure 3a,b). Three parameters have been measured (X, Y and Z; see Figure 3a) to enable quantitative omparisons to be drawn between P and SAP and either A or B. Figure 4 shows perfusion pressure minima (filled irles) immediately following injetion of eah drug (X) and the steady-state pressures attained after reovery (Z; open irles), both expressed as a fration of the initial agonist-indued pressure (Y) and plotted as a funtion of log1 injeted dose. The onentrations of SAP, P, A and B ( tm) required to produe half-maximal L MB (75 FtM) a I d Hb (15 F.M) b L L-MMA (1 FLm) Figure 2 (a) Vasodilator responses produed by miroinjetion (1 fl) of test doses of arbahol (CCh: 1O-2 M) and nitroprusside (P; 2.5x1-4 M) and by irradiation with laser light (L; A = nm; I mw; 2 min), ompared with T-type responses produed by lusters A (14 M) and B (5x1-4 M). Artery pre-ontrated with 5x1-6 M phenylephrine (agonist-indued pressure = 135 mmhg). Female rat, 274 g. Temp., 32'C. (b) Suppression of responses to A and B and to laser light by treatment with methylene blue (MB 7.5x1-5 M). Details as above. () Inhibition of responses to ompound A and B by haemoglobin (Hb, 15x1-6 M). Hb inreased agonist-indued pressure to 1.8x ontrol level (after a. 6 min). Artery pre-ontrated with 4x1-6 M phenylephrine (agonist-indued pressure = 114 mmhg). Female rat, 373 g. Temp. 32'C. (d) The O synthase (OS) inhibitor, G-monomethyl-L-arginine (L-MMA,1x lo M) inreased agonist-indued tone to around 1.28x the ontrol level. Drugs tested 2 min after beginning perfusion with L-MMA. Responses to arbahol attenuated by L-MMA, but not those produed by nitroprusside (P), A, B or laser light. Details as for () above. Calibration bars (all reordings): vertial, 2% agonist-indued tone (.2Y; see Figure 3a); horizontal, 4 min. Figure 3 Averaged pressure reordings obtained by suessive injetions of inreasing doses of ompound A, B or nitroprusside (P) (graphs a, b and, respetively). Mean values ( ± s.e., vertial bars) are plotted (n = 6 preparations for a, and n = 5 for both b and ). ote that injetion doses of A or B whih exeed DT (see text) produe a sustained drop in vessel tone (S-type response), whereas responses to P are fully reversible (T-type response). In (a), Y = agonist-indued pressure; X = pressure minimum following eah injetion; and Z = steady-state (plateau) pressure attained after reovery (see also Figure 4). Calibration bars: vertial =.2Y; horizontal = 2 min (A and B) or 1 min (C). /

4 VASODILATOR PROPERTIES OF IRO-SULPHUR ITROSYLS 845 a _.6 _.4 _.2 _- L i I I I I I I lo dereases of X/Y (EC5 values) are 36, 43, 126 and 399, respetively. The orresponding urves for Z/Y reflet the tendeny for A and B (but not P) to generate S-type responses. The effet is more pronouned for B than for A, as shown by the greater proximity of the X/Y and Z/Y urves. DT is estimated to be around 1-11 gm for B and 2-1fiM for A. Abolition of the 'plateau' phase of the S-type response by Hb or by MB The S-type response was remarkable in that even prolonged perfusion ('wash-out') with Krebs+PE solution alone (up to 5 h) failed to indue any further vasoonstrition, one the plateau phase was established. However, the addition of either Hb (15 gm) or MB (>1 jim; not illustrated) to the internal perfusate initiated a prompt and omplete restoration of all agonist-indued tone; indeed, as Figure 5 learly shows, the perfusion pressure rises well above the ontrol -Hb 1.2 b h.2.6 H.4 F-.2 H _ I I I I I I I r l Figure 5(a, b) Pressure reordings from two preparations showing responses to ferro-haemoglobin (Hb, 31.5x1-5 M) added to, and then later removed from, the internal perfusate during the plateau phase of an S-type response. S-type responses were produed by a single bolus injetion of 5x1-3 M A or B (upper and lower reordings respetively). The addition of Hb (upward arrows) eliits a prompt vasoonstrition whih auses the perfusion pressure to rise above the ontrol (pre-injetion) value. Removal of Hb produes a vasodilatation whih returns the pressure to the plateau value (urved dashed lines). Similar responses to these have been reorded when the time delay between the injetion of A or B and treatment with Hb was inreased to 6 h. Calibration bars: vertial =.2Y; horizontal = 4 min loglo [Ml, A, B or P -1 Figure 4 Mean ( + or - s.e., vertial lines values for X/Y (-) and Z/Y (; see Figure 3a) plotted as a funtion of logl injeted dose for A (a), B (b) and nitroprusside (P) (). Responses to P show omplete reovery, with some pressure overshoot at the higher doses (Z/Y > 1.). S-nitroso--aetylpeniillamine (SAP) (data not shown) gives results whih are qualitatively similar to those seen with P. The values of Z/Y for both lusters (measured 2 min after eah injetion) are < 1. at doses whih exeed DT (by definition; see text), refleting the tendeny for A and B to evoke S-type responses. Data obtained from same pressure reordings used to onstrut Figure 3 (a-).

5 846 F.W. FLITEY et al. level (Y), here to around 1.6Y (A) and 1.8Y (B). Even more striking, the subsequent removal of Hb (but not of MB) auses arteries to re-dilate fully (Figure 5a,b). These observations lead us to onlude that extremely brief exposure of arteries to A or B (the 'transit' time for a 1 p1l bolus passing through the vessel is a..3 s) an suffie to establish a durable soure of O within the tissue. Histohemial studies showing that iron sulphur luster nitrosyls penetrate endothelial ells and aumulate there The existene of a long-lasting soure (store) of O- in vessels exposed to either luster is supported by mirosopi studies of treated arteries. Solutions of A and B are intensely blak and visual inspetion of segments of artery subjeted to repeated bolus injetions (experiments of the type used to onstrut Figure 3) revealed some disolouration of the vessel lining. Mirosopi examination of freshly-frozen, unfixed setions of artery exposed to luster B (5 x 1IO3M, 5 min ontinuous perfusion, followed by 15 min washout with Krebs+ PE to remove exess drug from the lumen of the artery) showed that the disolouration was onfined to the endothelium (Figure 6a). Unfixed, frozen setions whih had been stained with either bathophenanthroline (Figure 6b) or ferroyanide (not shown), to detet iron(ii) derived from the lusters (Pearse, 1972), showed a similar distribution of reation produt, loated. predominantly in endothelial ells, but sometimes with some faint staining of adjaent smooth musle ells also. Control (untreated) arteries did not reat positively for iron. Disussion This paper establishes iron-sulphur luster nitrosyls as a novel lass of O donor drug with unusual vasodilator properties. The ompounds investigated here are able to generate two distint kinds of response, designated T-and S-type responses. In this respet they differ from either P or SAP, both of whih give fully-reversible responses only (Figures 3 and 4). The ability of Hb and of MB, but not OS inhibitors, to blok T- and S-type responses (Figure 2) shows that they are mediated by O derived from eah luster, and not by enhaned EDRF prodution through stimulation of the endothelial L-arginine:O pathway. Composite nature of the S-type response Our results lead us to postulate that the S-type response represents the sum of two, distint vasodilator omponents: a reversible omponent, attributable to free O generated by the spontaneous deomposition of lusters in solution and present at the time of injetion; superimposed on a 'nonreoverable' omponent, aused by the delayed release of O from lusters whih have beome trapped within the endothelium. This hypothesis is supported by histohemial studies (Figure 6), whih revealed the presene of Fe(II) within endothelial ells of treated (but not ontrol) arteries, and by our observation that the plateau phase of the S-type response an be abolished in a reversible manner by perfusing preparations with Hb (Figure 5). Vanin and o-workers (Vanin, 1991) have shown that irondinitrosyls (Fe(O)2) omplexed with low moleular weight thiols (e.g. ysteine or glutathione) an vasodilate isolated blood vessels and also inhibit platelet aggregation. Interestingly, the vasodilator ation exhibits two-phase kinetis when tested on intat animals, resembling the S-type response desribed in the present study: that is, an initial, rapid derease of arterial pressure, followed by a persistent (several hours) hypotensive ation. The rapid and sustained omponents of the response are attributed to O released from (i) irondinitrosyl low-moleular thiol omplexes (= rapid phase) and Figure 6 Transverse setions (2 pm) of a freshly-frozen artery perfused with B (5 mm, 2 ml min-') for 5 min, followed by 15 min perfusion (wash-out) with Krebs + PE solution only. (a) Unstained, unfixed setion showing disolouration of the vessel lining. (b) Freshly-frozen, unfixed setion after staining for 7 h with bathophenanthroline, followed by brief (4 min) ounterstaining with methylene blue (sale as in (a)). Endothelial ells stain positively (red-brown olouration) for iron(ii) derived from the luster. () Unfixed setion stained for 4 min with methylene blue to show up smooth musle ells. ote that brown olouration is onfined to the endothelial lining of the vessel.

6 VASODILATOR PROPERTIES OF IRO-SULPHUR ITROSYLS 847 (ii) from iron-dinitrosyl protein-thiol omplexes (= sustained phase) formed within the tissue. The latter are thought to result from the transfer of Fe(O)2 groups from low moleular weight to protein-borne ligands, forming a longlasting, moleular 'store' of O- in the vasular bed. Suppression of endogenous EDRF prodution did not our during the plateau phase of S-type responses The effets of Hb administered during established S-type responses lead us to onlude that the formation of endogenous O from L-arginine is not suppressed by either luster and that its ontinued release from endothelial ells helps to sustain the plateau phase. Thus, perfusing vessels with Hb does not merely re-establish the ontrol (preinjetion) perfusion pressure, but instead drives it to a level substantially greater than this (Figure 5). The inferene to be drawn is that Hb savenges O from both exogenous (luster-derived) and endothelial soures. The nature of the intraellular mehanism(s) underlying the vasodilator ations of iron-sulphur nitrosyls remains to be established before meaningful omparisons an be made with other nitrovasodilators. The fat that Hb and MB an blok responses to either luster (and also to SAP) but not those evoked by P (Figure 2) learly implies different ellular mehanisms of ation. This has not been investigated in the present study. However, there is evidene to suggest that P may relax vasular smooth musle through both guanosine 3':5'-yli monophosphate (yli GMP)- dependent and yli GMP-independent pathways (Otsuka et al., 1988). A omponent of the relaxant effet of glyeryltrinitrate may also be mediated by a mehanism whih does not involve elevated yli GMP levels (Diamond & Chu, 1983). Physiohemial properties of iron-sulphur luster nitrosyls The physiohemial properties of iron sulphur luster nitrosyls (Butler et al., 1988) provide some insight into their vasodilator ations. Both ompounds are potentially able to transport large quantities of O-, with 4 and 7 mol. of ligated O- per mol. of A and B, respetively. Their seletive retention and gradual deomposition from within the endothelium an evidently generate physiologially signifiant amounts of O at this most relevant of sites. The mehanism by whih O is released from lusters trapped inside the endothelium (whether spontaneous or enzymati) is not known. Extended Hukel moleular orbital alulations show that A and B are eletron preise (Sung et al., 1985): this means that addition of an eletron to A, and either the addition or removal of an eletron in the ase of B, will weaken the age bonding and ause the iron sulphur framework to disintegrate, releasing free O. The oxidative status of the endothelial ell may therefore prove to be an important fator in determining the rate of O release in vivo. The apparent ease with whih both lusters are able to penetrate the endothelial ell membrane is probably related to their high solubility in non-polar solvents (Butler et al., 1988). Compound B is espeially interesting in this regard beause it is ioni (A is neutral) and therefore soluble in polar solvents too. Curiously, though, the marked ontrast in the intensity of staining as seen between endothelial ells and overlying smooth musle ells (see Figure 6) shows that neither ompound an traverse the endothelial ell layer readily. The faint reation produt whih is sometimes disernible in smooth musle ells immediately adjaent to the endothelium is probably a diffusion artefat of the staining proedure, beause neither luster survives onventional fixation methods (inluding vapour fixation) and so histohemial observations were of neessity made on freshlyfrozen setions. This explanation is supported by preliminary X-ray miroprobe studies of rapidly-quenhed arteries following treatment with A or B, whih reveal elemental Fe and S within endothelial ells, but no evidene for either in nearby smooth musle ells (unpublished observations: Elder, Pediani, Megson & Flitney). This work is supported by a grant from the Caner Researh Campaign. We thank Dr Sheila Glidewell for her help and advie in synthesizing the lusters. Dr Rihard Palmer (Wellome Researh Laboratories, Bekenham, Kent) kindly provided the samples of L-MMA used in this study. The expert tehnial assistane of Mr David Rohe, Ms Karen A. Johnstone and Mr R.J. Stuart is gratefully aknowledged. Referenes AISAKA, K., GROSS, S.S., GRIFFITH, O.W. & LEVI, R. (1989). Gmethyl arginine, an inhibitor of endothelium-derived nitri oxide synthesis is a potent pressor agent in the guinea pig: does nitri oxide regulate blood pressure in vivo? Biohem. Biophys. Res. Commun., 16, BRAUER, G. (196). In Handbuh der Praparativen Anorganishen Chemie, 2nd Edition. Vol. 2, p Stuttgart: Enke. BREDT, D.S. & SYDER, S.H. (199). Isolation of nitri oxide synthase, a almodulin-requiring enzyme. Pro. atl. Aad. Si. U.S.A. 87, BUTLER, A.R., GLIDEWELL, C. & MI-HSI, L. (1988). itrosyl omplexes of iron-sulfur lusters. Adv. Inorg. Chem., 32, CHAPPELL, S.P., LEWIS, M.J. & HEDERSO, A.H. (1987). Effet of lipid feeding on endothelium-dependent relaxation in rabbit aorta preparations. Cardiovas. Res. 21, CHESTER, A.H., O'EIL, G.S., MOCADA, S., TADJKARIMI, S. & YACOUB, M. (199). Low basal and stimulated release of nitri oxide in atherosleroti epiardial oronary arteries. Lanet, 336, CHU, C.T-W. & DAHL, L.F. (1977). Strutural harateristis of [AsPh4]*[Fe4(O)7(Ap3-S)3]iI. Stereohemial and bonding relationship of the Roussin blak monoanion and the red ethyl ester, Fe2(O)4(j2-SC2H5)2, and Fe4(O)4 (p3-s)4. Inorg. Chem., 16, CHU, A., CHAMBERS, D.E., LI, C-C., KUEHL, W.D., PALMER, R.J., MOCADA, S. & COBB, F. (1991). Effets of inhibition of nitri oxide formation on basal vasomotion and endothelium-dependent responses of the oronary arteries in awake dogs. J. Clin. Invest., 87, CHU, C.T-W., LO, F.Y.-K., & DAHL, L.F. (1982). Synthesis and stereohemial analysis of the [Fe4(O)4 (ju3-s4)]n series (n =, - 1) whih possesses a ubane-like Fe4S4 ore: diret evidene for the antibonding tetrametal harater of the unpaired eletron upon a one-eletron redution of a ompletely bonding tetrahedral metal luster. J. Am. Chem. So., 14, DIAMOD, J. & CHU, E.B. (1983). Possible role for yli GMP in endothelium-dependent relaxation of rabbit aorta by aetylholine. Comparison with nitrogylerin. Res. Commun. Chem. Path. Pharmaol., 41, FEELISCH, M. & OAK, E.A. (1987). Correlation between nitri oxide formation during degradation of organi nitrates and ativation of guanylate ylase. Eur. J. Pharmaol., 139, FIELD, L., DILTS, R.V., RAVICHADRA, R., LEHERT, P.G. & CARAHA, G.E. (1978) An unusually stable thionitrite from -aetyl-d,l-peniillamine; X ray struture and moleular struture of 2-(aetylamino)-2-arboxy-1,l-dimethylethylthionitrite. J. Chem. So. Chem. Commun.,

7 848 F.W. FLITEY et al. FLITEY, F.W. & KEOVI, G.D. (1988). Haemoglobin suppresses the negative inotropi effet of photolysed nitroprusside on isolated frog ventriular trabeulae: identifiation of nitrosylhaem formation by eletron paramagneti resonane and absorption spetrosopy. J. Physiol., 49, 28P. FLITEY, F.W., MEGSO, I.L. & BUTLER, A.R. (1992). Iron-sulphur luster nitrosyls: long-ating, nitri oxide donor drugs. In Biology of itri Oxide. ed. Monada, S., Marletta, M.B., Hibbs Jnr., J.B. & Higgs, A.E. Portland Press, (in press). FLITEY, F.W., MEGSO, I.L., CLOUGH, T. & BUTLER, A.R. (199). itrosylated iron-sulphur lusters, a novel lass of nitrovasodilator: studies on the rat isolated tail artery. J. Physiol., 43, 42P. FORSTERMA, U., MUGGE, A., ALHEID, U., HAVERICH, A. & FROLICH, J.C. (1988). Seletive attenuation of endotheliummediated vasodilation in atherosleroti human oronary arteries. Cir. Res., 62, FOZARD, J.R. & PART, M.L. (1991). Haemodynami responses to -monomethyl-l-arginine in spontaneously hypertensive and normotensive Wistar-Kyoto rats. Br. J. Pharmaol., 12, FURCHGOTT, R.F., EHRREICH, S.J. & GREEBLATT, E. (1961). The photoativated relaxation of smooth musle of rabbit aorta. J. Gen. Physiol., 44, FURCHGOTT, R.F. & ZAWADSKI, J.V. (198). The obligatory role of endothelial ells in the relaxation of arterial smooth musle by aetylholine. ature, 288, GALL, R.S., CHU, C.T.-W. & DAHL, L.F. (1974). Preparation, struture and bonding of two ubane-like iron-nitrosyl omplexes, Fe4(O)4(p3-S)4 and Fe4(O)4(Q3-S)2(p3-C(CH3)3)2. Stereohemial onsequenes of bridging ligand substitution on a ompletely bonding tetrametal luster unit and of different terminal ligands on the ubane-like Fe4S4 ore. J. Am. Chem. So., 96, GARDIER, S.M., COMPTO, A.M., KEMP, P.A. & BEET, T. (199). Regional and ardia haemodynami effets of G-nitro- L-arginine methyl ester in onsious, Long Evans rats. Br. J. Pharmaol., 11, GUERRA, R. Jr., BROTHERTO, A.F.A., GOODWI, P.J., CLARK, C.R., ARMSTROG, M.L. & HARRISO, D.G. (1989). Mehanisms of abnormal endothelium-dependent relaxation in atheroslerosis: impliations for altered autorine and pararine funtions of EDRF. Blood Vessels, 26, HARRISO, D.G., ARMSTROG, M.L., FRIEMA, P.C. & HEISTAD, D.D. (1987). Restoration of endothelium-dependent relaxation by dietary treatment of atheroslerosis. J. Clin. Invest., 8, HERY, P.D., BOSSALLER, C. & YAMMOMOTO, H. (1987). Impaired endothelium-dependent relaxation and yli guanosine 3', 5' monophosphate formation in atherosleroti human oronary artery and rabbit aorta. Thromb. Res., Suppl. VII, 6. HOEFFER, U. & VAHOUTTE, P.M. (1989). Inreases in flow redue the release of endothelium-derived relaxing fator in the aorta of normotensive and spontaneously hypertensive rats. Am. J. Hypertens., 2, IGARRO, L.J., BURS, R.E., BUGA, G.M., WOOD, K.S. & CHAUD- HURI, G. (1988). Pharmaologial evidene that endotheliumderived relaxing fator is nitri oxide: use of pyrogallol and superoxide dismutase to study endothelium-dependent and nitri oxide-eliited vasular smooth musle relaxation. J. Pharmaol. Exp. Ther., 244, JACOBS, M., PLAE, F. & BRUCKDORFER, R. (199). Inhibition of endothelium-derived nitri oxide and atheroslerosis. In itri Oxide from L-Arginine. a Bioregulatory System. ed. Monada, S. & Higgs, E.A. pp Amsterdam: Elsevier. KREYE, V.A. (198). Sodium nitroprusside. In Pharmaology of Antihypertensive Drugs. ed. Sribine, A. pp ew York: Raven Press. LUSCHER, T.F. & VAHOUTTE, P.M. (1986). Endothelium-dependent ontrations to aetylholine in the aorta of the spontaneously hypertensive rat. Hypertens., 8, MYERS, P.R., MIOR, R.L., GUERRA, R., BATES, J.. & HARRISO, D.G. (199). Vasorelaxant properties of the endothelium-derived relaxing fator more losely resemble S-nitrosoysteine than nitri oxide. ature, 345, OTSUKA, U., DIPIERO, A., HIRT, E., BRAAMA, B. & LOCKErTE, W. (1988). Vasular relaxation and yli GMP in hypertension. Am. J. Physiol., 254, H PALMER, R.J., ASHTO, D. & MOCADA, S. (1988a). Vasular endothelial ells synthesise nitri oxide from L-arginine. ature, 333, PALMER, R.J., FERRIGE, A.G. & MOCADA, S. (1987). itri oxide release aounts for the biologial ativity of endothelium-derived relaxing fator. ature, 327, PALMER, R.J. & MOCADA, S. (1989). A novel itrulline-forming enzyme impliated in the formation of nitri oxide by vasular endothelial ells. Biohem. Biophys. Res. Commun., 158, PALMER, R.J., REES, D.D., ASHTO, D.S. & MOCADA, S. (1988b). L-arginine is the physiologial preursor for the formation of nitri oxide in endothelium-dependent relaxation. Biohem. Biophys. Res. Commun., 153, PEARSE, A.G.E. (1972). Histohemistry. Third edition. Vol. II. Edinburgh, London: Churhill Livingstone. REES, D.D., PALMER, R.J. & MOCADA, S. (1989). Role of endothelium-derived nitri oxide in the regulation of blood pressure. Pro. atl. Aad. Si. U.S.A., 86, REES, D.D., PALMER, R.J., SCHULTZ, R., HODSO, H.F. & MO- CADA, S. (199). Charaterisation of three inhibitors of endothelial nitri oxide synthase in vitro and in vivo. Br. J. Pharmaol., 11, ROUSSI, F.Z. (1858). Reherhes sur les nitrosulfures doubles de fer. Ann. Chim. Phys., 52, SUAO, S., OSUGI, S. & SHIMAMURA, K. (1989). Blood pressure and impairment of endothelium-dependent relaxation in spontaneously hypertensive rats. Experientia, 45, SUG, S-S., GLIDEWELL, C., BUTLER, A.R. & HOFFMA, R. (1985). Bonding in nitrosylated iron-sulfur lusters. Inorg. Chem., 24, TESFAMARIAM, B. & HALPER, W. (1988). Endothelium-dependent and endothelium-independent vasodilation in resistane arteries from hypertensive rats. Hypertens., 11, VALLACE, P., COLLIER, J. & MOCADA, S. (1989). Effets of endothelium derived nitri oxide on peripheral arteriolar tone in man. Lanet, 334, VA DE VOORDE, J. & LEUSE, I. (1986). Endothelium-dependent and independent relaxation of aorti rings from hypertensive rats. Am. J. Physiol., 25, H VAI, A. (1991). Hypothesis: Endothelium-derived relaxing fator is a nitrosyl-iron omplex with thiol ligands. Febs Lett., 289, 1-3. WALDMA, S.A. & MURAD, F. (1987). Cyli GMP synthesis and funtion. Pharmaol. Rev., 39, WIQUIST, R.J., BUTIG, P.B., BASKI, E.P. & WALLACE, A.A. (1984). Dereased endothelium-dependent relaxation in ew Zealand geneti hypertensive rats. J. Hypertens., 2, WOLFE, S.K. & SWIEHART, J.H. (1975). Photohemistry of pentayanonitrosylferrate(2-), nitroprusside. Inorg. Chem., 14, (Reeived April 8, 1992 Revised July 7, 1992 Aepted July 22, 1992)