Gold Drugs: Mechanism ofActionand Toxicity

Sabine L Best and Peter] Sadler

Department ofChemistry, Birkbeck College, University ofLondon, Gordon Houseand Christopher Ingold Laboratories, 29 Gordon Square, London WC}H opp, UK

Gold drugs are still amongst the most efficacious for the treatment of rheumatoid arthritis. Their mechanism ofaction, as well as the molecular basis of their side-effects, remain poorly understood. Current theories arereviewed, including recent potential breakthroughs. The interaction of gold(III) with peptides and proteins andits immunochemical implications are discussed.

Ninety elements occur naturally on earth of which 9are radioactive. Eighty-one elements are thereforepotentially available to support life, of which 61 aremetals. It is believed that about 25 are essential forhuman life but our knowledge of the biochemistry ofseveral of these (e.g.V; Ni, Sn) is poor. In general,therefore, there is enormous scope for the use ofinorganic compounds in medicine, and a need forresearch in this area. Metals currently used in medicineinclude gold in antiarthritic agents, platinum in anti­cancer drugs, lithium for manic depression, bismuth inanti-ulcer drugs, and silver and mercury in anti­microbial agents (1). The effectiveness of rutheniumcomplexes as anti-metastatic agents with potential usein cancer therapy is currently receiving much attention(2). However, the potential of inorganic metalcompounds as drugs has yet to be fully explored. Withthe exception of platinum, which has attracted themost research efforts because of its importance in thetherapy of some types of cancer, and which is nowgenerally thought to be effective by interaction withDNA, the mechanism of action of other inorganicdrugs is mostly unknown.

One of the problems of inorganic drugs is theirside-effects. For gold drugs used in the therapy of

* Hypersensitivity is an adaptive immune response occurring in anexaggerated or inappropriate form causing tissue damage. It is acharacteristic of the individual and is manifested on secondcontact with the particular antigen (in this case gold drugs).Delayed-type hypersensitivity takes more than 12 hours todevelop and is mediated primarily by T cell and macrophages.

(6i9' GoldBulletin 1996, 29(3)

rheumatoid arthritis (RA), about 30% of patientsexperience side-effects which often necessitatediscontinuation of therapy (3). The main side-effect isa delayed-type hypersensitivity" (4) which manifestsitself in an itchy skin rash. A more detailedunderstanding of the toxicity of gold drugs mightmake their use safer and could lead to the synthesis ofless toxic compounds.

STRUCTURES OF GOLD DRUGS

Gold(I) has a very high preference for 'soft' ligandssuch as sulfur (thiolares) and phosphorus (phosphines),with little affinity for oxygen ligands and only weaklyfor nitrogen. Linear two-coordination predominatesbut trigonal or tetrahedral coordination is possible.There are two main classes of gold drugs in currentchemical use:1 Injectable gold(I) thiolates such as

aurothiomalare, aurothioglucose andaurothiopropanol sulfonate which are majoringredients of MyocrisinTM, Solganol" andAllocrysin'['", respectively, and

2 The oral complex (2,3,4,6-tetra-O-acetyl-1-thio­B-D-glucopyranosato-S) (rriethylphosphine) gold(I) (auranofin).

The oral drug has a well defined structure withlinear two-coordinate Au(I), but the injectable drugsare amorphous non-crystalline solids thought tocontain thiolate sulfur-bridged oligomers and ringstructures (Figure 1). A hexameric ring structure hasrecently been demonstrated by X-ray crystallographyfor another 1:1 gold(I) thiolate complex (5).

87

Figure 1 Structuresofsome widely usedgold drugsand themetabolite dicyanogold(I)

The elevated copper levels in RA that can becorrelated to the severity of the disease, have led to thesuggestion that antiarthritic drugs like penicillamineand aurothiomalate act by sequestering labile andpotentially disease-inducing forms of copper, e.g.copper bound to human serum albumin (11).

The three widely used gold drugs, aurothiornalate,aurothioglucose and auranofin, have been shown to

inhibit protein kinase C (12), a metallo-enzymecontaining Zn2t bound to cysteine and histidineresidues, which plays a crucial role in intracellularsignal transduction by phosphorylatingserine/threonine residues in protein. The inhibition ofprotein kinase C has been suggested as a possible modeof action for the therapeutic antirheumatic action ofgold drugs (12, 13).

Gold binding to rhiol groups on the hexosetransport protein in the membranes of blood cells canalter glucose metabolism and thus may affect theactivity and viability of cells (14).

Recently, a new hypothesis for the mode of actionof gold drugs has been put forward. Au (I) may occupythe cysteine-rich metal binding site in transcriptionfactors containing the zinc finger motif, thus inhibitingbinding to their specific DNA response elements inpromoter and enhancer regions of genes (15).

Aurothiomalate may be activated throughinteraction with cyanide which is produced byactivated polymorphonuclear leukocytes (16). Cyanidecan readily replace the thiomalare ligand fromaurothiornalate in vitro (I 7) and aurocyanide[Au(CN)z]- has been shown to be a metabolite of golddrugs in vivo (9). Interestingly, the blood of smokerscontains a higher level of thiocyanide, the precursor ofcyanide, and the uptake of gold by red blood cells isalso enhanced in smokers (18). Aurocyanide may bethe active metabolite of gold drugs since it inhibitsvarious functions (e.g, the oxidative burst) ofpolymorphonuclear leukocytes and other cells, andthese cellular interactions may lead to the therapeuticeffects (3). Oxygen radical production (eg superoxide,O 2- by cells during the respiratory burst is essential to

the immune response towards invading bacteria, butexcessive production of these radicals has been linkedto the chronic inflammation encountered in RA.

Gold compounds may act by suppressing theimmunological processes that underlie the chronicinflammation of RA (19). Aurothiomalate (20) andauranofin (19) can inhibit the functional capacity ofmonocytes to induce T lymphocyte proliferation invitro. Both monocytes and T lymphocytes are whiteblood cells performing different functions in the

Oligomer

Aurocyan Ide

Aurothioglucose

[NC-AU-CNr

Hellamerlc structure

Auranofill

r--- -

o~ /O'Na'~C

IAu-S-C-H

Ir-oc~o~ 'O"Na'

'-- _ n

Sodium aurot/1'omalate

The mechanism of action of gold drugs is stillunknown. However, various models for theeffectiveness of chrysotherapy have been put forward.

Other possible uses of gold compounds in therapyhave been suggested. Apart from its antiarthriticproperties, aurothiomalate has been shown to haveanti-leishmanial activity in hamsters(6). More recently,auranofin has been proposed for the treatment ofpsoriasis (7). Bis(thioglucose)gold(I) has been foundto inhibit the replication of one strain of the HIV virusin vitro (8). The inhibition of the virus was suggestedto be due to the interaction of gold(I) with a cysteineresidue in a surface protein of the viral envelope (8). Itwas shown separately that bis(thioglucose)gold(I)inhibits solubilized reverse transcriptase in vitro (8). Ametabolite of all gold drugs is dicyanogold(I) (9),which has been proposed for the treatment of AIDSdue to its ability to penetrate cells rapidly and itssuggested low toxicity (10).

PROPOSED MECHANISMS OFACTION OF GOLD DRUGS

OTHER POTENTIAL USES OFGOLD IN THERAPY

88 (69' Gold Bulletin 1996,29(3)

C""",,­pel,lbide

" --­,,"ullll)-,(

Protein--­llfI_

b)

Stimulation of new set of T cells->side-effects

Inhibition of specific T cells-> therapeutic elTect

a)

APe =Anli.sen-presenling cellMHC = Map histocompatibility protein

Figure 2a Proposed MHC class 11 -peptide- gold- T cell

immune system. More recently, it has been shown thatAuCI 4- can inhibit the peptide-dependent proliferationof a T cell clone in vitro in a concentration dependentmanner (21). The hypothesis that gold has the abilityto alter the MHC-class Il-peptidet complex that isrecognised by T cells has the advantage of providing anexplanation for both the therapeutic and the toxiceffects of chrysotherapy. Modification of possible'autoimmune peptide-MHC complexes' by gold wouldinhibit the stimulation of autoaggressive T cells and, inturn, elicit gold-specific T cells that could beresponsible for the delayed type hypersensitivity side­effects (22) (Figure 2a). It has been suggested that goldbinding to the MHC class II protein rather than to thebound peptide inhibits Tcell proliferation in contrastto nickel(II) which, it has been suggested, interactswith the peptide (22). However, it has recently beenshown that aurothiornalate inhibits T cell recognitionof peptides containing two or more cysteine residues,possibly due to the formation of chelate complexes(23). Gold-specific T cell clones that proliferate whenexposed to gold(l) and gold(III) compounds in vitro

[t] In the cell mediated immune response, foreign proteins areprocessed in antigen-presenting cells such as monocyres byproteolytic enzymes. A resultant peptide is presented on thesurface of the cell bound to a membrane protein, the so-calledmajor histocompatibility complex (MHC) protein. Thecomplex formed by peptide and MHC protein is recognized byT cells. There are two main types of T cells recognizingdifferent MHC proteins: cytotoxic (CDS+) i' cells recognizeMHC class I proteins, while helper (CD4+) T cells recognizeMHC II Class proteins, T helper cells stimulate inflammatoryand antibody responses by releasing cyrokines and by directintercellular contact with B cells. Rheumatoid arthritis as wellas delayed-type hypersensitivity are linked to ;: helper cellswhich recognize peptides bound to MHC class II proteins.

Figure 2b Concept ofthescission ofdifferent peptides from aproteinalone or reacted with Au(11J), adaptedfrom[24J

have been isolated from patients who have developedsuch hypersensitivity reactions (2I). Two recent reviewssummarise the current view of the effect of gold on Tcell recognition (22, 24).

INTERACTION OF GOLD (I) WITHALBUMIN

The chemistry (25-27), biochemistry (28, 29),pharmacology (30) and the more clinical aspects (3) ofgold drugs have been reviewed. The major binding sitefor gold drugs in plasma is the free thiol group ofcysteine-34 of human serum albumin (28, 29).Evidence for an albumin-gold-glutathione complex asmetabolite of auranofin has been reported (31).

Albumin is the major protein in blood plasma,concentration ca 0.63 mM. It consists of a singlepolypeptide chain of 585 amino acids arranged inthree largely helical domains folded into a globularheart-shaped molecule. There are 6 disulfide bridgesin each domain, except domain I which is missing acysteine residue and in consequence has a free thiol atcysteine in sequence position 34. Binding of auranofinand derivatives to the free thiolate group of cysteine-34of albumin induces a structural transition, the effect ofwhich can be observed on lH-NMR resonances ofhistidine-3 of the protein (32, 33). Such acommunication between histidine-3 and cysteine-34may be explained by small changes in the arrangementof intervening helices 1 and 2 of domain I of theprotein. This may be mediated by a cis-transisomerization of proline-35, changing the environmentof cysteine-34 from a buried to an exposed one

(ei9' GoldBulletin 1996, 29(3) 89

TOXICITY OF GOLD DRUGS

Figure 3 Gold-induced ']lip-out" ofCys-34ofalbumin,adaptedfrom f35i

INTERACTION OF GOLD(III) WITHPEPTIDES

These findings recently led us to examine theinteraction of gold(III) with peptides. Mostinvestigations of the biological chemistry of gold have,for obvious reasons, focused on gold(I) and itsinteraction with naturally occurring thiols.Investigations concerning the biological chemistry ofgold(III) are scant despite the use of AuCV as a heavyatom label in protein X-ray crystallography (41).

Only a few investigations have been carried out onreactions of gold(III) with amino acids and peptidesand most of these have focused on the sulfur­containing amino acids cysteine (a thiol) andmethionine (a thioether). Gold(IlI) can oxidizedisulfide bridges in albumin (42), and the reaction oftetrachloroaurate with cysteine (RSH) follows theequation (43, 44):

with Au(III) results in T cell sensitization in mice toso-called 'cryptic peprides'j derived from this proteinwhile T cells from mice injected with RNase alone aresensitized only to a common, the immunodominant,peptide derived from RNase (38). The scission andpresentation of cryptic peptides may arise fromgold(III)-induced oxidation of methionine residuesand/or gold(III)-induced conformational changes inthe antigenic protein leading to a different intracellularprocessing of the protein and therefore to theformation of different antigenic peptides. The novelaspect of these data is that the metal is not involved inthe recognition process between MHC class II protein,peptide and T cell, but affects the prior processing stepof a protein altered by the effect of the metal (gold)(38), (Figure 2b).

Strong oxidants are required to convert gold(I) intogold(III) but they are potentially available in vivo ininflammatory situations. For example rnyelo­peroxidase, in the presence of H 20 2 and Cl, canoxidize aurothiomalate in vitro via production of ClO­[39]. Hypochlorite (CIO-) is synthesized by theenzyme myeloperoxidase of phagocytic cells fromH 20 2 and Cl" during the oxidative burst, and canoxidize gold(I) in aurothiomalare, aurothioglucose andauranofin to gold(III) (40).

S__ Au.PEI~ICys34

"''''"I J.N~

Cys34 Exposed

Auranolin..

C)'s34 Buried

It has recently been suggested that gold(III) producedfrom gold(I) drugs might be involved in the toxic side­effects encountered in chrysotherapy. The chronictreatment of mice with gold(I) drugs results in theproduction ofT cells that are stimulated not by gold(I)but by gold(III) (36) and patients with gold-induceddermatitis show significant lymphocyte proliferation inresponse to gold(III) but not to gold(I) [37]. Reactionof the model antigen bovine ribonuclease (RNase) A

(Figure 3) (33). This structural transition induced bygold(I) binding may have important consequences insignalling degradation of albumin and also in copperhomeostasis in patients undergoing chrysotherapy, ashistidine-3 is involved in the N-terminal copperbinding site of albumin. Kinetic investigations of thereaction of albumin with auranofin show that thereaction is first order with respect to albumin but zeroorder with respect to auranofin (34), findings that areconsistent with crevice-opening and exposure ofcysteine-34 being the rate-limiting step (35). The rateof reaction of auranofin with albumin is fast implyingthat the intact drug has only a short half-life in plasma(a few seconds) (34). The structural transition ofalbumin induced by gold drugs has also been observedby NMR studies of intact blood plasma (35). Goldmodulation of cysteine-34 could be important intherapy since albumin catabolism in RA patients isknown to be excessively high.

[:j:J Immunodominant peptides are those peptides of a proteinantigen that dominate the T cell response to this antigen. Incontrast, cryptic peptides are peptides that do not elicit a T cellresponse under normal conditions, but induce a response whentheir presentation is up-regulated for a particular reason (38).

NaAuCl4 + 3 RSH - RSAu + RSSR + 3 HCI + NaCI

Cystine (RSSR) can reduce KAuBr4 with theformation of metallic gold (45):

90 @ GoldBulletin 19%, 29(3)

- -+ -+

Figure 4 Structures a/some crystallizedgold(III)-peptide complexes:[Au(Gfy-L-His-H)CI]Cf-3H20, [Au(Gfy-L-His-HJI/ JOH20 [49] and [Au(Gfy-Gfy-L-His-H-JiCf-H20 [50]

3 RSSR + 10 KAuBrq + 18 H 20 - 6 RS03­

+ 10 AuG + 40 Br" + 10 K+ + 36 H'

The reaction is thought to proceed via immediate S-Sbond scission to form the sulfenic acid which is furtheroxidized to the sulfonic acid RS03H.

Insulin, which consists of two peptide chains (Aand B) joined by two disulfide bridges, can be oxidizedby tetrabromoaurate (42), and the oxidation ofdisulfides with consequent disruption of the secondaryand tertiary structures of proteins by gold (III)compounds could play an important role in thetoxicity of gold(III) compounds, hence precludingtheir use in chrysotherapy (45).

The amino acid methionine and methionineresidues in proteins can also be oxidized by gold(III)(46, 47) and the reactions are fast and stereospecific.The addition of tetrachloroaurate to native ribonucleasecauses the formation of aggregates. However, when theprotein is partially unfolded by lowering the pH, theaddition of tetrachloroaurate causes oxidation ofmethionine residues to methionine sulfoxide (48).

Little is known about the interaction of gold(III)with other amino acids in peptides and proteins. Onlytwo crystal structures of gold(III) complexes ofpeptides have previously been reported, bothcontaining the dipeptide glycyl-L-histidine (Gly-L­His): [Au(Gly-L-His-H_ t)Cl]Cl'3H20 and [Au(Gly-L­His-R3)k 10H20 [49]. The former was crystallizedfrom a 1:1 reaction mixture at low pH (1.5-2) and is asquare-planar complex in which Au(III) binds to theimidazole 8N of His, one deprotonated peptidenitrogen and the amino group of Gly1, with a chloride

~ GoldBulletin 1996, 29(3)

ion in the fourth coordination site. When these crystalswere dissolved at higher pH (6-7), a cyclic tetrarnerwith the metal bridging between the 8N and EN atomsof two different Gly-L-His molecules was obtained(Figure 4) (49). We have recently prepared the four­coordinate complex of gold(III) with the tripeptideGly-Gly-L-His (Figure 4) (50), the first Au(III)­tripeptide complex to be reported. This forms readilyvia one intermediate in strongly acidic solutions (pH1.5).

We have recently examined potential binding sitesfor gold (III) on pep tides using the monodentatecompound [Au(dien)Cl]CI2• In the simple tripeptideglycylglycylglycine, the amino group is the onlybinding site, with no evidence found for aninvolvement of the carboxyl group (51). In histidine­containing peptides, the binding sites for gold(III)include the Hisf and E nitrogens and the Nvrerminalamino group. Species in which the imidazole ring actsas a bridging ligand between two gold(III) moieties arefound at pH values as low as 5 (51).

Gold (III) is one of the strongest inducers ofacross-the-ring ionization of HisENH when bound toHis8N of an imidazole ring of histidine. It lowers thepKl of the so-called 'pyrrole nitrogen' (52) from above14 in free histidine to about 9, when gold(III) isbound to His8N, the so-called 'pyridine nitrogen'(50). Comparable values have been reported only forRu(III) with a pKa value of 8.7 (53), Co (III) with apK, value of 9.6 (54) and Hg(II) with a pKa value of9.6 also (55). For comparison, copper(II)complexation lowers the pK, value of the pyrrolenitrogen to values of about 11 (52) and palladium(II)to 10.8 (56) and 11.3 (50).

91

It is clear that gold(III) may undergo redox reactionswith pep tides and proteins, particularly involving sulfuramino acids, but also has an unusual ability amongstmetal ions to deprotonate and bind to peptide amidebonds and cross-link histidine imidazole rings.Interactions with other amino acid side-chains such astyrosine can also be envisaged, but these have not yetbeen investigated. In view of the possible involvementof Au(III) in the immunological side-effects of goldtherapy, further work on gold(III)-peptide and -proteininteractions may help to improve our understanding ofthe immunochemistry ofgold drugs.

CONCLUSIONS

It is possible that gold drugs act upon different stagesof rheumatoid arthritis, but a final statement mustawait further progress into the aetiology of thiscomplex disease. The first action might be on thecentral recognition process of T helper cells (whiteblood cells) which recognize a disease-inducing(auroimmune') peptide bound to an MHC class IIprotein on the surface of immunological cells. As RA isnow generally considered to be an auto-immunedisease, this might be a peptide derived from self­proteins. An inhibition of the disease-triggering T cellresponse by gold might be its first effect. Thisinhibition might be due to interaction of gold(I) orgold(III) with MHC class II protein or bound peptide.Toxic side-effects might in turn be elicited by this samemechanism or might be due to the expression ofcryptic peptides derived from a protein altered by gold.On a second level, the inhibition of the oxidative burstof polymorphonuclear lymphocytes by gold mightalleviate secondary effects of the chronic inflammationin the affected joints. The active metabolite responsiblefor such inhibition appears to be [Au(CN)2]- which isa universal metabolite formed from both injectable andoral gold drugs and may play a key role in thepharmacology of gold.

ABOUT THE AUTHORS

Sabine Best has recently completed her PhD atBirkbeck College, University of London on thebiological chemistry of gold, and is currently a researchfellow in the School of Biological Sciences at QueenMary and Westfield College, London. Peter Sadler hasjust taken up the Crum Brown Chair of Chemistry atthe University of Edinburgh after 23 years at BirkbeckCollege. His major research interests lie in thechemistry of metals in medicine.

92

REFERENCES

1 P. J. Sadler, Adv. Inorg. Cbem., 1991,36, 12 G. Sava, in 'Metal Compounds in Cancer

Therapy', ed. S.P. Fricker, Chapman and Hall,London, 1994, p.65

3 G.D. Champion, G.G. Graham and J.B. Ziegler,Baillieres Clin. Rbeumatol., 1990,4,491

4 L.Treagan, Metal Ions in Bioi. Syst., 1983, 16,275 1. Schroter and J. Strahle, Chern. Ber., 1991, 124,

21616 J.L. Abruzzo and R.J. DeHoratius, Am. I. Trop.

Med. Hyg., 1969, 18,5067 R.E. Thomas and R.A. Papandrea, Med. I.

At~tra!., 1993, 158, 7208 T. Okada, B.K Patterson, S.-Q. Ye and M.E.

Gurney, Virology, 1993, 192,6319 R.e. Elder, Z. Zhao, Y. Zhang, J.G. Dorsey,

E.V Hess and KJ. Tepperman, I. Rbeumatol.,1993,20,268

10 Y.F. Zhang, E.V Hess, KG. Pryhuber, J.G.Dorsey, K Tepperman and R.e. Elder, Inorg.Chim.Ac~, 1995,229,271

11 G.W. Rafter']' Inorg. Biochem., 1990,38,26112 e.W Mahoney, c.r. Hensey, and A. Azzi,

Biocbem. Pbarmacol., 1989,38,338313 T. Herlin, K Fogh, N.O. Christiansen and K

Kragballe,AgentsandActions, 1989,28,12114 P. E. McGowan, J.M. Campbell, J. Reglinski and

W.E. Smith, Inorg. Cbim. Acta, 1993,210, 115 M.L. Handel, A. deFazio, e.KW Watts, R.O.

Day and R.L. Sutherland, Mo!. Pbarmacol., 1991,40,613

16 R. Rudkowski, G.G. Graham, G.D. Championand J.B. Ziegler, Biocbem. Pbarmacol., 1990, 39,1687

17 G.G. Graham, J.R. Bales, M.e. Groorveld andP.J. Sadler,] Inorg. Biocbem., 1985,25, 163

18 G.G. Graham, T.M. Haavisto, P.J. McNaught,c.o. Browne, and G.D. Champion, I­Rbeumatol., 1982,9,527

19 G. Salmeron and P.E. Lipsky,] Rbeumatol.Suppl., 1982,8,25

20 P.E. Lipsky and M. Ziff,] Clin. Invest., 1977,59,455

21 P. Romagnoli, G.A. Spinas and F. Sinigaglia, J.Cun. Inv~t, 1992,89,254

22 F. Sinigaglia,] Invest. Dermatol., 1994, 102,39823 P. Griem, K Takahashi, H. Kalbacher and E.

Gleichmann,] Immunol., 1995,155, 157524 P. Griem and E. Gleichmann, Curro Opin.

Immunol., 1995,7,831

~ GoldBulletin 1996,29(3)

25 P.]. Sadler, Gold. Bull., 1976,9,11026 D.H. Brown and WE. Smith, Chem. Soc. Rev.,

1980,9,21727 P.]. Sadler and R.E. Sue, Metal-Based Drttgf,

1994,1, 10728 c.r. Shaw III, Inorg. Perspect. BioI. Med., 1979, 2,

28729 c.F. Shaw III, Comments Inorg. Cbem., 1989, 8,

23330 A]. Lewis and D.T. Walz, Prog. Med. Chem.,

1982, 19, 131 c.r, Shaw III, AA Isab, M.T. Coffer and C.

Mirabelli, Biochem. Pbarmacol., 1990, 40, 122732 a.M. Ni Dhubhghaill, P.]. Sadler and A Tucker,

[c Am. Chem. Soc., 1992, 114,111833 ]. Christodoulou, P.]. Sadler and A Tucker, Eur.

1 Biocbem., 1994, 225, 36334 ].R. Roberts,]. Xiao, B. Schliesman, D.].

Parsons and c.F. Shaw III, Inorg. Chem., 1996,35,424

35 ]. Christodoulou, P.]. Sadler and A. Tucker,FEBS Letters, 1995,376,1

36 D. Schuhmann, M. Kubickamuranyi, ].Mirtschewa, ]. Gunther, P. Kind and E.Gleichmann,j Immunol., 1990, 145,2132

37 ]. Verwilghen, G.H. Kingsley, L. Gambling andG.S. Panayi, Arthritis Rheum., 1992,35, 1413

38 P. Griem, K. Panthel, H. Kalbacher and E.Gleichmann, Eut: 1 Immunol., 1996,26,279

39 B. Beverly and D. Couri, Fed. Proc., 1987, 46,854

40 c.r, Shaw III, S. Schraa, E. Gleichmann, Y.P.Grover, L. Dunemann and A ]agarlamudi, MetalBased Drugs, 1994,1,351

§9' GoldBulletin 1996, 29(3)

41 T.L. Blundell and L.N. Johnson, 'ProteinCrystallography', Academic Press, New York, 1976

42 P.L. Witkiewicz and c.r. Shaw Ill, Chem.Commun., 1981, 1111

43 D.H. Brown and WE. Smith, Am. Chem. Soc.Symp. Se~, 1983,209,401

44 WE. Smith and ]. Reglinski, Perspec. Bioinorg.Chem., 1991, 1, 183

45 c.n Shaw III, M.P. Cancro, P.L. Witkiewicz and].E. Eldridge, Inorg. Chem., 1980, 19,3198

46 E. Bordignon, L. Cattalini, G. Natile and AScatturin, Chem. Commun., 1973, 878

47 G. Natile, E. Bordignon and L. Cattalini, Inorg.Chem., 1976,15,246

48 AA Isab and P.]. Sadler, Biochim. Biophys. Acta,1977,492,322

49 M. Wienken, B. Lippert, E. Zangrando and L.Randaccio, Inorg. Chem., 1992,31,1983

50 S.L. Best., T.K. Chattopadhyay, M.I. Djuran,R.A Palmer, P.]. Sadler, I. S6vago and K.Varnagy, in preparation

51 S.L. Best, M.1. Djuran and P.]. Sadler, III

preparation52 R.]. Sundberg and R.B. Martin, Chem. Reu.,

1974,74,47153 R.]. Sundberg and G. Gupta, Bioinorg. Chem.,

1973,3,3954 W]. Eilbeck and M.S. West, 1 Chem. Soc.,

Dalton Trans., 1976,27455 C.A Evans, D.L. Rabenstein, G. Geier and I.W

Erni,j Am. Chem. Soc., 1977,99,810656 T.P. Pitner, E.W. Wilson and R.B. Martin, Inorg.

Chem., 1972, 11, 738

93