1022 J.C.S. DaltonComparison of the Ligating Properties of Disulphides and Thioethers :Dimethyl Disulphide, Dimethyl Sulphide, and Related LigandsBy Helmut Sigel." Kurt H. Scheller, Volker M. Rheinberger, and Beda E. Fischer, Institute of InorganicChemistry, University of Basel, Spitalstrasse 51, CH-4056 Basel, SwitzerlandThe stability constants of 1 : 1 complexes between dimethyl disulphide (dmds) or dimethyl thioether (dms) andCa2+, Zn2+, Cd2+, Pb2+, or Ag+ (KML for Mn+ + L + [ML]"+) have been determined in aqueous solution by lHn.m.r. shift measurements. The results [e.g. log KCa({llnds) ca. -1.4, log &d(d,&) ca. -1.4, log KAg(dtnd@) = 2.01 f 0.09 ;log KCa(dma) ca. -1.6, log KCd(,,lne) = -0.3 f 0.2, and log KAy(dnls) = 3.7 f 0.31 show that the complexes with softmetal ions are stronger than those with borderline or hard metal ions.It is also evident that the ligating propertiesof the thioether moiety are somewhat more pronounced towards borderline and soft metal ions than those of thedisulphide group. Spectrophotometrically determined stability constants (in 50% aqueous ethanol) for the com-plexes between Mn2+, Cu2+, or one of the above metal ions and dmds, diethyl sulphide, or tetrahydrothiophenaccord with this. In addition, for several complexes of tetrahydrothiophen-2-carboxylate (thtc-) and 1,2-dithioI-an-3-carboxylate [ = tetranorlipoate (tnl-)], the dimensionless constants for the intramolecular equilibriumbetween the chelated isomer (which is bonded to the metal by the sulphur atom and the carboxylate group) and thesimple carboxylate-co-ordinated isomer have been calculated, together with the percentages of the chelatedisomer (e.g.93 f 1 for [Cu(thtc)]+ and 41 * 7 for [Cu(tnl)]+ ; 55 f 5 for [Cd(thtc)]+ and <20% for [Cd(tnl)]+;<20% for both [Mn(thtc)]+ and [Mn(tnl)]+}. Possible biological implications of such weak interactionsand the resulting intramolecular equilibria are briefly discussed.THE disulphide group is well recognized as a potentialbinding site for metal ions in biological systems. Thisgroup is found in cystine,l an important constituent ofmany proteins, and in the oxidized form of glutathione,2as well as in lipoic acid,3 which participates in manyenzymatic reactions, especially as a protein-boundcoenzyme in oxidative decarbo~ylations.~The interactions between the disulphide moiety andhard (e.g.Ca2+ or Mn2+) or borderline (e.g. Cu2+ or Zn2')metal ions are certainly rather weak,l but they do occur,as crystal-structure determinations reveal.6 Compari-: on; of stability constants have been used to suggest aCu2+ interaction with the disulphide bond of ~ - c y s t i n e , ~and there is spectral evidence for a possible Cu2'-disul-phide interaction in solutions with L-cystinediamidt.8A copper(r1) complex of oxidized glutathione has also beenchara~terized,~ and there is evidence that the disulphidemoiety of the 1,2-dithiolan ring of a-lipoate may partici-pate in complex formation.lO~ll However, all the deter-minations of stability constants have been performed withligands containing at least one proton-basic site inaddition to the disulphide moiety.29799-11Certainly, soft metal ions like Cu', Ag+, or Hg2+ havea greater tendency to co-ordinate with the disulphidegroup and crystal structures of disulphides, e.g.with Cut,are known,12p13 although co-ordination of these metal ionsoften causes cleavage of the disulphide bond.Ip2 Thiscleavage reaction is probably one of the reasons whypractically no stability data seem to exist for complexesbetween soft metal ions and simple disulphides, but aAg+-disulphide interaction has been suggested for thesystem Ag+-basic bovine pancreatic trypsin in1iibit0r.l~The reasons for the complete lack of stability data onthe interactions between the biologically important hardor borderline metal ions like Ca2+, Mn2+, Cu2+, or Zn2+and simple disulphides are that : (2) such complexes arevery weak,lT2 and (ii) a pH-metric determination ofstability constants is not possible, as these ligands arevery poor proton a~cept0rs.l~ In fact, the disulphidegroup is even less basic than the sulphur of thioether l6which is itself an extremely weak base with pK, =--6.8.17 Estimations of basicity suggested the orderMe,S > Me,S, _N MeSH,16 which indicates pK, < -6.8for the disulphide moiety.Hence, so far only disulphidedmds dmsdes thtinl- tMc'FIGURE 1 Structures of the sulphur ligandscomplexes with ligands containing an additional bindingsite, like an amino- or carboxylato-group, have beenFor the thioethers, the situation is similar: the bio-logical importance of the sulphur atom in these com-pounds as a potential binding site for metal ions is wellr e c o g n i ~ e d , ~ * ~ ~ ~ ~ ~ and indeed the X-ray crystal-structureanalysis 2o of plastocyanin, for example, revealed aCuz'-thioether bond with a methionine residue in this' blue ' copper protein.However, stability data haveonly very recently been published, based on spectro-photometric measurements,21 for complexes betweenstudied.L2,7,9-11023hard or borderline metal ions and simple thioetherligands like tetrahydrothiophen or diethyl sulphide.We now compare the stability of complexes formedbetween Ca2+, several 3d transition-metal ions,Zn2+, CdZ+, Ag+, or Pb2+ and the ligands shown inFigure 1.First the complexes of the simple dimethyldisulphide (dmds) were studied: the stability of. the[Cu(dmds)12+ complex was determined spectrophoto-metrically, while the other complexes with dimethyldisulphide and diamagnetic metal ions were investigatedby lH n.m.r. shift experiments. For comparison thestabilities of some complexes with dimethyl sulphide(dms) and related thioethers were also measured, and thepercentages of the chelated isomer of several M2+ 1 : 1complexes of 1,2-dithiolan-3-carboxylate and tetra-hydrothiophen-2-carboxylate were calculated.EXPERIMENTALMaterials .-The metal perchlorates (purum or puris-simum) were from Fluka AG, Buchs, Switzerland, with theexception of zinc and cadmium perchlorate which wereobtained from K and K Laboratories, Ohio, U.S.A.andfrom Ventron GmbH, Karlsruhe, Germany, respectively.The metal nitrates (pro analysi) were from Merck AG,Darmstadt, Germany, except silver nitrate which was fromB.D.H. Chemicals Ltd., Poole.Dimethyl disulphide (for synthesis) and tetrahydro-thiophen (pro analysi) were obtained from Merck-Schuchardt, Hohenbrunn near Munich, Germany ; bothligands were used in spectrophotometric measurements andtherefore were distilled twice before use over a Raschigcolumn under N, and then stored under N,. Dimethylsulphide (purum) was from Flulta AG and the solventethanol (for spectroscopic use) from Merck AG.The pH was measured with a Metrohm potentiometerE 353B using a EA 121 or a micro EA 125 Metrohm glasselectrode.Calibration was done with commercial aqueousbuffers (pH 4 and 7, from Metrohm AG, Herisau, Switzer-land) and direct readings for pH were used, except in D20solutions where the pD was obtained by adding 0.40 to thepH-meter reading.22 Unless otherwise specified the pH (orpD) of the solutions was adjusted to ca. 2.Spectrophotometric Measurements.-Absorbance spectrawere recorded with a Beckman DB spectrophotometerconnected to a Walz + Walz Electronic Hi-Speed recorder202, or 011 a Varian Techtron spectrophotometer (model635) connected to a Honeywell recorder (model 196).The experiments with the Cu2+-dimethyl disulphidesystem and their evaluations were done exactly as describedrecently 21 for Cu2+-thioether systems.The conditionsused in the Cu2+-dmds experiments are shown in Figures 2and 3.The stability constants of the cadmium(I1) and lead(n)complexes with tetrahydrothiophen (tht) were determinedfrom ' competition ' experiments with the Cu2+-tht system.This means that the absorption of the Cu2+--tht system a t359 nm decreases in the presence of Cd2+ or Pb2+, and fromthis decrease in absorption the stability of the cadmium(I1)or lead(rr) complexes may be calculated. The experimentalconditions and the calculations were identical to thosedescribed recently for several other M2+-tht systems.21Hydrogen- 1 N . M. R. Measurements.-For the Mn+-dmdssystems the lH n.m.r. spectra were recorded in D,O solu-tions with a Bruker WH-90 Fourier-transform spectrometer(90.025 MHz) at 27 "C or a Varian Anaspect EM-360spectrometer (60 MHz) at 34 "C using the centre peak of thetetramethylammonium triplet as reference ; all thesechemical shifts were converted into a trimethylsilylpropane-sulphonate reference by adding 3.188 ~ .p . m . , ~ From a plotof these chemical shifts against the increasing concentrationof the diamagnetic metal ion Mn+ (for details see Figure 4),the curve best fitting these experimental data can be com-puted, using a Hewlett-Packard model 9821A, connected toa model 9862A plotter, and log KM(dmds) thus obtained.The constants for the Mn+-dms systems were determinedin exactly the same way (Figures 4 and 5), but here allspectra were recorded in aqueous solutions with the VarianAnaspect EM-360 spectrometer (60 MHz) a t 34 "C.The mentioned curve-fit is based on equation (1) ,* where8, is the chemical shift of the free ligand (L) and 6, thechemical shift of the complex [ML]"+.Equation (1)together with the definition (2), which is valid for ourexperimental conditions where the species [ML,n+] withm >, 2 are negligible, gives equation (3). From the%bs. = 80([L1/[Lltot.) + 8m([MLl/[Llt~t.)[Lltot. = [LI + [MLI (2)%bs. = + (h - 80)([MLl/[Lltot.) (3)[Mltot. = [MI + WLI (4)KML = [MLl/[Wl [LI ( 5 )[MLI = S([MItot. + [LItot. + KML-~ -(6)definitions (2), (4), and (5) follows equation (6). Substi-tution of equation (6) in (3) gives an expression whichcontains only two unknown parameters, 6, and KMI,, andthese may now be determined by starting the iterativecalculation with estimated values, and varying these untilthe standard deviation reaches a minimum (least-squaresregression).In cases where 8, cannot be determined with asignificantly greater accuracy than the other experimentalpoints, 8, may also be treated as a variable parameter.All experiments were carried out a t least twice. Theerrors given throughout this work are twice the standarddeviation, unless stated otherwise.{([Wtot. + [LItot. + KML-?~ - 4[MItot.[LI,,t.P)RESULTSAs the stability of most of the complexes to be determinedwas rather small, perchlorate salts were used wheneverpossible. Only in the few cases where these were notavailable or where the perchlorate salts are relativelyinsoluble, as with Ag[ClOJ, were the nitrates used.The[ClOJ ion does not affect the stability constants since thereis no association between this anion and the metal ions used(or an extremely small one),25926 while with [NO,]- very weakcomplexes may form in some case^.^^^^' However, therecent results 21 for M2+-thioether systems show that theinfluence of [NO3]-, if present a t all, is very small indeed.Mn+ + L + [ML]n+ (7)The dmds-Mn+ Systems.-For the Cu2+-dmds system, thestability constant KML of the 1 : 1 complex [according toequilibrium (7)] could be determined spectrophotometricallyin 50% aqueous ethanol (I = 1.0 mol dmP3, Na(lC1OJ ; 25 "C)1024 J.C.S. DaltonFigure 2 shows how the absorption ( A ) of dmds decreaseswith increasing concentration of Cu2+.Since there are twoopposing factors affecting the experimental accuracy, i.e.increase of the concentration of Cu2+ leads t o larger spectrali \ IAlnmFIGURE 2 Absorption spectra (--) of dimethyl disulphide(2 x rnol dm-3) alone (1) and in the presence of: 0.075 (2),0.15 (3), and 0.30 rnol dm-3 Cu[CIO,], (4); the cell in thereference beam always contained the appropriate concentrationof Cu[ClO,],. The spectra were measured in 50% aqueousethanol (v/v) a t pH ca. 3, I = 1.0 rnol dm-3 (Na[CIO,]), and 25"C in 1-cm quartz cells. (a *), Uncertainty due to the largeabsorption of Cu[C104],; (- - - -) , spectra (under the aboveconditions) of 0.02 (5) and 0.30 rnol dm-3 Cu[CIO,], (6) solutionsdifferences while simultaneously the difference spectrabecome less accurate due to the increasing absorption of theCu2+ solution in the reference beam, the spectra wereevaluated a t three different wavelengths (270, 278, and 282nm).z1~2s An example of this is shown in Figure 3 where1/[Cu2+Itot.is plotted against l/AA ; the resulting straightlines confirm that 1 : 1 complexes are formed and that theirintercepts with the y axis are identical within experimentalM"+Ca2+Mn2+cu2+Zn2+Cd2+Pb2+Agierror. The average result of three independent experi-ments for the stability constant of [Cu(drnds)l2+ is logThe stability constants of other [M(dmds)12+ complexescould not be measured by the ' competition method ' usedearlier,,l since the spectral alterations were too poor to beexactly reproduceable.We therefore carried out lH n.m.r.KCu(dmds) == 0.49 f 0.22.-5Y 0 5 10 15 20 25 30 35 401 IAAFIGURE 3 Graphical determination of the stability constant,KC^(^^^^) (dm3 mol-l), of the Cu2+-dmds 1 : 1 complex in 50%aqueous ethanol (v/v) by plotting 1/[Cu2+]tot. against l/AA ;[dmds] = 2 x 10-3 rnol dm-3, I = 1.0 rnol dm-3 (Na[ClO,]);25 "C, and pH cu. 3. For concentrations of Cu[ClO,], 0.02-0.05 mol dm-3 the evaluation was done a t 270 nm (e), for0.04-0.15 rnol dm-3 at 278 nm ( O ) , and for 0.075-0.30 rnol~ l m - ~ at 282 nm (A), the intercepts with the y axis [-- K c ~ ( ~ ~ ~ ~ J aye -3.82 f 0.92, -4.96 f 1.25, and -2.594 1.13 respectively. The straight lincs were drawn accordingto the least-squares regression methodexperiments with some diamagnetic metal ions and tneasuredthe chemical shift of the methyl protons of dmds in D,Osolutions; under the influence of increasing amounts ofTABLE 1Logarithms of the stability constants K M ~ of 1 : 1 complexes of MnC with dmds, dms, des, or t h tdmdsAI dmsD,O 50% EtOH2 H,O- 1.4 e9-J-g - 1.6 8-9des t h t500/, EtOH" 50% EtOH-0.30 f 0.13 ' y i-0.31 f 0.11 ' 9 '.0.47 f 0.19 h , i 0.02 f 0.04 h*d-0.21 * 0.09 hpf-0.26 f 0.060.08 f 0.053.51 f 0.14 i , la Unless stated otherwise, the errors given are twice the standard error of the mean value.I - 0.1 rnol Na[NO,]. c 34 "C.d 25 "C. f The errorlimits for these results are estimated as follows: to see how the results depend on A8 (taken from the Ag+ systems) the calculationswere repeated for dmds complexes with A8 = 0.200 and 0.300 p.p.m.(2.e. A8 = 0.249 f 0.050 p.p.m.); the results are within i O . 1log units, as given above, but t o be on the safe side we allowed +0.2 log units for this error range (if A8 should be out by a factor of0.5 or of 2 the values of log KML would be out by + 0.3 or - 0.3 log units respectively). Furthermore, since the small shifts observedin the presence of M2+ might partly originate from the change in ionic strength from 2 to 4.5 (2.e. the difference calculated between6, at I = 2 rnol dm-, and the shift in the presence of M2+ a t I = 4.5 rnol dm-3 might be slightly too large, as the correspondingexperiments indicate), although certainly not more than by a factor of 4, the complexes might be up to 0.3 log units less stable thancalculated.g 1 = 2.0-4.5rnol dm-3, Na[ClO,]. 27 "C.1 I = 1.0 mol dm-3, Na[NO,].Estimated values with A8 = 0.249 and 0.334 p.p.m. for the dmds and dms systems respectively (see text).Hence, the lower error limit for the values given above is -0.5 and the upper limit +0.2 log units.h I = 1.0 mol dm-3, Na[CIO,]. Value taken from ref. 21. j The error range is estimated (cf. text).m I = 2.0-3.0 mol dm-l Na[NO,]1025Mn+ the signals are shifted downfield. These cheniical-shift differences may be used to obtain the stability con-stants of the [M(dmds)]n+ complexes.An example is shown in Figure 4 for the Ag+-dmdssystem ( I = 0.1 mol drn-,, Na[NO,]; 34 "C), where asignificant downfield shift is observed.The measured2.100 0.05 0.10Ag* 1 lot! mol dm-3FIGURE 4 Chemical shifts of the resonance of the methyl protonsof dimethyl disulphide (a) and of dimethyl sulphide (b) (each 5.7x 10P mol dm-3) as a function of [Ag(NO,)]t,t. The measure-ments ( I =- 0.1 rnol dm-3, Na[NO,]; 34 "C) were carried out inL>,O at pD ca. 2 for dmds and in H,O a t pH ca. 2 for dms.The curvcs shown arc the computer-calculated best fits of theexperimental data: KAg(dmdr) = 93.8 f 29.7 dm3 Inol-l,8, = 2.463 & 0.015 p.p.m., and 803 = 2.708 f 0.016 p.p.m.;KAg(dms) = 2 835 f 1 066 dm3 molP, 6, = 2.125 p.p.m., andSo0 = 8.461 1. 0.006 p.p.m.shifts were plotted against the increasing concentration ofAg+ and the curve best fitting the experimental data wascomputed (see Experimental section). The resultingstability constant (the average of two experiments) of[Ag(dm<ls)] ' is log KAg(dn,ds) = 2.01 f 0.09, the chemicalshifts of free dmds and [Ag(dmds)]+ being 8, = 2.461 f0.010 and 8, = 2.710 f 0.011 p.p.m., respectively.In these Ag'-dnids experiments we adjusted the pD toca.2 with DXO, to prevent cleavage of the disulphide bond,a reaction which occurs a t higher pH.,O In fact, a Ag+-dmds solution stored for 24 h gave the same lH n.m.r.spectrum as that of the freshly prepared solution. Attemptsto study the Hg2+-dmds system failed as a white precipitateformed; this could be the nitrate salt of a [Hg(dmds),12+complex or the result of fission of the disulphide b0nd.~0Furthermore, HgZ2+ seemed to undergo disproportionationin the presence of dmds, in accordance with related observa-tions.31 Hence, no mercury-dmds system could be studied.Under the influence of increasing amounts of Ca[C10,],,Zn[C10,12, C~l[C10,]~, or Pb[N0312 in D20 the signals ofdmds are slightly shifted downfield indicating weak inter-actions, but the shift differences are far too small to allow acurve-fitting procedure ( I = 2-4.5, Na[ClOJ, or 2-3mol dm-,, Na[NO,] ; 27 or 34 "C).Therefore, the assump-tion was made that A8 = 8, - 8, would be about the samefor these systems as for the Ag+-dmds system.* Based onA8 = 0.249 p.p.m. we estimated the stability constants ofthese [M(dmds)12+ complexes; the results with theirapproximate error limits are summarized in Table 1 togetherwith the stability data of related thioether complexes.Itis worth noting that an increase in A8 leads to a slightdecrease in KM(dmds), and vice versa.Thioether-Mn+ Systems.-Since the stability of the[Ag(dmds)]+ complex is significantly lower than that of theAg+-thioether system,21. 32 and the same behaviour isshown by the corresponding complexes 21 with e.g. Cd2+, wedecided to study for comparison a thioether ligand undersimilar conditions. Hence, the chemical shifts observed fordimethyl sulphide (dins) in aqueous solution and in thepresence of Ag+, Ca2+, Zn2+, Cd2+, or Pb2+ were measured.Again, Ag+ (I = 0.1 niol dmP3, Na[N03]; 34 "C) producedsignificant downfield shifts (Figure 4); in fact these werealready so large at low concentrations that the stabilityconstant could only be estimated, but the result log Kgg(d,s)3.7 f 0.3 is in excellent agreement with related data,21especially with the Ag+-2,2'-thiodiethanol (tde) systemstudied by Widmer and Schwarzenbach: 32 log KAg(tde) =3.60 f 0.03 ( I = 1.0 mol dmP3, K[NO,]; 20 "C); inaddition the value log' KAg(tde), = 2.46 f 0.05 for[Ag(tde)]+ + tde + [Ag(tde),]+ was determined.32As with dmds, the downfield shifts with dms in thepresence of Ca2+, Zn2+ (I = 2-4.5 mol dm-3, Na[C10,];34 "C), or Pb2+ ( I = 2-3 mol drn-,, Na[N03]; 34 "C) weretoo small for the curve-fitting procedure, but with A8 =0.334 p.p.m.1 the stability constants for the corresponding[M(dms)12+ complexes could again be estimated in the waydescribed for the M2+-dnids systems.With the Cd2+-dmssystem ( I = 2-4.5 mol Na[C10,]; 34 "C) the situationis more complicated : here significant shifts were observed(Figure 5), but the plot of 8 against [Cd2+], which is re-produceable, does not show the expected behaviour ; onlythe measurements a t [Cd2+] < 0.7 niol dm-3 can be fittedby equations (3) and (6) in a satisfactory way, while themeasurements at [Cd2+] 2 0.75 mol dmP3 deviate sig-nificantly from this calculated curve, which correspondsto log Kcd(dms) = -0.32. However, evaluation of themeasurements at large Cd2+ concentrations, in the waydescribed for Ca2+-, Zn2+-, or Pb2+--dnis, gives a verysimilar result: log KCd(dme) = -0.33. Hence, despite theaforementioned difficulties we conclude that log Kcd(drns, =-0.3 & 0.2 and that this is a reasonable estimate for thestability of [C~l(dms)]~'; moreover, if all experimental datashown in Figure 5 are forcefully fitted to equations (3) and(6) one still obtains a constant which is within the givenerror limits. The only way we see a t present to explainthe deviations indicated in Figure 5 would be to postulate aspecies [Cd,(dms)]*+ (which would formally correspond to asulphone, i.e.02SR,, while [Cd(dnis)I2+ corresponds to asulphoxide OSR,}, but without further evidence, whichseems difficult to obtain, we are very reluctant to do so,* To a first approximation this assumption is certainly justified(see also footnote c in Table 1) and is confirmed, e.g.by thesimilar changes observed in the chemical shifts of amino-acidsunder the influence of protonation and Zn2+ complexation (B. E.Fischer and H. Sigel, J . Amev. Chem. Soc., 1980, 102, in the press). t This value corresponds to the Ag+-dms system where 6,= 2.125 p.p.m. (average value) and 803 = 2.459 f 0.003 p.p.m1026 J.C.S. Daltonalthough it must be said that pchloro- or p-hydroxo-complexes, for example, are well known.All the results of the Mn+-dms systems are summarised inTable 1, together with the stability constants of severalMn+-tetrahydrothiophen systems. These latter systemswere studied by spectrophotometry (see ExperimentalI . . . . . . . , . . . .0 0.2 04 0.6 0.8 1-0 1.2 ' 1.4[Cd2*ltot,/md dm-3Chemical shift of the resonance of the methyl protonsof dimethyl sulphide (1.5 x mol dm-3) as a function of[Cd(ClO,),]t,t. The measurements (I = 2-4.5 mol dm-3,Na[C10,] ; 34 "C) were carried out in aqueous solution a t pHca.2. The curve shown is the computer calculated best fit ofthe experimental data a t [Cdz+] < 0.7 mol dmP3 (a) : Kcd(drns)= 0.47 5 0.25 dm3 moP, 6, = 2.143 p.p.m., and &O = 2.32f 0.08 p.p.m. For data (0) which deviate from the curve seetextsection and ref. 21) in 50% aqueous ethanol which corre-sponds to a mol fraction of 0.24 (I = 1 .O mol dm-3, Na[ClOJor Na[NO,]; 25 "C). The results demonstrate again therather high stability of the silver(1) complex compared withthe listed M2+ complexes.FIGURE 5DISCUSSIONComparison of the Stability of Disulphide and ThioetherComplexes.-A comparison of the stability constantslisted in Table 1 reveals that the ligating properties ofthe disulphide and the thioether groups appear to beabout the same for hard metal ions (e.g.Ca2+), while forseveral of the borderline metal ions (e.g. Cd2+, see refs. 1and 21, and Pb2+) they are somewhat different: thismeans the disulphide complexes are less stable than thecorresponding thioether complexes.* With the soft Ag+ion these properties are even more evident: the dif-ference in stability between the silver(1) complexes ofdimethyl disulphide and dimethyl sulphide is a factorof 50. This may indicate that the S,R, group is some-what less soft than the SR, moiety.All things considered, one may conclude that theresults of Table 1 are in accord with what one might havepredicted.Due to the formation of double bonds in adisulphide (p,-d, interaction) 4934 the electrons are lessavailable for co-ordination than at thioether ~ulphur,~* The stabilities of [Cu(dmds)12+ and [Cu(des)]z+ (des =diethyl sulphide) in 50% ethanol (see Table 1) are apparentlycontrary to this conclusion but one should remember thatdifferent solvents have different effects on the complex stability ofligands (probably due to steric restrictions in the solvated co-ordination sphere) 33 as was shown recently for Cuzf-thioethercomplexes.21 The stabilities of the copper(I1) complexes of thesterically very similar ligands 1,2-dithiolan-3-carboxylate andtetrahydrothiophen-2-carboxylate do fit into the above picture(see below and Table 2).hence complexes with S,R, are expected to be somewhatless stable than those with SR,.In line with this reason-ing is the so-called a-effect, i.e. the observation that e.g.OH- is a better base than OOH-, or NH, than NH,NH,,and indeed this is also true for SMe, and S,Me2.l6Chelate-formation Properties of Disulphide and Thio-ether Moieties.-In the light of the preceding results it isinteresting to compare the ligating properties of thedisulphide and thioether groups in ligands which arepotentially able to form chelates: this was done bycomparing the tetrahydrothiophen moiety in tetrahydro-thiophen-2-carboxylate (thtc-) with the 1,2-dithiolanresidue of 1,2-dithiolan-3-carboxylate [tetranorlipo-ate (tnl-); see Figure 11.It should be noted that thesteric properties of both ligands are practically identical.Taking into account the suggestion4,34 that the p,,-d,,interaction is more pronounced in simple aliphaticdisulphides and that this double-bond character issomewhat diminished in the 1,2-thiolan ring, one mightexpect some enhanced ligating properties for the ringsystem compared with simple aliphatic disulphi des,but this seems not to be the case. The 1,2-dithiolanmoiety is also a poorer ligating group than the tetra-hydrothiophen ring, i.e. the following results fit clearlyinto the picture already described for dimethyl di-sulphide and dimethyl thioether.For the complexes with tetrahydrothiophen-2-carboxyl-ate and 1,2-dithiolan-3-carboxylate (ML+) an equilibriumis expected between a chelated isomer in which the carb-oxylate moiety and a sulphur atom are co-ordinated tothe metal ion (depicted as MLos+) and a simple carboxyl-ate-co-ordinated isomer (MLo+).Hence, for theseisomers the intramolecular equilibrium 35936 (8a) can beMLo+ + MLop,+ (84Kch. = [MLOS] [MLO] (8b)defined. The corresponding equilibrium constant Kchelate(= Kch) is dimensionless and therefore independent ofthe absolute concentration of ML+. The simpleequilibrium (7) may now be rewritten as equilibrium (9a)M2+ + L- + (MLo+ + MLos+)KML = ([MLoI + [MLosl)/[Ml[LlIn addition a constant KMLO may be defined whichquantifies equilibrium (lOa), i.e. the reaction between(9a)(9b)MZ+ + L- + MLo+ (104KML~ = [MLol/[Ml [LI (lob)M2+ and L- for the simple carboxylate co-ordination.Substitution of (8b) and (lob) into equation (9b) givesthe relations (11) and (12).If one assumes that in aKML = KMJJo + KMLo Kch. = KMLo (1 + K&.) (11)Kch. = (KML/KML~) - 1 (12)theoretical case 20% of ML+ exists in the chelated formMLos+, one calculates from equation (8b) Kch. = 0.25,and then from equation (12) the difference log A = log-KML - logKMLo = 0.1. As a difference of 0.1 log unit1027corresponds approximately to the lower limit whichmay be determined from potentiometric pH titrationdata with some reliability, for the considered complexsuch interactions may occur stereospecifically as wasshown 28938*39 earlier for the tetrahydrothiophen moietyof d-biotin with Mn2+ and Cu2+.TABLE 2Dimensionless equilibrium constants K,, [equation (8b)l for the intraniolecular equilibrium between the chelated isomer(MLosi ) and the simple carboxylate-co-ordinated isomer (MLo+) for several metal(1r)-ion complexes of thtc- and tnl-,together with the calculated percentages of the chelated isomer, in 50% aqueous dioxan at I = 0.1 mol dm-" and 25 "CL M log KYL* log KMLO log A Kch.yo MLos+ dthtc- Mn 1.80 f 0.05 1.88 ca. 0 ca. 0 <2Ofc u 4.31 f 0.03 3.15 1.16 13.5 93 f 1Zn 2.35 f 0.03 2.20 0.15 0 . 4 29 f 8Cd 2.68 f 0.02 2.34 0.34 1.2 54 Ik 4Pb 3.32 0.03 2.92 0.40 1.5 60 f 5tnl- Mn 1.87 f 0.03 h 1.83 ca. 0 ca. 0 < % O JZn 2.14 & 0.03 1.99 h 0.15 0 . 4 29 & 8Cd 2.10 & 0.04 g 2.16 ca.0 ca. 0 <20fc u 3.07 & 0.03 2.84 0.23 0.7 41 3- 7Pb 2.76 & 0.04 2.45 g 0.31 1.0 50 & 7The errors given are three times the standard error of the mean value. These values are expected on the grounds of thebasicity of the carboxylate group (see the corresponding ' reference lines ' in the original papers,10.11.37*38 which were obtained fromthe complexes with formate, acetate, etc.). As the extrapolation of these values is based on several independent systems the error isestimated to be s0.02 log units. The error limits were calculated by adding 0.02 log units to thtigiven errors of log KML. From refs. 37 and 38. f See text. From ref. 1 1 . From ref. 10.log A = log KYL - logKMLo.systems, this means that the lower limit of formationwhich may be estimated for MLos+ is 20%; all lowervalues cannot be determined by this method.Further-more, a difference of log A = 1 means that 90% of thecomplex is present as the chelated isomer MLos+, andlog A = 2 corresponds to 99% MLos+. I t must beadded that the values of log A (Table 2) may not bedirectly compared with the values of log KMTJ given inTable 1.With this background we may now view the resultslisted in Table 2 which have been calculated from pub-lished data.10.11.37*38 It is immediately evident that inall systems a considerable percentage of the simplecarboxylate co-ordinated isomer MLo+ is still present,but also that in the M2+-thtc- systems the percentageof the chelated isomer MLos+ is larger than for thehl, k-tnl systems, thus confirming the somewhat betterligating properties of thioethers compared with di-sulphides.In addition, by closer consideration, finerdetails are confirmed such as the similar co-ordinationtendency o f Zn2+ towards SR, and S,R,, or the tendencyof Cd2+ to favour SR, (see also Table 1).To summarize, the ligating properties of both thedisulphide and the thioether 21 groups favour soft metalions, while their tendency to co-ordinate with the bio-logically important hard or borderline metal ions isobviously very slight and rather unspecific (Table 1).However, it must be pointed out that these seeminglynegligible co-ordinating properties may lead to enhancedcomplex formation with ligands which have theseweakly co-ordinating moieties in addition to strongerbinding sites in the appropriate steric position, as wehave seen for the M2+-thtc- or -tnl- systems (Table 2),a fact which also holds for mixed-ligand complexes.lO*=We believe that the most important point, with regard tobiological systems, is that equilibria between isomers ofcertain complexes may be attained rather easily. Thisfactor could well be a kev to selectivitv, especiallv asWe thank Mr. K.Aegerter for recording the 90-MHzn.m.r. spectra, Miss K. Baunibusch for the skilful under-taking of some of the spectrophotometric measurements,and the Swiss National Science Foundation for a researchgrant.[9/892 Heceiued, 11th June, 19791REFERENCESSigel, Marcel Dekker, New York, Basle, 1979, vol.9, p. 1.p. 103.McCormick, J . Hrterocyclic Chem., 1974, 11, 119.1965, 77, 900; Angew. Chem. Znternat. Edn., 1965, 4, 846.R. B. Martin, in ' Metal Ions in Biological Systems,' ed. 13.1). L. Kabenstein, R. Guevremont, and C. A. Evans, in ref. 1 ,J. C. H. Shill, P. B. Williams, L. I). Wright, and L). B.U. Schmidt, P. Grafen, and H. W. Goedde, Angew. Chem.,K. G. Pearson, J . Chem. Educ., 1968, 45, 581, 643.L. G. Warner, T. Ottersen, and K. Seff, Inorg. Chetn., 1974,C. J . Hawkins and D. D. Perrin, Znorg. Chem., 19631, 2, 843.* E. W. Wilson, jun., and K . U. Martin, Arch. Biocliem.13 P. Kroneck, J. Amer. Chrm. SOL., 1976, 97, 3839.13, 2529.Biophys., 1971, 142, 445.lo H. Sigel, B. Prijs, D. B.McCormick, and J . C. 11. Shill,l1 H. Sigel, B. Prijs, and D. B. McCormick, J . Inovg. Nuclearl4 C.-1. Branden, Acta Chern. Scand., 1967, 21, 1000.l3 L. G. Warner, T. Ottersen, and K. Seff, Inorg. Chew., 1974,Arch. Biochem. Biophys., 1978, 187, 208.Chem., 1978, 40, 1678.13, 2819.K. Burger, in ref. 1, p. 213.l5 T. L. James and R. J . Kula, J . Phys. Chem., 1969, 73, 634.l6 P. Bonvicini, A. Levi, V. Lucchini, G. Modena, and G.Scorrano, J . Amer. Chem. SOC., 1973, 95, 5960.P. Bonvicini, A. Levi, V. Lucchini, and G. Scorrano, J . C . S .Pevkin I I , 1972, 2267.la D. B. McCormick, R. Griesser, and H. Sigel, in ' Metal Ionsin Biological Systems,' ed. H . Sigel, Marcel Dekker, New York,Basle, 1974, vol. 1 , p. 213.A. K. Amundsen, J . Whelan, and B. Bosnich, Inorg. Chew.,1979, 18, 206.2o P. M. Colman, H. C. Freeman, J . M. Guss, M. hlurata, Lr. *\.Norris, J. A. M. Ramshaw, and M. P. Venkatappa, ,'Vatwe, 1978,272, 319.21 H. Sigel, V. M. Rheinberger, and B. E. Fischer, Inorg. Chem.,1979, 18, 3334.22 P.K. Glasoe and F. A. Long, J . Pliys. Chem., 1960, 64, 188.23 P. R. Mitchell and H. Sigel, J . Amev. Chpm. Soc., 1978, 100,15641028 J.C.S. Daltona4 P. R. Mitchell, J.C.S. Dalton, 1979, 771.25 L. Johansson, Co-ordination Chem. Rev., 1974, 12, 241.26 L. G. Sillen and A. E. Martell, ' Stability Constants ofMetal-Ion Complexes,' Special Publ., The Chemical Society,London, 1964, no. 17; 1971, no. 25.2' M. H. Hutchinson and W. C. E. Higginson, J.C.S. Dalton,1973, 1247.28 H. A. Benesi and J. H. Hildebrand, J . Amer. Chem. SOC.,1949, 71, 2703.29 R. Cecil and J. R. McPhee, Biochem. J . , 1957, 66, 538;Adv. Protein Chem., 1959, 14, 255.30 S. Blackburn and F. Challenger, J . Chem. Soc., 1938, 1872.3 l M. M. David, R. Sperling, and I. 2. Steinberg, Biochim.32 M. Widmer, Thesis, ETH Zurich, 1972; M. Widmer and G.Biophys. Acta, 1974, 359, 101.Schwarzenbach, Chimia, 1970, 24, 447.8s H. B. Silber and L. U. Kromer, American Chemical Society34 G. Bergson, Arkiv. Kemi, 1958, 12, 233.36 Y. H. Mariam and R. B. Martin, Inorg. Chim. Acta, 1979,J. B. Orenberg, B. E. Fischer, and H. Sigel, J . Inorg.37 H. Sigel, R. Griesser, B. Prijs, D. B. McCormick, and M. G.38 H. Sigel, D. B. McCormick, R. Griesser, B. Prijs, and L. D.39 R. Griesser, B. Prijs, H. Sigel, W. Fory, L. D. Wright, andMeeting, Honolulu (Hawaii), 1979, Abstract INOR no. 518.35;623.Nuclear Chem., 1980, in the press.Joiner, Arch. Biochem. Biophys., 1969, 130, 514.Wright, Biochemistry, 1969, 8, 2687.D. B. McCormick. Biochemistry, 1970, 9, 3285