J. CHEM. SOC. FARADAY TRANS., 1994, 90(5), 739-743 Metal locyclodextr ins of 6A-(3-Aminopropylam in0)=6~-deoxy=~- cyclodextrint: Their Formation and Enantioselective Complexation of (R)-and (S)-Tryptophan Anions in Aqueous Solution Susan E. Brown, John H. Coates, Christopher J. Easton and Stephen F. Lincoln* Department of Chemistry, University of Adelaide , South Australia 5005,Australia From a pH titration study, the complexation of divalent metal ions (M2+)by 6A-(3-aminopropylamino)-6A-deoxy-~-cyclodextrin (BCDpn) to form the metallocyclodextrins, [M(/3CDpn)12+, is characterized by log(KJdm3 mol-') = 4.22& 0.02, 5.2 f0.1,7.35& 0.04and 4.96f0.08when M2+ = Co2+, Ni2+, Cu2+ and Zn2+, respec- tively, in aqueous solution at I = 0.10 (NaCIO,) and 298.2K.The complexation of the tryptophan anion (Trp-) by [M(/3CDpn)12+ is enantioselective for (S)-Trp-as indicated by log(K,,/dm3 mol-') and Iog(K,,/dm3 mol-') = 4.04& 0.03 and 4.32f0.05, 4.1f0.2 and 5.1 f0.2,and 7.85& 0.07 and 8.09f0.05, where the first and second magnitudes refer to the stability constants for [M(/3CDpn)(R)-Trp] + and [M(/3CDpn)(S)-Trp]+, respec- tively, when Mz+= Co2+,NiZ+ and Cu2+, respectively. The corresponding magnitudes for M2+= Zn2+ are both 5.3 & 0.1, indicating no enantioselectivity. The role of M2+ and other factors affecting complexation and enantio- selectivity are discussed. Natural and modified cyclodextrins exist in single enantio- meric forms and, when acting as host molecules, may prefer- entially complex one enantiomer of a chiral guest to produce two diastereomeric host-guest complexes of differing ther- modynamic stability.1-7 The degree of enantioselectivity varies substantially with the nature of the cyclodextrins and guests. In the case of aromatic guest molecules with polar substituents, this may be understood in terms of a model in which the aromatic moiety enters the cyclodextrin annulus and one or more polar substituents of the guest interact with the hydroxy or other polar groups of the cyclodextrin with varying degrees of inten~ity.'-~*'-l' Thus, R-and S-guests experience different geometric and electrostatic interactions with the cyclodextrin which may generate differing stabilities in the two diastereomeric host-guest complexes. In this study we are particularly interested in the combined effects of the coordinating ability of the metal centres and the chirality of the modified cyclodextrins or metallocyclodextrins on the enantioselectivity between guest enantiomers in host-guest c~mplexes.'~-'~ we showed that In a preliminary rep~rt'~ 6A-(3-aminopropylamino)-6A-deoxy-/3-cyclodextrin(BCDpn) formed a nickel(@ metallocyclodextrin ([Ni(BCDpn)] ),+ which was enantioselective for the (S)-tryptophan anion [(S)-Trp-)] in a ratio of 10:1 over (R)-tryptophan anion [(R)- (Sigma) were dried to constant weight and stored in the dark over P205in a vacuum desiccator prior to use.The enantio- meric purities of (R)-and (S)-Trp were determined to be >99% by HPLC analysis [Pirkle covalent (S)-phenylglycine column] of the respective esters formed with thionyl chloride pretreated methanol at 348 K.These purity limits were used in calculations of error limits of the stability constants char- acterizing the complexation of these enantiomers.Metal per- chlorates (Fluka) were twice recrystallized from water, and were dried and stored over P,O, under vacuum. (Caution: Anhydrous perchlorate salts are potentially powerful oxi- dants and should be handled with care.) All solutions were prepared using deionized water purified with a MilliQ-reagent system to produce water with a specific resistance of >15 Ma cm, which was then boiled to remove CO, . Equilibrium Studies Titrations were performed using a Metrohm Dosimat E665 titrimator, an Orion SA 720 potentiometer, and an Orion 8103 Ross combination pH electrode which was filled with 0.10 mol dm- NaC10,.Throughout a titration a stream of fine nitrogen bubbles (previously passed through aqueous 0.10 mol dm-3 NaC10,) was passed through the titration Trp-] in forming the ni~kel(rr)-6~-(3-aminopropylamino)-6~-solution which was magnetically stirred and thermostatted at deoxy-/3-cyclodextrin-tryptophananion host-guest complex, [Ni(/3CDpn)Trp] +,in aqueous solution. This appears to be the greatest degree of enantioselectivity between chiral guests so far reported for a metallocyclodextrin, and we now examine the effects of the variation of the nature of the metal ion on this enantioselectivity, and make comparisons with related systems.Experimental Preparation of Materials 6A-(3-Aminopropylamino)-6A-deoxy-/3-cyclodextrinprepared as in the literature,', and (R)-, (S)-and (RS)-tryptophan$ 7 B-Cyclodextrin = cycloheptamaltaose. 1Protonated tryptophan, tryptophan zwitterion and tryptophan anion are denoted as TrpH+, Trp and Trp-, respectively, prefixed by (R)-or (S)-as appropriate. 298.2 & 0.1 K in a water-jacketed 20 an3 titration vessel which was closed to the atmosphere with the exception of a small exit for the nitrogen stream. The 0.100 mol dm- Ni(ClO,), ,Cu(ClO,), and Zn(C10,), stock solutions were standardized by EDTA titration in the presence of Murexide indicator in the first two cases and Eriochrome Black T in the case of Zn(C10,),.l5 Ion exchange of Co2+ on an Amberlite HRC-120 cation-exchange column in the acid form followed by back titration of the liberated acid was used as the standardization method for the 0.100mol dm-3 Co(ClO,), stock solution. In all titrations, standardized 0.100 mol dm-3 NaOH was titrated against the species of interest in solutions 0.010 mol dmP3 in HClO, and 0.090 mol dm-3 in NaClO,. Thus the pK, values of /3CDpnHi+ and TrpH' were determined from titrations of 10.00cm3 aliquots of their 0.001 mol dm-' solu-tions. The stability constants for the formation of the /3CDpn. (R)-Trp- and /3CDpn + (S)-Trp-complexes were J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 determined by titration of 5.00 an3 each of 0.001 mol dm-3 solutions of either (R)-TrpH' or (S)-TrpH + and /3CDpnH;+, and the stability constants for the formation of the +[M(Trp)] +, [M(/3CDpn)12 and related complexes were determined by titration of 10.00 cm3 aliquots of 0.001 mol dm-3 solutions of either TrpH+ or j?CDpnH$+ with either 0.095 cm3 or 0.045 cm3 of M(ClO,), solution added.The sta- bility constants for the formation of [M(BCDpn)(R)-Trp] + and [M(/3CDpn)(S)-Trp] and related complexes were deter- + mined by titration of 5.00 cm3 each of 0.001 mol dm-3 solu- tions of either (R)-TrpH' or (S)-TrpH+ and /3CDpnHi+ with 0.045 cm3 of M(ClO,), solution added. E, and pK, values were determined by titration of 0.010 mol dm-3 HClO, (0.090 mol dm-3 in NaClO,) against 0.100 mol dmP3 NaOH.Derivations of the stability constants were carried out using the program SUPERQUAD.16 At least three runs were performed for each system, and at least two of these runs were averaged ;the criterion for selection for this averag- ing being that X2 for each run was e12.6 at the 95% con- fidence level. ' Results and Discussion General Aspects Several complexes formed in aqueous solutions of BCDpn, M2+ and tryptophan in the 2.5-11.5 pH range of this study. Their stabilities were determined from the differences between the pH profiles arising from titration against NaOH of solu- tions containing different combinations of the complexing species using the program SUPERQUAD. The sequence of these titrations was : (i) pK, determinations of /3CDpnHi+ and TrpH+, followed by determination of the stability con- stants of complexes in solutions of (ii) M2+and TrpHf, (iii) M2+ and BCDpnH$+, (iv) /3CDpnHq+ and either (R)-TrpH+ or (S)-TrpH+, and (v) M2+, PCDpnHi'and either (R)-TrpH' or (S)-TrpH+.The pK,s determined in (i) were employed as known values in the determination of stability constants in (ii)-(v), and the stability constants determined in (ii)-(iv) were employed as known values in the determination of stability constants in (v). The titration data were fitted to equilibria containing the minimum number of species required for a good fit, and any newly determined species found to be <5% of the total BCDpn or tryptophan concen- trations were considered to be insignificant.Two such pH titration profiles are shown in Fig. 1. Plots of the major species present in the Cu2 +-PCDpn-(S)-tryptophan system are shown in Fig. 2 and 3. The stability constants of the major M2+ complexes appear in Table 1, and those for other 12 2 1.30 1 , 1.40 , , 1 , , , , , , , , .50 1.60 1.70 1.80 , , 1.90 2 10 volume of NaOH added/cm3 Fig. 1 Titration profiles for (a) BCDpnH:+ (5.04x lo-, mol dmP3) and (R)-TrpH+ (5.08 x mol dm-3), and (b) BCDpnH:+ (5.02x mol dm-3), (R)-TrpH+ (5.05 x mol dm-j) and Cu(ClO,), (4.50x lo-, mol dm-3), each in aqueous 0.010 mol dm-3 HClO, and 0.090 mol dm-3 NaClO,, against 0.101 mol dm-3 NaOH. h 40 v m .-0 n 20 0 4.5 5.5 6.5 7.5 8.5 PH Fig. 2 Plot of Cu2+ species in a solution 0.00095, 0.001 and 0.001 rnol dm-' in total Cu2+, BCDpn and (S)-tryptophan concentrations, respectively, plotted relative to total [BCDpn] = total [(S)-Tryp-tophan] = 100%.(a) [Cu((S)-Trp)]+,(b) CU", (c) [Cu((S)-Trp),], + '. ,(4 CCNBCDP~XS)-T~PI (4 CCU(BCDP~XS)-T~PI (f[Cu(BCDpn)I2+,(9)[Cu((S)-?rp)OH], (h) [Cu(/3CDpn)OH]+ and (i) [Cu(/3CDpn)((S)-Trp)OH]. Table 1 Stability constants" characterizing metallo-6A-(3-aminopropylamino)-6A-deoxy-~-cyclodextrinsand related complexes in aqueous solution at 298.2K and I 0.10(NaCIO,)= M2+ coz+ Ni2+ b cu2+ Zn2+ log(K2/dm rnol -') 4.22f0.02 5.2 f0.1 7.35 _+ 0.04 4.96 _+ 0.08 10g(K3/dm3 mol-') 2.5 & 0.2 3.1 f0.1 3.09 f0.04 3.0f 0.1 log(KJdm3 mol-') 4.41 f0.05 5.42f0.03 8.11 f0.03 4.90f0.04 log(KJdm mol -') iog(~,./dm~ mol-') iog(~',,/dm~ mol-') log(K7,/drn3 mol- ') log(K7,/dm3 mol-') 4.01 0.08 4.04f0.03 (0.1) 4.32 0.05 (0.09) 4.67 f0.03 5.1 f 0.2(0.2) 4.1 f 0.2 (0.2) 7.20f 0.07 8.09 f 0.05(0.06) 7.85 f0.07 (0.07) 5.4 * 0.1 (0.2) 5.29 f 0.05 (0.1) 5.3 f0.1 (0.1) 5.3 f0.1(0.1) " Errors quoted for K (the mean of N runs) represent the standard deviations, CT = J{[Z(K,-K)']/(N -1)) where Kiis a value from a single run for the best fit of the variation of pH with added volume of NaOH titrant obtained through SUPERQUAD, and i = 1,2,...,N.When a K derived in this way was employed as a constant in the subsequent derivation of another K, the error associated with the first K was propagated in the derivation of the second K.For the diastereomers, the first and second errors quoted are calculated assuming 100 and 99% enantiomeric purity of tryptophan, respectively. 'Ref. 14. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 741-100, I Similar deprotonations were not reliably detected for the Co2+ and Zn2 + analogues, because the precipitation of a metal hydroxide species above pH 8.5 and 7.5, respectively, h2 60 \\ I In .-0 !? 40 20 0. 4.5 5:5 6.5 7.5 8.5 9.5 PH Fig. 3 Plot of non-Cu2+ species in a solution 0.00095, 0.001 and 0.001 mol dm-3 in total Cu2+, j?CDpn and (S)-tryptophan concen- trations, respectively, plotted relative to total [j?CDpn] = total [(S)-tryptophan] = 100%. (a) BCDpnH;+, (b) (S)-Trp, (c) BCDpnH+ and (4BCDpn.species appear below. A detailed discussion of the various equilibria now follows. Cyclodextrin Equilibria The acid dissociations of the diprotonated amino-propylamino substituent of BCDpnH;+ : Ka 1 BCDpnH;' =BCDpnH+ + H+ (1) K.2 j?CDpnH+ BCDpn + H+ (2) derived from data in the pH range 6.0-11.5 are characterized by pK,, = 7.39 0.04 and pK,, = 9.9 f 0.1. The acid disso- ciations of TrpH' characterized by pK,s of 2.40 f 0.02 and 9.28 0.01, and derived from data obtained in the pH ranges 2.0-3.0 and 8.0-10.5, respectively, are similar to those in the literature.' For complexations of either (R)-Trp- or (S)-Trp-by PCDpn : KIR BCDpn + (R)-Trp-BCDpn -(R)-Trp-(3) Kis BCDpn + (S)-Trp-BCDpn.(S)-Trp- (4) 1og(KlR/dm3 mol-') = 3.41 f. 0.02 (0.05) and log(Kls/dm3 mol-') = 3.40 f 0.07 (0.1) were derived from data in the pH range 8.5-11.5, where the first and second errors are calcu- lated on the basis of tryptophan being 100 and 99% pure, respectively. Complexation of 6"-(3-Aminopropylamino)-6*-deoxy-/?-cyclodextrin and the Tryptophan Anion by Divalent Metal Ions The stability of the metallocyclodextrin formed by BCDpn : K2 M2+ + BCDpn [M(BCDpn)12+ (5) varies with the nature of M2+ as shown by the variation of the magnitude of K, in the sequence Co2+ < Ni2+ < Cu2+ > Zn2+ (Table 1) as anticipated from the Irving-Williams series.' In the cases of [Ni(BCDpn)I2+ and [Cu(/?CDpn)]'+, pK,s of 9.20 f 0.04 and 7.84 f 0.03, respec- tively, were determined and are thought to correspond to the deprotonation of aqua ligands bound to the metal centres.obscured the formation of [Co(BCDpn)OH] and+ [Zn(fiCDpn)OH] and rendered titrations above these pHs + impractical. The formation of [M(PCDpnH)13+ : M2+ + BCDpnH+ K3 [M(BCDpnH)I3+ (6) is less favoured (Table 1) as anticipated from the charge repulsion between M2 + and BCDpnH and the monodentate + nature of BCDpnH'. The pK, of [M(BCDpnH)l3+ is 8.3 & 0.1, 7.83 f. 0.02, 5.74 & 0.05 and 8.1 & 0.1 when M2+ = Co2+, Ni2+, Cu2+ and Zn2+, respectively. These values probably characterize the deprotonation of the mono- protonated aminopropylamino substituents of BCDpnH+ in the metallocyclodextrins. The stability constants K, and K, and the corresponding pK,s were derived from data obtained in the pH ranges 6.0-8.5, 5.5-8.5, 5.5-9.0 and 5.5-7.5, when M2+ = Co2+, Ni2+, Cu2+ and Zn2+, respectively. The formation of [M(Trp)] and [M(Trp),] also occurs: + K4-M2+ + Trp-4[M(Trp)]+ (7) K5 [M(Trp)l+ + TrP-=CM(Trp),l (8) The stability constants, K, and K, ,determined in this study (Table 1) are in reasonable agreement with those in the liter- ature,17 and also exhibit variations anticipated from the Irving-Williams series." For [Ni(Trp)] and [Cu(Trp)] +,+ pK,s of 9.1 & 0.1 and 7.28 & 0.07, respectively, were deter- mined, which probably correspond to the deprotonation of aqua ligands bound to the metal centres.Similar deprotona- tions were not reliably detected for the Co2+and Zn2+ ana- logues, because the precipitation of a metal hydroxide species above pH 8.5 and 7.5, respectively, interfered with the titra- tions. The stability constants K, and K, and the correspond- ing pK,s were derived from data obtained in the pH ranges 6.5-8.5, 5.0-9.0, 3.0-6.5 and 5.5-7.0, when M2+ = Co2+, Ni2+, Cu2 + and Zn2+, respectively.Enantioselectivity in the Complexation of (R)-and (8-Tryptophan Anion by Divalent Metal Complexes of 6"-(3-Aminopropylamino)-6"deoxy-/?~yclodextrin The stability of the complexes formed by [M(j?CDpn)]'+ with (R)-Trp- and (S)-Trp- : K6R [M(PCDpn)12 + (R)-Trp-[M(BCDpn)(R)-Trp]++ (9) Kss [M(BCDpn)12++ (S)-Trp-[M(BCDpn)(S)-Trp]+ (10) also varies with the nature of M2+ as shown by the variation of the magnitude of K6, and K6s in the sequence Co2+d Ni2+ < Cu2+> Zn2+ (Table 1).In addition, there is a ten-fold enantioselectivity for (S)-Trp- when M2+ = Ni2+, as a comparison of K6, with K6, shows. When M2+ = Co2+ and Cu2+, there is a moderate enantioselectivity for (S)-Trp-, but when M2+ = Zn2+, no enantioselectivity is observed. The effect of enantioselectivity on the concentrations of several species in the Ni2+ system is shown in Fig. 4. The lower stabilities of /?CDpn -(R)-Trp- and BCDpn (S)-Trp- by comparison with those of [M(BCDpn)(R)-Trp] and+ [M(BCDpn)(S)-Trp]+,demonstrate that M2 + strengthens the complexation of Trp- . However, as [M(PCDpn)(R)-Trp]+ and [M(BCDpn)(S)-Trp]+ (KbR and K6s) are less stable than J.CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 60 I h $ 40 v 20 0 7.0 7.5 8.0 8.5 9.0 PH Fig. 4 Plot of selected species in a solution 0.00095,0.001 and 0.001 mol dmP3 in total Ni2+, BCDpn and either (R)-tryptophan or (S)-tryptophan (indicated by primed letters) concentrations, respectively, plotted relative to total [BCDpn] = either total [(R)-tryptophan] or total [(S)-tryptophan] = 100%. (a) flCDpnH+, (af) BCDpnH+, (b) “i((R)-Trp)I +7 (4 “i((S)-Trp)l+ (4CNi((R)-Trp),l, (4“i((S)-9 Trp),l, (4CNi(BCDpn)Y+r (4“i(BCDPn)lZ +r (4 CNi(BCDPd (R)- Trp] and (e’)[Ni(flCDpn)(S)-Trp] +.+ [M(Trp)]+ (K4)when M2+ = Co2+, Ni2+ and Cu2+, it appears that the factors stabilizing complexation of (R)-Trp- and (S)-Trp- by BCDpn and M2+ in [M(BCDpn)(R)-Trp]+ and [M(BCDpn)(S)-Trp] do not reinforce each other.In + +contrast, [Zn(BCDpn)(R)-Trp] and [Zn(BCDpn)(S)-Trp]+ are more stable than [Zn(Trp)]+, consistent with mutual reinforcement of the complexation of (R)-Trp- and (S)-Trp- by BCDpn and Zn2+, but there is no enantioselection between (R)-Trp- and (S)-Trp-. The structure envisaged for [M(BCDpn)(R)-Trp]+ and [M(BCDpn)(S)-Trp]+ has the indole moiety of Trp-inside the cyclodextrin annulus with the Trp- chiral centre in the vicinity of the primary hydroxy groups of the cyclodextrin, and the Trp- amine and carboxylate groups coordinated to M2+,as shown in Fig. 5. It has been argued that the indole moiety is only inside the cyclodextrin annulus for the dia- stereomer with the higher stability in the Cu2+ complexes of 6A-[4-(2-aminoethyl)imidazoyl]-6A-deoxy-/.?-cyclodextrin and 6A-deoxy-6A-[2-(4-imidazoyl)ethylamino]-/3-cyclo-dextrin, which preferentially complex (S)-Trp- and (R)-Trp-, re~pectively.’~.’~We have no evidence for such major struc- tural differences in the complexes studied here.The influence of the nature of M2+ on the stabilities of [M(jICDpn)(R)-Trp]+ and [M(BCDpn)(S)-Trp] + reflects the Fig. 5 A possible structure for [M(pCDpn)(S)-Trp]+, where the cyclodextrin annulus is shown as a truncated cone with the narrow and wide ends representing the circles delineated by primary and sec- ondary hydroxy groups, respectively. variation in the ionic radii of six-coordinate Co2+, NiZ+, Cu2+ and Zn2+, which are 0.745, 0.69, 0.73 and 0.74 A,” respectively, and the geometric constraints arising from ligand-field effects in Co2+, Ni2+ and CU”.~O It is particu- larly interesting that [Zn(BCDpn)(R)-Trp]+ and [Zn(BCDpn)(S)-Trp]+ are of the same stability, whereas the analogous diastereomers for the other three metals are of dif- ferent stability.This suggests that the absence of ligand-field- generated geometric constraints on d” Zn2+ allows more flexibility in the structures of [Zn(BCDpnXR)-Trp]+ and [Zn(BCDpn)(S)-Trp]+ and as a result enantioselectivity is decreased. In contrast, the d9 electronic configuration for the similar sized Cu2 + imposes a tetragonally distorted octa- hedral geometry which may place greater constraints on the interaction of the chiral centres of (R)-Trp- and (S)-Trp- with the BCDpn moiety and decrease the stability of [Cu(BCDpn)(R)-Trp]+ by comparison with that of [Cu(BCDpn)(S)-Trp]+.Similar arguments may be applied in the cases of d7 Co2+ and d8 Ni2+ whose six-coordinate geometries more closely approach regular octahedrons. The greater enantioselectivity caused by Ni2 + indicates that the size of the metal centre is important, and that a difference of 0.04 A can result in a substantial change in the degree of enantioselectivity.The stabilities of [Cu(BCDpn)(R)-Trp] and [Cu(BCDpn) + (S)-Trp]’ + : K~R [Cu(BCDpn)12+ + (R)-Trp [Cu(/?CDpn)(R)-Trpl2+ (11) [Cu(PCDpn)12+ + (S)-Trp Kis [Cu(BCDpnXS)-Trp] + (12) are lower than those of [Cu(flCDpn)(R)-Trp] and+ [Cu(BCDpn)(S)-Trp]+ (Table 1) and there is no significant enantioselectivity, probably because (R)-Trp and (S)-Trp act as monodentate ligands and are less sterically constrained than bidentate (R)-Trp- and (S)-Trp-.The deprotonations +of [Cu(/3CDpn)(R)-Trp] +, [Cu(BCDpn)(S)-Trp12 , + +[C@CDpn)(R)-Trp] and [Cu(pCDpn)(S)-Trp] are char- acterized by pK,s of 6.72 f0.08 (O,l), 6.6 & 0.1 (0.2), 9.48 & 0.07 (0.09) and 9.37 & 0.04 (0.05), respectively. The pK,s for the first pair may characterize the deprotonation of either the TrpH+ or the BCDpnH’ ligand, but an unam- biguous assignment is not possible. The pK,s for the second pair probably characterize the deprotonation of an aqua ligand.These reactions were -not detected when M2+ = Co2+, Ni2+ and Zn2+. The stability constants for the complexations shown in eqn. (9)-(12) were derived from data obtained in the pH ranges 7.5-8.7, 7.0-9.2, 4.5-9.5 and 6.5-8.0, when M2+ = Co2+, Ni2+, Cu2+ and Zn2+, respectively. We gratefully acknowledge funding for this research from the University of Adelaide and the Australian Research Council, and the award of an Australian Postgraduate Priority Research Award to S.E.B. References 1 N. J. Smith, T. M. Spotswood and S. F. Lincoln, Carbohydrate Res., 1989, 192,9. 2 S. E. Brown, J. H. Coates, S. F. Lincoln, D. R. Coghlan and C. J. Easton, J. Chem. SOC.,Faruday Trans., 1991,87,2699.3 S. E. Brown, J. H. Coates, P. A. Duckworth, S. F. Lincoln, C. J. Easton and B. L. May, J. Chem. Soc., Faraday Trans., 1993, W, 1035. J. CHEM. SOC. FARADAY TRANS., 1994, VOL. 90 743 4 I. Tabushi, Y. Kuroda and T. Mizutani, J. Am. Chem. SOC., 1986, 14 S. E. Brown, J. H. Coates, C. J. Easton, S. J. van Eyk, S. F. 5 108,4514. I. Tabushi, Y. Kuroda, M. Yamada and T. Sera, J. Zncl. Lincoln, B. L. May, M. A. Stile, C. B. Whalland and M. L. Wil- liams, J. Chem. SOC., Chem. Commun., 1994,47. Phenom., 1988,6,599. 15 A. I. Vogel, Quantitative Inorganic Analysis, Longmans, London, 6 K. Harata, Bull. Chem. SOC. Jpn., 1987,60,2763. 3rd edn., 1961. 7 D. Greatbanks and R. Pickford, Mag. Reson. Chem., 1987, 25, 16 P. Gans, A. Sabatini and A. Vacca, J. Chern. SOC., Dalton Trans., 208. 1985,1195. 8 R. J. Clarke, J. H. Coates and S. F. Lincoln, Adv. Carbohyd. 17 Critical Stability Constants, ed., R. M. Smith and A. E. Martell, 9 Chem. Biochem., 1989,46,205. J. Szejtli, Cyclodextrin Technology, Kluwer, Dordrecht, 1988. 18 Plenum Press, New York, 1975, vol. 1. H. Irving and R.J. P. Willliams, J. Chem. SOC., 1953, 3192. 10 J. F. Stoddard, Cyclodextrins, Royal Society of Chemistry, Cam- 19 R. D. Shannon, Acta Crystallogr., Sect. A, 1976,32, 751. bridge, 1990. 20 A. Cotton and G. Wilkinson, Advanced Inorganic Chemistry, 11 W. Saenger, Incl. Comp., 1984, 2, 231. New York, 4th edn., 1980. 12 G. Impellizzeri, G. Maccarrone, E. Rizzarelli, G. Vecchio, R. Corradini and R. Marchelli, Angew. Chem., Int. Ed. Engl., 1991, 30,1348. 13 V. Cucinotta, F. DAlessandro, G. Impellizzeri, G. Maccarrone and G. Vecchio, J. Chem. SOC., Chem. Commun., 1992, 1743. Paper 3/060201; Received 8th October, 1993