J . Chem. SOC., Faraday Trans. I, 1986,82, 2333-2343 Complexation of Roccellin by B- and y-Cyclodextrin Ronald J. Clarke, John H. Coates* and Stephen F. Lincoln* Department of Physical and Inorganic Chemistry, Uniuersity of Adelaide, South Australia 5001, Australiu Measurements of the u.v.-visible, fluorescence and circular dichroic spectra of roccellin (RO) in the presence of a-, p- and y-cyclodextrin (aCD, PCD and yCD) have been carried out. From equilibrium u.v.--visible spectral measurements, in the case of PCD, a single 1 : 1 complex was observed (298.2 K): RO+/ICD+RO*PCD ( K l ) where Kl = (7.20k0.88) x 10, dm3 mol-'. In the case of yCD a single 2: 1 complex was observed (298.2 K): 2RO+yCDe(RO),.yCD ( K J where K,, = (9.0+ 1.8) x 1Olo dm6 rnolk2. No complexation was observed between roccellin and aCD.Measurements of induced circular dichroism and of fluorescence were consistent with the reaction schemes proposed above for both /ICD and yCD, but no interaction was observed with aCD. Measurements of the u.v.-visible spectra of aqueous solutions of roccellin allowed characterisation of the dimerisation equilibrium : 2R0 G (RO), (Kd) where K , is (1.64 & 0.12) x 1 O4 dm3 mol-l. Measurements of the u.v.-visible and fluorescence spectra of roccellin, in increasing concentrations of lithium sulphate up to 0.5 mol dm-3, showed marked changes, consistent with the presence of increasing concentrations of aggregates of dye molecules with increasing electrolyte concentration. There has been considerable interest recently in the interactions between cyclodextrins and a variety of aromatic molecules, which form complexes by inclusion within the cyclodextrin ~avity.l-~ Such studies are of importance as there are analogies between this series of inclusion compounds and both drug-receptor and enzyme-substrate inter- actions.Indeed a number of attempts have been made to simulate enzyme behaviour by chemical modification of cyclodextrin molecules.6 In addition, the possibility exists for controlled sterically directed chemical syntheses by including one or more reactant molecules within a cyclodextrin cavity. Finally, it should be noted that a variety of drugs are suitable for inclusion within cyclodextrins, thus enabling improvement of their pharmacological proper tie^.^ In view of the above possibilities, it is worthwhile examining the selectivity of aCD, PCD and yCD, for a variety of included molecules.aCD, PCD and yCD are a-l,4-linked cyclic oligomers of D-glucopyranose with internal radii of 5-6, 7-8 and 9-10 A, respectively. We have investigated the diazo dyes methyl orange (MO)7 and tropaeolin 000 No. 2 (TR)8 in terms of both their kinetic and equilibrium properties on inclusions by the three cyclodextrins. A degree of size selectivity is shown7 by these molecule. Methyl orange, with two benzene rings, has been shown to form inclusion complexes with all three cyclodextrins, whereas TR,8 which possesses one naphthalene and one benzene ring, and roccellin, with two naphthalene rings, do not form any inclusion 23332334 Complexation of Roccellin by Cyclodextrins complexes with aCD, the smallest of the three cyclodextrins.Tropaeolin has been shown to be included as a 1 : 1, a 2: 1 and as a 2: 2 (dye-cyclodextrin) complex with yCD, in the case of DCD the 2 : 2 complex is not formed. Since roccellin possesses two naphthalene rings, but is otherwise quite similar to tropaeolin, it was decided to investigate the inclusion complexes which it will form with DCD and with yCD. It is a notable feature of the inclusion of azo dyes by the larger cyclodextrins that any tendency which they may have to form dimers in solution is enhanced by the possibility of inclusion in a suitably large cavity. In the case of roccellin, the dye itself forms dimers and higher oligomers readily, particularly in the presence of electrolytes.This property is also associated with the dye having a very low solubility in the presence of added salts, such as those used to provide a conductive solution for electrical Joule heating in a temperature-jump experiment. Consequently, measurements of the kinetics of the inclusion process were not possible for this dye using the usual temperature-jump methods. The kinetics of reactions occurring in non-conducting solutions can be studied using a pressure jump rather than a temperature jump, provided that the reaction of interest or some coupled reaction has a suitably large associated volume ~ h a n g e . ~ Experiments were attempted using a pressure-jump apparatus, but it was found that the amplitudes of the pressure-induced perturbations were too small and too rapid for satisfactory evaluation.However, it has been found possible to characterise the equilibrium properties of the RO-cyclodextrin complexes quantitatively. Experimental The a-, D- and y-cyclodextrins were obtained from the Sigma Chemical Co. and were used without further purification. They were stored as the anhydrous materials over phosphorus pentoxide in a vacuum desiccator. Rocellin (Sigma) was purified by salting out from hot distilled water with sodium acetate, recrystallised twice from distilled water and the final crystals were rinsed with distilled water. Elemental analysis was consistent with the dye being present as the monohydrate. The analytical grade salts potassium sulphate (B.D.H.), sodium chloride (Univar) and lithium sulphate (B.D.H.),, were used without further purification. All measurements were made on freshly pre ared dye solutions and exposure to light was kept to a minimum.No adsorption o f t f e dye to glass or quartz was detected. All volumetric glassware was cleaned by prolonged soaking in Decon 90 solution, followed by extensive rinsing in distilled water. The quartz spectrophotometer cells were carefully rinsed with each solution to be studied, prior to the recording of its spectrum. All solutions were diluted by weight from stock solutions prepared using A-grade volumetric apparatus. Visible spectra were measured in quartz cellsusingazeiss DMRlOdouble-beam spectrophotometer equipped witha thermostatted ( f 0.1 K) cell block. Spectra were run in duplicate at 298.2 K, recorded digitally at 2 nm intervals in the range 350-650 nm, punched onto paper tape and analysed using a Cyber 173 computer.Circular dichroic spectra were measured on a JASCO J40-CS spectropolarimeter, equipped with a microprocessor for averaging repeated measurements made at each wavelength. Linear dichroic spectra were determined on stretched poly(viny1 alcohol) films using a modified Zeiss PMQII spectrophotometer. Fluorescence spectra were measured using a Perkin-Elmer 3000 fluorescence spectrometer, equipped with a thennostatted cell holder. All measurements were made in a 1 cm path-length quartz fluorescence cell at 298.2 f 0.1 K. Results and Discussion Aqueous solutions of roccellin exhibit a red colour which deepens on the addition of a small amount of 1 mol dm-3 sodium hydroxide solution.This colour change occurs in the pH range 11-12. Thus, by analogy with tropaeolin8 it is likely that the pK, of theR. J. Clarke, J . H. Coates and S . F. Lincoln 2335 hydroxy group of roccellin is close to 1 1.4. Since this value is higher than that ofp-naphthol (pK, = 9.51), it is likely that the OH group is involved in a hydrogen bond in a similar fashion to the equivalent group in tropaeolin. The results described in this paper, except where specifically noted, were obtained in aqueous solution at ca. pH 7. Thus the monoanion is normally assumed to be the species present throughout. The similarity between the structures of roccellin and tropaeolin, shown below, suggests that the former, as well as containing an intramolecular hydrogen bond, should also undergo azo- hydrazone tautomerism (scheme 1).When in aqueous solution the hydrazone is probably the predominant species.lO tropaeolin d 7 roccellin azo hydrazone Scheme 1. Measurements of the u.v.-visible spectra of aqueous solutions of roccellin showed marked departure from the Beer-Lambert law. Determination of the dependence of the apparent molar absorbance on concentration for solutionscontaining no added electrolyte allowed characterisation of the dimerisation equilibrium : 2RO (RO), (&) (1) where Kd = (1.6450.12) x lo4 dm3 mol-l at 298.2 K. In view of the strong tendency for dimerisation, and in order to avoid contributions to the total absorbance from dimers, subsequent experiments involving cyclodextrins were carried out at roccellin concentrations of ca.2 x lop6 mol dm-3. At these concen- trations the bulk of the dye was in the monomer form. The nature of the aggregation in roccellin solutions was further investigated by measuring the visible absorption spectra of a series of roccellin solutions ([RO] = 5 x lop5 mol dm-3) containing increasing concentrations of lithium sulphate ( c 0 . 5 mol dm-") to increase the ionic strength. The results are shown in fig. 1. It can be seen that there is a large decrease in the intensity of the band at 505 nm, and a subsequent increase in the intensity of a band at 440 nm, as the ionic strength is increased. The curves in fig. 1 show that several spectroscopically distinct species must be present, since there is no isosbestic point. The diminution of the absorption band at longer wavelength, associated with increase in intensity of a band at shorter wavelength, is characteristic of the association of planar molecules into stacked aggregates accompanied by excition interaction.ll According to a number of workers,12' l3 higher cation concentrations aid aggregation in such systems by promoting ion-pair formation with negatively charged groups, thus decreasing intermolecular repulsions.Roccellin solutions in water do not give rise to any fluorescence emission. However,2336 Complexation of Roccellin by Cyclodextrins - I 20 " I - z m E 2 * 10 8 2 P m --- i 4 c( z 0 4 0 0 500 600 h/nm Fig. 1. Visible absorption spectrum of roccellin ( 5 x loe5 mol dmP3) in the presence of increasing concentrations of lithium sulphate at 298.2 K [Li,SO,]: (a) 0, (b) 0.01, (c) 0.05, ( d ) 0.075, ( e ) 0.10, v) 0.15, ( g ) 0.20, ( h ) 0.225, (i) 0.25, ( j ) 0.30 and ( k ) 0.50 mol dm-3.120 100 n +., -r( 5 80 d 3 v x c.' .4 Y 60 5 3 2 40 2 E: .r( Q) 2 0 0 1 I I (dl 600 700 h/nm Fig. 2. Fluorescence spectrum (Aex = 420 nm) of roccellin (4.9 x mol drnp3) in the presence of increasing concentrations of lithium sulphate at 298.2 K [Li,SO,]: (a) 0.15, (b) 0.20, (c) 0.30 and (d) 0.50 mol dmP3R. J . Clarke, J . H. Coates and S . F. Lincoln 2.0 1 . 5 1 . o 0 . 5 R . . . ...............:..* ..... . . . . . ...-.*a . - . . ... . . . . .. 0 . 0 2 50 4 50 hlnm 650 2337 Fig. 3. Linear dichroic spectrum of roccellin in stretched poly(viny1 alcohol) film (left-hand ordinate). (-) Light polarised parallel to stretch.(---) Light polarised perpendicular to stretch. The ratios of the parallel and perpendicular absorbances, R, are shown as individual points and refer to the right-hand ordinate. in the presence of increasing concentrations of lithium sulphate, a band appears at 600 nm, which increases in intensity and moves towards 620 nm (fig. 2). Since this behaviour parallels the u.v.-visible spectral behaviour in the presence of lithium sulphate, the fluorescence may be attributed to the presence of dye aggregates, within which quenching by oxygen or water is attenuated. In order to estimate the manner in which planar aromatic molecules aggregate, it is of interest to determine the geometrical relationship between the transition moment for an absorption band and the long axis of the molecule.This may be achieved by measuring the linear dichroism of a dye in a stretched poly(viny1 alcohol) film.14 It can be seen from fig. 3 that over the range 40s580 nm the absorbance parallel to the direction of stretch is substantially greater than that perpendicular to the direction of stretch and that the dichroic ratio is approximately constant. Assuming that the roccellin molecules align with the molecular long axis along the direction of stretch, these results indicate that the absorption bands in this wavelength range are polarised approximately along a line interconnecting the centres of the two naphthyl groups. Before proceeding to discuss the interaction between roccellin and the cyclodextrins in terms of molecular inclusion, it must be established that interaction between the glucose residues and the dye does not lead to spectral changes in the absence of inclusion.To this end the spectra of roccellin (2.0 x lop6 mol dmp3) in water and in a 0.01 mol dmP3 glucose solution were determined. It was found that only very small increases in molecular absorbance were obtained compared with those induced by the cyclodextrins, which are described later.2338 20 10 0:. Complexation of Roccellin by Cyclodextrins - . . . - L . - . . Eo - I 0 - E “E 0, -0 m . . d . I - . 20 5 . E * I - . m . E 2 - 10 s 5 -fl 0, P, . E ‘0 -u m . . h - . 0 . 400 50 0 600 A/ nm Fig. 4. Visible absorption spectrum of roccellin (2.0 x mol drnp3) in the presence of ( a ) BCD and (6) yCD at 298.2 K.In both cases the molar absorbance at 500 nm decreases systematically as the cyclodextrin concentration increases. The B-and y-CD concentration ranges were 0-2 x 10- and 0-5 x rnol dm-3, respectively. a-Cyclodextrin Interactions The addition of aCD (4.04 x lop3 mol dmp3) to an aqueous solution of roccellin (4.01 x mol dm-3) resulted in no significant changes in the dye’s absorption spectrum and no measurable fluorescence, neither was any induced circular dichroism apparent. Thus, it appears that aCD is not able to include the roccellin anion to any significant degree. This conclusion is in agreement with the results of attempts to construct possible inclusion of complexes using space-filling molecular models, where it is apparent that the naphthyl rings are too large to fit into the aCD cavity. b-Cyclodextrin Interactions The u.v.-visible absorption spectrum of roccellin (2.0 x lop6 rnol dm-3) alone, and in the presence of PCD concentrations from 5 x lop5 to 2 x lop3 mol dm-3, is shown in fig.4. A small hypochromic effect was observed on complexation, but no wavelength shift was detectable. The data are adequately described by the 1: 1 complex formation equilibrium : RO +pCD + RO *PCD (Kl). For this scheme the observed absorbance is given by A = &RO[RO] +EEO.BCD[RO*PCD]. (3) The equilibrium spectra of fig. 5 were fitted to eqn (3) by using the non-linear least-squares data fitting routine DATA FIT,^^ at all measured wavelengths, except thoseR. J. Clarke, J . H. Coates and S . F. Lincoln 2339 l-l d 0 E Fig.5. I I 10 - - 0 1 I 500 600 h/nm spectrum of RO alone (a). Derived spectra of the RO -BCD (b) and (RO), . yCD (c) complexes compared to the 24 2 0 I -12 - 1 6 300 50 0 700 X/nm Fig. 6. Induced circular dichroic spectrum of roccellin in the presence of p- and y-CD at 298.2 K. (---) [RO] = 4.0 x lop5, [DCD] = 4.0 x mol drnp3. (-) [RO] = 8.0 x [yCD] = 8.0 x mol dm-3.2340 Complexation of Roccellin by Cyclodextrins 100 8 0 h + .3 r: e 2 6 0 5 3 2 4 0 x 2 2 x Y .r( 2 0 0 550 6 50 X/nm 750 Fig. 7. Fluorescence spectrum (Aex = 420 nm) of roccellin (4.0 x lop5 mol dm-3) in the presence of 4.0 x lop3 mol dm-3 PCD (---) and yCD (--) at 298.2 K. where small changes in absorbance prevented DATAFIT converging to a best-fit value. The Kl values, calculated at 2 nm intervals in the range 462-570 nm, were weighted according to their estimated uncertainties, and averaged to give Kl = (7.20 & 0.88) x lo2 dm3 mol-l.Using this value of Kl, together with the directly determined molar absorptivities of roccellin, the spectrum of the RO.QCD complex was derived (fig. 5 ) . Comparison of the Kl values of roccellin and tropaeolin (Kl = 7.1 x lo2 mol dm-3 from temperature-jump measurements)* suggests that a similar process may be occurring in both cases, possibly the preferential encapsulation of the a-naphthol moiety. Com- parison of the spectrum of roccellin with that of PCD encapsulated roccellin, shows negligible shift in the wavelength of the absorption maximum, coupled with a diminution of the absorbances across the band. This behaviour is consistent with a change in the local environment of the chromophore, such as would be experienced on encapsulation.Fig. 6 shows the induced circular dichroic spectrum of roccellin in a 100-fold excess of Q-cyclodextrin. The spectrum exhibits only positive signals and there is no evidence of splitting due to exciton interaction. Thus, the spectrum appears to be consistent with the formation of a simple 1 : 1 complex and, since the CD signals are all positive, the transition moments of the dye molecule across the wavelength range studied must, according to Kajtar et a1.,16 lie within a 30" cone centred on the axis of symmetry of the QCD. The fluorescence spectrum of roccellin in a 100-fold excess of /3-cyclodextrin is shown in fig. 7. In the absence of PCD the dye exhibits no fluorescence.The fluorescence in the presence ofQCD no doubt arises as a consequence of the protection which the PCD cavity confers against quenching caused by solvent water and dissolved oxygen.17R. J . Clarke, J . H . Coates and S . F. Lincoln 234 1 y-CyclodextrieRoccellin Interaction The visible absorption spectra of roccellin (2.0 x mol dm-3) alone and in the presence of yCD concentrations ranging from 2.5 x lop7 to 5 x lop4 mol dmp3 are shown in fig. 4. A large hypochromic effect and a blue shift of cu. 15 nm were observed. In contrast to the roccellin-PCD system, significant changes are observed in the spectrum even at concentrations at which the dye is in excess, indicating that the affinity of the dye for the yCD is much greater than for the PCD.In addition, the form of the spectral changes suggests that the dye is included as a dimer rather than a monomer. In fact the data are best described by the equilibrium: 2 R 0 + yCD + (RO), - yCD (Kl,). (4) The methodology used previously for fitting the spectroscopic data was found to be inappropriate, since the very large value of K , , resulted in a shallow minimum which prevented convergence. Instead it was assumed that the system could be described by eqn (4), and a computer program MOLABS~~ was devised to calculate the absorbance of the dimer at every wavelength of interest by extrapolation of a double reciprocal plot of the change in the apparent molar absorptivity of the dye Germs the cyclodextrin concentration, to infinite cyclodextrin concentration.Program ROCEQU~~ was then used to calculate first the equilibrium concentrations of dye, cyclodextrin and complex, from the molar absorbances of the complex and of the free dye. The equilibrium concentrations were then used to obtain a value of the equilibrium constant for each wavelength of interest. Finally, these values were averaged over the range 580-464 nm to give the quoted value: K12 = (9.0 1.8) x 1Olo dm6 mol-,. Using this, and the directly determined values of the molar absorbances of roccellin, the spectrum of the (RO), yCD complex, shown in fig. 5, was derived. The form of the spectrum, which is characterised by a pronounced shift to shorter wavelengths of the maximum and the appearance of a shoulder at 540 nm, is consistent with the formation of an included dimer, accompanied by exciton interaction.The induced circular dichroic spectrum of roccellin in a 1 00-fold excess of yCD is shown in fig. 6. The positive and negative signals observed are characteristic of exciton splitting caused by dimerisation of the dye within a chiral environment, since the dye itself was shown by linear dichroism studies to have only long-axis polarised transitions associated with its visible absorption band. The necessary chiral environment is presumably provided by inclusion within the cyclodextrin. Two degenerate transition moments in close proximity and in the appropriate orientation are known to produce high-intensity exci ton spectra. The fluorescence spectrum of roccellin in a 100-fold excess of yCD is shown in fig.7. The intensity is almost twice that observed for PCD under the same conditions, and is presumably enhanced by the presence of the dye in the dimer form since, as has been observed earlier, aggregation of the dye is associated with fluorescence enhancement. The results presented here suggest that, where a complex is formed between any of the three cyclodextrins and roccellin, its stoichiometry and stability are determined mainly by the relative sizes of the guest molecule and the host cavity. It appears that the naphthyl groups of roccellin are too large to allow encapsulation by aCD. On the other hand, PCD is able to encapsulate one naphthyl group, whereas yCD allows two naphthyl groups to be encapsulated simultaneously, since the complex has the spectroscopic properties of a dimer.Furthermore, the tendency for dimerisa- tion of the dye is considerably enhanced by the presence of yCD in solution. Thus, the effective dimerisation constant (Kl, JyCD]) is ca. 4.5 x lo7 dm3 mol-1 at [yCD] = 5 x lop4 mol dmp3, compared with Kd = 1.64 x lo4 dm3 mol-l. The results des- cribed here, particularly when considered in the light of our previously published data7 for similar azo-dye systems, show that larger stability constants are associated with guests2342 Complexat ion of Roccellin by Cyclodextrins Table 1. Log Ka values for selected azo dyes and cyclodextrins (298.2 K) - methyl orange tropaeolin roccellin Kl, K2, 4 . 2 4, Kz, K1,Z K,, K . 2 _ - - - _ aCD 3.9,-, -b 9 7 , DCD 3.3,-, -b 2.8,6.6,9.4" 2.8,d ,- - 7 yCD 1.6,6.3, 7.9e 2.6,6.2,8.8" -, 10.9d a K , refers to DYEi-CDeDYE-CD.K2 refers to DYE * CD + DYE + (DYE), * CD. K l , refers to 2(DYE) + CD=(DYE);CD. Ref. (18). " Ref. (8). This study. Ref. (7). which fit most closely into the host. The results for three dyes are shown in table 1. It can be seen that MO, the dye with the smallest aromatic groups, is included by all three cyclodextrins. Although for a- and P-cyclodextrins the equilibrium spectra are not perfectly consistent with the total absence of other species, their spectra in the presence of MO suggest that the predominant species is the 1 : 1 complex and that the smaller aCD forms a more stable complex than the larger PCD. In the case of yCD, its 1 : 1 MO complex is considerably less stable than the 1 : 1 complexes of MO with the two smaller cyclodextrins.However, yCD is sufficiently large to allow the formation of an included MO dimer species of considerable stability. Tropaeolin is apparently unable to form an inclusion complex with aCD, probably because of the size of the naphthyl group. However, a dimer is included in both yCD and PCD. It is also notable that in the examples illustrated in table 1, the stability constants for the formation of a 1 : 1 complex (K,) are markedly less than for the corresponding 2: 1 complex (&), no doubt a consequence of the looseness of fit between the monomer and the cyclodextrin, which contrasts with the tightness of fit between the dimer and the cyclodextrin. Roccellin, with two naphthyl groups, is not surprisingly unable to form any inclusion complex with aCD, and forms only a 1 : 1 complex with PCD.The stability constants for the 1 : 1 complexes of both tropaeolin and roccellin with PCD are very similar, possibly indicating that in both cases it is the naphthyl moiety which is entering the cyclodextrin torus. It appears that there is insufficient room for a second roccellin molecule to enter the PCD molecule. y-Cyclodextrin is large enough to allow the entry of two roccellin molecules and the resulting complex has the greatest stability constant of those in table 1, again probably as a consequence of the closeness of fit between the large dimer and the large yCD cavity. We thank the Australian Research Grants Scheme for partial support of this research, and Dr Tom Kurucsev for the use of his program DATAFIT and for advice on some spectroscopic aspects of this study. References 1 M. L. Bender and M. Komiyama, Cyclodextrin Chemistry (Springer, Berlin, 1978). 2 W. Saenger, Angew. Chem., Int. Ed. Engl., 1980, 19, 344. 3 I. Tabushi, Acc. Chem. Res., 1982, 15, 66. 4 J. Szejtli, Cyclodextrins and their Inclusion Complexes (Akademiai Kiado, Budapest, 1982). 5 R. Breslow, Chem. Br., 1983, 126. 6 I. Tabushi and Y. Kuroda, J . Am. Chem. Soc., 1984, 106,4580.R. J. Clarke, J. H. Coates and S. F. Lincoln 2343 7 R. J. Clarke, J. H. Coates and S. F. Lincoln, Carbohydr. Res., 1984, 127, 18 1. 8 R. J. Clarke, J. H. Coates and S. F. Lincoln, J.Chem. Soc., Faraday Trans. 1, 1984, 80, 3 1 19. 9 J. S. Davis and H. Gutfreund, FEBS Lett., 1976, 72, 199. 10 R. L. Reeves and R. S. Kaiser, J . Org. Chem., 1970, 35, 3670. 11 M. Kasha, H. R. Rawls and M. Ashraf El Bayoumi, Pure Appl. Chem., 1965, 11, 371. 12 C. H. Giles, V. G. Agnihotri and K. McIver, J . Colloid Interface Sci., 1975, 50, 24. 13 B. R. Craven, J. C. Griffith and J. G. Kennedy, Aust. J. Chem., 1975, 28, 1971. 14 C. C. Bott and T. Kurucsev, J . Chem. Soc.. Faraday Trans. 2, 1975, 71. 749. I5 M. E. Gal, G. R. Kelly and T. Kurucsev, J . Chem. Soc., Faraday Trans. 2, 1973, 69, 395. 16 M. Kajtar, Cs. Horvath-Toro, E. Kuthi and J. Szejtli, Acta. Chim. Acad. Sci. Hung., 1982, 110, 327. 17 J. R. Lakowicz, Principles of Fluorescence Spectroscopy (Plenum Press, New York, 1983). 18 R. J. Clarke, Ph.D. Thesis (University of Adelaide, 1984). Paper 5 / 1373, Received 6th August, 1985