Bulletin of the Institute for Chemical Research, Kyoto University (1986), 64(3): 94-103
Departmental Bulletin Paper
Bull. Inst. Chem. Res., Kyoto
Vol. 64, No. 3, 1986
Analysis of Hydrophobic Effect in the Complex Formation of Beta-Cyclodextrin-mono-(6-deoxy-6-sulfonate) with Quarternary Ammonium Ions Hisashi
ReceivedJuly 24, 1986
$-Cyplodextrin substituted with a negatively charged sulfonate group at 6- position (8CDSO3 ) accelerated the alkaline hydrolysis of a positively charged meta-trimethylammoniophenyl (Me3N+-Ph) acetate by 3.5-fold when compared with J3-cyclodextrin(8CD). To analyze the interactions in the complex formation, dissociation constants, Kd(am), were determined for complexes of SCD and 8CDSO3- with a variety of quarternary ammonium ions as an inhibitory effect of ammonium ions on the association of cyclodextrins with a dye, crystal violet. Quantitative analysis using substituent parameters of N-substituent groups revealed that ammonium ions were included from the most hydrophobic N-substituent side first into the cavity of (3CDSO3 and that the positively charged N-atom remains out of the cavity. In the hydrolysis of positively charged acetate, however, meta-Me3N+group was included into the cavity to fit the acetyl group near the catalytic 2- or 3-hydroxy group of the host molecule and the electrostatic interaction between guest and host charged groups worked effectively to stabilize an inclusion complex.
Cyclodextrins have been studied as enzyme-models in the field of bioorganic chemistry because of their ability to form an inclusion complex with various types of organic compounds. In the inclusion reactions of 13-cyclodextrin (pCD), hydrophobic and van der Waals interactions are mainly responsible for complex formation)) Cyclodextrins have been chemically modified to improve their selectivities to a guest molecule and to increase catalytic functions.2'3) Electrostatic interactions between charged groups play an important role in an enzymatic system such as acetylcholinesterase and trypsin.4) Enzymes are macromolecular electrolytes, whereas cyclodextrins contain no charged group. Positively charged cyclodextrins, a (3CD flexibly capped with a metal ion5) and mono-(6-trimethylammonio-6-deoxy)-pCD,6) have been synthesized to develop new host systems which include an electronically charged guest molecule. Complex formation with hydrophobic anions was significantly increased by charged cyclodextrins and their results were explained in terms of electrostatic and hydrophobic interactions between host and quest molecules. In the present work, a negatively charged /3CD-mono-(6-deoxy-6-sulfonic acid), * JIj =VA , i,EEITtA : Departmentof AgriculturalChemistry,KyotoUniversity,Kyoto606. ** Presentaddress: Institutefor ChemicalResearch,KyotoUniversity , Uji-city,Kyoto611. ( 94 )
with 4° Ammonium
/3CDSO-3, was prepared as a simple model of trypsin and acetylcholinesterase containing a negatively charged group in the vicinity of the active site. Electrostatic interaction between /3CDS0-3 and positively charged meta-trimethylammoniophenyl (Me3N+-Ph) acetate was observed in the hydrolysis reaction of the acetate catalysed by /CDSO3-. To investigate hydrophobic and electrostatic interactions in the complex formation of j3CDSO3 , we measured dissociation constants, Kd(am), of inclusion complexes of pcDso3- with 15different quarternary ammonium ions and analyzed the substituent effects of N-substituent groups on log 1/Kd(am) values.
Materials /3CD(Hayashibara Biochemical Laboratories Inc.) was recrystalized from water and dried over phosphorous pentoxide in vacuo.Quarternary ammonium salts used in this study were those reported in our previous papers7'8)recrystalized before use from McOH, McOH/EtOAc, or McOH/ether. meta- and para-Me3N+Ph acetates were prepared from correspondingdimethylaminophenolby esterification with acetyl chloride, followed by methylation of dimethylamino group with MeI. Preparationof f3CDSO3 pCD was tosylated by p-toluenesulfonylchloride.6) Monotosylated /3CDwas separated from other tosylates on hydrophobic adsorption chromatography,9"1°) then substituted by Na2SO3 to fCD-sulfonate.11) 6-Deoxy-6sulfonation was confirmed by HCl hydrolysisof /CDSO-3 to give a product chromatographically identical with authentic 6-deoxy-glucose-6-sulfonic acid which was synthesized from diacetone glucose.12) Mono-6-tosyl-g-cyclodextrin A solution of p-toluenesulfonyl chloride (5 g) in dry pyridine was added to a solution of /3CD (26.2 g) in dry pyridine (250 ml) cooled on ice. The mixture was stirred for 1 hr and kept at 5°C for 1 hr. Solvent was removed in vacuoat 40°C to dryness. Diethyl ether was added to the residue. The presipitate was collectedand recrystallized twice from water to afford 13.5 g of crude product. HPLC analysis (LS410, 60% aq. McOH, UV at 254 nm) of the product indicated a mixture of mono- and poly-tosyl gal This crude product (8.23g) was dissolved in water and chromatographed on porous polystyrene gel (Amberlite XAD-2, 4,4X23 cm). Fractions eluted by 30% aq. EtOH gave a single product (4.19 g) on HPLC and TLC (silica gel, 8:1:1 AcOH/CHC13/H2Oand 100:30:27 EtOAc/MeOH/H2O). Recrystallization from water gave colorlesspowder of/3CDTs 3.91 g. The monosubstitutionof /3CDwas confirmedby 1H NMR spectra in DMSOd6; 83.0-3.8 (m, 28H), 4.8 (s, 7H), 5.7 (m, 14H), 7.43 (d, 2H), 7.73 (d, 2H). The ratio of peak areas due to the aromatic ring hydrogens (8 7.43 and 7.73) and the anomeric hydrogens (8 4.8) were 4:7.13) /CD-mono-(6-sulfonic acid) ammonium salt A solution of /CDTs (2 g) in DMF (40 ml) was added to an aq. solution of Na2SO3 (3.2 g/120 ml). The mixture was stirred at 50°C until a spot of (3CDTsdisappeared on TLC (silica gel, 8:1:3 AcOH/CHC13/ H2O), and condenced in vacuoat 40°C to syrup. The syrup was solidified in EtOH: 5.3 g colorlesssolid. Chromatography on a DEAE-SephadexA-25 column (j2.7 X 52 cm, linear gradient of NH4HCO3 from 0.02 M to 0.15 M) gave fractions free from ( 95 )
,8CDTs and /3CD on TLC. Water and NH4HCO3 were removed under reduced pressure to dryness. The residue was gel-filtrated on Sephadex G-15 (4)2.7X 55 cm) to afford 1.49 g of /3CDS03 ammonium salt as a colorless powder. Purity of the product was confirmed by silica gel TLC (Rf=0.41, 8:1:3 AcOH/CHC13/H20 and Rf=0.43, 3:1:2 BuOH/DMF/H20) and HPLC (IEX 260 SA SIL, 0.05 M NaC1, RI). 1H NMR (D20) 84.9 (s, 7H), 3.5-4.0 (m, 28H). Anal. Calcd for C42H73037 SN.10H2O: C, 36.14; H, 6.71; S, 2.30. Found: C, 36.22; H, 7.04; S, 2.33. The 6-deoxy-6-sulfonylation of /3CD was confirmed by the acid hydrolysis of /3CDS03 ammonium salt with 1 N HC1. After neutralization with aqueous ammonium bicarbonate, the products of hydrolysis were identified as D-glucose and 6-deoxy-Dglucose-6-sulfonate on TLC (Avicel, 2:1:1 BuOH/DMF/H20, detection with aniline hydrogen phthalate: Rf=0.57 for D-glucose; Rf=0.40 for 6-deoxy-D-glucose-6-
sulfonate) and on HPLC (LS460K, elution with 0.025-0.05 M NaC1, detection with RI). 6-Tosyl-1,2-0-isopropylidene-D-glucofuranose To a solution of diacetone-D-glucose (Sigma, 20 g) in methanol (10 ml) was added 10 ml of 0.8% sulfuric acid. The solution was stirred for 3 hr at room temp, neutralized with IR45, and filtered. The solvent was removed in vacuoto dryness. Crystallization from methanol/ether provided isopropylidene (940 mg). The isopropylidene, dried over phosphorus pentoxide at 60°C in vacuo,was dissolvedinto 4 ml of dry pyridine. To this solution, a solution of p-toluenesulfonyl chloride (188 mg) in 5 ml of CHC13was added and stirred at 0°C for 3 hr, then at room temp until the isopropylidene disappeared. The solvent was evaporated to afford oily residue. Following standard workup wiht chloroform, the oily residue was purified by chromatography on silica gel (elution with CHC13),then crystalized from CHC13/hexane to afford 1.0 g of tosylate. 1H NMR (CDC13)8 1.28-1.4 (s, 6H), 2.4 (s, 3H), 4.0-4.6 (m, 6H), 5.80(d, 1H,J=4 Hz, anomeric), 7.23(d,2H), 7.70 (d, 2H). 6-Deoxy-D-glucose-6-sulfonic acidammonium salt Aqueous solution of sodium sulfite (0.6 g/18 ml) was added to a solution of tosylate (1 g) in ethanol (10 ml) and refluxed for 24 hr. After cooling, the solution was passed through Amberlite IR 120 (H+ type) ion exchange resin, and condensedto oily residue. The oily residue was disolved in water, neutralized with ammonium bicarbonate, condensed in vacuo,and purified by ion exchange chromatography on DEAE-Sephadex A-25 (HCO3- type, elution with 0.02-0.07 M sodium bicarbonate). The yield was 190 mg. Anal. Calcd for C6H1508NS: 0,27.59 ;H,5.79;S,12.27. Found: C,27.55;H,5.93;S,12.11. Determination of the dissociation constantsof fCD- and/3CDS03--quarternary ammonium salt complexesDissociationconstants of inclusion complexesof cyclodextrin with quarternary ammonium salts were determined by a method of spectral competitive inhibition in which inclusion of quarternary ammonium salt was observed as the change in the spectra of cyclodextrin-dye complex due to a decrease in the concentration of cyclodextrin-dyecomplex. Crystal violet was used as a dye for the spectral competitive inhibition method. Crystal violet (ca. 20 mg) was dissolved into water (100ml) and diluted with ( 96 )
with 4° Ammonium
0.1 M phosphate buffer (pH 7.0) to 50-fold by volume. Quarternary ammonium salt was dissolved into this dye solution (ammonium-dye solution). Concentrations of quarternary ammonium salt were 10-40 mM in the ammonium-dye solution. Sample and reference cells were filled with 2.5 ml of the ammonium-dye solution and kept at 25+0.1°C for 5 min in Shimadzu UV-360 spectrophotometer equipped with a magnetic stirring and thermoelemental cell compartment. A solution of cyclodextrin dissolved in the ammonium-dye solution (pCD, 40 mg/5 ml; /3CDSO3-, 40 mg/2.5 ml) was successively added by a volume of 50-500 pi to a solution in the sample cell. The difference spectra were recorded at each addition of cyclodextrin. Since inclusion complexes with p-propylphenyltrimethyl ammonium iodide and quinoline hydrochloride showed difference spectra in the absence of a dye, their dissociation constants were determined without crystal violet. RESULTS AND DISCUSSION CatalyticEffectsof /CDS03- on the Alkaline Hydrolysisof meta- and para-Me3N+-Ph Acetates Time course of alkaline hydrolysis of Me3N+-Ph acetates was spectrophotometrically followed by the increase in the absorption of Me3N+-phenol produced. Bimolecular rate constants, k,,,,, of uncatalyzed alkailne hydrolysis of meta- and para-Me3N+-Ph acetates were 5.98 and 4.86 x 10-3 s-1, respectively, in 0.05 M carbonate buffer (pH 10.6) at 25°C. In the presence of cyclodextrins, apparent alkaline hydrolysis rate constant, k°bs, linearly increased with the concentration of cylcodextrins (0-15 mM). meta-Me3N+-Ph acetate was hydrolyzed about 3.5 times faster by pCDSO3 than by pCD (Figure 1). The slopes in Figure 1 were 5.69 x 10-3 s-1 mM-1 for pCDSO3 and 1.60 x 10-3 s-1 mM-1 for pCD. para-Me3N+Ph acetate, however, was not effectively acceralated by cyclodextrins; 0.24 x 10-3 s-1 mM-1 for /3CDSO3- and 0.34 x 10-3 s-1 mM-1 for /3CD. Since the acetyl group of phenyl acetates have to locate near the catalytic 2or 3-hydroxy group of the host molecule,14) the Me3N+-Ph group of acetate is included first into the cyclodextrin cavity. meta-Dervative easily takes a more favorable posikobs 10 - x10-2 S-1•
13CDS03 • •• 5-••f3CD •
05 Fig. 1.
Dependence of alkaline hydrolysis rate constant, k°bs, of meta-trimethylammoniophenyl acetate on the concentration of pCD and $CDSO3- in 0.05 M carbonate buffer (pH 10.6) at 25°C. (97)
H. MIYAGAWA, T. NISHIGKA and T. FUJITA tion in the cavity for the hydrolysis than para-derivatives. At this relative position of the guest in the cavity, meta-Me3N+ group is very close to 6-sulfonate group of the host. Then, electrostatic interaction between positively charged ammonium group and negatively charged sulfonate is responsible for 3.5 times greater acceralation of meta-Me3N+Ph acetate hydrolysis catalyzed by /3CDSO3 than hydrolysis by /3CD. DifferenceSpectra of a Cyclodextrin-CrystalViolet Complex Absorption spectra of crystal violet (Amax=589nm) showed a red shift by complexing with cyclodextrins. In the difference spectra of crystal violet with succesive additions of /3CD and 13CDSO3 appeared two peaks at 581 and 616 nm for /3CD, 581 and 615 nm for /3CDSO3 (Figure 2). /3CD- and f3CDSO3-crystal violet complexes were almost equal in the degree of red shift. Change in the absorption at minimun (581 nm) and maximum (615 or 616 nm) peaks followed the stoichiometry of the 1:1 inclusion complex formation between crystal violet and cyclodextrin. The value of dissociation constant, Kd(dye), of a dye-cyclodextrin complex determined at 581 nm was slightly different from that determined at 616 nm for 13CD(or 615 nm for pCDSO3) In this study, difference spectra were measured as the difference between the absorptions at 581 nm and at 616 or 615 nm (dAbs=dAbssis-1Abs581 for /3CD and dAbs=4Abs615—dAbs581 for f CDSO3 ). Kd(dye.) values were determined as 0.198+0.01 mM (number of data points=12) and 0.400±0.01 mM (number of data points=10) for /3CD and pCDSO3, respectively. Crystal violet formed a complex more stable with /3CD than with /3CDSO3. ComplexFormation of cyclodextrin with QuarternaryAmmoniumIons In the presence of quarternary ammonium ions, concentration of cyclodextrin-crystal violet complex decreases by the amount of cyclodextrin-ammonium ion complex formation. First, concentration of cyclodextrin-ammonium ion complex was estimated as the best fitting value in Scheme 1 by using dAbs, total concentrations of cylcodextrin and ammonium ion, and Kd(dye) value. Kd(am) in Scheme 1 is the dissociation constant of cyclodextrin-ammonium ion complex. Then, the value of Kd(am) was determined by the method of weight regression analysis from the estimated concentrations of cyclodextrin and its ammonium ion complex. Table 1 summarizes the values of Kd (am) for 15 quarternary ammonium ions determind from difference spectra. Kd (am) values for unsubstituted and p-OH-phenyltrimethyl ammonium, quinuclidine, and N-methylquinuclidine were less for the complex with /3CDSO3 than /3CD,
1nmnm700 Fig. 2. Differencespectra due to the crystal violet-cyclodextrininclusion complex formation. a) Difference spectra of the 13CD-crystalviolet system in the presence of 12.35 mM N-methylquinuclidine iodide. Concentrations of 13CDwere 0.0140.506 mM. b) Difference spectra of the 13CDSO3 -crystal violet system in the presence of 11.08mM para-chlorophenyltrimethylammonium iodide. Concentrations of pCDSO3 were 0.025-0.881 mM. ( 99 )
H. MIYAGAWA, T. NISHIGKA and T. FUJITA Table 1. DissociationconstantsKd(am) of SCD and SCDSO3 complexes withquarternaryammoniumionsdetermined by spectralcompetitiveinhibitionmethod. AmmoniumionsKd(am) PhN+(Me)3 p-H58. p-F62.6112 p-Cl27. p-CF311.420. p-Me38.428.9 p-Pr1. p-OMe31.233. p-OPr5.035. p-OH31.229. Me4N+199373 EtMe3N+31. isoPrMe3N+ QuinuclidineHC1 N-Mequinuclidine QuinolineHCl
mM with/3CD Kd(am) with19CDS03mM 1
2 30. 1 58.416.6 41.4 12.4
1.24 7 58 5 39.6 35.0 24.8 18.5
as expected from electrostatic interaction between a positively charged gest and a negatively charged host. Other quarternary ammonium ions listed in Table 1, however, formed more stable complex with /3CD rather than with negatively charged /CDSO3 . To find a general difference in the inclusion mechanism between two cyclodextrins, logarithmic values of 1/Kd(am) of /ICDSO3 complexes were plotted against those of pCD complexes with the corresponding ammonium ions (Figure 3). Almost linear relationship of the plot suggests that critical inclusion forces are the same between the two cyclodextrins. Several intermolecular interactions are simultaneously rseponsible for complex formation of cyclodextrins. Experiments and theoretical considerations have shown that hydrophobic interaction and van der Waals forces probably dominate in the complex formation.1,3,14,15) Since diameter of pCD is greater than that of alphacyclodextrin, van der Waals interaction is less important in the inclusion reactions of /3CD and its derivatives. We tried to estimate the contribution of hydrophobic interaction in the inclusion reaction of ammonium ions by /3CD and pCDSO3 in Table 1. In the structure-reactivity analyses, partition coefficient determined in the 1octanol/water system is used as a linear-free-energy related hydrophobic parameter (Hansch-Fujita analysis).17)Parameter r, estimated from the difference of partition coefficients of substituted derivative and its parent compound, is defined for the hydrophobicity parameter of a substituent group. Takayama et al.8) have measured partition coefficients of quarternary ammonium salts as those of ion-pair formationpartition equilibrium constants with picrate and separted them into the sum of hy(100)
0123 Fig. 3. Plot of log 1/Kd(am),dissociationconstants,of inclusioncomplexof quarternary ammoniumionsby SCD againstthoseof by pCDSO3. dropohbic, electronic, steric, and hydrogen-bonding effects of N-substituents. They suggested that the effects other than hydrophobic effect were derived from the interaction at the ion-pair formation between quarternary ammonium ion and picrate. Hydrophobicity of ammonium ions is mostly represented as the sum of hydrophobic effects of N-substituent groups (Table 2). Dissociation constants, log 1/Kd(am) values, were first plotted against hydrophobic substituent parameter 71-values of the N-substituent groups attached to a trimethylammonio radical (Figure 4a). Complex of quinoline HC1 was not included in the plot because of the lack of ,c value. The plot in Figure 4a was biphasic. Complexes of ammonium ions substituted with hydrophobic groups such as p-Pr-, CF3-, Cl-, Me-, and PrO-phenyltrimethylammonium ions formed a stable complex with cyclodextrins and their log 1/Kd(am) values increased linearly with their 7rvalues. Complexes with as less hydrophobic ammonium ions as p-OH-phenyltrimethy, tetramethy, ethyltrimethyammonium ions, however, deviated from the linear relation between the hydrophobic ammonium ions and their n values. Parametres such as electronic Qi and steric Esc effects did not improve the biphasic correlation in Figure 4a. Biphasic behavior of the plot reflects the inclusion direction of a guest ion in the cavity of a host cyclodextrin. Linear part of the plot proposes that ammonium ions are predominantly included into the cyclodextrin cavity from the substituted phenyl group side first, Me3N+ groups remain extended out of the cavity. For these ions, the substituted phenyl group is the most hydrophobic group among four N-substituent groups : 7r values of substituted phenyl groups are greater than the sum of x values for three methyl groups, 0.54 x 3=1.62. On the other hand, any of four N-alkyl groups of tetramethyl, ethyl- and isopropyltrimethyl ammonium ions are less hydrophobic and less bulky than a substituted phenyl group. Three N-alkyl substituents (101)
H. MIYAGAWA,T. NISHIOKAand T. FUJITA Table 2.
Substituent parameters used for the analysis of log 1/Kd(am).1)
1) Unless otherwise noted, values were from Ref. 8. 2) 7r values of the most hydrophobic group among N-substituents. 3) rr values of substituted phenyl group or the sum of three N-substituent groups.
4) rr valueswerecorrectedfor the electroniceffectof trimethylammonio group on the hydrogen bonding of oxygen atom with solvent (Ref 17). p-OMe and p-OPr were corrected by the equation rr(OR)+0.27 x 0.80. p-OH by the equation Tr(OH) 0.94 x 0.80. 5) M. Charton, Prog. Phys. Org. Chem., 13, 119 (1981).
0 ----------------------------------------------312 3 Fig. 4. Plot of log 1/Kd(am) values against hydrophobic parameter revalues. a) Plot against rr values of N-substituent groups attached to a trimethylammonio radical. Complexee by /3CD (0) and by SCDSO3 (•). b) Plot against rr values of substituted phenyldimethyl and alkyldimethyl groups. Complexes by RCD (0) and by /3CDSO3 (•). (102)
with 4° Ammonium
of these ammonium ions—trimethyl, ethyldimethyl, and isopropyldimethyl groups— are able to be included into the cyclodexrin cavity. When log 1/K(am) values were plotted against the sum of n values for three N-substituent groups, the deviation from the linearity was effectively improved (Figure 4b). Quinuclidine HC1 and Nmethyl quinuclidine were included from a bicycic alkyl group as expected from their ,r values. Although most ammonium ions studied formed a more stable complex with unmodified /3CD than with pCDSO3 , inside of the cavity /3CDSO3 was as hydrophobic as that of /3CD when estimated from the degree of red-shit of crystal violet complex. The analysis of log 1/Kd(am) values of quarternary ammonium ioncyclodextrin complexes by a hydrophobic substituent parameter ,r showed that predominant inclusion force of pCDSO3 was hydrophobic interaction, the dependency on which was same to that of /3CD, and that ammonium ions were included from the most hydrophobic substituent sides first into the cyclodextrin cavity. The positively charged nitrogen atom remained out of the cavity, so that the distance between nitrogen atom of the guest and 6-SO3 group of the host was large. This is the reason why electrostatic charge attraction between the negatively charged cyclodextrin 6SO3- group and the positively charged nitrogen atom of ammonium ions did not so effectively strengthen the stability of inclusion complexes as hydrophobic interaction. REFERENCES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) (17)
Y. Matsui, T. Nishioka,and T. Fujita, in `Topicsin Current Chemistry," F.L. BoschkeEd., Vol 128,Springer-Verlag, Berlin, 1985,pp. 61-89. I. Tabushi,K. Shimokawa,N. Shimizu,H. Shirakata,and K. Fujita, J. Am. Chem.Soc.,98, 7855(1976). W. Saenger,Angew. Chem.Int. Ed. Engl.,19, 344 (1980). G. E. Schulz and R. H. Schirmer,"Principlesof Protein Structure", Springer-Verlag,New York,1979. I. Tabushi,N. Shimizu,T. Sugimoto,M. Shinozuka,and K. Yamamura,J. Am.Chem.Soc., 99, 7100(1977). Y. Matsuiand A. Okimoto,Bull.Chem.Soc.Jpn., 51, 3030(1978). C. Takayama,T. Fujita,and M. Nakajima,J. Org.Chem.,44, 2871 (1979). C. Takayama,M. Akamatsu,and T. Fujita, Quant.Struct.-Act. Relat.,4, 149(1985). L. D. Meltonand K. N. Slessor,Carbohyd. Res.,18, 29 (1971). S. Onozuka,M. Kojima,K. Hattori, and F. Toda, Bull.Chem.Soc.Jpn., 53, 3221 (1980). M. Miyanoand A.A. Benson,J. Am.Chem.Soc.,84, 59 (1962). A.S. Meyerand T. Reichstein,Hela.Chem.Acta,29, 139 (1946). Y. Matsui,T. Yokoi,and K. Mochida,Chem.Lett.,1976,1037. M. L. Benderand M. Komiyama,"CyclodextrinChemistry,"Springer,Berlin, 1978. I. Tabushi,Y.-I.Kiyosuke,T. Sugimoto,and K. Yamamura,J. Am.Chem. Soc.,100,916(1978). C. Hansch and T. Fujita, J. Am. Chem.Soc.,86, 1616 (1964). T. Fujita,Frog.Phys.Org.Chem.,14, 75 (1983).