pubs.acs.org/Langmuir © 2009 American Chemical Society
Aggregation Behavior of Tetracarboxylic Surfactants Derived from Cholic and Deoxycholic Acids and Ethylenediaminetetraacetic Acid Mercedes Alvarez Alcalde,† Aida Jover,† Francisco Meijide,† Luciano Galantini,‡ Nicolae Viorel Pavel,‡ Alvaro Antelo,† and Jose Vazquez Tato*,† †
Departamento de Quı´mica Fı´sica, Facultad de Ciencias, Universidad de Santiago de Compostela, Avda. Alfonso X El Sabio s/n, 27002 Lugo, Spain, and ‡Dipartimento di Chimica, Research Center SOFT-INFM-CNR, Universit a di Roma “La Sapienza”, P.le A. Moro 5, 00185 Roma, Italy Received March 4, 2009. Revised Manuscript Received June 1, 2009
The reaction of 3β-aminoderivatives of cholic and deoxycholic acids (steroid residues) with dimethyl ester of ethylenediaminetetraacetic acid (bridge) leads to the formation of dimers carrying four carboxylic organic functions, two of them located on the side chain of each steroid residue and the other two on the bridge. As tetrasodium salts, these new compounds behave as surfactants and have been characterized by surface tension, fluorescence intensity of pyrene (as a probe), and static and dynamic light scattering measurements. Thermodynamic parameters for micellization were obtained from the dependence of the critical micelle concentration (cmc) with temperature. For both surfactants, the fraction of bound counterions is close to 0.5. The aggregation behavior is similar to one of their bile salt residues [i.e., sodium cholate (NaC) and sodium deoxycholate (NaDC)] and can be summarized as follows: (i) molecular areas at the interface for the new surfactants are fairly close to twice the value for a single molecule in a monolayer of natural bile salts; (ii) the environment where pyrene is solubilized is very apolar, as in natural bile salt aggregates; (iii) Gibbs free energies (per steroid residue) for micellization are not far from published values for NaC and NaDC, and the differences can be understood on the basis of less hydrophobicity of the new surfactants due to the charges in the bridge; and (iv) as for NaC and NaDC, aggregates have rather low aggregation numbers (which depend on the amount of added inert salt, NaCl). A structure based on the disklike model accepted for small bile salt aggregates is proposed.
Introduction Bile salts are natural amphiphilic compounds of great physiological importance.1 They have a fascinating molecular structure derived from their steroid nucleus and the organic functions (mainly hydroxyl groups) attached to it, commonly at positions 3, 7, and/or 12 in bile salts from mammals. These hydroxyl groups and the side chain supporting a carboxylic acid group have made them very attractive building blocks for forming supramolecular structures,2-4 novel antibiotics,5 prodrugs,6 new surfactants,7 organogelators,8,9 etc. Bile salts have a great surface activity and form aggregates (generally named micelles) in water as long as their concentration is above a critical concentration or critical micellization concentration (cmc).10 This surface activity is related to their amphipathic nature, which rises from the existence of a hydrophilic side, R (toward which the hydroxyl groups are oriented with the exception of ursodeoxycholate), and a hydrophobic side, β, where the hydrophobic methyl groups (C18 and C19) are located. The balance between hydrophilic and hydrophobic characteristics and geometry aspects of the molecules (1) Monte, M. J.; Garcia Marin, J. J.; Antelo, A.; Vazquez Tato, J. World J. Gastroenterol. 2009, 15, 804. (2) Tamminen, J.; Kolehmainen, E. Molecules 2001, 6, 21. (3) Soto Tellini, V. H.; Jover, A.; Meijide, F.; Vazquez Tato, J.; Galantini, L.; Pavel, N. V. Adv. Mater. 2007, 19, 1752. (4) Soto Tellini, V. H.; Jover, A.; Galantini, L.; Pavel, N. V.; Meijide, F.; Vazquez Tato, J. J. Phys. Chem. B 2006, 110, 13679–81. (5) Savage, P. B.; Li, C.; Taotafa, U.; Ding, B.; Guan, Q. FEMS Microbiol. Lett. 2002, 217, 1. (6) Sievanen, E. Molecules 2007, 12, 1859. (7) Alvarez Alcalde, M.; Jover, A.; Meijide, F.; Galantini, L.; Pavel, N. V.; Antelo, A.; Vazquez Tato, J. Langmuir 2008, 24, 6060. (8) Babu, P.; Sangeetha, N. M.; Maitra, U. Macromol. Symp. 2006, 241, 60. (9) Nonappa; Maitra, U. Soft Matter 2007, 3, 1428. (10) Small, D. M. Adv. Chem. Ser. 1968, 84, 31.
Langmuir 2009, 25(16), 9037–9044
govern both the adsorption at interfaces and the nature of the aggregates. This balance can be modified by attaching hydrophobic residues, which enlarge the hydrophobic area of the molecule and may dramatically affect their cmc and the nature of the aggregates formed. This is the case of the compounds obtained by attaching the p-t-butylphenyl or adamantyl groups (through an amide bond) to the 3-aminocholate.3,4 In the case of the p-t-butylphenyl derivative,3 temperature and surfactant concentration control the architecture of the resulting aggregates: vesicles, molecular tubes, or micelles. In this context, it must be remembered that only sodium or ammonium litocholate can form nanotubes,11-13 probably due to its higher hydrophobicity in comparison to dihydroxy and thihydroxy natural bile salts. In the case of the adamantyl derivative,4 SAXS and TEM measurements evidence the formation of a lamellar phase in aqueous solution. To enhance the surface activity of single-chain conventional surfactants, gemini (where the polar groups are located on the bridge that links two hydrophobic units) and bolaforms (having their polar groups at the extreme positions of the alkyl chains) have gained increasing attention.14-17 This is a nice strategy that should be much more explored with natural bile salts. After the synthesis in 1977 by McKenna et al.18 of the first bis-steroid containing two cholic acid molecules, other dimers (11) Terech, P.; De Geyer, A.; Struth, B.; Talmon, Y. Adv. Mater. 2002, 14, 495. (12) Terech, P.; Sangeetha, N. M.; Bhat, S.; Allegraud, J.-J.; Buhler, E. Soft Matter 2006, 2, 517. (13) Terech, P.; Jean, B.; Ne, F. Adv. Mater. 2006, 18, 1571. (14) Menger, F. M.; Keiper, J. S. Angew. Chem., Int. Ed. 2000, 39, 1907. (15) Menger, F. M.; Littau, C. A. J. Am. Chem. Soc. 1991, 113, 1451. (16) Zana, R. Adv. Colloid Interface Sci. 2002, 97, 205. (17) Li, Y.; Li, P.; Dong, C.; Wang, X.; Wang, Y.; Yan, H.; Thomas, R. K. Langmuir 2006, 22, 42. (18) McKenna, J.; McKenna, J. M.; Thornthwaite, D. W. J. Chem. Soc., Chem. Commun. 1977, 809.
Published on Web 07/16/2009
DOI: 10.1021/la9007813
9037
Article
[tail-to-tail (bolaform),19 head-to-head (gemini),7,20 and head-totail21,22], oligomers (lineal polymers,23,24 molecular baskets,25,26 molecular umbrellas,27,28 molecular clefts,29 and dendrimers,30,31), and cyclic oligomers32-37 have also been reported. Concerning dimers, most of the published papers refer to their synthesis18,29 or biological effects,20-22 but their aggregation behavior is not generally reported. The only exhaustive study of the aggregation of a bile salt dimer refers to a new gemini surfactant derived from 3R,12R-dihydroxy-5β-cholan-24-amine and ethylenediaminetetraacetic acid (EDTA).7 As expected, this compound is a much better surfactant than sodium deoxycholate (NaDC) and forms both micelles and vesicle-like aggregates. On the other hand, dimers of different compounds7,38-40 have been obtained by using EDTA as a bridge. Two of the four carboxylic acids are used to link the residue (normally through amide bonds), while the other two functions remain free, allowing an extraordinary enhancement of the solubility of low soluble hydrophobic residues, as for instance, adamantyl derivatives.38 Solubilization up to 6 mM cholesterol in water can be easily obtained when a β-cyclodextrin dimer (with the two cyclodextrin residues linked by EDTA) is used.40 The bile salt gemini mentioned above is another example.7 If this bridge is used to link two molecules of a natural bile salt by its 3-position, a third possibility, attractive to be investigated, results since the four carboxylic groups are located at far positions to each other, two of them belonging to the bridge while the other two are located on the side chain of the two bile acid residues. The resulting structure does not fulfill the architectures of common surfactants including bolaform and gemini ones, but it can be considered as a hybrid of a gemini and a bolaform surfactants (Figure 1).41 With all of this information in mind, two bile salt derivatives belonging to this new type of surfactant have been synthesized. For this purpose, two molecules of the 3β-amine derivatives of cholate and deoxycholate are linked by their amine groups to EDTA through amide bonds (Figure 2). The abbreviated names bis(C)-EDTA and bis(DC)-EDTA are used for the derivatives from cholate and deoxycholate, respectively. The new surfactants (19) Gouin, S.; Zhu, X. X. Langmuir 1998, 14, 4025. (20) Ronsin, G.; Kirby, A. J.; Rittenhouse, S.; Woodnutt, G.; Camilleri, P. J. Chem. Soc., Perkin Trans. 2 2002, 1302. (21) Enhsen, A.; Kramer, W.; Wess, G. Drug Discovery Today 1998, 3, 409. (22) Kramer, W.; Wess, G.; Bewersdorf, U.; Corsiero, D.; Girbig, F.; Weyland, C.; Stengelin, S.; Enhsen, A.; Bock, K.; Kleine, H.; Le Dreau, M.-A.; Schafer, H.-L. Eur. J. Biochem. 1997, 249, 456. (23) Gouin, S.; Zhu, X. X.; Lehnert, S. Macromolecules 2000, 33, 5379. (24) Gouin, S.; Zhu, X. X. Can. Polym. Prepr. 1997, 38, 586. (25) Ryu, E.-H.; Yan, J.; Zhong, Z.; Zhao, Y. J. Org. Chem. 2006, 71, 7205. (26) Ryu, E.-H.; Zhao, Y. J. Org. Chem. 2006, 71, 9491. (27) Janout, V.; Lanier, M.; Regen, S. L. J. Am. Chem. Soc. 1996, 118, 1573. (28) Janout, V.; Lanier, M.; Regen, S. L. J. Am. Chem. Soc. 1997, 119, 640. (29) Joachimiak, R.; Paryzek, Z. J. Inclusion Phenom. Macrocycl. Chem. 2004, 49, 127. (30) Balasubramanian, R.; Rao, P.; Maitra, U. Chem. Commun. 1999, 2353. (31) Ropponen, J.; Tamminen, J.; Lahtinen, M.; Linnanto, J.; Rissanen, K.; Kolehmainen, E. Eur. J. Org. Chem. 2005, 73. (32) Bonar-Law, R. P.; Sanders, J. K. M. Tetrahedron Lett. 1992, 33, 2071. (33) Bonar-Law, R. P.; Sanders, J. K. M. S. Tetrahedron Lett. 1993, 34, 1677. (34) Bhattarai, K. M.; Davis, A. P.; Perry, J. J.; Walter, C. J.; Menzer, S.; Williams, D. J. J. Org. Chem. 1997, 62, 8463. (35) Davis, A. P. Coord. Chem. Rev. 2006, 250, 2939. (36) Davis, A. P. Molecules 2007, 12, 2106. (37) Pandey, P. S.; Rai, R.; Singh, R. B. J. Chem. Soc. Perkin Trans. 1 2002, 7, 918. (38) Soto Tellini, V. H.; Jover, A.; Carrazana Garcia, J.; Galantini, L.; Meijide, F.; Vazquez Tato, J. J. Am. Chem. Soc. 2006, 128, 5728. (39) Yan, J.-M.; Atsumi, M.; Yuan, D.-Q.; Fujita, K. Helv. Chim. Acta 2002, 85, 1496. (40) Alvarez Alcalde, M.; Antelo, A.; Jover, A.; Meijide, F.; Gancedo, C.; Galantini, L.; Vazquez Tato, J. J. Inclusion Phenom. Macrocyclic Chem. DOI: 10.1007/s10847-008-9524-3. (41) Chevalier, Y. Curr. Opin. Colloid Interface Sci. 2002, 7, 3.
9038 DOI: 10.1021/la9007813
Alvarez Alcalde et al.
Figure 1. Molecular architectures of bolaform, gemini, and hybrid surfactants.
Figure 2. Structure of the new surfactants, bis(C)-EDTA (R = OH) and bis(DC)-EDTA (R = H), derived from EDTA (as the linking bridge) and the 3β-amino derivatives of cholate and deoxycholate, respectively.
have been characterized by surface tension, fluorescence of pyrene (as probe), and light scattering techniques. A priori, it can be expected that the four carboxylate groups will enhance the hydrophilicity of the dimer with respect to their related bile salts.
Experimental Section Synthesis. Following is the typical procedure for the synthesis of dimers bis(C)-EDTA and bis(DC)-EDTA from the dimethyl ester of EDTA and methyl esters of 3β-aminocholic and 3βaminodesoxicholic acids: Methyl esters of 3β-amino bile acids were synthesized from cholic acid and deoxycholic acids according to literature procedures.42 The dimethyl ester of EDTA (1.0 g, 3.1 mmol) was dissolved in 10 mL of dried dimethyl formamide (DMF). Diethyl cyanophosphonate (DEPC, 1.1 mL, 7.36 mmol) was added to this solution under stirring. After 20 min, the solution was cooled at 0 °C, and a solution of the methyl ester of the 3β-aminocholic acid (3 g, 7.1 mmol) in 5 mL of dried DMF and 2.2 mL of triethylamine was added dropwise with stirring under Ar. After 30 min, the ice bath was removed, and the reaction was maintained for 24 h at room temperature (42) Anelli, P. L.; Lattuada, L.; Uggeri, F. Synth. Commun. 1998, 28, 109.
Langmuir 2009, 25(16), 9037–9044
Alvarez Alcalde et al.
Article
[with thin-layer chromatography (TLC) control]. After the solvent was removed under vacuum, 200 mL of chloroform was added and washed with water (2 50 mL) to remove DMF. The organic phase was dried (Na2SO4) and partially evaporated under reduced pressure. Finally, the crude product was purified by column chromatography (silica gel 70-230 mesh; eluent ethyl acetate:methanol 9:1). The identity of the compound was confirmed by NMR, and the purity was confirmed by TLC. The overall yield was 73%. To remove the methyl ester groups, the compound was refluxed with 1 M KOH in methanol for 1 h at 80 °C. The solvent was evaporated, and the solid was redissolved in water (200 mL) and acidified with HCl (pH ∼1). When the solution was cooled, the compound precipitated, and it was filtered and dried in a vacuum oven. Finally, the corresponding sodium salt was obtained by adding a stoichiometric amount of NaOH. A similar procedure was employed for bis(DC)-EDTA. Bis(C)-EDTA. 13C NMR (300 MHz, DMSO). C1 (CH2) 31.30, C2 (CH2) 25.23, C3 (CH) 45.05, C4 (CH2) 35.09, C5 (CH) 37.48, C6 (CH2) 34.12, C7 (CH) 67.00, C8 (CH) 40.14, C9 (CH) 26.32, C10 (C) 35.44, C11 (CH2) 29.41, C12 (CH) 71.79, C13 (C) 46.53, C14 (CH) 42.07, C15 (CH2) 23.47, C16 (CH2) 27.90, C17 (CH) 46.85, C18 (CH3) 13.01, C19 (CH3) 23.60, C20 (CH) 35.68, C21 (CH3) 17.67, C22 (CH2) 31.52, C23 (CH2) 31.57, C24 (C) 175.54, -N-CH2-CH2-N- 53.74, -CH2-COOH 56.96, -CH2-CONH- 59.27, -COOH 170.01, -CONH 173.35 ppm. 1 H NMR (300 MHz, DMSO). 7.81 (d, 2H, amide bond), 3.86 (bs, 2H, H3), 3.77 (bs, 2H, H12), 3.61 (bs, 2H, H7), 3.35 (s, 4H, -CH2-CONH-), 3.13 (s, 4H, -CH2-COOH), 2.67 (s, 4H, -N-CH2-CH2-N-), 0.8-2.2 (m, Haliphatic), 0.58 (s, 6H, H18) ppm. Bis(DC)-EDTA. 13C NMR (300 MHz, DMSO). C1 (CH2) 31.47, C2 (CH2) 25.22, C3 (CH) 44.91, C4 (CH2) 31.55, C5 (CH) 32.93, C6 (CH2) 27.80, C7 (CH2) 26.47, C9 (CH) 35.53, C10 (C) 34.88, C11 (CH2) 29.46, C12 (CH) 71.86, C13 (C) 46.77, C14 (CH) 46.94, C15 (CH2) 24.14, C16 (CH2) 27.14, C17 (CH) 48.15, C18 (CH3) 13.13, C19 (CH3) 24.08, C21 (CH3) 17.65, C22 (CH2) 31.07, C23 (CH2) 31.21, C24 (C) 175.49, -N-CH2-CH2N- 53.67, -CH2-COOH 56.90, -CH2-CONH- 59.17, -COOH 169.95, -CONH 173.34 ppm. 1 H NMR (300 MHz, DMSO). 7.91 (d, 2H, amide bond), 3.96 (bs, 2H, H3), 3.77 (bs, 2H, H12), 3.37 (s, 4H, -CH2-CONH-), 3.16 (s,4H, -CH2-COOH), 2.668-2.482 (s, 4H, -N-CH2CH2-N-), 0.8-2.4 (m, Haliphatic), 0.59 (s, 6H, H18) ppm. Instrumental Techniques. Surface tension measurements (Wilhelmy plate method) were carried out in a KRUSS model K-10-ST tensiometer. Steady-state fluorescence measurements were recorded on a Hitachi model F-3010 (excitation wavelength, 336 nm; excitation slit, 3 nm; and emission slit, 1.5 nm). A Brookhaven instrument constituted by a BI-2030AT digital correlator with 136 channels and a BI-200SM goniometer was used for static (SLS) and dynamic (DLS) light scattering measurements. The light source was a Uniphase solid-state laser system model 4601 operating at 532 nm. Dust was eliminated by means of a Brookhaven ultrafiltration unit (BIUU1) for flow-through cells, the volume of the flow cell being about 1.0 cm3. Nuclepore filters with a pore size of 0.1 μm were used. The samples were placed in the cell for at least 30 min prior to the measurement to allow for thermal equilibration. Their temperature was kept constant within 0.5 °C by a circulating water bath. In the DLS experiments, the intensity-intensity autocorrelation function was measured at a particular value of the scattering vector q and related to the normalized electric field autocorrelation function g1(q,τ) by the Siegert relation. Therefore, g1(q,τ) was analyzed through the cumulant expansion, and the so-called apparent diffusion coefficient Dapp was obtained from the first cumulant by the relation Dapp ¼ -
1 d ln g1 ðq, τÞ jτ ¼ 0 q2 dτ
Langmuir 2009, 25(16), 9037–9044
ð1Þ
As a check, an analysis by CONTIN of g1(q,τ) was also performed for verifying multimodal distributions. In the SLS measurements, the excess Rayleigh ratio ΔR values were analyzed by means of the equation c0 K 1 ¼ ΔR Mapp
ð2Þ
where Mapp is the apparent molecular weight; K is a constant that depends on the solvent refractive index, the solution refractive index increment, and the laser wavelength; and c0 represents the micelle concentration (total solute concentration minus cmc) as weight/volume. The refractive index measurements were performed by an Atago differential refractometer model DD7. The observed excess Rayleigh ratios and the apparent diffusion coefficients did not depend on the exchanged wave vector in the scattering angle range 30-150° under our experimental conditions; therefore, only the results at 90° were analyzed. Accounting for particle interactions, the following concentration dependence were assumed for Dapp and Mapp 1 1 ð1þλI c0 Þ ¼ Mapp M
ð3Þ
Dapp ¼ Do ð1þλD c0 Þ
ð4Þ
where M and Do are the particle molecular weight and free diffusion coefficient, respectively. λI is a parameter that depends on static interaction, whereas both static and hydrodynamic interaction affect the λD value. From the Do value, the hydrodynamic radius Rh can be calculated by the Stokes-Einstein equation Rh ¼
kT 6πηDo
ð5Þ
where k is the Boltzmann constant, T is the absolute temperature, and η is the solvent viscosity.
Results and Discussion The surface tension, γ, was studied at 25 °C in water and in bicarbonate/carbonate buffer solution (pH = 9.75) of different ionic strengths. Figure 3 shows some experimental series. The plots do not show any minima in the surface tension, indicating the absence of strong surface-active impurities in the samples.43 From the breaking points of the linear dependence, the cmc is determined (Table 1). It can be observed that cmc decreases with increasing the ionic strength, as in typical surfactants, allowing the estimation of the fraction of bound counterions to the micelle, β (see below). The cmc values are very similar to those reported for sodium cholate (NaC) and NaDC by Almgren et al.44 (19.9 mM and 7.8 mM, respectively, in 0.05 M NaCl) and others.45 Furthermore, the cmc values for bis(DC)-EDTA are lower than those for bis(C)-EDTA, evidencing a more hydrophobic nature. The surface behavior may be analyzed in terms of the adsorption Gibbs equation, which relates the change in the equilibrium surface tension with changes in the chemical potentials of all of the solutes at constant temperature dγ ¼ -
n -1 X
Γi dμi
ð6Þ
i¼1
(43) Kratohvil, J. P.; Hsu, W. P.; Jacobs, M. A.; Aminabhavi, T. M.; Mukunoki, Y. Colloid Polym. Sci. 1983, 261, 781. (44) Swanson-Vethamuthu, M.; Almgren, M.; Hansson, P.; Zhao, J. Langmuir 1996, 12, 2186. (45) Coello, A.; Meijide, F.; Rodrı´ guez Nunez, E.; Vazquez Tato, J. J. Pharm. Sci. 1996, 85, 9.
DOI: 10.1021/la9007813
9039
Article
Alvarez Alcalde et al.
Figure 3. Surface tension vs ln c (molar concentration) at 25 ( 0.1 °C. (a) Bis(C)-EDTA aqueous solution (9), 0.1 M NaHCO3/Na2CO3 buffer (2), and 0.2 M NaHCO3/Na2CO3 buffer (O). (b) bis(DC)-EDTA aqueous solution (9), 0.1 M NaHCO3/Na2CO3 buffer (2), and 0.2 M NaHCO3/Na2CO3 buffer (O). Table 1. Surface Properties of the Bile Salt Dimers Determined at 25.0 ( 0.1 °C compound bis(C)-EDTA
bis(DC)-EDTA
cmc (mM)
107Γ (mol/m2)
As (A˚2)
H2O 0.1 M buffer 0.2 M buffer
19.6 11.8 8.4
9.73 9.55
171 174
H2O 0.1 M buffer 0.2 M buffer
6.7 5.8 4.0
9.94 9.52
167 174
solvent
where Γi is the surface excess concentration of the i-th component. The complete dissociation of the surfactant in the bulk solution is MQ f ν - Qz - þνþ M zþ
ð7Þ
Q and M being the surfactant ion and the counterion, respectively. The dissociation of an inert electrolyte having the same counterion as the surfactant, M zþ, and a nonadsorbing co-ion, zsX , is s
s
MX f νs- X z - þ νsþ M zþ
ð8Þ
The superscript “s” indicates added salt, and the summation in eq 6 is expanded over all of the ionic species in solution, that is, Q, M, and X. By considering (i) that the solution is ideal, (ii) that the surface excess concentration of the co-ion is near zero, (iii) mass balance equations, and (iv) the electroneutrality conditions, the well-known eq 9 is obtained 1 dγ Γ¼ ð9Þ nRT d ln c cs where c and cs are the known bulk concentrations of surfactant and added salt, respectively, R is the universal gas constant, T is the absolute temperature, and n is given by n ¼ ν- þ
ν2þ νþ þ νsþ ccs
ð10Þ
This equation shows that the prefactor n in the Gibbs equation depends on the stoichiometry of the surfactant and also on both the stoichiometry and the concentration of an added inert 9040 DOI: 10.1021/la9007813
electrolyte. When no electrolyte is added (measurements in water), cs=0, and n=ν- þ νþ. For present surfactants, n would be 5. In buffer solutions, the concentration of electrolyte is in a large excess with respect to the surfactant concentration and the prefactor becomes n=ν-, that is, n=1. Furthermore, for weak electrolytes, cs is also a function of the degree of dissociation, which must be known. The high pH value used in this work was chosen to prevent the existence of protonated species in the solution, in particular those derived from the protonation of the nitrogen atoms of the bridge, since pKa values in the ranges 3.54.4 and 6.7-7.3 have been published for the dissociation of the tertiary amino groups of EDTA diamides.46 Therefore, a full dissociation of carboxylate groups and no protonation of amino groups is guaranteed. Consequently, although some experimental measurements have been carried out in water, we shall only discuss the results obtained in buffer solutions since in water there is some uncertainty on the actual species present in solution. This will prevent possible errors in the choice of the value for n for calculating molecular areas at the interface. The analysis of the experimental data in Figure 3 obtained in buffer solutions leads to the surface excess concentrations listed in Table 1. The corresponding molecular areas at the interface, As, are also included and have been calculated with the equation Γ=1/(AsNA). Surface studies of insoluble monolayers of various bile acids have been reviewed by Small.47 Monohydroxy and dihydroxy bile acids have nearly identical compression isotherms, with monolayer collapse points corresponding to surface area values of about 80-90 A˚2/molecule. For cholic acid, the collapse point occurs at a greater surface area, although in this case the collapse point is not well-defined as the curve after the collapse shows a gradual rise of pressure instead of a flat portion as in the case of other bile acids. Thus, the values obtained here (around 170 A˚2/molecule) are fairly close to twice the value for a single molecule in a monolayer of natural bile acids at the point in which the single liquid monolayer collapses. This coincidence suggests that the surfactant molecule is lying flat at the air-water interface. Calculations with space models show that the areas for fully extended conformations are around ≈230-240 A˚2 (Figure 4). (46) Danil , A. F.; Pacheco Tanaka, D. A. J. Chem. Soc. Faraday Trans. 1998, 94, 3105. (47) Small, D. M. In The Bile Acids, Chemistry, Physiology, and Metabolism; Nair, P. P., Kritchevski, D., Eds.; Plenum Press: New York, 1971; Chapter 8.
Langmuir 2009, 25(16), 9037–9044
Alvarez Alcalde et al.
Article
Figure 4. Conformations of the bis(C)-EDTA at the air/water interface. (a) Upright view of the fully extended molecule. (b) “V” conformation with a superficial area depending on the angle formed by the two steroid moieties. (c) Molecule lying in a flat position with polar groups oriented toward water.
The difference may be understood on the basis that the flexibility of the bridge would allow the adoption of angular (V) conformations. It is interesting to notice that the area corresponding to a closed V conformation is around 75 A˚2, which is not far from twice the value (about 40-44 A˚2/molecule)47 for a bile acid molecule standing up at the interface. Furthermore, the contribution of the bridge to the total flat area is probably overestimated in the calculations. Therefore, it may be concluded that the most probable conformation for a surfactant molecule at the air-water interface would correspond to an almost fully extended molecule, the R faces (hydrophilic) of the two steroid residues and the double-charged bridge pointing toward water and the β faces toward the air, far from a closed V conformation. This would be favored by the screening of the electrostatic repulsions by the counterions from added buffer. The relation between the fluorescence intensities of the first and third vibronic peaks of pyrene, I1/I3, is a widely used criteria for estimating the local polarity of the environment around the probe inside of surfactant micelles,48 including bile salt aggregates.49 Thus, the aggregation of both dimers in water was also investigated by the fluorescence of pyrene. Experimental I1/I3 values are plotted in Figure 5 as a function of the dimer concentration. Typical sigmoid curves were obtained, although the decrease in the I1/I3 ratio is not abrupt, since it ranges from approximately 1 to 50 mM. Values for cmc were taken from the inflection points50 and are shown in Table 2. They are in agreement with those obtained from surface tension measurements. Measurements were carried out at three temperatures, covering the range 1040 °C. This allows the estimation of thermodynamic parameters associated to the aggregation process. Upper limits in Figure 5 (I1/I3 ≈ 1.9) correspond to pyrene dissolved in the bulk solution. In these conditions, surfactant molecules are not associated. On the other hand, the plateaus reached at the highest surfactant concentrations [I1/I3 ≈ 0.75 for bis(DC)-EDTA and 0.77 for bis(C)-EDTA] are very close to those obtained for NaC and NaDC (0.70 and 0.75, respectively).49 As in the case of the natural bile salts, these values suggest that the probe is solubilized in a nonpolar environment, inside the aggregates. In what follows, the structure of the anionic tensioactive (Qz_ in eq 7) is considered as composed by i charged groups of valency zi of the EDTA residue and j bile salt (BS) units of valency zj linked by a neutral spacer, that is, z-=izi þ jzj and Qz-=Qizj i þ jzj. For the surfactants studied here, i=j=2, zi=zj=-1, and zþ=1. (48) Kalyanasundaram, K.; Thomas, J. K. J. Am. Chem. Soc. 1977, 99, 2039. (49) Jover, A.; Meijide, F.; Rodrı´ guez Nun~ez, E.; Vazquez Tato, J.; Mosquera, M.; Rodrı´ guez Prieto, F. Langmuir 1996, 12, 1789. (50) Aguiar, J.; Carpena, P.; Molina-Bolivar, J. A.; Carnero Ruiz, C. J. Colloid Interface Sci. 2003, 258, 116.
Langmuir 2009, 25(16), 9037–9044
The micellization process, in which N molecules of surfactant and p counterions of valency zþ are implied, is described by eq 11: iz þ jzj
Nðiz þ jzj Þ -pzþ
þ pX zþ SQj , N i
NQj i
ð11Þ
If the concentrations of species are expressed in mol of BS units per volume unit, the equilibrium constant takes the form given in eq 12, where the terms in brackets are molar concentrations. Nðiz þ jzj Þ -pzþ
K¼
½Qj, N i
=jN
iz þ jz ð½Qj i j =jÞN ½X zþ p
ð12Þ
Therefore, the free energy of micellization per mol of BS residue is given by: 0 1 izi þ jzj ½Q RT ln K RT @N ln j þ p ln½X zþ A ð13Þ ≈ ΔGoM ¼ j jN jN In deducing this equation, it has been taken into account that the term associated with the molar micelle concentration is very small at concentrations very slightly above the cmc, and consequently, it has been neglected.51 Equation 15 is obtained by substitution of the fraction of charge of micellized surfactant neutralized by the counterions (eq 14) in the equation expressing the electroneutrality at the cmc (≈[Qizj i þ jzj]) in terms of mol of steroid residues per volume unit. β¼
pjzþ j Nðijzi jþ jjzj jÞ
ijzi jþ jjzj j izi þ jzj ½Qj ¼ jzþ j½X zþ j
ð14Þ
ð15Þ
The term [Qizj i þ jzj] can be extracted and substituted in eq 13, thus providing ΔGoM ¼ RT
1 ijzi jþ jjzj j þβ ln cmc j jjzþ j
ð16Þ
In the deduction of this equation, terms with a negligible contribution have not been considered. The application of the Gibbs-Helmholtz equation to eq 12 allows the determination of the enthalpy of micellization per mol of BS residue, ΔHoM. By (51) Zana, R. Langmuir 1996, 12, 1208.
DOI: 10.1021/la9007813
9041
Article
Alvarez Alcalde et al.
Figure 5. Plot of the relative pyrene fluorescence intensity, I1/I3, vs ln c [bis(DC)-EDTA] (a) and ln c [bis(C)-EDTA] at several temperatures (c). Dependence of ln cmc with temperature for bis(DC)-EDTA (b) and bis(C)-EDTA (d). Table 2. Aggregation Parameters for Bis(DC)-EDTA and Bis(C)-EDTA in Watera compound bis(C)-EDTA
T (K)
cmc (mM)
ΔGoM (kJ mol-1)
ΔHoM (kJ mol-1)
ΔSoM (J K-1mol-1)
283 298 313
10.63 12.18 14.34
-15.6 -15.9 -16.1
-10.8 17.1 16.9
17.0
-20.5 -10.1 -9.9
-9.5
283 4.89 -17.8 298 6.85 -17.5 313 8.92 -17.4 a Values of cmc were obtained from fluorescent measurements in the presence of pyrene.
bis(DC)-EDTA
accepting that β is independent of T, eq 17 is derived. DðΔGoM =TÞ o ΔHM ¼ Dð1=TÞ P 1 ijzi j þ jjzj j D ln cmc ¼ -R þ β j jjzþ j Dð1=TÞ P
ð17Þ
Finally, the values of the entropy of micellization per mol of BS residue are obtained from ΔSoM =(ΔHoM - ΔGoM)/T. So, for the calculation of the thermodynamic quantities, the knowledge of β is required. Its value can be estimated from the variation of the cmc with the ionic strength of the solution (eq 18).52 log½Qcmc = log cmc ¼
1 ½micelles log -β log½X N K
(52) Carey, M. C.; Small, D. M. J. Colloid Interface Sci. 1969, 31, 382.
9042 DOI: 10.1021/la9007813
ð18Þ
From cmc values in Table 1, values of β=0.46 and β=0.48 are obtained for bis(DC)-EDTA and bis(C)-EDTA, respectively. The results are in agreement with those of natural bile salts, for which values in the range β=0.3-0.6 have been published.45,53-59 It must be remembered that the palisade layer, in the sense given to it in classical surfactant micelles, does not exist in bile salt aggregates.45,60 (53) Coello, A.; Meijide, F.; Rodrı´ guez Nun~ez, E.; Vazquez Tato, J. J. Pharm. Sci. 1994, 83, 828. (54) Galantini, L.; Giampaolo, S. M.; Mannina, L.; Pavel, N. V.; Viel, S. J. Phys. Chem. B 2004, 108, 4799. (55) Galantini, L.; Giglio, E.; Leonelli, A.; Pavel, N. V. J. Phys. Chem. B 2004, 108, 3078. (56) Leggio, C.; Galantini, L.; Zaccarelli, E.; Pavel, N. V. J. Phys. Chem. B 2005, 109, 23857. (57) Cozzolino, S.; Galantini, L.; Leggio, C.; Pavel, N. V. J. Phys. Chem. B 2005, 109, 6111. (58) Matsuoka, K.; Moroi, Y. Biochim. Biophys. Acta 2002, 1580, 189. (59) Matsuoka, K.; Suzuki, M.; Honda, C.; Endo, K.; Moroi, Y. Chem. Phys. Lipids 2006, 139, 1. (60) Zana, R.; G€uveli, D. J. Phys. Chem. 1985, 89, 1687.
Langmuir 2009, 25(16), 9037–9044
Alvarez Alcalde et al.
Article
Figure 6. (a) Inverse of apparent molar mass and (b) apparent diffusion coefficient of bis(DC)-EDTA (filled symbols) and bis(C)-EDTA (open symbols) aggregates in aqueous solutions at [NaCl] 0.6 (circles) and 1.0 M (squares). In panel b, lines are guides for the eye.
Once β is known, ΔGoM can be obtained. Values are given in Table 2. It can be noticed that, within experimental error, ΔGoM is fairly constant for both surfactants with average values of -15.9 ( 0.3 [bis(C)-EDTA] and -17.6 ( 0.2 kJ mol-1 [bis(DC)EDTA]. ΔGoM is 1.7 kJ mol-1 more favorable for bis(DC)-EDTA than for bis(C)-EDTA, following the same tendency as in natural bile salts. Blume et al.61,62 have measured the thermodynamic parameters of demicellization of NaC and NaDC, by isothermal titration calorimetry. We will refer to values reported in the later paper since the authors consider that they are more precise. For both bile salts, ΔGdemic (demic=demicellization) slowly increases with temperature, although this increment represents a maximum deviation of only 8% from the average values of ΔGdemic (21.8 and 24.9 kJ mol-1 for NaC and NaDC, respectively) for a 60 °C interval. At the same temperature, the average difference between ΔGdemic values for both bile salts is 2.9 ( 0.5 kJ/mol (around 12% from absolutes values). This indicates that micellization is favored when the steroid skeleton is more hydrophobic and the difference in Gibbs free energy for micellization between these surfactants and natural bile salts can be understood on the basis of a less hydrophobicity due to the charges in the bridge. For both dimers, Figure 5 shows that the plot of ln cmc vs T -1 is linear, within the range of temperatures studied here. Substitution of slopes and fraction of bound counterion values in eq 18 leads to values for ΔHoM shown in Table 2. However, it must be taken into consideration that the determination of the micellization enthalpy from slopes of van’t Hoff plots is very difficult and to get reliable results, cmc data have to be very accurate. Furthermore, when micelles have low aggregation numbers (see below), this accuracy cannot be obtained regardless of the method used, as Blume et al.62 have pointed out. Therefore, the hypothesis that ΔHoM is constant is probably a rough approximation since the dependence of ΔHdemic (measured by Blume et al.61,62) for natural bile salts with temperature is rather complex. Above a given temperatue (approximately 290-300 K), ΔHdemic is endothermic, ΔHdemic(NaC)
