Jörn
Schmidt-Lassen
and
Thisbe K.
Lindhorst
*
Otto Diels Institute of Organic Chemistry, Christiana Albertina University of Kiel, Otto-Hahn-Platz 3-4, D-24098 Kiel, Germany. E-mail: tklind@oc.uni-kiel.de
First published on 30th May 2014
A series of octyl α- and β-glycosides of the manno- galacto- and gluco-series were synthesized and employed in formation of homo- and hetero-micelles in water. Critical micelle concentrations (cmc), thermodynamic quantities of demicellation and, to some extent, the hydrodynamic radii of glycomicelles were determined by isothermal titration calorimetry (ITC) and diffusion NMR studies. The goal of this work was to determine the significance of anomeric configuration as well as of epimerisation at the sugar ring for supramolecular features of the respective glycoside. A new projection of glycoside structures is proposed to facilitate interpretation of structure–property relationships in this regard.
Glycomicelles are robust and well investigated supramolecular systems that have been used earlier in biological chemistry6 such as for the study of self-assembly properties of glycolipids, for example.7–10 They are amenable to different biophysical techniques, e.g. surface tension measurements, fluorescence spectroscopy, or various scattering methods (DLS, SANS, SAXS).8 In addition, methods that allow to measure kinetic and thermodynamic parameters of micelles are of relevance in a biological context.
Here, a collection of six different octyl glycosides were synthesized and employed (Fig. 1) in isothermal calorimetry titrations and NMR-spectroscopic studies, including well-known octyl β-glucoside (6), which is commercially available and regularly used for the formation of glycomicelles. Two series of octyl glycosides were tested, α-D-mannoside (1), α-D-galactoside (3), and α-D-glucoside (5), and the respective β-glycosides 2, 4, and 6. Hence, the influence of the anomeric configuration on micelle formation can be compared, as well as the meaning of inversion of configuration at C2 and C4 of the sugar ring, respectively. In addition, mixed, so-called hetero-glycomicelles were formed and compared with the respective homo-glycomicelles.
Fig. 2 Integrated ITC titration curves of homo-glycomicelles obtained at 298 K. Effect of anomeric configuration: (a) α- and β-octyl mannoside micelles formed of 1 and 2; (b) α- and β-octyl galactoside micelles formed of 3 and 4. Effect of configuration of the sugar ring: (c) α-anomeric micelles formed of 1 and 3, respectively (5 was not sufficiently soluble in water); (d) β-anomeric micelles formed of 2, 4, and 6, respectively. Representative curves are shown. ΔRH: reaction enthalpy per injection. ITC starting concentrations c and obtained mean cmc values are listed in the table insert; SD: standard deviation. (a) 6.0;19 (b) 16;19 (c) 6.3,18 12;19 (d) 20,19 23,16 25.8.20 |
Thus, the effect of anomeric configuration on micelle formation could be studied in the manno- as well as the galacto-series (Fig. 2a and b). No significant difference between the α- and β-glycomicelles was detected in the galacto-series, but the manno-series, on the other hand, exhibits a distinctive effect with a much lower cmc for the α-micelle (10.9 mM) than for the respective β-micelle (22.9 mM). Accordingly, formation of α-mannomicelles starts at ∼half the concentration which is required for formation of the corresponding β-micelle. In the gluco-series, we determined a cmc of 28.3 mM for the β-glucoside. Literature reports on cmc values of octyl glucosides diverge to some extent, which can be due to different determination methods employed. Focher18 and Matsumura19 and co-workers used a surface tension method for cmc determination, where effects of dilution and all intermolecular interactions are not taken into account. Consequently lower cmc values were found. Focher et al. determined the cmc value for the β-glucoside 6 as 19 mM and for the α-glucoside 5 as 6.3 mM.18 Paula et al., on the other hand, used titration calorimetry and obtained a cmc of 23.0 mM (ref. 16) for the β-glucoside 6. Regardless of these differences, the gluco-series shows, in analogy to the manno-case, that α-glucomicelles are formed at lower concentrations than β-glucomicelles. This effect is less pronounced in case of the galactoside micelles.
Next, the influence of configurational characteristics of the sugar ring on micelle formation was studied (Fig. 2c and d). The two α-glycosides 1 and 3 show a significant difference in glycomicelle formation, with the cmc of mannoside 1 (10.9 mM) being much smaller than the cmc of galactoside 3 (30.2 mM). In the series of the three β-glycosides 2, 4, and 6, again the cmc of the galactoside (4, 31.7 mM) was found to be the highest. Overall, within both series of α- and β-octyl glycosides a carbohydrate-specific dependency of cmc values was observed, namely cmc(αMan) ≈ cmc(αGlc) < cmc(αGal), and cmc(βMan) < cmc(βGlc) < cmc(βGal).
Next, thermodynamic data of micelle formation were deduced from the measured ITC curves, using the so-called phase-separation model.16 In this model, the glycomicelle is considered as a separate microphase. At high aggregation numbers of the investigated micellar systems (assumed according to the literature16,20–22) the Gibbs free energy of demicellation is defined by the chemical potential of the amphiphile in water and the amphiphile in micellar phase (eqn (1)). To obtain the Gibbs free energy ΔG from measured cmc values, molar fractions cmc′ are pluged in.
ΔG°demic = μ0w − μ0mic = −RTlncmc′ | (1) |
Demicellation enthalpy, ΔH°demic, and demicellation entropy, ΔS°demic, are obtained from ΔG°demic according to the Gibbs–Helmholtz relation (2).
ΔG = ΔH − TΔS | (2) |
The enthalpy of demicellation (ΔH°demic) can be calculated according to eqn (3).
(3) |
For this calculation, csyringe was considered as equal to cmono + cmic. Here the approximation is used that above the cmc, the concentration of the monomers remains constant. Thus, the thermodynamic parameters of demicellation could be deduced from measured cmc values and are collected in Table 1.
αOctylMan 1 | βOctylMan 2 | αOctylGal 3 | βOctylGal 4 | βOctylGlc 6 | |
---|---|---|---|---|---|
cmc′ (×10−4) | 1.95 ± 0.13 | 4.11 ± 0.06 | 5.42 ± 0.07 | 5.68 ± 0.06 | 5.07 ± 0.09 |
lncmc′ | −8.55 ± 0.07 | −7.80 ± 0.03 | −7.52 ± 0.04 | −7.47 ± 0.03 | −7.59 ± 0.05 |
ΔG°demic [kJ mol−1] = −ΔG°mic | 21.2 ± 0.2 | 19.3 ± 0.1 | 18.6 ± 0.1 | 18.5 ± 0.2 | 18.8 ± 0.2 |
ΔH°demic [kJ mol−1] | −10.3 ± 1.7 | −9.7 ± 0.9 | −9.2 ± 1.0 | −8.4 ± 0.7 | −6.4 ± 1.2 |
TΔS°demic [kJ mol−1] | −31.5 ± 1.9 | −29.1 ± 0.8 | −27.9 ± 1.1 | −26.9 ± 1.2 | −25.2 ± 1.4 |
ΔS°demic [J K−1 mol−1] | −105.6 ± 6.5 | −97.5 ± 2.8 | −93.5 ± 3.3 | −90.2 ± 3.1 | −84.6 ± 4.6 |
The thermodynamic data summarized in Table 1 show that micelle formation is an exergonic process with negative ΔG°mic values. Enthalpy changes of micelle formation, ΔH°mic, are positive though, but enthalpy–entropy compensation23 results in an overall exergonic process. Positive ΔS°mic values indicate that micelle formation is entropically favoured, an effect which has been explained by desolvation of monomers during micelle formation.24 This behavior is observed for all investigated glycomicelles.
Here, we have employed 1:1-mixtures of two amphiphilic glycosides of the same anomeric configuration in a first series of experiments. The standard demicellation protocol was followed to obtain ITC curves of the investigated six mixed micelles, three α-hetero-micelles (Fig. 3a) and three β-hetero-micelles (Fig. 3b). Strikingly, in all cases the titration curves show exactly one point of inflection (reflecting one cmc value) indicating the formation of just one type of mixed micelle. Hence, it can be concluded that the employed 1:1-mixtures form homogeneous hetero-glycomicelles and that phase separation or homo-micelle formation is not effectively competing in this self-aggregation process.
Also in hetero-micelle formation, the investigated mannosides exerted a special effect. All mannoside-containing hetero-micelles had lower cmc values than mannoside-free hetero-micelles. In the α-series, the mixed micelles containing octyl mannoside had cmc values of 13.0 ± 0.7 mM and 15.4 ± 0.3 mM, whereas the mannose-free hetero-micelle had cmc = 25.4 ± 0.7 (Fig. 3, table insert). In the β-series, the analogous behaviour was observed. It should be noted, that calculation of cmc values of mixed micelles has been reported in the literature.30,31 In these cases, cmc values were either obtained according to a model for ideal mixing behaviour, or the employed relation was adapted for non-ideal mixing of the components using regular solution theory by Rubingh and colleagues.32
In our experiments it was striking, that it was possible to employ the weakly water-soluble octyl α-D-glucoside 5 in a binary system and to determine the cmc values of the respective hetero-glycomicelles. This demonstrates that the supramolecular properties of mixed micelles can differ significantly from pure homo-micelles. Similar effects have found technical applications in other binary systems.33
In Table 2 the thermodynamic parameters of the investigated hetero-micelles are summarized which were calculated in analogy to what has been described for the homo-glycomicelles. Again, formation of hetero-glycomicelles is entropically driven leading to overall exergonic processes.
αOctylMan 1, αOctylGlc 5 | αOctylMan 1, αOctylGal 3 | αOctylGal 3, αOctylGlc 5 | βOctylMan 2, βOctylGlc 6 | βOctylMan 2, βOctylGal 4 | βOctylGal 4, βOctylGlc 6 | |
---|---|---|---|---|---|---|
cmc′ (×10−4) | 2.34 ± 0.14 | 2.76 ± 0.06 | 4.48 ± 0.16 | 4.18 ± 0.04 | 4.39 ± 0.07 | 5.71 ± 0.06 |
lncmc′ | −8.36 ± 0.06 | −8.20 ± 0.03 | −7.71 ± 0.07 | −7.78 ± 0.01 | −7.73 ± 0.04 | −7.47 ± 0.03 |
ΔG°demic [kJ mol−1] = −ΔG°mic | 20.7 ± 0.2 | 20.3 ± 0.1 | 19.1 ± 0.2 | 19.3 ± 0.1 | 19.2 ± 0.2 | 18.5 ± 0.1 |
ΔH°demic [kJ mol−1] | −12.3 ± 0.1 | 5.2 ± 0.3 | −10.6 ± 0.1 | −8.6 ± 0.4 | −8.5 ± 0.2 | −7.2 ± 0.9 |
TΔS°demic [kJ mol−1] | −33.1 ± 0.1 | −25.2 ± 0.3 | −29.7 ± 0.1 | −27.9 ± 0.4 | −27.7 ± 0.3 | −25.7 ± 0.9 |
ΔS°demic [J K−1·mol−1] | −110.9 ± 0.4 | −85.7 ± 1.3 | −99.8 ± 0.3 | −93.4 ± 1.3 | −92.8 ± 0.9 | −86.1 ± 3.1 |
Here, the two anomeric glycosides, α- and β-octyl mannoside 1 and 2, were selected for diffusion NMR experiments, as they showed the most pronounced effect of anomeric configuration on micelle formation. For determination of the diffusion coefficients Dobs, four different regions in the 1H NMR spectra of the investigated compounds were used (δ = 4.00–3.30, 1.65–1:37, 1.36–1.00, and 0.90–0.60 ppm). Samples were measured at concentrations between 2.5 and 250 mM for 1, and 10 and 250 mM for 2, employing the PFGSE experiment and measured Dobs values plotted against sample concentration (Fig. 4). As expected,34Dobs values decrease rapidly with concentrations increasing beyond the cmc, resulting from the limited molecular motion within the micellar assembly.
All diffusion NMR experiments showed just one diffusing species. Thus, it can be concluded that the equilibrium between monomeric and micellar molecules is fast on the time scale of the NMR experiment. Consequently, the observed diffusion coefficient Dobs is a weighted average from two states of the investigated octyl mannoside (1 or 2, respectively), monomeric and micellar. Hence, Dobs can be expressed by eqn (4).34,37
(4) |
Eqn (4) is valid for concentrations >cmc, and thus only respective data points were used. Dmono was inserted into the equation as a fixed parameter that was obtained by NMR measurements at concentrations below the cmc.34ctotal was inserted as an independent variable.
However, in eqn (4) obstruction effects that occur at finite aggregation concentrations, are not taken into account. Therefore, according to Nilsson and Söderman,34eqn (4) was corrected using relation (5).
Dmic = Dmic,0 (1 − kϕmic) | (5) |
In (5), k is an interaction constant, representing dynamic interactions of hard spheres. According to the literature, k was chosen as 1.73.38ϕmic represents the volume fraction of the micelles. It was obtained using eqn (6).
(6) |
In eqn (6), M is the molecular weight of the octyl mannoside (1 or 2, respectively) and d its density (cf. ESI).
By insertion of eqn (5) and (6) into (4), expression (7) is obtained that was used for fitting of the data points of the PGSE NMR experiment (cf.Fig. 5).
(7) |
Fig. 5 Determination of cmc values and diffusion coefficients Dmono and Dmic,0 for aggregation systems formed with α- and β-mannosides 1 and 2, respectively, at 298 K. For fitting, eqn (7) and data points of ctotal > cmc were used; (a) Dmono at ctotal = 5 and 7.5 mM; (b) Dmono at ctotal = 15 mM. |
Thus, self-diffusion coefficients Dmic,0 for the micellar state of octyl mannosides could be determined and the obtained data are summarized in Fig. 5.
Both experimental methods, ITC and DOSY NMR spectroscopy, led to very similar cmc values for 1 and 2. However, the cmc values determined by diffusion NMR spectroscopy for mannoside 1 (9.1 mM) and 2, (18.8 mM) were found somewhat lower than the respective values determined by ITC (10.9 mM for 1 and 22.9 mM for 2). This finding has in principal been described before for association constants of calixpyrrols39 and can be explained by additional intermolecular interactions during demicellation occurring in the ITC experiment.
The diffusion coefficients Dmono determined for both mannosides, 1 and 2, were found equal within error limits (4.35 × 10−10 m2 s−1 for 1 and 4.21 × 10−10 m2 s−1 for 2) and similar to published data for βOctylGlc (6).40 On the contrary, the diffusion coefficients Dmic,0 of the respective micelles were found to differ significantly with Dmic,0 (1) of 3.27 × 10−11 m2 s−1 and Dmic,0 (2) of 4.42 × 10−11 m2 s−1. Hence, the α-mannosidic micelle diffuses more slowly than the β-mannosidic micelle. This corresponds to larger aggregates in the α-case than in the β-case. Thus, again a significant influence of the anomeric configuration of the mannosidic amphiphile on self-aggregation was revealed.
To obtain an estimate of the size of the investigated micelles the Stokes–Einstein eqn (8) was employed.
(8) |
In (8)Rh is the hydrodynamic radius of the micelle, η the dynamic viscosity of the used solvent, T the temperature and kB the Boltzmann constant. With the dynamic viscosity of D2O = 1.10 mNs m−2,41Rh of the α-mannosidic micelle was calculated as 60.7 Å and for the corresponding β-mannosidic micelle as 44.9 Å. Thus, the α-mannosidic micelle is significantly bigger than its β-analogue. It has to be kept in mind, however, that the Stokes–Einstein equation assumes a spherical shape for micelles, while the investigated systems might have a different form and thus other sizes than the calculated ones (vide infra).
The employed mannosides were also unique when employed in the formation of mixed micelles. Thus, the cmc value that is obtained for αOctylMan/αOctylGal hetero-micelles is significantly affected by the α-mannosidic component (Fig. 3). Equally, the cmc values obtained for βOctylMan/βOctylGlc as well as for βOctylMan/βOctylGal hetero-micelles are particularly affected by the β-mannoside component βOctylMan, hence increasing the tendency to micelle formation.
The micellar hydrodynamic radii Rh obtained for 1 and 2 using a PFGSE NMR experiment, namely 60.7 Å and 44.9 Å, are rather large compared to reported micelle sizes (∼23–27 Å).42–44 As occurring obstruction effects were taken into account in our fitting procedure, the reason for the large Rh values may be found by re-considering the shape of the investigated micelles. In the employed Stokes–Einstein equation a spherical shape of micelles is assumed, however, the herein investigated micelles might adopt a different form, probably rodlike or ellipsoidal. We will perform additional studies to specify the shape of octyl mannoside micelles and understand differences caused by the anomeric configuration (α or β).
It has become obvious in our study, that subtle structural changes such as epimerisation at C-2 of octyl glycosides lead to significant effects in a supramolecular context, including hetero-micelles However, how structural details determine the self-aggregation process, remains a secret until this day.
It has been reasoned earlier that configuration and different tilt angles of α- and β-glycosides influence the geometry as well as the hydrophilicity of glycosides, factors that can determine glycomicelle formation.18,19,45 In addition, theoretical studies have highlighted the importance of variable hydrogen bonding in glycomicelle formation of glucosides.46 MD simulations with octyl β-glucoside and octyl β-galactoside suggested that the β-galactoside has a slightly higher tendency to form inter- and intramolecular hydrogen bonds than the corresponding β-glucoside.47 How this results in lower cmc values for the glucoside is not completely clear given that micelle formation is an entropically driven process.
To facilitate the interpretation of structure–property relationships, we suggest an alternative structure view on glycosides to assess the configurational characteristics of the investigated α- and β-octyl glycosides in new light (Fig. 6). In the proposed projections, the relative positions of the hydroxy groups at C-2 and C-4 are depicted at respective edges of two fused triangles, forming the ‘front face’ of the glycoside, in relation to the octyl aglycone moiety in the rear of the molecule. Interestingly, this structural representation suggests that the glucosides can be considered as ‘transition’ between mannoside (low cmc) and galactoside (high cmc) features. The two front triangles of the mannosides on the one hand and the glucosides on the other, are mirror images of one another, whereas the relative position of the aglycone moiety at the far end of the molecule remains constant. The different positioning of the 2,4-dihydroxy group pair relative to the octyl aglycone in the α- and β-case, respectively, also becomes very obvious using the suggested projection. The depicted projection could even be further expanded by including the direction of the O–H bonds. This would demonstrate how the hydroxyl hydrogen atoms are exposed to the surrounding of the glycoside, featuring a particular hydrogen bonding behavior. The proposed projection could equally support the interpretation of self-aggregation of mixed micelles, which were shown to display different properties than the homo-micelles. In our further work, we will expand the present study in both an experimental and a theoretical context and we will make an attempt to draw inspiration from the proposed ‘trinity projection’ of glycosides. In particular, we have commenced MD simulations to understand the notable effects of octyl α- and β-mannosides on micelle formation.
Footnote |
† Electronic supplementary information (ESI) is available: Containing details of synthesis, ITC and DOSY NMR measurements. See DOI: 10.1039/c4md00122b |
This journal is © The Royal Society of Chemistry 2014 |