Sugar-Based Ionic Micelles in the Dilute-to-Condensed Regime

Jul 17, 2014 - preferential interactions exist among like-conformer head- groups that can keep the ganglioside micelles in a trapped configuration. We...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Langmuir

Optimizing the Crowding Strategy: Sugar-Based Ionic Micelles in the Dilute-to-Condensed Regime Elena Del Favero,† Paola Brocca,† Valeria Rondelli,† Simona Motta,† Antonio Raudino,‡ and Laura Cantu’*,† †

Department of Medical Biotechnologies and Traslational Medicine, University of Milano, LITA, Via F.lli Cervi 93, 20090 Segrate, Milano, Italy ‡ Department of Chemical Sciences, University of Catania, Viale Andrea Doria 6, 95125 Catania, Italy S Supporting Information *

ABSTRACT: In the present study, we explore the effect of concentration on micelles made by different gangliosides, which are ionic biological glycolipids bearing multisugar headgroups with huge steric hindrance. Moreover, strong preferential interactions exist among like-conformer headgroups that can keep the ganglioside micelles in a trapped configuration. We extend the well-known ionic-amphiphiles paradigm, where local condensation and micelle crowding are matched by forming larger aggregates at increasing concentration. In fact, we force the balance between interparticle and intraparticle interactions while allowing for like conformers to modulate rebalancing. In the vast experimental framework, obtained by Small Angle X-ray scattering (SAXS) experiments, a theoretical model, accounting for a collective conformational transition of the bulky headgroups, is developed and successfully tested. It allows us to shed some light on the nature and coupling of the intermolecular forces involved in the interactions among glycolipid micelles. Energy minimization leads to complex behavior of the aggregation number on increasing concentration, fully consistent with the experimental landscape. From a biological perspective, this result could be reflected in the properties of ganglioside-enriched rafts on cell membranes, with a nonlinear structural response to approaching bodies such as charged proteins. biological glycolipids bearing unusually large headgroups).6−8 There we showed that N keeps the same value over a wide range of concentration, spanning almost all of the L1 region and then suddenly decreases under rather concentrated conditions, close to the boundary with a liquid-crystal micellar cubic phase. We argued that this anomalous behavior could be allowed for amphiphiles with bulky headgroups that may assume several spatial arrangements and undergo strong likeconformer preferential interactions. The physicochemical properties of glycoamphiphiles have sometimes been shown to be very sensitive to small details in their sugar composition, in putative connection to extensive hydrogen bonding between neighboring headgroups and with the surrounding water. Also, the sugar-headgroup conformation was hypothesized to be effective in favoring or disfavoring side−side interaction between sugar rings, also influencing the extent and kinetics of water penetration into the hydrophilic layer.9−13 Extensive experimental work has shown that structural bistability is a typical feature of ganglioside assemblies,

1. INTRODUCTION Unlike solid objects, amphiphilic micelles use to face crowding by increasing their size. As the volume fraction ϕ increases, in fact, intra-aggregate head−head repulsion is overwhelmed by interaggregate repulsion. For a given ϕ, the larger the individual micelles, the further apart they are from each other, contributing to overall energy minimization and space-filling optimization. The case of ionic surfactants, such as sodium-alkyl sulfates (SAS), is paradigmatic in that sense. At low concentration but above the critical micellar concentration, only small spherical micelles are observed. On increasing the surfactant concentration, the SAS micellar radius increases,1 rapidly approaching a value corresponding to the maximum elongation of the surfactant hydrocarbon chain. Then, aggregates change to a cylindrical shape, a structure that can virtually accommodate an infinite number of monomers.2 The increase in aggregation number N with concentration over the entire phase diagram is experimentally observed and reproduced by molecular dynamics computer simulations.3,4 This general behavior has been revisited and corrected in a previous paper,5 where we investigated, both experimentally and theoretically, the crowding behavior of micelles formed by a GM1 ganglioside (belonging to an important family of natural © 2014 American Chemical Society

Received: May 20, 2014 Revised: July 8, 2014 Published: July 17, 2014 9157

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

Table 1. Micellar Parameters in the Very Dilute Regime (Volume Fraction ϕ < 0.002, 100 mM NaCl) for Different Gangliosides in the Two Different Packing Arrangements, If Presenta A (Å2)

N GM1 GD1a GalNacGD1a GT1b

301 226 246

205 172 207

95.4 98.1 97

176

99.5 101.7 99.5 100.8

Vpol (Å3)

Vapol (Å3)

Nsial

1000 1290 1580 1580

1066 1066 1066 1066

1 2 2 3

a

Aggregation number N, area per headgroup at the interface A, polar and apolar volumes of the ganglioside molecule Vpol and Vapol, and number of dissociable sialic acid groups per molecule Nsial. Results, obtained by laser light scattering measurements, are from ref 9.

connected to the cooperative transition between two different packing arrangements for the ganglioside headgroups within the hydrophilic layer of the aggregate,14−16 displaying strong cooperativity. Both packing states are “stable”; that is, spontaneous interconversion does not occur. The cooperative transition between the two states is driven, for example, by a temperature rise.8 Nonetheless, peculiar behaviors, seemingly connected to packing bistability, have also been observed at room temperature at very high GM1 concentration, ϕ ≈ 80%, where a lamellar phase15 exists. The physical mechanism for the conformational interconversion is not yet known, but it is supposed to be connected to the balance between head− solvent and head−head interactions. When dissolved in aqueous solution, most gangliosides selfaggregate in micelles from the very dilute region (ϕ = 0.001) to the concentrated region (above ϕ = 0.25), where micelles enter a Pn3m ordered phase, extending to ϕ ≈ 0.45. As ganglioside headgroups contain at least one sialic acid residue, micelles are partially dissociated, thus undergoing electrostatic interparticle repulsion. In the very dilute region, interactions can be suppressed by increasing the ionic strength of the solution (typically to physiological values, in the range of 100 mM) so as to obtain micellar parameters in the noninteracting case, as reported in Table 1 for the gangliosides sketched in the left panel of Figure 1. In these low-interactions conditions, the packing properties of different gangliosides, all bearing the same hydrophobic ceramide moiety and increasing sugar headgroup complexity, very nicely follow the model of Israelachvili,17 predicting that when the packing parameter P is close to 1/2 even small differences in the headgroup size are hugely reflected in the aggregation number.18 N is progressively lower for gangliosides with a greater number of sugar groups. These observations concurred in suggesting that intramicellar interactions due to ganglioside interheadgroups requirements, with a prominent steric component, easily prevail over intermicellar and electrostatic interactions. Table 1 reports, when applicable, two different values for N, relative to the two spatial arrangements of ganglioside headgroups, characterized by different values for the packing parameter P, as also shown pictorially in the right panel of Figure 1. Structural bistability is a feature shared by the various gangliosides with the interesting exception of GT1b, under the same low-interactions conditions. GT1b is a very complex ganglioside, with a huge seven-sugar headgroup including three dissociable sialic acid residues (Figure 1). This suggests that different regimes could be crossed, where either the hindrance or the charge or the spatial disposition of a sugar along the headgroup is crucial to the balance between intermicellar (mainly electrostatic) and intramicellar (electrostatic, steric, preferential) requirements. In the present study, we explore the response to an increasing concentration of micelles of different gangliosides and at different ionic strengths by small angle X-ray scattering

Figure 1. Left: Molecular structure of the investigated gangliosides, displaying the same hydrophobic portion, a ceramide, but different oligosaccharide headgroups. GD1a, GalNac-GD1a, and GT1b structures can be schematically obtained by the sequential addition of neutral sugars and sialic acid units to GM1. Svennerholm nomenclature for gangliosides helps in identifying the number of dissociable sialic acid units (monosialo (M), disialo (D), and trisialo (T) gangliosides). Right: Pictorial sketch of the different packing arrangements gangliosides can assume in the micelle. The tighter (looser) packing corresponds to a larger (smaller) micelle with protruding (laying) headgroups in a thicker (shallower) hydrophilic shell.

(SAXS) experiments. The progressive increase in the steric hindrance of the headgroup and in the number of potentially dissociated residues within the ganglioside series allows us to obtain a wide experimental landscape. Extending the SAS paradigm, we force the balance between interparticle and intraparticle interactions to extreme conditions, allowing for like-conformers modulating the rebalance. In the vast experimental framework, an improved theoretical model, first developed in a previous paper,5 accounting for a collective conformational transition of the bulky headgroups, is successfully tested and allows us to shed some light on the nature and coupling of the intermolecular forces involved in the interactions among glycolipid micelles. Energy minimization leads to a complex behavior of the aggregation number, fully consistent with the experimental results.

2. EXPERIMENTAL SECTION High-purity noncommercial gangliosides (Figure 1), prepared as sodium salts, were obtained as described in the paper by Tettamanti et al.19 SAXS experiments were performed on the ID02 high-brilliance beamline at the ESRF Synchrotron Facility (Grenoble, France). Experimental details concerning measurements and data analysis20,21 are reported in the Supporting Information. 9158

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

3. MODEL SECTION 3.1. Accounting for Headgroup Conformational Bistability. In this paper we extend and improve a simple analytical model of interacting micelles5 that is able to capture the most relevant physics. At variance with the model developed in ref 5, here we decompose the total intermicellar energy into the sum of electrostatic, hydration, and dispersion components, each of them having its own strength and decay rate. However, more importantly, the decay rate is linked to the conformational population of the different headgroup arrangements (see below). This fact reduces or expands the radius of influence around the micelle in a fashion that depends on the different arrangements of the bulky gangliosides’ headgroups. Here, we report the main ingredient of the model, leaving the detailed derivation in the Supporting Information (SI). Starting from the simplest model of an isolated spherical micelle17 arising from the balance between surface energy and repulsion energy among the heads, we considered a lattice of interacting micelles. The total free energy ETOT now contains two terms: ETOT = EMIC + EINT, the former, EMIC, refers to the isolated micelle and the latter, EINT, accounts for the intermicellar interactions and is a function of the amphiphile volume fraction ϕ, molecular volume v, and surface area A. Different energy terms contribute to the intermicellar interactions: electrostatic, van der Waals, and hydration forces arising from water ordering at the micelle surface.22 The headgroups of ionic amphiphiles within a micelle are not fully dissociated, with the effective charge of the micelles being only a fraction of the aggregation number N. In the case of gangliosides, the effective charge of the micelle corresponds to 15−35% of N.23 Minimization of the total energy with respect to A yields a relationship between the optimum surface area and the micelle concentration ϕ. Lastly, by relating the mean aggregation number N to the amphiphile surface area we obtain a compact equation for a spherical micelle ⎛ γ (ϕ) ⎞3/2 ⎟ N ≈ No⎜ eff ⎝ γ ⎠

in the SI, leads to a set of two nonlinear equations for the surface area A and conformational population −1 < η < 1 (when η = 0 the micelle contains the same number of protruding and lying-down conformers). Once A has been obtained, the mean aggregation number N is easily calculated. In some limiting cases, the set of equations is reduced to a single Landau−Ginzburg-type equation describing the evolution of an “order parameter” under the effect of an external “field” H(ϕ).28 (H(ϕ) encompasses both intra- and intermicellar interactions.) As described in the SI, we calculated the evolution of the aggregation number N with the surfactant volume fraction ϕ, as shown in Figure 2. Three different regimes are observed with

Figure 2. General behavior for the variation of the aggregation number N of ganglioside micelles versus the surfactant concentration ϕ. The two curves refer to aggregation numbers of two different conformational states of the micelle. At a critical surfactant concentration ϕ = ϕcrit, the relative stability of the conformations is reversed. This causes a sharp drop in the aggregation number from the upper curve (more stable in the region ϕ < ϕcrit) to the lower curve (more stable in the region ϕ > ϕcrit). All curves have been calculated by assuming positive cooperativity among the headgroups.

increasing concentration. At very low micelle concentration, the mutual repulsion is small, and N is that of the isolated micelle. Then, on increasing ϕ, because of the intermicellar interactions, N increases to NMAX = 36πν2/AMIN3 corresponding to all of the heads in the tighter (protruding) arrangement. (Interestingly, the slope depends on the amphiphile charge and geometry, being steeper at low net charge.) Upon further increases in ϕ, however, the conformational population is expected to change in order to minimize the repulsion among the approaching micelles. The number of protruding heads should decrease, in favor of the lying-down heads. The net result is an increase in A and, consequently, a decrease in N. Since the surface of ganglioside micelles describes an overcrowded environment, the change in the conformer population exhibits a cooperative behavior, as shown in the SI. It is noteworthy that the transition-like behavior (from the upper to lower curve in Figure 2) occurs when H(ϕ) reverses its sign at a critical value of ϕ = ϕcrit. On the contrary, when H(ϕ) is always positive (or negative) over the whole range of ϕ, the variation in N is smooth. This behavior is particularly evident in the case of nonideal mixing between different headgroup arrangements. When the mixing of the amphiphiles is nearly ideal, the decrease in N with ϕ is continuous. On the contrary, when interactions among identical arrangements are stronger than those among dissimilar ones, the variation in N behaves like a first-order phase transition. Lastly, at even higher ϕ, the intermicellar repulsion is so strong that basically only the lying-

(1)

where No is the aggregation number of the isolated micelle (ϕ → 0), γ is the water−micelle interfacial tension in the limit ϕ → 0, and γeff(ϕ) is the effective interfacial tension as modified by the intermicellar interactions: γeff(ϕ) = γ + ∂EINT/∂A|A=Ao. Equation 1 shows a monotonous increase of N with ϕ, provided that the close-packing concentration, the limit for micelle−micelle interpenetration, is far. A more intriguing behavior is found if we introduce a conformational flexibility into the water−micelle interfacial region, as for gangliosides. In fact, theoretical, molecular modeling, and NMR data24−27 suggest that, along with some undeformable regions, gangliosides also have flexible spots where different allowed torsion angles result in molecular arrangements of comparable energy but different packing hindrances. This effect modifies both the intramicellar selfenergy and their mutual interactions. Since two conformers occupy different areas, a variation in their relative population will change the micelle surface area. Besides, for micelles carrying out bulky headgroups, monomers in the tighter arrangement protrude toward the aqueous medium more than those in the looser arrangement, where the head is lying nearly parallel to the micellar surface. As a consequence, the intermicelle repulsion varies accordingly. The theory, developed 9159

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

down arrangement is present. Any further reduction of the intermicellar distance does not change the conformational population. Hence, bistable micelles behave again like ideal ones, and their size might increase further with concentration. 3.2. Ionic Strength in the Intramicellar and Intermicellar Bargain. We also investigated the salt effect on N. Both intra- and intermicellar terms modulate the salt effect. Intramicellar effects arise from the salt modulation of the repulsion among the charged headgroups belonging to the same micelle. In isolated micelles, the net salt effect is a decrease in the surface area and then an increase in N (e.g., a recent molecular dynamics simulation on SDS anionic micelles29). Roughly speaking, classical electrostatics shows that the repulsion energy of an isolated charged micelle embedded in an electrolyte solution of concentration c ̅ behaves as c−1/2 ̅ . By assuming that the most relevant contribution to the head−head interaction (eq 10 in the SI) is electrostatic in nature, we find that the surface area Ao per lipid molecule in the limit of noninteracting micelles behaves as Ao ≈ c−1/4 ̅ . By relating Ao to N by eq 1 in the SI, we obtain a simple scaling relationship (green curve of Figure 3) N ≈ c̅ ϕ→ 0

Figure 4. Experimental SAXS spectra relative to different ganglioside micellar solutions (tighter packing) at increasing ganglioside volume fraction (curves from bottom to top in each panel). GalNAc-GD1a: 0.0037, 0.0074, 0.015, 0.037, 0.074, 0.11, 0.149. GD1a: 0.0013, 0.0038, 0.0077, 0.015, 0.038, 0.077, 0.115, 0.153, 0.19. GT1b: 0.0013, 0.0037, 0.0074, 0.015, 0.037, 0.074, 0.11, 0.149, 0.209, 0.223. GM1: 0.001, 0.004, 0.04, 0.08, 0.12, 0.22, 0.25 (GM1: data from ref 5). All of the spectra show a pronounced minimum at some qmin, related to the micelle size.

3/4

(2)

dilute solution, high ionic strength), S(q) = 1 for all q’s, and I(q) is simply proportional to P(q). As an example, the intensity spectrum relative to GT1b micelles in the noninteracting limit is reported, together with the fit obtained with the form factor of an oblate ellipsoid in Figure S1. For a globular aggregate made of a given amphiphile, the q position of the first pronounced minimum of its form factor is related to the particle size. At a glance, prior to detailed analysis, if qmin does not change, then the micellar size is very likely constant. A look at Figure 4 immediately suggests that the size of GM1 micelles does not change over the entire L1 phase, the disordered micellar phase, from the very dilute region up to ϕ = 0.25, covering almost three orders of magnitude in concentration. The same is also observed for GalNacGD1a micelles in the range of 0.004 < ϕ < 0.15. Spectra relative to GD1a and GT1b display more complex behavior. A more focused look at experimental spectra obtained for the GT1b micellar system is reported in the Supporting Information. The values of N, as derived by fitting the experimental I(q) of Figure 4, are reported in the four panels of Figure 5 as a function of ganglioside volume fraction. 4.2. Intermicellar Distance. Independently but of course consistently, the behavior of the average intermicellar distance, derived from the structure factors of the micellar solutions, confirms the same evolution of the aggregation number for different gangliosides. The structure factor S(q) can be obtained by dividing I(q) by the particle form factor P(q) for each sample. All of the experimental structure factors relative to the different ganglioside micellar solutions (GD1a, GalNacGD1a, GT1b) in the one-decade range of 0.024 < ϕ < 0.2 at T = 25 °C, recovered from the spectra of Figure 4, are reported in Figure S5 of the Supporting Information. The detection of the structure-factor peak at low ϕ’s reveals that intermicellar

Figure 3. Aggregation number, N, versus salt concentration c ̅ for a semidilute solution (ϕ = 0.15) of charged micelles. Green curve: contribution of the intramicelle interactions; blue curve: contribution of the intermicellar interactions. The resulting behavior arises from the superposition of green and blue curves.

Intermicellar interactions lead to an opposite behavior: salt weakens the electrostatic repulsion, the micelle surface area increases, and the aggregation number N decreases (blue curve of Figure 3). Combining intermicellar and intramicellar effects, we get the blue curve. The overall behavior is almost linear in salt concentration c.̅ At low salt concentrations, however, the increase is smoother and a shallow minimum could be observed.

4. RESULTS 4.1. Micellar Size. SAXS spectra were obtained for micellar solutions of different gangliosides, namely, GD1a, GalNacGD1a, GT1b, and GM1 (sketched in Figure 1) on increasing concentration, as summarized in Figure 4. Micelles were prepared to be obtained in their tighter-packing arrangement in pure water solution (no added salt). The intensity spectra bear the fingerprints of both the form factor of the micelle, P(q), and the structure factor of the solution, S(q), with the total intensity being I(q) ≈ P(q) × S(q). In the noninteracting limit (very 9160

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

GT1b micelles in the very dilute region. Figure 7 reports the aggregation number of GT1b micelles at ϕ = 0.0015 as a

Figure 5. Micellar aggregation number N as a function of ganglioside volume fraction ϕ (tighter packing, protruding conformation). Very low and low surfactant concentration: the micellar aggregation number N at first (ϕ < 0.015) eventually increases (GD1a and GT1b) with ϕ (yellow region) and reaches a plateau value, extending up to 0.15−0.2 volume fraction for all of the investigated gangliosides (white region). Intermediate surfactant concentration: a reduction in N occurs at a concentration that depends on the specific gangliosidestill in the disordered micellar L1 phase for both GD1a and GT1b (green region) and close to the boundary with the liquid-crystalline micellar cubic phase for GM1.

Figure 7. Aggregation number N of GT1b micelles at very low ϕ (ϕ = 0.0013) as a function of c3/4 ̅ , where c ̅ is the molar concentration of added salt (0−100 mM NaCl), in the two packing arrangements: tighter packing (diamonds) and looser packing (dots). In the range of 0−10 mM NaCl, the dependence of the aggregation number N on c3/4 ̅ is linear.

interactions are significant in this regime. Analogous results were obtained for GM1.5 The position of the peak, qmax, reflects a characteristic length l that scales with the average distance of the interacting neighboring micelles in solution. In Figure 6, the

function of added NaCl (in the range of 0−100 mM). The variation of the ionic strength of the solution has a double effect on the aggregation behavior of the GT1b ganglioside: (1) different from what observed at high ionic strength (100 mM NaCl, Table 1), at each lower ionic strength GT1b micelles can adopt two different packing structures and (2) in both aggregative states, an increase in the ionic strength results in a growth of the aggregation number N of the GT1b micelles. In Figure 7, the values of N, for the two GT1b packing arrangements, are reported as a function of c3/4 ̅ , where c ̅ is the molar concentration of added salt. A range of ionic strength can be identified (0−10 mM NaCl) where the aggregation number N grows as c3/4 ̅ . For higher c,̅ the dependence of N on c ̅ becomes less pronounced until, at 100 mM NaCl, N reaches a unique plateau value, and packing bistability is no longer observed.

Figure 6. Swelling behavior of the inverse of the characteristic length l, corresponding to the scaling of the structure factor peaks position qmax, versus ϕ for GD1a (blue diamonds), GalNac-GD1a (green squares), and GT1b (red triangles) on the log−log scale. The fitting lines have the same slope s = 1/3, as expected for the three-dimensional swelling of globular objects of fixed size l/ϕ1/3. The two points that deviate from the expected behavior correspond to GT1b at ϕ = 0.209 and 0.223 (intermediate surfactant concentration).

5. DISCUSSION The experimental landscape is wide. It has been carefully designed to extend the sodium-alkyl sulfate (SAS) paradigm to include interheadgroup interactions of a different nature and major importance. Besides, we force the interplay among different interactions to the boundaries where each one is expected to become the leading one. In this respect, the family of gangliosides offers a playground where electrostatic, steric, and preferential interactions can be effectively and finely tuned. First, the optimal aggregation number that ganglioside micelles adopt at each volume fraction is strongly affected by the headgroup conformational bistability. In fact, the preferential like−like interactions occurring among headgroups with similar conformations concur in defining the intraparticle interactions (together with electrostatic, steric, nad hydrophobic contributions) that balance interparticles interactions (mainly electrostatic) in optimizing the space-filling geometry. In the bistable ganglioside micelles, molecules have an additional degree of freedom, the collective switch of their packing arrangement, in order to minimize the total energy. The virtual lower limit for the aggregation number of

corresponding inverse l values are reported versus ϕ on a log− log scale. In this representation the slope s of the fitting line corresponds to the swelling exponent of the collection of particles, l−1/ϕs. The swelling behavior is three-dimensional (slope = 1/3) for globular objects of constant size. Results confirm that GD1a and GalNacGD1a micelles maintain the same size and N for 0.037 < ϕ < 0.19. For GT1b micelles, instead, the swelling behavior deviates from what is expected for three-dimensional crowding of given particles at ϕ = 0.209 and 0.223, standing for a reduction in micelle size at these volume fractions. 4.3. Influence of the Ionic Strength. Finally, we investigated the effect of ionic strength on the tricharged 9161

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

gangliosides is Nc = 56, according to the scheme of Israelachvili.17 Second, as the cmc of gangliosides (∼10−8 M) is much lower than that of SDS (∼8 × 10−3 M), a true low micelle concentration together with a small residual ionic strength can be attained. In the present experiments, the lowest micelle concentration is ∼10 μM, and the ionic strength coming from micelle dissociation is ∼100 μM in the absence of added salt, with a negligible contribution coming from dissociated free monomers, ∼10 nM. At this concentration, the center-to-center intermicellar distance is on the order of 100 nm, say, 10 times the ganglioside micelle size and 50 times the micelle hydrophilic thickness. Third, ganglioside micelles do not display any propensity toward rodlike elongation, a geometry that can virtually settle an infinite number of monomers while preserving high local curvature. This makes the tuning most tricky, as the system is not likely to escape from uncomfortable crowding in the usual way. Despite the care in logical design and the wide experimental set, the results are not straightforward. Looking at Figure 5, only an extended central steady regime is common to all of the investigated gangliosides. At very low surfactant concentration, ϕ < 0.015, GD1a and GT1b display a growing regime where N increases with ϕ before reaching the central plateau. The stationary regime, found also for GM1 and GalNacGD1a, extends up to ϕ ≈ 0.15−0.2, say, for low surfactant concentration. Beyond ϕ ≈ 0.2, the intermediate surfactant concentration range is attained, where micelles come close to one another to intermicellar distances on the order of the micellar size (hydrodynamic diameter ∼10−12 nm). At some point in this regime (the actual concentration depending on the specific ganglioside) a reduction of N occurs. For GD1a and GT1b, it occurs still in the L1 micellar disordered phase. For GM1, instead, the sharp N reduction occurs at ϕ ≈ 0.27 (Figure 5), where intermicellar interactions are so strong that the system crosses the boundary from L1 to a liquid-crystalline cubic phase (namely, Pm3n), as shown in a previous work.30 In the cubic region, the average aggregation number of GM1, as calculated according to the Fontell model for Pm3n,31 is dramatically lower than in the L1 phase. For GM1, moreover, after the sharp transition, a slight increase in N is observed along the cubic phase. A rational hint to the experimental set comes from the careful comparison with the theoretical model that was especially developed to account for conformational bistability and strong like-conformer preferential interactions occurring in gangliosides. Comparing the results on ganglioside micelles shown in Figures 2 and 5, a clear correspondence, albeit qualitative, between the theoretical model and the experimental data is seen. The interaction balance can be different for different gangliosides. The variation of molecular parameters, on going from GM1 to GT1b (sketched in Figure 1), does not change the general shape of the curve; rather, it shifts the different regimes at different ϕ values. 5.1. Growth Regime, Very Low Concentration. At very low ϕ, the model predicts a monotonous increase in N with ϕ, which is faster at low micelle net charge. Parallel experimental results on the two gangliosides GalNac-GD1a and GT1b, with the same number of sugars but different numbers of ionizable units, show that the growing N regime is quite wide in the case of the more charged GT1b (up to ϕ = 0.015) while it is out of the investigated ϕ range for the less-charged GalNac-GD1a or

maybe compressed to a more dilute region. For GM1 as well, with just one dissociable unit in the headgroup, two less than GT1b, this region is out of the experimental range. 5.2. Stationary Regime, Low Concentration. This is the region where preferential interactions among tight-packing like conformers with strong steric hindrance keep the micelles of all gangliosides in a packing trap, highly resistant to interparticle interactions. In this regime, the aggregation number of micelles of different gangliosides is highly indicative of monomer hindrance for the smallest details. 5.3. Intermediate Regime and the N Drop. At intermediate concentration, the intermicellar distances become comparable to the micellar diameters. The average surface-tosurface distance is different for different ganglioside micelles at the same ϕ, as it depends on N and on the thickness of the headgroup corona. From the experimental data, we can place the intermediate region at ϕ ≥ 0.15. At this ganglioside volume fraction, we can estimate that the center-to-center intermicellar distance is ∼17 to 18 nm for GalNac-GD1a, GD1a, and GT1b, to be compared to the micellar hydrodynamic diameter, in the range of 10−12 nm. For GM1, the same intermicellar distance is reached at higher volume fraction (ϕ ≈ 0.22). Above this ϕ, the surface-to-surface semidistance between two neighboring micelles becomes progressively comparable to the extension of the ganglioside headgroup, maybe smaller somewhere because of the nonspherical shape of the micelles. Thus, in this range, a concentration is reached where the repulsion among the crowding micelles can be reduced, if applicable, by pushing down the heads on the micelle surface so as to increase the surface-to-surface distance. Gangliosides, at some point in this regime, change the conformational population of the heads, cooperatively adopting the lying-down monomer instead of the protruding-monomer packing. Correspondingly, N decreases through a first-order phase transition, as predicted by the model for nonideally mixing headgroups, with strong like-conformer preferential interactions. This sudden decrease in N at some point for increasing ϕ in the intermediate regime, thus increasing the number of charged micelles, is counterintuitive and opposite to the SAS paradigm (larger micelles at higher concentration). It suggests that the distance among charged loci, as measured on the interaggregate and intraaggregate scale, has come to a turning point. A rough quantitative estimate can be made, neglecting the possibility of changing the degree of dissociation of the headgroups. In the case of the three-charge GT1b, the 1 nm reduction in the interparticle distance, induced by the N drop from 165 to 135, as observed, can be compensated for by the reduction in the protrusion of the heads. On the other hand, with the charged groups distributed along the headgroup, the conformational change could alter or not alter the mutual arrangement of the charges within the micelle corona, a central parameter in defining the optimal interfacial area per headgroup at all concentrations. As an example, in the two-charge GD1a micelle, two charges belonging to two adjacent monomers are ∼0.8 nm apart, as well as the two charges along the same sugar chain. So bistable ganglioside micelles face crowding through a peculiar alternative route. The sharp transition in N occurs at lower ϕ for more charged and bulky heads. For the three-charged GT1b, the drop in N occurs at ϕcrit ≈ 0.2, in the L1 disordered micellar phase. For the singly charge GM1, with a right-shifted intermediate regime, the drop in N occurs at higher ϕ, at the boundary with the micellar ordered cubic phase. Nonetheless, predicting the ϕcrit for any specific ganglioside is not immediate, 9162

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

forces. The study concerning this complex issue is actually in progress. In particular, the role of multiple charged sites along the ganglioside headgroups is an interesting issue as it results in charges being distributed within the hydrophilic shell of the micelle, with its discreteness becoming important on a short length scale. Apart from its implications in defining the amphiphile phase diagram, the role of charge decoration of the headgroup coupled to its conformational bistability is interesting in connection to lipid rafts, where gangliosides are enriched. Residing in the outer leaflet of plasma membranes, ganglioside-enriched domains constitute a platform for cell−cell recognition, protein receptor functions, coordination seeding, and structural templating.32−34 Interface reshaping forced by an approaching body could imply a modification in the local dipole, favoring locking, thus stabilizing temporary interactions or forcing packing on a structurally encoded matrix.

being connected to the headgroup hindrance and the number of charges but surely also depending on their spatial disposition and degree of dissociation. 5.4. Ionic Strength and the Transition State. Finally, let us focus on another extreme boundary for complex interaction interplay, that is, the three-charge GT1b in the very dilute regime and vanishing ionic strength. As noted above, the extreme dilution and the extremely low cmc of gangliosides allow for reaching a very low residual ionic strength of the solution due to counterion and co-ion concentrations. We recall (Table 1) that GT1b at ϕ ≈ 0.001 in 100 mM added salt does not exhibit any bistable behavior, suggesting that, whatever the extent or type of interaction, just one type of packing is allowed.6 This fact, clearly contrasting with the behavior displayed by other gangliosides, suggested that the huge hindrance of the GT1b headgroup was forcing the widest possible packing, providing enough room for large and repelling charged groups. The present experimental results extend the scenario to a crucial balance between electrostatics and hindrance. In fact, due to the high number of charges of GT1b, the decrease in ionic strength to very low values (below 10 mM) makes intraparticle interactions very dependent on the ionic strength itself. This is proved by the c3/4 ̅ behavior of N versus the ionic strength, as expected for interacting amphiphiles (Figures 3 and 7). On the other hand, under these conditions, the GT1b micelle can be trapped in either lying-down or protruding conformations, revealing again the role of like-conformer interactions in determining N. This suggests that the energy level of the conformational transition state for ganglioside monomers, to be crossed in order to pass from protruding to lying down, is not only lowered by cooperativity but is also modulated by the whole set of interactions, including electrostatic.

■

ASSOCIATED CONTENT

S Supporting Information *

Detailed experimental section. Derivation of the analytical model of interacting micelles in the case of amphiphiles displaying headgroup conformation bistability. Focus on experimental results on GT1b. Experimental structure factors relative to the different gangliosides micellar solutions (GD1a, GalNacGD1a, and GT1b). This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +39 02 503 30362. Fax: +39 02 503 30365. Notes

The authors declare no competing financial interest.

■

6. CONCLUSIONS Walking across the micellar phase (L1) of different complex gangliosides, we explored, both experimentally and theoretically, a rich interaction landscape where multiple crossover between intramicellar and intermicellar predominance occurs. Gangliosides are anionic glycolipids where one or more charged sites decorate bulky flexible sugar headgroups that display a strong preferential interaction with like-conformer neighbors on the condensed curved surface of a micelle. We found that, on crowding, ganglioside micelles do not follow the smooth behavior usually found for ionic amphiphiles (the SAS paradigm). We developed a theoretical model parallel to this experimental behavior by assuming the predominance of either interaction in different regimes. Long-range and short-range contributions remix on interparticle distances on the order of the micellar size, producing a stepwise transition contrasting with the usual crowding rule. The occurrence and position of different regimes depend on the specific ganglioside, in agreement with the theoretical scheme. In order to further explore the basis of the collective behavior of ganglioside headgroups, detailed molecular dynamics simulations have been undertaken with two main goals: (a) to prove the variation of the intermicellar repulsion upon a conformational transition from a lying-down to a protruding arrangement of the headgroups (through, for example, a displacement of the charge distribution within the micelle corona); (b) to unveil the cooperative response of the headgroup region to tiny variations of the intermolecular

ACKNOWLEDGMENTS We thank T. Narayanan for his valuable support at the ESRF ID02 beamline and Prof. S. Sonnino for generously supplying huge amounts of highly purified noncommercial gangliosides.

■

REFERENCES

(1) Quina, F.H.; Nassar, P.M.; Bonilha, J. B. S.; Bales, B.L. Growth of sodium dodecyl sulfate micelles with detergent concentration. J. Phys. Chem. 1995, 99, 17028−17031 (and references therein). (2) Mazer, N.A.; Benedek, G.B.; Carey, M.C. An Investigation of the Micellar Phase of Sodium Dodecyl Sulfate in Aqueous Sodium Chloride Using Quasielastic Light Scattering Spectroscopy. J. Phys. Chem. 1976, 80, 1075−1085. (3) Nelson, P.H.; Rutledge, G.C.; Hatton, T.A. On the size and shape of self-assembled micelles. J. Chem. Phys. 1997, 107, 10777−10781. (4) Lazaridis, T.; Mallik, B.; Chen, Y. Implicit solvent simulations of DPC micelle formation. J. Phys. Chem. B 2005, 109, 15098−15106. (5) Brocca, P.; Cantu’, L.; Corti, M.; Del Favero, E.; Raudino, A. Intermicellar interactions may induce anomalous size behaviour in micelles carrying out bulky heads with multiple spatial arrangements. Langmuir 2007, 23, 3067−3074. (6) Sonnino, S.; Mauri, L.; Chigorno, V.; Prinetti, A. Gangliosides as components of lipid membrane domains. Glycobiology 2007, 17, 1R− 13R. (7) Ledeen, R.; Wu, G. New findings on nuclear gangliosides: overview on metabolism and function. J. Neurochem. 2011, 116, 714− 720. (8) Kolter, T. Ganglioside Biochemistry. ISRN Biochem. 2012, Article ID 506160, doi:10.5402/2012/506160.

9163

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164

Langmuir

Article

(9) Köberl, M.; Schöppe, A.; Hinz, H.- J.; Rapp, G. Cooperativity of glycolipid phase transitions: a critical comparison of results from equilibrium X-ray, density and microcalorimetric measurements. Chem. Phys. Lipids 1998, 95, 59−82. (10) Hinz, H.-J.; Kuttenreich, H.; Meyer, R.; Renner, M.; Fründ, R.; Koynova, R.; Boyanov, A.I.; Tenchov, B.G. Stereochemistry and size of sugar head groups determine structure and phase behavior of glycolipid membranes: densitometric, calorimetric, and X-ray studies. Biochemistry 1991, 30, 5125−5138. (11) Mannock, D.A.; McElhaney, R.N. Thermotropic and lyotropic phase properties of glycolipid diastereomers: role of headgroup and interfacial interactions in determining phase behaviour. Curr. Opin. Colloid Interface Sci. 2004, 8, 426−447. (12) Hato, M.; Minamikawa, H.; Tamada, K.; Baba, T.; Tanabe, Y. Self-assembly of synthetic glycolipid/water systems. Adv. Colloid Interface Sci. 1999, 80, 233−270. (13) Corti, M.; Cantù, L.; Brocca, P.; Del Favero, E. Self-assembly in glycolipids. Curr. Opin. Colloid Interface Sci. 2007, 12, 148−154. (14) Cantu’, L.; Corti, M.; Del Favero, E.; Digirolamo, E.; Sonnino, S.; Tettamanti, G. Experimental evidence of a temperature-related conformational change of the hydrophilic portion of gangliosides. Chem. Phys. Lipids 1996, 79, 137−145. (15) Cantu’, L.; Corti, M.; Del Favero, E.; Raudino, A. Tightly Packed Lipid Lamellae with Large Conformational Flexibility in the Interfacial Region May Exhibit Multiple Periodicity in their Repeat Distance. A Theoretical Analysis and X-ray Verification. Langmuir 2000, 16, 8903−8911. (16) Cantu’, L.; Corti, M.; Del Favero, E.; Raudino, A. Bistable Molecular Selfassembly. Curr. Opin. Colloid Interface Sci. 2000, 5, 13− 18. (17) Israelachvili, J. N.; Mitchell, D. J.; Ninham, B. W. Theory of selfassembly of hydrocarbon amphiphiles into micelles and bilayers. J. Chem. Soc., Faraday Trans. 2 1976, 72, 1525−1568. (18) Cantu’, L.; Del Favero, E.; Brocca, P.; Corti, M. Multilevel structuring of ganglioside-containing aggregates: from simple micelles to complex biomimetic membranes. Adv. Colloid Polym. Sci. 2014, 205, 177−186. (19) Tettamanti, G.; Bonali, F.; Marchesini, S.; Zambotti, V. A new procedure for the extraction, purification and fractionation of brain gangliosides. Biochim. Biophys. Acta 1973, 296, 160−170. (20) Klein, R. Interacting Colloidal Suspensions. In Neutrons, X-rays and Light: Scattering Methods Applied to Soft Condensed Matter; Lindner, P., Zemb, Th., Eds.; Elsevier Science B.V.: Amsterdam, 2002; pp 351−379. (21) Pedersen, J. S. Modelling of Small-Angle Scattering Data from Colloids and Polymer Systems. In Neutrons, X-rays and Light: Scattering methods applied to soft condensed matter; Lindner, P., Zemb, Th., Eds.; Elsevier Science B.V.: Amsterdam, 2002; pp 391− 420. (22) Leikin, S.; Parsegian, V.A.; Rau, D.C.; Rand, R.P. Hydration Forces. Annu. Rev. Phys. Chem. 1993, 44, 369−395. (23) Cantu’, L.; Corti, M.; Degiorgio, V.; Piazza, R.; Rennie, A. Neutron scattering from ganglioside micelles. Prog. Colloid Polym. Sci. 1988, 76, 216−220. (24) Venkateshwari, S.; Veluraja, K. Molecular Modelling and Molecular Dynamics studies of GD1A, GD1B and their complexes with BoNT/B − Perspectives in interaction and specificity. J. Struct. Biol. 2012, 180, 497−508. (25) Rodgers, J. C.; Portoghese, P. S. Molecular modeling of the conformational and sodium ion binding properties of the oligosaccharide component of ganglioside GM1. Biopolymers 1994, 34, 1311− 1326. (26) Bock, K. The Solution Conformation of Gangliosides Inferred from HSEA Calculations and High Field NMR Spectroscopy. FIDIA Res. Ser. 1986, 6, 47−56. (27) Acquotti, D.; Poppe, L.; Dabrowski, J.; von der Lieth, C.; Sonnino, S.; Tettamanti, G. Three-Dimensional Structure of the Oligosaccharide Chain of GM 1 Ganglioside Revealed by a DistanceMapping Procedure: A Rotating and Laboratory Frame Nuclear

Overhauser Enhancement Investigation of Native Glycolipid in Dimethyl Sulfoxide and in Water-Dodecylphosphocholine Solutions. J. Am. Chem. Soc. 1990, 112, 7172−7118. (28) Landau, L. D.; Lifshitz, E. M. Statistical Physics; Pergamon Press: New York, 1985. (29) Sammalkorpi, M.; Karttunen, M.; Haataja, M. Ionic surfactants in saline solutions: Sodium Dodecyl Sulphate (SDS) in the presence of excess NaCl or CaCl2. J. Phys. Chem. B 2009, 113, 5863−5870. (30) Boretta, M.; Cantu’, L.; Corti, M.; Del Favero, E. Cubic Phases of Gangliosides in Water: Possible Role of the Conformational Bistability of the Headgroup. Physica A 1997, 236, 162−176. (31) Fontell, K.; Fox, K.K.; Hansson, F. On the structure of the cubic phase II in some lipid-water systems. Mol. Cryst. Liq. Cryst. Lett. 1985, 1, 9−17. (32) Lingwood, D.; Simons, K. Lipid rafts as a membrane-organizing principle. Science 2010, 327, 46−50. (33) Yoshizaki, F.; Nakayama, H.; Iwahara, C.; Takamori, K.; Ogawa, H.; Iwabuchi, K. Role of glycosphingolipid-enriched microdomains in innate immunity: microdomain-dependent phagocytic cell functions. Biochim. Biophys. Acta 2008, 1780, 383−392. (34) Matsuzaki, K.; Kato, K.; Yanagisawa, K. Abeta polymerization through interaction with membrane gangliosides. Biochim. Biophys. Acta 2010, 1801, 868−877.

9164

dx.doi.org/10.1021/la501963y | Langmuir 2014, 30, 9157−9164