Influence of Long-Tailed Alcohols on the Solubilisation of Cationic DHAB Bilayers

The effects of n-decanol and cetyl alcohol incorporated into cationic di-n-hexadecyldimethylammonium bromide (DHAB) bilayers on the packing and the resistance against solubilisation by Triton X-100 have been investigated. Solubilisation experiments revealed that the mismatch between the C10-tail of n-decanol with the DHAB-tails does not affect the structural integrity of the vesicles. On the contrary, upon increasing amounts of cetyl alcohol, of which the tail matches in size with the DHAB-tails, breakdown of the vesicles is promoted, whereas complete solubilisation is strongly hampered. The system is further investigated by DSC and cryo-EM experiments. In addition, the effects of n-decanol and cetyl alcohol embedded into DHAB bilayers on the vesicular rate constants for the decarboxylation of 6-nitrobenzisoxazole-3-carboxylate (6-NBIC) suggest that both alcohols affect the structure of the Stern region in an equal manner, leading to a decrease in the catalysis. Therefore, it is concluded that addition of the alcohols leads to changes in properties of the interior of the bilayer, rather than the polar-apolar interface.


Introduction
Complex mixtures of steroids, lipids and proteins are the main components in biological membranes [1][2][3]. Both leaflets in biomembranes have different compositions, which makes these biomembranes very complex. In addition to these complexities, the distribution of the lipids over the leaflets is a nonrandom one [4][5][6][7][8][9][10]. The composition of these membranes plays a key role in several biological processes, which are influenced by the interactions between (different) components of the membranes [11].
Synthetic bilayer vesicles are interesting model systems because of their structural relationship with biological membranes. These vesicles have therefore been used as cell membrane mimics. For instance, Eliasz et al. [12] investigated the interactions between various alcohols, carboxylic acids and single-tailed quaternary ammonium salts with 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC). They showed, using differential scanning calorimetry (DSC), that the main phase transition temperature (T m ) is strongly influenced by the membrane composition. Linear alcohols and linear carboxylic acids with short tails in comparison with the phospholipid tails lower the T m drastically.
In comparable studies the effects of various n-alkanols on different phospholipids have been examined [13][14][15][16][17]. It has been found that short-tailed alcohols lower the T m (by ca. 20°C) and are inducing a broadening of the T m transitions. In addition to these phenomena, it has also been found that the binding of these alcohols can result in a disappearance of pre-transitions. This phenomenon has also been observed for the long-tailed alcohol dodecanol, but in contrast to the short-tailed alcohols, addition of dodecanol can lead to an increase in T m by ca. 15°C. Depending on the phospholipids used, n-decanol lowers or increases T m . Generally, the nature of the alkyl chain of the alcohols also plays a role in the incorporation of alcohols. For instance, vesicles are more strongly affected by branched alcohols in comparison with linear alcohols [18]. The effects described so far are mole-fraction dependent.
In the present study our interest is focused on the incorporation of increasing amounts of n-decanol and cetyl alcohol into cationic DHAB bilayers. Due to their large hydrophobic part and poor solubility in water, the alcohols bind strongly to membranes [12,19]. These alcohols were chosen on the basis of the match, in the case of cetyl alcohol, and mismatch, for n-decanol, with the length of the C 16 -tails of DHAB. A major goal of this study is to determine the effects of these matching-and mismatching tails.
Most studies on mixed (DHAB) aggregates mainly focus on the determination of the main phase transition temperature (T m ) [20,21], as well as on the phase diagrams [14,22] and other related characteristics of the aggregates. Here we report the kinetic effects of two types of DHAB/C n OH (n= 10, 16) mixtures on the unimolecular decarboxylation of 6-nitrobenzisoxazole-3-carboxylate (6-NBIC) (Scheme 1). This kinetic probe has been widely used in biophysical studies and is sensitive to the local micropolarity [23][24][25][26], which makes it a proper model reaction for investigation of composition changes in membranes. In addition, the reaction is also highly sensitive to hydrogen-bonding interactions which stabilize 6-NBIC by forming hydrogen-bonds with the carboxylate group [23][24][25][26]. By contrast, an increase in medium polarity will increase the first-order rate constant. Scheme 1. Decarboxylation of 6-nitrobenzisoxazole-3-carboxylate anion (6-NBIC).
The pseudophase model [27,28] is used for determination of the amphiphilic rate constants: where [Amph] is the concentration of the amphiphile, K NBIC is the binding constant of 6-NBIC to the vesicles and k' w and k' amph represent the first-order rate constants in, respectively, the aqueous and amphiphilic pseudophases.
When 6-NBIC is fully bound to the vesicles, as in our case, (i.e., K NBIC [Amph]»1 and k' amph K NBIC [Amph]»k' w ), equation [1] is reduced to the following equation: [2] Solubilisation of the bilayer vesicles by detergents (e.g. TritonX100, sodium dodecyl sulfate (SDS)) can give valuable information about the packing of the membrane and its resistance against detergents. A typical solubilisation profile is shown in Figure 1. This profile can be divided into three stages [29,30]. In the first stage the detergent is adsorbed by the vesicles, resulting in swelling. At a certain detergent concentration (R sat ) the bilayers are saturated with detergent and from this point the vesicles are starting to solubilise upon addition of more detergent. This is the start of the intermediate region which consists of mixed micelles and vesicles. When all the vesicles are solubilised (R sol ) by the detergent only mixed micelles are present and the turbidity levels off. Adding more detergent induces no more changes in the turbidity.
Although turbidity is a non-discriminating technique, its sensitivity is good enough to determine the saturation and solubilisation point accurately as a consequence of the large increase of the scattered intensity of light with particle size [31]. Therefore, almost any "large" particle (e.g. small vesicles, bilayer fragments) will be detected, hence the reported R sol values indicate the point of complete solubilisation. Additional evidence of partial vesicle solubilisation between R sat and R sol comes from cryo-electron microscopy experiments involving detergent-induced vesicle solubilisation of vesicles formed from phospholipids [32,33]. Above R sat , besides spherical micelles, vesicles that have holes are observed, whereas above R sol no longer vesicles or bilayer fragments are found. It should be noted that in the present paper we are interested in trends, rather than absolute values.
Besides the kinetic and solubilisation measurements, we have also used other techniques for characterisation of the DHAB/C 10 OH mixtures: DSC measurements to determine the main phase transition temperature of different DHAB/C 10 OH mixtures and cryo-electron microscopy for a morphological identification of the different DHAB/C 10 OH mixtures.

Results and Discussion
Kinetics Figure 2 shows k obs values for the decarboxylation of 1 versus the amount of n-decanol and cetyl alcohol present in the DHAB bilayers. The rate profiles for both alcohols show that the rate is slowed down with increasing amounts of n-decanol or cetyl alcohol in the DHAB/C n OH (n=10,16) mixtures. This pattern probably originates from a change in the hydrogen-bonding interactions between water molecules in the vicinity of the vesicular surface and the hydroxyl group of the alcohols with the carboxylate unit of the reactant and the activated complex. As already noted, hydrogen bonds lower the Gibbs energy of the initial state of 6-NBIC relatively to the activated complex. The net result of this behaviour is that the Gibbs energy of activation will rise upon an increase in n-decanol or cetyl alcohol in the DHAB/C n OH (n=10,16) mixtures. The data in Figure 2 show that the differences between DHAB/C 10 OH and DHAB/C 16 OH mixtures are negligible. This observation makes it unjustified to attribute the differences to a difference in the type of alcohols present. For clarity, the k obs found for 0 mol% C n OH and 50 mol% C n OH (n=10,16), differ by a factor of about 3.5, which is quite small.
The rate constants observed are comparable with those found for hydrophobically and electrostatically modified polyacrylamides as cosolutes. These compounds induce rate constants of an order of magnitude of 10 -3 s -1 [34][35][36]. In addition, the k obs found for cationic micelles formed from nhexadecyltrimethylammonium bromide (CTAB) at 30°C is in the order of magnitude of 10 -4 s -1 [25,37], which is comparable with the k obs found for DHAB at 35°C. Comparable k obs values at a temperature of 30° C are also found for vesicles containing different amounts of di-noctadecyldimethylammonium chloride (DODAC) [28]. The comparisons observed, suggest that DHAB bilayers do not differ significantly in micropolarity and hydrogen-bonding propensity from CTAB and DODAC aggregates.  In equations [4] and [5] [amph] is the molar concentration of amphiphile and additive. As can be seen in equation [4], the slopes of the plots in Figure 3 and 4 represent the R b values, which are the amounts of TritonX-100 bound to the amphiphilic pseudophases. The intercept with the y-axis represents the aqueous concentration of TritonX-100. From the ratio of the slope to the intercept the binding constant of detergent molecules to the vesicles (K det ) can be calculated via equation [6]: [TX] amph K det = ------ [6] [TX] w [amph] However, due to scattering in the experimental data, extrapolation to zero amphiphile concentration and the relatively small value of the intercept compared to the slope, the error in K det will be large. Therefore, we refrained from calculating the binding constant.
The data in Figure 3 indicate that the composition of the bilayer (R b ) at the point of saturation (i.e. the slope of the plots in Figure 3) is independent of the mole fraction of incorporated n-decanol or cetyl alcohol. Except that the presence of over 15 mol% of cetyl alcohol lowers R b at the saturation point by a factor of around 2. This means that less TritonX-100 is required to saturate the vesicles with detergent, if cetyl alcohol is present.
In Figure 4 it can be seen that for the solubilisation point R b is constant upon increasing mole fraction of n-decanol, but increases linearly with increasing mole fraction of cetyl alcohol. At 35 mol% of cetyl alcohol R b is more than twice as large. All data is summarised in Figure 5. It is surprising to notice that increasing mole fraction of n-decanol does not influence detergent saturation and solubilisation of DHAB vesicles. On the contrary, upon addition of cetyl alcohol less TritonX-100 is needed to saturate the membrane, but that more TritonX-100 is required to completely solubilise all material into mixed micelles.
The values observed in Figure 5 are comparable with those of TritonX-100 bound to PC vesicles. 37 These vesicles are quite sensitive for the nature of the surfactant used. For instance, the R b values for the ionic surfactants sodium cholate [40], sodium alkylsulfates, as well as alkyltrimethylammonium bromides [39] in PC membranes vary between 0.3 and 4. These values depend strongly on the tail length of the surfactant.
The R b values at the solubilisation point (between 1 and 3.5) are comparable with those found for solubilisation of 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) vesicles by SDS [41], as well as solubilisation of phosphatidylcholine (PC) vesicles by different sugar-based surfactants [38]. The observed values for R b upon addition of n-decanol are about 1.5, which is quite small. These values are comparable with the solubilisation of PC vesicles by dodecyl-α-D-maltoside, as well as by some other sugar surfactants [38].
If the data would be analysed in terms of solely the concentration of amphiphile and not the concentration of amphiphile and alcohol, the slopes in Figures 3 and 4 would increase progressively with increasing alcohol content, leading to larger values of R b . However, such an approach underestimates the contribution of the alcohols to the total hydrophobic volume to which TX-100 binds. In the absence of knowledge about the molar volume of the alcohols and DHAB (and hence the binding cannot be expressed as a function of the total hydrophobic volume), it is a more accurate approach to use the concentration of amphiphile and alcohol.     Differential Scanning Calorimetry Figure 6 shows the dependence of the main phase transition temperature on the mole fraction of C 10 OH in DHAB/C 10 OH vesicles. The T m found for pure DHAB is 28°C, which agrees with the literature [19,20]. Up to 20 mol% of C 10 OH T m decreases until about 23°C. Generally a lowering in T m is explained by a decreased packing of the chains [21,42]. Upon further addition of C 10 OH the T m increases again, reaching 30°C at 50 mol% of C 10 OH. In the literature it has been reported that in mixtures of n-alcohols and phospholipids n-alcohols behave like amphiphiles leading to changes in the phase transition temperature that are linearly correlated with the melting point of the alcohols. A decrease in the phase transition temperature has been found upon the addition of alcohols with a short ( C 8 ) chain and an increase for alcohols with a long ( C 12 ) chain [12][13][14][15][16][17]. For mixtures of various n-alcohols and a synthetic quaternary ammonium amphiphile a decrease in the T m at low mole fraction has been found, but an increase at higher mole fractions. The extent of the increase depends on the length of the alcohol [43].
The origin of the observed effects remains unclear. In general, the DSC scans show more than one peak, indicating several types of transitions, domain formation, or a combination of both. In the absence of more detailed information we refrain from an interpretation of these data.

Cryo-electron microscopy
The micrographs shown in Figure 7, suggest that the solutions contain closed vesicles, flat bilayers, and structures in between these extremes. In comparison with DHAB bilayers, the average size of the DHAB-30 mol% C 10 OH aggregates is smaller. For synthetic amphiphiles open vesicle structures or bilayer fragments have been reported in the literature when the amphiphiles are below the main phase transition temperature, as is the case for the structures shown in Figure 7 [44,45]. The size of the bar represents 100 nm.

Matching and mismatching tails
The influence of match and mismatch of cetyl alcohol and n-decanol, respectively, with the C 16tails of DHAB is most apparent in the solubilisation experiments ( Figure 5). Whereas, mismatch does not lead to changes in membrane composition (R b is constant) at the points of saturation and solubilisation, match leads to a strong increase in R b for complete solubilisation of the vesicles upon increasing the amount of incorporated cetyl alcohol. In addition R b decreases for the saturation point. On the contrary, the decrease in vesicular rate constant upon increasing alcohol content of the vesicles does not depend on the match or mismatch. It should be noted that the decarboxylation reaction takes place in the Stern region, whereas solubilisation of the bilayer depends on the membrane as a whole. Apparently, addition of both alcohols affects the Stern region is a similar manner, but the packing deeper in the bilayer is affected in a dissimilar manner. Considering that the hydroxyl group resides in the head group region of the amphiphile, it is reasonable that differences in packing are most pronounced at position remote from the polar-apolar interface. The decrease in catalysis is most likely a result of an increase in water concentration due to decrease in the (head group) packing.
The cryo-EM pictures show that n-decanol embedded into DHAB bilayers induces small morphological changes. Considering the DSC results, these changes are probably accompanied by changes in the packing of the DHAB/C 10 OH bilayers [42,45].

Conclusions
In this study we have outlined the different effects observed between matching and mismatching tails in vesicles formed from DHAB and C n OH (n=10 or 16). The results show that, besides small morphological changes and changes in the main phase transition temperature, upon increasing amounts n-decanol the composition of the membrane (R b ) at the points of saturation and solubilisation by TritonX-100 are not affected up to 50 mol% of n-decanol. On the contrary, upon increasing amounts of cetyl alcohol the point of saturation is reached at a lower detergent to amphiphile ratio, but the point of solubilisation is reached at a higher detergent to amphiphile ratio.
The monotonic and alcohol-structure independent decrease in the vesicular rate constant of the unimolecular decarboxylation reaction of 1 upon increasing alcohol content suggests that changes in the bilayer mainly occur in the interior of the bilayer, rather than at the polar-apolar interface. In addition, the decrease in rate constant suggests a slight increase in the water content at the interface of the vesicles.

Acknowledgements
Dr. Marc Stuart is kindly acknowledged for his skilful cryo electron microscopy measurements on DHAB/C 10 OH mixtures.

Solubilisation measurements
TritonX-100 was added from a freshly prepared aqueous stock solution to the solutions containing DHAB aggregates. The solubilisation processes were followed by monitoring the turbidity (absorption) at 410 nm. Measurements were performed above the T m at 35.0 ±0.1 °C and were carried out using a Perkin Elmer λ5 spectrophotometer. The R sol and R sat were determined graphically.

Kinetic measurements
The decarboxylation of 1 was monitored at 35.0±0.1°C by measuring the increase in the absorption at 410 nm for at least 5 half-lives after addition of 4 µl of a 6 mM stock solution of 1 in acetonitrile into a 0.9 ml solution containing an alkaline surfactant solution. The stock solution of 1 was stored at -20°C. The concentration of sodium hydroxide was 100 fold higher than the concentration of 6-NBIC. Measurements were performed using a Perkin Elmer λ5 or a Perkin Elmer λ2 spectrophotometer. The average error observed in the rate constants was 10%.

DSC measurements
Differential Scanning Calorimetric (DSC) measurements were carried out with a VP-DSC microcalorimeter. A typical experiment was performed with a scan rate of 60°C/h and an amphiphile concentration of 2 mM. Each scan was repeated at least 5 times.

Cryo-Electron Microscopy
The following procedure was used to perform cryo-electron microscopy: aliquots of amphiphile solutions were deposited on holey carbon grids; the excess solution was blotted off using filter paper. The samples were vitrified by rapid plunging into liquid ethane. The grids were transferred to a Gatan model 626 cryo holder and examined at ca. -170 °C in a CM120 Philips microscope operating at 120 kV. Micrographs were recorded under low dose conditions.