Tuning Curvature and Stability of Monoolein Bilayers by Designer Lipid-Like Peptide Surfactants

This study reports the effect of loading four different charged designer lipid-like short anionic and cationic peptide surfactants on the fully hydrated monoolein (MO)-based Pn3m phase (Q224). The studied peptide surfactants comprise seven amino acid residues, namely A6D, DA6, A6K, and KA6. D (aspartic acid) bears two negative charges, K (lysine) bears one positive charge, and A (alanine) constitutes the hydrophobic tail. To elucidate the impact of these peptide surfactants, the ternary MO/peptide/water system has been investigated using small-angle X-ray scattering (SAXS), within a certain range of peptide concentrations (R≤0.2) and temperatures (25 to 70°C). We demonstrate that the bilayer curvature and the stability are modulated by: i) the peptide/lipid molar ratio, ii) the peptide molecular structure (the degree of hydrophobicity, the type of the hydrophilic amino acid, and the headgroup location), and iii) the temperature. The anionic peptide surfactants, A6D and DA6, exhibit the strongest surface activity. At low peptide concentrations (R = 0.01), the Pn3m structure is still preserved, but its lattice increases due to the strong electrostatic repulsion between the negatively charged peptide molecules, which are incorporated into the interface. This means that the anionic peptides have the effect of enlarging the water channels and thus they serve to enhance the accommodation of positively charged water-soluble active molecules in the Pn3m phase. At higher peptide concentration (R = 0.10), the lipid bilayers are destabilized and the structural transition from the Pn3m to the inverted hexagonal phase (H2) is induced. For the cationic peptides, our study illustrates how even minor modifications, such as changing the location of the headgroup (A6K vs. KA6), affects significantly the peptide's effectiveness. Only KA6 displays a propensity to promote the formation of H2, which suggests that KA6 molecules have a higher degree of incorporation in the interface than those of A6K.

The MO/water system displays two different types of bicontinuous cubic phases depending on the water content [28][29][30][31][32][33][34]: cubic assemblies with Ia3d (the gyroid type, C G ) and Pn3m (diamond type, C D ) symmetries, respectively. These cubic phases exhibit the lowest curvature inhomogeneity [1] and are complemented by the cubic phase with Im3m symmetry (the primitive type, C P ), which is found in various other lipid systems [49,50] and their aqueous dispersions [40][41][42]50]. The water channel connectivity and the surface topology of these bicontinuous cubic phases are illustrated in Figure 1. MO-based bicontinuous cubic phases and their aqueous dispersions have been of great interest for various novel applications including for the crystallization of membrane proteins [51][52][53][54], and for the solubilization of different hydrophilic and lipophilic guest molecules such as vitamins, essential oils, and drugs [55][56][57][58]. In particular, these nanostructured systems are promising for the formation of effective drug delivery systems and have great potential applications in food, pharmaceutical, and cosmetic industry. The stability of such cubic phases depends on various parameters such as water content, temperature as well as on the type and the amount of the solubilized guest molecules [55][56][57][58][59]. For instance, the addition of lipophilic molecules [45,60] leads to a structural transition from the V 2 to H 2 , while amphiphilic molecules with a propensity of having positive spontaneous curvatures such as sodium oleate [60] lead to a transition from V 2 to L a (in literature, the sign of curvature is defined to be positive for surfaces of normal oil in water aggregates and negative for surfaces of inverse mesophases).
Moreover, the effect of various peptides on the stability of MObased cubic phases has been subject of several studies [61][62][63][64]. It was found that the electrostatic interactions in the membranes incorporating negatively (or positively) charged lipids are controlling the structural lamellar-nonlamellar transitions, i.e. the nature of these guest peptides, the peptide-MO ratio, and salt concentration play an important role in modulating the selfassembled nanostructure.
In this work, we report on the effect of four different short lipidlike designer peptide surfactants on the fully hydrated Pn3m-phase of monoolein (MO). Our aim is to understand the key factors that tune the bilayer curvature and to gain further insight into the mechanism of the structural transformations induced by the addition of these molecules. We also checked the possibility to functionalize the bicontinuous cubic phase by the addition of small amount of peptide surfactant. It should be noted that the concept of functionalization [65] means to control the solubilization capacity of the liquid crystalline phases by the inclusion of specific anchors such as charged or long-chain amphiphilic molecules.
The used peptides [66][67][68][69][70][71][72] are short cationic (A 6 K & KA 6 ) and anionic (A 6 D & DA 6 ) with an approximate length of 25 Å , which have been designed to mimic biological phospholipids ( Figure 2). These peptides contain seven amino acid residues and are amphiphilic with a hydrophilic headgroup and a hydrophobic tail. For example, in the peptide surfactant A 6 D, the hydrophilic head has an aspartic acid (D) at the C-terminus, whereas the hydrophobic tail consists of six consecutive hydrophobic amine acids (alanine, A) with an acetylated N-terminus, eliminating the positive charge. We note that this 7-residue peptide has two negative charges at the C-terminus. Thus, these peptides exhibit self-assembling behavior akin to phospholipids with distinct critical aggregate concentration (CAC) values, also commonly referred to as critical micelle concentration (CMC) [69]. The CAC value [69,71] corresponds to approximately 1.6 mM for A 6 D in pure water, and A 6 K has a CAC of <1.5 mM. Moreover, these designer peptide surfactants not only self assemble to form nanotubules and nanovesicles in water [66][67][68] with diameters of ,300-500 Å ( Figure 1D & E), but these ''peptergents'', referring to peptides with detergent properties [72], represent a new class of biomaterials with excellent potential to solubilize, stabilize, and crystallize membrane proteins and enzymes [70][71][72].
The present study is organized in the following 4 sections: in the first two the effects of anionic and cationic peptides on the stability of Pn3m phase of the MO/water system are discussed. In section 3, the impact of headgroup location for the anionic A 6 D vs. DA 6 and the cationic A 6 K vs. KA 6 surfactants is described. In section 4, the phase behavior is examined on the basis of peptide concentration versus temperature diagrams. Finally, we propose hypotheses concerning the biological relevance and future applications.  . From top to bottom: the molecular models of the peptide surfactants A 6 D, A 6 K, DA 6 , and KA 6 . Aspartic acid (D) bears two negative charges, and lysine (K) bears one positive charge. Alanine (A) constitutes the hydrophobic tails. The peptides A 6 D and A 6 K were synthesized with the head group at the C-terminus. In the opposite arrangement, the peptides DA 6 and KA 6 were synthesized with the head group at the N-terminus. All used peptides are similar to biological phospholipids: each peptide is 25 Å in length. The color scheme as follows: carbon, cyan; nitrogen, blue; oxygen, red; and hydrogen, white. doi:10.1371/journal.pone.0000479.g002

Effect of anionic peptide surfactant
The effect of anionic peptide surfactants on the stability of the MO-based Pn3m phase (Q 224 ) has been examined by SAXS. In Figure 3A, the scattering curves for A 6 D-loaded MO/water phases with 4 different R values (in the range of 0 to 0.10) are shown at 25uC: in the absence of peptide (at R = 0), the diffraction pattern is indexed in accordance to a cubic Pn3m lattice. At low A 6 D concentration (at R = 0.01), the cubic Pn3m structure is preserved, but the peaks are shifted to lower q values, thus the corresponding lattice parameter, a, increases (Table 1). At R = 0.05, the diffraction pattern displays a coexistence of the former Pn3m cubic phase with a newly-formed H 2 phase. A further increase in the peptide content (R = 0.1, which is the highest R value used for this peptide) leads to a complete structural transformation. The three observed peaks are identified by the (100), (110), and (200) reflections of the H 2 phase. In short, we found that increasing the peptide concentration leads to structural transitions in the order V 2 (Pn3m)RPn3m & H 2 RH 2 . In addition, the diffraction peaks shift as a function of peptide concentration to higher q values, which mean a decrease in the lattice parameters.
It is well known that the influence of guest molecules on phase transitions in liquid crystalline phases is closely related to the degree of their penetration in the membrane interface as well as to their ability to alter the spontaneous curvature of the monolayer leaflets [45,[56][57][58][59][60]. To shed some light on the impact of the designer peptides, we will particularly discuss the 'effective' molecular geometry of the membrane constituents and analyze its influence on forming diverse supramolecular structures. In a basic approach, the molecular shape can be described by the critical packing parameter (CPP) or the molecular wedge shape factor, which is defined as: where v s is the hydrophobic chain volume, a 0 is the headgroup area, and l is the hydrophobic chain length [73]. This packing parameter is controlled by various factors such as surfactant's molecular shape, temperature, hydration, the presence of hydrophilic or hydrophobic guest molecules, and electrostatic effects [16,[29][30][31][32][33]35,[45][46][47][48][49][50]61,62]. In our study, at low A 6 D content (R = 0.01), the increase in the unit cell parameter, a, of the Pn3m phase (Table 1) is most probably attributed to the electrostatic repulsions between the negative charges of the peptides, which are incorporated into the electrically neutral MO-based membrane interface. The electrostatic forces will increase the distance of the negative charges and hence lead to an increase in the a 0 value at the water-lipid interface (see Figure 4). However, the electrostatic repulsions are not strong enough to increase significantly the CPP value or in other words to flatten the bilayer completely. The cubic phase Pn3m remains stable. In contrast, at higher peptide concentrations the lattice Pn3m lattice shrinks and at a certain peptide's concentration the structural transformation from Pn3m to H 2 is observed. This can be attributed to an altered water solubilization in the selfassembled structure.
When the content of A 6 D is high, a considerable fraction of A 6 D molecules are incorporated into the MO matrix. As a consequence the amount of solubilized water at the lipid water interface decreases, i.e. CPP value increases due a reduction in a 0  However, the formation of the H 2 phase is not only governed by a reduction in bending energy. The release of curvature frustration has a price, which is the build-up of packing frustration within the newly-formed hexagonal phase [74]. As seen in Figure 4, the closest packing of rods demands that each lipid cylinder has to fill out the space of a hexagon. This signifies that the lipid lengths must vary, i.e. lipids oriented towards the hexagon corners have a maximum length and those oriented perpendicular to the hexagon faces are the shortest. This variation in lipid length l does not depend on the water core radius, but can directly be determined from the lattice parameter of the H 2 phase [75]: It is important to note that the used designer peptides in this study ( Figure 2) are about 7 Å longer then that of MO molecule, which at full hydration conditions has a length of about 18 Å [34,76]. Therefore, it suggests that the designer peptides preferentially are localized in the corners of the hexagon and hence reduce the packing frustration within the hexagonal lattice. It is known that the interstitial regions [77] in H 2 phase account for a volume fraction of about 9%. Our results show that the threshold concentration of the anionic peptides to induce the H 2 phase is in the same order of magnitude (R = 0.05-0.1). Moreover, applying equation (2) to the A 6 D induced H 2 phase with a ,64 Å (R = 0.05, Table 1) results in a lipid length variation of 5 Å , which resembles just the difference in length of MO and peptide. A strong reduction of interstitial energy contributions would also explain in part the strongly reduced transition temperatures from Pn3m to H 2 . This transition occurs in the binary MO/water system at ,90uC [32]. An additional interesting point is related to the peptide chains, which should have also an influence onto the spontaneous curvature of the MO monolayers. It is known that the hydrophobic tail packing for the surfactant-like peptides is quite different from that reported for conventional surfactants, most likely because of the intermolecular interactions due to the hydrogen bonds between peptide backbones [66][67][68]. Therefore, in pure peptide domains hydrophobic tails are rather tightly packed, leading to a local increase in the membrane rigidity. At this stage, however, we are not able to predict the overall rigidity of the composite MO/peptide leaflets and how rather rigid peptide chains influence the overall lateral chain pressure.
Interestingly, in aqueous medium our short designer peptide surfactants [68,70] such as A 6 D favors the formation of bilayer structures with an estimated thickness of 50 Å (Figures 1D & 1E). Therefore, it is worthy noting that in our study the presence of bilayer-forming peptide in the MO/A 6 D mixture promotes the formation of H 2 instead of the planar L a (structure with zero curvature). Different findings were reported by Yamazaki and his coworkers [62] on the effect of the positively charged peptide WLFLLKKK, which was loaded with similar concentrations, onto the MO-based Pn3m phase. This peptide with a relatively large positively charged headgroup (KKK) induces the structural transformation from V 2 (Pn3m) via V 2 (Im3m symmetry) to L a . The reduction in bilayer curvature was explained in terms of increased electrostatic repulsions due to the partition of the charged peptide molecules in the MO interfacial film. Likewise, Chupin et al. [63] observed in a MO-based system that cubic to L a phase transition can also be facilitated in the presence of a transmembrane peptide. However, transmembrane peptides are also known to induce non-lamellar structures when inserted into various phospholipid/water systems [17,[78][79][80]. For example, WALP and KALP (a-helical transmembrane peptides), which with 16 to 31 amino acids are significantly larger than our peptide surfactants, promote the L a RV 2 and/or H 2 structural transitions [17,78,79]. In particular, temperature and the peptide content dominate the lamellar/non-lamellar preference [17,78]. Their propensity to form non-lamellar phases was rationalized by hydrophobic mismatch effects. When the hydrophobic thickness of the bilayer significantly exceeds the hydrophobic length of the transmembrane peptides [17,[78][79][80], a local disordering of the lipid acyl chains can occur, which serves as precursor for the formation of non-lamellar structures. However, in our study we are dealing with surfactant-like peptides and thus our results can not be explained within the framework of hydrophobic mismatch concept, but the transition from bicontinuous cubic to H 2 phase is rather driven by the peptide structure and its amphiphilicity.
It comes clear that the impact of the designer peptide surfactants, which form well-ordered nanostructures in water ( Figures 1D & 1E), is different from most short peptides reported in literature. In general, the previously studied peptides are not surfactant-like and thus do not form well-defined nanostructures in water. On the contrary, A 6 D must be classified as highly surfaceactive material, which can be exploited as a ''tuneable inducer'' of spontaneous membrane curvature. Figure 3B shows the effect of the cationic peptide surfactant A 6 K on the stability of the fully hydrated MO-based Pn3m phase. SAXS patterns of the ternary MO/A 6 K/water system are displayed for various R values (in the range of 0 to 0.2) at 25uC. In the presence of A 6 K, all diffraction patterns are consistent with indexing a cubic lattice of the type Pn3m. This means that the structure is preserved after loading the cationic peptide surfactant. As shown in Figures 3A & 3B, similar behavior to the anionic A 6 D peptide is observed at low peptide content (at R = 0.01). In both systems, an increase of the unit cell parameter is observed, which is attributed to strong electrostatic repulsion between the positive charges of the peptide molecules incorporated into the interface (R#0.1). However, a further increase of this cationic peptide concentration (R = 0.2) causes a counter effect, i.e. there is a slight decrease in the corresponding structure parameter of the Pn3m phase. Intriguingly, there is no indication for the formation of the H 2 phase. This means that A 6 K is not as efficient as the negatively charged peptide A 6 D. It is plausible that the difference in their effectiveness is attributed to their degree of incorporation in the polar-apolar MO interface region. Both peptides are designed with a charged headgroup in the C-terminus to increase their solubility in water, using aspartic acid or lysine, respectively. However, our results suggest that A 6 K is only partially incorporated into the interface (a considerable part of it prefers to reside in the excess aqueous phase) and thus, most likely there is no significant reduction in the amount of solubilized water at the polar interface. Whereas, it seems that A 6 D prefers mainly to penetrate the membrane interface and thus is more efficient in inducing increasingly negative spontaneous monolayer curvature, which leads at a certain peptide concentration to the Pn3mRH 2 transition.

The impact of headgroup location on tuning the membrane curvature
To investigate the role of the headgroup location and the importance of the headgroup charge, we compared the efficiency of the peptides A 6 D & A 6 K (the headgroup at the C-terminus, Figures 3A & 3B) with the peptide surfactants DA 6 & KA 6 (the headgroup at the N-terminus, Figure 5). It is well-known that a simple modification of the amino sequence of these designer peptide surfactants has a significant impact on their solubility in water [68][69][70]. For enhancing solubility, it is favorable for the peptides to have the headgroup at the C-terminus [68][69][70]. However, Zhang and his coworkers [68,70] reported that the ordered nanostructures of these peptides in water are independent on the headgroup's location. Keeping this in mind, we checked the possible effect of slightly modified peptide structures on the MObased Pn3m phase. Figure 5A illustrates the impact of loading DA 6 on the Pn3m structure. At low concentrations (R = 0.01 & 0.05), this peptide has a similar influence on the Pn3m phase as A 6 D does. Furthermore, it is also efficient at higher peptide concentrations in inducing the formation of H 2 phase. Thus, the headgroup location plays only a minor role for the short anionic peptides. This means that our findings are similar to those reported on the aggregation behavior of the designer peptides in water [68,70]. In contrast, the headgroup location does matter for the impact of cationic peptides. There are pronounced differences when A 6 K is replaced at the same peptide's concentration by KA 6 . As shown in Figure 5B, upon addition of KA 6 the mean lattice parameter for the Pn3m phase increases as the R values changes in the range of 0-0.05. With a further increase in the peptide content (R = 0.1), the system exhibits a two-phase region of coexisting Pn3m and H 2 structures, and at R = 0.20 a transition from Pn3m to H 2 phase takes place. For this system, an additional broad peak appears at low q values ( Figure 5B, * ). This diffuse scattering contribution disappears when the sample is heated from 25 to 40uC (data not shown). Therefore, we believe that it arises from traces of a disordered bicontinuous diamond phase. The behavior of A 6 K ( Figure 3B) contrasts sharply with that of KA 6 . A 6 K does not destabilize the MO-lipid bilayer. This can be attributed to the higher solubility [68] of A 6 K in the aqueous medium as compared to KA 6 , leading to a lower degree of incorporation in the MO bilayers and thus being less effective.
In brief, the positively charged KA 6 has similar impact on the MO/water system as the negative peptides A 6 D & DA 6 , but it is less effective. Thus, the hydrophilicity of the headgroup seems to play an important role on tuning the membrane curvatures: replacing D with the less hydrophilic K decreases significantly the peptide's effectiveness. Figure 4 summarizes the main findings. Low peptide concentrations increase slightly the interface curvature due to the electrostatic repulsion at the headgroups, while at higher concentrations especially the anionic designer peptides induce a nanostructure transition from Pn3m to H 2 due to a decrease in the amount of solubilized water in the interface region, and secondly to a release of the curvature frustration (compare with the discussion on anionic peptide surfactants).
In Figures 3 & 5, it is also worthy noting that all peptide-loaded samples display an increased diffuse scattering at low q values. An example for such effect is shown in Figure 5B. If we take into consideration that these surfactant peptides undergo self-assembly [65][66][67][68][69][70] in water and form nanovesicles and nanotubes ( Figures 1D  & 1E), this scattering contribution might be attributed to the formation of small peptide surfactant aggregates in co-existence with the ternary MO/peptide/water phases.
Effect of temperature on MO/peptide/water systems As we have discussed in the previous sections, the formation of H 2 phase is promoted most efficiently by the anionic peptides. In contrast, the addition of cationic peptides to the MO/water system displays a reduced efficiency in varying the monolayer curvature. In Figure 6A, we compare the temperature-peptide concentration diagram of MO/A 6 D/water with that of MO/KA 6 /water ( Figure 6B). For better distinction, the cubic phase regions in these diagrams have been highlighted in light grey. It is worth noting that the phase boundaries are only very roughly estimated and do not allow any deduction of phase transition temperatures or phase transition concentrations, respectively. It is clear though that especially in the low temperature regime A 6 D is more efficient than KA 6 in destabilizing the membrane bilayer. However, rising temperature increases the CPP value and therefore enhances the spontaneous negative curvature by two ways: it reduces the value of a 0 due to the headgroup's dehydration and it enhances simultaneously the value of v s . Thus, in this scenario the bilayer thickness as well as the water core radii decrease monotonously with temperature. At higher temperatures, our results for both ternary systems ( Figure 6) indicate that the differences in peptide efficiency are not that pronounced any more. Here, the hydrophobic chain pressure [81] becomes the dominating driving force and hence gains importance in tuning membrane curvature. In other words, a further change in the interfacial area due to headgroup dehydration seems to play a less significant role at higher temperatures.
In Figure 7A, we take a closer look to the temperature behavior of the Pn3m phase at R = 0.01. Our results reveal that heating the A 6 D-and KA 6 -loaded systems display a very similar trend of behavior when compared to the unloaded binary MO/ water system [32]. As shown also in Table 2, the decline of the lattice parameter with temperature is nearly identical in all cases. The interpretation is straightforward: the CPP value increases with temperature and hence the spontaneous curvature at the interface becomes increasingly negative. This holds true also for the H 2 phase, but the situation is more complex: it is important to recall that the H 2 phase in the binary MO/water system [27,32] exists only within a small regime at high temperatures (,90-100uC). Further, increasing temperature from 90 to 100uC causes a slight decrease in lattice parameter of this mesophase from 54 to 52 Å ( Figure 7B, +). At even higher temperatures (.100uC), a fluid isotropic fluid (inverted micellar solution, L 2 ) is formed.
As can be judged from Figure 7B, there is a significant impact of the studied peptides on the behavior of the H 2 phases when compared to that of the binary system: firstly, there is a strong reduction in the Pn3mRH 2 transition temperatures. At R = 0.05, loading A 6 D and KA 6 reduces the transition temperature from approximately 90 to 25, and 40uC, respectively ( Figure 6). Secondly, the H 2 region in the phase diagram is drastically increased. Thirdly, the addition of the designer peptides enlarges the diameter of the formed hydrophilic cylinders (Tables 1 & 2 and Figure 7B). For instance at R = 0.1, for the KA 6 -loaded system the smallest lattice parameter measured has a value of 58 Å (Figure 7B,  N). However, an interpolation of the a(T) plot would lead to similar lattice parameter values as recorded for the binary MO/water system at 90-100uC. The behavior is different, when the most efficient peptide A 6 D is loaded ( Figure 7B, O). While a(T) decreases with approximately 20.2 Å /uC, which is almost the same as for the binary MO/water system, the a value drops down to 52 Å already at 70uC (compare O with + in Figure 7B). This underlines once more the exceptional strong ability of the designer peptide A 6 D to reduce the effective interface area per molecule, since an increased chain pressure alone can not cause such high curvatures between 60 and 70uC.

Relevance of the short designer peptide surfactants
The self-assembly of biological molecules for the design novel materials with well-defined nanostructures is increasingly exploited in biotechnology [70]. For instance, the designer peptide surfactants of this work undergo self-assembly and form welldefined systems useful for various potential applications such as the encapsulation, the solubilization, or the crystallization of active biomolecules [66][67][68][69][70][71][72]. Here, we have shown that these peptide surfactants can also be used to stabilize different non-lamellar mesophases, which have also biological relevance. For instance, Bechinger and Lohner [82] pointed out that peptide surfactants, which are present in plant and human cells, modulate the antimicrobial activities of biomembranes.
The short designer peptide surfactants are tuneable nanobiomaterials. It is easy to modify the peptide hydrophobic tail as well as its headgroup. For instance, the degree of hydrophobicity can be fine-tuned by replacing alanine (A) with more hydrophobic amino acids such as valine (V), or leucine (L). Furthermore, the degree of hydrophilicity can be varied by increasing the number of the hydrophilic amino acids on the peptide's backbone and by replacing the negatively charged aspartic acid or the positively charged lysine by other hydrophilic amino acids such as the negatively charged glutamic acid (E) or the positively charged histidine (H) or arginine (R).
As a final point we would like to show up a practical route for functionalizing bicontinuous cubic phases by the addition of peptide surfactants. We anticipate that the inclusion of small amount of the charged short designer surfactants A 6 D (or A 6 K) into the MO interfacial film will enhance the loading capacity of positively (or negatively) charged water-soluble active molecules in the hydrophilic channels of the Pn3m phase. Enhancement of solubilization capacity of active guest molecules is an important issue in various pharmaceutical, food, and cosmetic applications. In particular, today the formation of suitable carriers for loading charged drugs or peptides is of great interest.

Conclusions
The present work focuses on the influences of short charged designer peptides, which mimic the properties of anionic and cationic surfactant molecules, on the fully hydrated Pn3m nanostructure of MO. The following major conclusions can be drawn from our investigations:

1.
Charged designer peptide surfactants: the structure of the bicontinuous cubic phases can be modulated by adding short negatively or positively charged peptide surfactants. First, they allow to tune the water channel size of the crystalline phases at low peptide's content, and second, the transition of the nanostructure from Pn3m to H 2 can be induced either by augmenting the peptides' concentration at room temperature or by increasing the temperature at a fixed peptide concentration. We found that the effectiveness of these peptide surfactants depends on the headgroup location and its structure (the type of the hydrophilic amino acid). 2. Model systems as tools for studying biological systems: lipids (such as monoglycerides and phospholipids) form in water various mesophases (lamellar, hexagonal or cubic structures). The nanostructures of these mesophases depend on temperature, water content and the molecular structure of the lipids (single or double chained, saturation degree, and length of fatty acyl chain). Our present lipid/ peptide surfactant membrane model provides the basis to learn more about possible effects of active molecules in biomembranes and helps to understand structural transitions that occur within biological cells. 3. Potential applications: functionalization of bicontinuous cubic phases is possible by the addition of charged short peptides. These serve as anchors in the water/lipid interface and allow enhancing the loading capacity of charged active molecules.

Peptide Surfactants: Design and Synthesis
The lipid-like peptide surfactants were synthesized by CPC Scientific (San Jose, CA) and characterized by the Biopolymers Laboratory at MIT (MA, USA). The design is based on creating a peptide of approximately phospholipid size containing a hydrophilic head and a hydrophobic tail [66][67][68][69]. The synthesis of these surfactant peptides are described in previous reports of Zhang and his coworkers [66][67][68]. The four peptides studied are A 6 K, A 6 D, KA 6 , and DA 6 . The first two are synthesized with the headgroup at C-terminus, whereas the last two with the headgroup at the free N-terminus (K = lysine, A = alanine, D = aspartic acid).

Preparation of MO/Peptide Surfactant Samples
Lipid-like peptide surfactants with appropriate concentrations were dissolved in TFE while MO was dissolved in chloroform. These solvents were mixed and then evaporated using a gentle stream of nitrogen, followed by drying under vacuum for at least 12 hours in order to remove completely the residual organic solvent. The dry lipid-peptide film was hydrated by adding the PBS buffer and carrying out at least 5 freeze-thaw cycles between liquid nitrogen and room temperature and then homogenizing several times during the thawing steps by vigorous vortexing. The fully hydrated samples coexist with excess water and were formed with a fixed total lipid concentration of ,18 wt%. In our study, we investigated the effect of varying the molar ratio R (R = moles of surfactant peptide/mole of MO) on the stability of the MO-based Pn3m phase. The prepared samples were incubated at room temperature for two to three weeks before carrying out SAXS measurements.

X-Ray Scattering Measurements
The X-ray measurements were carried out on a small-and wide angle X-ray scattering camera with Kratky collimation [83] (SWAXS, System3, Hecus X-ray Systems, Graz, Austria) using a 4 kW rotating Cu-anode X-ray generator (Rigaku-Denki, MA, USA). The system incorporates a pulse-height discriminator, which is used in combination with a 10 mm Ni foil to obtain Cu K a radiation (l = 1.542 Å ). Further, the camera is equipped with a Peltier-controlled variable-temperature cuvette (temperature resolution 0.1uC) and a linear one-dimensional position-sensitive detector (PSD 50-M, Hecus X-ray Systems, Graz, Austria) covering the q-range of 0.004 to 0.5 Å 21 (q = 2p sinh /l ). The system allows automatic serial exposures by a programmable temperature unit and time frame generator. The temperature scans were performed in heating direction. After equilibration of the samples for at least 600 s (waiting time) at the respective temperature, the SAXS pattern were recorded with exposure times of 1000 s. For indexing the different mesophases and calculating the corresponding unit lattice parameter a, we applied the respective reflection laws for the cubic and the hexagonal phases [8].