- Email: [email protected]
Aggregation behaviour of a dirhamnolipid biosurfactant secreted by Pseudomonas aeruginosa in aqueous media
Journal of Colloid and Interface Science 307 (2007) 246–253 www.elsevier.com/locate/jcis
Aggregation behaviour of a dirhamnolipid biosurfactant secreted by Pseudomonas aeruginosa in aqueous media Marina Sánchez a , Francisco J. Aranda a , María J. Espuny b , Ana Marqués b , José A. Teruel a , Ángeles Manresa b , Antonio Ortiz a,∗ a Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Universidad de Murcia, E-30100 Murcia, Spain b Laboratorio de Microbiología, Facultad de Farmacia, Universidad de Barcelona, Joan XXIII s/n, E-08028 Barcelona, Spain
Received 6 October 2006; accepted 27 November 2006 Available online 1 December 2006
Abstract The process of micelle formation, along with the formation of higher order aggregates, is described for a dirhamnolipid extracellular biosurfactant secreted by Pseudomonas aeruginosa. As determined by surface tension measurements, at pH 7.4 the CMC of dirhamnolipid is 0.110 mM, whereas at pH 4.0 it falls to 0.010 mM, indicating that the negatively charged diRL has a much higher CMC value than the neutral species. Centrifugation and dynamic light scattering measurements show formation of larger aggregates at concentrations above the CMC. These aggregates have been shown by electron microscopy to be mainly multilamellar vesicles of heterogeneous size. X-ray scattering gave a value of 32 Å for the interlamellar repeat distance of these vesicles. Taking into account the experimental data, a molecular modelling of the dirhamnolipid moiety has been carried out, which details the size of the hydrophilic and hydrophobic portions, and suggests the possible intermolecular interactions responsible for the stabilisation of dirhamnolipid aggregates. The relevance of this aggregation behaviour is discussed with respect to the molecular basis of its activities. © 2006 Elsevier Inc. All rights reserved. Keywords: Rhamnolipids; Bacterial glycolipids; Biosurfactants; CMC; Aggregation behaviour
1. Introduction Biosurfactants are amphiphilic compounds with surface activity produced by a number of microorganisms, including bacteria, yeasts and fungi. These compounds are structurally diverse and most of them are of lipidic nature: glycolipids, lipoaminoacids and lipopeptides. Because of their interesting properties, low toxicity, biodegradable character, and effectiveness at extreme temperature and pH values, there is a growing interest in considering biosurfactants as potential alternatives to compounds obtained by chemical synthesis [1–4]. Biosurfactants have gained importance in various fields, to the extent that the commercial production of these compounds is being encouraged in order to develop more economically attractive processes [5]. * Corresponding author. Fax: +34 968 364147.
E-mail address: [email protected] (A. Ortiz). 0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.11.041
Pseudomonas aeruginosa, a Gram-negative bacterium, is environmentally versatile which can cause disease in particular susceptible individuals and is resistant to antibiotics. The bacterium can use several organic compounds as food sources, which gives it an exceptional ability to colonise ecological niches where nutrients are limited. Pseudomonas aeruginosa secretes rhamnolipids when grown under the appropriate conditions [6]. Rhamnolipids are a group of biosurfactants of glycolipid nature, composed of a hydrophilic head formed by one or two rhamnose molecules, known respectively as monorhamnolipid (monoRL) and dirhamnolipid (diRL), and a hydrophobic tail which contains one or two fatty acids (Fig. 1). The type and proportion of the rhamnolipids produced depends on the bacterial strain, the carbon source used and the culture conditions [7]. Rhamnolipids constitute one of the most interesting classes of biosurfactants because of their advantageous characteristics. With respect to their production, they show high yields as compared to other biosurfactants, and several renewable ma-
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
247
Fig. 1. The chemical structure of the diRL compounds produced by Pseudomonas aeruginosa. For Rha–Rha–C10 –C10 , m, n = 6 and for Rha–Rha–C10 –C12 , m = 8 and n = 6.
terials, like used oils or wastes from the food industry, can be used as carbon sources [7–9]. The whole process of bacterial biosynthesis of rhamnolipids can be classified, therefore, as a green process. Rhamnolipids are surface-active, reducing the surface tension of water from 72 mN m−1 to values close to 30 mN m−1 [10]. The critical micellar concentration of pure rhamnolipids and their mixtures depends to a great extent on the chemical composition of the various species, and varies from 50 to 200 mg l−1 [11]. Rhamnolipids have also been shown to present several interesting biological activities. Both monoRL and diRL behave as exotoxins, restricting the growth of microorganisms like Bacillus subtilis [12,13], and presenting zoosporicidal activity on species of three genera of zoosporic phytopathogens [14]. It is an accepted fact that the majority of the activities mentioned must be related to the interaction of the rhamnolipids with the lipid constituent of biological membranes, as has been shown for other biosurfactants which also affect the structure of phospholipid membranes [15,16]. The compounds secreted by Pseudomonas aeruginosa constitute a heterogeneous mixture of mono- and dirhamnolipids which has been used to obtain most of the published data. However, it is interesting to evaluate the individual contribution of each homologue to the biological properties of the mixture, and thus obtain a molecule with the desired properties for specific uses. We have previously shown that a purified diRL modifies the physicochemical characteristics of phosphatidylethanolamine membranes [17]. In this paper we describe diRL micelle formation as well as the appearance of higher order aggregates, multilamellar membrane vesicles, as a function of the concentration of the glycolipid. 2. Experimental 2.1. Materials All the reagents were of the highest purity available. Purified water was deionised in a Milli-Q equipment from Millipore
(Millipore, Bedford, MA, USA) and had a resistivity of ca. 18 M. Stock solutions of the diRL were prepared in chloroform/methanol (1:1) and stored at −20 ◦ C. The buffers used were 5 mM Hepes pH 7.4 and 100 mM sodium citrate pH 4.0, with the desired concentration of NaCl as indicated. These buffers were chosen for an optimal buffering capacity at both pH values. Most measurements were carried out in 150 mM NaCl in order to work under physiological ionic strength conditions. Water and all the buffer solutions used in this work were filtered through 0.2 µm filters prior to use. 2.2. Rhamnolipid production and purification Biosurfactant producer, strain 47T2 (NCIB 40044), was isolated from contaminated soil samples from Barcelona (Spain) and was selected for its capacity to accumulate surface active rhamnolipids from hydrophobic substrates [18]. This strain was maintained by fortnightly cultures and preserved in cryovials at −20 ◦ C. From previous morphological and biochemical tests the isolate was identified as Pseudomonas aeruginosa [19]. Strain 47T2 secreted rhamnolipids when it was grown in a mineral medium of the following composition (g l−1 ): NaNO3 4.64; K2 HPO4 /KH2 PO4 (ratio 2:1) 1; CaCl2 0.01; KCl 0.1; MgSO4 ·7H2 O 0.0074; yeast extract 0.1 ml l−1 ; and supplemented with 0.05 ml l−1 of a trace mineral solution containing (g l−1 ): H3 BO3 0.26; CuSO4 ·5H2 O 0.5; MnSO4 ·H2 O 0.5; MoNa2 O4 ·2H2 O 0.06; ZnSO4 ·7H2 O 0.7. 50 g l−1 of waste fried oil (1:1 v/v olive/sunflower), which was mainly composed of C18:1 (74.4%), C18:2 (20%), C16:0 (2.7%) and C18:0 (1.4%) was used as carbon source. The strain was incubated at 30 ◦ C and 120 rpm in a reciprocal shaker. Rhamnolipids were recovered from the culture supernatant as follows. Cells were removed from the culture by centrifugation at 8000g in a Kontron centrifuge (Milano, Italy) for 20 min at 15 ◦ C. Purification of rhamnolipids was achieved, following a modification of the method of Rieling et al. [20], by adsorption chromatography on a polystyrene resin, Amberlite XAD2
248
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
(Sigma, St. Louis, USA). The resin (60 g) was placed in a glass column (60 × 3 cm), yielding a bed volume of 200 ml. The column was equilibrated with 0.1 M phosphate buffer, pH 6.1. The culture supernatant was acidified to pH 6.1 and applied through a sieve placed on top of the resin to prevent whirling up. The adsorption of the active compounds on the resin was monitored measuring the surface tension (γst ) of the column outlet. The saturation of the resin was determined when γst of the effluent dropped below 40 mN m−1 . Thereafter, the column was rinsed with three volumes of distilled water until γst of the effluent approached 72 mN m−1 , at which no free fatty acids remained in the column. Biosurfactants were eluted with methanol and the solvent was evaporated under vacuum. A 5 mg aliquot was dissolved in methanol and analysed by HPLC-MS to check the purity of the extract, i.e., the absence of residual fatty acids [21]. The diRL component was separated from the heterogeneous mixture as follows. A slurry of silicagel 60 in chloroform was poured onto a glass chromatography column (2 × 40 cm). Two grams of the crude rhamnolipid mixture was dissolved in 4 ml of chloroform and loaded onto the column. The column was washed, at a flow rate of ca. 1 ml min−1 , with chloroform until neutral lipids were totally eluted, followed by chloroform/methanol 50:3 and 50:5 (which eluted the monoRL component), and chloroform/methanol 50:50 and pure methanol (to elute the diRL component). The composition of the fractions was checked by thin layer chromatography on silica gel plates using chloroform/methanol/H2 O (65:15:2) as mobile phase. The purified diRL component showed a single spot by thin-layer chromatography. Rhamnolipids were quantified by weighing after extensive desiccation under high vacuum until constant weight. The chemical characterisation of the compounds was confirmed by ES-MS as previously described [22]. The composition of this DiRL component has been reported before [22] and it consists mainly of Rha–Rha–C10 –C10 (ca. 50%) and Rha–Rha–C10 –C12 (ca. 29%), with small contributions of three other minor species. 2.3. Sample preparation The diRL aqueous samples were prepared by dispersion of the required amount of biosurfactant in the appropriate buffer, as indicated. Briefly, the desired amount of diRL was dissolved in chloroform/methanol. The solvent was gently evaporated under a stream of dry N2 , to obtain a thin film at the bottom of a glass tube. The last traces of solvent were removed by a further minimum 3 h desiccation under high vacuum. The appropriate buffer was added to the dry samples and these were shaken at room temperature until a homogeneous solution or suspension was obtained.
and the concentration of the glycolipid was determined in both samples with the anthrone reagent [23], with rhamnose being used as standard. 2.5. Surface tension measurements Equilibrium surface tension (γst ) was measured at 25 ◦ C with a Krüss K9 digital tensiometer (Krüss, Helsinki, Finland) by the ring method. The instrument was calibrated against ultrapure water (γst 72 mN m−1 ), pure ethanol (γst 22.7 mN m−1 ) and 30.5% ethanol/water (γst 34.4 mN m−1 ) to ensure accuracy over the entire range of surface tension. Prior to use the platinum ring and all the glassware were sequentially washed with chromic acid, deionised water, acetone and finally flamed with a Bunsen burner. 2.6. Estimation of the surface area per molecule of dirhamnolipid A plot of γst as a function of the ln[diRL] was fitted to a straight line with a slope of dγst /d ln[diRL]. Using the Gibbs equation for the surface [24], the surface excess of an adsorbed molecule, Γ (mol cm−2 ) = −(1/mRT )(dγst /d ln[diRL])T,P , was calculated. In this equation m is the number of independent components. Although the issue about the value of m for ionic surfactants has not been fully clarified [24,25], given that our experiments were performed in the presence of an added electrolyte (150 mM NaCl), then m can be taken as 1, since the concentration of only one species, i.e., the ionised surfactant monomer, changes with the bulk surfactant concentration. The surface area per molecule of dirhamnolipid at the CMC, S (Å2 molecule−1 ) = 1/(Γ × N ), where N is Avogadro’s number, was thus obtained under different conditions. This experiment was performed three times with a SE of ±2.5%. 2.7. Dynamic light scattering measurements Dynamic light scattering measurements were made in a Malvern Autosizer 4800 (Worcester, UK) equipped with a 488 nm argon-ion laser light source, at a power of ca. 30 mW. Samples were placed in a 10 mm diameter cylindrical quartz cuvette placed in a thermostated holder, and data were collected at 90◦ angle. Particle sizes were obtained with the cumulants method and calculated from the diffusion coefficients obtained from a computer analysis using the software provided with the equipment. 2.8. Transmission electron microscopy
2.4. Centrifugation experiments In order to determine the fraction of diRL that could be pelleted by centrifugation, samples prepared at different concentrations of the biosurfactant were centrifuged in a 55.2Ti rotor of a Beckman (California, USA) ultracentrifuge at 100,000g for 2 h at 20 ◦ C. The supernatants and the pellets were separated,
A sample of diRL in aqueous buffer was prepared as described above. Samples were placed on carbon-coated grids for different time periods, stained using 5% ammonium heptamolybdate and washed with distilled water. The grids were examined in an electron microscope EM10 (Zeiss, Germany) shortly after their preparation.
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
249
2.9. Small angle x-ray scattering measurements Small angle X-ray scattering (SAXS) measurements were carried out using a Kratky compact camera (MBraum–Graz Optical Systems, Graz, Austria) and a linear position sensitive detector (PSD; MBraum, Garching, Germany) monitoring the s-range (s = 2 sin θ/λ, 2θ = scattering angle, λ = 1.54 Å) between 0.0075 and 0.07 Å−1 . Nickel-filtered CuKα X-rays were generated by a Philips (Eindhoven, The Netherlands) PW3830 X-ray generator operating at 50 kV and 30 mA. The position calibration of the detector was performed by using silver stearate (d-spacing at 48.8 Å) as reference material. Samples for X-ray scattering analysis were prepared essentially as described above. The diRL suspensions were placed in a steel holder, which provided good thermal contact with the Peltier thermostating unit, with cellophane windows. Typical exposure times were 5 min, allowing 10 min prior to the measurement for temperature equilibration.
Fig. 2. Effect of pH on diRL CMC as determined by surface tension measurements. A plot of the surface tension, γst , as a function of the diRL concentration for a series of diRL solutions prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 (") or in 100 mM sodium citrate, pH 4.0 (!) buffer, at 25 ◦ C.
2.10. Molecular modelling The molecular modelling of the structure and the interactions of two molecules of diRL were generated using Hyperchem software from Hypercube Inc. Geometry optimisation was calculated by combining Fletcher–Reeves [26] and Polak– Ribiere [27] algorithms. The molecular mechanics force field was settled as Amber [28] with distance dependent dielectric constant. The polar headgroups of the lipids were placed inside a box of water molecules and the hydrocarbon chains in vacuum, resembling a non-polar medium allowing hydrophobic interactions. Dirhamnolipids were in their non-protonated state with one negative charge on the carboxylic group, thus corresponding to the pH 7.4 structure. 3. Results and discussion Surfactants can assemble into a wide variety of morphologically different structures [29]. In this work we report a detailed study of the aggregation behaviour of a purified diRL biosurfactant produced by Pseudomonas aeruginosa in aqueous media, and we show a concentration dependent micelle-to-vesicle transformation. The main aim was to characterise the process of micelle formation, as well as other type of higher order aggregates, in order to contribute to the understanding of the molecular basis of its biological activities. 3.1. Monomeric and micellar diRL The CMC of rhamnolipids has been previously determined for various mixtures of heterogeneous composition, including mono- and dirhamnolipids [11,21,30]. We have recently reported the first determination of the CMC of a purified diRL biosurfactant produced by Pseudomonas aeruginosa, both by surface tension and by isothermal titration calorimetry [17]. In order to characterise the process of micelle formation by diRL we have determined the CMC of the purified compound
under different conditions. Fig. 2 shows a plot of γst as a function of the concentration of diRL at two different pH values and equal ionic strength. A dilute solution of the biosurfactant had a value of γst close to 55 mN m−1 . As the concentration of diRL was increased, γst decreased to reach values below 36 mN m−1 . The break point of these plots corresponded to the value of the CMC, which was 0.110 mM at pH 7.4 and 0.010 mM at pH 4.0. According to the reported value of 5.6 for the pKa of the diRL [31], at pH 7.4 more than 98% of the molecules will be negatively charged, whereas at pH 4.0 more than 98% of the molecules will be neutral. Thus, it is clear that the negatively charged diRL has a much higher CMC value (one order of magnitude) than the uncharged species. Most probably, the presence of electrostatic repulsions between the negatively charged diRL molecules makes its association into micellar aggregates more difficult, as compared to the neutral molecules present at pH 4.0. The literature values for the CMC of different rhamnolipid mixtures ranged from 53 to 230 mg l−1 [11,21,30]. The value of 0.110 mM (71.5 mg l−1 ) that we have obtained for pure diRL at pH 7.4 is in good agreement with these values. As compared to other relevant biosurfactants like surfactin (CMC 0.0075 mM) the CMC of diRL is one order of magnitude higher, suggesting that dirhamnolipids behave as weak detergents. From the linear portions of the plots shown in Fig. 2 the slope, dγst /d ln c, was used to calculate the excess surface concentration, Γ , at the interface using the equation Γ (mol cm−2 ) = −(1/mRT )(dγst /d ln[diRL])T,P (see Section 2). A value of 76 ± 1.9 Å2 molecule−1 was obtained for the surface area per molecule at both pH values, indicating that the charge of the polar head of the diRL molecule does not strongly influence its conformation, in contrast to the pattern observed for the CMC. Therefore the differences in CMC at both pH values are not a consequence of a change in surface area. This surface area is of the same order of the typical value for a phospholipid like DPPC in the fluid phase, which is ca.
250
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
Fig. 3. Dependence of diRL CMC on ionic strength. The CMC were determined by surface tension measurements carried out in 5 mM Hepes, pH 7.4 (") or in 100 mM sodium citrate, pH 4.0 (!) buffer, at 25 ◦ C.
72 Å2 molecule−1 [32], which suggests that aqueous diRL solutions might show phospholipid-like characteristics. The effect of ionic strength on diRL CMC, shown in Fig. 3, provided additional interesting information. Whereas at pH 4.0 the CMC was not affected by the ionic strength between 0 and 500 mM NaCl, at pH 7.4 it was strongly reduced as ionic strength was increased within the same range. This was an expected behaviour and confirmed the presence of neutral diRL molecules at pH 4.0 and of negatively charged ones at pH 7.4. 3.2. Vesicle formation Our results thus indicate that, under physiological conditions of pH and ionic strength, diRL had a relatively high CMC, close to 0.1 mM, and a rather large surface area similar to that of a typical phospholipid. These data suggested that diRL might exhibit interesting aggregation characteristics upon increasing concentration, and we therefore carried out a detailed study of the aggregation behaviour at pH 7.4 and 150 mM NaCl. Increasing the concentration of diRL above the CMC led to a sharp increase of the turbidity of the suspensions (Fig. 4, panel A), which went up very rapidly to a concentration ca. 0.5 mM and then essentially levelled off. This increase in turbidity might be due to an increase of the amount of micelles, or to the formation of larger aggregates which could be separated by centrifugation. To check this point the corresponding centrifugation experiment was carried out, and the results are shown in Fig. 4, panel B. It can be observed that upon increasing diRL concentration, the amount of glycolipid in the supernatant decreased with the concomitant increase in the pellet. It is thus clear that, at low concentrations close to the CMC, small structures were present which, upon increasing concentration, began to form larger ones. To obtain an overall picture of this diRL aggregation behaviour, dynamic light scattering measurements were performed at various concentrations close to and above the CMC (Fig. 5), under the same conditions as in the
Fig. 4. Pelletable and non-pelletable diRL aggregates as a function of diRL concentration. Panel A: the turbidity (absorbance at 600 nm) of diRL suspensions of different concentrations prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer, prior to centrifugation. Panel B: the suspensions were centrifuged as described under Section 2, and the diRL concentration was determined in the pellets (") and the supernatants (!). Data correspond to the average of five different experiments ± standard error (error bars).
previous experiment. As a general pattern, aggregates within three size ranges (hydrodynamic diameter) were observed: 43– 66, 350–550 and >1500 nm (Fig. 5, panel A). The smallest ones, 43–66 nm, were most likely diRL micellar structures. Up to concentrations around 1 mM, micelles and 350–550 nm aggregates coexisted essentially in the same proportion. However, a further increase in concentration resulted in the disappearance of these structures in favour of much larger ones (>1500 nm), which were the only ones present at 2.5 mM and above. Altogether, these results indicate that diRL undergoes a micelle-tovesicle transition on increasing the concentration. This type of transitions has been described before for other amphiphiles and the mechanism has been thoroughly discussed (see [33] for a review). The relatively large hydrodynamic radius of the micelles (around 50 nm) could indicate that these structures might be of a different shape than the ellipsoid, perhaps a cylindrical shape, which has been reported before for other surfactants presenting or not presenting a micelle-to-vesicle transition [34,35]. The characterisation of diRL non-micellar aggregates was carried out first by electron microscopy. Fig. 6 shows rep-
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
Fig. 5. Size distribution of diRL aggregates as a function of concentration. The samples were prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer, and measurements were carried out at 25 ◦ C. In panel A the numbers beside the curves indicate the concentration of diRL (mM). Panel B shows the contribution of the different size aggregates, quantified by integration of the plots shown in panel A, as a function of diRL concentration.
resentative negative-staining electron micrographs of concentrated aqueous suspensions of diRL. The samples were predominantly heterogeneous, clearly showing the presence of vesicles with sizes in the range of 150–200 nm. The vesicles were multilamellar (bottom), and elongated vesicles with lengths >1000 nm could also be seen (top), as well as aggregated structures of larger size (bottom). Electron microscopy was thus very much in accordance with the data obtained by dynamic light scattering (Fig. 5), and unequivocally showed the presence of diRL multilamellar vesicles of heterogeneous size at pH 7.4. In fact, it has been reported before that diRL stabilises the lamellar structure in systems containing non-lamellar forming lipids, like phosphatidylethanolamine [16], a result which supports our observations. Vesicle formation has been reported before for a monorhamnolipid also produced by Pseudomonas aeruginosa [36]. In this case the proportion of vesicles present at pH 7.4
251
Fig. 6. Negative staining electron micrographs of diRL aggregates. DiRL suspensions were prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer at 25 ◦ C. Bars = 500 nm.
was only 20%, but 100% at pH 6.8. The differences with our results (we found a proportion of vesicles higher than 80% as shown in Fig. 4) could be due, on the one hand to the presence of just one rhamnose ring which changes the CMC of the monorhamnolipid to ca. 62 µM [31], as compared to the value of 110 µM that we have determined for our diRL or, on the other, to the fact that the former authors [36] performed a statistical quantitation of vesicles from cryo-TEM results, which may not be accurate, as they themselves suggest in their article. 3.3. SAXS and molecular modelling We have made SAXS measurements of diRL aqueous dispersions at pH 7.4 and physiological ionic strength (150 mM NaCl) at 25 ◦ C (Fig. 7). In the investigated q range (q is the scattering modulus, q = (4π sin θ )/λ, where 2θ is the scattering angle) a broad peak was observed, centred at 0.2 Å−1 (corresponding distance 32 Å, d = 2π/q). The absence of a sharp Bragg diffraction peak, characteristic of a multilamellar structure, indicates that diRL under these conditions is organised in
252
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
out that various other combinations are also possible. In bond a the carboxyl group of the first diRL moiety (wireframe model) interacts with the C-3 –OH group of the second rhamnose ring of the second moiety (balls and sticks model). Bond b binds the –OH in the C-4 of the first moiety with the –OH in the C-3 of the second one. Finally the carboxyl group of the second moiety may interact with the C-2 –OH (bond c) or C-3 –OH (bond d) of the first moiety. It is thus clear that the possibility of formation of such a large number of hydrogen bonds between the headgroups constitutes the driving force to keep diRL molecules together and to explain their strong tendency to self-aggregate in aqueous solution. 3.4. Conclusions Fig. 7. SAXS curve of diRL suspension at 25 ◦ C. DiRL was prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer at 20 mM concentration. The continuous line is the corresponding theoretical fit according to the equation described in Section 2.
Dirhamnolipid has been shown to behave as an exotoxin which restricts the growth of Bacillus subtilis [12,13] and presents zoosporicidal activity on species of three genera of zoosporic phytopathogens [14]. It has also been reported that diRL presents immiscibilities and alters the structure of model phospholipid membranes [17,38]. In addition to these interesting biological activities, in this work we have shown that diRL exhibits a micelle-to-vesicle transition upon increasing concentration. The formation of stable lamellar vesicles from this compound would perhaps meet the criteria for drug delivery applications, provided that these vesicles are able to trap an aqueous compartment and that they show the required stability and toxicity properties. Acknowledgments This work was funded by Ministerio de Ciencia y Tecnología, Spain (Grant CTQ2004-00107 to A.O.).
Fig. 8. Molecular modelling of dirhamnolipid and intermolecular interactions. Molecular modelling was performed as described in the text. Two molecules of diRL are shown to depict the molecular interactions between them. One molecule has been represented with wireframe for the sake of simplicity. The hydrogen bonds are indicated with dashed lines.
single bilayers or, as is more likely and supported by the electron microscopy results, in non-organised multilayers. Taking d = 32 Å (the full thickness of the bilayer) the experimental data shown in Fig. 7 were fitted as described before [37], assuming that the bulk aqueous electron density, ρw , was 0.33 e Å−3 . This fitting procedure yielded a similar electron density in the polar as well as in the acyl chain region of diRL. The polar part of the diRL moiety had a thickness of 6.0 Å and the hydrophobic acyl chain region was 10 Å thick (including 0.3 Å of the final methyl group). Taking all these data together, a molecular modelling procedure was carried out to yield the structures shown in Fig. 8. In this figure we have shown two diRL molecules to depict the main intermolecular interactions keeping the molecules together. Whereas the hydrophobic core is mainly maintained by van der Waals interactions between the methylene groups of the acyl chains, a number of hydrogen bonds could be envisaged in the polar region. Here we have shown four particular hydrogen bonds, but it should be pointed
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]
J.D. Desai, I.M. Banat, Microbiol. Mol. Biol. Rev. 61 (1997) 47. S.S. Cameotra, R.S. Makkar, Appl. Microbiol. Biotechnol. 50 (1998) 520. P. Singh, S.S. Cameotra, Trends Biotechnol. 22 (2004) 142. L. Rodrigues, I.M. Banat, J. Teixeira, R. Oliveira, J. Antimicrob. Chemother. 57 (2006) 609. S. Mukherjee, P. Das, R. Sen, Trends Biotechnol. 24 (2006) 509. F.G. Jarvis, M.J. Johnson, J. Am. Chem. Soc. 71 (1949) 4124. G. Soberón-Chávez, F. Lépine, E. Déziel, Appl. Microbiol. Biotechnol. 14 (2005) 1. S. Lang, D. Wullbrandt, Appl. Microbiol. Biotechnol. 51 (1999) 22. I.M. Banat, R.S. Makkar, S.S. Cameotra, Appl. Microbiol. Biotechnol. 53 (2000) 495. J.L. Parra, J. Guinea, A. Manresa, M. Robert, M.E. Mercade, F. Comelles, M.P. Bosch, J. Am. Oil Chem. Soc. 66 (1989) 141. M. Benincasa, A. Abalos, I. Oliveira, A. Manresa, Antonie van Leeuwenhoek 85 (2004) 1. S. Lang, E. Katsiwela, F. Wagner, Fat Sci. Technol. 91 (1989) 363. S. Lang, F. Wagner, Biosurfactants, Dekker, New York, 1993. M.E. Stanghellini, R.M. Miller, Plant Dis. 81 (1997) 4. A. Grau, A. Ortiz, A. De Godos, J.C. Gómez-Fernández, Arch. Biochem. Biophys. 377 (2000) 315. A. Grau, J.C. Gómez-Fernández, F. Peypoux, A. Ortiz, Biochim. Biophys. Acta 1418 (1999) 307. M. Sánchez, J.A. Teruel, M.J. Espuny, A. Marqués, F.J. Aranda, A. Manresa, A. Ortiz, Chem. Phys. Lipids 142 (2006) 118.
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
[18] M. Robert, M.E. Mercadé, P. Bosch, J.L. Parra, M.J. Espuny, M.A. Manresa, J. Guinea, Biotechnol. Lett. 11 (1989) 871. [19] E. Haba, M.J. Espuny, M. Busquets, A. Manresa, J. Appl. Microbiol. 88 (2000) 379. [20] H.E. Rieling, U. Thanei-wyss, L.H. Guerra-Santos, R. Hirt, O. Käppeli, A. Fiechter, Appl. Environ. Microbiol. 51 (1986) 985. [21] A. Abalos, A. Pinazo, R. Infante, M. Casals, F. García, A. Manresa, Langmuir 17 (2001) 1367. [22] E. Haba, A. Pinazo, O. Jauregui, M.J. Espuny, M.R. Infante, A. Manresa, Biotech. Bioeng. 81 (2003) 316. [23] J.E. Hodge, B.T. Hofreiter, Methods in Carbohydrate Chemistry, Academic Press, New York, 1962. [24] S. Bae, K. Haage, K. Wantke, H. Motschmann, Phys. Chem. B 103 (1999) 1045. [25] G. Andersson, T. Krebs, H. Morgner, Phys. Chem. Chem. Phys. 7 (2005) 136. [26] R. Fletcher, C. Reeves, J. Comput. 7 (1964) 149. [27] B. Polak, G. Ribiere, Rev. Fr. Imform. Rech. Oper. 16 (1969) 35.
253
[28] S.J. Weiner, P.A. Kollman, D.T. Nguyen, D.A. Case, J. Comput. Chem. 7 (1986) 230. [29] S. Svenson, Curr. Opin. Colloid Interface Sci. 9 (2002) 201. [30] J. Mata-Sandoval, J. Karns, A. Torrens, J. Chromatogr. 864 (1999) 211. [31] Y. Ishigami, Y. Gama, H. Nagahora, M. Yamaguchi, H. Ankara, T. Kamata, Chem. Lett. 5 (1987) 763. [32] R.L. Thurmond, S.W. Dodd, M.F. Brown, Biophys. J. 59 (1991) 108. [33] S. Svenson, Curr. Opin. Colloid Interface Sci. 9 (2004) 201. [34] M. Johnsson, A. Wagenaar, J.B.F.N. Engberts, J. Am. Chem. Soc. 125 (2003) 757. [35] S. Zhou, X. Chang, J. Wang, W. Gao, R. Akhverdiyeva, V. Shah, R. Gross, Langmuir 20 (2004) 7926. [36] J.T. Champion, J.C. Gilkey, H. Lamparski, J. Retterer, R.M. Miller, J. Colloid Interface Sci. 170 (1995) 569. [37] K.A. Riske, L.Q. Amaral, M.T. Lamy-Freund, Biochim. Biophys. Acta 1511 (2001) 297. [38] A. Ortiz, J.A. Teruel, M.J. Espuny, A. Marqués, A. Manresa, F.J. Aranda, Int. J. Pharm. 325 (2006) 99.
Aggregation behaviour of a dirhamnolipid biosurfactant secreted by Pseudomonas aeruginosa in aqueous media Marina Sánchez a , Francisco J. Aranda a , María J. Espuny b , Ana Marqués b , José A. Teruel a , Ángeles Manresa b , Antonio Ortiz a,∗ a Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Universidad de Murcia, E-30100 Murcia, Spain b Laboratorio de Microbiología, Facultad de Farmacia, Universidad de Barcelona, Joan XXIII s/n, E-08028 Barcelona, Spain
Received 6 October 2006; accepted 27 November 2006 Available online 1 December 2006
Abstract The process of micelle formation, along with the formation of higher order aggregates, is described for a dirhamnolipid extracellular biosurfactant secreted by Pseudomonas aeruginosa. As determined by surface tension measurements, at pH 7.4 the CMC of dirhamnolipid is 0.110 mM, whereas at pH 4.0 it falls to 0.010 mM, indicating that the negatively charged diRL has a much higher CMC value than the neutral species. Centrifugation and dynamic light scattering measurements show formation of larger aggregates at concentrations above the CMC. These aggregates have been shown by electron microscopy to be mainly multilamellar vesicles of heterogeneous size. X-ray scattering gave a value of 32 Å for the interlamellar repeat distance of these vesicles. Taking into account the experimental data, a molecular modelling of the dirhamnolipid moiety has been carried out, which details the size of the hydrophilic and hydrophobic portions, and suggests the possible intermolecular interactions responsible for the stabilisation of dirhamnolipid aggregates. The relevance of this aggregation behaviour is discussed with respect to the molecular basis of its activities. © 2006 Elsevier Inc. All rights reserved. Keywords: Rhamnolipids; Bacterial glycolipids; Biosurfactants; CMC; Aggregation behaviour
1. Introduction Biosurfactants are amphiphilic compounds with surface activity produced by a number of microorganisms, including bacteria, yeasts and fungi. These compounds are structurally diverse and most of them are of lipidic nature: glycolipids, lipoaminoacids and lipopeptides. Because of their interesting properties, low toxicity, biodegradable character, and effectiveness at extreme temperature and pH values, there is a growing interest in considering biosurfactants as potential alternatives to compounds obtained by chemical synthesis [1–4]. Biosurfactants have gained importance in various fields, to the extent that the commercial production of these compounds is being encouraged in order to develop more economically attractive processes [5]. * Corresponding author. Fax: +34 968 364147.
E-mail address: [email protected] (A. Ortiz). 0021-9797/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2006.11.041
Pseudomonas aeruginosa, a Gram-negative bacterium, is environmentally versatile which can cause disease in particular susceptible individuals and is resistant to antibiotics. The bacterium can use several organic compounds as food sources, which gives it an exceptional ability to colonise ecological niches where nutrients are limited. Pseudomonas aeruginosa secretes rhamnolipids when grown under the appropriate conditions [6]. Rhamnolipids are a group of biosurfactants of glycolipid nature, composed of a hydrophilic head formed by one or two rhamnose molecules, known respectively as monorhamnolipid (monoRL) and dirhamnolipid (diRL), and a hydrophobic tail which contains one or two fatty acids (Fig. 1). The type and proportion of the rhamnolipids produced depends on the bacterial strain, the carbon source used and the culture conditions [7]. Rhamnolipids constitute one of the most interesting classes of biosurfactants because of their advantageous characteristics. With respect to their production, they show high yields as compared to other biosurfactants, and several renewable ma-
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
247
Fig. 1. The chemical structure of the diRL compounds produced by Pseudomonas aeruginosa. For Rha–Rha–C10 –C10 , m, n = 6 and for Rha–Rha–C10 –C12 , m = 8 and n = 6.
terials, like used oils or wastes from the food industry, can be used as carbon sources [7–9]. The whole process of bacterial biosynthesis of rhamnolipids can be classified, therefore, as a green process. Rhamnolipids are surface-active, reducing the surface tension of water from 72 mN m−1 to values close to 30 mN m−1 [10]. The critical micellar concentration of pure rhamnolipids and their mixtures depends to a great extent on the chemical composition of the various species, and varies from 50 to 200 mg l−1 [11]. Rhamnolipids have also been shown to present several interesting biological activities. Both monoRL and diRL behave as exotoxins, restricting the growth of microorganisms like Bacillus subtilis [12,13], and presenting zoosporicidal activity on species of three genera of zoosporic phytopathogens [14]. It is an accepted fact that the majority of the activities mentioned must be related to the interaction of the rhamnolipids with the lipid constituent of biological membranes, as has been shown for other biosurfactants which also affect the structure of phospholipid membranes [15,16]. The compounds secreted by Pseudomonas aeruginosa constitute a heterogeneous mixture of mono- and dirhamnolipids which has been used to obtain most of the published data. However, it is interesting to evaluate the individual contribution of each homologue to the biological properties of the mixture, and thus obtain a molecule with the desired properties for specific uses. We have previously shown that a purified diRL modifies the physicochemical characteristics of phosphatidylethanolamine membranes [17]. In this paper we describe diRL micelle formation as well as the appearance of higher order aggregates, multilamellar membrane vesicles, as a function of the concentration of the glycolipid. 2. Experimental 2.1. Materials All the reagents were of the highest purity available. Purified water was deionised in a Milli-Q equipment from Millipore
(Millipore, Bedford, MA, USA) and had a resistivity of ca. 18 M. Stock solutions of the diRL were prepared in chloroform/methanol (1:1) and stored at −20 ◦ C. The buffers used were 5 mM Hepes pH 7.4 and 100 mM sodium citrate pH 4.0, with the desired concentration of NaCl as indicated. These buffers were chosen for an optimal buffering capacity at both pH values. Most measurements were carried out in 150 mM NaCl in order to work under physiological ionic strength conditions. Water and all the buffer solutions used in this work were filtered through 0.2 µm filters prior to use. 2.2. Rhamnolipid production and purification Biosurfactant producer, strain 47T2 (NCIB 40044), was isolated from contaminated soil samples from Barcelona (Spain) and was selected for its capacity to accumulate surface active rhamnolipids from hydrophobic substrates [18]. This strain was maintained by fortnightly cultures and preserved in cryovials at −20 ◦ C. From previous morphological and biochemical tests the isolate was identified as Pseudomonas aeruginosa [19]. Strain 47T2 secreted rhamnolipids when it was grown in a mineral medium of the following composition (g l−1 ): NaNO3 4.64; K2 HPO4 /KH2 PO4 (ratio 2:1) 1; CaCl2 0.01; KCl 0.1; MgSO4 ·7H2 O 0.0074; yeast extract 0.1 ml l−1 ; and supplemented with 0.05 ml l−1 of a trace mineral solution containing (g l−1 ): H3 BO3 0.26; CuSO4 ·5H2 O 0.5; MnSO4 ·H2 O 0.5; MoNa2 O4 ·2H2 O 0.06; ZnSO4 ·7H2 O 0.7. 50 g l−1 of waste fried oil (1:1 v/v olive/sunflower), which was mainly composed of C18:1 (74.4%), C18:2 (20%), C16:0 (2.7%) and C18:0 (1.4%) was used as carbon source. The strain was incubated at 30 ◦ C and 120 rpm in a reciprocal shaker. Rhamnolipids were recovered from the culture supernatant as follows. Cells were removed from the culture by centrifugation at 8000g in a Kontron centrifuge (Milano, Italy) for 20 min at 15 ◦ C. Purification of rhamnolipids was achieved, following a modification of the method of Rieling et al. [20], by adsorption chromatography on a polystyrene resin, Amberlite XAD2
248
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
(Sigma, St. Louis, USA). The resin (60 g) was placed in a glass column (60 × 3 cm), yielding a bed volume of 200 ml. The column was equilibrated with 0.1 M phosphate buffer, pH 6.1. The culture supernatant was acidified to pH 6.1 and applied through a sieve placed on top of the resin to prevent whirling up. The adsorption of the active compounds on the resin was monitored measuring the surface tension (γst ) of the column outlet. The saturation of the resin was determined when γst of the effluent dropped below 40 mN m−1 . Thereafter, the column was rinsed with three volumes of distilled water until γst of the effluent approached 72 mN m−1 , at which no free fatty acids remained in the column. Biosurfactants were eluted with methanol and the solvent was evaporated under vacuum. A 5 mg aliquot was dissolved in methanol and analysed by HPLC-MS to check the purity of the extract, i.e., the absence of residual fatty acids [21]. The diRL component was separated from the heterogeneous mixture as follows. A slurry of silicagel 60 in chloroform was poured onto a glass chromatography column (2 × 40 cm). Two grams of the crude rhamnolipid mixture was dissolved in 4 ml of chloroform and loaded onto the column. The column was washed, at a flow rate of ca. 1 ml min−1 , with chloroform until neutral lipids were totally eluted, followed by chloroform/methanol 50:3 and 50:5 (which eluted the monoRL component), and chloroform/methanol 50:50 and pure methanol (to elute the diRL component). The composition of the fractions was checked by thin layer chromatography on silica gel plates using chloroform/methanol/H2 O (65:15:2) as mobile phase. The purified diRL component showed a single spot by thin-layer chromatography. Rhamnolipids were quantified by weighing after extensive desiccation under high vacuum until constant weight. The chemical characterisation of the compounds was confirmed by ES-MS as previously described [22]. The composition of this DiRL component has been reported before [22] and it consists mainly of Rha–Rha–C10 –C10 (ca. 50%) and Rha–Rha–C10 –C12 (ca. 29%), with small contributions of three other minor species. 2.3. Sample preparation The diRL aqueous samples were prepared by dispersion of the required amount of biosurfactant in the appropriate buffer, as indicated. Briefly, the desired amount of diRL was dissolved in chloroform/methanol. The solvent was gently evaporated under a stream of dry N2 , to obtain a thin film at the bottom of a glass tube. The last traces of solvent were removed by a further minimum 3 h desiccation under high vacuum. The appropriate buffer was added to the dry samples and these were shaken at room temperature until a homogeneous solution or suspension was obtained.
and the concentration of the glycolipid was determined in both samples with the anthrone reagent [23], with rhamnose being used as standard. 2.5. Surface tension measurements Equilibrium surface tension (γst ) was measured at 25 ◦ C with a Krüss K9 digital tensiometer (Krüss, Helsinki, Finland) by the ring method. The instrument was calibrated against ultrapure water (γst 72 mN m−1 ), pure ethanol (γst 22.7 mN m−1 ) and 30.5% ethanol/water (γst 34.4 mN m−1 ) to ensure accuracy over the entire range of surface tension. Prior to use the platinum ring and all the glassware were sequentially washed with chromic acid, deionised water, acetone and finally flamed with a Bunsen burner. 2.6. Estimation of the surface area per molecule of dirhamnolipid A plot of γst as a function of the ln[diRL] was fitted to a straight line with a slope of dγst /d ln[diRL]. Using the Gibbs equation for the surface [24], the surface excess of an adsorbed molecule, Γ (mol cm−2 ) = −(1/mRT )(dγst /d ln[diRL])T,P , was calculated. In this equation m is the number of independent components. Although the issue about the value of m for ionic surfactants has not been fully clarified [24,25], given that our experiments were performed in the presence of an added electrolyte (150 mM NaCl), then m can be taken as 1, since the concentration of only one species, i.e., the ionised surfactant monomer, changes with the bulk surfactant concentration. The surface area per molecule of dirhamnolipid at the CMC, S (Å2 molecule−1 ) = 1/(Γ × N ), where N is Avogadro’s number, was thus obtained under different conditions. This experiment was performed three times with a SE of ±2.5%. 2.7. Dynamic light scattering measurements Dynamic light scattering measurements were made in a Malvern Autosizer 4800 (Worcester, UK) equipped with a 488 nm argon-ion laser light source, at a power of ca. 30 mW. Samples were placed in a 10 mm diameter cylindrical quartz cuvette placed in a thermostated holder, and data were collected at 90◦ angle. Particle sizes were obtained with the cumulants method and calculated from the diffusion coefficients obtained from a computer analysis using the software provided with the equipment. 2.8. Transmission electron microscopy
2.4. Centrifugation experiments In order to determine the fraction of diRL that could be pelleted by centrifugation, samples prepared at different concentrations of the biosurfactant were centrifuged in a 55.2Ti rotor of a Beckman (California, USA) ultracentrifuge at 100,000g for 2 h at 20 ◦ C. The supernatants and the pellets were separated,
A sample of diRL in aqueous buffer was prepared as described above. Samples were placed on carbon-coated grids for different time periods, stained using 5% ammonium heptamolybdate and washed with distilled water. The grids were examined in an electron microscope EM10 (Zeiss, Germany) shortly after their preparation.
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
249
2.9. Small angle x-ray scattering measurements Small angle X-ray scattering (SAXS) measurements were carried out using a Kratky compact camera (MBraum–Graz Optical Systems, Graz, Austria) and a linear position sensitive detector (PSD; MBraum, Garching, Germany) monitoring the s-range (s = 2 sin θ/λ, 2θ = scattering angle, λ = 1.54 Å) between 0.0075 and 0.07 Å−1 . Nickel-filtered CuKα X-rays were generated by a Philips (Eindhoven, The Netherlands) PW3830 X-ray generator operating at 50 kV and 30 mA. The position calibration of the detector was performed by using silver stearate (d-spacing at 48.8 Å) as reference material. Samples for X-ray scattering analysis were prepared essentially as described above. The diRL suspensions were placed in a steel holder, which provided good thermal contact with the Peltier thermostating unit, with cellophane windows. Typical exposure times were 5 min, allowing 10 min prior to the measurement for temperature equilibration.
Fig. 2. Effect of pH on diRL CMC as determined by surface tension measurements. A plot of the surface tension, γst , as a function of the diRL concentration for a series of diRL solutions prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 (") or in 100 mM sodium citrate, pH 4.0 (!) buffer, at 25 ◦ C.
2.10. Molecular modelling The molecular modelling of the structure and the interactions of two molecules of diRL were generated using Hyperchem software from Hypercube Inc. Geometry optimisation was calculated by combining Fletcher–Reeves [26] and Polak– Ribiere [27] algorithms. The molecular mechanics force field was settled as Amber [28] with distance dependent dielectric constant. The polar headgroups of the lipids were placed inside a box of water molecules and the hydrocarbon chains in vacuum, resembling a non-polar medium allowing hydrophobic interactions. Dirhamnolipids were in their non-protonated state with one negative charge on the carboxylic group, thus corresponding to the pH 7.4 structure. 3. Results and discussion Surfactants can assemble into a wide variety of morphologically different structures [29]. In this work we report a detailed study of the aggregation behaviour of a purified diRL biosurfactant produced by Pseudomonas aeruginosa in aqueous media, and we show a concentration dependent micelle-to-vesicle transformation. The main aim was to characterise the process of micelle formation, as well as other type of higher order aggregates, in order to contribute to the understanding of the molecular basis of its biological activities. 3.1. Monomeric and micellar diRL The CMC of rhamnolipids has been previously determined for various mixtures of heterogeneous composition, including mono- and dirhamnolipids [11,21,30]. We have recently reported the first determination of the CMC of a purified diRL biosurfactant produced by Pseudomonas aeruginosa, both by surface tension and by isothermal titration calorimetry [17]. In order to characterise the process of micelle formation by diRL we have determined the CMC of the purified compound
under different conditions. Fig. 2 shows a plot of γst as a function of the concentration of diRL at two different pH values and equal ionic strength. A dilute solution of the biosurfactant had a value of γst close to 55 mN m−1 . As the concentration of diRL was increased, γst decreased to reach values below 36 mN m−1 . The break point of these plots corresponded to the value of the CMC, which was 0.110 mM at pH 7.4 and 0.010 mM at pH 4.0. According to the reported value of 5.6 for the pKa of the diRL [31], at pH 7.4 more than 98% of the molecules will be negatively charged, whereas at pH 4.0 more than 98% of the molecules will be neutral. Thus, it is clear that the negatively charged diRL has a much higher CMC value (one order of magnitude) than the uncharged species. Most probably, the presence of electrostatic repulsions between the negatively charged diRL molecules makes its association into micellar aggregates more difficult, as compared to the neutral molecules present at pH 4.0. The literature values for the CMC of different rhamnolipid mixtures ranged from 53 to 230 mg l−1 [11,21,30]. The value of 0.110 mM (71.5 mg l−1 ) that we have obtained for pure diRL at pH 7.4 is in good agreement with these values. As compared to other relevant biosurfactants like surfactin (CMC 0.0075 mM) the CMC of diRL is one order of magnitude higher, suggesting that dirhamnolipids behave as weak detergents. From the linear portions of the plots shown in Fig. 2 the slope, dγst /d ln c, was used to calculate the excess surface concentration, Γ , at the interface using the equation Γ (mol cm−2 ) = −(1/mRT )(dγst /d ln[diRL])T,P (see Section 2). A value of 76 ± 1.9 Å2 molecule−1 was obtained for the surface area per molecule at both pH values, indicating that the charge of the polar head of the diRL molecule does not strongly influence its conformation, in contrast to the pattern observed for the CMC. Therefore the differences in CMC at both pH values are not a consequence of a change in surface area. This surface area is of the same order of the typical value for a phospholipid like DPPC in the fluid phase, which is ca.
250
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
Fig. 3. Dependence of diRL CMC on ionic strength. The CMC were determined by surface tension measurements carried out in 5 mM Hepes, pH 7.4 (") or in 100 mM sodium citrate, pH 4.0 (!) buffer, at 25 ◦ C.
72 Å2 molecule−1 [32], which suggests that aqueous diRL solutions might show phospholipid-like characteristics. The effect of ionic strength on diRL CMC, shown in Fig. 3, provided additional interesting information. Whereas at pH 4.0 the CMC was not affected by the ionic strength between 0 and 500 mM NaCl, at pH 7.4 it was strongly reduced as ionic strength was increased within the same range. This was an expected behaviour and confirmed the presence of neutral diRL molecules at pH 4.0 and of negatively charged ones at pH 7.4. 3.2. Vesicle formation Our results thus indicate that, under physiological conditions of pH and ionic strength, diRL had a relatively high CMC, close to 0.1 mM, and a rather large surface area similar to that of a typical phospholipid. These data suggested that diRL might exhibit interesting aggregation characteristics upon increasing concentration, and we therefore carried out a detailed study of the aggregation behaviour at pH 7.4 and 150 mM NaCl. Increasing the concentration of diRL above the CMC led to a sharp increase of the turbidity of the suspensions (Fig. 4, panel A), which went up very rapidly to a concentration ca. 0.5 mM and then essentially levelled off. This increase in turbidity might be due to an increase of the amount of micelles, or to the formation of larger aggregates which could be separated by centrifugation. To check this point the corresponding centrifugation experiment was carried out, and the results are shown in Fig. 4, panel B. It can be observed that upon increasing diRL concentration, the amount of glycolipid in the supernatant decreased with the concomitant increase in the pellet. It is thus clear that, at low concentrations close to the CMC, small structures were present which, upon increasing concentration, began to form larger ones. To obtain an overall picture of this diRL aggregation behaviour, dynamic light scattering measurements were performed at various concentrations close to and above the CMC (Fig. 5), under the same conditions as in the
Fig. 4. Pelletable and non-pelletable diRL aggregates as a function of diRL concentration. Panel A: the turbidity (absorbance at 600 nm) of diRL suspensions of different concentrations prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer, prior to centrifugation. Panel B: the suspensions were centrifuged as described under Section 2, and the diRL concentration was determined in the pellets (") and the supernatants (!). Data correspond to the average of five different experiments ± standard error (error bars).
previous experiment. As a general pattern, aggregates within three size ranges (hydrodynamic diameter) were observed: 43– 66, 350–550 and >1500 nm (Fig. 5, panel A). The smallest ones, 43–66 nm, were most likely diRL micellar structures. Up to concentrations around 1 mM, micelles and 350–550 nm aggregates coexisted essentially in the same proportion. However, a further increase in concentration resulted in the disappearance of these structures in favour of much larger ones (>1500 nm), which were the only ones present at 2.5 mM and above. Altogether, these results indicate that diRL undergoes a micelle-tovesicle transition on increasing the concentration. This type of transitions has been described before for other amphiphiles and the mechanism has been thoroughly discussed (see [33] for a review). The relatively large hydrodynamic radius of the micelles (around 50 nm) could indicate that these structures might be of a different shape than the ellipsoid, perhaps a cylindrical shape, which has been reported before for other surfactants presenting or not presenting a micelle-to-vesicle transition [34,35]. The characterisation of diRL non-micellar aggregates was carried out first by electron microscopy. Fig. 6 shows rep-
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
Fig. 5. Size distribution of diRL aggregates as a function of concentration. The samples were prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer, and measurements were carried out at 25 ◦ C. In panel A the numbers beside the curves indicate the concentration of diRL (mM). Panel B shows the contribution of the different size aggregates, quantified by integration of the plots shown in panel A, as a function of diRL concentration.
resentative negative-staining electron micrographs of concentrated aqueous suspensions of diRL. The samples were predominantly heterogeneous, clearly showing the presence of vesicles with sizes in the range of 150–200 nm. The vesicles were multilamellar (bottom), and elongated vesicles with lengths >1000 nm could also be seen (top), as well as aggregated structures of larger size (bottom). Electron microscopy was thus very much in accordance with the data obtained by dynamic light scattering (Fig. 5), and unequivocally showed the presence of diRL multilamellar vesicles of heterogeneous size at pH 7.4. In fact, it has been reported before that diRL stabilises the lamellar structure in systems containing non-lamellar forming lipids, like phosphatidylethanolamine [16], a result which supports our observations. Vesicle formation has been reported before for a monorhamnolipid also produced by Pseudomonas aeruginosa [36]. In this case the proportion of vesicles present at pH 7.4
251
Fig. 6. Negative staining electron micrographs of diRL aggregates. DiRL suspensions were prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer at 25 ◦ C. Bars = 500 nm.
was only 20%, but 100% at pH 6.8. The differences with our results (we found a proportion of vesicles higher than 80% as shown in Fig. 4) could be due, on the one hand to the presence of just one rhamnose ring which changes the CMC of the monorhamnolipid to ca. 62 µM [31], as compared to the value of 110 µM that we have determined for our diRL or, on the other, to the fact that the former authors [36] performed a statistical quantitation of vesicles from cryo-TEM results, which may not be accurate, as they themselves suggest in their article. 3.3. SAXS and molecular modelling We have made SAXS measurements of diRL aqueous dispersions at pH 7.4 and physiological ionic strength (150 mM NaCl) at 25 ◦ C (Fig. 7). In the investigated q range (q is the scattering modulus, q = (4π sin θ )/λ, where 2θ is the scattering angle) a broad peak was observed, centred at 0.2 Å−1 (corresponding distance 32 Å, d = 2π/q). The absence of a sharp Bragg diffraction peak, characteristic of a multilamellar structure, indicates that diRL under these conditions is organised in
252
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
out that various other combinations are also possible. In bond a the carboxyl group of the first diRL moiety (wireframe model) interacts with the C-3 –OH group of the second rhamnose ring of the second moiety (balls and sticks model). Bond b binds the –OH in the C-4 of the first moiety with the –OH in the C-3 of the second one. Finally the carboxyl group of the second moiety may interact with the C-2 –OH (bond c) or C-3 –OH (bond d) of the first moiety. It is thus clear that the possibility of formation of such a large number of hydrogen bonds between the headgroups constitutes the driving force to keep diRL molecules together and to explain their strong tendency to self-aggregate in aqueous solution. 3.4. Conclusions Fig. 7. SAXS curve of diRL suspension at 25 ◦ C. DiRL was prepared in 150 mM NaCl, 5 mM Hepes, pH 7.4 buffer at 20 mM concentration. The continuous line is the corresponding theoretical fit according to the equation described in Section 2.
Dirhamnolipid has been shown to behave as an exotoxin which restricts the growth of Bacillus subtilis [12,13] and presents zoosporicidal activity on species of three genera of zoosporic phytopathogens [14]. It has also been reported that diRL presents immiscibilities and alters the structure of model phospholipid membranes [17,38]. In addition to these interesting biological activities, in this work we have shown that diRL exhibits a micelle-to-vesicle transition upon increasing concentration. The formation of stable lamellar vesicles from this compound would perhaps meet the criteria for drug delivery applications, provided that these vesicles are able to trap an aqueous compartment and that they show the required stability and toxicity properties. Acknowledgments This work was funded by Ministerio de Ciencia y Tecnología, Spain (Grant CTQ2004-00107 to A.O.).
Fig. 8. Molecular modelling of dirhamnolipid and intermolecular interactions. Molecular modelling was performed as described in the text. Two molecules of diRL are shown to depict the molecular interactions between them. One molecule has been represented with wireframe for the sake of simplicity. The hydrogen bonds are indicated with dashed lines.
single bilayers or, as is more likely and supported by the electron microscopy results, in non-organised multilayers. Taking d = 32 Å (the full thickness of the bilayer) the experimental data shown in Fig. 7 were fitted as described before [37], assuming that the bulk aqueous electron density, ρw , was 0.33 e Å−3 . This fitting procedure yielded a similar electron density in the polar as well as in the acyl chain region of diRL. The polar part of the diRL moiety had a thickness of 6.0 Å and the hydrophobic acyl chain region was 10 Å thick (including 0.3 Å of the final methyl group). Taking all these data together, a molecular modelling procedure was carried out to yield the structures shown in Fig. 8. In this figure we have shown two diRL molecules to depict the main intermolecular interactions keeping the molecules together. Whereas the hydrophobic core is mainly maintained by van der Waals interactions between the methylene groups of the acyl chains, a number of hydrogen bonds could be envisaged in the polar region. Here we have shown four particular hydrogen bonds, but it should be pointed
References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17]
J.D. Desai, I.M. Banat, Microbiol. Mol. Biol. Rev. 61 (1997) 47. S.S. Cameotra, R.S. Makkar, Appl. Microbiol. Biotechnol. 50 (1998) 520. P. Singh, S.S. Cameotra, Trends Biotechnol. 22 (2004) 142. L. Rodrigues, I.M. Banat, J. Teixeira, R. Oliveira, J. Antimicrob. Chemother. 57 (2006) 609. S. Mukherjee, P. Das, R. Sen, Trends Biotechnol. 24 (2006) 509. F.G. Jarvis, M.J. Johnson, J. Am. Chem. Soc. 71 (1949) 4124. G. Soberón-Chávez, F. Lépine, E. Déziel, Appl. Microbiol. Biotechnol. 14 (2005) 1. S. Lang, D. Wullbrandt, Appl. Microbiol. Biotechnol. 51 (1999) 22. I.M. Banat, R.S. Makkar, S.S. Cameotra, Appl. Microbiol. Biotechnol. 53 (2000) 495. J.L. Parra, J. Guinea, A. Manresa, M. Robert, M.E. Mercade, F. Comelles, M.P. Bosch, J. Am. Oil Chem. Soc. 66 (1989) 141. M. Benincasa, A. Abalos, I. Oliveira, A. Manresa, Antonie van Leeuwenhoek 85 (2004) 1. S. Lang, E. Katsiwela, F. Wagner, Fat Sci. Technol. 91 (1989) 363. S. Lang, F. Wagner, Biosurfactants, Dekker, New York, 1993. M.E. Stanghellini, R.M. Miller, Plant Dis. 81 (1997) 4. A. Grau, A. Ortiz, A. De Godos, J.C. Gómez-Fernández, Arch. Biochem. Biophys. 377 (2000) 315. A. Grau, J.C. Gómez-Fernández, F. Peypoux, A. Ortiz, Biochim. Biophys. Acta 1418 (1999) 307. M. Sánchez, J.A. Teruel, M.J. Espuny, A. Marqués, F.J. Aranda, A. Manresa, A. Ortiz, Chem. Phys. Lipids 142 (2006) 118.
M. Sánchez et al. / Journal of Colloid and Interface Science 307 (2007) 246–253
[18] M. Robert, M.E. Mercadé, P. Bosch, J.L. Parra, M.J. Espuny, M.A. Manresa, J. Guinea, Biotechnol. Lett. 11 (1989) 871. [19] E. Haba, M.J. Espuny, M. Busquets, A. Manresa, J. Appl. Microbiol. 88 (2000) 379. [20] H.E. Rieling, U. Thanei-wyss, L.H. Guerra-Santos, R. Hirt, O. Käppeli, A. Fiechter, Appl. Environ. Microbiol. 51 (1986) 985. [21] A. Abalos, A. Pinazo, R. Infante, M. Casals, F. García, A. Manresa, Langmuir 17 (2001) 1367. [22] E. Haba, A. Pinazo, O. Jauregui, M.J. Espuny, M.R. Infante, A. Manresa, Biotech. Bioeng. 81 (2003) 316. [23] J.E. Hodge, B.T. Hofreiter, Methods in Carbohydrate Chemistry, Academic Press, New York, 1962. [24] S. Bae, K. Haage, K. Wantke, H. Motschmann, Phys. Chem. B 103 (1999) 1045. [25] G. Andersson, T. Krebs, H. Morgner, Phys. Chem. Chem. Phys. 7 (2005) 136. [26] R. Fletcher, C. Reeves, J. Comput. 7 (1964) 149. [27] B. Polak, G. Ribiere, Rev. Fr. Imform. Rech. Oper. 16 (1969) 35.
253
[28] S.J. Weiner, P.A. Kollman, D.T. Nguyen, D.A. Case, J. Comput. Chem. 7 (1986) 230. [29] S. Svenson, Curr. Opin. Colloid Interface Sci. 9 (2002) 201. [30] J. Mata-Sandoval, J. Karns, A. Torrens, J. Chromatogr. 864 (1999) 211. [31] Y. Ishigami, Y. Gama, H. Nagahora, M. Yamaguchi, H. Ankara, T. Kamata, Chem. Lett. 5 (1987) 763. [32] R.L. Thurmond, S.W. Dodd, M.F. Brown, Biophys. J. 59 (1991) 108. [33] S. Svenson, Curr. Opin. Colloid Interface Sci. 9 (2004) 201. [34] M. Johnsson, A. Wagenaar, J.B.F.N. Engberts, J. Am. Chem. Soc. 125 (2003) 757. [35] S. Zhou, X. Chang, J. Wang, W. Gao, R. Akhverdiyeva, V. Shah, R. Gross, Langmuir 20 (2004) 7926. [36] J.T. Champion, J.C. Gilkey, H. Lamparski, J. Retterer, R.M. Miller, J. Colloid Interface Sci. 170 (1995) 569. [37] K.A. Riske, L.Q. Amaral, M.T. Lamy-Freund, Biochim. Biophys. Acta 1511 (2001) 297. [38] A. Ortiz, J.A. Teruel, M.J. Espuny, A. Marqués, A. Manresa, F.J. Aranda, Int. J. Pharm. 325 (2006) 99.