Catanionic Systems from Conversion of Nucleotides into Nucleo-Lipids

Feb 1, 2008 - Chimica, UniVersita` di Bari, V. Orabona 4, I-70126 Bari, Italy. ReceiVed August 21, 2007. In Final Form: December 4, 2007. This article...
0 downloads 0 Views 455KB Size
Download PDF
2348

Langmuir 2008, 24, 2348-2355

Catanionic Systems from Conversion of Nucleotides into Nucleo-Lipids Ruggero Angelico,*,† Andrea Ceglie,† Francesca Cuomo,† Cosimo Cardellicchio,‡ Giuseppe Mascolo,§ and Giuseppe Colafemmina| Consorzio per lo sViluppo dei Sistemi a Grande Interfase (CSGI) c/o UniVersita` del Molise (DISTAAM), V. De Sanctis, I-86100 Campobasso, Italy, CNR (ICCOM), Dipartimento di Chimica, UniVersita` di Bari, Via Orabona 4, I-70126 Bari, Italy, CNR (IRSA), Via F. De Blasio 5, I-70123 Bari, Italy, and Dipartimento di Chimica, UniVersita` di Bari, V. Orabona 4, I-70126 Bari, Italy ReceiVed August 21, 2007. In Final Form: December 4, 2007 This article focuses on reactions performed in nanostructured environments where the pair of complementary nucleotides, 5′-AMP and 5′-UMP, are converted into their amphiphilic derivatives. The synthesis is carried out by using the hydrophobic reactant dodecyl epoxide (DE) dispersed in a micellar solution based on the cationic surfactant cetyltrimethylammoniumbromide (CTAB). Novel nucleo-lipids monomers and CTAB molecules give rise to the spontaneous self-assembly of catanionic supramolecular structures in water, showing typical Maltese crosses in optical microscopy. In the final colloidal suspensions, mono- and dichained derivatives have been identified in the system incubated with 5′-UMP through LC-QqTOF-MS analysis, whereas only mono-alkylated adducts are found in the analogue reaction with 5′-AMP. A new di-alkylated 5′-UMP adduct is obtained from the 1:1 mixture of both complementary nucleotides, in addition to the nucleo-lipids found in separate systems. Time-resolved DLS measurements reveal very different kinetic processes for aggregates’ formation when 5′-UMP, 5′-AMP, or their equimolar combination are used in the reaction mixture. This system as a whole represents a potential experimental model where the effect of both intermolecular interactions and self-association processes can be investigated by tuning the type of nucleobases in the reaction mixtures.

Introduction The self-assembly of amphiphilic molecules into nanostructured supramolecular architectures is a central feature in the chemistry of life.1 Concerning the well-known nucleotide molecules, structural units of nucleic acids, the physical-chemical behavior of their correspondent lipophilic derivatives represents, today, a well-defined research area with promising margin of progress. One of the general aims is to confine nucleic bases at the surfaces of amphiphilic aggregates, to investigate the interplay between the cooperative self-assembly properties2 typical of amphiphilic molecules and specific molecular recognition features characterizing biological macromolecules.3 Yet, a relevant issue concerns the study of mechanisms of formation of alkylated DNA bases4 as a consequence of the action of highly electrophilic molecules such as, e.g., aliphatic and aromatic epoxides, which are highly reactive organic compounds with mutagenic and in some cases carcinogenic properties.5 Some authors have also suggested that alkylated nucleobases could have provided additional functional groups in the primitive versions of ribozymes, especially for the construction of * To whom correspondence should be addressed: Consorzio per lo sviluppo dei Sistemi a Grande Interfase (CSGI) c/o Universita` del Molise (DISTAAM), v. De Sanctis, I-86100 Campobasso, Italy. Phone: +390874404632; fax: +39-0874404652; e-mail: [email protected]. † DISTAAM. ‡ ICCOM. § IRSA. | Dipartimento di Chimica, Universita ` di Bari. (1) Fuhrhop, J. H.; Koning, J. In Membranes and Molecular Assemblies: the Synkinetic Approach; Fraser, J., Ed.; The Royal Society of Chemistry: Cambridge, U.K., 1994. (2) Evans, D. F.; Wennerstro¨m, H. in The Colloidal Domain Where Physics, Biology, and Technology Meet; VCH Publishers, Inc.: Weinheim, 1994. (3) Lehn, J. M. In Supramolecular Chemistry; VCH: Weinheim, 1995. (4) Iwai, S. Nucleosides, Nucleotides Nucleic Acids 2006, 25, 561-582. (5) Koskinen, M.; Plna´, K. Chem. Biol. Interact. 2000, 129, 209-229 and references therein.

hydrophobic binding pockets.6 These hydrophobically modified nucleotide analogues can be obtained through a widespread number of synthetic routes, depending on the nature of the alkyl/ acyl group(s) and sites of attachment on nucleotide molecular structures, being a phosphate group,7 sugar moiety,8 or within the heterocyclic bases themselves.9 Baglioni and co-workers, for example, have found the interfacial properties of nucleoside-based phosphocholine amphiphiles to be modulated by intermolecular base-base interactions between surfactant molecules in a variety of self-assembled structures.10-14 Wang et al.15 followed an alternative approach in which nucleotides have not been covalently bonded to lipophilic chains but confined electrostatically, like counterions, through their phosphate residues at the surface of cationic gemini (bisquaternary ammonium) surfactants. Shimizu et al.16 produced nucleotide bolamphiphiles capable of water gelling ability through the spontaneous formation of a fibrous network. In this paper, we report on reactions performed in nanostructured environments, specifically microemulsions, between comple(6) Levy, M.; Miller, S. L. J. Mol. EVol. 1999, 48, 631-637. (7) Zandomeneghi, G.; Luisi, P. L.; Mannina, L.; Segre, A. HelV. Chim. Acta 2001, 84, 3710-3725. (8) Moreau, L.; Barthe´le´my, P.; El Maataoui, M.; Grinstaff, M. W. J. Am. Chem. Soc. 2004, 126, 7533-7539. (9) Nowick, J. S.; Chen, J. S.; Noronha, G. J. Am. Chem. Soc. 1993, 115, 7636-7644. (10) Berti, D.; Pini, F.; Teixeira, J.; Baglioni, P. J. Phys. Chem. B 1999, 103, 1738-1745. (11) Baldelli Bombelli, F.; Berti, D.; Keiderling, U.; Baglioni, P. J. Phys. Chem. B 2002, 106, 11613-11621. (12) Baglioni, P.; Berti, D. Curr. Opin. Colloid Interface Sci. 2003, 8, 55-61. (13) Fortini, M.; Berti, D.; Baglioni, P.; Ninham, B. Curr. Opin. Colloid Interface Sci. 2004, 9, 168. (14) Milani, S.; Baldelli Bombelli, F.; Berti, D.; Hauss, T.; Dante, S.; Baglioni, P. Biophys. J. 2006, 90, 1260-1269. (15) Wang, Y.; Desbat, B.; Manet, S.; Aime`, C.; Labrot, T.; Oda, R. J. Colloid Interface Sci. 2005, 283, 555. (16) Shimizu, T.; Iwaura, R.; Masuda, M.; Hanada, T.; Yase, K. J. Am. Chem. Soc. 2001, 123, 5947-5955.

10.1021/la702580j CCC: $40.75 © 2008 American Chemical Society Published on Web 02/01/2008

Catanionic Systems

Langmuir, Vol. 24, No. 6, 2008 2349

QqTOF-MS, which is followed by optical microscopy investigations and time-resolved turbidity measurements obtained by the DLS technique. Finally, several concluding remarks are outlined. A complete description about materials and methods has been included in the Supporting Information. Figure 1. Schematic representation of the transition from an initial state constituted by positively charged CTAB spherical micelles, incubated with a hydrophobic epoxide (R ) -C10H21) and nucleotides, to a final state where the synthesis of alkylated anionic analogues of nucleobases, due to an irreversible ring-opening reaction, trigger the formation of a planar mixed catanionic interface.

mentary nucleotides and a hydrophobic epoxide, with emphasis on intrinsic base-base recognition properties in the implementation of abiotic selective syntheses17 of novel amphiphilic molecules. Suitable nucleophiles are known to open the epoxy ring leading to the formation of an alcohol-nucleophile adduct.18 Yet, charged interfaces have been reported to favor the ring opening reaction of 1,2-epoxyoctane19 and 1,2-epoxydodecane.20 Therefore, to overcome the mutual immiscibility between the apolar epoxide and aqueous solutions of nucleotides, we performed the reaction described above in the presence of a micellar interface (CTAB) to produce alkylated nucleobases, hereafter called nucleo-lipids. The spontaneously formed amphiphilic monomers should incorporate into the preexistent cationic surfactant film, owing to the favorable electrostatic interactions with the oppositely charged headgroups (see Figure 1). As a consequence, self-assembled catanionic supramolecular structures, characterized by more planar interface curvature such as, e.g., multilamellar mixed vesicles, (MLMVs), are expected to appear in the reaction solutions. Therefore, we propose this system to find out if molecular recognition and self-association processes may cooperate to the formation of complex supramolecular structures triggered by smart reaction products, which are generated in a starting microemulsion system. According to the previous working hypotheses, we have designed our experiments (1) to identify, through high-resolution mass spectrometry, various kinds of nucleo-lipids yielded in the reaction mixtures, and (2) to investigate, through time-resolved turbidity measurements, the rate of formation of micrometer-sized catanionic aggregated, whose size and density per unit volume can be directly related to the kinetics of formation of nucleolipid monomers. We anticipate that the onset of turbidity, linked to the appearance of these large aggregates, strongly depends on the type of nucleobase present in the starting solutions, both as single type and in equimolar combination. In principle, the presence of different nucleophilic centers in the molecular structures of both nucleotides may give rise to mixtures of regioisomers, due to a specific nucleophilic attack of phosphate group, hydroxyl groups of ribose, or O- and N-positions of heterocyclic bases. On the other hand, aliphatic alkyl epoxides should react preferentially at sites with high relative nucleophilicity such as, e.g., the ring nitrogen positions.21,22 The remainder of the paper is organized as follows. First, we illustrate the details about experimental systems under investigation; then, we illustrate the characterization of nucleo-lipids through LC(17) Monnard, P.-A. Cell. Mol. Life Sci. 2005, 62, 520-534. (18) Pasc-Banu, A.; Blanzat, M.; Belloni, M.; Perez, E.; Mingotaud, C.; RicoLattes, I.; Labrot, T.; Oda, R. J. Florine Chem. 2005, 126, 33-38. (19) Anderson, K.; Kizling, J.; Holmberg, K; Bystro¨om, S. Colloids Surf., A: Physicochem. Eng. Aspects 1998, 144, 259-266. (20) Conde-Frieboes, K.; Blochliger, E. Biosystems 2001, 61, 109. (21) Mayr, H.; Kempf, B.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66-77. (22) Brotzel, F.; Kempf, B.; Singe, T.; Zipse, H.; Mayr, H. Chem.sEur. J. 2007, 13, 336-345.

Preparation of CTAB/DE/NMP Systems We have decided to fix the composition of the starting mixtures throughout our experimental investigation as well as the order of addition of compounds. First, weighted amounts of CTAB and disodium salts of 5′-AMP, 5′-UMP, or their 1:1 molar combination were dissolved in 5 mL distilled water to obtain [CTAB] ) 26 mM (1 wt %) and [NMP]tot. ) 14 mM, where [NMP]tot. ) [5′-AMP], [5′-UMP], or [5′-AMP] + [ 5′-AMP] (1:1). Then, 40 µL of dodecyl epoxide (DE) was added to the water solution by Hamilton syringe and mixed for ∼1 h with a magnetic stirrer until the solutions turned optically transparent or slightly translucent. The final overall [DE] was 37 mM, corresponding to the following molar ratios DE/NMPtot. ) 2.6 and DE/CTAB ) 1.4. The former was chosen to achieve the condition of limiting agents for nucleotides involved in the bimolecular kinetic process, whereas the latter was the DE solubility limit in a one-phase oil-in-water microemulsion formulated with 1 wt % CTAB. Actually, the composition of starting mixtures has been experimentally determined to fit the criteria of a reproducible, thermodynamically stable initial state, whose optical properties can be fixed by the dimensions of constituent particlessin this case, spherical submicrometer-sized monodisperse micelles. Indeed, the choice of optically transparent microemulsions as starting solutions was very useful to detect unambiguously the onset of turbidity through dynamic light scattering technique (vide infra). We defined system A, system U, and system AU as the reaction mixtures incubated, respectively, with 5′-AMP, 5′-UMP, and 5′-AMP + 5′-UMP (1:1). Reaction solutions included a trace of a microbial inhibitor. Measured values in both starting mixtures and final dispersions fell in the neutral-alkaline range (pH ≈ 7-8). All preparations were continuously sheared with a magnetic stirrer over 10 days at steady speed of 350 rpm and controlled temperature of 25 °C. After 1 month of ripening at rest, solutions were subjected to LC-QqTOF-MS analyses and optical microscopy investigations (see Supporting Information for experimental details). Parallel reactions have been carried out as reference systems to follow up the formation of nucleo-lipid in the absence of CTAB. For this purpose, we modified a simple synthetic protocol by Koskinen et al.,23 used to alkylate DNA bases with aromatic epoxides, by substituting the styrene 7,8-oxide with DE. For our references, we chose the same molar ratio of DE/NMP used in presence of CTAB, (NMP ) AMP or UMP). Reactants were dispersed in 50 mM Tris-HCl (pH 7.5) and 30% methanol and stirred at rt for 10 days. Then, as before, mixtures have been aged at rest for a month. Finally, the excess DE was extracted twice with ethyl acetate (1 vol), and the mixtures were first evaporated and then dissolved in water/MeOH (4:1 v/v) to perform LC-Qq-TOFMS-MS analyses (see Supporting Information for details).

Results and Discussion LC-QqTOF-MS-MS identification of Nucleo-Lipids. Several hydrophobic derivatives of 5′-AMP have been identified as reaction products in system A. In the HPLC separation of the reaction mixture, a prominent peak was detected together with (23) Koskinen, M.; Schweda, E. K. H.; Hemminki, K. J. Chem. Soc., Perkin Trans. 2 1999, 129, 2441-2445.

2350 Langmuir, Vol. 24, No. 6, 2008

Figure 2. Top panel: LC-QqTOF-MS-MS chromatogram (negative ions) of alkylated 5′-AMP adducts identified in system A. Middle panel: LC-QqTOF-MS-MS product ion spectrum of [M - H]- at m/z 530.2 of 5′-AMP adduct alkylated at the phosphate group. Bottom panel: LC-QqTOF-MS-MS product ion spectrum of [M - H]- at m/z 530.2 of 5′-AMP adduct alkylated at the N- site of purine ring.

two secondary peaks at retention times of 18.2, 19.7, and 20.1 min, respectively (top Figure 2). QqTOF-MS-MS spectra of those peaks showed the same [M - H]- ion correspondent to monohydroxydodecyl derivative of 5′-AMP (C22H37N5O8P-; m/z 530). From QqTOF-MS-MS spectra, we interpreted the product (a) as resulting from the alkylation at level of phosphate group (see middle panel of Figure 2) characterized by fragment ions at m/z 281 (loss of adenosine group) and m/z 395 (depurinated product). Other components coming from peaks (b) and (c) gave fragmentation patterns correspondent to alkylation at level of one of the accessible N-sites on the purine ring (see a typical spectrum in bottom panel of Figure 2). Indeed, an alkylated heterocyclic adenine ring has been attributed to the presence of m/z 318, although it was not possible from QqTOF-MS-MS spectra to specify the site of attachment of the alkylic chain. However, the above-reported fragmentations were consistent with accurate mass measurements of fragment ions (see Table S1 in Supporting Information). Those results were qualitatively reproduced in the reference system, based on the reaction between 5′-AMP and DE in the absence of CTAB. Thus, a preliminary conclusion was that the addition of DE to 5′-AMP led to the formation of a mixture of monochained derivatives, characterized by two sites of alkylation at level of phosphate residue and purine ring, respectively. A completely different situation was found in the analogous system U. Indeed, a single component was identified for the monohydroxydodecyl derivative of 5′-UMP, ([M - H]- at m/z 507, C21H36N2O10P-), characterized by a retention time of 14.8 min (red chromatogram in top panel of Figure 3) whose product ion spectrum revealed the fragment at m/z 295, corresponding to the

Angelico et al.

Figure 3. Top panel: Overlaid LC-QqTOF-MS-MS chromatograms (negative ions) showing two alkylated 5′-UMP adducts identified in the reference system (blue trace) and a single alkylated 5′-UMP adduct identified in system U (red trace). Middle panel: LC-QqTOFMS-MS product ion spectrum of [M - H]- adduct at m/z 507.2 of 5′-UMP alkylated at the N-site of pyrimidine ring. Bottom panel: LC-QqTOF-MS-MS product ion spectrum of [M - H]- adduct at m/z 507.2 of 5′-UMP alkylated at the phosphate group.

alkylated uracil base after the cleavage of the glycosidic bond (middle panel of Figure 3). The previous result was not reproduced for the analogous reference system, where 5′-UMP reacted with DE in the absence of CTAB micelles. Rather, nonselective alkylation was recorded, giving rise to a pair of 5′-UMP adducts (blue chromatogram in top panel of Figure 3), with retention times of 14.8 and 16.2 min, respectively, both having [M - H]at m/z 507. In fact, QqTOF-MS-MS spectrum of the component at t ) 16.2 min showed a fragmentation pattern originating from the loss of uridine (see the hydroxyldodecylphosphate ion at m/z 281 in the spectrum of Figure 3, bottom panel), while for the second peak (t ) 14.8 min), a molecular structure with alkylation on the pyrimidine ring was assigned. Again, the above-reported fragmentations were consistent with accurate mass measurements of fragment ions (see Table S1 in Supporting Information). So far, we have shown that the starting compositions set up for both single-base systems, i.e., system A and system U, triggered the evolution at the molecular level of the reagents (nucleotides and epoxide) to new amphiphilic molecules having nucleobases in their polar groups. However, the site of (mono-)alkylation seemed to be more affected by the presence of cationic micelles for chemical reactions with 5′-UMP than the analogous solutions incubated with 5′-AMP. Yet, a new component was detected at higher m/z ratios in system U. Indeed, by following different chromatographic conditions, (described in Supporting Information), HPLC separation gave a peak at t ) 12.2 min assigned to m/z 693 in positive ion mode (see red chromatogram in top panel of Figure 4).

Catanionic Systems

Figure 4. Top panel: Overlaid LC-QqTOF-MS-MS chromatograms (positive ions) showing a single bi-alkylated 5′-UMP adduct identified in system U (red trace) and two bi-alkylated 5′-UMP products detected in system AU (blue trace). Middle panel: LC-QqTOF-MS-MS spectrum of [M + H]+ ion at m/z 693.4 for 5′-UMP alkylated at both the phosphate group and the pyrimidine ring. Bottom panel: LCQqTOF-MS-MS spectrum of [M + H]+ ion at m/z 693.4 showing fragment ions compatible with mono-alkylation at level of the uracil ring. Fragments including the other site of alkylation were lacking in those conditions (positive ion) due to the presence of a free ribose 5′-phosphate negatively charged group.

The QqTOF-MS-MS spectrum was consistent with a dialkylation product, showing two fragment ions at m/z 279 and 379, for epoxy ring-opening reactions occurring at the level of both the nitrogenous base and the phosphate group, respectively (see middle panel in Figure 4). The top panel of Figure 4 also shows the chromatogram (blue trace) measured in the same conditions for system AU. Here, a different component (b) is clearly resolved (t ) 12.7 min) from the identical dialkylation adduct (a) found in system U. However, from the QqTOFMS-MS spectrum (bottom panel in Figure 4) one of the sites of alkylation was straightforwardly assigned at the level of the uracil base while fragments related the other site of alkylation were lacking in positive ion mode. This, in fact, was consistent with the proposed structure having a nonalkylated 5′-phosphate group. The above-reported fragments were also confirmed by accurate mass measurements. Finally, by switching in negative ion mode, we deduced the second site of attachment, occurring at the ribose moiety of the 5′-UMP molecule. Indeed, the chromatogram obtained from system AU (blue trace in top panel of Figure 5) showed the component (b) whose fragmentation pattern (bottom panel in Figure 5) revealed ions identified as hydroxyl-dodecyl derivatives of ribose ring 5′phosphate (mixture of isomers). It is worth noting that in negative ion mode the fragments related to the other site of alkylation were present as well as the ion O3P- at m/z 79, suggesting the absence of alkylation at phosphate group.

Langmuir, Vol. 24, No. 6, 2008 2351

Figure 5. Top panel: Overlaid LC/QqTOF-MS-MS chromatograms (negative ions) showing a single bi-alkylated 5′-UMP adduct identified in system U (red trace) and two bi-alkylated 5′-UMP products found in system AU (blue trace). Middle panel: LC/QqTOFMS-MS spectrum of [M - H]- ion at m/z 691.4 for 5′-UMP alkylated at both the phosphate group and the pyrimidine ring. Bottom panel: LC/QqTOF-MS-MS spectrum of [M - H]- ion at m/z 691.4 showing charged fragments consistent with alkylation at level of sugar moiety. Fragments including the other site of alkylation were lacking in those conditions.

Molecular structure underlying peak (a) was also verified (top and middle panels in Figure 5). Moreover, identical products detected in system A have been also identified in the mixed system AU. Although no further analysis has been performed to assign the positional isomerism of hydroxydodecyl chains, the preferred site of nucleophilic attack on DE molecules should be primarily at the less-substituted and more accessible β-carbon. This hypothesis is supported by the fact that DE must orient into the hydrophobic core of CTAB micelles with the slightly hydrophilic three-membered ring facing at the Stern layer, showing a more accessible β position than the alternative site. Quantitative separation and purification analytical procedures of novel amphiphilic compounds were found intricate due to the strong electrostatic interaction between cationic CTAB surfactant and anionic nucleolipid monomers. The results obtained from MS data indicate that the presence of a cationic interface (CTAB micelles) favors the alkylation at nitrogen ring of 5′-UMP, giving rise to a single molecular species for its mono-alkylated derivative. However, a pair of di-chained 5′-UMP regioisomers are produced if the complementary 5′-AMP is present in the starting reaction mixture (system AU). In other words, the copresence of both nucleotides seems to be the preferred combination to enhance the potential chemical activity of the 5′-UMP nucleobase. Optical Microscopy of MLMV Liquid Dispersions. The onset of a diffuse turbidity in freshly made samples left under stirring at room temperature for the whole reaction course was observed after ∼1 h, 25 h, and 6 days, in systems U, AU and A, respectively. After 10 days reaction under stirring, the alkylated

2352 Langmuir, Vol. 24, No. 6, 2008

Figure 6. Microphotographs through cross polarizers (enlargement 100×) of 1-month aged samples taken from systems A, U, and AU left at rest after 10-day reaction under stirring. System A. Upper left: cluster of birefringent spherulites with the typical Maltese cross patterns, indicating the presence of large multilayered vesicles. Upper right: same optical field but observed under normal light showing a coarse vesicular dispersion. System U. Bottom left: ensemble of multilayered structures evidenced by Maltese crosses, also found in the mixed system AU (bottom right).

derivative of nucleobase monomers accumulated in the solutions, transforming the initial transparent dispersions of CTAB micelles into milky suspensions. By optical analysis through microscopy under polarized light, a heterogeneous collection of LC patterns was found. Among them, spherulites showing Maltese crosses have been recognized as the most representative LC mesoscopic structures (Figure 6), which indicate the presence of an onionlike arrangement of lamellae forming multilamellar spherical vesicles. These structures have been found very often in mixed catanionic surfactant systems.24 However, in some cases, other birefringent textures were recognized as lamellar and cholesteric phases, indicating that helical LC arrangements of these hydrophobically modified nucleotides may occur similarly to analogous supramolecular structures caused by stacking of free monomers.25 MLMV Formation in the Absence of Mechanical Forces. To investigate the possible influence of shear stress forces onto the final state of MLMV phase produced by the chemical reaction, we set up the same preparations but avoided stirring the samples over the whole reaction process. Thus, after having stirred the freshly made mixtures for 1 h, we stopped shearing the samples, and left them to rest at 25 °C. We found that the final state was history-independent at equilibrium. Indeed, over 6 months aging, the MLMV suspensions collapsed very slowly to dense heterogeneous layers that had a lower density than water and showed up creaming. Interestingly, Hoffmann et al.26 reported experimental evidence of spontaneous vesicle formation in a dilute catanionic surfactant system where one of the components was produced by a chemical reaction and not by mixing of the components. On the other hand, they found that, in concentrated solutions of the same system, nonspontaneous vesicles were produced in sheared samples, whereas a thermodynamically stable LR lamellar phase was observed without application of shear forces. We have observed a similar behavior, in the absence of shear forces, in our system as well. Indeed, (24) Tondre, C.; Caillet, C. AdV. Colloid Interface Sci. 2001, 93, 115-134. (25) Wong, A.; Ida, I.; Spindler, S.; Wu, G. J. Am. Chem. Soc. 2005, 127, 6990-6998. (26) Hao, J.; Yuan, Z.; Liu, W.; Hoffmann, H. J. Phys. Chem. B 2004, 108, 5105-5112.

Angelico et al.

preparations characterized by higher concentrations of reactants evolved to a stable lamellar phase in equilibrium with an isotropic solution.27 Dynamic Light Scattering Investigation. If, during the course of reaction, hydrophobically modified anionic nucleotide monomers accumulate within the cationic CTAB micellar interface, which by itself prefers a positive spherical curvature,2 a transition to a bilayer arrangement would be expected for the final catanionic mixed surfactant film.28 This was reflected in the appearance of the bluish color in the samples, a typical manifestation of the Tyndall effect due to large scatterers in solutions. The rate of conversion of spherical, nanosized, oil-in-water monodisperse micelles into large MLMVs has been conveniently followed by measuring the turbidity of reaction mixtures through dynamic light scattering (DLS). This has been accomplished by following the time evolution, over time scales of about 10 days depending on samples, of hydrodynamic Z-average diameters, intensity of scattered light at fixed 90° angle, and intensity-weighted size distributions of aggregates produced in system A, system U, and system AU, respectively. First, we shall describe DLS results obtained from systems U and A separately; then, a comparison will be made taking into account analogous results collected in the mixed system AU (for experimental details, see Supporting Information). System U. Time courses of measured scattered intensity at fixed scattering angle of 90° and derived Z-average values for hydrodynamic aggregate diameter (in µm) are shown in Figure 7A. The onset of the process coincides with the sharp rise in sample turbidity, about an hour after addition of DE to CTAB + 5′-UMP aqueous solution. Indeed, in less than 24 h a second broad peak appears at the large-particle side of the intensityweighted size distribution (see diameter distribution in nanometers in inset of Figure 7A). This rapid process of formation of large aggregates can be empirically described by a first-order kinetic process. Around t* ≈ 90 h, size goes through a maximum, corresponding to a mean diameter of about 1.5 µm. If the observed monotonic increase of Z-average dimensions of mixed liposomes was directly linked to an increase of alkylated 5′-UMP/CTAB molar ratio from zero to close to charge neutrality, t* would roughly indicate that this condition has been satisfied. Assuming the monohydroxydodecyl derivative of 5′-UMP (alkylated at level nitrogenous ring) as the most abundant anionic nucleo-lipid formed in system U (see Figure 3)sat least in the early stage of the overall processsa conversion of about 84% of initial [5′-UMP] ) 14 mM, i.e., 12 mM, into corresponding mono-alkylated monomers can be estimated from the pseudofirst-order kinetic relationship UMP% ) 1 - e-Kt/, with UMP% being the percentage of converted nucleotide, t* ≈ 90 h, and KU ) 0.02 h-1 (see Figure 7A). Since in the neutral-alkaline reaction solution, alkylated 5′UMP monomers have the 5′-phosphate groups mostly in the dianionic state, a 1:2 anionic lipid/CTAB molar ratio corresponds to electroneutrality of the mixed surfactant film. In fact, as stated above, we get 12 mM for this limit condition, which is very close to [CTAB]/2 ) 13 mM. Then, a partial dissolution process, very common in vesicle-micelle transitions,29 is observed as the (27) Three reaction mixtures were prepared using the concentrations: [CTAB] ) 137 mM (5 wt%); [NMP]tot. ) 47 mM and [DE] ) 82 mM. The initial turbid solutions stirred for 1 h were left at rest for six months at 25 °C. Then, the following multiphase equilibria were experimentally found: a low-viscosity isotropic gel phase (system AU); isotropic solution in equilibrium with upper liquid-crystalline phase (system A); a three-phase body constituted by liquidcrystalline upper phase, isotropic solution (middle phase) and dense isotropic gel as lower phase (system U). (28) Kaler, E. W.; Murthy, A. K.; Rodriguez, B. E.; Zasadzinski, J. A. N. Science 1989, 245, 1371-1374.

Catanionic Systems

Langmuir, Vol. 24, No. 6, 2008 2353

System A. Upon replacing 5′-UMP with 5′-AMP, there are notable changes. Scattering intensity and Z-average diameters have been plotted in Figure 7B over the course of a typical experiment. The size distribution is monodispersed (D ≈ 10 nm from oil-in-water micelles) over a long induction period (∼6 days), followed by a pronounced exponential growth phase of aggregate sizes. This sharp increment follows a first-order kinetic process whose characteristic constant is KA ) 0.03 h-1, (see Figure 7B). The presence of a discontinuity in the time dependence of scattering intensity localizes the end of lag phase (150 h) and, consequently, the onset of macro-aggregate formation. The huge difference between the kinetics of aggregates growing with 5′UMP and 5′-AMP nucleotides may be explained in terms of a higher energy barrier to the ring-opening reaction of the latter than the former. System AU. Here, the kinetics of the MLMV growth process can be described as a sigmoidal curve (see Figure 7C) defined by an initial lag time of about 25 h, a subsequent exponential growth phase in which effective hydrodynamic diameter increases from 10 nm up to ∼3 µm, and a final equilibrium phase, characterized by stable liposomial dispersion. The correlation data for samples with bimodal size distributions were force-fitted to double-exponential relaxations with acceptable statistical accuracy (see Figure S1 in Supporting Information). However, data collected for t > 150 h were affected by higher errors in the fitting procedures, probably due to multiple scattering effects. Measurements were plotted as a function of time and fitted to a curve described by the following equation:

〈D〉 ) Df + at -

Figure 7. Variation of Z-average diameters in µm (left axes - blue diamonds) and intensity at 90° scattering angle in Giga cps (right axes - red circles) as a function of reaction time for the investigated systems. A: system U. Dashed line is a first-order nonlinear leastsquare fit, KU ) 0.02 h-1. Inset: bimodal size distribution at reaction time t ) 23.3 h. B: system A. The onset of sample turbidity starts after a long induction period of about t ≈ 150 h. Dashed line is a first-order nonlinear least-square fit, KA ) 0.03 h-1. Inset: bimodal size distribution at reaction time t ) 230.6 h. C: system AU. Increment of aggregate dimensions vs time can be fitted through a sigmoid function (eq 1, see text) starting after a lag-time of 25 h, followed by a rapid rise and subsequent final drift (slow coarsening process). Inset: size distribution by intensity at t ) 44 h.

reaction yield moves toward completeness giving rise to a slight reduction of size in the sub-micrometer scale while the scattering intensity increases further with reaction time. (29) Cohen, D. E.; Angelico, R.; Carey, M. C. J. Lipid Res. 1990, 31, 55-70.

(

Df - Di

)

(t - to)β 1+e τ

(1)

where 〈D〉 is the average diameter, to is the time to 50% of maximal size, and β the steepness of the function. The final drift after the growth phase has ended from initial value Di is described by Df + at, where Df and a are final average diameter and slope of drift, respectively. The apparent first-order rate constant, KAU for the growth of liposomal aggregates is calculated as 1/τ. Relevant parameters derived from the fit are to ) 64 h, KAU ) 0.12 h-1, and β ) 0.05. Interestingly, time dependence of intensity of scattered light, which is extremely sensitive to very large particles,30 shows a discontinuity close to t* ) 90 h corresponding to the time of maximum aggregate size observed in system U (see Figure 7A). All together, DLS data clearly show very different growth rates of aggregate dimensions in separate systems U and A, characterized by a long lag time for system A compared to the that of analogous system U. This evidence not only suggests a higher reactivity of 5′-UMP than 5′-AMP toward the alkylation agent but also indicates that chemical activities of both may be mutually dependent in the mixed system AU. Lag Time Dependence on Nucleobase Concentration. If we define tl as the lag time at which the onset of size increment is experimentally measured, we note that tlA . tlAU > tlU. This means that the activation energy for the rate-limiting step of the whole kinetic process is much lower when 5′-UMP interacts directly with DE than 5′-AMP. Thus, in the mixed system AU, where both [5′-AMP] and [5′-UMP] are one-half the respective concentrations in the parent systems A and U, we would expect a lag-time dependence on the initial quantity of 5′-UMP. To verify this effect, we have experimentally followed the time dependence of the hydrodynamic diameter in the two reaction (30) Pecora, R. In Dynamic Light Scattering; Plenum Pubblishing Corp.: New York, 1985.

2354 Langmuir, Vol. 24, No. 6, 2008

Angelico et al.

Figure 8. Time evolution of Z-average diameters of aggregates produced with [CTAB] ) 26 mM and [5′-UMP] ) 7 mM + [Na2HPO4] ) 7 mM (blue diamonds); [CTAB] ) 26 mM and [5′-UMP] ) 7 mM (hollow diamonds); system AU (red symbols). Long and medium dashed lines represent guides for eyes.

mixtures with the usual oil-in-water micellar solution formulated with [CTAB] ) 26 mM, composed by (1) [5′-UMP] ) 7 mM and (2) [5′-UMP] ) 7 mM + [Na2HPO4] ) 7 mM, respectively. In the latter mixture, the addition of disodium salt of the inorganic phosphate in equimolar quantity reproduces the ionic strength of system AU, with the evident difference of the lack of the complementary adenosine residue. The resulting profiles have been compared to the analogous values reported in Figure 7C and plotted together in a new graph (Figure 8). Interestingly, the induction period of 25 h found in system AU is reproduced in the solution where 5′-AMP has been substituted by phosphate ions, while it is shifted to longer times if [5′-UMP] is halved in the absence of both 5′-AMP and inorganic phosphate. A relevant feature highlighted in Figure 8 is represented by the final Z-average dimensions, which show the tendency to level off toward a common threshold value of about 0.25 µm. The experimental determination of a different trend observed when the complementary nucleotide 5′-AMP is added to 5′UMP in equimolar ratio (system AU), may be interpreted in terms of a mechanism operating at the molecular level and acting cooperatively at larger length scales. Indeed, DLS data of Figure 8 show that (1) the early stage of the alkylation process in system AU can be predominantly due to 5′-UMP ions; and (2) Z-average hydrodynamic diameters in system AU increase well beyond the maximum dimension compatible with aggregates obtained for [5′-UMP] ) 7 mM. On the contrary, the very long induction period (∼150 h) recorded prior to the appearance in solution of A-type aggregates (Figure 7B) is not varied if [5′-AMP] is doubled in the initial reaction mixture (data not shown). The mechanism underlying the appearance of aggregates in the mixed system AU may be too complex to be described in detail without further investigation. Thus, the whole process may be represented as the sum of two consecutive events: a rapid production of alkylated 5′-UMP monomers leading to the formation of a first “generation” of mixed CTAB-liposomes, followed by a growth phase due to the synthesis of alkylated 5′-AMP monomers. Those events are not an additive combination of both systems A and U separately. Indeed, a new di-alkylated 5′-UMP adduct has been identified in system AU (see Figure 5, component b) but not found in system U. Moreover, mono-alkylated adducts of the complementary base 5′-AMP are produced in system AU in a much

Figure 9. Molecular structures identified for mono- and disubstituted alkylated adducts of 5′-AMP and 5′-UMP, produced through ring-opening reaction of DE incorporated in aqueous solution of CTAB cationic micelles. The residue R- is R-(or β-)hydroxydodecyl hydrophobic moiety.

shorter reaction time than system A, although no other new components have been identified. The combination of MS and DLS data underscores that in this system the production of amphiphilic derivatives of 5′-UMP and 5′-AMP, together with their large-scale self-association in mesoscopic structures (MLMVs), is clearly intertwined with a local molecular recognition.

Concluding Remarks To sum up, in Figure 9, molecular structures of relevant products identified through mass spectrometry have been collected. Due to the asymmetry of R-carbon, the reaction generates a diastereomeric pair of products formed through nucleophilic attack on either R and β carbons. The striking result obtained by the present investigation concerns the binding positions of the hydroxyl-dodecyl apolar chains which are found, respectively, at level of nitrogenous ring and phosphate residue for 5′-UMP when the reaction is carried out in system U, i.e., in the absence of the complementary 5′AMP nucleotide. However, a new di-alkylated adduct is produced in system AU where a chain is attached at the sugar ring of the 5′-UMP molecule. This result may be ascribed by an intermolecular effect occurring between complementary nucleobases. This feature has not been observed for the other complementary nucleotide, where mono-alkylation has been ascertained to occur either at the phosphate group and the purine ring. A more favored orientation of the uracil ring in the micellar Stern layer, leading to partial micellar incorporation of the uridine moiety, may be most probably a very significant factor. This dynamical configuration could be emphasized by the presence of 5′-AMP, whereas the inverse situation does not occur. In favor of this hypothesis, it is worth remarking that the six-membered ring of the uracil base is associated with a dipole moment value of 4.37

Catanionic Systems

D, which is almost twice the value for adenine.31 Mono- and di-alkylated nucleotides accumulate in the reaction media giving rise to final aqueous dispersions of MLMVs, independent of the mechanical history (in the low concentration regime). To investigate the effects on the morphological changes of the vesicular system induced by the chemical coupling of the precursors in the new amphiphilic monomers, a study of the kinetic process of vesicle formation has been carried out by monitoring the time progress of the increase of scattered light intensity at 90°. When both complementary nucleotides react with the lipophilic epoxide in the same mixture, the observed sigmoidal kinetic increment of aggregate sizes can be correlated to a cooperative process, indicating an underlying molecular recognition mechanism. In comparison, systems A and U separately show very different kinetic trends. A subsequent implementation of this study is currently addressed in our lab to reveal the role played by the complementary nucleotides in the stability of the related supramolecular aggregates from one side and, from the other side, to quantitatively ascertain the reaction yields in all systems A, U, and AU as well as their dependence on the concentrations of reactants in the starting reaction mixtures. This will be the subject of a forthcoming paper. Yet, since it is known that epoxides are formed biogenically by the epoxidation of alkenes in reactions catalyzed by a number (31) Sponer, J.; Leszczynski, J.; Hobza, P. Biopolymers 2001, 61, 3.

Langmuir, Vol. 24, No. 6, 2008 2355

of enzymes with monooxygenase activity,32 the study of the formation of these hydrophobically modified nucleotides in the presence of an interface can provide a potential tool to investigate the membranogenic effect of alkylated bases as potential destabilizing agents of cell membranes. As a final remark, we think that the present system can offer a potential tool through which to test the substantiation of the role charged colloidal surfaces might have played in prebiotic macromolecular evolution.33,34 Acknowledgment. R.A. acknowledges Prof. G. Palazzo and Dr. F. Lopez for fruitful discussions and suggestions. The authors wish to thank Mr. Vito Locaputo for his support in running LC-QqTOF-MS analyses. Consorzio per lo sviluppo dei Sistemi a Grande Interfase (CSGI), MUR-PRIN 2006 and CNR-FUSINT 2007 are also acknowledged for financial support. Supporting Information Available: Detailed information about Materials and Methods, integrated by exposition of LC-QqTOF-MS analytical procedures and DLS measurements used for the present study. This material is available free of charge via the Internet at http://pubs.acs.org. LA702580J (32) Ensign, S. A.; Allen, J. R. Annu. ReV. Biochem. 2003, 72, 55-76. (33) Deamer, D. W. Microbiol. Mol. Biol. ReV. 1997, 61, 239-261. (34) Mascolo, G.; Giustini, M.; Luisi, P. L.; Lang, J. J. Colloid Interface Sci. 1990, 140, 401-407.